INNOVATIONS IN GAMMA SPECTROSCOPY SOFTWARE or More reasons

INNOVATIONS IN GAMMA SPECTROSCOPY SOFTWARE or More reasons why we should appreciate Mathematics Teachers Frazier Bronson CHP Canberra 18 th Annual RETS-REMP Workshop June 23 -25, 2008 Charlotte, NC

Agenda u A brief review of the first millennium of gamma spectroscopy in 5 minutes u Efficiency Calibrations u Cascade Summing Corrections u Future changes 2 2 2

Stone Age Gamma Spectroscopy [when I started doing this] u Otto Frisch, 1951 w 30 ch, 100 cps, 6 Hz u Tullamore – 1954 w 20 SCAs, x 5 switch = 100 ch u State of the art: w Human readout w Graph paper display w Adding machine net peak area w Radioactive efficiency standards w Graphical efficiency curve fitting 3 3 3

Middle Ages of Gamma Spectroscopy u Wilkinson ADC – ‘ 55, RIDL u Transistorized MCA – ’ 59 ND, TMC u TMC Blue Box 1 ft 3 MCA – ’ 61 u Teletype ASR 33 – ’ 63 u Central Computers – big, expensive, remote u State of the art: w Human peak search w Computer net peak area w Radioactive efficiency standards w Graphical efficiency curve fitting 4 4 4

Renaissance – life is getting easier u Ge(Li) detectors – ’ 64 Ortec, RCA, SSR, Nuclear Diodes u Portable Multi-Attitude Cryostat – ’ 68 Harshaw u HPGe detectors – ’ 73 PGT u Battery powered Field MCAs – ’ 81 ND u Computer as part of MCA – ’ 71 Tennecomp u State of the art: w DEC computers w Computer based peak analysis w Computer assisted graphical efficiency curve fitting w Radioactive efficiency standards in lab w HASL-258 infinite field efficiency calibration – ‘ 72 5 5 5

How we used to calibrate large things “in the olden days” when we had to walk 2 miles to school up-hill both ways The “Lego” calibration method - radioactive building blocks u Animal counters w Convertible sheep/cow/pig/snake phantom What’s wrong with u 32 detector WBC w Adjustable man – 70 to 400 lbs u Large “Box” counters this picture? Plenty u Soil conveyor assay system 6 6 6

Computers have changed everything they allow us to do mathematical modeling, efficiency calibrations, and data corrections u ’ 40 s - Fermi, von Neumann, et. al. started work at Los Alamos on Monte Carlo process for nuclear modeling u ’ 77 – MCNP released to general public u Canberra’s MCNP uses w ’ 80 s – used to model neutron assay systems w ’ 90 s – used to model gamma assay systems w ’ 94 – used to calibrate customer gamma systems [waste, In. Situ] w ’ 96 – used in production process for ISOCS software u Mathematical efficiency calibration software w Canberra: ‘ 97 – ISOCS for In. Situ; ’ 98 – Lab. SOCS for Laboratory geometries w Ortec: ’ 99 – Program Isotopic for boxes and drums u ’ 99 – Cascade summing corrections – Ortec, Canberra 7 7 7

Calibration methodologies Radioactive source calibrations Mathematical calibrations u Need appropriate licenses u Construct exact mathematical model of the item to be measured u Need radioactive sources w Sources with wide range of energies and proper activity w Without cascade summing u Need qualified and trained personnel u Construct exact duplicate of item to be measured w Size, shape, density, Z, w The most difficult part u Count the calibration source and construct the calibration curve u Interpolate to energy of interest u Compute the efficiency calibration at exact energies of interest u MCNP and others w Can be quite accurate as long as detector is well known l Helmer and Hardy – 0. 2% w Complicated and slow u ISOCS/Lab. SOCS w Generally good enough ~5% w Easy and fast w Accepted - About 1000 Ge detectors Characterized for ISOCS/Lab. SOCS use 8 8 8

Calibration process with ISOCS or Lab. SOCS u Select or create sample container u Enter characteristics of sample w height, density, material u Define any absorbers w e. g. sample shelf u Select detector from list u Enter source-detector distance u Compute energy/efficiency/error datapoints [5 -10 sec] u Compute efficiency curve u Use for sample analysis 9 9 9

ISOCS Calibrations 20 different sample shapes to choose from n n n n 10 Wide variety of sample shapes Multiple adjustable parameters for each sample shape Multiple sources & locations within each sample shape Sample sizes from points to VERY large Any location within 500 meter radius of detector Any energy from 45 - 7000 ke. V Collimators, both cylindrical and rectangular 10 10

Lab. SOCS Calibration Templates u box u cylinder from side u cylinder or cone from end u sphere u Marinelli beaker u complex shape [cylindrically symmetric] 11 11 11

If pre-defined materials not adequate, create your own with the Material Editor u Any material can be created by the user u All elements and all cross-sections are in library u Materials can be defined by chemical formula [H 2 O] or previous material [drydirt] to make combination [mud] 12 12 12

