Annual Report Fission Time Projection Chamber Projectoregonstate.edu/.../TPC/2009_Annual_Report_Final-Draft.pdf · 2011-07-21 · Annual Report Fission Time Projection Chamber Project

15 November 2009

Annual Report

Fission Time Projection Chamber Project Fiscal Year 2009

Nuclear Energy Research Initiative

The Time Projection Chamber is an innovative approach to carry out precision fission measurements at the Los Alamos Neutron Science Center. This novel 4! detector system will provide

unprecedented data about the fission process. Shown here, in front of a TPC CAD drawing, is the DOE Nuclear Energy Fuel Cycle R&D Excellence Award that NIFFTE received at the FC R&D Annual Review Meeting in Albuquerque, New Mexico.

ii Fission Time Projection Chamber Annual Report Fiscal Year 2009

Blank Page

iii Fission Time Projection Chamber Annual Report Fiscal Year 2009

NERI-c Annual Technical Report

Fission Time Projection Chamber Project

Fiscal Year 2009

Approvals

Date_____10/25/09

Mike Heffner, Principal Investigator

Date_____10/25/09

Nolan Hertel, Principal Investigator

Date_____10/25/09

Tony Hill, Principal Investigator

iv Fission Time Projection Chamber Annual Report Fiscal Year 2009

Blank Page

v Fission Time Projection Chamber Annual Report Fiscal Year 2009

Table of Contents

Major Contributors ........................................................................................................... vii!

2009 Submitted Papers, Reports and Summaries......................................................... vii!

2009 Presentations.......................................................................................................... viii!

2009 Posters .................................................................................................................... viii!

2009 Student Participation ............................................................................................. viii!

2009 Student Presentations ............................................................................................. ix!

2009 Student Posters ........................................................................................................ ix!

In Memory of John D. Baker (August 9, 1949-April 23, 2009)........................................ xi!

Acronyms and Symbols.................................................................................................. xiii!

Time Projection Chamber Project..................................................................................... 2!

Introduction ............................................................................................................................ 2!

2009 Summary ...................................................................................................................... 3!

TPC Hardware................................................................................................................ 4!

Scope..................................................................................................................................... 4!

Highlights ............................................................................................................................... 4!

Time Projection Chamber [LLNL, LANL] ............................................................................... 4!

Pressure Vessel .......................................................................................................... 4!Field Cage ................................................................................................................... 5!High Voltage Feed-through ......................................................................................... 6!Pad Plane Readout ..................................................................................................... 7!MicroMegas ................................................................................................................. 8!TPC Stand ................................................................................................................... 9!

Data Acquisition System [LLNL, LANL, ACU]...................................................................... 12!

Preamplifier Boards ................................................................................................... 12!

Preamplifier Test Stand ....................................................................................................... 23!

EtherDAQ Digitization Boards ................................................................................... 25!MIDAS Evaluation ..................................................................................................... 30!

Gas Handling and Temperature Control Systems [CSM] .................................................... 30!

Gas System Computer Control ................................................................................. 30!Mock TPC Chamber .................................................................................................. 31!Rest Gas Analyzer..................................................................................................... 33!83

Kr Calibration .......................................................................................................... 36!HEPA Filteration ........................................................................................................ 37!

Slow Controls System [ACU] ............................................................................................... 37!

Target Design and Fabrication [OSU, INL] .......................................................................... 38!

Target Fabrication [OSU]........................................................................................... 40!Target preparation ..................................................................................................... 44!

vi Fission Time Projection Chamber Annual Report Fiscal Year 2009

TPC Software ............................................................................................................... 49!

Scope ...................................................................................................................................49!

Overview ....................................................................................................................49!TPC Packet Format Design .......................................................................................50!Event Builder Development .......................................................................................51!TPC Raw Data File Format........................................................................................54!Data Acquisition Simulation Chain.............................................................................54!

Offline Software [CalPoly, LLNL, ACU] ................................................................................56!

Overview ....................................................................................................................56!Reorganization...........................................................................................................57!Monte Carlo truth association ....................................................................................60!TPC Reconstruction Library.......................................................................................61!Geometry Conversion and Event Display ..................................................................66!NIFFTE Virtual Machine.............................................................................................68!NIFFTE Software Manual ..........................................................................................69!

Data Acquisition Software [ACU, LLNL] ...............................................................................69!

Overview ....................................................................................................................69!MIDAS Development .................................................................................................70!

Simulation [GIT, LANL, CalPoly, ACU].................................................................................70!

Overview ..............................................................................................................................71!

TPC Detector Simulation (‘DetSim’) Library...............................................................74!Mock Data Challenge.................................................................................................75!

MDC Hardware.....................................................................................................................76!

Collimation Simulations..............................................................................................76!

Advanced Neutron Detection ..................................................................................... 81!

Scope ...................................................................................................................................81!

Highlights..............................................................................................................................81!

The Hydrogen Standard.............................................................................................. 95!

Scope ...................................................................................................................................95!

Highlights..............................................................................................................................95!

Hydrogen Standard [OU]......................................................................................................95!

H(n,n)H Experiment ...................................................................................................96!Experimental Status...................................................................................................97!

Facilities and Operation............................................................................................ 107!

Scope .................................................................................................................................107!

Highlights............................................................................................................................108!

Livermore [LLNL]................................................................................................................108!

Los Alamos [LANL].............................................................................................................108!

FIREhouse ...............................................................................................................108!TPC Lab Space........................................................................................................108!Physical and Environmental Interface......................................................................109!FIREHouse Upgrades..............................................................................................109!

Ohio University [OU]...........................................................................................................111!

Management .............................................................................................................. 112!

References ................................................................................................................. 112!

vii Fission Time Projection Chamber Annual Report Fiscal Year 2009

Major Contributors

Universities (NERIc funded)

N. Hertel (GIT)

E. Burgett (GIT)

W. Loveland (OSU)

J. Klay (CalPoly)

R. Towell (ACU)

D. Isenhower (ACU)

S. Watson (ACU)

H. Qu (ACU)

S. Grimes (OU)

T. Massey (OU)

U. Greife (CSM)

Laboratories (AFCI and NNSA funded)

M. Heffner (LLNL)

T. Hill (INL)

F. Tovesson (LANL)

C. McGrath (INL)

2009 Submitted Papers, Reports and Summaries

NERI Monthly Reports for October-September, edited by T. Hill, submitted to

NERI Program Office

NERI 2nd Quarter Report, edited by T. Hill, submitted to NERI Program Office

NERI 3rd Quarter Report, edited by T. Hill, submitted to NERI Program Office

NERI 4th Quarter Report, edited by T. Hill, submitted to NERI Program Office

Target Preparation for the Fission TPC, W. Loveland and J.D. Baker,

J. Radioanal. Nucl. Chem. 282, 361 (2009).

viii Fission Time Projection Chamber Annual Report Fiscal Year 2009

2009 Presentations

The Fission TPC Project, M. Heffner (LLNL), contributed talk presented at the

Fission Workshop at Los Alamos, NM, February 2-5 2009.

The Fission TPC Project, M. Heffner (LLNL), invited talk at the AFCI Nuclear

Physics Working Group Meeting at Coronado Island, CA, May 2009.

Target Preparation for the Fission TPC, W. Loveland (OSU), contributed talk

presented at the ANS Conference on Methods and Applications of Radioanalytical

Chemistry, MARC VIII, Kona, HA, April 2009.

A Time Projection Chamber for Precision Fission Cross Sections, R. Towell

(ACU), invited talk at the LANSCE User Group Meeting in Santa Fe, NM, October

1, 2009.

The Time Projection Chamber Project, T. Hill (INL), contributed talk at the APS

Division of Nuclear Physics meeting in Waikoloa, HI, October 17 2009.

The Time Projection Chamber Status, M. Heffner (LLNL), contributed talk at the

APS Division of Nuclear Physics meeting in Waikoloa, HI, October 17 2009.

The Data Acquisition System for the NIFFTE Fission TPC Project, R. Towell

(ACU), contributed talk at the APS Division of Nuclear Physics meeting in

Waikoloa, HI, October 17 2009.

2009 Posters

The Time Projection Chamber Project, presented by M. Heffner (LLNL) at the

Boost Initiative Review meeting, Albuquerque, NM, April 7, 2009.

Fission Time Projection Chamber Software Update, presented by Dr. H. Qu

(ACU) at the 2009 AFCI Annual Review Meeting, Albuquerque, NM, October 22,

2009 – received DOE Nuclear Energy Fuel Cycle R&D Excellence Award.

2009 Student Participation

Undergraduate Graduate

Abilene Christian University 6

Cal Poly San Luis Obispo 2

Colorado School of Mines 2

Georgia Institute of Technology 4

Oregon State University 1

ix Fission Time Projection Chamber Annual Report Fiscal Year 2009

2009 Student Presentations

GEANT4 Simulation of the NIFFTE TPC, R. Thornton (ACU), contributed talk at

the 2008 Texas/Four Corners APS Sections, El Paso, TX, October 2008

Introduction to NIFFTE and its Data Acquisition System, A. White (ACU),

contributed talk at the 2008 Texas/Four Corners APS Sections, El Paso, TX,

October 2008 – received award

Track Reconstruction for the NIFFTE TPC, S. Sharma (ACU), contributed talk

at the 2008 Texas/Four Corners APS Sections, El Paso, TX, October 2008 –

received award

Design of a Gas Delivery System for the NIFFTE TPC, L. Snyder (CSM)

contributed talk at the APS Division of Nuclear Physics, Waikoloa, HI, October 17,

2009

The Offline Software for the NIFFTE TPC, R. Kudo (CalPoly) contributed talk at

the APS Division of Nuclear Physics, Waikoloa, HI, October 17, 2009.

Track Reconstruction Techniques for the NIFFTE Time Projection Chamber,

S. Sharma (ACU) contributed talk at the Joint Fall 2009 Meeting of the Texas

Sections of the APS, AAPT and SPS, October 24, 2009.

Simulations of the NIFFTE Time Projection Chamber, R. Thornton (ACU),

contribute talk at the Joint Fall 2009 Meeting of the Texas Sections of the APS,

AAPT and SPS, October 24, 2009

2009 Student Posters

Fission Time Projection Chamber Status Report, presented by L. Snyder

(Colorado School of Mines) at the 2008 GNEP Annual Review Meeting, Idaho

Falls, ID, October 2008 – received award

Fission Time Projection Chamber Project, presented by E. Burgett (Georgia

Institute of Technology) at the 2008 GNEP Annual Review Meeting, Idaho Falls,

ID, October 2008 – received award

TPC tracking software for NIFFTE: the Neutron Induced Fission Fragment

Tracking Experiment, presented by R. Kudo (CalPoly) at the 2008 APS-DNP

Meeting, Oakland, CA, 2008 – received award

NIFFTE Overview and Goals, presented by S. Stewart (ACU) at the 2008 APS-

DNP Meeting, Oakland, CA, October 2008 – received award

x Fission Time Projection Chamber Annual Report Fiscal Year 2009

GEANT4 Simulation of the NIFFTE TPC, presented by R. Thornton (ACU), at

the 2008 APS-DNP Meeting, Oakland, CA, October 2008 – received award

Track Reconstruction for the NIFFTE TPC, presented by S. Sharma (ACU), at

the 2008 APS-DNP Meeting, Oakland, CA, October 2008 – received award

Slow Controls for the NIFFTE TPC, presented by N. Pickle (ACU) at the APS

Division of Nuclear Physics, Waikoloa, HI, October 15, 2009 – received award

Track Reconstruction for the NIFFTE TPC, presented by S. Sharma (ACU) at

the APS Division of Nuclear Physics, Waikoloa, HI, October 15, 2009 – received

award

Simulation of the NIFFTE TPC, presented by R. Thornton (ACU) at the APS

Division of Nuclear Physics, Waikoloa, HI, October 15, 2009 – received award

Fission Time Projection Chamber Status Report, presented by L. Snyder

(CSM) at the 2009 AFCI Annual Review Meeting, Albuquerque, NM, October 22,

2009 – received DOE Nuclear Energy Fuel Cycle R&D Excellence Award

xi Fission Time Projection Chamber Annual Report Fiscal Year 2009

In Memory of John D. Baker (August 9, 1949-April 23, 2009)

It is with deep regret and a profound sense of loss that we note the passing of one of the original members of the NIFFTE collaboration, John D. Baker, of the Idaho National Laboratory. John received his B.S. degree from Northern Illinois University and his M.S. degree from the University of Idaho. John was a long time employee of INL and its predecessors.

John was known as an extremely accomplished radiochemist. His work on chemical separation techniques led to the discovery of five new radionuclides and was important in studies of fundamental processes related to neutrino production. John held five US patents and was the author or co-author of over 50 journal articles. John was known for his very careful, exacting work and his vast knowledge of many practical aspects of nuclear and radiochemistry.

His vibrant personality and his constant joking and laughing endeared him to a wide range of his colleagues. We will deeply miss his contributions to our project and our lives.

xii Fission Time Projection Chamber Annual Report Fiscal Year 2009

Blank Page

xiii Fission Time Projection Chamber Annual Report Fiscal Year 2009

Acronyms and Symbols

" Alpha particle ACU Abilene Christian University ADC Analogue to Digital Converter AFCI Advanced Fuel Cycle Initiative AFM Atomic Force Microscopy Al Aluminum ALEXIS Accelerator at Livermore for EXperiments in Isotope Sciences Am Americium ANL Argonne National Laboratory ANS American Nuclear Society ASIC Application Specific Integrated Circuit ASME American Society of Mechanical Engineers Atm Atmosphere (pressure unit) Ba Barium Be Beryllium Bi Bismuth BNL Brookhaven National Laboratory BRONDL Evaluated Nuclear Data File (Russian) C Carbon CAD Computer Assisted Drawing CalPoly California Polytechnic State University, San Luis Obispo CAMAC Computer Automoated Measurement And Control Ce Cerium Cf Californium Cl Chlorine Cm Curium CS cross section Cs Cesium CSM Colorado School of Mines Cu Copper CVD Chemical Vapor Deposition d deuteron DANCE Detector for Advanced Neutron Capture Experiment DAQ Data Acquisition System dE/dx Differential Energy loss per Differential path-length DetSim Detector Simulation (software) DLC Diamond-Like Carbon DOE Department of Energy dpa Displacements per Atom EIS Environmental Impact Statement ENDF Evaluated Nuclear Data File (US) EOF End Of File ES&H Environmental, Safety, and Health EtherDAQ Ethernet-based Data Acquisition System (TPC specific) Eu Europium EvB Event Builder F Flourine Fe Iron FIREHouse FIssion Ratio Experimental House (WNR neutron beam facility) FPGA Field-programmable gate array FWHM Full Width Half Maximum Ga Gallium

xiv Fission Time Projection Chamber Annual Report Fiscal Year 2009

Gd Gadolinium GEANT4 Geometry And Tracking monte carlo program from CERN GIT Georgia Institute of Technology GNASH Nuclear Reaction Code GTRI Georgia Tech Research Institute H Hydrogen He Helium HEPA High-Efficiency Particulate Air HEU Highly enriched uranium Hg Mercury HPCC High-Performance Computing Cluster IAC Idaho Accelerator Center In Indium INL Idaho National Laboratory I/O Input/Output JAERI Japan Atomic Energy Research Institute JENDL Evaluated Nuclear Data Library (Japan) K Potassium keV Kilo (thousand) electron Volt Kr Krypton LA150n Los Alamos generated nuclear data library, extending up to 150 MeV LANL Los Alamos National Laboratory LANSCE Los Alamos Neutron Science Center Li Lithium LLNL Lawrence Livermore National Laboratory µC micro-Curie

µg micro-gram

µm micro-meter mA milli-Ampere MA Minor actinide MAC Media Access Control mb Millibarn mCi Millicurie MC Monte Carlo MCNP Monte Carlo N-Particle Transport Code MCNPX Merged code—Los Alamos High-Energy Transport (LAHET) and Monte Carlo

N-Particle Codes (MCNP) MCP Multi-Channel Plate (detector) MDC Mock Data Challenge MeV Mega (million) electron Volts MFC Mass Flow Controller Micro-megas Micro-MEsh Gaseous Structure MIDAS Maximum Integrated Data Acquisition Software mips Minimum ionizing particles mL Milliliter Mo Molybdenum MOCVD Metal Oxide Chemical Vapor Deposition mR Millirad (a measure of radiation) n neutron N Nickel Np Neptunium NE213 Proprietary liquid organic scintillator material NEPA National Environmental Protection Agency NERAC Nuclear Energy Research Advisory Committee NERI Nuclear Energy Research Initiative

xv Fission Time Projection Chamber Annual Report Fiscal Year 2009

NIFFTE Neutron Induced Fission Fragment Tracking Experiment (TPC Collaboration name) nm nano-meter ns nano-second O Oxygen or Oxide O&M Operations and Maintenance ORNL Oak Ridge National Laboratory OSU Oregon State University OU The Ohio University PACS Personnel Access Control System Pb Lead PC Personal Computer PCB Printed Circuit Board Pd Paladium pF pico-Farad PID Particle IDentificatoin PMT Photo-Multiplier Tube PNNL Pacific Northwest National Laboratory Pu Plutonium QA Quality Assurance R Rad (a measure of radiation) Rb Rubidium rms root mean square RGA Residual Gas Analyzer ROOT an object oriented data analysis framework from CERN RSICC Radiation Safety Information Computational Center Ru Ruthenium sccm standard cubic centimeters per second SDRAM Synchornous Dynaminc Random Access Memory SEM Scanning Electron Microscopy Sn Tin STP Standard Temperature and Pressure svn SubVersioN (software version control system) Ta Tantalum Tc Technitium TEM Transmission Electron Microscopy Th Thorium Ti Titanium TJNAF Thomas Jefferson National Accelerator Facility TPC Time Projection Chamber TTL Transistor-Transistor Logic U Uranium V Vanadium W Tungsten WBS Work Breakdown Structure WNR Weapons Neutron Research (facility at LANSCE) Xe Xenon XRD X-ray Diffraction Zn Zinc Zr Zirconium

1 Fission Time Projection Chamber Annual Report Fiscal Year 2009

Blank Page


Time Projection Chamber Project

Introduction

Reactor core calculations are dependent on nuclear physics for cross sections and kinematics. Design codes interface the nuclear data through nuclear data libraries, which are a culmination of experimental results and nuclear theory and modeling. Uncertainties in the data contained in those libraries propagate into uncertainties in reactor performance parameters, such as: criticality, peak power, temperature reactivity, transmutation potential, radiotoxicity and decay heat in a repository. The impact of nuclear data uncertainties has been studied in detail for transmutation systems and sensitivity codes have subsequently been developed that provide nuclear data accuracy requirements based on adopted target accuracies on crucial design parameters. The sensitivity calculations have been performed for a number of candidate systems. These sensitivity studies provide specific requirements for uncertainties on many fission cross sections, many of which are beyond the reach of current experimental tools. The sensitivity codes are proving to be very useful for identifying the highest impact measurements for the AFCI program and the TPC measurement program will help provide those data. The result of these new precision measurements will be a reduction in the reactor core performance margins, thus reducing the construction and operating costs of the new reactor fleet. The new class of precision fission measurements will not be easy. The proposed method is to employ a Time Projection Chamber and perform fission measurements relative the H(n,n)H elastic scattering. The TPC technology has been in use in high-energy physics for over two decades - it is well developed and well understood. However, it will have to be optimized for this task that includes miniaturization, design for hydrogen gas, and large dynamic range electronics. The TPC is the perfect tool for minimizing most of the systematic errors associated with fission measurements. The idea is to engineer a TPC specifically for delivering fission cross-section measurements with uncertainties below 1.0%.

