Author's Accepted Manuscript

Application of Neutron Tomography in Culture Heritage Research T. Mongy

www.elsevier.com/locate/apradiso

PII: DOI: Reference:

To appear in:

S0969-8043(13)00448-X http://dx.doi.org/10.1016/j.apradiso.2013.11.028 ARI6411

Applied Radiation and Isotopes

Received date: 31 July 2013 Revised date: 2 October 2013 Accepted date: 12 November 2013 Cite this article as: T. Mongy, Application of Neutron Tomography in Culture Heritage Research, Applied Radiation and Isotopes, http://dx.doi.org/10.1016/j. apradiso.2013.11.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ApplicationofNeutronTomographyinCultureHeritageResearch

T.Mongy AtomicEnergyAuthority(AEA)ofEgypt,EgyptSecondResearchReactor(ETRR2),P.O.Box 13759,AboZaabal,Cairo,Egypt Email:tmongy@gmail.com Phone:+201112247943

Abstract: Neutron Tomography (NT) investigation of Culture Heritages (CH) is an efficient tool for understanding the culture of ancient civilizations. Neutron Imaging (NI) is astatoftheart nondestructivetoolintheareaofCHandplaysanimportantroleinthemodernarcheology. NItechnologycanbewidelyutilizedinthefieldofelementalanalysis.

AtEgyptSecondResearchReactor(ETRR2),acollimatedNeutronRadiography(NR)beamis employed for neutron imaging purposes. A digital CCD camera is utilized for recording the beam attenuation in the sample. This helps for detection of hidden objects and characterization of material properties. Research activity can be extended to use computer software for quantitative neutron measurement. Development of image processing algorithmscanbeusedtoobtainhighqualityimages.

In this work, full description of ETRR2 was introduced with up to date neutron imaging system as well. Tomographic investigation of a clay forged artifact represents CH object was studied by neutron imaging methods in order to obtain some hidden information and highlight some attractive quantitative measurements. Computer software was used for imaging processing and enhancement. Also Astra image 3.0 Pro software was employed for highprecisemeasurementsandimagingenhancementusingadvancedalgorithms.

ThisworkincreasedeffectiveutilizationofETRR2NeutronRadiography/Tomography(NR/T) techniqueincultureheritagesactivities.

KeyWord:NeutronImaging,NeutronTomography,NeutronRadiography,CultureHeritage.

1.Introduction: ERRR2 Neutron Radiography (NR) facility was commissioned in 1999. The first experiments carried out were determination of characterization parameters, such as, flux map, n/ ratio, Cd ratio and spatial resolution. The results were compared with MCNP calculations. Internal details were detected for different samples using nitrocellulose film. Photographic film was replaced by nitrocellulose film to get high quality image formation. A lot of experiments were performed toward scientific research and quality assurance as well. Welding inspection,measuringwaterpermeabilityinbuildingmaterialsandimagingenhancementby scatteredneutrondeblurringwerethemajorexperimentsachievedinthepast.

Fast Neutron Resonance Radiography (FNRR) was also introduced using portable neutron source.Thecharacterizationparametersweredeterminedfortheneutronsource.Detection of hydrogenous components was implemented to study water permeability in building materials using quantitative FNRR (T. Mongy, 2011). Capabilities of underground water migration in soil were also investigated and implemented to preserve cultural heritage monuments.

Gammarayradiography(R)techniquewasusedtoinspectweldingmaterials.NRtechnique was compared with R in 2011. Dynamic System Neutron Radiography (DSNR) was pre commissioned using DELCam camera. High definitions, high resolution and high quality imageswereobtainedusingimagingprocessingtechnique.

2.DescriptionofETRR2:

ETRR2 is an open pooltype research reactor with variable core arrangement (F. Esposto, 1999). The core power is 22 MWth cooled by light water, moderated by water and with beryllium reflectors. The design concept is based on versatile utilizations, It has been mainly designedfor:

1Basicandappliedresearchinreactorphysicsandnuclearengineering, 2Neutronradiographyforresearchandindustrialpurpose, 3Radioisotopeproductionformedicalandindustrialpurposes, 4 Beam holes experimentation for neutron scattering experiments and neutron radiography, 5Materialtesting, 6Materialirradiation, 7Activationanalysis, 8Trainingofscientificandtechnicalpersonnel.

ETRR2 has four neutron beams and a thermal column as the main experimental devices allowneutronsourcesoutsidethereactorcore.Thefourneutronbeamstubesare:

1TheNeutronRadiographyFacility, 2TheRadialBeamTube, 3TheTangentialBeamTube, 4TheUnderwaterNeutronRadiographyFacility.

