Tim Vidmar

SCK•CEN, Belgian Nuclear Research Centre

Workshop “Metodi Avanzati di spettrometria gamma”

Milan, 10-12 November 2010

Efficiency Transfer

Milan 2010

#2 / 8

T. Vidmar (Belgium)

Introduction

� Samples may vary in size, composition and density

� Preparation of standards time consuming and costly

� Can we use a computational approach?

� Efficiency transfer – “transfer” of the efficiency of a

standard to another sample

Milan 2010

#3 / 8

T. Vidmar (Belgium)

Introduction

� Standards still required - but only a few!

� The efficiency of the measured sample is calculated

� The measured sample can differ from the standard in:

� Size – geometry correction

� Density – density correction

� Composition

� Any given combination of these

� Calculation can only be done with a computer

� Model of the detector and the sample are required

Milan 2010

#4 / 8

T. Vidmar (Belgium)

Moens theorem

� Virtual peak-to-total ratio independent of the sample for a given

gamma-ray energy

� Virtual efficiencies do not take scattering into account

� Virtual peak efficiency equivalent to the usual full-energy-peak

efficiency

� Virtual total efficiency can only be calculated (except for a point

source)

� P1/T1 = P2/T2 for two different samples 1 and 2

Milan 2010

#5 / 8

T. Vidmar (Belgium)

Moens theorem

� P/T = (τ+κσ)/µ

� τ = photo-effect absorption coefficient

� σ = Compton absorption coefficient

� µ = Total absorption coefficient

� κ κ = percentage of Compton scattered gamma-rays which eventually end

up in the full-energy peak (do not escape)

� κ can be shown to be independent of the source position

� T.Vidmar, A. Likar On the invariability of the total-to-peakratio

in gamma-ray spectrometry. Applied Radiation and Isotopes 60

(2004) 191–195

� Moens theorem not valid for planar detectors!

Milan 2010

#6 / 8

T. Vidmar (Belgium)

The Method

� P2 / T2 = P1 / T1 � P2 = P1 ( T2 / T1 )

� P1 measured, Reference sample (calibrated standard)

� T2 and T1 calculated

� P2 efficiency for the desired (analyzed) sample

� In the calculated ratio (T2/T1) many inaccuracies of the detector

model cancel out!

� Detector data can be taken from the manufacturers sheet

� Simultaneous “correction” for geometry and self absorption

� Alternative: P2 = P1 (ε2/ε1)

� ε2, ε1 full-energy-peak efficiencies � Monte Carlo simulation

Milan 2010

#7 / 8

T. Vidmar (Belgium)

Virtual Total Efficiency

Absorption:

F = exp(-µss) exp(-µww) …

Interaction:

R = 1-exp(-µGeL)

Registration:

pi = Fi Ri

Efficiency:

T = (1/N) Σ pi

Milan 2010

#8 / 8

T. Vidmar (Belgium)

Study

� M-C. Lepy et al., EUROMET Action 428: Transfer of Ge detector efficiency

calibration from point source geometry to other sources

� Various approaches

� Direct Monte Carlo with and without optimization

� Efficiency transfer with Monte Carlo efficiencies

� Efficiency transfer with the Moens method

� Semi-empirical methods

� Point-to-point, point-to-extended and extended-to-extended source transfers

� Efficiency transfer found to be much more consistent and reliable than direct

calculation

� Transfer with peak efficiencies (involves Monte Carlo calculations) best

� Accuracy sufficient for environmental measurements, especially for self-

absorption correction

Milan 2010

#9 / 8

T. Vidmar (Belgium)

The Originators

� Moens, L., De Donder, J., Xi-lei, L., De Corte, F., De

Wispelaere, A., Simonits, A., Hoste, J., 1981. Calculation of the

absolute peak efficiency of gamma-ray detectors for different

counting geometries. Nuclear Instruments and Methods in

Physics Research, 187, 451-472.

� Point-source reference measurements

� Sensitivity study

� Method very sensitive to the detector diameter!

� Otherwise quite robust

� Sensitivity increases with decreasing energy

Tim Vidmar

SCK•CEN, Belgian Nuclear Research Centre

Workshop “Metodi Avanzati di spettrometria gamma”

Milan, 10-12 November 2010

Efficiency Transfer –

Experimental Verification

Milan 2010

#11 / 8

T. Vidmar (Belgium)

Study

� Water solution samples with known activities: Am-241, Cd-109,

Ce-139, Sn-113, Co-57, Cr-51, Sr-85, Co-60 and Y-88

� Point source measurements

� Transfer from point source and between different extended

samples

� Efficiency transfer between extended sources.

