Industrial X-ray Tomography as a Tool for Shape and ...

1

Industrial X-ray Tomography

as a Tool for Shape and Integrity Control of SRF Cavities

Hans-Walter Glock, Jens Knobloch, Axel Neumann, Adolfo Velez-Saiz

2021 Int. Conf. on RF Superconductivity (SRF’2021) – THOTEV07

Hans-Walter GlockTHOTEV07

Acknowledgements to:

2

– the trigger:

Michele Bertucci (INFN-LASA, Segrate, Italy) et al.:

“Test, diagnostics and computed tomographic inspection of a large grain 3.9 GHz prototype cavity”, JACoW-IPAC2017-MOPVA062

– the industry partners:

J. Kinzinger (XRAY-LAB, Sachsenheim, Germany),

M. Böhnel, N. Reims, M. Salamon (Fraunhofer Institute for Integrated Circuits IIS, Development Center X-ray Technology EZRT, Fürth, Germany):

“Operational Experiences with X-ray Tomography for SRF Cavity Shape and Surface Control”, JACoW-IPAC2019-WEPRB017

– the SRF’2021 Program Comittee …

… who’s talk invitation caused a reviewed and detailed summary of available data.


X-ray tomography (in a nutshell)

3

electron acc. ?tungsten

target X-ra

y im

age

dete

ctor

test object on rotating table

Is it a “magic eye” to look inside a cavity ?

“look” = surface inspection, weld integrity,

thickness and shape control

Tomography of VSR-1-Cell test cavity, colors indicate deviation from design


X-ray tomography: X-rays

4


target X-ra

y im

age

dete

ctor


incoherent broadband (~10 keV … 10 MeV) X-rays, core property is intensity-proportional absorption with material-dependent

absorption coefficient µ

assumed to happen along a straight line between source and detector.


X-ray tomography: absorption

5

material-dependent absorption coefficient µ


target X-ra

y im

age

dete

ctor


data from: J. H. Hubbell, S. M. Seltzer, X-Ray Mass Attenuation Coefficients, doi: 10.18434/T4D01F

Niobium attenuates much stronger than iron up to 200 keV.


X-ray tomography: source

6

we tried: – DC-gun 300 kV– DC-gun 587 kV– 9-MeV Linac (Siemens Silac)


target X-ra

y im

age

dete

ctor


Most photons have significantly less energy than the electron.

placed in front of the scintillator at 1461mm from thesource. The scintillator, the mirror, and the CCD cameraare placed inside a metallic box (detector box) of overalldimensions of 490! 510! 1000mm3 (width!height!depth) with walls made of a 3 mm thick steel layer (innerlayer), a 3 mm thick lead layer and a 5 mm thick steel layer(external layer). The source–detector distance is 1500mm.A lead housing is placed around the X-ray tube to reducethe leakage of the X-ray tube housing. The dimensions ofthe X-ray shielding room are 3.6! 2.9! 3.7m3 (width!height! depth). Two walls of the room are made ofconcrete and the other two of a sandwich of steel (3mm),lead (25mm), and steel (3mm).

2.1.1. X-ray sourceThe X-ray unit employed is a 450 kV X-ray generator

manufactured by Comet AG (Model MXR 451) with2.3mm iron and 1.0mm copper inherent filtration. Theangle of the target, made of tungsten, is 301. The sizeof the focal spot is 2.5mm. The emission cone of the X-raysource is 401. A 1.0mm tungsten (alloy: HPM1750)attenuator is used to reduce the primary flux of low-energyphotons.

2.1.2. X-ray source–collimatorThree source–collimators were manufactured to investi-

gate the influence of beam aperture on image contrast: (i) a100 mm thick rectangular lead source–collimator withangular aperture 8.871! 6.091 (horizontal! vertical), (ii) a180 mm thick brass source–collimator with angularaperture 5.611! 5.611 (big aperture), and (iii) a brasssource–collimator with angular aperture 3.151! 3.151(small aperture). The brass source–collimators can beinserted inside the lead source–collimator. Moreover theconfiguration without any source–collimator, which corre-sponds to a beam aperture of half angle 201, is considered.In each studied configuration the test object was comple-tely irradiated by the primary beam.