Figure 1 represents the reactor tank with its internals. The Figure shows: passing mechanism, pool cooling pipe, thermal column, beryllium block, tangential tube, guide box, control rods, chimney, core cooling system pipe, pool cooling system diffuser, fuel elements, reflectors, irradiation grid, irradiation chamber shield, suctionbox, radialtubeand UnderWaterNeutronRadiography(UWNR)tube.

3.StateoftheArtNeutronImagingFacilityatETRR2:

ThecommissioningofstateoftheartnewNeutronImaging(NI)systemisstartedattheend of March 2012 under the frame work of TC communication between AEA and IAEA. The lay out of the system is shown in Figure 2. The NI system was installed to replace static based filmneutronradiography.
2

Figure1.ETRR2tankwithitsinternals.

Figure2.Layoutforneutronradioscopysystemwithascintillationscreenand cooledCCDcameraasusedinETRR2.

For NI, the neutrons are attenuated and a sufficient amount of light is produced by a scintillation screen that detected with a CCD camera. The scintillation screen used had a composition of LiF:ZnS:Cu (L.G.I. Bennett, et al., 2001). The neutrons interact with lithium due its large cross section to produce an alpha particle (4He++) and tritium (3H) daughter products. The energy from alpha particle is deposited into zinc sulfide (ZnS) and efficient phosphor producing visible light. The copper element acts as a wave length shifter to producelightintheyellowgreenregion,whichhasanaveragewavelengthof525nm.

TheuseofCCDcamerasystemhasanadvantageoftomographyexperimentspossibility.The CCDcamera with itslenses,mirrorand integratedcooling unitarehoused in ashielded light tight aluminum box. Table 1 summarizes the CCD camera system with technical specifications.

Camware program was employed for camera control and image acquisition. Image acquisitionandimageprocessingwasaccomplishedbyusingImageJ.VGStudioMAXsoftware from Volume Graphics was used for data processing, visualization andanimation of image capturing.Forimagereconstruction,Octopussoftwarewasalsoapplied.

Table 2 summarizes the used software for data reconstruction with specific technical functions.

Table1EquipmentusedforNTinETRR2withspecifications

Equipment PCO2000CCDthermoelectricalcooling camerasystemcompactwithpowersupply Lenses Mirror Scintillator Lighttightbox

Nikon,50mmfocallength. 45degree, highreflectivitypolishedsilicon. 6LiF+ZnS. Aluminum.

Technicalspecifications Highresolution(2048x2048pixel), 14bitdynamicrange, Imagerateof14.7fps@full resolution, Lownoise, Lowdarkcurrent, Pixelsize(hor.xver.)is7.4x7.4 m2, 4GBcameramemory.

Camware ImageJ

Table2Softwarefordatareconstructionwithspecificfunctions. Software Functions Cameracontrol, Imageacquisitionandarchivingin variousfileformat. Display.Edit,analyze,process,save, print8bitcolorandgrayscale, Readimagewithdifferentimage formatting.

LabVIEW Octopus VGStudioMax

Dateacquisition, Motioncontrol. Imageprocessing, Imagereconstruction, Imageanalysis, Singlesliceevaluation, Tuningreconstructionparameters withoutprocessingthecomplete volume, 3Dviewer, Readimagewithdifferentimage formatting. 3Dimageprocessingandanimation, VolumetricDataanalysisand measurement, Highqualityinnerstructure visualization, Volumeelementcharacterization.

4.ApplicationofNTinCultureHeritageResearch:

The experiment was carried out at ETRR2 Neutron Radiography/Tomography Facility (NR/TF). The facility characterization parameters are summarized in table 3 (M. A. Abou Mandouretal.,2007).

Table3.CharacterizationparametersofETRR2NR/TF

Parameters Inletaperture(D)ofthecollimator Outletapertureofthecollimator(beam openingatsampleposition) Sourcetoobjectdistance(L) L/D(resolutionparameter) Thermalfluxatsampleposition(nth) (nth/)ratio Cdratio(a) a Ratiobetweentotalandfastfluxes.

Characterizations 3cm 22cm 351.8cm 117.3 1.5*107neutrons/cm2s 105n.cm2.mR1 10

Figure 3 illustrates scheme diagram of neutron radiography beam tube showing collimator inletaperture(D)andsourcetoobjectdistance(L).

Figure3.Schemediagramofneutronradiographybeamtube

A simulated forged pottery Peruvian artifact, shown in Figure 4, provided by IAEA was scanned to enhance reconstructed image formation resulting from Octopus reconstruction software. For a three dimensional tomographic reconstruction, transmission images of the artifactwere taken fromdifferent view angles(1.2 degree per projection). A rotational table is driven by a stepper motor connected to a computerized motion control system, was installed at the sample position. The number of projections to full scan the artifact with 360 degree were 300 projections. To synchronize the Camware software (CCD camera software) and the rotational table; a new software based on Copley Controls macro language was designedandused.

