Discovery of a big void in Khufu's Pyramid by observation of cosmic-ray muons

14 Pages·2017·7.52 MB·English

Discovery of a big void in Khufu’s Pyramid by
observation of cosmic-ray muons
Kunihiro Morishima, Mitsuaki Kuno, Akira Nishio, Nobuko Kitagawa, Yuta
Manabe, Masaki Moto, Fumihiko Takasaki, Hirofumi Fujii, Kotaro Satoh,
Hideyo Kodama, et al.
To cite this version:
KunihiroMorishima,MitsuakiKuno,AkiraNishio,NobukoKitagawa,YutaManabe,etal.. Discovery
of a big void in Khufu’s Pyramid by observation of cosmic-ray muons. Nature, 2017, 542, pp.386-390.
￿10.1038/nature24647￿. ￿hal-01630260v2￿
HAL Id: hal-01630260
https://hal.inria.fr/hal-01630260v2
Submitted on 22 Nov 2017
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est
archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Thismanuscriptisthepre-submissionmanuscriptprovidedbytheauthors.
Forthefinal,post-reviewversion,pleasesee:
http://www.nature.com/nature/journal/vaap/ncurrent/full/nature24647.html
Discovery of a big void in Khufu’s Pyramid by
observation of cosmic-ray muons
Kunihiro Morishima1, Mitsuaki Kuno1, Akira Nishio1, Nobuko Kitagawa1, Yuta Manabe1,Masaki Moto1 — Fumihiko Takasaki2,
Hirofumi Fujii2, Kotaro Satoh2, Hideyo Kodama2, Kohei Hayashi2, Shigeru Odaka2 — Sébastien Procureur3, David Attié3, Simon
Bouteille3,DenisCalvet3,ChristopherFilosa3,PatrickMagnier3,IrakliMandjavidze3,MarcRiallot3 —BenoitMarini5,PierreGable7,
YoshikatsuDate8,MakikoSugiura9,YasserElshayeb4,TamerElnady10,MustaphaEzzy4,EmmanuelGuerriero7,VincentSteiger5,Nico-
lasSerikoff5,Jean-BaptisteMouret11,BernardCharlès6,HanyHelal4,5,MehdiTayoubi5,6
CorrespondingAuthors:KunihiroMorishima(morishima@flab.phys.nagoya-u.ac.jp)andMehdiTayoubi([email protected])
1F-lab,NagoyaUniversity.Furo-cho,Chikusa-ku,Nagoya,464-8602,Japan—2KEK,1-1oho,Tsukuba,Ibaraki305-0801Japan—3IRFU,CEA,
UniversitéParisSaclay,91191Gif-sur-Yvette,France—4CairoUniversity,9AlGameya,Oula,GizaGovernorate,Egypt—5HIPInstitute,50
rue de Rome 75008 Paris, France — 6Dassault Systèmes, 10 Rue Marcel Dassault, 78140 Vélizy-Villacoublay, France — 7 Emissive, 71 rue
deProvence75009Paris,France—8 NEP,4-14Kamiyama-cho,Shibuya-ku,Tokyo150-0047,Japan—9Suaveimages,N-2MaisondeShino,
3-30-8Kamineguro,Meguro-Ku,Tokyo,153-0051,Japan—10AinShamsUniversity,Kasrel-Zaafaran,Abbasiya,Cairo,Egypt—11InriaNancy
-GrandEst,615rueduJardinBotanique,Villers-lès-Nancy,54600,France
The Great Pyramid or Khufu’s Pyramid was built on the Giza In1986,ateamsurveyedthepyramidusingmicrogravimetry17,thatis,
Plateau (Egypt) during the IVth dynasty by the pharaoh Khufu themeasurementofslightvariationsingravityduetolargevariations
(Cheops), whoreignedfrom2509to2483BC1. Despitebeingone intheamountofmatter18.Basedonthesedata,theteamdrilled3holes
of the oldest and largest monuments on Earth, there is no con- inthecorridortotheQueen’schamberinthehopeoffindinga“hidden
sensus about how it was built2,3. To better understand its inter- chamber”; the team only observed sand17. A more recent analysis of
nal structure, we imaged the pyramid using muons, which are the same data dismissed the theory of a “hidden chamber” where the
by-products of cosmic rays that are only partially absorbed by holeshadbeendrilled17.In1988,aGround-PenetratingRadarsurvey19
stone4,5,6. The resulting cosmic-ray muon radiography allows us suggested that an unknown corridor could be parallel to the Queen’s
tovisualizetheknownandpotentiallyunknownvoidsinthepyra- chambercorridor. Toourknowledge,thistheoryhasbeenneithercon-
mid in a non-invasive way. Here we report the discovery of a firmednorrefuted.
large void (with a cross section similar to the Grand Gallery and Here we follow in the footsteps of Alvarez et al.5 who used spark
a length of30 m minimum) above theGrand Gallery, which con- chambers as muon detectors in Khafre’s pyramid (Kephren) and con-
stitutesthefirstmajorinnerstructurefoundintheGreatPyramid cludedthatthereisnounknownstructurewithavolumesimilartothe
sincethe19th century1. Thisvoid,namedScanPyramidsBigVoid, King’schamberabovetheBelzonichamber5. Muonparticlesoriginate
was first observed with nuclear emulsion films7,8,9 installed in the from the interactions of cosmic rays with the atoms of the upper at-
Queen’s chamber (Nagoya University), then confirmed with scin- mosphere,andtheycontinuouslyreachtheEarthwithaspeednearto
tillatorhodoscopes10,11 setupinthesamechamber(KEK)andre- thatoflightandafluxofaround10,000perm2perminute4. Similarto
confirmedwithgasdetectors12outsideofthepyramid(CEA).This X-rayswhichcanpenetratethebodyandallowboneimaging,theseele-
large void has therefore been detected with a high confidence by mentaryparticlescankeepaquasi-lineartrajectorywhilegoingthrough
threedifferentmuondetectiontechnologiesandthreeindependent hundreds of meters of stone before decaying or being absorbed. By
analyses. These results constitute a breakthrough for the under- recordingthepositionandthedirectionofeachmuonthattraversestheir
standingofKhufu’sPyramidanditsinternalstructure.Whilethere detectionsurface,muondetectorscandistinguishcavitiesfromstones:
iscurrentlynoinformationabouttheroleofthisvoid,thesefindings while muons cross cavities with practically no interactions, they can
showhowmodernparticlephysicscanshednewlightontheworld’s be absorbed or deflected by stones. Put differently,muons traversing
archaeologicalheritage. aregionwithlower-than-expecteddensitywillresultinahigher-than-
The pyramid of Khufu is 139 m high and 230 m wide1,13. There expectedfluxinthedirectionoftheregion. Intherecentyears,muon
arethreeknownchambers(Fig. 1),atdifferentheightsofthepyramid, detectors have been successfully deployed in particle accelerators, in
whichalllieinthenorth-southverticalplane1: thesubterraneancham- volcanology6, to visualize the inner structure of the Fukushima’s nu-
ber,theQueen’schamber,andtheKing’schamber.Thesechambersare clear reactor11,20, and for homeland security21. In heritage buildings,
connectedbyseveralcorridors, themostnotableonebeingtheGrand detectorshavebeenrecentlysetupinarchaeologicalsitesnearRome22
Gallery(8.6mhigh×46.7mlong×2.1to1.0mwide). TheQueen’s andNaples23(Italy),wheretheywereabletodetectsomeknownstruc-
chamberandtheKing’schamberpossesstwo“airshafts”each,which tures from underground, and in the Teotihuacan Pyramid of the Sun
were mapped by a series of robots14,15 between 1990 and 2010. The (Mexico)24. However,sincethemuonsgeneratedbycosmicrayscome
original entrance is believed to be the “descending corridor”, which fromthesky, thedetectorscanonlydetectdensityvariationsinsome
starts from the North face, but today tourists enter the pyramid via a solidangleabovethem(theexactacceptancedependsonthedetection
tunnelattributedtoCaliphal-Ma’mun’s(aroundAD820)1. technology).Inaddition,themuonimagingtechniquemeasurestheav-
MostofthecurrentunderstandingofKhufu’sPyramidcomesfrom eragedensityofthestructureinagivendirection: whileavoidresults
architecturalsurveysandcomparativestudieswithotherpyramids1,2,13. inalowerdensity,itsimpactonthemuonradiographyisdeterminedby
InHistories,HerodotusdescribedtheconstructionofKhufu’sPyramid, theratiobetweenthevoidandthelengthofmaterialthatistraversed.