The long term goal is to fill theTPC with hydrogen gas and measure fission cross sections relative to H(n,n)H elastic scattering, thus removing the uncertainties associated with using the U-235 fission cross section for normalization. In fact, we will provide the world's best differential measurement of the U-235 fission cross section and this will impact nearly all fission library data, since it has been used as a standard in much of the available fission experimental data.

The immediate objective of this effort is to implement a fission cross section measurement program with the goal of providing the most needed measurements with unprecedented precision using a time projection chamber. The first three years of this program will provide all the groundwork and infrastructure for a successful measurement campaign. Shortly following, we will provide precision fission ratio measurements for Pu-239/U-235 and U-238/U-235 along with a full design proposal to measure 235U/n(n,p)p. The 235U/H(n,n)H measurement will provide the best single measurement of the U-235 fission cross section and will allow us to convert the initial, and any subsequent, ratio experiments to worlds best absolute measurements. After completion of the U-238 and Pu-239 ratio measurements, the experimenters will


move on to measurement of the minor actinide cross section, fission fragment distribution and neutron yield measurements. This information will play a crucial role in the long term AFCI reactor R&D campaign.

The reporting for this project is broken down into four categories:

• TPC Hardware activities include design, testing and operation of the complete time projection chamber, including gas system and electronics.

• TPC Software activities will provide the project with the required programming for the online data acquisition system, data reduction and analysis as well as simulation.

• The Hydrogen Standard will be used to minimize total cross section errors. The ability to accurately and precisely determine fission cross sections hinges on the H(n,n)H total cross section and angular distributions.

• Facilities and Operations will need to be identified and prepared for the construction, testing and operation of the TPC. This activity is spread amongst the collaborators, based on the work they are performing, such as target fabrication, computing, design, component testing, and operation.

• Management section describes the organizational work required for a project this size.

2009 Summary

All of the TPC hardware components have been designed and most of them constructed and tested. The TPC pressure vessel has been completed, including the high voltage feed-through and pressure testing. The field cage and cathode have been fabricated and successfully tested in place in the chamber vessel. The preamplifier and EtherDAQ electronics cards have been designed, reviewed, constructed and successfully tested. The anode plane has been designed, reviewed and sent out for fabrication. The jig for completing the construction of the anode planes has been designed, constructed and tested. A complete electronics chain test that includes a fully constructed TPC chamber and 128 channels of data will be conducted in the first quarter of 2010.

The TPC online software has also completed an end-to-end test of the components necessary to process data from the TPC Ethernet outputs. The online data packet formats have been implemented to adapt data from the front-end electronics to the receiving computers. Initial event builder functionality has been developed to accumulate and pack online TPC events for storage in the newly defined TPC raw data format. An online DAQ simulation chain has been implemented for the online system validation and performance testing. The new codes will be implemented in the full electronics chain testing to be carried out at LLNL next quarter.

The TPC offline software made several key advances that enabled the Mock Data Challenge exercise to run successfully and advanced our preparedness for analyzing real data. The software repository was re-organized and the entire framework was redesigned to allow detector response simulation, reconstruction and analysis algorithms to run within the same executable. The NIFFTETaskManager was completely overhauled to allow serial processing of multiple events within one executable session. The new code system will also be implemented during the initial TPC testing at LLNL next quarter.


TPC Hardware

Scope

The components that make up the TPC proper are included in this section. This includes the pressure vessel, field cage, pad-plane, gas amplifier, laser alignment system, targets, electronics and the engineering required to integrate all of the parts into a working system.

Highlights

• The TPC chamber, field cage and cathode have been constructed and fully tested. Safety basis documentation is complete for operation at LLNL.

• The TPC preamplifiers have been fully tested and redesigned for final testing.

• The first EtherDAQ cards have been designed, reviewed and initial testing demonstrates full functionality.

• The TPC pad plane design has been completed and an order placed for construction. The micro-megas assembly fixture has also been constructed.

• The TPC gas system is complete and fully tested.

• The TPC mechanical stand design is well underway.

Time Projection Chamber [LLNL, LANL]

The TPC is the centerpiece of the experiment and consists of a number of parts that have to each be designed and integrated into a working whole. This section will describe the progress on each of the subsystems and tasks.

Pressure Vessel

The pressure vessel is designed to contain up to 5 bar of hydrogen or other drift gas. In this quarter, the safety document for the vessel was written and approved. The pressure vessel passed a pressure test and is now ready for operation. A small leak in a high pressure feed-through was found and will be redesigned by an outside company who has already been contracted for the work.

Blank carbon foils from OSU were received at LLNL and LANL to test the shipping and the procedure to insert them into the TPC (see Figure 1 and Figure 2). A fixture is being designed to install the actinide-covered targets into the field cage and minimize the chances of breaking them in the process.


Figure 1: Shown here are the blank carbon foils shipped from Oregon State University and

Lawrence Livermore National Laboratory. One of the foils was broken during transit.

Figure 2: Shown here is a blank carbon foil inserted into to the TPC. The carbon foil is

manufactured on an aluminum support ring that fits into the TPC copper cathode.

Field Cage

The field cage assembly was designed for a maximum cathode drift potential of 27 kV with hydrogen at 5 bar of pressure. A test was conducted in air, which has a lower dielectric strength than hydrogen, at 1 bar of pressure. The equivalent breakdown to exceed in air is 13.5 kV. A Bertan 225 high voltage power supply (see Figure 3) was


purchased for this test (will also be used for nominal TPC operations during data taking). With the new supply, the TPC field cage could be rapidly ramped to over 20 kV and held 23 kV for more than one hour with no indication of breakdown. This is a large safety margin on the design and therefore the field cage design passes.

Figure 3: Here is the Bertan power supply used in the field cage high voltage testing. This unit will

also provide power to the TPC field cage during the planned experiments.

With the high voltage system design verified, focus has been on the pressure vessel and obtaining the safety and engineering documentation to fill the vessel with hydrogen at pressure. This process involves writing an "engineering safety note" that describes how all of the potential hazards are mitigated. A first drift has been written and it is currently under review of the pressure safety officer at LLNL.

High Voltage Feed-through

The TPC field cage and requires high voltage to operate and those voltages are delivered through a small opening in the TPC aluminum bulkhead. A feed-through has been designed to provide passage for three cables through the pencil-sized opening and provide a seal for hydrogen at 5 bar operation pressure and withstand 30 kV applied voltage on the cables. The latest feed-through design has been constructed, installed and successfully tested (see Figure 4).


Figure 4: Shown here is the high voltage feed-through installed on the endplate of the TPC pressure

vessel during the bubble test. The blue plastic is the feed-through and the white wires (0.100”

diameter) deliver 30 kV to the TPC field cage. There is a pool of water around the base to detect

leaks as the detector is pressurized with argon to 5 bar.

Pad Plane Readout

The pad plane is the central piece of the TPC and the design of the first version was finished this year and sent out for manufacturing (see Figure 5). It is expected that the new part will be delivered in late October. Once received, construction of the micromegas layer will commence, followed by testing of the pad plane up to and including the full assembly of the TPC. Once this has been accomplished, readout of single sectors can be started and we can look at the response to sealed sources.

The structure that holds the electronics also supports the pad plane and directs airflow over the electronics is getting the final touches before it is set out for machining. It is expected that we will get the first version of this sometime in November.


Figure 5: Here is a view of the design for the TPC pad plane. It is a 12 layer “computer board” that

incorporates the hexagonal copper read-out pads (central region), the traces from the pads to the

preamplifier board connectors (connectors are along the outer edges) and a variety of support

services, both electrical and mechanical, for the operation and construction of the TPC.

MicroMegas

The foil used in the MICROMEGAS is very delicate, yet must be stretched and glued down to the pad plane to form the gas gain structure of the TPC. Tests were carried out this year using a small, 210-pad prototype. A fixture was also designed (see Figure 6) to accommodate the fabrication of a full-sized pad plane. The fixture will be constructed and tested in the next quarter.

Figure 6: A glue-down fixture was designed to facilitate micro-megas fabrication for a full sized pad

plane.


Figure 7: Shown here is a close-up of the pad plane and the micro-megas structure. The large copper

hexagonal pads have 100 mm tall insulating pillars placed at the corners. These pillars suspend a

fine mesh, which is hard to see in this picture, above the pads at a uniform distance across the entire

plane. A high voltage is applied between the pads and the mesh creating a low-noise, high-gain gas

amplifier that drives signals from the pads to the discrete component preamplifiers that are attached

to the outer edges of the pad plane board.

TPC Stand

A moveable stand will be designed to hold the TPC detector along with the supporting equipment. There are a number of different requirements for the stand that makes the design challenging, such as versatility, mobility and accessibility. A student assigned to this portion of the project has come up with a number of possible solutions.

A document outlining the mechanical and environmental plan for TPC placement at LANL was prepared, meeting a milestone for the Advanced Detector Development package of the Nuclear Data program of the AFCI. The stand is designed to be easily positioned on the neutron flight path, and to hold much of the support equipment for the TPC. The design requirements for TPC stand are

• Ability to accurately place the detector in the neutron beam. The design criteria is the TPC will need to be placed with 1/16” accuracy in the x,y,z positions.

• The stand needs to house some of the supporting equipment, such as gas handling system and readout electronics.

• Due to space limitation on the flight path the stand cannot exceed 3 ft in width.

• There should be a “bag-line” which includes all the equipment that will be removed in case of contamination inside the TPC.

• Ease of transport. The TPC and supporting equipment will need to be moved outside the flight when other experiments are running on the flight path, and in it should also be reasonable straight forward to ship the TPC to another facility when needed.


The current TPC stand design consists of a table made from aluminum 80/20 bars and .25 in aluminum plates. The TPC will be attached to a positioning table located on top of the stand with the whole stand rolling on casters. The stand is designed to hold much of the TPC associated electronics, as well as the gas handling system. Equipments housed in the current design are:

• 8 Computer switches - Dell Powerconnect 6224

• 2 Power Supplies - RPS-600 Redundant Power Supply

• 3 Slow Control units – IOtech DaqScan 2000 series. One controller unit, one thermocouple reader, and one voltage reader.

• Gas Controls – MKS PR 4000, MKS Multi Gas Controller 647 C

• Gas handling system

• Rest gas analyzer (RGA)

The gas handling system, excluding the controller units, are located under the table in the current design. The design of the gas handling system is expected to change in the final version, so the current stand design will be modified to accommodate the final gas handling system. The likely scenario is that the “side table” currently holding the RGA will house the gas handling system (see Figure 8 and Figure 9), and that a smaller chamber will be used for the RGA to save space. A laser unit that will be used for TPC calibration purposes will be mounted on the wall of the flight path, and thus not be part of the TPC stand.

The table is designed to be adjustable along the the x,y and z axis, with the z-axis defined as the neutron beam axis, and the y-axis is up/down. Position adjustments of the TPC along the x- and y-axis is controlled by a table attached to the top of the stand, and the z-axis adjustment is achieved by rolling the whole stand along rails that are attached to the floor of the experimental hall. Any rotational adjustment of the TPC will be done manually, and it will be locked into place on the table with bolts.

The x-axis is adjusted using hand wheels on both ends of the positioning table. The dial is graduated in 0.02 mm increments, and has 12” of travel. Table size is 18-5/8” long and 6-3/16” wide. It has two 9/16" T-slots that are 2 3/8" center-to-center. Also has a load capacity of 150 lbs. The y-axis is controlled by 4 hex body threaded studs at each corner of table, and has 3-1/2” of vertical travel. It has 16 threads per inch, and is made of aluminum. One end is right hand thread, and the other is left handed thread so that it pushes or pulls the table up or down.


Figure 8: CAD drawing of the TPC stand. This is an earlier design where the stand rolls on ball

casters. Figure 9 show the latest design with wheels that rolls on rails in the floor of the flight path.

The TPC stand rolls along the beam axis on rails in the floor using four v-shaped casters. The casters are made from stainless steel, and are thus rigid enough to avoid any sinking of the stand. Bolts in the corners of the bottom plate of the stand will lock it into position along the rails.

The stand design work is being carried out using the proEngineer 3D CAD software. This versatile program made it possible to import the TPC mechanical drawings prepared by LLNL and the CAD drawings of the FIREhouse experimental hall prepared by ACU (see Figure 10), and thus inserting the stand designs with these other elements.

Figure 9: Side view of the TPC stand design. The table to the right holds the TPC itself, computer

switches and electronics. To the left is the chamber for the rest gas analyzer.


Figure 10: CAD drawing of the FIREhouse. The shielding blocks shown here are movable, and might

be reconfigured to allow the TPC to be moved around inside the flight path.

Data Acquisition System [LLNL, LANL, ACU]

The TPC will have over 6000 pads, each of which need to be instrumented with a preamplifier, ADC and digital readout. The challenge of a large number of densely packed high-speed channels has been met in the past with custom ASIC chips. The technology of both ADC and FPGA has improved considerably over the last decade and it is now possible to use off-the-shelf components to accomplish the same task for considerably less development cost, less time to working prototypes, and considerably more flexibility in the final design.

Preamplifier Boards

The design files for the preamp boards were laid out last quarter and were sent to a PCB manufacturer and built. Twenty boards were built for initial testing and so far, two boards have been loaded with 4 channels (out of 32) each. The connectors on these boards are very small so they can fit into the small TPC connectors (see Figure 11 and Figure 12). Due to their small size, it is difficult to power up and test the preamps without the TPC, or a breakout board that plugs into the preamp and provides larger connections for test equipment (see Figure 13 and Figure 14).


Figure 11: Shown here is a 32-channel preamp card alongside a quarter for scale.

Figure 12: This is a close up view of one of the preamp channels on the preamp board. The width of

the channel is approximately 3.5 mm.


Figure 13: This board was designed and built to provide access for delivering test signals to each

channel of the preamp for bench testing.

Figure 14: This board was designed and built to provide access to each of the output signals from the

preamp for bench testing measurements.


A variety of tests were carried out on the initial preamplifiers and a few initial issues were identified for correction:

• Only 1/2 dynamic range of ADC spanned. We calculated the resistors (and ran spice simulation) to use more of the range and will load a channel for testing.

• The output pins were swapped on the schematic and silkscreen of the breakout boards. This will not be a problem for the digital board since it is not laid out yet.

• Small undershoot. This is acceptable, but we are investigating if it is possible to reduce the effect.

The second, and most recent, version of prototyped preamplifier board with four populated channels (channels numbers are 28, 29, 30, and 31) was also tested this year. The few minor problems identified in the first prototype were rectified in the new preamp release. During testing, all response parameters of the preamp boards were measured. The same setup as before was used for preamp test. Tests were completed on all four populated channels of the preamp board, see accumulated traces in Figure 15. Comparison of amplifying coefficients and output pedestals was done using the same input signal for charge injector. Numeric results are presented in Table 1. It was found:

• All populated channels are functional,

• The under-damped condition of the output signal has been corrected in the updated version of preamp,

• The amplifying coefficient is slightly lower for updated version. For example, for channel #28 it is 10.8 vs 12.9 for initial version at signal of 400 mV inputted into charge injector,

• All channels show good similarity in their response parameters except for channel #29, which shows output signal about 30% higher than others.

• Detailed testing of preamp response parameters was carried out using channel #28. Input and output signal traces were accumulated for different pulse height signals in the extended dynamic range of 50 – 1200 mV as input into charge injector, see Figure 16.

The following preamp performance features were investigated:

• Output pulse height vs input,

• Amplifying coefficient dependence on input pulse height,

• Pedestal in output signal dependence on input pulse height,

• Output signal rising slope vs input pulse height,

• Timing properties of output signal.


Figure 15: Shown here are the signal traces from all four populated channels on the updated version

of prototyped preamp board using the same input signal.


Table 1: Shown here are the response parameters of the populated channels #28, 29, 30, and 31 of the

updated prototyped preamp board.

Preamp # Input pulse

height, V

Output pulse

height, V

Output pulse

pedestal, V

Amplifying coefficient

28 0.0244 0.262 0.0074 10.7

29 0.0226 0.431 0.0079 19.1

30 0.0247 0.272 0.0080 11.0

31 0.0247 0.265 0.0083 10.7

Figure 16: Shown here are signal traces measured for different pulse height signals in the extended

dynamic range. The upper plot is the preamp input signal and the bottom plot is the preamp output

signal.


The results obtained are presented in Figure 17 and from these results we can conclude:

• Updated preamp is more linear in response (linear fit has smaller errors). It has wider dynamic range of inputted pulse height: saturation in preamp output signal starts at charge injector input of 1000 mV (corresponding preamp input pulse height of 0.065 V),

• Pedestal of output signal does not show significant change below input pulse height of 0.05 V, it is slightly decrease with input pulse height. But above 0.05 V it changes noticeably,

• Amplifying coefficient is much stable in the unsaturated dynamic range of input pulses 0 – 0.065 V. Graph vertical scale is the same as for the similar plot for initial prototype, see Fig. 11 in the 3rd Quarterly Report,

• Rising slope of output signal is pretty linear with input pulse height but above 0.05 V some saturation begins,

• It is very noticeable improvement in a time value for output pulses to reach its 90% level for updated preamp board in comparison with initial one. It is very visible the same graph vertical scale as for the similar plot for initial prototype. At 0.065 V (charge injector input of 1000 mV) output signal is a bit slower, 2 ns difference from average level, but it does not look too important for our application.


Figure 17: Shown here are the performance features of the updated preamp, channel #28 in the

extended dynamic range of input signal. The first plot shows the preamp input versus the input

pulse at the charge injector. The second plot shows the preamp output pulse height versus the

preamp input. The third plot shows the pedestal values of the output pulse. The fourth plot shows

the amplifying coefficient as a function of the input signal. The fifth plot shows the rising slope of the

output signal as a function of the input signal. The sixth plot shows the time to get to the 90% level of

output signal (rise-time).

In general we can conclude that preamp board has good design and is ready to be used in the TPC experiment.

Cross talk can be measured between different channels of the board. For that measurement a signal was provided only to channel #28 (see second plot for input


signal and third for output one in Figure 18 and measure output signals in channels # 29, 30, and 31. By eye cross-talk pulse is linear with input pulse height in channel #28. Scope picture, Figure 19, shows output cross-talk height in channel # 30 of about 2.8 mV. Comparing with “normal” output signal in channel #30 of 272 mV (Table 1), cross-talk pulse is about 1%, which is slightly higher than that of initial prototype (0.6%). Estimation of cross-talk in channels 29 and 31 was done using twice higher input pulse. Scope picture of signal in channel # 29 shows output cross-talk height of about 2.3 mV, Figure 20. Estimation of cross-talk pulse height for 400 mV charge injector input is 1.15 mV or 0.3% of “normal” output signal of 0.431 V in channel #29 (Table 1). No signal was observed at output of channel #31 because channels 28 and 31 are neighbor channels, which are placed on different sides of the circuit board. We may conclude from these measurements:

• The new version of the preamp board was tested for inter-channel crosstalk. It was determined that 1% of the nominal signal is measured on the same side neighboring channels and 0.3% on the diametrically opposed channel on the opposite side of the board,

• The crosstalk is limited to only these three nearest neighbors,

• The amount of cross talk measured will not result in any problem for future measurements.

Figure 18: Shown here is a picture of the signal trace for preamp channel #30 (4th

plot) and #29

(5th

plot) while only channel #28 is being pulsed (2nd

and 3rd

plots are its input and output signals).


Figure 19: Shown here is a picture of the scope trace for preamp channel #30 while only channel #28

is being pulsed. The yellow trace is the synch pulse for input signal to channel #28 and the blue trace

is the signal measured in channel #30, the nearest neighbor to channel #28.

Figure 20: Shown here is a picture of the scope trace for preamp channel #29 while only channel #28

is being pulsed. The yellow trace is the synch pulse for input signal to channel #28 and the blue trace

is the signal measured in channel #29, the diametrically opposed channel on the opposite side of the

board.