Reasonable exposure time was adjusted at 10 Sec. per projection with 10 Mwatt reactor power operation, it means about one hour was sufficient to full scan the object including imageprocessingcorrection(darkcurrentandflatfieldcorrection).

Dark current images are offset images taken at the same irradiation time (10 sec.) averaged over 10 images, in this experiment. Flat field images are open beam images taken at the sameirradiationtime(10sec.)averagedover10images.

To correct recorded beam intensity inhomogeneities, image normalization was calculated, byOctopussoftware,accordingtoequation(1):

Normalizedimage=(TomographyimageAveragedarkcurrentimages)/(Flatfieldimage Averagedarkcurrentimages)(1)

For dark and flat field (open beam) images, the correction was done by applying a median filter over 10 images by Octopus software. For each pixel the median pixel value (within the 10 values of the 10 images) was taken to obtain the resulting dark current and flat field (openbeam)images.
Figure4.ThePeruvianforgedpotteryartifactonitsscanningpositionatthebeamopening.The artifacthas26cmheight,12cmwidthand9cmside.

As chargeless particles, neutrons can deeply penetrate objects and nondestructively provide information about the inner structure and composition of materials (Chadwick, 1932). The resultant 3D images reconstructions of the artifact after applying imaging processingenhancementbyOctopusareshowninFigure5.

Figure5.ImagereconstructionandenhancementusingOctopussoftware.

For imaging enhancement resulting from Octopus, Astra Image 3.0 Pro was used. Astra is image processing software used for, debluring, sharpening and analyzing 3D image formation.

To extract features from the scanned article, Astra Image Pro was used by implementing Point Spread Function (PSF), Cauchy type, with 2.0 PSF size and 20 iterations, high quality results were provided. Figure 6 shows 3D image reconstruction and enhancement by Octopus software (left). Right one represents Imaging enhancement of the article by MaximumEntropyDconvolutiontoolinAstraimagePro.

Figure6.ImagingenhancementusingAstraImagePro.

From the figure, enhanced, sharp and high quality image we obtained by Astra Image Pro. A lotofprotrusionshaveappeared,i.e.,thearticleisfullofunlimitedprotrusions.

Astra Image 3.0 Pro is powerful software to calculate line profile and image statistics. The scanned line profile(AA), inFigure 7, was plotted forthe 3D image formation resulting from Octopus reconstruction software. The enhancement was carried out by the former mentioned tool using Astra Image Pro (right image in Figure 7). The line profile is shown in Figure8.

AstraImage prohasfunctionofcalculating imagestatistics,suchastheminimum(minimum pixel value in the image) and maximum (maximum pixel value in the image), the mean (the mean of all pixels in the image) and the standard deviation (Std. Dev.). From the definition, the standard deviation of all pixels in the image measures how much variation or dispersion fromthemean.Alowstandarddeviationindicatesthatthedatapointstendtobeveryclose tothemean;andhencehighqualityimageformationisobtainwithnoisesuppression(IanT. Youngetal.,2007),ontheotherhand,highstandarddeviationindicatesthatthedatapoints are spread out over a large range of values, and hence weak image formation is obtained withpresenceofnoise.

TheStd.Dev.ofthered(R),green(G)andblue(B)colorsaretabulatedintable4.Thereisno imagedegradationfromStd.Dev.viewpoint. Figure7.ComparisonofenhancedImagereconstructionbyOctopus(left)andimaging enhancementtechniquebyAstraImagePro(right).

Figure8.LineprofileofthescannedlineAAofFigure7(theconversionfactoris0.28 mm/pixel).

Table4Std.Dev.ofenhancedimagereconstructionbyOctopusandAstraImagepro.

Software Octopus AstraImagepro

Color R G B R G B

Std.Dev. 39.330 59.090 72.877 39.330 59.090 72.877

From Figure 7, the article side measured value (9 Cm) represents by pixel number from 125 to 450. i.e., each pixel represents 0.28 mm. Table 5 represents actual, experimental values andpercentageoferrorfordifferentdimensionsofarticlesideviewthatshowninFigure9.

Figure9.Differentlocationsofactualandexperimentalvaluesofarticlesideview.