butthisaccountwaswrittenabout2,000yearslater(in440BC).The Forthisreason,smallcavitieslikeairshaftscannotbedetectedbythis
onlyknowndocumentswrittenduringKhufu’sreignwerediscoveredin techniquewithinareasonableexposuretime.Lastly,inlargeanddense
201316,butthesepapyridescribeonlythelogisticsoftheconstruction, buildingslikeKhufu’sPyramidwhereonlyabout1%ofmuonsreach
suchashowthestonesweretransported,andnottheconstructionitself. thedetectors,thedataneedtobeaccumulatedoverseveralmonths.
1
a b c d
Nagoya and KEK detectors Nagoya and KEK detectors CEA G2 (Brahic) Nagoya KEK
Pyramid central axis 23 m NE1 NE2 H1 H2
South North N
o
rth
QKiunege'sn c'sh cahmabmebrer Grand Gallery South Grand gallery axis
m
7
Subterranean chamber 5.5 m 1.0 m 4.0 m
CEA detectors CEA detectorsGrand Gallery axis 17 m 1.0 m
CEA G1 (Alhazen)
f g h i e King's chamber
Grand Gallery
Queen's chamber
Figure1|MuondetectorsinstalledforKhufu’sPyramid.a,Sideviewofthepyramid,withsensorpositionsandindicativefieldofview.b,Topview.c,Close
viewofthepositionofthegasdetectors(CEA).d,OrthographicviewofQueen’schamberwithnuclearemulsionfilms(NagoyaUniversity,redpositionsNE1and
NE2)andscintillatorhodoscopes(KEK,greenpositionsH1andH2). e,Orthographicviewofthemainknowninternalstructuresf,Nuclearemulsionplatesin
positionNE1(NagoyaUniversity).g,NuclearemulsionplatesinpositionNE2(NagoyaUniversity).h,ScintillatorhodoscopessetupforpositionH1(KEK).i,Gas
detectors(muontelescopes,CEA).
We first installed nuclear emulsion films (Fig. 1f-g), developed by then performed a triangulation using the data from the two positions
NagoyaUniversity,becausetheyarecompactanddonotrequireelec- and four points along the new void (Methods). The results show that
tricpower,whichmakesthemwellsuitedforinstallationinthepyramid thenewvoidhasanestimatedlengthofmorethan30mandislocated
(ExtendedDataTable1). Anuclearemulsionfilmisaspecialphoto- between40mand50mawayfromthedetectorpositions,21mabove
graphicfilmthatcandetectmuontrajectoriesinthree-dimensionalim- thegroundlevel.
ageswithsub-micrometricaccuracy7,8,9.Intheseexperiments,weused Theseconddetectiontechnology,designedbyKEK,iscomposedof
unitfilmsof30×25cm2andcoveredamaximum8m2atatime. Each four layers of scintillator hodoscopes arranged in two units of double
filmhasa70micronsthickemulsioncoatingonbothsidesofa175mi- orthogonallayers10,11 (Fig. 1h). Eachlayeriscomposedof120plastic
cronsthicktransparentplasticbase(Methods).Themuonmeasurement scintillatorbarsmeasuring1×1cm2incrosssectiontocoveranareaof
accuracyisaround1microninpositionandaround1.8mrad(~0.1de- 120 × 120 cm2. Two units are vertically separated to allow the mea-
gree)forverticaltrackbyusingtheinformationobtainedwithonefilm surementofthetwo-dimensionalincomingdirectionofthemuons. We
(ExtendedDataTable1). placedthedetectorsneartheSouth-Eastcorner(positionH1,Fig1d)of
The films were installed near the south-west corner of the Queen’s theQueen’schamberinAugust2016.Theseparationbetweentwounits
chamber (position NE2, Fig. 1d) and in the adjacent narrow hand- wassetto1.5m. Unfortunately,thenewlydetectedvoidwasoverlap-
excavated corridor called “Niche” (position NE1, Fig 1d) on the east pingwiththeGrandGallery,whichmadeitdifficulttoidentify.Aftera
sideoftheQueen’schamber.Thedistancebetweenthecentresofthese stableoperationforfivemonths,wemovedthedetectortoanotherposi-
twodetectorsisabout10monaverage,whichallowsustoperforma tionnearthesouth-westcornerofthechamberinJanuary2017(position
stereoanalysisofthedetectedstructures. TheexposurestartedinDe- H2,2.9mfrompositionH1,Fig.1d)andreducedtheunitseparationto
cember2015. Duringtheexposure,wemodifiedthefilmconfiguration 1.0mtoenlargetheangularcoverageofthemeasurement.Thedetector
several times when we changed the films (every two months on aver- operationhasbeenverystableformorethanoneyearandisstillbeing
age);onlyasubsetofthefulldatasetwasusedforthisanalysistomit- continued(Methods).
igatetheeffectoftheparallaxontheresolution(ExtendedDataTable Whenwenormalizetheresultsbythesimulationofthesolidpyra-
1andMethods). Aftereachexposure,thenuclearemulsionfilmswere midwithouttheknownstructures,weseethemclearly(Fig. 3a,b). By
processedbyphotographicdevelopmentinthedarkroomoftheGrand normalizing with a simulation that includes the known structures, we
Egyptian Museum Conservation Center. After the development, they observe a muon excess that is consistent with Nagoya’s result (Meth-
were transported to Nagoya University and read-out by an automated ods,Fig.3c-f).
nuclearemulsionscanningsystem25,26,27. Thethirdkindofinstrument,designedbytheCEA,ismadeofmicro-
We compared the resulting muon radiographs (Fig. 2a,b) with the pattern gaseous detectors (Micromegas) based on an argon mixture12
expected results computed using Monte Carlo simulation that include (Methods). Theyarerobustenoughtobeinstalledoutsideandcanrun
knownstructuresinsidetheobservationregion(Fig.2c,d). Thesecom- for unlimited time, but they have a larger footprint than the emulsion
parisonsclearlyshowthatthelargeknownstructures(theGrandGallery plates(Fig. 1i,ExtendedDataTable1). Each“telescope”isbuiltfrom
andtheKing’schamber)areobservedbythemeasurementsandthere- fouridenticaldetectorswithanactiveareaof50×50cm2,andasignal
sultsmatchtheexpectation(Fig. 2a-d). However, fromthetwoposi- onatleastthreeofthemisrequiredtotriggertheacquisition.Anonline
tions,wealsodetectedanunexpectedandsignificantexcessofmuons analysis is performed to extract the muon track parameters which are
in a region almost parallel to the image of the Grand Gallery. Statis- thentransferredatCEA,inFrance,witha3Gconnection(Methods).