Three irradiation tests of the initial preamp board were carried out in direct beam at the flight path 4FP90L at LANSCE. Irradiation times were 0.5, 2 and 7.8 hours. Detailed test of preamp properties was done after each irradiation run. There were no visible changes in preamp performance features. In Figure 21 we tracked change of some derivative parameters of preamp to find any tendency to failure. Plots in Figure 21 are built based on the parameters, measured after each irradiation test, and which are similar to those presented in Figure 17. We don’t see any crucial change in preamp performance. That clearly demonstrates preamp radiation stability. Coarse estimation indicates that reached duration of irradiation in direct beam is exceeded by many times expected irradiation of preamp boards of the TPC over full life of experiments. This test will continue running until radiation failure thresholds are discovered.

Figure 21: Shown here is a picture of the preamp parameters change (for initial prototype) with

irradiation time.


Preamplifier Test Stand

The TPC anodes are segmented into nearly 6000 pads, each of which are read out through an independent preamplifier channel on a board that contains 32 channels. A fully instrumented TPC will have 192 preamplifier boards. All of the boards and each independent channel will need to be tested and measured before installation. It will also be necessary to test boards that develop performance issues during use. To carry out these types of test in bulk fashion, a test stand is being developed for this important task. The preamplifier test stand will be located near the experiment at LANL to make detailed tests of preamp properties in manual mode or mass testing in automatic mode.

The test stand was designed to make tests in two modes - manual and automatic. Manual mode is used mostly for a detailed inspection of a few channels on a board. This mode was developed for the initial tests of the partially populated preamp boards. The data obtained in these tests were used as the basis for the requirements in the development of the automatic mode testing, which will be used to measure key performance parameters of each channel before installation in the TPC. Some of these performance parameters will bee used as the initial calibration for each of the channels.

Automatic mode is used for mass tests of boards in use. The test stand in use in the automatic mode is presented in Figure 22. The idea implemented in the test stand design is simultaneous loading of an input signal to the individual preamplifier channels with a specific and uniform pulse for measuring the output signal of all channels with meaningful inter-comparisons. Based on this information, the gain parameter and crosstalk are measured as a function of input signal, which is enough to make confident conclusion about preamp board quality and performance.

Figure 22: General view of the Test Station use in the automatic mode.


The preamplifier test stand in automatic mode uses information about preamp characteristics measured in the manual mode as a starting point. Each preamp input is injected with the same 3 pF charge. Due to significant scattering in electronic component parameters, it is complicated to deliver the same input signal from 32 different circuits. Instead, a singe charge injection is split out to all the preamp channels. To realize this charge injection method, the new charge injection board was designed. The block-scheme of the test station is presented in Figure 23.

Figure 23: Test Station schematic in the automatic mode.

The charge injection distribution board consists of four pairs of parallel to serial decoders of eight channels each to cover all 32 preamps on a board. The schematic for the board is presented in Figure 24. The parallel port TTL signals are buffered using a 74HCT573 (Octal transparent buffer). These TTL signals are used to select 1 of 32 relays to activate. This is done using four 74HCT238 (3 to 8 line decoders). The decoder output can drive up to 25-30 mA to activate the relays coil. Every single relay has a diode in parallel with the coil.


Figure 24: The Distributive board schematic.

This shorts out inductive voltage spikes from the collapsing magnetic field. The diode is a 1N914 with transition time of 4 ns. The Distributive board connects to preamp board using the 80R-JMDSS-G-1-TF connector.

The designed printed circuit board is a standard 0.062 inch thick four layer PCB. It is fabricated using FR4 dielectric and 1 oz thick copper layers. There is solder mask and silkscreen on both the top and bottom. Both through hole and surface mount technologies (SMT) were used in the design of this board. To accommodate the 32 SMT relays, there are components placed on both sides of the board. The four-layer stack up, starting from top to bottom, consists of the top layer (Component side), the ground plane, the power plane, and then the bottom (Solder side). Every trace is routed on either the top or bottom. The overall size of the board is 5 x 3.5 inches.

The automated preamplifier test stand is controlled by a Linux PC. The MIDAS software package is used for storing and viewing the data collected by ADCs. The MIDAS front-end code programs and collects data from ADCs while the slow control features are used to program the channel selection thru a parallel port. The time required to test one channel and switch the charge injector to another one is about 1 ms. A complete preamplifier board takes less than a second to run through the testing.

EtherDAQ Digitization Boards

The first version of the EtherDAQ layout has been completed and the digital board has been through an engineering review. The board has been sent out for a limited fabrication run. Documentation of the hardware is complete and has been distributed to the collaboration.


Figure 25: Shown here is the top side of the TPC EtherDAQ digital board along with the locations of

the major components.

The firmware for the digital board was improved to allow individual threshold setting of individual channels, which will provide very fine control of the triggering. This change will be useful in optimizing the signal to noise performance of the TPC.

A full 2.2 ms of post-place-and-route simulation has been completed which allows us to look at the FPGA performance after some counters roll over. The simulation takes a few days of computer time to complete and is generally considered a very good simulation of the hardware. Usually, if the simulation works, the hardware will work. This exercise has uncovered a few bugs and they were fixed. The firmware is in good shape and the focus was then placed on the hardware and software to read out the card.

Software written in C++ has been developed to communicate with the EtherDAQ cards and has been tested on a windows PC over a 100BaseT network. Simulated output from the FPGA simulation was sent over a real network and reconstructed by this code. This is a full chain test of the readout demonstrating that it should work once all the hardware is in place.

The first version of the EtherDAQ board has been built and populated with components (see Figure 26 and Figure 27). In addition, a power supply test board has also been built and the power supply section tested. The power supplies work correctly but one of them showed excessive noise initially, which was fixed.


Figure 26: Shown here is the top side of the EtherDAQ board. A US quarter is also shown for scale.

Figure 27: Shown here is a close-up of the EtherDAQ board. One of the ADC chips is shown in the

center with the Virtex5 FPGA chip on the left.

After having trouble with the first loading of the digital boards, we tried another shop and had slightly better results but they still shorted some components, see Figure 28. The part placement was within each of the shop’s tolerances. They could not provide a reason why they shorted. To avoid this in the future, the design was modified so the shorted components are further apart.


Figure 28: Micrograph showing a short on a board from the parts loading shop.

We found another short under a part from a “via” that is too close to a component pad. A “via” is the means by which connections are made between layers of a multi-layer board. After fixing this last short, the whole board works, and we now have 4 out of the 5 boards working. The one board that is not working was powered up before we knew about the shorts and it was probably permanently damaged. These 4 boards should be enough to do the initial tests with the TPC while the final work is done with the electronics before producing all 192.

After a moderate amount of debugging all of the main features of the EtherDAQ card that have been tested appear to be working with one exception. The ADC will only work up to 41MHz. We see a lot of errors when pressed to 62.5MHz. The TPC will operate at 41MHz as the fall back position if further debugging can't get the operating speed up to 62.5MHz. This will have no impact on the cross section measurement performance.

Full tests of a few pieces have been carried out and the initial noise measurements show that the EtherDAQ cards are performing well, (see Figure 29). With the exception of a few channels, the noise is below that specified by the manufacture of the ADC chip. The resulting energy resolution is quite good for a mid range signal. There are a few channels that exhibit higher noise and they are closer to the power supply chip. Since this test was done with the input floating we need to repeat this with a preamp attached to see if the noise is still there. If necessary, we will have to look into moving the power section further from the ADC.


Figure 29: The noise for all channels of one EtherDAQ board. The blue line shows the spec from the

manufacturer of the ADC chip and the system does better than this on most channels.

We connected 3 channels to a preamp card and completed a full end-to-end test of the system. The noise is shown in Figure 30. If this noise level can be maintained for the full set of electronics, the energy resolution will not be limited by the electronics.

Figure 30: Noise on one channel with preamp connected.


MIDAS Evaluation

A Maximum Integration Data Aquisition System (MIDAS) slow control front-end program has been written that is capable of controlling the DaqScan2001 and receiving data from it. The front-end program is accessed via MIDAS web browser interface. The front-end was written using packet captures taken from communication between the DaqView program (running on a Windows machine) and the DaqScan2001. Soon it will be possible to plot received data. Work is needed to address problems with the packets, and uploading the data to the MIDAS Online DataBase (ODB). There has been continual work on integrating all of the parts being developed into the MIDAS framework. So far the choice of MIDAS has shown to be a good one.

Gas Handling and Temperature Control Systems [CSM]

One dominant source of error in the absolute measurement of fission cross sections is the normalization of fission data to the U-235 standard. Any campaign that wants to improve on existing data will have to include a new absolute U-235 measurement. The U-235 reference is used to determine the total neutron beam flux, however this method incurs an error of no less than 1% due to the cross section error. The most promising alternative is the reaction 1H(n,n)1H, which is known to 0.2%. This would though require the use of either pure Hydrogen or a very well known admixture as both target and detection gas in the TPC. In order to be not the limiting factor for the precision of the experiment, the hydrogen density will need to be known and kept constant within 0.1%. For calibration purposes, a gas admixture of Kr-83 will be used and in a later stage, one experimental possibility would be the use of a gaseous actinide target. We are planning for a system that allows for the use of three gases being mixed and supplied to the TPC.

Gas System Computer Control

A LabVIEW program was written to implement procedures for the slow fill of the TPC (see Figure 31). The program communicates with the gas system’s main controller, which includes the proportional-integral-derivative controller (PID), via an RS-232 connection. Procedures for the slow fill of the TPC were developed to minimize the risk of thin target breakage and to avoid “spin-up” of the integral term in the PID. Thin target breakage can occur when the PID controller sets the mass flow controllers (MFC) to high flow rates to fill the TPC. High flow rates across the target may result in target backing breakage. Spin-up occurs when the measured process variable, in this case the pressure, takes an exceedingly long time to reach the set-point. This results in a large overshoot and increases the amount of time it takes for the PID to stabilize. Simultaneously keeping flow rates low and avoiding spin-up requires a program to bypass the PID until the gas system is near the desired set-point. The program allows the user to set a flow rate and mixture ratio on the MFCs and a pressure set-point. It maintains the flow rate until the gas system pressure is within 50 Torr of the set-point. The program then transfers control of the gas system to the PID. The program also operates as a general remote user interface, reporting system status and allowing the user to send all of the commands that would normally be input at the main controller. In addition the program can simultaneously send commands over a second RS-232 connection to the exhaust side MFCs that are controlled by a separate unit.


Figure 31: Shown here is the graphical display of the LabVIEW program used to control the

gas system.

Mock TPC Chamber

A mock TPC chamber, field cage, and target holder has been constructed to test the effectiveness of the slow fill procedures (see Figure 32). Carbon foils with thicknesses of 30 and 50 !g/cm2 were used to simulate the target. The target was held in the center of the field cage in the TPC chamber. The field cage has twelve 5 mm diameter holes around its circumference to allow the gas to flow around the target. The setup was tested under various flow rates to determine when and if target breakage would occur. It was also used to test if any other gas system procedures that involve the opening and closing of valves or the changing of flow rates or pressure damaged the target.

It was found that both the 30 and 50 !g/cm2 could withstand the full flow of the 1000 sccm MFC without damage. The actuating of valves in the system caused a noticeable movement in the target but did not result in breakage. When HEPA filters were inserted in the gas line between the target and the valve, the effects of the actuation were reduced but still present. To completely eliminate the problem the valves directly upstream and downstream of the TPC will be replaced with zero volume change valves.


Figure 32: Shown here is the mock TPC chamber with field cage and carbon foil used for

determining flow safety envelope to avoid target breakage.

The mock TPC chamber was also used to test the PIDs ability to maintain a stable pressure over a long period of time. Measurements from the high accuracy pressure gauge were recorded over two periods of approximately 72 hours and 120 hours, at a pressure of 1500 Torr with and exhaust flow rate of 5 sccm. It was found that once the pressure had stabilized after the initial filling process it fluctuated within a range of 0.6 Torr or 0.04%. The reported accuracy of the pressure gauge is 0.08% of reading therefore the pressure is essentially constant. Figure 33 and Figure 34 show the initial stabilization period of the PID and the long-term pressure measurement for the 120-hour run.

Figure 33: Initial stabilization period of PID.


Figure 34: Pressure after stabilization. The apparent double values of the pressure at given

times is a result of the density of the data points. The offset from 1500 Torr is a result of the

calibration of the high accuracy gauge compared to the PID gauge.

Rest Gas Analyzer

The components for the Residual Gas Analyzer (RGA) system were installed and tested. The gas sampling system was tested in the static sampling mode, where higher precision is expected, compared to the usual flow through mode. The measurements led to the decision to carry out parallel development of a constant-flow sampling mode to handle out-gassing problems in the RGA as they became apparent. All components of the RGA system that were not initially metal sealed have been replaced. This has reduced the out-gassing rate and allowed the RGA system to achieve a lower vacuum pressure (9"10-8 Torr), improving the performance of the static sampling mode (see Figure 36). Tests were performed in which the composition of a mixture of two gases was varied by the mass flow controllers and measurements were taken with the RGA in continuous flow mode. The intent of the experiment was to determine if the RGA could accurately track the changes in composition. It was discovered that, because of the slow flow rates used and the amount of piping between the gas system and the RGA, there was a significant time delay in the RGA registering a change in gas composition. To overcome this problem the gas flow exiting the TPC will have to be passed as close to the inlet of the RGA as is possible. The design of the gas system and the positioning of the RGA will be finalized in coordination with the team at LANL to ensure that the system fits properly in the allocated space.


Figure 35: Rest Gas Analyzer vacuum chamber connected to the previously constructed gas system.

A test of the RGA was performed to determine the lowest measurable partial pressure. A background reading was taken and then argon was leaked into the system until a peak at 40 AMU was discernible above the background noise (see Figure 37). A pressure of approximately 4 " 10-10 Torr was measurable above background. The RGA can operate at pressures above 10-5 Torr, allowing it to measure approximately 1 part in 100,000.

Figure 36: The RGA background noise has been lowered to 1 x 10-10

Torr.


Figure 37: The argon peak of 4x10-10

Torr at 40 AMU can easily be seen over the

background.

An additional test was performed to determine the RGA accuracy when measuring a known gas mixture. A P-10 gas consisting of 10% methane and 90% argon was injected into the RGA system to test its response. Figure 38 and Figure 39 show the methane and argon spectra that were produced by the RGA system.

Figure 38: A partial pressure versus mass plot of the methane cracking pattern.


Figure 39: A partial pressure versus mass plot of the argon pattern.

The area under the peaks corresponds to the total partial pressure of each gas species. The ratio of the area under the peaks of the argon spectrum with respect to those of the methane spectrum was expected to equal 9:1. The actual ratio measured was 9.54:1. This discrepancy is a result of the fact that every species of gas has a different sensitivity to ionization by the RGA filament. Calibrating the RGA to each specific molecule or element that needs to be measured can compensate this effect. Even with calibration, the accuracy of ion gauges usually cannot be expected to be better than 10%. Because of this lack of accuracy the RGA will be most effective at qualitative analysis of contaminants rather than quantitative analysis of composition.

83Kr Calibration

Additionally, literature research was performed on a Kr-83m delivery system. The selected approach will be based on the absorption of a Rb-83 solution (commercially available) into zeolite beads. Research indicates that nearly 100% of the Kr-83m will escape the zeolite beads in gaseous form, allowing it to be carried by the gas system flow for calibration purposes into the TPC volume. In the selected procedure, the zeolite was first baked at 320° C under a flow of pure nitrogen to remove impurities. After the zeolite had absorbed the Rb-83 solution it was baked again to remove the solvent. A measurement of the 520 keV gamma rays from Rb-83 indicated that all of the Rb-83 was retained in the zeolite during the baking. In order to test the stability of the Rb-83 in the zeolite, the sample was placed under a fore vacuum and pumped for 1 day at a pressure of 10-1 Pa, while a filter (20 !m pore size) was placed between the chamber and pump. Subsequent measurements of the 520 keV gamma rays from the sample and the filter indicate that no Rb-83 escaped the zeolite and there was no contamination of the filter. In a further test the sample was exposed to a pressure of 6 x 10-4 Pa and heated to a temperature of 300° C for 4 hours. No measureable loss of


Rb-83 was detected. The release of Kr-83m from the zeolite matrix was investigated by measuring 32 keV gamma rays. The sample was first placed in an airtight polyethylene container. A measurement of the 32 keV gamma rays was performed with the chamber exposed to ambient air and with the chamber closed. Only with the chamber closed were 32 keV gamma rays detected. The same test was performed with the sample being placed in a glass test tube equipped with a vacuum valve connected to a vacuum pump. The results were the same as those for the test with ambient air. The 32 keV gamma rays were only detectable when the vacuum valve was closed. A further test was conducted whereby the intensity of the 32 keV line was measured as a function of time after the vessel was closed. The resulting data was in good agreement with the expected exponential law. It is our intention to perform a similar experiment, and if the results are acceptable, use an Rb-83 solution absorbed in zeolite as our Kr-83m delivery system.

HEPA Filteration

HEPA filters and the associated filter holders were purchased and tested in the gas system. A test was conducted to determine if the filters caused any significant impediment to gas flow. Calibrated pressure gauges were placed upstream and downstream of the filter and monitored as gas flowed through to a chamber and the pressure increased. No measurable pressure difference between the two gauges was detected, therefore there is no flow restriction (at our rates) caused by the filters. The leak integrity of the filter holder was also tested. A leak was detected and could not be repaired. A new filter holder has been designed and is currently being manufactured as CSM.

Slow Controls System [ACU]

The slow control system for NIFFTE is being designed for control and read-out via Ethernet using the MIDAS DAQ framework. Most of the read-out modules have been purchased from IOtech. These include various modules such as thermocouple inputs, digital I/O, strain gauge read-out, differential ADCs for different input voltage ranges, and relay control modules. Among the items completed this year are:

• The IOtech slow controls are now set up in the experimental area at LANSE.

• Continued preparing documentation of the IOtech DaqScan/2001 device for use in the DAQ auxiliary system

• Continued work on modeling the electronics part of the detector; zero suppress algorithm (trigger module) and lookback + pedestal noise.

• Created device and class driver for IOTech DBK90 (temperature reading) via MIDAS.

• Put a parser into the device driver to get rid of headers from IOTech Ethernet packets.

• Added conversion function from raw IOTech temperature data into usable temperature data.

• Created data logger (mlogger) to work with raw and real IOTech temperature data.


Target Design and Fabrication [OSU, INL]

A well-prepared set of targets is very important for high quality measurements of fission cross sections. Uncertainties in fission cross section measurements with fission chambers can be attributed, in part, to uncertainties in the target mass, non- uniformities in the target, surface defects in the targets and surface contaminants in the targets, as well as impure target materials. While the proposed TPC for fission studies will allow detailed corrections for many of these problems, it is of great benefit to start with the highest quality actinide targets.

Summary of 2008 progress

In 2008, we evaluated the possibility of using gaseous plutonium targets and presented a method for their preparation. We made preliminary investigations of the use of molecular plating for the preparation of high specific activity actinide targets including measuring typical non-uniformities (~3.7%) in thickness in these targets. We developed a low geometry method of preparing targets by vacuum volatilization that yielded non-uniformities in thickness of < 0.5% over the target area. We acquired 29 g of high purity (99.7%) 235U. We prepared a 232Th target for use at LLNL along with some 238U targets by vacuum volatilization. We investigated several materials as possible target backings, settling on 30-100 µg/cm2 C as the best choice. We

performed AFM on these materials to determine their surface roughness. We designed and tested vacuum shipping containers for the targets. We successfully shipped these targets to other laboratories.