Tim Vidmar, Branko Vodenik, Marijan Nečemer, Applied

Radiation and Isotopes, 68 (2010) 2352–2354

Milan 2010

#12 / 8

T. Vidmar (Belgium)

Standard solutions

39.4111387.2I

37.390240.9G

19.990128.3F

10.19065.2E

19.36055.5D

11.76033.6C

4.96014.0B

5.05010.0A

h [mm]d [mm]m [g]Sample

Milan 2010

#13 / 8

T. Vidmar (Belgium)

Detector Data

AlHousing material

1Housing thickness

5Crystal-to-window distance

AlWindow material

1Window thickness

80Window diameter

45Central cavity depth

10Central cavity diameter

1Side dead layer thickness

1Top dead layer thickness

0Crystal rounding (bulletization)

58.8Crystal length (including the top dead layer)

60.5Crystal diameter (including the side dead layer)

GeCrystal material

pCrystal type

Value Parameter

Milan 2010

#14 / 8

T. Vidmar (Belgium)

Point source measurements

Milan 2010

#15 / 8

T. Vidmar (Belgium)

Transfer from point source

7.25.78.97.29.09.18.67.31836

5.65.67.28.07.47.86.25.91333

5.74.15.44.66.47.87.24.91173

6.65.99.87.89.011.910.28.1898

7.16.68.38.69.710.510.28.9662

8.98.610.710.911.312.611.811.0392

8.37.89.29.29.210.810.58.9320

15.112.714.615.015.917.416.614.2166

17.016.018.018.018.019.519.017.2122

11.19.911.314.512.812.215.112.788

10.08.08.811.29.08.811.25.560

IGFEDCBAE [keV]

Milan 2010

#16 / 8

T. Vidmar (Belgium)

Transfer between extended samples

1.2-2.91.6-1.6-0.10.51.11836

0.0-1.5-0.80.6-0.31.50.31333

1.5-1.20.8-1.7-1.30.52.21173

0.6-3.61.9-1.1-2.61.61.9898

0.6-1.6-0.3-1.0-0.80.21.3662

0.3-2.0-0.2-0.3-1.20.70.7392

0.5-1.20.0-0.1-1.40.31.4320

2.1-1.6-0.3-0.8-1.30.72.1166

0.8-1.70.1-0.1-1.20.41.5122

1.0-1.2-2.81.50.5-2.52.188

2.0-0.6-2.21.90.2-2.25.460

I/GG/FF/EE/DD/CC/BB/AE [keV]

Tim Vidmar

SCK•CEN, Belgian Nuclear Research Centre

Workshop “Metodi Avanzati di spettrometria gamma”

Milan, 10-12 November 2010

Efficiency Transfer –

Computer Codes

Milan 2010

#18 / 8

T. Vidmar (Belgium)

General purpose packages

� Written for large experiments in particle physics, medical

physics, shielding applications, criticality calculations, etc

� Mostly require programming, quite complex to use

� Track various kinds of particles

� Perform full Monte Carlo calculations

� Can handle arbitrary geometries – unusual objects

� Run times between minutes and hours

� Variance reduction techniques available

Milan 2010

#19 / 8

T. Vidmar (Belgium)

General purpose packages

� MCNP – 40 years of development, no programming (input decks only

� GEANT3 - Fortran, geometry description in terms of volumes, no longer developed

� GEANT4 – C++, general purpose, developed at CERN, no input decks

� PENELOPE – Fortran, e- and gamma, very precise interaction treatment, limited programming required

� EGS – Mortran, popular in medical community

� FLUKA – developed at CERN, Fortran, rather recent

Milan 2010

#20 / 8

T. Vidmar (Belgium)

Specialized codes

� Specifically aimed at gamma-ray spectrometry

� No programming, user friendly interfaces

� Various approaches

� Full Monte Carlo (spectra available)

� Virtual total efficiencies – Effective solid angle

� Much quicker than general packages

� Limited geometry

� Not all of them (yet) integrated with major manufacturer’s

platforms

Milan 2010

#21 / 8

T. Vidmar (Belgium)

Specialized codes

� SOLANG – “grand daddy” of coincidence correction codes,

developed by the method’s originators

� GESPECOR – developed in the 90’s, commercial, widely used,

also by NMIs, constant improvements, all major detector and

sample types covered, abundant references, full Monte Carlo

� ETNA – Effective solid angle, well-type and cylindrical

detectors, Marinelli beakers

� DETEFF – Monte Carlo, special attention to low energies

� EFFTRAN – Effective solid angle, fast, limited to cylindrical

symmetry

� Your own code? – use XCOM material data!