tical significance of the excess is higher than 10 sigma at the highest InordertoconfirmNagoya’sdiscoveryandtoprovideanadditional
difference direction (Fig. 2c). The muon excess is similar to the one pointofview,weplacedtwosuchtelescopesinfrontoftheNorthface
generated by the Grand Gallery, which means that the volume of the ofthepyramid(Fig.1a-c),lookinginthedirectionoftheGrandGallery
twovoidsisofthesameorder(Fig.2e,fandExtendedDataFig.2).We (Methods),andcloseenoughtoeachothersothattheirdatacanbecom-
2
1) a c e
on New void B B
siti
po New void
NE 1 ulsions, B
m
ar e A A
e
cl
u
n
(
2) b d f
n B New void B
o
siti
o
2 ns, p B New void
E sio
Nul
m A A
e
ar
e
cl
u
n
(
nal analysis gQueen’s chamber [m] B h Queen’s chamber [m] A i West [m] C
3-dimensio Height from the floor of C,SDoAuth North [m] Height from the floor of EDast B CWest [m] East SoAuth D B North [m]
Figure2|Resultsoftheanalysisofthenuclearemulsionfilms.(A:King’schamber,B:GrandGallery,C:Queen’schamber(positionNE2),D:Niche(position
NE1),NewVoid:theunexpectedmuonexcessregion;thesearecommoninallplots)a,b,Two-dimensionalhistogramsofthedetectedmuonflux(/cm2/day/sr)at
positionsNE1andNE2. Theresolutionofthishistogramis0.025×0.025(Methods). Inthisfigure,rightiswest,topisnorth. Thefouredgesofthepyramidare
clearlyseenasacrosspattern.c,d,ResultsofsimulationwiththeknowninnerstructuresusingtheGeant4software28frompositionsNE1andNE2.e,f,Histogram
oftypicalangularregionasshownbythewhitesquare(0.4≤tanθy<0.7). Theredlineshowsthedata;theblacksolidlineshowsthesimulationwiththeinner
structures;thegraydashedlineshowsthesimulationwithouttheinnerstructures. Errorbarsindicate1sigmaofstatisticalerror(standarddeviation). Moreslices
areavailableontheExtendedDataFig.2.g-i,Resultsofthetriangulationanalysis(threesectionalviews).Eachfigureshowstheinnerstructures(grayline)andthe
results.Foreachposition,wedividedtheregionofinterest(0≤tanθy<1)intofourslicesandextractedthecenterofthemuonexcessforeachofthem,resulting
in4pairsofdirection(Methods).Eachofthefourpointsrepresentstheresultofthetriangulationforapairofslicesandtheassociatedstatisticalerror(Methods).
ThedetectorpositionsareshownasablackboxforpositionNE1andawhiteboxforpositionNE2;gshowsaverticalsection(rightisnorth);hshowsavertical
section(rightiswest);ishowsahorizontalsection(upiswest,rightisnorth).
bined(Fig. 1c). Aprevious3-monthmeasurementcampaignwithone 1. Lehner,M.,Thecompletepyramids: SolvingtheAncientMysteries
telescopeonthenorthsidealreadyrevealedananomalyinthisregion, (ThamesandHudson,2008).
above 3 sigma, but the telescope was shifted from the main vertical 2. Hawass, Z., Pyramid Construction: New Evidence Discovered in
axis,andsonotoptimallypositioned. Aftertwomonthsofacquisition Giza.Stationen:BeiträgezurKulturgeschichte,1998.
from the new position, the data analysis reveals two statistically sig- 3. Smith,C.B.,Hawass,Z.&Lehner,M.,Howthegreatpyramidwas
built(HarperCollins,2006).
nificant excesses of muons in the core of the pyramid (Fig. 4): one
correspondingtotheGrandGalleryandonecorrespondingtotheun- 4. ParticleDataGroup, Reviewofparticlephysics. ChinesePhysics
C40(10),100001(2016).
expectedvoid(Methods). The3Dmodelconfirmsthatthetelescopes’
5. Alvarez,L.W.etal.,Searchforhiddenchambersinthepyramids.
observationconvergestothesameregionastheoneobtainedfromthe
Science167(3919),832-839(1970).
Queen’schamberwithemulsionplates(Fig.4h).Theoverallcombined
6. Tanaka,H.K.M.,Nakano,T.,Takahashi,S.,Yoshida,J.&Niwa,K.,
excessesyield8.4(resp.5.8)sigmafortheGrandGallery(resp.thenew
Developmentofanemulsionimagingsystemforcosmic-raymuon
void)(Methods). Toourknowledge,thisisthefirsttimeaninstrument radiography to explore the internal structure of a volcano, , Mt.
hasdetectedadeepvoidfromoutsideapyramid. Asama.Nucl.Instrum.MethodsPhys.Res.A575,489-497(2007).
Three techniques of cosmic-ray muon imaging were applied to in- 7. Morishima, K., Nishio, A., Moto, M., Nakano, T. & Nakamura, M.,
Development of nuclear emulsion for muography. Annals of Geo-
vestigatetheinnerstructureofthepyramid. Theknownvoids(King’s
physics60(1),0112(2017).
chamberandGrandGallery)areobservedaswellasanunexpectedbig
8. Nakamura,T.A.etal.,TheOPERAfilm: Newnuclearemulsionfor
void,whichfullydemonstratestheabilityofcosmic-raymuonradiogra- large-scale,high-precisionexperiments.".NuclearInstrumentsand
phytoimagestructures.Thecentreofthevoidislocatedbetween40m MethodsinPhysicsResearchSectionA:Accelerators,Spectrome-
and50mfromtheflooroftheQueen’schamber.Itslengthismorethan ters,DetectorsandAssociatedEquipment 556(1),80-86(2006).
30manditscrosssectioniscomparabletothatoftheGrandGallery. 9. Nishio,A.,Morishima,K.,Kuwabara,K.&Nakamura,M.,Develop-
Therearestillmanyarchitecturalhypothesestoconsider;inparticular, ment of nuclear emulsion detector for muon radiography. Physics
Procedia80,74-77(2015).
thebigvoidcouldbemadeofoneorseveraladjacentstructures,andit
10. Fujii,H.etal.,Detectionofon-surfaceobjectswithanunderground
couldbeinclinedorhorizontal.Thedetailedstructureofthevoidshould
radiography detector system using cosmic-ray muons. PTEP (5),
befurtherstudied. Overall,thisdiscoveryshowshowthemethodsde-
053C01(2017).
velopedinparticlephysicscanshedlightononeofthemostimportant
11. Fujii,H.etal.,Performanceofaremotelylocatedmuonradiography
heritagebuildings,anditcallsformoreinterdisciplinarycollaborations system to identify the inner structure of a nuclear plant. PTEP 7,
tohelpunderstandingthepyramidanditsconstructionprocess. 073C01(2013).
3
Figure3|Resultsoftheanalysisofthescintillationhodoscopes. a-b,Two-dimensionalhistogramsofthedetectedangleofmuonsafternormalizationbythe
simulationwithoutinnerstructuresatpositionsH1andH2(Methods). Theimagespresentatopviewandthenorthistheupward. ∆xand∆ycorrespondto
thechannelnumberdifferencebetweentheupperandlowerlayersalongxandycoordinates(Methods). Thebinnumberhencerangesfrom-120to120and
providesthetangentoftheincidentanglewhendividedby150(resp.100)forpositionH1(resp.H2).Thedetectorintroducesadarkcross-shapedartefactvisible
onthetwo-dimensionalhistograms,whichaddsasmallsystematicerrorof3%totheanalysis(Methods). c-d,Two-dimensionalhistogramsofdetectedangleof
muonsafteranormalizationbythesimulationwithinnerstructures(whicharethereforeremoved). Theimagespresentthetopviewandthenorthistheupward.
Theunexpectedstructureisvisible. e-f,Histogramsoftypicalangularregion(-25<∆x<75,25<∆y<47)correspondingtoasliceofpanelscandd(yellow
rectangle). MoreslicesareavailableinExtendedDataFig. 4. Blackpointswitherrorbarsshowtherelativeexcessafterdividingthedatabythemodelwiththe
knownstructures(theKing’schamberandtheGrandGallery): aperfectagreementbetweenthedataandthemodelwouldcorrespondtoahorizontalline;apeak
correspondstoanunexpectedexcessofmuonsinthedata.