2009 Progress

At present we are preparing targets on a backing of 30-100 µg/cm2 carbon. The

energy loss of a 80Br fission fragment emerging normal to the target surface and originating at the center of the target is estimated (SRIM) to be 1.8 MeV in a 200 µg/cm2 UF4 deposit and 5.1 MeV in a 100 µg/cm2 target backing. It is felt that

this energy loss must be reduced. Furthermore, we note that the estimation of energy losses of fission fragments interacting with light materials has been found to be problematicali. The use of semi-empirical methods, such as SRIM or LSS stopping power theory, for estimating energy losses of fission fragments in foils of low Z materials, like carbon, leads to errors of 15-80% in the predicted values of dE/dx. Accordingly we investigated the use of thinner target backings beginning with carbon nanotube foils.

The assessment of carbon nanotube foils as possible target backings was completed. We acquired three carbon nanotube foils from D. Shapira at ORNL. (The Houston group prepared the foils by interleaving layers). Two of the foils were mounted on frames with a 2.54 cm diameter hole and one was mounted on a frame with a 1.00 cm diameter hole. Attempts to evaporate UF4 onto the large area foils were unsuccessful due to tears that developed where the foil was attached to the target frame. UF4 was successfully evaporated onto the 1.00 cm diameter foil. However, visual inspection of the deposit shows holes in the target due to holes in the underlying carbon foils. We conclude that the carbon nanotube foils are robust and thin but unsuitable as target backings due to holes and non-uniformities.

A set of 10 µg/cm2 diamond-like carbon (DLC) foils was obtained for evaluation

as a possible target backing. Shown in Figure 40 are two views of a frame mounted 10 µg/cm2 DLC foil. One can see the foils are transparent to light and the use of such


foils would significantly reduce the energy loss in the target backing to 0.6 MeV. Attempts have been made to deposit UF4 on these foils. The deposits are not uniform because the evaporated material adheres poorly to the DLC surface. A rough rule of thumb is that for unsupported target backings, one needs 1 µg/cm2

carbon of support for each millimeter of diameter of the apertureii. The TPC targets have a 20 mm diameter aperture, thus requiring a supporting backing of roughly 20 µg/cm2 of carbon. Such design parameters are consistent with the experience of

others in preparing actinide targetsiii. We conclude that 30-100 µg/cm2 C foils will

be adequate as target backing materials.

Figure 40: Two views of a mounted 10 µg/cm2 diamond-like carbon (DLC) foil.

A first test of the radiation stability of the targets was completed. Depleted uranium targets of thickness 120-150 µg/cm2 of 238U were put in special irradiation

containers and irradiated in a pseudo fission spectrum neutron flux in a core position of the OSU TRIGA reactor for 3600 s. The neutron spectrum is shown in Figure 41iv. The total flux at this position was 4.7 x 1013 n/cm2/s (total dose = 1.7x1017 n/cm2). The target foils tore during the irradiation probably due to the high temperatures (250-300°C) of the samples in the corev. A material loss of 4.8% was observed (the total

dose anticipated in the real TPC experiments is ~ 1012 n/cm2). Two 238U targets of thickness 160 and 280 µg/cm2 were then prepared and shipped to INL where they


were coated with 4 µg/cm2 of Ti. The foils were irradiated in an epithermal neutron

flux in a cooled core position. The dose given to the foils was ~10x the fast neutron dose expected in one week of irradiation at LANL in TPC experiments. No material loss was observed in the foils with a 1# upper limit of 0.5%. A significant color

change was observed in the foils and the foils did become more brittle. We speculate the possible formation of F-centers in the titanium-coated foils upon irradiation.

Figure 41: Shown here is the neutron spectrumiv

at the core position of the OSU TRIGA reactor.

Target Fabrication [OSU]

An order was placed for a new evaporator. The new compact evaporator replaces the old unit at OSU. It allows us the capability of mounting the evaporator in a glove box as needed for the evaporation of 239Pu and other high specific activity materials, high vacuum to reduce target contamination, and allows an increase in productivity and reliability over the old unit, which had reached the limit of its useful lifetime.

The new evaporator arrived, was unpacked and mounted in the lab. It was discovered that the liquid nitrogen trap was omitted from the unit. Since the liquid nitrogen unit is needed to achieve pressures of 10-7 torr or below, which are needed for uniform coating free from light element impurities, the unit was returned to the factory for installation of the missing trap. The new evaporator was modified to include a new trap by Denton Vacuum and returned to us (see Figure 42). Bids have been obtained on a ventilated enclosure for the evaporator for the handling of 239Pu and other high specific activity radionuclides.


Figure 42: Shown here is the new vacuum evaporator.

Tests of the vacuum performance of the evaporator (P~10-7 torr) after 15 min pumping) and the operation of the sample shutter, rotator and resistive evaporation were satisfactory. During a test of a large (1500 W) boat to evaporate Au (whose melting point is similar to that of the actinide tetrafluorides), the wires carrying the current (~300 amps) overheated, melting the heat shield on the wires and an air hose that operated the vent valve of the bell jar, causing the release of an entire cylinder of N2. All the air hoses that operate the system valves were securely relocated to avoid proximity with the cables carrying high current and all heat shields were replaced. The separator was operated with a current of 400 amps without any damage to anything. Tests of the automatic operation of the valves, pump-out and venting showed that our thin targets could survive without damage.

A four-position target holder was constructed for our new evaporator. Since all the controls of the evaporator, i.e. boat current, rotation speed, pressure, etc. are digitally controlled, and we embarked on making a series of calibration curves for evaporator operation. Our typical throughput for targets on 100 ug/cm2 C backings is 12 targets per every 4 hours.

Due to problems in breaking foils during evaporation, a new geometry for evaporation was prepared using a mechanical shutter and calibrated. The calibration curve is shown in Figure 43. The actual yields are better than one would predict from simple geometrical considerations (the dimple boats are focusing). The overall efficiency is about 0.4% in making very uniform targets.


Figure 43: Evaporator calibration curve.

A multi-isotope target consisting of 232Th and 238U (see Figure 44) was prepared to demonstrate the ability to prepare well-defined deposits of different actinides on the same target.

Figure 44: Shown here is a multi-isotope target consisting of 232

Th and 238

U on a thin carbon backing.

Because much of our time involves "-counting the target foils for uniformity as well as

total activity, we have embarked on a design of an "-assay system involving a double-

sided Si strip detector (of the Micron W design) to provide crude auto-radiographs of each target produced within a fraction of a day. Initially we will instrument this counting


system using the high-density electronics associated with our heavy element program. If the system is successful, we will need to explore a more permanent alternative.

In making a 235U target for LANL, we had to convert some of a sample of 99.91% 235U3O8 to UF4. (This material was received earlier from LLNL). We tested the chemical procedure with depleted 238U3O8 and achieved a conversion efficiency of > 90%. (The procedure involves dissolving the oxide in HF, heating it to boiling, reducing any UO2

++ to U(IV) with SnCl2, and recovering the insoluble UF4.) The yield with the LLNL 235U oxide was less than 30% and the reactions and color of the material indicates it is not U3O8 and contains substantial amounts of impurities. Evaporation of the nominal 235UF4 was not efficient with the presence of low melting impurities.

We received a request from LANL to prepare some large area (6 x 3 cm), thick (2 mg/cm2) 232ThF4 deposits on an Al target holder. There is the additional requirement that the targets be completely free of 235U. We built a new holder for these targets and attempted to make 2 demonstration targets using evaporation. We exceeded the current capacity of the evaporator and blew two holes in the evaporator base plate (actually melted the seals). We fixed the seals and have now vowed to make these targets by molecular plating. We are constructing a special cell to make these large area targets with cooling of the electrodes. A sample cell is shown in Figure 45.

Figure 45: Electroplating cell for Th-232 targets.

A number of attempts were made to evaporate 238UF4 on 30 !g/cm2 C foil backings. The success rate for target construction was only 0.083. The foils broke during evaporation and during pump-out.

An alternate strategy was employed in which the UF4 was evaporated onto the C backing while the backing was still attached to a supporting glass slide. The target backing with the UF4 coating was floated free of the glass slide and picked up on a target frame. The success rate was 1.000. A sample glass slide is shown below in Figure 46.


Figure 46: Glass slide coated with 30 !g/cm2 C foil upon which two circles of

238UF4

have been evaporated.

6LiF was received from Georgia Tech along with a glass scintillator. The glass scintillator arrived in pieces, not having survived transport. In a test experiment, LiF was successfully evaporated onto a glass slide and a piece of the broken scintillator. We await a determination by Georgia Tech group as the thickness of desired LiF coating and a supply of unbroken scintillator.

The background in the alpha spectrometers used to assay the targets was reduced by a factor of 14 by replacing all aluminum components of the counting chambers with components machined from electrolytically refined copper.

An application was made at Oregon State University, defended and granted to prepare targets of “exotic” actinides, 235U (large quantities), 239Pu (large quantities), 240Pu, 242Pu, 244Pu, 243Am, and 248Cm.

Target Fabrication [INL]

An aggressive search has been underway looking for highly enriched actinide isotopes. INL received the enriched actinide isotope shipment from ORNL that included ~1.2 µg of Cf-252, ~1.5 µg of Cm-244 (81.442% enriched), ~3 mg of Am-241

(100% enriched) and ~3 mg Am-243 (99.987% enriched). A shipment containing 99.99% enriched Pu-239 and 99.8% enriched Am-243 is expected from Argonne National Laboratory. Additionally, highly enriched isotopes of U-233, U-238, Pu-244 are expected to be delivered from a facility that is closing at INL.

Target preparation

Six target holders with 1 cm diameter apertures were coated with 30 !g/cm2 carbon foil. Three of these were shipped to LLNL and three were shipped to LANL to test the shipping and handling of these very thin foils. One of the six foils broke during


shipment while five survived intact. This success ratio is consistent with previous experiences. Two of the remaining foils were broken in handling tests at LLNL.

Six new target holders with 2 cm apertures were coated with 30 !g/cm2 carbon foil. Three of these foils along with a 2 cm aperture target holder with a 100 !g/cm2 carbon foil were shipped to LLNL. The targets arrived intact at LLNL.

A dozen 238UF4 targets were prepared on 100 !g/cm2 C foil. The U thicknesses ranged from 150-250 !g/cm2. Work continued on reducing the thickness variations across the targets to better than 0.5 %.

A pair of 238UF4 targets on 100 !g/cm2 C foil was received from INL. These targets had been prepared some time ago with a 20 nm Ti coating on them and each spotted with 5 small spots of 252Cf (see Figure 47). These targets are intended to check the position resolution of the TPC. Upon unpacking the shipping container that held these targets under vacuum, 252Cf contamination was discovered in the shipping container at the level of a few hundred counts per minute above background. The targets were placed in a new clean container, pumped out and allowed to age for a week. Upon opening the new container, similar levels of 252Cf contamination were found. One of the spotted targets was returned to INL for an increase in the thickness of the Ti coating to reduce the self-transfer from the foil before using it in the TPC.

Figure 47: Shown here are two targets with five 252

Cf spots on them.

We received an urgent request from LANL to prepare some large area deposits of 235,238UF4 on stainless steel target backings. We rebuilt the sample holders of the evaporator to accommodate these larger target areas (2.54 cm radius instead of


1.0 cm). We prepared one 238U target and one 235U target. (See Figure 48 and Figure 49.)

Figure 48: 238

U target on stainless steel foil.

Figure 49: 235

U target on stainless steel foil.

Six new 238U targets on 100 ug/cm2 C were prepared and are being stored at OSU. See Figure 50 for a “family” portrait.


Figure 50: Six 238

UF4 targets on 100 µg/cm2 C foils.

Four target frames with apertures ranging from 1.0 to 2.0 cm were coated with 30 !g/cm2 C foil backings. The foils were successfully shipped to the TPC assembly lab at LLNL for further evaluation as regards mechanical stability

We have made five 238UF4 targets on 30 !g/cm2 C foil backings. These targets are being shipped to collaboration members for testing.

Target Preparation [INL]

During this period we chemically purified ~2 mg of 243Am and fabricated a fission target for use at LANSCE. This target along with a 232Th fission target (prepared by OSU) was shipped to and received by LANL, thereby completing an AFCI milestone. Additionally, we dissolved the 252Cf, purchased from ORNL, and have started to produce several fission calibration sources. The sputterer/coater was installed in a hood and the coating of these calibration sources began, using various amounts of titanium. One 252Cf source was alternately fission counted and coated with continually thicker levels of titanium to measure the impact of the titanium confinement layer. The fission spectra are shown in Figure 51. A quartz micro-balance is being calibrated in order to better quantify the absolute thicknesses.


Figure 51: Shown here are 252

Cf fission fragment distributions with various thicknesses of a titanium

confinement layer.

Two stippled PPAC 252Cf (1 !Ci each) calibration sources were fabricated and coated with 100 nm of titanium for use in the DANCE detector. A picture of these targets is shown in Figure 52. The fission spectra through both the titanium containment layer and through the titanium coated Mylar are shown in Figure 53.

Figure 52: Shown here are two 252

Cf DANCE targets prepared with a 100 nm thick titanium

containment layer to mitigate the spread of 252

Cf into the PPAC detector.


Figure 53: Shown here are the fission spectra of the DANCE 252

Cf targets as measured from either

side of the target. As can be seen, the titanium containment layer has less impact on the fission

fragment energy than the Mylar backing, which also has a conductive titanium coating.

TPC Software

Scope

The TPC Software effort is comprehensive and will include online and offline coding, FPGA programming for the data acquisition system, and simulation.

In an extension of the modeling effort, a mock data challenge will be produced. The idea behind the mock data challenge is to run the virtual experiment with enough detail to understand the impact of parameter choices - from collimation to pad layouts. This will entail running the full 4-D model of the experiment in the experimental facility, and the fission as well as background events are scored. These results will be processed through the electronics system by use of a pulse generation system and analyzed by the online software. This model will allow for the optimization of several experimental parameters such as beam flux, collimation, shielding, gas pressure, gas temperature, etc. This will also give the users the ability to study and optimize the design parameters of the TPC and the front-end electronics, including the zero-suppress (sparsification) algorithms.

Overview

In FY09, progress has been made in several important areas of the online software development. The online data packet formats have been implemented to adapt data from the front-end electronics to the receiving computers. Initial event builder


functionality has been developed to accumulate and pack online TPC events for storage in the newly defined TPC raw data format. An online DAQ simulation chain has been implemented for the online system validation and performance testing. Each of these elements will be described in more detail in the following subsections.

TPC Packet Format Design

The TPC data packets will be sent from each FPGA card through the Ethernet switches to computers running software to acknowledge received of packets and build events. Several basic C++ objects have been designed for retrieval and manipulation of data packets either from an input binary file or socket.

TPCEtherDAQDataPacket is designed for making standard Ethernet data packets from the input data stream. It includes the Ethernet MAC address of the packet sender and the intended recipient (The MAC address is also wrapped in the C++ object as the data member object of the TPCEtherDAQDataPacket for easy access and manipulation). It has a fixed packet length that matches the standard Ethernet packet format. The packet capture time is recorded when the packet is received by the computer. Several public methods are defined for accessing and modifying the data members, which includes:

unsigned int GetPacketNumber() const // access packet number

int SetPacketNumber(const unsigned int pnumber)

unsigned int GetBlockNumber() const // access block number

int SetBlockNumber(const unsigned int bnumber)

int AppendData(const std::string edata) // append data string

bool GetACK() const // acknowledge receiving of packet

void SetACK(const bool val)

unsigned int GetTimeSeconds() const // access capture time

unsigned int GetTimeMicroseconds() const

void SetTimeSeconds(unsigned int sec)

void SetTimeMicroseconds(unsigned int usec)

TPCChannelDataPacket is designed for storing the data from each ADC channel. The data members consist of 5 bytes fast time stamp for recording the event happen time for up to 4 hours time period; 4 bytes packet type for recording different type of packet, such as ADC, buffer, busy, housekeeping or temperature; 4 bytes channel number to represent the data origin (e.g., which ADC channel on the FPGA card) and the data body in string format. The size of the data body varies and depends on the trigger setup and the look back samples for each triggered event. The FPGA card number is also recorded and to be saved in the raw data for offline analysis usage. The public functions include:

uint64_t GetFastTime() const //access fast timer

void SetTimeStamp(const unsigned int starttimehigh, const unsigned int starttimelow);

unsigned int GetPacketType() const // access packet type


void SetPacketType(const unsigned int type)

unsigned int GetChannelNumber() const // access channel number

void SetChannelNumber(const unsigned int channel)

unsigned int GetSize() const // access packet size

int GetCardID() const // access FPGA card number

void SetCardID(const int card)

TPCEthernetAddr is the Ethernet address object for storing the Ethernet MAC address and converting it into different formats.

Event Builder Development

The critical part of the online system is the event builder (EvB). The EvB takes data from the packet receivers (one or multiple computers to collect the data packets from the front end electronics) and assembles the data into events based on recorded timestamps. The data fragments from a physics event (spontaneous fission, neutron induced fission, alpha decay, etc.) are distributed over numerous front-end EtherDAQ cards. The EvB is responsible for correlating and collating all associated fragments into complete event records. It also packages the events to NIFFTE standard format and passes them to online storage components.

The main processes of the EvB are:

• Interface with packet receiver(s) to receive unique sets of EtherDAQ data packets (EDPs) and sort them by MAC address (this corresponds to EtherDAQ card number) and packet number (sequence information). The channel data packets (CDPs) from sets of EDPs will be assembled and lined up in sequence.

• Group assembled CDPs into events or event fragments based on an event time window (use internal 40-bit 65 MHz fast timestamps, and wall clock time to resolve issues related to # 4 hour fast timestamp rollover). Keep track of the state of incoming data and close finished event fragments.

• Pass events to online storage and stream a fraction of assembled events to an analyzer for data quality monitoring or real-time online analysis.

An overview of the EvB functionality is shown Figure 54 and Figure 55. High data rate capability will require the online data assembly system to support parallelization for horizontal scalability (e.g., scale from one event builder process on one computer building events to multiple processes on multiple computers). It will also allow other post processing of the event stream by an event rebuilder (e.g., to split compound events, combine adjacent partial events, or add extremely late data to an event).


Figure 54: Conceptual design of the data flow at the Event Builder stage.

Figure 55: Flow chart of the Event Builder design.


The development of the EvB system has been broken into several stages, with Phase one completed in FY09, Phase two currently in progress and Phase three to begin in the early part of FY10.

• Phase one: [Complete] A basic reusable EvB library for performing CDP reassembly and EDP sorting tasks has been developed. The initial prototype only supports CDPs with ADC channel data directly. For example, all CDP types will be saved, but any specific support for passing temperature and housekeeping information to other systems will be postponed. A performance benchmark on common baseline CPU has been used to test the efficiency of the algorithms. Single threaded implementation of the event process is used and the completed event stream is dumped to data disk.

• Phase two: [In progress] Extend EvB library codes to improve performance (re-implement any component that is considered a significant bottleneck). Extend the event builder to multiple processes and threads with an inter-process communication scheme that allows high performance on a single machine and good performance over the network (e.g., Unix Domain Sockets for local and TCP/IP for remote). The interface of the EvB with the packet receiver is being implemented. A simulation using data from the offline software has been developed to estimate the EvB performance relative to the ability to process Californium spontaneous fission and alpha decay events in real time.

• Phase three: [Early FY10] Improve realism of library code to accurately match incoming data (remove any consequential simplifications present from previous phases). Complete data logging and storage tasks. The interface of EvB with online control and monitor networks will be implemented.

The initial reference implementation (Phase one) event builder performance is shown in

Table 2. The Ethernet packet files are generated with 10 thousand fission events (each contains two fission fragments and one alpha) in various event per trigger sizes. The computer wall time shows the used CPU time. The input packet data parsing and event assembly rates are also shown.