Table5Actual,experimentalvaluesandpercentageoferrorforsideviewlocationsof Figure8.
Location Line1 Line2 Line3 Porosity(1) Porosity(2) Grit(1) Grit(2)

Actual Values (Cm) 4.00 7.00 8.00 NotMeasured(NM) NM NM NM

Experimental Values (Cm) 4.10 6.92 8.10 0.266 0.212 0.12(height) 0.028(Radius) 0.86(height) 0.084(Radius)

Percentage of error (%) 2.50 (Calculated) 1.40 (Calculated) 1.25 (Calculated) 1.95(assumingMax.error) 1.95 (assumingMax.error) NM 1.95(assumingMax.error) NM 1.95(assumingMax.error)

3D surface plot was created by the software to find out some detail hidden information. To measure the depth of the two porosities 1 and 2 of the enhanced image formation of the article, the rectangular selection tool is used to draw surface plot of the selected surface. Thesurfaceplotgraylevelrangesfrom64.67to255,thisgraylevelprofilerepresents0.5cm depth(measuredvalue). Thusthe conversion factor is0.002 cm/gray level.Figure 10 shows theselectedrectangularsurfacetoobtain3Dsurfaceplot.

Figure103Dsurfaceplotoftherectangularselectedsurface. The distinction of NT for detection of presence of air voids inside material (porosities) is pronounced, as clearly shown in Figure 9. Surface plots tool was employed for the two porosities 1 and 2 to investigate the porosities shapes. Figure 11 shows 3D surface plot of porosities 1 (upper) and 2 (down). Referring to the porosity 1, the deepest depth was found tobe2.66mmandfortheporosity2itis2.12mm.

10

From Figure 11, valuable notifications must be mentioned. First, the porosity 1 is not a normal void; it has a prominence of height 0.665 mm. Second, the porosity 2 is differing fromporosity1andithasaneggshapedwithconcavedepth.

The same tool was used to explore features of grit 1 and 2 of Figure 9. The surface plots were drawn showing 1.2 mm and 0.86 mm heights for the grits 1 and 2 respectively. Also, thelineprofilesweredrawntofindoutgritsradius(theconversionfactoris0.28mm/pixel), itwasfoundtobe0.28and0.84mmrespectively.

Figure113Dsurfaceplotofporosities1(upper)and2(down)

For high quality 3D animation and visualization, VGStudioMax software was used for the creation of impressive animation of the article from volumetric data set imported from Octopus software. Videos from the slice image stacks are created providing another possibility to easily display and share analysis results. Figure 12 shows VGStudioMax

11

providinghighaccuracydeterminationfeaturesonthebasisofvoxelsdatasetfromOctopus software.

The functionality of VGStudio MAX for volumetric data analysis and animation was implemented.Forbackvoidinvestigation,3Dsurfaceplottoolwasusedagaintoexplorethe voidshapeoftheselectedareainFigure12.Theintensityofeachpixeldeterminesitsheight in threedimensional space. Figure 12 shows the 3D surface plots. The void extends through thearticlethickness.

Figure12Backvoidsurfaceplotsofthearticle.

5.Conclusions:

Neutron Tomography (NT) has been developed as a mature scientific method in ETRR2. Nondestructive investigation of Culture Heritages (CH) is important for understanding the culture of ancient civilizations. Neutron Tomography (NT) is astatoftheart non destructivetoolintheareaofCH,andplaysanimportantroleinthemodernarcheology.

6.Acknowledgements:

The author would like to express his deep thanks to Nikolay Kardjilov, Helmholz Institute, Germany, for his assistance and advices. Also, a great thanks from deep heart to D. Ridikas, IAEA,forfacilitatingalldifficultiesinfinancialsupports.

12

References:

F.Esposto,1999.Characterizationoftheneutronbeamfacilities,076713303TABL7021O. Ian T. Young, et al., 2007. Fundamentals of Image Processing Handbook. Delft University of Technology. L.G.I. Bennett, et al., 2001. Neutron Radioscopy Inspection of Composite Flight Control Surfaces.10thAsiaPacificConferenceonNonDestructiveTesting. M. A. Abou Mandour, et al., 2007. Characterization and Application of the Thermal Neutron Radiography Beam in the Egyptian Second Experimental and Training Research Reactor (ETRR2). Hindawi Publishing Corporation, Science and Technology of Nuclear Installations Vol.1. T. Mongy and Mohamed A.Gaheen,2011.DevelopmentofNeutronRadiographyFacility for Detection of Hydrogenous Components Using Recoil Proton Technique. Arab Journal of NuclearSciencesandApplications,44(1),322329.

Research Highlights This manuscript highlighted the following: 1- Neutron tomography is an efficient tool in the field of cultural heritage research, 2- The full description of the ETRR-2 and state-of-the- art neutron tomography system, 3- Implementation of using computer software package in image reconstruction and imaging processing, 4- Precise measurements that was impossible by traditional methods, 5- The manuscript opens the door to investigate ancient Egyptian treasures.

13