12. Bouteille,S.etal.,AMicromegas-basedtelescopeformuontomog- 24. Menchaca-Rocha, A., Searching for cavities in the Teotihuacan
raphy: The WatTo experiment (and references therein). Nuclear Pyramid of the Sun using cosmic muons experiments and instru-
Instruments and Methods in Physics Research Section A: Accel- mentation.InternationalCosmicRayConference4,325(2011).
erators,Spectrometers,DetectorsandAssociatedEquipment 834,
25. Yoshimoto,M.,Nakano,T.,Komatani,R.&Kawahara,H.,Nuclear
223-228(2016).
emulsion readout system HTS aiming at scanning an area of one
13. Dash, G., The Great Pyramid’s Footprint: Results from our 2015 thousandsquaremeters.e-Print:arXiv,1704.06814(2017).
Survey.Aeragram16(2),8-14(2015).
26. Morishima,K.,Hamada,K.,Komatani,R.,Nakano,T.&Kodama,K.,
14. Hawass,Z.etal.,Firstreport:Videosurveyofthesouthernshaftof Developmentofanautomatednuclearemulsionanalyzingsystem.
thequeen’schamberinthegreatpyramid. AnnalesduServicedes RadiationMeasurements50,237-240(2013).
Antiquitésdel’Egypte84,203-216(2010).
27. Hamada, K. et al., Comprehensive track reconstruction tool
15. Richardson,R.etal.,The"Djedi"RobotExplorationoftheSouth- "NETSCAN 2.0" for the analysis of the OPERA Emulsion Cloud
ern Shaft of the Queen’s Chamber in the Great Pyramid of Giza, Chamber.JournalofInstrumentation7(7),P07001(2012).
Egypt.JournalofFieldsRobotics30(3),323-348(2013).
28. Agostinelli,S.etal.,GEANT4—asimulationtoolkit. Nuclearinstru-
16. Tallet, P., LespapyrusdelaMerRouge1: Le"journaldeMerer" ments and methods in physics research section A: Accelerators,
(InstitutFrançaisd’ArchéologieOrientale,2017). Spectrometers,DetectorsandAssociatedEquipment 506(3),250-
17. Bui,H.D.,ImagingtheCheopspyramid (SpringerScience&Busi- 303(2003).
nessMedia,2011).
18. Butler, D. K., Microgravimetric and gravity gradient techniques for
AuthorcontributionsK.M., M.K., A.N., N.K., Y.M.andM.M.per-
detection of subsurface cavities. Geophysics 49 (7), 1084-1096
formedtheexperimentsandanalysedtheresultsforthenuclearemulsion
(1984).
films;F.T.,H.F.,K.S.,H.K.,K.H.andS.O.performedtheexperiments
19. Yoshimura,S.,Nakagawa,T.,Tonouchi,S.&Seki,K.,Nondestruc- andanalysedtheresultsforthescintillatorhodoscopes. S.P.,D.A.,S.
tivePyramidsinvestigation,Part1and2. StudiesinEgyptianCul- B.,D.C.,C.F.,P.M.,I.M.andM.R.performedtheexperimentandanal-
ture6(1987). ysedtheresultsfortelescopesgasdetectors.B.M.,P.G.,E.G.,N.S.,Y.
D.andM.S.createdthe3Dmodelsusedformuographysimulationsand
20. Morishima,K.etal.,Firstdemonstrationofcosmicraymuonradiog-
RTMS.B.M.designedandimplementedtheRealTimeMuographySim-
raphyofreactorcoreswithnuclearemulsionbasedonanautomated
ulator(RTMS)andcontributedtotheanalyses;Y.E.,T.E.,M.E.andV.S.
high-speed scanning technology. Proc. of the 26th Workshop on
coordinatedthedifferentexperimentsoperationsonthefield(muography,
RadiationDetectorsandTheirUses(2012).
3Dscans);thepaperwasmainlywrittenbyK.M.,S.P.,F.T.,M.T.,B.M.
21. Borozdin, K. N. et al., Surveillance: Radiographic imaging with andJ.-B.M.,withcontributionsofalltheotherauthors;H.H.,M.T.,B.C.,
cosmic-raymuons.Nature422(6929),277(2003). B.M.andY.E.designedandcoordinatedtheproject(ScanPyramids);
22. Menichelli,M.etal.,Ascintillatingfibrestrackerdetectorforarchae- Authorinformation Correspondenceandrequestsformaterialsshould
ologicalapplications. NuclearInstrumentsandMethodsinPhysics be addressed both to K. M. (morishima@flab.phys.nagoya-u.ac.jp) and
Research Section A: Accelerators, Spectrometers, Detectors and M.T.([email protected]).
AssociatedEquipment572(1),262–265(2007).
Acknowledgements This experiment is part of the ScanPyramids
23. Saracino,G.etal.,Imagingofundergroundcavitieswithcosmic-ray project, which is supported by NHK, La Fondation Dassault Systèmes,
muonsfromobservationsatMt. Echia(Naples). ScientificReports Suez, IceWatch, le Groupe Dassault, Batscop, Itekube, Parrot, ILP,
7(1),1181(2017). Kurtzdev, Gen-G, Schneider Electric. The measurement with nuclear
4
a G1 (Alhazen) 1600 b d G2 (Brahic) 2200 e
1400 2000
1800
1200 1600
1000 141.3 ± 28.3 1400 167.1 ± 44.6
800 11020000
600 800
400 600
400
200 200
00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
tan(θ) tan(θ)
1000 c 1200 f
800 1000
800
600 141.8 ± 24.7 163.0 ± 26.5
600
400
400
200 200
g 00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 h 00 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
tan(θ) tan(θ)
Triangulation positions
from Nagoya University
G1 (Alhazen)
G2 (Brahic)
Figure4|Resultsoftheanalysisofthegasdetectors. a,2DimagefrompositionG1(Alhazentelescope,Fig. 1a,b)inlogarithmicscale,withwhiterectangles
indicatingthe2muonexcessesseenbythetelescope.b,Horizontalsliceshowingthemuonexcessthatcorrespondstotheupperdottedblackrectangleinpanela.
Thesolidblacklineisafitofthedata(Methods). c,Horizontalsliceshowingthemuonexcessthatcorrespondstothelowerdottedblackrectanglein(a). d,2D
imagefrompositionG2(Brahictelescope,Fig1a,b)inlogarithmicscale,withwhiterectanglesindicatingthe2muonexcessesseenbythetelescope.Somenoiseis
visibleatthecentreoftheimage,butitdoesnotinterferewiththescannedregions.e,Horizontalsliceshowingthemuonexcessthatcorrespondstotheupperblack
rectanglein(d).f,Horizontalsliceshowingthemuonexcessthatcorrespondstothelowerblackrectanglein(d).g,Yellowconeandredcone,tracedfromposition
1,indicatingtheangularareafortheobservedmuonexcesses(a-c).ThetwoconesintersecttheverticalplaneoftheGrandGallery(RTMSsoftware).h,Sideview
ofthesameconesasing,forthetwopositions,confirmingthatthe2excessescorrespondtotheGrandGalleryandthenewvoid.Theblackcrossescorrespondto
thepositionestimatedwiththenuclearemulsionplates(Fig.2g-i):thetwoanalysesareperfectlycompatible.Gridsizeis10m(RTMSsoftware).Moreslicesare
availableontheExtendedDataFig.6.