Table 2: Event building performance test for reference implementation.

lookback (samples)

Ethernet packet file size (MB)

Wall time (seconds)

Input Data Rate (MB/s)

Event rate (Events/s)

8 359.46 17.961 20.0 556.8

16 632.36 21.724 29.1 460.3

32 1177.58 28.644 41.1 349.1


TPC Raw Data File Format

The raw data file consists of file header, data block (block header and event block) and end-of-file (EOF) block. The file header and EOF block will be of fixed size (subject to change from version to version). The file header will record common information for the data taking, such as run number, data taking dates, data size and the comments from shifts. The EOF block is designed to have an additional handle on data integrity. Data blocks are inherently of variable size to reduce the amount of disk space usage. The ADC raw data is written in the event block in chunks of time specified in the event builder. The trigger, housekeeping and buffer status and TPC temperature data are written in the other blocks with different block ID to be distinguished from the event block. Each data block consists of a block header to specify the data type and the version information.

Figure 56: TPC raw data file format (units of bytes).

Data Acquisition Simulation Chain

In order to develop and test the event builder and other major online components, a full DAQ simulation chain has been implemented. The procedures consist of:

• Generate fission events from Geant4 simulation package

• Run through NIFFTE detector simulation software to make offline event digits

• Convert the offline digits to Ethernet packets in the same format as the FPGA based front-end data acquisition cards.


• Read the Ethernet packets and build online events which contain the data samples recorded by the TPC within a time interval.

• Write out the events in the NIFFTE raw data file format

The simulated data packets are generated from offline digits and saved in the libpcap format (http://wiki.wireshark.org/Development/LibpcapFileFormat). The libpcap format is a simple standard file format commonly used for network captures on UNIX systems. The file format has a global header containing information common for all of the captured packets, e.g. timezone, format version number and maximum packet size; then it is followed by each captured packet with a certain packet header defined to save the capture time and actual packet length. The packet format is shown in Figure 57.

Global

Header

Packet 1

Header

Packet 1

Data

Packet 2

Header

Packet 2

Data ...

Packet N

Header

Packet N

Data

Figure 57: The libpcap packet format.

The file header and packet header structures are shown below. struct tpc_pcap_file_header { uint32_t magic; //= 0xa1b2c3d4; used for byte order checks uint16_t version_major; //= 2; uint16_t version_minor; //= 4; int32_t thiszone; //= 0; timezone offset, use GMT uint32_t sigfigs; //= 0; uint32_t snaplen; //= 65535; maximum packet size uint32_t linktype; //= 1; Ethernet };

struct tpc_pcap_file_pkthdr { uint32_t sec; // seconds since 1970 uint32_t usec; // microseconds since start of current second uint32_t caplen; // size (byte) of captured packet data that follows this header (e.g., for partial captures) uint32_t len; // actual size (bytes) of packet on the wire };

The digits from offline simulation record the energy depositions in the unit volume of the TPC. The simulation does not represent the energy to charge conversion, signal amplification and digitization processes. The processes such as preamp response, 1/f noise, pedestals, and EtherDAQ digital semi-Gaussian filter and trigger have been primarily implemented to simulate the real signal generation. Figure 58 shows the signal output through those processes. Since the current NIFFTE preamps have a nominal RC time constant of # 620 !s (which is much longer than the event length of # 5 !s considering the drift distance = 5.4 cm, drift velocity # 1 cm/!s), the decay time


of the preamps is large which makes the signal pile-up. A pile-up correction and ballistic deficit correction will also be implemented to represent the real operating conditions.

Figure 58: Pulse charge input (top), signal after preamplifier with 1/f noise (middle) and semi-

Gaussian filter and trigger output (bottom).

Offline Software [CalPoly, LLNL, ACU]

The thrust of this task is to transform the data from their raw form to final calibrated results, which requires a complete data analysis chain. The online software required to perform this analysis must be designed, organized, written and documented. In order to achieve maximum flexibility, the design should focus on providing simple interfaces within a modular framework. For ease of use by collaborating experimenters, the software should also be well documented and maintained in a central repository available to the entire collaboration.

Overview

In FY09, offline software made several key advances that enabled the Mock Data Challenge exercise to run successfully and advanced our preparedness for analyzing real data. The software repository was re-organized and the entire framework was redesigned to allow detector response simulation, reconstruction and analysis


algorithms to run within the same executable. The NIFFTETaskManager was completely overhauled to allow serial processing of multiple events within one executable session.

Monte Carlo truth information from simulated particles was added to the offline data file format and the TPCDigits and TPCHits were augmented with references to their MC truth information. The transformation of the first detector simulation code to the NIFFTE framework was completed and the TPC geometry definition was ported to the ROOT library’s TGeo classes in preparation for the eventual transition to the ROOT Virtual Monte Carlo paradigm. New classes for defining the specific TPC geometry outside of GEANT 4, accessible by the TPC library, were also created.

One of the most stunning new developments from FY09 is the development of the NIFFTE event display, which allows visualization of the detector geometry and fission events in ROOT’s OpenGL-based TEve classes. In addition, the NIFFTE Virtual Machine, a complete linux distribution with all of the NIFFTE software libraries pre-installed, was created to facilitate the use of the NIFFTE offline software and event display by new users. The NiffteVM and instructions for using it are available online at http://nuclear.calpoly.edu/~niffte/NiffteVM/. Finally, a software manual covering all of the features of the NIFFTE framework and installation procedures was developed and is available as part of the subversion repository. All of these updates will be described in more detail below.

Reorganization

In order to better accommodate the expected use cases for NIFFTE software in online, offline and simulation contexts, the subversion repository was completely reorganized in FY09. The new structure recognizes the fact that all code that is not explicitly ‘online' is inherently ‘offline' code. The reorganization has enabled more flexibility in the software. In particular it has made the TPC library more independent, ensuring the ease of future re-use for other projects. Details on the implementation of the detector response simulation modules will be discussed in the Simulation section, while details regarding the TPC reconstruction library are included here.

Task Management

As part of the code re-organization and improvements to the independence of the TPC library, some functionality from the NIFFTE framework was abstracted to base classes in the TPC library. TPCTaskConfig and TPCTaskManager ($NIFFTE/tpc/base/) now handle the XML parsing of the configuration file and the management of any of the three computing tasks: analysis, detector simulation or reconstruction selected in the configuration file. The base class owns the transient data objects that are passed between the modules and configures and runs the requested modules. In order to create a working executable, a project must derive an inherited class from this base class and overload the I/O methods to retrieve and store persistent data. For NIFFTE, the class NiffteTaskManager, located in $NIFFTE/common/framework/, performs this function and is the object instantiated in the $NIFFTE/common/framework/main.cc file.

Online to offline channel mapping

The electronics channels in the online data acquisition system are addressed according to their assignment to specific readout hardware. This numbering scheme


is appropriate for the physical detector but is not necessarily the most convenient for use in offline software. A more user-friendly numbering system that is visually simpler and generalized for use with any TPC geometry was devised for offline software. The mapping to transform one to the other was also developed. Figure 59 shows the schematic transformation of online to offline numbering. The offline system utilizes rectilinear coordinates (row,column,bucket) to address adjacent channels. The hexagonal geometry of the NIFFTE TPC fits inside a larger square with unused channels in the four corners. The first potentially non-empty channel in the column direction occurs at (0,16), while the first potentially non-empty channel in the row direction is at (31,0). Since the vector of digits to be reconstructed will be filled with only non-zero (after zero-suppression of real channels with data below threshold) locations, the unused channels will never appear in offline data.

Figure 59: Shown here is the mapping of online electronics channel numbering to offline software

indexing for the NIFFTE hexagonal geometry.

Depending on the orientation of the hexagonal pads, adjacent pads in one of the rectilinear directions will be offset with respect to one another. Therefore a convention for the location of the zeroth elements in the (row,column) directions must be adopted. In the NIFFTE TPC, the hexels are oriented such that neighboring hexels in the horizontal direction share a straight side while neighboring cells in the vertical direction are staggered with respect to one another. In Figure 60 the numbering convention adopted for NIFFTE is illustrated. The first channel in row 1 is south-east of the first channel in row 0 while the first channel in row 2 is south-west of it. The bucket numbering starts at the anode (pad plane) and increases toward the target. This is true for both sides of the TPC. The “upstream'” side (where the neutron beam enters) is volume 0, while the “downstream” side (where the beam exits) is volume 1.


Figure 60: Shown here is the offline channel numbering convention for the NIFFTE TPC geometry.

Monte Carlo hit information

The base class TPCMCHit was introduced to store information from physics simulation (not tied to the originating MC generator) that can be passed to the detector response modules to produce the final digits. It is meant to be a generic container for energy deposited by a simulated track within a voxel of any TPC. The names of the data members were chosen accordingly and are described below. In this model, there can be more than one TPCMCHit per voxel if there are multiple MC tracks that deposit energy into it. Figure 61 shows the elements of an MC hit for the NIFFTE hexagonal TPC geometry.

Figure 61: Visual explanation of Monte Carlo hit indexing for a hexagonal TPC geometry.


int fVolume; // Side of target in TPC drift volume

int fColumn; // Column of drift volume between target and plane

int fSlice; // Location in drift direction within column

int fDivision; // Division within a (column, slice) pair

//(NIFFTE: 0=central hex,1-6=boundary trapezoids)

double fEdep; // Energy deposited in division by MC particle

int fTrackId; // Track id from simulation that created this hit

The fDivision data member allows one to subdivide a voxel into a central area and border regions that can be used to determine how charge is shared among neighboring channels due to charge diffusion during drift. A corresponding persistent data class, NiffteMCHit, was added to the dataIO directory for reading/writing these objects on disk.

Monte Carlo truth association

A new typedef has been included in the TPCConstants.h header file that holds the track ID and the weight of a simulated track's contribution to a given reconstructed object:

/*! \brief typedef for simulation info - int for Id, double for weight */

typedef std::pair<int, double> TPCSimTrackIdWgt;

The TPCDigit and TPCHit classes (as well as their corresponding persistent twins NiffteTPCDigit and NiffteTPCHit) contain data members and accessor/modifier methods to hold information about Monte Carlo particles which have contributed to them. They store a vector of the typedef'd TPCSimTrackIdWgt objects:

/*! \brief For simulated digits/hits only, Track ID and weight

of each simulated track's contribution to digit */

std::vector< TPCSimTrackIdWgt > fSimInfo;

If the vector size is zero, then the digit/hit does not come from simulation. If it is non-zero, the elements contain info on which Monte Carlo tracks and how much they contributed to the given digit or hit. If the sum of the weights adds up to less than unity, the remaining amount should be considered to come from “real” data. This model allows the possibility of using “embedding” for correction studies, in which simulated tracks are embedded into real data and then the embedded files are reconstructed using the standard tools. An association algorithm can then be run to determine efficiencies, etc.


Monte Carlo kinematics information

NiffteMCParticle, which inherits from ROOT's TParticle class, defines persistent data objects to hold the kinematic information about simulated particles. The only additional data member they contain beyond those in the TParticle class is fTrackId, which is their identifier for associating them with reconstructed tracks.

ROOT uses the Particle Data Group’s standard particle numbering scheme to uniquely identify particles of different types, but this database does not by default include nuclei above deuterium. Additional particles can be defined and added to the database as needed or the database can be initialized with a file containing a list of additional particles. The file $NIFFTE/common/utils/nuclei_pdg.txt defines a long list of nuclei compiled from the list of known isotopes in the RIPL-2 database. This file is read by niffte executable and can also be loaded in a ROOT session to retrieve the Z,A, and mass of all fission fragments that might appear in our data stream.

TPC Reconstruction Library

Two reconstruction models were developed for NIFFTE. One utilizes the Hough transform for track-finding followed by the Kalman filter for track fitting, while the other implements a traditional two-dimensional cluster and hit-finder with a subsequent “follow-your-nose” track finder and straight-line least-squares minimization track fitter. The components of these parallel methods are somewhat interchangeable. For example, one can use the follow-your-nose finder with the Kalman fitter and vice versa. The ultimate goal is to evaluate both methods on simulated data and capitalize on their strengths under various conditions for real data. Having multiple reconstruction algorithms is also essential for estimating systematic uncertainties on the final output, especially critical for the high precision results expected from NIFFTE.

Both models output a set of track parameters to uniquely specify the trajectory of a given track. Since the NIFFTE TPC is not operated with an external applied magnetic field, the track trajectories should be approximately straight lines. The straight-line parameterization used in NIFFTE is shown in Figure 62. Two parameters, the perpendicular bisector distance, $, from the origin to the track, and the angle % from the positive x-axis to the bisector are used. In order to represent the track in three dimensions, these two parameters must be specified in the two planes x-z and y-z. This model avoids the problem of infinities when dealing with large slopes that is inherent in a y = mx + b track parameterization.

Figure 62: Straight-line track parameterization in the x-y plane.


The TPC modules that make up the two reconstruction algorithms will be listed and described briefly, including their configurable parameters.

TPCClusterFinder looks for clusters of adjacent digits in each bucket plane of two-dimensional pads and columns by employing a nearest neighbor search algorithm.

• Threshold (int) – the value in ADC counts below which a digit should not be included in a cluster, default value = 1

• Geometry (string) – specifies which pad plane geometry to use for nearest neighbor searching, default value = “hex”

• MinNumOfDigits (int) – minimum number of digits for a cluster to be kept, default value = 2

TPCHitFinder implements deconvolution of 2-D clusters into peaks and Gaussian fitting of the peaks to determine the x,y position and energy deposition of hits in the z-direction. Two modes are currently available – with and without multiple hit deconvolution, however the deconvolution algorithm requires additional tuning and development to work, so it is turned off by default.

• PeakADCThreshold (int) – threshold value in units of ADC for isolating peaks within a cluster, default value = 4

• PeakRadius (int) – how far in units of row, column around a given peak to include digits in a peak for Gaussian fitting, default value = 1

• HitFindMode (string) – two options ‘Simple’ and ‘Standard’. Standard performs both deconvolution and peak fitting. Simple takes the center of gravity of the cluster as the position and the sum of all ADC values as the total charge, default value is “Simple”

TPCTrackFinder employs a “follow-your-nose” algorithm to identify tracks. Starting from the least dense regions of the TPC in either volume, near the pad planes, it takes the outermost hit and iteratively predicts the expected position of the hit in the next adjacent bucket-plane according to the slope of the line through all the hits in the track and the search radius window. The initial slope is zero but is recalculated at each step.

• LeeWay (double) – controls how much separated two adjacent hits can be to be included in one track, default value = 0.2 cm

• SearchRadius (double) – defines the starting search radius for seeking hits, default value = 10. Cm

• NarrowRatio (double) – by what fraction the search radius will be narrowed as each subsequent bucket-plane is traversed, default value = 0.9, corresponding to a narrowing to 90% of the previous search radius

An example configuration file for running the standard TPC reconstruction chain called configStd.xml is included in the NIFFTE/build directory. Complete testing of the TPCTrackFinder and implementation of a simple TPCTrackFitter module are currently underway. Improvements to the TPCTrackFinder that will also be developed include merging broken tracks, splitting incorrectly merged tracks, reconsideration of orphan hits after the first pass of track-finding and widening of the search radius if no tracks are found in the first pass.


TPCHough/TPCHoughD implements a Hough transform algorithm to identify tracks. It works on either digits (TPCHoughD) or reconstructed hits (TPCHough) from the TPCHitFinder. Each point (digit or hit) is discretized into ($, %) curves in two planes (starting with y,z) by drawing all possible straight lines through each point and filling a histogram in ($, %). The bin with the highest occupancy corresponds to the most probable track from the list of input points. The points contributing to that bin are removed and the algorithm repeats for the remaining hits. An example of the frequency distributions from a simulated event with three tracks is shown in Figure 63.

• HoughRadius (double) - first Hough radius used to pick tracks in the y-z plane, default value = 0.6 cm

• HoughRadius2 (double) – second Hough radius to pick tracks in the x-z plane, default value = 0.4 cm

Figure 63: Hough Transform frequency distributions. The bins with the highest occupancy (circled

in red) represent the values of rho and theta for a highly probable track. Points can be iteratively

removed from the high occupancy bins and assigned to tracks, with remaining points re-used for

further track finding.

TPCKalmanFilter is a recursive track state estimator based on the well-characterized Kalman filter method to determine the best line fit parameters to a set of measured track points. The state vector is updated after each recursion and the error matrices evolve at each step, providing the confidence interval and &2 contribution for each measurement. The algorithm proceeds in three steps: predict, filter and smooth. Visual representations of the algorithm are shown in Figure 64 and Figure 65.

• BlowupFactor (double) – Rescales the initial error matrices, default value = 2.0

• ProcessNoise (double) – process noise assumed from multiple scattering, default value = 0.0

•


Figure 64: An example of the three steps (prediction, filtering, smoothing) involved in the Kalman

Filter method of track fitting.

Figure 65: Illustration of Kalman Filter operation. From the initially large uncertainty (green

cones) each step in the Kalman Filter improves knowledge of the trajectory (blue cones). As each

subsequent step is added, the track reconstruction rapidly converges.

An example configuration file called configHough.xml is available in the $NIFFTE/build directory for running the TPCHoughD and TPCKalmanFilter algorithms. The performance of the Hough Transform + Kalman Filter alorithms have been evaluated using data generated during the Mock Data Challenge. The results


are shown in Figure 66. The top panel shows the difference absolute error on the reconstructed theta angle vs. the angle from a large event sample. The bottom panel show the “pull” distribution, or difference between reconstructed and MC theta angle in the y-z plane. The fact that the distribution is centered at zero with a width of one indicates that even prior to any tuning, the algorithm is fully capable of reproducing the correct track parameters for simulated NIFFTE data.

Figure 66: Absolute error vs. angle (top) and deviation of reconstructed theta angle from MC angle

or "pull" (bottom) for NIFFTE Hough Transform + Kalman Filter algorithms.


In addition to the Hough Transform track finder, another algorithm based on the Binary Space partitioning algorithm has been developed and will be incorporated into the TPC reconstruction library soon. Beyond the identification of tracks in the TPC, a TPCVertex algorithm is also under development to combine tracks coming from the same location within the target. This algorithm will help better identify correlated fission fragments and ultimately allow detailed investigation of fragment emission.

The progress on the reconstruction code during FY09 in preparation for the Mock Data Challenge exercise has placed us in excellent position to be able to evaluate in detail the TPC performance with a reasonably mature and realistic event generator, particle transport and detector simulation plus multiple reconstruction algorithms. These modules will continue to develop and be improved during FY10 as we prepare for first data from the TPC data.

Geometry Conversion and Event Display

The NIFFTE detector geometry description is currently maintained in two places, within the GEANT4 simulation code and using ROOT’s TGeo classes in $NIFFTE/common/geometry/NiffteTPCDetectorConstruction. Both utilize the class $NIFFTE/common/geometry/NiffteGeometryMap for setting constants and NIFFTE specific geometry mappings. Eventually the GEANT4 geometry will be removed once the transition to full ROOT Virtual Monte Carlo model is completed. The development of the NiffteTPCDetectorConstruction class was necessary in order to utilize the ROOT TEve package for visualization. Although with two geometries there exist possibilities for inconsistencies, these are minimized by the use of NiffteGeometryMap, where key parameter values can be set and retrieved.

The origin of the NIFFTE coordinate system has been placed at the center of the TPC, within the target. Figure 67 shows the coordinate axes superposed on half of the detector using the new NIFFTE event display. When the TPC is placed in the beamline at LANSCE, the neutron beams will exit the detector on the positive z-side (right side of Figure 67).

Figure 67: Here is the TPC with pad plane detail from NIFFTE Event Display tool utilizing ROOT's

TEve and TGeo packages with OpenGL graphics.


The NIFFTE event display is based on TEve, ROOT’s OpenGL visualization tool, and TGeo, ROOT’s 3D geometric modeling package. The level of detail seen, in Figure 67 and Figure 68, can be adjusted at the command line. The image can also be zoomed, rotated, panned, scanned and otherwise manipulated with the mouse. In Figure 68 the pad plane detail is turned on to illustrate the positioning of the readout sector divisions. Each colored area represents the channels read out by separate front-end boards. The honeycomb shape of the individual pads can also be seen.