Mathematical Calibrations are better in many cases u Calibrate difficult matrices w Concrete, soil, gas, steel, oil, u Calibrate things at any density, don’t waste time adjusting sample to match calibration density w Vegetation, soil, u Calibrate at any Z Ratio of efficiency compared to efficiency at 0% Fe w Important at low E Fig 2 Efficiency error vs Fe soil content 1. 2 1. 1 1. 0 0. 9 0. 8 0. 7 60 ke. V 88 ke. V 122 ke. V 0. 6 0. 5 0. 4 0 13 5 10 Wt% of Fe in soil 15 13 13

Use Lab. SOCS to make your detector seem larger Used Lab. SOCS to determine optimum sample container shape for 500 cc sample for 3 different shapes of detectors l l Typical container used in laboratory l 7. 5 cm dia x 11. 3 cm deep Optimum 500 cc container l 12 cm dia x 4. 5 cm deep - Can gain ~30% higher efficiency with optimum container - Can then cut count time approx in half 14 14 14

Simplify your sample preparation with infinite geometry calibrations [1] u Conventional calibration w Units: counts/gamma = cps/gps w Calibration behavior Sample diameter Sample thickness Efficiency Sample density 15 15 15

Simplify your sample preparation with infinite geometry calibrations [2] u Infinite calibration [thickness shown here] w Units: cps/gram = efficiency * mass w Calibration behavior Sample diameter Sample thickness Efficiency Sample density u Fill container more than 10 cm u No volume or mass measurements u No density corrections u One efficiency calibration u Results in activity/gram u Saves time and labor 16 16 16

More tricks for Efficiency Calibrations Don’t do it !! u Traditional method: w Create calibration file for common geometries l Container type, shelf, volume, matrix, density w Prepare sample to match exact calibration geometries l Takes time and labor, curve fitting and interpolation errors u Today: w Fill container with any material, volume, density l Use all of sample; w Place it anywhere w Start count w Tell software the geometry l Container type, material, volume or mass, w Software finds peaks, computes efficiency at exact energy 17 l No curve fitting or interpolation l No calibration process to maintain, just QA to show nothing has changed 17 17

Summing Errors in Gamma Spectroscopy Random Summing w loss of peak area at very high count rate Cs-137 w independent of energy, sample-to-detector distance, number of nuclides in the sample w correctable with the use of a pulser or a stationary reference source g 1 True Coincidence Summing [Cascade Summing] w loss or gain of peak area as a function of nuclide decay scheme and geometry w independent of count rate - a problem for all samples Co-60 w different effect for different gamma energies of even the same nuclide w correctable by increasing the sample-to-detector distance, if you want to pay the lower MDA penalty w Many common nuclides have this problem l Co-60, Y-88, Ce-139, Eu-152, Cs-134, Ba-133 l Self-correcting if calibration and assay nuclide are SAME l l 18 g 1 g 2 For other nuclides with nearby energies, calibrations with these nuclides will be wrong [Cs-134, Zn-65, Fe-59, K-40, …] The higher the efficiency, the worse the problem 18 18

Magnitude of Cascade Summing Errors u Example is Eu-152 with 60% efficiency HPGe u 400 ml Marinelli beaker geometry [shown below] w from (-45%) for 244 ke. V to (+10%) for 1086 ke. V. u Point geometry: w from (-60%) for 244 ke. V to (+15%) for 1086 ke. V u For many other nuclides with many coincidence transitions the effect may be even larger. u Current Canberra and Ortec software can correct these errors 19 19 19

Genie 2000 Cascade Summing Correction Input needed Done once per detector P/T Calibration Done once per sample geometry Geometry Description From analysis Automatically done Cascade Summing Correction Corrected NID Results 12/1/2020 20

Geometry Description Works best with Lab. SOCS-Characterized detector and efficiency, but that isn‘t required Select sample shape Enter parameters describing sample Geometry Composer Creates file used in analysis Geometry Description (*. GEO) Internally, the calculations use the ISOCS/Lab. SOCS engines 12/1/2020 21 21 21

P/T (Peak to Total) Calibration automated feature of v 2. 0 241 Am 109 Cd 54 Mn Quick and easy to do Once per detector [not per geometry] Low cost uncalibrated button sources single gamma preferred no cascade gammas Just count at approx. 10 cm the software does the rest 137 Cs 113 Sn P/T calibration curve 12/1/2020 22 65 Zn 22 22

Current Cascade Summing Correction Results u Typical contact filter paper before and after CSC w Within 5% u Accuracy as a function of detector size w Works for all size detectors u Limitations: w Only for gamma-gamma coincidences w X-rays a problem on low energy detectors w Requires simple P/T calibration 23 23 23

What does the future hold ? ? ? u Improvements in ISOCS/Lab. SOCS w 3 -D graphical user interface for better visualization of the exact geometry modeled w 2 new templates for more complex modeling w 4 new complex collimators for waste assay applications w Updated mass attenuation data [2003 vs. 1982] w Lower energies – down to 15 ke. V w Uncertainty Estimator and Assay Planner w Lab. SOCS Complex Beaker Viewer and Editor w More features in Complex Beaker template u Cascade Summing improvements w Updated CSC nuclide library – now 205 nuclides and 2366 lines w Now include Xray-Gamma summing and Xray-511 summing w Total efficiency automatically calibrated – no sources need 24 24 24

25 25 25