emulsionswassupportedbytheJST-SENTANProgramfromJapanSci- atNagoyaUniversity. Inthisdesign,silverbromidecrystalswithadiam-
enceandTechnologyAgency,JSTandJSPSKAKENHIGrantNumbers eterof200nmaredispersedina70µmthicksensitiveemulsionlayer,
JP15H04241. The CEA telescopes were partly funded by the Région whichiscoatedonbothsidesofa175µmthicktransparentpolystyrene
Ile-de-France and the P2IO LabEx (ANR-10-LABX-0038) in the frame- plastic base7,29 (Extended Data Fig. 1a,b). When a charged particle
work “Investissements d’Avenir” (ANR-11-IDEX-0003-01) managed by passesthroughthisemulsionlayer,itsthree-dimensionaltrajectory(track)
the Agence Nationale de la Recherche (ANR, France). The detectors isrecordedandcanberevealedthroughthechemicaldevelopmentpro-
werebuiltbytheELVIAcompanyandtheCERNMPGDworkshop. The cess (Extended Data Fig. 1c,e). Thanks to the precise grain size and
authorswouldliketothankthemembersandbenefactorsoftheScan- structure, tracks can be reconstructed with sub-micron accuracy in 4π
Pyramidsproject,andinparticular:T.Hisaizumi,themembersofCairo steradiansbyusinganopticalmicroscope(ExtendedDataFig. 1d). The
University, the members of F-lab Nagoya University, Y. Doki, Aïn El trackreconstructionqualitydependsonthegraindensityperlengthalong
ShamsUniversity3Dscanningteam,themembersofEgyptianMinistryof the line of a track. In this experiment, films with a grain density of 37
Antiquities,K.ElEnany,M.ElDamaty,T.Tawfik,S.Mourad,S.Tageldin, grainsper100µmandanoiselevelofabout1grainper1000µm3were
E.Badawy,M.Moussa,T.Yabuki,D.Takama,T.Shibasaki,K.Tsutsum- used. These films can be used for long term (2-3 months) measure-
ida, K. Mikami, J. Nakao, H. Kurihara, S. Wada, H. Anwar, T. de Ter- mentinanenvironmentat25°C(temperatureintheQueen’schamber)
sant,P.Forestier,L.Barthès,M.-P.Aulas,P.Daloz,S.Moignet,V.Raoult- bytuningofvolumeoccupancyofsilverbromidecrystals(35%)30. Each
Desprez,S.Sellam,P.Johnson,J.-M.Boursier,,T.Alexandre,V.Ferret, 30×25cm2filmwasvacuum-packedinanaluminiumlaminatedpackage
T. Collet, H. Andorre, C. Oger-Chevalier, V. Picou, B. Duplat, K. Guil- forlightshieldingandhumiditycontrol(30%RH)(ExtendedDataFig.1f).
bert,J.Ulrich,D.Ulrich,C.Thouvenin,L.Jamet,A.Kiner,M.-H.Habert, An acrylic plate with 2 mm thickness was also packed together for the
B.Habert, L.Gaudé, F.Schuiten, F.Barati, P.Bourseiller, R.Theet, J.- controlofnoiseincrease9. Thepackedfilmwasfixedontoanaluminum
P. Lutgen, R. Chok, N. Duteil, F. Tran, J.-P. Houdin, L. Kaltenbach, M. honeycombplate(ExtendedDataFig. 1g,h)atNagoyaUniversity,and
Léveillé-Nizerolle, R. Breitner, R. Fontaine, H. Pomeranc, F. Ruffier, G. thentransportedtoCairobyairplane. Toavoidtheelevatedcosmic-ray
Bourge, R.Pantanacce, M.Jany, L.Walker, L.Chapus, E.Galal, H.A. flux during the flight, we transported the two emulsions layers of each
Mohalhal,S.M.Elhindawi,J.Lefaucheux,J.-M.Conan,E.M.Elwilly,A. detectorseparatelyandassembledtheminEgypt. Sincewerejectthe
Y.Saad, H.Barrada, E.Priou, S.Parrault, J.-C.Barré, X.Maldague, C. trajectoriesthatarenotcrossingthetwolayers,themuonsaccumulated
Ibarra Castenado, M. Klein, F. Khodayar, G. Amsellem, M. Sassen, C. duringtheflightarerejectedintheanalysisandareconsideredasback-
Béhar, M. Ezzeldin, E. Van Laere, D. Leglu, B. Biard, N. Godin, P. der groundtracks.
Manuelian,L.Gabriel,P.Attar,A.DeSousa,F.Morfoisse,R.Cotentin.
Dataacquisition(Nuclearemulsions-NagoyaUniversity).Theobser-
vationwithnuclearemulsionfilmswaslaunchedintheQueen’schamber
in December 2015 and the films were periodically replaced. Each film
wasalignedwithareferencelinedrawnwithalasermarkerandaspirit
level,whichledtoanangularerroroflessthan10mrad. Aftereachex-
Methods
posure,weprocessedthefilmsbyphotographicdevelopmentinthedark
roomattheGrandEgyptianMuseumConservationCenter. Forthede-
Detector Design (nuclear emulsions - Nagoya University). Nuclear velopingsolution,weusedtheXAAdeveloper(FUJIFILMCo.Ltd.)for25
emulsion is a special photographic film that is able to detect minimum minutesat20°C.Afterthedevelopment,wecarriedthefilmstoNagoya
ionizing particles such as cosmic-ray muons (Extended Data Fig. 1). University,wheretheywereread-outbyanautomatednuclearemulsion
The films used in this experiment have been developed and produced scanningsystemdevelopedsinceearly1980’sinNagoyaUniversity31,32.
5
In this experiment, tracks recorded in films were scanned by a Hyper Dataacquisition(scintillatorhodoscopes-KEK).Rawdata-timeand
TrackSelector25whichcanread-outtracksataspeedof4700cm2/hwith positionofallhitchannels-arefirststoredinaPCandregularlyretrieved
angularaccuracyof1.8mradforverticaltracks,andsavedinacomputer withUSBmemorytobesenttoKEKthroughtheInternet. Intheoff-line
storagedeviceasdigitaldata(position,angle,pulseheight). Theangu- analysis,amuoneventisdefinedbythecoincidenceofthe4layers,with
laracceptanceisapproximately|tanθ|≤1.3whereθisthezenithangle atmost2neighboringhitsineachofthem. Eventsareaccumulatedin
relativetotheperpendicularoftheemulsionsurface. two-dimensionalbins(∆x,∆y)givenbythechannelnumberdifferences
betweentheupperandlowerlayersalongxandycoordinates. Thebin
Dataanalysisandstatistics(nuclearemulsions-NagoyaUniversity).
numberhencerangesfrom-120to120andprovidesthetangentofthe
Themuontracksarereconstructedbythecoincidencebetweenthetwo
incidentanglewhendividedby150(resp.100)forpositionH1(resp.H2).
stackedfilmswithincriteriaofsignalselectionandthencountedasde-
WeinstalledthedetectoratpositionH1inAugust2016, andcontinued
tectedmuons26. Inthisanalysis,asubsetofthefulldatasetwasused
thedatatakingforfivemonthsuntilJanuaryin2017. Duringthisperiod,
toavoiddecreasingtheresolutionbecauseofimagingparallax: 4.4mil-
we accumulated 4.8 M events. We then moved the detector by 2.9 m
liontrackswereaccumulatedfor98daysatpositionNE1and6.2million
west in order to better investigate the newly observed void. The data
tracks for 140 days at position NE2, with an effective area of 0.45 m2.
taking is still continuing for more than eight months and 12.9M events
Subsequently,detectedmuonswereintegratedintotwo-dimensionalan-
wereaccumulatedatpositionH2attheendofSeptember2017,withan
gularspace(tanθx-tanθy)withthebinsizeofaspecifiedsize(e.g.,0.025 overallsmoothacquisitionformorethanayear.