Figure 68: NIFFTE TPC with pad plane detail. The different colors identify the channels that are

common to a particular hardware read-out card.

Beyond visualization of the detector geometry, the event display’s true power lies in its ability to visualize data events and allow the user to manipulate the information. Figure 69 shows the output from a simulated event with the geometry detail turned off to highlight the data. On the left-hand side of the window, the user can select which data elements to plot. In this case, NiffteTPCDigits are shown in black while the trajectories of the simulated particles are shown with the red lines. This event contains four alphas and two fission fragments plus detector noise. Some aspects of the NIFFTE event display are still under development; such as the ability to click on a reconstructed track and have a histogram of the energy deposition vs. track length appear. We expect to include several such improvements in FY10.


Figure 69: NIFFTE Event Display showing the energy deposited by two fragments and four alphas

plus random noise.

NIFFTE Virtual Machine

It was realized at the Summer MDC workshop in Los Alamos that the burden of configuring new machines for running NIFFTE software could be lifted for the end user by distributing a single virtual machine (VM) image file with all of the external libraries and NIFFTE software pre-installed. Upon investigation and discussion, it was decided that VMware plus Fedora linux was the best choice for virtualization since the VMware Player is a free application that anyone using Linux or Windows can download to run the NIFFTE VM. For Mac users, a relatively inexpensive ($40 academic license) alternative called VMware Fusion is available.

This solution was implemented and tested following the MDC workshop and found to produce a small performance penalty compared to running natively, which varies depending on the hardware on which it is run. The NIFFTE VM will be useful for those users who wish to do small analysis and visualization tasks on their desktop, but it remains to be seen whether the graphics acceleration is sufficient for large event visualization.


Already the usefulness of this distribution model is apparent. In August of 2009, representatives from the Active-Target TPC collaboration at Michigan State University were provided with the NIFFTE Virtual Machine in order to facilitate collaboration on the TPC software library. They were able to very quickly adapt the NIFFTE software to their detector geometry and begin evaluating the framework for their experiment. We expect fruitful collaboration on the generic TPC library will occur in FY10.

NIFFTE Software Manual

A complete software manual for the NIFFTE framework, including detailed installation instructions was prepared by a group of CalPoly students as the final project of their summer technical writing course. This manual is available as part of the subversion repository and will be maintained and updated as the software matures.

Data Acquisition Software [ACU, LLNL]

This effort will develop all the software required to control the TPC experiment and log the data. An experiment control interface will be developed to allow collaborators to run and monitor the experiment from remote sites, including a slow control system with appropriate interfaces. The front-end cards for the TPC will be quite powerful and flexible because of the Field Programmable Gate Arrays (FPGA). The FPGAs do require programming which we will organize in a framework of modules (each module representing one task) for easy reconfiguration of the device. The modules that would be written for the FPGA would include (1) an ADC receiver that interfaces with the ADC chip, sending and receiving clock signals, receiving the serial data and presenting the data in a pipeline for the next module, (2) preprocessing modules would work with the data before zero suppression, and would include functions such as, ballistic deficit correction, fast proton timing, rebinning, and digital shaping.

Overview

Work tasks completed for reading out the data once it exits the FPGA cards is described in the Online Software section. This part of the DAQ system is integral part of the online software.

Slow Control System [ACU, LANL]

The slow control system for NIFFTE will be controlled and read out via Ethernet and the MIDAS DAQ framework. Most of the readout modules have been purchased from IOtech. These include various modules such as thermocouple inputs, digital I/O, strain gauge readout, differential ADCs for different input voltage ranges, and relay control modules. Among the items completed in FY09 are:

• The IOtech slow controls are now set up in the experimental area at LANSCE.

• Continued preparing documentation of the IOtech DaqScan/2001 device for use in the DAQ auxiliary system.

• Continued work on modeling the electronics part of the detector; zero suppress algorithm (trigger module) and lookback + pedestal noise.

• Created device and class driver for IOTech DBK90 (temperature reading) via MIDAS.

• Put a parser into the device driver to get rid of headers from IOTech Ethernet


packets. • Added conversion function from raw IOTech temperature data into usable

temperature data. • Created data logger (mlogger) to work with raw and real IOTech temperature

data.

MIDAS Development

The software for the DAQ system is based on MIDAS running over Ethernet on a LINUX based system. The MIDAS driver for the CMC100 CAMAC controller was committed to the collaboration's SVN repository. At its current state, the driver is already complete enough to read out 32 ADC channels from two CMC081 charge integrating ADC CAMAC modules through MIDAS. The driver implements basic functionality needed for the preamplifier card test stand. An example front-end program and ROOTANA based analyzer have also been developed and included. A DAQ system configuration document has been created to assist in the verification that all of the different online systems are brought together in a complete and seamless system. The Slow Control system is being integrated into the MIDAS Ethernet based online software to avoid dealing with a separate path for the control and monitoring of the experiment.

The DAQ software system has also been installed for a preamplifier test stand. A computer has been sent to LANL with the necessary operating system and DAQ libraries. This has demonstrated the flexibility of the MIDAS based Ethernet online software for use in testing and development the various components that must come together for NIFFTE.

Simulation [GIT, LANL, CalPoly, ACU]

In order completely understand how the TPC will respond to various neutron environments and to accurately determine the fission parameters of uranium and the minor actinides, a complex simulation effort will be undertaken. The environments that the TPC will be used in will require accurate modeling of the detector systems used as well as the neutron production. MCNPX will be used to model the experimental setup at both the LANSCE, the quasi-monoenergetic neutron source at LLNL and Ohio University mono-energetic experimental facilities. These fully detailed four-dimensional models (3D space and time) will be used to create the source term for the GEANT4 modeling of the detector itself. Since MCNPX does not have the ability to transport heavy fission fragments, GEANT has been selected for this task. GEANT only has data for uranium fission events in the G4NDL library and the data for the remaining fissionable isotopes is based on a low precision neutron yield model. GEANT will need to be modified to use the Los Alamos model, also know as the Madland-Nix model, in which fission data will be added for U-238 and Pu-239 and the minor actinides. The modified GEANT module will allow the user to select the Los Alamos model or a fission distribution file supplied by the user. The fission fragmentation model will also be added to this module. To allow for a full model of the detector, another GEANT modification will be the addition of a static electric field modeling capability. This module will be used to accurately model the gas electron amplification inside the detector system. This will allow GEANT to completely model the detection system from birth (through MCNPX) to charge collection in the TPC pads. Using the high fidelity models of the experimental setup facilities, a series of


databases will be created for various isotopes. This will allow for rapid comparison with experimental data.

Overview

In FY09 we completed several milestones related to the simulations. In preparation for the Mock Data Challenge, a number of improvements and expansions were created for the simulation software. The fission generator was retooled to allow more flexible event generation and the resulting output of the GEANT4 particle transport is now written to ROOT files using the NIFFTE I/O classes. The TPC detector simulation library was developed and tested with new modules to handle the various physical effects that need to be modeled. Monte Carlo simulated TPC data can now be converted to raw TPC data processing Monte Carlo generated data through the complete offline software chain. An overview of the full NIFFTE simulation chain is shown in Figure 70, with the yellow-highlighted elements indicating those that were completed for the Mock Data Challenge. In addition to the TPC simulation, two neutron beam collimator designs were also simulated and evaluated to estimate the backgrounds from room return neutrons. Each of these elements of the simulation effort will be described in more detail in the following sections.

Figure 70: Overview of the NIFFTE simulation and reconstruction chain. Elements highlighted in

yellow will be included in the FY2009 Mock Data Challenge.


Fission Generator

The Fission generator portion of the simulation, written in JAVA, samples the fragment probabilities from JENDL cross-section libraries and the neutron and gamma spectra with the correct energy balance. The angular distributions are carried forward and an angular correction of the two fission fragments is made. This module can run in either spontaneous fission or neutron induced fission mode. An example of the fragment probability distributions for Cf-252 spontaneous fission and 5 MeV neutron-induced fission of U-235, U-238 and Pu-239 are shown is in Figure 71.

Figure 71: Fission fragment probability distributions from JENDL for Cf-252 spontaneous fission

and 5 MeV neutron-induced fission of U-235, U-238 and Pu-239.

Ternary fission, along with an unphysical option that produces multiple alpha particle emission with binary fission is also included in the simulator. Two modes of operation are possible. The first is accurate sampling in which ternary fission is sampled according to the data provided in the libraries. This method strictly samples the energies, momenta and mass splits of the fragments according to the library data. If the user desires unphysical events with more alphas to be emitted per fission event, the user may select either the correct or forced sampling algorithm (this option is valuable in tracking resolution and occupancy studies). If the “correct” sampling option is chosen, mass and energy splits are calculated from the fission libraries. This can result in artificial and unrealistic atomic numbers and masses for the fission fragments that are far from the neutron drip line. The user is encouraged to use the


“forced” sampling algorithm that ensures that the two fission fragments have realistic atomic numbers and masses (generated as a pair) but breaks the correlation of mass and energy. In this mode, alpha particles are generated in addition to the fission event.

In the end, a correctly sampled fission fragment pair along with neutrons and gamma rays are sampled and written to a file suitable for input into GEANT4. The code used to generate these files is located in the repository in the sim/tpc/TPCSimNiffteIO/data directory. For flexibility, the code also has the ability to override the basic physics of the problem. Special parameters can be set on the command line to produce specific effects for testing the fission tracking codes. These include the suppression of ternary fission, sampling ternary fission event-by-event, or the unphysical production of some number of alphas with a binary fission event. The basic physics of the problem are maintained but can produce user desirable properties for testing purposes.

New experimental fission data for fragments, energy and their correlation have been obtained from Los Alamos National Lab. This new data has a 121 x 121 element fragment mass as a function of total kinetic energy (TKE) matrix of the most recent evaluations for 1 and 6 MeV incident neutrons. This data will be interpolated with fission fragment/energy models until more data becomes available. Correlated “sawtooth” neutron multiplicities and energies versus fragment yield are also being prepared for integration into the generator. Once these elements are complete, the standalone fission generator will be incorporated into the GEANT4 TPC active volume simulator (see below).

On a parallel path, another simplified standalone GEANT4 simulation has also been written which includes a fission generator capable of simulating Cf-252 fission fragments. A pair of back-to-back fragments is generated randomly on the surface of a disk with the direction of the fragments also chosen randomly. Momentum is split between the fragments according to conservation laws. Although the model does not include the emission of neutrons and gammas, the output for the fission fragments will be used in conjunction with the standard fission generator for cross-checking and validation purposes.

GEANT4 Active TPC Volume Simulation

The output of the fission generator forms the input to the GEANT4 active TPC volume simulation. The GEANT4 simulator can read in an ASCII text file containing a list of events with particles and their starting momentum components from the fission generator (or any desired model) and feed that to the G4 simulation. The input data file is formatted such that the first line indicates the number of particles in the event, followed by that number of lines specifying the A,Z of the particle and the px, py, pz of the starting momentum, etc. for as many events as the user desires.

Maintenance of two separate programs for the fission generator and the active volume simulation, in two separate programming languages, is an undesirable long-term situation for the software. In FY10, the standalone fission generator will be integrated into a GEANT4 fission physics package that correctly handles the fragments as well as the neutron and gamma emission. This package will then be used together with the active volume simulation to create fission events and transport them through the detector volume in one go. We also plan to convert the standalone


GEANT4 application currently used into a NIFFTE framework based extension of ROOT’s Virtual Monte Carlo model. Simultaneously, the detector geometry will be fully transitioned to the ROOT TGeo classes and the geometric detector model used in a given simulation will be stored persistently with the data in NIFFTE ROOT files.

The G4 simulation uses standard GEANT4 transport physics to accumulate the energy deposited by the simulated particles into each volume element of the TPC. The output of the active volume simulation is a NIFFTE ROOT file containing the list of Monte Carlo Particles (NiffteMCParticles) and the list of TPCMCHits described in the offline section. This output data can then be fed into the modules comprising the TPC Detector Simulation Library to simulate the detector response.

TPC Detector Simulation (‘DetSim’) Library

The TPC detector simulation ‘DetSim’ library is included as a set of configurable modules within the generic TPC library. The user can specify with an XML configuration file which modules to run and with what parameters. The list of available modules, a brief description and the configurable parameters associated to each one follows.

TPCLatching simulates the asynchronous timing between ADCs and the fission event. It randomly selects a time-offset of between zero and ten bucket divisions and then sums all subsequent bucket divisions into groups of ten to create the ’buckets’ that are the fundamental unit of a hexel digit in the z-direction.

• NumberOfClocks (int) - how many independent clocks that run the TPC, default value = 1

TPCDiffusion models the diffusion of the ionization clouds as they drift toward the pad plane. To handle lateral diffusion, the NiffteMCHits output by GEANT are subdivided into a central hexagonal zone and six bordering trapezoids. Charge deposited within the bordering trapezoids is shared with neighboring hexels while that within the central hexagonal subdivision is not. The thickness of the trapezoidal area and the amount of sharing is dictated by the diffusion calculation and can be tuned appropriately. Longitudinal diffusion is also included.

• GeometryUsing (string) - which pad plane geometry is implemented, default value = “hex”

TPCChargeSharing simulates the cross-talk and charge sharing between neighboring pads during readout summing.

• GeometryUsing (string) - which pad plane geometry is implemented, default value = “hex”

• SharedFractionPerNeighbor (double) - fraction of charge or energy deposition to be capacitively shared with each neighboring pad, default value = 0.02

TPCDigitizer converts the oversampled TPCMCHits into TPCDigits by grouping all TPCMCHits for a given time-slice into one digit bucket. The electronic gain will be adjustable by changing the voltages on the preamps, but this has not yet been modeled. For now, a simple conversion factor of 1 ADC = 1.0e'09 GeV is used.

• No parameters currently configurable

TPCRandomDigits models electronic noise in the detector by randomly selecting hexels with no signal in them and adding a Gaussian-distributed amount of noise


in ADC values to them. One can specify the number of noise digits to include or allow the noise multiplicity to also be randomly selected.

• Sigma (double) - width of Gaussian used for random noise sampling, default value = 50 ADC counts

• NumberOfDigits (int) - number of random digits to be added to event, default value = 0 corresponds to random number of digits added

NiffteCheckPad is a NIFFTE-specific afterburner that verifies the output of the detector simulation modules matches the physical dimensions of the detector. For example, the diffusion, charge sharing and noise modules may produce digits outside of the instrumented regions of the pad-plane. These boundary conditions are NIFFTE-specific and must be imposed before the simulated data is written out for subsequent reconstruction.

• No parameters currently configurable

TPCRaw contains a simple preamp model, 1/f noise generator, semi-Gaussian filter, and threshold trigger. It generates the final list of TPC “raw” digits from all of the simulation steps

• TriggerLookbackSamples (int) – max number of extra digits to add before and after each charge deposition

• TriggerThreshold (int) – threshold for triggering

• TriggerHysteresis (int) – hysteresis parameter for trigger

• NoiseSigma (double) – Gaussian width of the noise distribution

• ADCMinVoltage (double) – minimum voltage for each ADC

• ADCMaxVoltage (double) – maximum voltage for each ADC

• ADCSteps (int) – number of ADC steps

• MicroMegasGain (double) – gain of the charge amplification system

• PreampGain (double) – gain of the preamplifier

• PreampRCTimeConstant (double) – decay constant of the preamplifier

• PreampPedestal (double) – pedestal value for the preamplifier

• PreampPedestalSigma (double) – Gaussian width of the preamplifier pedestal

An example XML configuration file, called testDetSimConfig.xml, implementing each of these elements is included in the $NIFFTE/build directory. The output from the detector response simulation forms the input to the reconstruction modules described in the offline software section.

Mock Data Challenge

A Mock Data Challenge (MDC) is a computing and analysis exercise that is a standard in high energy physics experiments. The goals of an MDC are to

• Accurately simulate the performance of the detector(s) for a variety of key operational parameters to evaluate and benchmark the design.

• Validate the simulation and reconstruction software.


• Test under realistic data analysis conditions the performance of the computing hardware resources, storage and job management.

The analysis of the output from the MDC will help determine any necessary adjustments to the TPC design. The completion of the MDC depends on the availability of the hardware and software resources needed to perform it. Most of FY09 was spent in finalizing the software, as described in detail in the other software sections of this report, procuring and configuring the hardware on which it is to be run (see below) and performing as much of the analysis as possible on the generated data. In order to manage these efforts and to keep track of the progress towards our goals, weekly MDC meetings were convened starting in the second quarter. In addition, a two-day workshop was convened on July 9-10 at Los Alamos in conjunction with the second collaboration meeting of the year. At the workshop, the principal players presented status reports summarizing all of the progress on simulations and offline software and a full-scale tutorial on software installation and operation was conducted. At the end of the workshop all participants had a working copy of the full NIFFTE software suite running on their laptops and were given example exercises for creating analysis macros to investigate MDC output.

MDC Hardware

In order to complete the MDC, a high-performance computing cluster (HPCC), dubbed “NEATO” was created at GIT. NEATO is a 192-processor compute cluster running the CentOS 5.2 operating system. The cluster is built as a parallel computing machine to service the NIFFTE collaboration. NEATO has a dynamic load balancing software package installed for ease of use by the collaborators that is used for run scheduling and load balancing. A complete software package for the NIFFTE simulations and data analysis is maintained there. Remote login capabilities allow users to use remote desktops to provide full graphics rendering capabilities. A Postgre-SQL database was created and a backend server established to store all of the identifying information about the MDC runs. An interactive load monitoring web page created using the Ganglia package can be viewed to check the load on the cluster at any time using the ganglia user interface. Example job submission scripts and documentation were included in the svn repository and a NIFFTE production account; “niffteprod” was created to centrally control the submission and storage of MDC data.

MDC Analysis

Results from the analysis of MDC data using the Hough Transform plus Kalman filter tracking algorithms was shown in the offline software section. Further simulations and analysis are ongoing and will be completed in early FY10.

Collimation Simulations

The simulation effort this quarter focused on finalizing the collimator design and beginning the room return contributions to the TPC. Finalized models for three collimator designs have been created. Three different collimator designs were presented for fabrication and installation at the LANSCE. Two different designs are being proposed, a divergent collimator design and a convergent collimator design. Each design is shown below in Figure 72, Figure 73 and Figure 74.


Figure 72: The large orifice divergent collimator. This provides the largest beam spot possible with

the shutter assembly installed on 90 L. The black lines represent the primary radiation field, the

orange lines represent the divergent neutron radiation field. The light blue is the bulk shield, and the

blue cylinder is the tungsten target.

Figure 73: The image above is a simplified design for the small 2 cm divergent collimator. The black

lines represent the primary neutron field and the orange lines represent the secondary neutron field.

Notice that orange lines cannot penetrate the bulk shield.

Figure 74: Here is the convergent beam image. In the convergent beam image, the primary beam

spot is very intense but the secondary orange neutron scatter fields are quite large and degrade

system performance.


Two divergent collimators were designed. One collimator provides the largest beam spot the shutter assembly allows. The second is a perfect 2 cm beam spot at 10 meters from the target. The convergent collimator provides a much higher fluence rate designed for a 2 cm beam spot. However, the secondary radiation field is significantly increased. This results in a vastly degraded energy versus time profile. Several time dependent neutron beam images of the divergent collimator are shown in Figure 75 through Figure 79.

Figure 75: The 100 keV to 20 MeV neutron beam image at 20 shakes (200 nan0seconds) after the

proton pulse. The neutron pulse is arriving at the right initially producing the initial neutron pulse.

Figure 76: The 100 keV to 20 MeV neutron beam image at 35 shakes (350 nanoseconds) after the

proton pulse. The neutron beam spot is moving across the tungsten target.