Milan 2010

#22 / 8

T. Vidmar (Belgium)

Major manufacturers

� CANBERRA – ISOCS and LABSOCS

� All imaginable sample geometries and materials

� Detector characterization required

� Sufficient accuracy?

� ORTEC – ANGLE, part of GammaVision

� Fully integrated with the spectrum analysis

� Planar, coaxial and well-type detectors

� Cylindrical, rectangular and Marinelli samples

� Calculates effective solid angle

Milan 2010

#23 / 8

T. Vidmar (Belgium)

The choice is yours …

� Do I really need efficiency transfer?

� Am I willing to do some programming?

� Would I rather stick to major manufacturers’ products?

� Do I work with an (semi)-automated analysis?

� How well do I know my detectors?

� Is there not a simple, tested, one-fit-all solution?

Tim Vidmar

SCK•CEN, Belgian Nuclear Research Centre

Workshop “Metodi Avanzati di spettrometria gamma”

Milan, 10-12 November 2010

Efficiency Transfer –

Computer Codes II

Milan 2010

#25 / 8

T. Vidmar (Belgium)

Study of efficiency transfer codes

� Codes were compared to each other for a set of well defined

detector and sample geometries

� No reference to experimental data

� Testing Efficiency Transfer Codes for Equivalence.

T. Vidmar et al., Applied Radiation and Isotopes, 68 (2010)

355-359.

Milan 2010

#26 / 8

T. Vidmar (Belgium)

Codes and Modes

xxPENELOPE PENCYL

xxPENELOPE 2008

xxPENELOPE 2003

xxMCNPX

xxMCNP5

xxMCNP4C

xxGESPECOR 4.2

xxxGEANT 3.21

xxEGS4

xxETNA

xxEFFTRAN

xxxDETEFF 4.2

xxANGLE

GeneralSpecializedTEFEPE

Code TypeET ImplementationComputer Code

Milan 2010

#27 / 8

T. Vidmar (Belgium)

Detector models

AlAlHousing material

11Housing thickness

8080Housing length

55Crystal-to-window distance

BeAlWindow material

0.51Window thickness

8080Window diameter

4040Hole depth

1010Hole diameter

0.00031Dead layer thickness (top and side)

6060Crystal length (including the top dead layer)

6060Crystal diameter (including the side dead layer)

GeGeCrystal material

npCrystal type

Detector BDetector AParameter

Milan 2010

#28 / 8

T. Vidmar (Belgium)

Samples

0.00.01.00.0Container-to-detector-window distance

CelluloseQuartz-WaterSample material

3400.020Sample height

(including the container)

80900.060Sample diameter

(including the container)

FilterSoilPointReferenceParameter

Milan 2010

#29 / 8

T. Vidmar (Belgium)

Materials

C8H

91.050Plastics (polystyrene)

C6H

12O

60.200Cellulose

SiO2

1.400Quartz

H2O1.000Water

Be1.848Be

Al2.700Al

Ge5.323Ge

Chemical formulaDensityMaterial

Milan 2010

#30 / 8

T. Vidmar (Belgium)

Water sample, Detector B

Milan 2010

#31 / 8

T. Vidmar (Belgium)

Results – Standard Deviations

0.80.80.60.71.00.72000

0.90.80.50.51.10.71000

0.90.60.30.80.90.8500

0.80.60.60.60.70.7200

0.60.60.50.70.40.5120

0.50.80.70.90.40.680

0.51.00.90.90.50.660

0.80.90.50.90.90.545

1.22.51.420

Filter BFilter ASoil BSoil APoint BPoint AEnergy

Milan 2010

#32 / 8

T. Vidmar (Belgium)

Results – Maximum Deviations

1.31.40.91.51.91.92000

1.51.41.20.92.01.61000

1.51.10.51.71.91.4500

1.41.21.61.21.41.1200

1.11.21.61.80.80.9120

1.12.01.62.01.01.280

1.02.52.31.91.11.260

1.42.21.33.11.71.245

3.4-7.6-3.1-20

Filter BFilter ASoil BSoil APoint BPoint AEnergy