× 0.025) and the angular acceptance of |tanθ|≤ 1.0, and converted to
muonflux(/cm2/day/sr)ineachbin(Fig.2). Dataanalysisandstatistics(scintillatorhodoscopes-KEK).Thefirst
WeusedtheMonteCarlosimulatorGeant4Version10.228tocompute step of the analysis is the normalization of the data by a Monte Carlo
the expected muon flux at detector position. In these simulations, the simulation that takes into account the cosmic ray muon flux and muon
physical process of electromagnetic interactions and decays of muons interactions35,36(energylossandmultiplescattering)inthePyramid.We
were included, Miyake’s formula of integrated intensity of cosmic-ray assumeaconstantenergylossof1.7MeVperg/cm2(ref.4),ameanden-
muons was utilized as a flux model, and only muons were generated sity of 2.2 g/cm3, and a radiation length of 26.5 g/cm2 for the stones.
asprimaryparticlesinthissimulator33. Inordertoreducetheprocess- Muonsarepropagatedinstepsof0.1m. Becausetheknownstructures
ingtime, onlymuonswerepropagatedandtherangeofincidentmuon ofthePyramidaresimulated,theireffectsareremovedafterthenormal-
energy was limited to be 20 to 1000 GeV in the zenith angular range izationofthedataandtheremainingmuonexcessshowstheexistence
0 to 70 degree (-2.75 < tanθ < 2.75). For the pyramid simulation, we ofanunknowncorridor-likenewstructure. Thesuccessfuleliminationof
modeledtheshapeandthelocationoftheknownstructures(theGrand theknownstructuressuggeststhereliabilityofoursimulation. Slicesof
Gallery, the King’s chamber, the corridor that connects them, and the theimagesalong∆xarepresentedinExtendedDataFig.4a,b.Thever-
Queen’schamber)usingthesurveyofDormion34. Wedefinedthatany ticalscaleistherelativeyieldtothesimulationresult. Thenewstructure
emptyvoidwouldbefilledwithairandthatthestonesarelimestone(2.2 isclearlyseenineachslice. Thesignificanceofthemuonexcesswas
g/cm3)exceptaroundtheKing’schamber,wheretheyaregranite(2.75 obtainedbyaGaussianfit:atpositionH1theexcessheightsintheslices
g/cm3). Theexposureperiodinthesimulationiscompatiblewith1000 rangesfrom5.2%to8.9%andmorethan10sigmaexceptforthemost
days,whichisapproximately10timeslongerthanthatofanalyzeddata. outerslice,inwhichtheheightisstillmorethan5sigma. AtpositionH2
Weestimatedtherockthicknessdistributionin2Dfromthedetectorpo- theheightrangesfrom8.9%to11%andareagainabove10sigmaex-
sition:theminimumthicknessis65mandthemaximumthicknessis115 ceptforthemostouterslice,inwhichtheexcessisabove7sigma.From
m. Ifweassumethefluctuationofsurfacestructureis1mscale(stone theseslices,wefoundthatthestructurestartsalmostatthecentreofthe
size),theeffectoftherelativefluctuationislessthan2%. Pyramidandendsatananglewhosetangentis0.8totheNorth. Asa
Normalization was performed to compare real and simulated data in result,thelengthofthemainpartofthenewstructureisapproximately
the region without the analyzing target (the Grand Gallery, the King’s 30m. ResultsfrombothpositionH1andH2showthatthenewvoidis
chamber,andtheanomalyregion). Theexcessregionofmuonfluxwas abovetheGrandGallery,whichisconsistentwithNagoya’sresult.
clearlyapparentintheimages(Figure2a-d).Twohistograms(Figure2e,f) Detector design (gas detectors - CEA). A telescope (Extended Data
showmuonfluxextractingfromtheslicein0.4≤tanθy <0.7. Fromthe Fig.5a)iscomposedof4Micromegas(ExtendedDataFig.5b),aMicro-
comparisonbetweendataandthesimulation,thesignificancesofeach PatternGaseousDetector(MPGD)inventedatCEA-Saclay37(Extended
anomalyregionwereevaluatedat13.7sigma(statistical)forpositionNE1 DataFig. 5c). Allthedetectorsareidentical,withanactiveareaof50×
and12.7sigmaforpositionNE2. 50cm²(ref.12). Theyarebuiltfromthebulktechnology38 withascreen-
Inordertolocatethenewlydiscoveredstructure,weperformedatri- printingresistivefilmontopofthereadoutstripstoallowforstableop-
angulationfromthetwopositions. ThecentreofdetectorpositionNE1 erationandhighergain39. EachdetectorprovidesXandYcoordinates
waslocatedat5.8meastfromtheaxisoftheGrandGalleryandat4.5 througha2Dreadoutinsertedontotheprintedcircuitboard(Extended
mwestforpositionNE2. ThedistancebetweenpositionNE1andNE2 DataFig. 5d). The1037readoutstrips(482micronpitch)ofeachco-
is1.1minnorth-south(Figure1d). Inordertodeterminethedirection ordinatearemultiplexedaccordingtoapatentedscheme40. AnArgon-
towardtheanomalyregion,weperformedthefittingoftheexcessregion iC4H10-CF4,non-flammablegasmixture(95-2-3)isflushedinseriesin
toaGaussianfunctionbydividingtheregion(0≤tanθy<1)intofourre- allthedetectorsofatelescope,withaflowbelow0.5L/h,thankstoatight
gionswithasegmentof0.25intanθy,becausethenewstructureseems seal(measuredgasleakbelow5mL/hperdetector).
toalignalongthetanθyaxisdirection(ExtendedDataFig. 2). Thefitted Each telescope is operated with a Hummingboard nano-PC running
center value was used for triangulation and the errors of the estimated GNU/Linux,whichcontrolsalltheelectronics41:adedicatedhighvoltage
positionsweredefinedfromtheerrorsonthesightlinescomingfroma power supply (HVPS) card with 5 miniaturized CAEN modules, which
halfofthebinwidth,i.e.0.0125intanθxand0.125intanθy(Fig.2g-i). provideupto2kVwith12Vinputs,andtheFrontEndUnit(FEU)readout
electronicsbasedontheDREAMASIC42.Aparticularlyimportantfeature
Detector design (scintillator hodoscopes - KEK). The detector con-
of DREAM is its self-triggering option to generate the trigger from the
sistsoftwounitsofdoublelayers, i.e., xandylayers, ofplasticscintil-
detectorsthemselves. Adedicatedsoftwarepackagewasdevelopedto
lator arrays (Extended Data Fig. 3a-c). A single scintillator element is
10 ×10 mm2 in cross section and 1200 mm long. Each layer has 120 interface all these electronic components to the Hummingboard which
performsthedataacquisitionwiththeFEU.Italsomonitorsandsetsthe
elementstightlypacked,andhenceitsactiveareais1200×1200mm²
highvoltagesthroughtheHVPSandapatentedamplitudefeedbackto
(ExtendedDataFig.3d).Theelementhasacentralholealongitslength,
keepthegainconstantinspiteoftheextremeenvironmentalconditions
throughwhichawave-shifteropticalfiberisinsertedtoefficientlytrans-
oftheGizaplateau(ExtendedDataFig.5e-g).Theoverallconsumption
ferthescintillationlighttoaMulti-PixelPhotonCounterSensor(MPPC,
ofatelescopeisonly35W.
Hamamatsu). ThebiasvoltageoftheMPPCwasselectedaccordingto
thetemperatureoftheQueen’sChamber, whichisconstantregardless A trigger is formed by the FEU when at least 5 coordinates out of 8
oftheweatheroutside. EachlayerhasitsownDAQbox,whichdigitizes observeasignalaboveaprogrammablethreshold.Thesampledsignals
theinformationofsensorsignalsandsendsthemtoacommonPCinside (50samplesatafrequencyof100/6MHz,ExtendedDataFig. 5h)ofall
thedetectorframe. Thetotalpowerconsumptionofthedetectorsystem theelectronicchannels(64×8)arethendirectlyconvertedtoaROOT
is300W.Theverticaldistancebetweenthetwounitsis1500mmatpo- file43, a format commonly used in particle physics. The nano-PC per-
sitionH1and1000mmatpositionH2,andgivesanangularresolution formstheonlinereconstructionofmuontrajectories,andthemuontrack
around7and10mradrespectively.Thetangentacceptancerangesfrom parametersaresenttoFrancewithsomeenvironmentaldata(tempera-
0(vertical)to0.8and1.2radrespectively. ture,pressure,humidityinthegas)througha3Gconnection.