Figure 77: The 100 keV to 20 MeV neutron beam image at 40 shakes (400 nanoseconds) after the

proton pulse. Here the beam spot is no longer a spot at all, but a image of the target itself.

Figure 78: The 100 keV to 20 MeV neutron beam spot at 20 shakes (200 nanoseconds) for the large

aperture collimator.

Figure 79: The 100 keV to 20 MeV neutron spectrum at 30 shakes (300 nanoseconds) after the proton

pulse for the large aperture collimator.


Figure 80: The small aperture 2 cm beam spot from 100 keV to 20 MeV. The color band has been

multiplied by a factor of 100 to enhance the ratios.

Figure 81: The small aperture 2 cm beam spot 100 keV to 20 MeV after tertiary collimation at

10 meters.

It can be seen in Figure 80 and Figure 81 that the divergent collimator produces a varying time dependent neutron energy profile. The fluence rate is significantly higher, however due to the time varying profile, this design isn’t an optimal one. The large aperture beam profile provides a nice clean beam spot with a 10 cm diameter produces a uniform large diameter beam profile with a slight target outline shape. The small diverging collimator with tertiary collimation produces a very uniform without any time dependence and for the highest accuracy measurements, this design is recommended at the cost of neutron fluence rate.


MCNPX Model Validation

The MCNPX model for room return, collimator and air scattering has been evaluated using a series of time dependent neutron and gamma-ray measurements. A BC501a neutron/gamma scintillator was used to measure the time-dependent neutron field around the proposed TPC location in a reference configuration with no collimation in the inner and outer shutters. These time-dependent measurements will be analyzed separating the neutron from gamma-ray responses and unfolding the neutron spectra. This will provide a benchmark configuration for comparison to the MCNPX scatter calculations. Future work will be needed to ascertain the origins of each component of the scatter field in order to lower the scatter contribution to the fission cross section measurements. The MCNPX model will provide insight into the various components of the scattered neutron component at the TPC location.

Advanced Neutron Detection

Scope

In support of the goal to begin to measure the neutron emission from the fission target in the TPC, a campaign to develop new neutron detectors has begun this year. The goal of this neutron detector campaign is to develop new, high speed, high efficiency, low optical density materials which could be incorporated into a double time of flight neutron detector array. Previous designs have relied on organic scintillators because of their traditionally fast speeds and high hydrogen content. These detectors however rely on pulse shape discrimination to separate neutrons from gamma rays. This system has a high dead time due to the requirement to resolve the entire pulse as well as a pulse height non-linearity. These detectors rely on hydrogen scattering as the neutron detection mechanism. This mechanism however is only good above 300 keV on an extremely sensitive, low noise system. The introduction of a new target material is needed. To accomplish this, two novel detection schemes have been envisioned and are being pursued. Both methods are a thin film device that is engineered to produce an optimal response for neutron detection while producing a minimal gamma ray response.

Highlights

• A variety of ZnO scintillators were grown (bulk growth using a high pressure melt growth method and MOCVD precision deposition techniques) and tested using thermal, fission and fast neutrons. High levels of a variety of dopants (Li, Ga, In, Al, N and C) have been achieved with excellent crystalline quality.

• 1 mm2 GaN semiconductor devices have been successfully created with Gd dopants and the response to photon, neutron and alpha sources measured.

The first design is lithium or boron doped ZnO scintillator. ZnO material is a scintillation material was studied in the 1950s and 1960s and thought to be a promising detector material. However, at that time, only polycrystalline powders were available and light collection was difficult. Now, with resurgence in interest in the use


of ZnO as a room temperature semiconductor, research into the growth and fabrication of it has progressed and ZnO’s potential as a radiation detector bears a new look. It should be noted that ZnO also is currently being investigated as a ultraviolet LED material.

The ultrafast ZnO scintillator has several unique characteristics such as a 30 pS pulse height rise time and a 0.65 !S decay time(Neal). High light yields are advantageous, with yields greater than 100,000 optical photons per MeV of energy deposited. These detectors, as are most neutron scintillators, are sensitive to gamma rays. The scintillator response to gamma rays, however, can be minimized by the use of thin crystals. Although the thin detectors lead to decreased efficiency of the detector, very high dopant levels can be used to increase the efficiency and surface treatments (coatings) can be use to increase the detection sensitivity. Evaporated coatings of LiF enriched in 6Li or Gd coatings can be applied to the external surface to generate recoils that can be used to create scintillations in the crystal.

Values of this that can be found in literature on previous work on ZnO can be seen in Figure 82. The ultrafast rise and fall times are advantageous to the TPC project since it utilizes the double time of flight system. The faster the timing on the front end, the closer the detectors can be placed to the TPC. This minimizes scattering of the neutrons as well as reduces the overall cost of the detector array. The cost minimization is two fold. Fewer channels need to be monitored as the number of separate scintillators decrease. Second, the amount of scintillators decreases thus lowering the overall cost of fabrication.

Figure 82: Pulse height rise (in millivolts) and fall time measurement (in nanoseconds) of the ZnO

scintillator (open boxes) and plastic scintillator (open circles).


Two distinct growth methods exist for obtaining high crystalline quality ZnO. The first is through a bulk melt growth procedure. Using a custom high pressure melt growth oven, large charges of zinc are introduced, and heated to melting point. An oxidizing atmosphere is introduced and completely oxidizes the zinc to ZnO. During the growth-melt process, lithium dopants are added to the ZnO melt integrating into the lattice structure of the ZnO. The ZnO is allowed to cool slowly after introduction of a seed crystal. Up to 10 weight percent dopant levels have been achieved using this process. A 10 weight-percent scintillator wafer can be seen in Figure 83 through Figure 85.

Figure 83: ZnO lithium doped scintillator with 1” diameter 0.1 mm thickness.

Figure 84: ZnO lithium doped scintillator with 1” diameter 0.1 mm thickness.


Figure 85: ZnO lithium doped scintillator under black light illumination showing its strong

absorption in the UV range.

A second growth method for ZnO that is being investigated is through the use of Metallo-Organic Chemical Vapor Deposition (MOCVD). This MOCVD process allows for very fine control over the growth parameters and detector thickness. MOCVD allows for thin, crystalline ZnO to be grown with conformal coatings of lithium containing compounds. Through this process, control of the growth of the crystal is strictly regulated. The growth is controlled by precision regulation of the addition of the different precursor materials. The growth is conducted inside a specialized growth chamber shown in Figure 86. Diethyl zinc is cracked in a vacuum growth chamber and reduced with precise amounts of oxygen. A nitrogen blanket gas is used to clear away all of the impurities. The by-products of the growth process are CO2, H2O, and ZnO, all of which are environmentally friendly. This scintillator is one of the only scintillators to be grown in an entirely environmentally friendly fashion. Thus precision control of the dopant levels and crystalline quality are possible. In the MOCVD growth process, precision 5 !m layers can be grown over wurzite structure substrates such as silicon, lithium niobate, lithium tantalite, or sapphire. Using a pulsed growth method and direct liquid precursor injection, high lithium dopant levels are possible with growth rates in excess of 4 microns per hour. Large planar structures can be grown with this method with production runs up to 6-inch diameter wafers possible.

Figure 86: The ZnO Growth tool.


Through the MOCVD process several different dopants have been investigated. Dopants have been initially selected based on their ability to adjust the donor electron concentrations and effectively engineering the band-gap of scintillation to move the scintillation wavelength to a longer wavelength. This helps overcome the strong affinity for reabsorption. Dopant materials such as nitrogen (precursor of ammonia), aluminum (precursor of diethyl aluminum) gallium (triethyl gallium precursor), indium and cobalt in varying dopant concentrations has been investigated and compared to the response of lithium doped ZnO. These scintillators were tested by observing the pulse height distribution for a fixed Pu-239 alpha source configuration and fixed counting time. A plot of the resulting pulse height spectra can be seen in Figure 5. Here, lithium alone is providing the best results. The combination of doped scintillators can be seen below in Figure 87.

Figure 87: Several doped and undoped scintillators. Moving from the top left, a 1” diameter ZnO:Li

doped scintillator, a 1cm x 1cm x 0.1 mm ZnO:Li doped scintillator, 1 cm x 1 cm x 0.1 mm ZnO:In,

An initial 2” MOCVD grown 5 um thickness scintillator. Second row, 1 cm x 1 cm x 1 mm ZnO:Li, 1

cm x 1 cm x 1 mm undoped ZnO, 2” MOCVD grown ZnO:Al, 2” MOCVD grown scintillator with a

800 C anneal to improve quality, and bottom row, a piece of GaN, and a ZnO:Co doped piece of

material.

Figure 88: Pulse height distribution of several different dopants inside the ZnO matrix. Lithium

dopant has shown the best neutron versus gamma discrimination capabilities with the pulse height

distribution for the alpha peak being the furthest from the gamma ray low energy tail.


Several tests have already begun on these scintillators. The first, experimental pulse height rise times were measured. An example of the measured pulse height rise and fall time can be seen in Figure 89. Here, the pulse height distribution is limited to the speed of the photomultiplier tube. The peak is a uniform Gaussian 2 us wide which is expected from this PMT. If a faster PMT or micro-channel plate PMT (MCP-PMT) were available, the faster pulse shapes would dominate. These detectors were tested initially with several alpha particle sources of different energies.

Figure 89: An anode pulse coming out of the PMT for ZnO. The major divisions are 1 nS/division.

The resulting pulse is currently limited to the rise and fall times of the PMT which is limited at 2 nS

rise and fall times.

Work has begun on developing a functional prototype for a lithium-doped thermal neutron ZnO scintillator. Seven different prototype scintillators have been developed by Georgia Tech and grown either by Cermet Inc or by MOCVD at Georgia Tech. Some images of prototype detectors are shown in Figure 87. Preliminary testing of prototype structures has shown reasonable response to fast neutrons (PuBe source) and a limited response to gamma rays. Crystals that were 1 cm x 1 cm x 0.1 mm and 1 cm x 1 cm x 1 mm have been tested. The results are shown below in Figure 90 and Figure 91. The scintillators have also been tested with Pu-239 alpha particles. The detector should have a similar response to ~5-MeV alphas as it would to the charged particles produced in the 6Li(n,!" reaction. A 239Pu generated pulse-height spectrum can be seen in Figure 92. The resolution of the alpha particle peak is distorted since the test was conducted in air with an electroplated source, so there is a distribution of energies due to energy losses in the intervening materials. The efficiency for detecting these alpha particles is < 86 % intrinsic efficiency.


Figure 90: Experimental neutron response of lithium doped ZnO crystal to PuBe neutrons.

Responses are shown for 60Co gamma rays and neutron response spectra. The dopant levels are

0.1% by mass 6Li.

Figure 91: Experimental neutron response of undoped ZnO. Responses are shown for a 60

Co

spectrum and PuBe neutrons. This sample is undoped ZnO.


Figure 92: Pu-239 alpha particle spectrum of lithium doped ZnO scintillator. The red curve is the

alpha particle events while the blue curve is > 100 mR/hr gamma ray field. These preliminary

scintillators show exceptional neutron gamma discrimination.

These scintillators show good neutron versus gamma ray discrimination as shown in Figure 93. Gamma discrimination is readily accomplished by using simple low-level pulse-height discriminator settings. Due to their ultrafast speed, pulse pileup events which might lead to false positives in other detectors, will not occur with ZnO scintillator except for extremely intense neutron sources which are not likely to be found in practical applications.. Prototype detectors have been compared to standard detectors for thermal neutron detection such as 3He tubes and lithiated glass scintillators. The intrinsic efficiencies of the doped ZnO scintillators are ~2 that of a lithiated glass scintillator of the same diameter and can be over 5 times that of a 10 atm 3He tube. Additional ZnO scintillators are constructed and will be tested in a variety of neutron fields. The detectors performance will be compared to that of a lithiated glass scintillator and currently available 3He tubes (E. Burgett; Reynolds; Reynolds).

Figure 93: Pu-239 alpha particle spectrum of lithium doped ZnO scintillator compared to the

neutron response. The red curve is the alpha particle events while the blue curve is the response to a

fission like neutron spectrum. This involves a large thrmal peak, a fission peak and a hard !,n peak.

The higher energy neutrons lead to the forward skewed nature of the response showing the

scintillators spectroscopic capabilities. These preliminary scintillators show exceptional neutron

gamma discrimination.


Based on the response of the Li-doped ZnO detectors to different energy alpha sources, there is a potential to also use the scintillator as a neutron spectrometer if the thermal neutron response can be suppressed. The neutron field used in Figure 93 was a spectrum with a fission peak, a fast (!,n) peak, and a large thermal neutron peak. This can be seen in the forward skew of the alpha peak in the scintillator. This spectrum potentially could be unfolded resulting in an ultra thin, ultra fast spectroscopic neutron scintillator which discriminates gamma ray interactions by pulse-height discrimination.

The second neutron detector technology being investigated is a neutron detecting semiconductor detector. This device is based on a GaN diode structure. GaN was selected because of its high tolerance to radiation damage. It has been tested at CERN as a potential radiation detector. The researchers at CERN have exposed GaN diodes to very high fluence rates of neutrons, pions, and other hadrons such as protons. Radiation tolerances shown in Figure 94 show that the GaN structures would survive far longer than the course of the experiment run at LANSCE. 500 nm intrinsic regions have successfully been grown. Wire bonded P-i-N structures have been made and tested at Georgia Tech. Neutron testing of these devices has proven that they are sensitive to neutron radiation. To improve their efficiency, work is being done to increase dopant levels as well as through the use of target nuclei layers. Evaporated coatings of Lithium, Boron, Gadolinium, and Uranium are being investigated. To optimize each of these designs, a buffer layer is needed to protect the diode structure as well as to help range out the energetic particles produced from each of the different target nuclei. A side view and top view of the proposed devices can be seen in Figure 95. These structures are grown off of the 0,0,1,0 plane. The chemical structure of the GaN device can be seen in Figure 96.

Figure 94: Radiation hardness of GaN semiconductors in an unpowered state for neutrons pions and

other hadrons.


Figure 95: Side and top down configuration of the proposed GaN semiconductor structures. SiN

spacer layers are needed to help range out the alpha particles.

Figure 96: Chemical molecular structure of GaN.

Several new GaN masks were created for the etching process to create larger semiconductor structures. This includes a 1 cm x 1 cm chip, several 5 mm x 5 mm chips and many smaller sized detectors. The new mask structure that has been designed can be seen in Figure 97. Two detector sizes were introduced, 1 square mm devices in addition to the traditional 350 square nanometer devices. These larger devices are a step in the right direction towards a fieldable, viable semiconductor neutron detector.


Figure 97: New mask structure for GaN structures. The large devices are 1 mm2 while the smaller

devices are 350 mm2.

Work also on the incorporation of Gadolinium into the matrix was investigated. The gadolinium content was not experimentally verified with X-ray Defraction XRD techniques. Since GaN has similar characteristics to ZnO it can function as a scintillator as well. GaN was tested as a scintillator structure to compare doped and undoped responses. Since the incorporation of gadolinium into the matrix raises the effective Z number significantly, it was easy to determine the change in effective Z number by comparing the increased response to alpha particles. By using the Bethe-Bloch slowing down equations, dopant levels were calculated to be 7.5% +/- 0.5%.

The pulse height spectra resulting from doped and undoped GaN can be seen in Figure 7. The incorporation of the higher Z number material has produced an adverse effect in that it increased the low energy gamma ray tail that encroaches on the GaN’s alpha peak. This unwanted effect will be investigated this next quarter to try to mitigate these effects.

Figure 98: The pulse height distribution of GaN undoped (light blue and with the alpha peak

illiuminated in red) compared to the doped GaN pulse height spectra. The discernable alpha peak

has been smeared leaving a smoother continuum.


Electronic structure testing has been completed on the completed P-i-N structures that have been grown. A 500 nm i region material has been grown and tested. This structure has been wire bonded to a DIP package for easy integration into a circuit system for testing. The DIP packaged GaN structure can be seen in Figure 100. This detector had 5 working 350 nm structures on it. These structures were tested and a sample I/V curve resulting from this set of detectors can be seen in Figure 101. These are a result of the band diagram also shown in that figure. These detectors are quite rugged and show good promise as a neutron detector system.

Several different intrinsic region detectors were created and tested. Intrinsic regions ranging from 100 nm to 5000 nm were created. IV curves for these devices are shown in Figure 99. Simulations were conducted to reconcile the lower performance of the larger intrinsic regions. These simulations can be seen in Figure 103. It was found that the lower efficiency comes from the inability to fully deplete the intrinsic region producing an insulator structure. This may be able to be reconciled in the near future by utilizing a Shotky style device. Even though these devices had a difficult time functioning as P-i-N structures, they were still of excellent crystalline quality. The samples were tested by rocking mode XRD. The resulting information, shown in Figure 102, demonstrates good crystalline quality.

Figure 99: IV curve for five different intrinsic region thicknesses in GaN that were grown as neutron

detectors.


Figure 100: Wire bonded DIP package for GaN semiconductor detector.

Figure 101: Band diagram and I/V curve for GaN large intrinsic region semiconductor detectors.

Figure 102: Rocking mode XRD profile of GaN structures. The crystalline quality is still ok.


Figure 103: Simulation of the depletion zone in a GaN PN junction. This is the origin of the 200 nm

limitation.

The advanced neutron detection effort within NIFFTE has progressed significantly this year and is on par to meet the next generation needs of the fission measurement campaign.


The Hydrogen Standard

Scope

The project to accurately and precisely determine fission cross sections hinges on the H(n,n)H total cross section and angular distributions. The H(n,n)H total cross section is well determined with errors less than 0.5%. In the planned TPC measurements, hydrogen will to be used as the working gas for a new standards measurement of U-235. The H(n,n)H angular distribution must be known to connect the total cross section to the measured elastic recoils in the TPC.

Highlights

• The tandem belt that was damaged as a result of sparking down the vertical column has been replaced and extensively tested during the last quarter and is ready for running at low voltages.

• The test experiment for neutron scattering at 4.0, 10.0 and 14.9 MeV as been set up and tested using neutron sources.

• A draft of a paper has been completed on the H(n,n)H angular distribution at 14.9 MeV. The proton recoils were measured using an 11 member $E-E

silicon solid state angular array.

• The active target for H(n,n)H scattering is being reengineered.

• The initial phase of neutron efficiency determination with a Cf-252 source has been completed.

Hydrogen Standard [OU]

The TPC offers a major advance in technology for measuring the H(n,n)H angular distribution. The solid angle for this detector is a factor of 100 larger than that used in our current measurements at 14.9 MeV. The target thickness would be comparable. This would result in a counting rate increase by a nearly a factor of 100. This may allow high accuracy measurements of the angular distributions. A further factor is that the beam need not be as strongly collimated which would give as much as another factor of 10 statistical improvement. This may reduce the experimental running time for a 1% measurement from 3 months to days. The results would be binned into an angular distribution in the center of mass system. The fitted angular distribution can then be compared to calculations based on potential models such as Bonn or on phase shifts such as Arndt. This may eventually provide a test of the QCD-based model of nuclear forces. We are proposing to collaborate on the modeling of hydrogen scattering in the TPC chamber. We will look at the minimum energy detected. We will also investigate the angular resolution for H(n,n)H scattering in the chamber. A pure hydrogen atmosphere will be investigated along with addition of quenching and/or scintillating gases. Further work will also be done on determining the gas composition and density to better than 0.2%. The systematic errors in a standard measurement must be fully explored in order to reach the desired goal. The


modeling work will be focused on these problems. Consideration will also be given to possible inter-comparison of the neutron standards such as 6Li(n,")3H, 10B(n,")7Li

and 235U(n, f).