The detector introduces a dark cross-shaped artefact visible on the Dataacquisition(gasdetectors-CEA).Thedatawerecollectedfrom
two-dimensionalhistograms(Fig. 3a-d),whichaddsasmallsystematic May4thtoJuly3rd,2017withastabledatataking. Twotelescopeswere
errorof3%totheanalysis. Accordingtoouranalyses,theerrorislikely installedinfrontofthechevrons(Northface),atadistanceof17and23
tobecausedbytheverynarrowgapbetweenneighboringscintillatorel- m,respectively(Fig. 1b,c). Theaxisofbothtelescopesdeviatedslightly
ements,butthiseffecthasnotbeenfullyunderstoodyet.Thissystematic fromthenorth-southaxis,towardtheeast,topreventtheGrandGallery
errorisnotrelevantinthepresentanalysis,whichonlyexaminestheex- frombeingatthecentreoftheimage,wheresomecorrelatednoisecan
istenceofthenewvoid. showup. Duringtheacquisitiontimethetwotelescopes(calledAlhazen
6
and Brahic, at positions G1 and G2, see Fig. 1b,c) recorded 15.0 and 29. Morishima,K.,LatestDevelopmentsinNuclearEmulsionTechnol-
14.5 million triggers, respectively, from which 10.6 and 10.4 million of ogy.PhysicsProcedia80,19-24(2015).
track candidates were identified. After the chi² quality cut, 6.9 and 6.0
30. Nishio,A.etal.,LongTermPropertyofNuclearEmulsion.Program
million good tracks were reconstructed, and form the images shown in
andProceedingsofThe1stInternationalConferenceonAdvanced
Figure4(a)and(d).
Imaging(ICAI2015),668-671(2015).
Dataanalysisandstatistics(gasdetectors-CEA).Fromtheacquired
31. Aoki, S. et al., The Fully Automated Emulsion Analysis System.
tracks,wesearchedforanomaliesbyextractingslicesintan(phi),i.e.hor-
NuclearInstrumentsandMethodsinPhysicsResearchSectionB:
izontalslices. Togetmorestatistics,thethicknessoftheslicesislarger
BeamInteractionswithMaterialsandAtoms51,466-472(1990).
thanthebinningshowninthe2Dimages. Wechoseaslicethicknessof
0.10fortheAlhazentelescope,andweincreasedto0.11fortheBrahic 32. Morishima, K.&Nakano, T., Developmentofanewautomaticnu-
telescopetokeeproughlythesamestatistics.TheExtendedDataFig.6 clearemulsionscanningsystem,S-UTS,withcontinuous3Dtomo-
illustratesalltheslicesmadewithAlhazenfrom0.21to-0.19. Fromone graphicimageread-out. JournalofInstrumentation 5(4), P04011
histogramtoanother,theslicepositionisshiftedby0.02,whichmeans (2010).
thedataofconsecutivehistogramslargelyoverlap. Thegoalistoscan
33. Miyake,S.,Rapporteurpaperonmuonsandneutrinos. 13thInter-
thepyramidanddetectanydeviationfromstatistics,beingfluctuationsor
nationalcosmicrayconference(1973).
not.
AscanbeseenonExtendedDataFig.6,theslicesshowsmoothdis- 34. Dormion, G., La chambre de Cheops: analyse architecturale (Fa-
tributions,exceptaroundhistograms5-6and15. Thehistograms6and yard,2004).
15correspondtoFig. 4band4crespectively. Thesedistributionswere 35. Reyna,D.,Asimpleparameterizationofthecosmic-raymuonmo-
fittedwithdifferentfunctions,inparticularpolynomials.Thoughsuchfunc- mentumspectraatthesurfaceasafunctionofzenithangle. arXiv
tionsareessentiallyempirical,aCRY/Geant4simulationwasperformed preprint,hep-ph/0604145(2006).
tofurthervalidatethischoice,leadingtoagoodagreementusinga2nd
orderpolynomialwithareducedχ²of1.4.Thesamefunctionreproduces 36. Jokisch, H., Carstensen, K., Dau, W., Meyer, H. & Allkofer, O.,
datadistributionsfairlywell-withareducedχ²valueof1.6and2.0re- Cosmic-raymuonspectrumupto1TeVat75zenithangle.Physical
spectivelyforhistograms6and15-exceptinaregionwhereanexcess reviewD19(5),1368(1979).
isclearlyobservedonbothslices,withsinglebinexcessof4.2and5.3 37. Giomataris,I.Y.,Rebourgeard,P.,Robert,J.P.&Charpak,G.,MI-
sigma respectively. Re-fitting with a 2nd order polynomial and a Gaus- CROMEGAS:ahigh-granularityposition-sensitivegaseousdetector
siansignificantlyreducestheχ²to1.2and1.4,withaGaussianintegral forhighparticle-fluxenvironments. NuclearInstrumentsandMeth-
of141.3±28.3(5.0sigma, histogram6)and141.8±24.7(5.7sigma, ods in Physics Research Section A: Accelerators, Spectrometers,
histogram 15). Similarly, Brahic data show 2 significant excesses, cor- DetectorsandAssociatedEquipment376(1),29-35(1996).
responding to Fig. 4e and 4f. A 2nd order polynomial alone results in
reduced χ² value of 1.8 and 2.4 respectively, while adding a Gaussian 38. Giomataris,I.Y.etal.,Micromegasinabulk. NuclearInstruments
and Methods in Physics Research Section A: Accelerators, Spec-
reducesthemto1.5and1.6. TheGaussianintegralis167.1±44.6(3.7
trometers, Detectors and Associated Equipment 560 (2), 405-408
sigma,Fig.4e)and163±26.5(6.1sigma,Fig.4f).
(2006).
The3Dmodelconfirmsthatthe(wellcompatible)excessesfromFig.
4c(Alhazen)and4f(Brahic)pointtothesameregionofthepyramidand 39. Alexopoulos, T. et al., A spark-resistant bulk-micromegas cham-
overlapverywellwiththeGrandGallery(Fig. 4h). Thequasi-fulloverlap berforhigh-rateapplications. NuclearInstrumentsandMethodsin
ofthecones(duetothepurposefulproximityofthetelescopes)justifies PhysicsResearchSectionA:Accelerators, Spectrometers, Detec-
adding the 2 excesses, leading to 304.8 ± 36.2 (i.e., 8.4 sigma). This torsandAssociatedEquipment640,110-118(2011).
fullyvalidatestheabilityofthetelescopestounambiguouslydetectlarge
40. Procureur,S.,Dupré,R.&Aune,S.,Geneticmultiplexingandfirst
structuresinthecoreofthepyramid.
results with a 50x50cm2 Micromegas. Nuclear Instruments and
The 3D model also confirms that the excesses from Fig. 4b and 4e
Methods in Physics Research Section A: Accelerators, Spectrom-
pointtothesameregion. Likebefore,thequasi-fulloverlapofthecones
eters,DetectorsandAssociatedEquipment729,888-894(2013).
justifiesaddingthe2excesses, leadingto308.4±52.8(5.8sigma). A
3DcomparisonwiththetriangulationmadebyNagoyaUniversityfurther 41. Bouteille,S.etal.,AMicromegas-basedtelescopeformuontomog-
confirmsalargeoverlapoftheseregions.Theotherslicesshownoother raphy:TheWatToexperiment.NuclearInstrumentsandMethodsin
anomalyexceeding5sigma. PhysicsResearchSectionA:Accelerators, Spectrometers, Detec-
Itisworthmentioningthatthisanalysisreliesdirectlyonrawdatawith- torsandAssociatedEquipment834,223-228(2016).
outmodelsubtraction, whichmeansthesystematicsaremuchsmaller,
42. Flouzat, C. et al., Dream: a 64-channel Front-end Chip with Ana-
andcanonlyoriginatefromthefittingfunction.Asanexercise,a3rdpoly-
logue Trigger Latency Buffer for the Micromégas Tracker of the
nomialfitwasusedforthehistogramsshowingthenewvoid,resultingin
CLAS12Experiment.Proc.ofTWEPPconference(2014).
excessesof5.2and3.6sigmaforAlhazenandBrahic,andacombined
excessof303.5±52.1(5.8sigma). 43. Brun,R.&Rademakers,F.,ROOT—anobjectorienteddataanalysis
framework. NuclearInstrumentsandMethodsinPhysicsResearch
3-dimensional model of the pyramid. We designed an accurate 3-
SectionA:Accelerators,Spectrometers,DetectorsandAssociated
dimensional model of the pyramid by combining existing architectural
Equipment389(1-2),81-86(1997).
drawings34,44photogrammetry45, and laser scanner measurements45,
both inside and outside of the pyramid. After merging these data, the 44. Maragioglio, V. & Rinaldi, C., L’architettura delle piramidi Menfite.
modelcontainsabout500,000triangles(ExtendedDataFig.7b-d).This ParteIV,LaGrandepiramidediCheope(Rapallo,1965).
modelwasmainlyusedintheRTMSsimulator(seebelow)andasref-
45. Kraus,K.,Photogrammetry:geometryfromimagesandlaserscans
erence for the simplified models used in the other simulators. The full
(WalterdeGruyter,2007).
modelhasaprecisionofapproximately30cmfortheinternalstructures
andapproximately1mfortheexternalcasing.
RealTimeMuographySimulator.TheRealTimeMuographySimulator Extended Data Table 1 | Comparison of the three muon detection
(RTMS,ExtendedDataFigure7a)isafast,interactivesimulatorthatwas technologies.
mainlyusedforpreliminaryanalyses,toaidpositioningthegassensors
(telescopes),andforconfirmingtheresultsobtainedfromtheotherssim- Nuclear emulsion Hodoscopes Gas detectors
ulators. Itallowstheusertoplaceasensorinthedetailed3Dmodelof Nagoya University KEK CEA
thepyramidandtosimulatetheobservedmuonrateinrealtime. Muon Angular Resolution 2-14 mrad 7-10 mrad 0.8 - 4 mrad
scatteringisnotsimulated.Foreachpixelofthesensor,whichrepresents Angular Acceptance 45 degrees 34 - 45 degrees 45 degrees
adirectionrelativetothesensor,thesimulatorcomputestheopacity(in- Active area 30 cm x 25 cm / unit: 1.2 m x 1.2 m 50 cm x 50 cm
tegral of the density along the path) from the sensor to the outside of
(for this analysis) 0.75 m x 0.6 m (NE1)
thepyramid,alongthisdirection. Weusedadensityof2.2g/cm3forthe
limestone, and2.6g/cm3 forthegranite. Weconsiderthatmuonslose 0.9 m x 0.5 m (NE2)
Position Resolution 1 μm 10 mm 400 μm
energyataconstantrateof1.69MeVperg/cm2 (ref4) whichallowsus Height 0.2 mm 1-1.5 m 60 cm
tocomputetheminimalenergyEmintocrossthepyramidgiventhevalue Power requirement No Yes (300W) Yes (35W)
oftheopacity. Finally,weuseMiyake’sformula33 tocalculatethedistri- Data taking Need development Real time Real time
butionofmuonsthathavegreaterenergythanE comingatazenith
min
angleθ. Thisvalueiscomputedforeachpixeloftheimage,leadingtoa
2-dimensionalhistogramsimilartothoseobtainedwiththedetectors.
Data availability. The data that support the findings in this study are
availablefromthecorrespondingauthorsonreasonablerequest.
7
a b Silber bromide crystals c Track
(diameter 200 nm)
Emulsion layer 70 μm
Transparency 175 μm
plastic base plate
Photographic
70 μm development
Emulsion layer process
d
e
100 μm
g h
Aluminum honeycomb
plate : 17 mm
f
Rubber plate (5 mm)
Nuclear emusion (0.3 mm)
Aluminum honeycomb
plate : 17 mm
ExtendedDataFigure1 |Overviewofthenuclearemulsionfilms(NagoyaUniversity).a,Acrosssectionalschematicviewofnuclearemulsion.b,Enlarged
schematicviewoftheemulsionlayer.Silverbromidecrystalsaredispersedingelatin.Thereddashedarrowshowsthetrajectoryofthechargedparticle.c,Afterthe
photographicdevelopmentprocess,silvergrainsarealignedalongthelineswiththetrajectory(or,track)ofchargedparticle.d,Anopticalmicroscopicphotograph
ofatrackofminimumionizedparticlerecordedinanuclearemulsion.e,Anuclearemulsionafterdevelopment.f,Avacuumpackednuclearemulsion.g,Schematic
viewofthedetectorconfiguration: sixpackednuclearemulsionfilmswithadetectionareaof30×25cm2each(yellow)arefixedbetweenaluminumsupporting
plates(honeycombplate,ingrey).Twofilmsstackedontopofeachotherarepressedwitharubbersheet(black)byfourshortscrews.Threeadditionallongscrews
areusedaslegstocorrecttheinclinationofthedetector. h,Crosssectionalschematicviewofthenuclearemulsiondetectorasshowning. Twopackedfilmsare
stackedbetweentwohoneycombplatesandrubbersheet.
8
NE1 (nuclear emulsions, position 1) NE2 (nuclear emulsions, position 2)
ExtendedDataFigure2 |Slicesofthedataforthenuclearemulsionplates. Eachfigureshowslicesfortanθyevery0.25units(intangent)andseparatedinto
fourrangesat0≤tanθy<1(Fig.2a,c).Toppanelshowsmuonfluxdistributionandbottompanelshowsdifferenceofmuonfluxineachfigure.Inthetoppanels,
theredlineshowsthedata,theblacksolidlineshowsthesimulationwiththeinnerstructures,andthegreydashedlineshowsthesimulationwithoutanyinternal
structure. Inthebottompanels,theredlineshowsthesubtractionbetweenthedataandthesimulationwiththeinnerstructures,theblacklineshowssubtraction
betweenthesimulationwiththeinnerstructuresandwithoutthem,sothattheGrandGalleryappearsasamuonexcess. Errorbarsindicate1sigmaofstatistical
error(standarddeviation).ThecomparisonbetweentheexcessthatcorrespondtotheGrandGalleryandtheonethatcorrespondstothenewvoidshowsthatthetwo
structuresareofasimilarscale.Foreachprojectionofdifferenceofmuonflux,weperformedaGaussianfittingtoestimatethedirectionofanomalies.Thefitting
zonewas0≤tanθx≤0.2forpositionNE1and-0.2≤tanθx≤0forpositionNE2.Thesefittedcentreswereusedforthetriangulation.
a
b
d
c
ExtendedDataFigure3 |Overviewofthescintillatorhodoscopes(KEK).a,Verticalviewofthedetector,consistingoftwounitsoforthogonaldoublescintillator
layers.Abluearrowindicatesamuontrackpassingthroughthewholeinstrument.b,Crosssectionofascintillatorelement,showingthecentralholefortheoptical
fiber.b,Gridmadeofdoublelayers,detectingthepositionofincidentmuons.d,Planeviewofthedetectorhavinganactiveareaof1.2×1.2m².
9