H(n,n)H Experiment

A draft of a paper of H(n,n)H scattering at 14.9 MeV with protons detected has been completed and will be submitted soon. The background and the error estimates are under review. The error bars on the individual points are now between 1-2.5%. This review has reduced the error estimate substantially on some of the more important data points.

The tandem belt was damaged as a result of sparking down the vertical column. The column has been chemically cleaned to remove a film that had developed on the surface. A new plastic belt has been hung and normal operation has now resumed at low voltages. The belt initially had large instabilities that cleared up with additional use. The belt has been extensively tested during the last quarter and is ready for running at low voltages.

The problem with the belt was diagnosed as a piezoelectric effect from the polymer that composed the belt. The problem was circumvented by changing the point where charge is applied in the region of the pulley to the first plane. Using this charging method, stable operation of the accelerator was demonstrated to be at least a factor of five better than ever previously achieved.

The work of setting up to measure the neutron scattered by hydrogen has been completed. There has only been one previous attempt at measurement of the neutrons under 20 MeV. That work was carried out using the associated particle technique at 14 MeV. As shown in Figure 107, these data do not constrain the angular distribution. The error bars on the current evaluation are significantly larger at the forward angles.

The experiment to measure the angular distribution of the neutrons from H(n,n)H scattering was begun with a test experiment to determine the errors involved. The ultimate plan is to measure the neutron scattering at 4.0, 10.0 and 14.9 MeV and at least 5 angles. The electronics for H(n,n)H test experiment have been designed to allow high event rates for the scattering detectors. The data acquisition program has been tested using neutrons from radioactive sources.

A swinger shadow bar for has been set up for neutron scattering. This was resurrected from the various hiding places around the lab. The shadow bar has been adjusted to reduce first order scattering (see Figure 104 and Figure 105).

Neutrons are generated in either a d(d,n) gas cell or a titanium tritide target. Neutrons are then scattered by an active detector, such a scintillator such as a plastic scintillator, stilbene, or a liquid scintillator cell filled with benzene, deuterated benzene, or NE213. The scattered neutrons are then detected in well-shielded neutron detectors in the tunnel (see Figure 106).

The electronics for H(n,n)H test experiment have been designed to allow high event rates for the scattering detector. The data acquisition system was tested using sources. The neutron source used was a deuterium gas cell for 4.0 MeV neutrons.


Experimental Status

The swinger shadow bar for neutron scattering has been put in place. This was resurrected from the various hiding places around the lab. The shadow bar was adjusted to eliminate any neutrons from the source being seen by the detectors in the tunnel and to reduce first order scattering.

Test experiments were completed to check the neutron backgrounds in the experiment. In the test experiment, a d(d,n) gas cell was used, as the maximum voltage of the accelerator was limited during the wear in time of the new belt. The source neutron was scattered by an active detector, either a plastic scintillator or stilbene. The neutron was then detected in well-shielded neutron NE213 liquid scintillator detectors in the tunnel. Three 5”x2” NE213 liquid scintillator detectors were used at 6.97 meters in the tunnel for the detection of the scattered neutron. Mesytec Modules were used for the detector pulse shape discrimination in the tunnel detector signals. The Mesytec modules were controlled remotely using their USB port.

A thin 3 mm thick stilbene monitor detector was mounted on the swinger. This monitor was mounted at a fixed distance from the center of the swinger axis. The monitor was routed into the data acquisition system with the tunnel neutron detectors. This was done to allow for a common dead-time, which is the included when the ratio of the detector to monitor is calculated.

The data acquisition system was modified for this experiment to have a fast clear and enable for the accepted data. The data acquisition set up such that no data is accepted until an electronic event signal is received from the tunnel detectors. Using the timing of the neutrons, a gate is set so that the slowest neutrons possible in the experiment are accepted and all later signals are rejected. With this system, it is possible reach triggering rates of 20,000 events/second on the scattering detector with no harm to the data acquisition. Results of the current experiments are shown in Figure 111 and Figure 112.

The test of neutron scattering at 4.0 MeV has revealed that the multiple scattering from the source detector is unacceptably large. The significance of multiple scattering was found to increase with angle. At an angle of 110 degrees, the elastic scattering of carbon was observed. At the higher energies of 10 and 14 MeV, the elastic scattering from carbon is easily observed. The carbon scattering may be removed, at least in part, by requiring larger pulse heights in the scattering detector.

The use of the scattering detector as a monitor was tested. Raw net counts versus angle were compared to the integrated current. It was demonstrated that the scattering from the shadow-bar causes more than 20% excess counts in the scattering detector. A silicon surface barrier %E-E telescope or a CsI(Tl) detector with

pulse shape discrimination is being considered. Both of these methods would have a hydrogen target and a detector for the recoil proton. Both have the advantage of being sensitive only to the target area at which they are pointed.

We have worked on improving the measurement of the H(n,n)H angular distribution. This has consisted of two parts: improving the experimental setup and increasing the accuracy of the neutron calibration.

The work to improve the detector calibration had been completed. The Cf-252 prompt neutron spectrum following spontaneous fission is believed to be known to better than 1%. A Cf-252 source was borrowed from OSU. These measurements have been


finished and the Cf-252 source returned to OSU. We have ordered a 50 microcurie target of Cf-252 target on platinum to achieve better separation of the fission fragments from the alpha particles.

The active target for H(n,n)H scattering was reengineered. We have used a plastic scintillator and a stilbene detector in the trials of the experiment last January. We now have 2.54 cm diameter by 2.54 cm long targets of deuterated benzene, benzene and a NE213 equivalent scintillator. We are now using a 1.25 cm diameter plastic scintillator which is suspended by thread inside a thin aluminum cylinder. This suspension is expected to reduce the contribution from the (n,Z) reactions on the walls. The smaller diameter will also reduce the contribution to the energy width of the observed neutron scattering.

The first phase work on neutron efficiency determination with a Cf-252 source has been completed. A low mass fission chamber has been fabricated for conducting neutron calibrations (see Figure 113, Figure 114 and Figure 115). These measurements use a low mass fission chamber, the energies are determined by time of flight, and the neutron flux is determined by the fission counting rate and the neutron multiplicity to determine the efficiencies. We have found and eliminated many sources of timing errors in the time of flight determination. Several electronic modules have been found to have variations in the timing signal of up to 20 ns. We now have the time variation down to less than 0.5 ns in a month. The “typical” response of a NE213 liquid scintillator to neutrons and gammas is shown in Figure 116. The separated neutron and gamma spectra are shown in Figure 117. A comparison to a Monte Carlo simulation to the measured efficiency is show in Figure 118. A comparison of efficiency for three similar detectors is shown in Figure 4. The increase in efficiency at high energy is thought to be an artifact due to the shoulder of the prompt gamma ray peak shown in Figure 117.

With further work on this method, we hope to improve the accuracy of the neutron efficiency calibration to 1.2% at 1.0 MeV and 2.4% at 8 MeV. We have compared the results of these calibrations with the stopping target calibrations performed with Al(d,n) and B(d,n).

We have estimated the influence of the time resolution on the high-energy range of the detector efficiencies. We also tried a number of current sensitive pre-amplifier to remove this shoulder in the prompt gamma in and improve the time resolution.

We have conducted several preliminary measurements at 14.9 MeV. We have used the 2.54 cm diameter by 2.54 cm long stilbene detector as a monitor detector. This detector has excellent pulse shape discrimination between gammas and neutrons. The number of neutrons scattered through the scattering detector and detected in coincidence at the monitor can be related to the incident neutron flux regardless of the angle of the swinger. This will allow a direct normalization of the angular scattering for this experiment.

The preliminary runs show very clean spectra for neutron scattering for both small angle of 20 degrees and our largest angle of 70 degrees. No unexpected backgrounds have been found.

Work has been completed in simplifying the electronics. They now are less dependent on all of the delays being properly set. We have also developed some checking procedures to make sure that the pulse height and the pulse shape spectra are monitored to avoid unwanted gain changes. The high voltages of the tunnel


detectors (and possibly the scattering detector) will be monitored at least twice a minute by an automatic logging system.

Figure 104: A view through the shadow bar and shadow box at the neutron detectors in the tunnel.

The detector in the foreground is a 2.54 cm high by 2.54 cm diameter stilbene detector. This detector

will be used as a scattering detector for hydrogen in this experiment.

Figure 105: A Side View of the neutron scattering experiment. The gas cell is on the end of the

swinger and an alignment piece is shown above the copper block. The main shadow bar block is in

position to block neutrons from the gas cell from reaching the detectors in the tunnel.


Figure 106: Shown here are three 5" diameter by 2" thick neutron detectors in the accelerator

tunnel.

Figure 107: The H(n,n)H scattering data and evaluations at 14 MeV.


Figure 108: Looking through the shadow-gar and shadow box at the detectors in the tunnel. The

detector in the foreground is a 2.54 cm high by 2.54 cm diameter stilbene detector. This detector will

be used as a scattering detector for hydrogen in this experiment.

Figure 109: A Side View of the Neutron Scattering Experiment. The gas cell is on the end of the

swinger an alignment piece is shown above the copper block, and the main shadow bar block is in

position to block neutrons from the gas cell from reaching the detectors in the tunnel.


Figure 110: Three 5" diameter by 2" thick neutron detector in the OUAL tunnel.

Figure 111: Neutron spectra from H(n,n)H scattering with plastic and stilbene scintillators.


Figure 112: Comparison of gamma ray spectra from plastic and stilbene detectors.

Figure 113: Shown here is a thin wall fission chamber to be used for neutron calibration.


Figure 114: Shown here are the inner electronics of the fission chamber. A dummy fission foil (dark

disk) is held in place by the electrode.

Figure 115: This is the top view of the inside of the fission chamber. The distance between the disk

electrodes is 3 mm.


Figure 116: Pulse shape (x-axis) versus pulse height (y-axis) for NE213 detector measured with 252

Cf

source.

Figure 117: Time of flight spectra measured with Cf-252 source for neutrons (blue) and gamma-rays

(pink). The small peak in channel ~1460 is a detector delayed signal for measurement of time

stability.


Figure 118: Detector efficiency measured with 252Cf source (pink) and result calculated by MC

method (blue line). The method for light output parameterization is described in N.V.Kornilov, I.

Fabry, S. Oberstedt, F.-J. Hambsch, NlM A599 (2009) 226.

.

Figure 119: Detector efficiencies for different neutron detectors applied for np-scattering

experiment.


Facilities and Operation

Scope

Due to the necessity to have a finely tuned neutron beam, with as little contamination as possible, the experimental area needs to be groomed for TPC installation and running. This will mean additional collimation will be needed to adjust the 90L flight path to work with the TPC. MCNPX simulations will be made of the 90L flight path. The collimation system will be manufactured at the Georgia Tech Research Institute (GTRI) machine shop. This world class machining facility is available to machine items up to 10m x 10m in size. GTRI employs a large number of undergraduate and graduate students to provide designing and drafting capabilities.

The TPC mount will need to be fabricated. The TPC mount will consist of a 3-axis positioner that the TPC will mount to that will allow for precise positioning of the TPC in the neutron beam. The design specifications will come from the TPC design team, as well as 3 axis movement specifications for fine tuning in the LANSCE beam. The mount will be designed at Georgia Tech and machined at GTRI to match the exact specifications provided by the design team. Introducing the problem in a joint nuclear engineering and mechanical engineering student’s senior design project will foster student participation. The experimental infrastructure will be partially provided by facilities currently at the LANSCE facility. Georgia Tech graduate and undergraduate students will participate in the scoping and design of the experimental facilities that will house the TPC at LANSCE. A good working rapport will be established with the facilities personnel at LANSCE through this collaborative effort.

The TPC experiment will be maintained and monitored while located at LANSCE. The Nuclear Science group employs a number of qualified technicians who will perform the required upkeep and maintenance of the TPC and related systems. The facilities will be maintained such that the instrument will function properly and beams can be supplied to the area. The TPC detector and associated electronics will be maintained as necessary. The gas system will be monitored and maintained, including gas bottle replacements and any required periodic testing. Experimenters will maintain the experimental specifics of the data acquisition system and a LANSCE supplied computer technician will provide the necessary computer support.

In addition to running at LANSCE, the TPC will also run at other facilities to cross check systematic errors. This will be critical to achieve the small systematic errors that are the goal of this experiment. One possibility is the ALEXIS facility under construction at LLNL. This mono energetic neutron source is notable for the low cost ($150/hr to have the whole facility) and high luminosity (108 n/s at 10 cm) neutron beam that will complement the LANSCE facility.

Another notable resource is the accelerator at Ohio University, which will be used to study the hydrogen standard for this project and develop the data required to extend the small uncertainties in the H(n,n)H total cross section to the actinide measurements.


Highlights

• The LANSCE FIREhouse facility, which will house the TPC, is being prepared to accept the device.

• TPC lab space has been secured and upgrades to include a clean bench and fume hood are underway.

• The flight path design and environmental interface has been completed.

Livermore [LLNL]

There are numerous facilities at LLNL that are of interest to this project. The biggest is the construction of ALEXIS, Accelerator at Livermore for EXperiments in Isotope Sciences, which may be available in FY10. This facility will generate pseudo-monoenergetic neutrons up to 108n/s/cm2 at energies from 100 keV up to 14 MeV at low operating cost. The LC computing system has large CPU clusters and storage systems that have been successfully utilized by similar computing projects such as Phenix at RHIC, MIPP at FNAL, and is currently working on setting up ALICE at CERN.

Los Alamos [LANL]

The Nuclear Science group at Los Alamos Neutron Science Center operates and maintains the Weapons Neutron Research facility that provides spallation neutrons to five flight paths. The group also maintains and operates two moderated neutron flight paths in the Lujan Center. The group operates and maintains the Blue Room facility, with access to an 800 MeV proton beam and a Lead Slowing Down Spectrometer. The Nuclear Science team will provide the floor space and neutron beam access to the TPC project primarily on the 90Left flight path at the WNR and flight path 5 of the Lujan Center. The 90L flight path experimental area is inside a new construction that contains an overhead crane, light lab space, a vented hood, source safes, computers and easy access to the neutron beam line. Flight path 5 experimental area includes an overhead crane, light lab space, source safes, computers and easy access to the neutron beam line. Recently refurnished light lab space will also be available for TPC work. Monitored stacks are in the vicinity of the two flight paths for TPC gas system and hood exhausts. Radiological shipments and handling facilities are also available. LANSCE provides outside users with all necessary facility and safety training, a cafeteria, and meeting rooms.

FIREhouse

Preparations have been made to modify the FIREhouse flight path for housing the TPC. Beam time has been allocated in November to test a collimation system designed for the TPC by GIT. A new hydrogen gas supply system has been designed and in the process of installation at the FIREhouse. Internet and phone lines have recently been added to accommodate the TPC project.

TPC Lab Space

Appropriate light lab space has been identified at LANSCE for use by the TPC project and the work to make it usable has started. The clean bench that will be used to perform open operations on the TPC has been ordered. A fume hood that will be


used for storing targets and working with the more radioactive samples has also been ordered.

Physical and Environmental Interface

The specific interface of the TPC to its environment is another important task that needs to be completed before the detector can be placed in the experimental hall at LANSCE. To begin this effort, a CAD layout of the experimental hall was created to assist with the physical interface of all TPC support equipment, which includes slow control equipment and probes, racks, switches, laser, gas system and power supplies.

Figure 120: Shown here is a CAD representation of the experimental hall. The steel structure

contains a 5-ton crane and shielding blocks that separate the beam area from the light lab space. The

blue box represents the TPC placement volume.

FIREHouse Upgrades

The FIREhouse flight path will be modified in order to support the TPC experiments. Figure 121 is a CAD drawing of the planned upgrades. A gas manifold will be placed outside the building as gas lines installed to supply gas to the gas handling system. A


vent line will vent the gas outside after it has passed through HEPA-filters. The proposed design of the gas supply system is currently being reviewed for pressure safety. In addition to gas lines, there will also be other cables laid out around the building. Cable trays will be mounted on all walls of the FIREhouse for this reason.

Figure 121: Shown here is the TPC stand as it will be mounted in the FIREhouse neutron flight path.

The stand will roll on rails parallel to the beam axis. The rails can be easily removed to

accommodate other activities on the flight path.

The rails used to move the TPC stand along the neutron beam axis will be bolted to the floor in the FIREhouse. They will be relatively easy to remove when necessary for other experiments. A sketch of the FIREhouse with the gas manifold, gas lines and rails are shown in Figure 122.

The hardware for network connections at the FIREhouse was installed in September. This provides different options for the types of networks available in the building, and the so-called “green” LANL network will be used for TPC experiments since it allows for outside connections from non-LANL computers.

The hood and clean bench that was ordered earlier this year has still not been delivered to LANL, but the purchasing request is in progress. Different options for placement in available laboratory space are being evaluated.


Figure 122: Limited drawing of the FIREhouse flight path. The position of the rails used for moving

the TPC stand is shown. Also shown are the gas manifold and gas lines used to supply gas to the TPC

gas handling system.

Ohio University [OU]

The work proposed will be undertaken in the Edwards Accelerator Laboratory of the Department of Physics and Astronomy at Ohio University. The laboratory includes a vault for the accelerator, two target rooms, a control room, a thin film preparation and chemistry room with a fume hood, an electronics shop, a teaching laboratory for small non-accelerator based nuclear experiments, and offices for students, staff, and faculty. The Laboratory building supplies approximately 10,000 square feet of lab space and 5,000 square feet of office space. In the Clippinger Research Laboratories, the Department of Physics and Astronomy has a 3000 square foot mechanical shop, staffed by two machinists, that supports all the experimental work of the department. The machinists have numerically controlled machines that they use in the fabrication of apparatus used in experiments, they are accomplished at making parts from exotic materials such as refractory metals, and can perform heli-arc welding and other sophisticated joining techniques.

The heart of the Edwards Accelerator Laboratory is the 4.5-MV tandem Van de Graaff accelerator and six beam lines. This machine is equipped with a sputter ion source and a duo-plasmatron charge-exchange ion source for the production of proton, deuteron, 3,4He, and heavy ion beams. DC beams of up to 30 µA are routinely available for protons, deuterons and many other species from the sputter ion source. Pulsing and bunching equipment are capable of achieving 1 ns bursts for proton and deuteron beams, 2.5 ns bursts for 3,4He beams, and 3 ns bursts for 7Li. The accelerator belt was replaced most recently in March 2004; the accelerator has performed very well since that time with good stability for terminal voltages up to 4.0 MV. The SF6 compressor and gas-handling system were refurbished in April 2005. The Laboratory is very well equipped for neutron time-of-flight experiments. The building is very well shielded thus allowing the production of neutrons from reactions


such as d(d,n). A beam swinger magnet and time-of-flight tunnel allow flight paths ranging from 4 to 30 m. The tunnel is well shielded, and the swinger-magnet assembly allows angular distributions to be measured with a single flight path.

Management

The NIFFTE university collaborators are funded under a NERI-c contract. The PI of this contract has reporting responsibilities for that grant. The laboratories are being funded directly through the AFCI Reactor R&D campaign to not only participate but to provide guidance and project oversight, including reporting requirements within the AFCI management system.

References i G.N. Knyazheva, et al., Nucl. Instru. Meth. Phys. Res. B 248, 7 (2006)

ii Stoner, Arizona Carbon Foils, 2008

iii B. Lommel, et al., NIM A590, 141(2008); B. Kindler, et al., NIM A 590, 126 (2008);

B. Lommel, et al. NIM A480, 199 (2002)

iv S.T. Keller, “Modeling the OSU TRIGA reactor, TRTR Conference, September, 2007

v C. S. Jolivet and J.O. Stoner, NIM A 590, 51 (2008)

Annual Report Fission Time Projection Chamber Projectoregonstate.edu/.../TPC/2009_Annual_Report_Final-Draft.pdf · 2011-07-21 · Annual Report Fission Time Projection Chamber Project

Documents