2.1.3. DetectorThe X-ray converter is a Thallium-doped Cesium Iodine,

CsI(Tl), scintillator with a thickness of 2mm manufa-ctured by Hamamatsu. The effective scintillator area is480! 280mm2. The 1 mm thick back plate is made ofaluminum. The converted photons are projected on a 451mirror and reflected on a CCD camera (Apogee Alta U32)of 2184! 1472 pixels. The pixel size is 6.8! 6.8 mm2.A NIKON 28mm lens is mounted on the CCD camera.The field of view is 524! 353mm2.

2.2. Description of the GEANT4 based Monte Carlo CTsimulation

The low-energy extension of the electromagnetic pro-cesses version 2.3 of the simulation toolkit GEANT4[20–22] was used to model the interactions of photons andelectrons with matter down to 250 eV. The processes

activated in the physics list for electrons were: ionization,bremsstrahlung, and multiple scattering; for photons:Rayleigh scattering, Compton scattering, and photoelectriceffect [23]. The cut value in range was set to 0.1mm forphotons and electrons.The simulation is performed in two steps: the generation

of the X-ray spectrum of the tube taking into account theanode angle, inherent and external filtration of the tube;and the generation of the projection of the object accordingto the acquisition setup (i.e. image of the energy depositedwithin the detector).The simulations were run on a Pentium-IV-based

personal computer with a 2.80GHz microprocessor. Thecomputing time to reach a good statistic strongly dependson the X-ray beam aperture; it goes from 1 day in case ofsmall objects (approx. 5 cm path length, beam aperture of61) to several days in case of large objects (approx. 20 cmpath length, beam aperture of 181).

2.2.1. Generation of the X-ray spectrumThe generation of the spectrum involves the simulation

of a monochromatic pencil electron beam hitting the tung-sten target at an angle of 301 with respect to the normal ofthe anode surface and the passage of the produced X-rayspectrum through inherent filtration (2.3mm Fe and1.0mm Cu) and external filtration (1.0mmW, alloyHPM1750). The radiation is retrieved within an angle of201 with respect to the central axis of the beam. Fig. 2shows the spectrum at 450 kV for the MXR-451 CometX-ray tube simulated using the X-ray tube characteristicsmentioned above. The spectrum was simulated with 2! 109

primary electron histories. The validation of the simulatedspectra has been assessed through comparison withexperimental data [24].

2.2.2. Image of the deposited energy distributionThe X-ray photons are emitted from the focal spot of

diameter 2.5mm, with energy sampled randomly from thesimulated spectrum, towards the object. Their direction isselected randomly from an isotropic distribution of angles

ARTICLE IN PRESS

0

0.5

1

1.5

2

2.5

3

3.5

4

0 50 100 150 200 250 300 350 400 450

Pho

tons p

er

ke

V

Energy (keV)

Fig. 2. Simulated spectrum of the X-ray tube MRX-451. The electronenergy was 450 keV, the inherent filtration was 2.3mm Fe+1.0mm Cuand the external filtration was 1mm W (alloy HPM1750).

A. Miceli et al. / Nuclear Instruments and Methods in Physics Research A 583 (2007) 313–323 315

picture freely taken from: A. Miceli et al., “Monte Carlo simulations of a high-resolution X-ray CT system for industrial applications”, NIM A, doi:10.1016/j.nima.2007.09.012, Fig. 2

example: electron energy 450 keV

bremsstrahlungch

arac

teris

tic


X-ray tomography: detector and object size

7

we used square panels~ (2000 x 2000) pixels ~ (0.3 x 0.3) m2 @ 300 kV,~ (0.4 x 0.4) m2 @ 587 kV, ~ (0.5 x 0.5) m2 @ 9 MeV


target X-ra

y im

age

dete

ctor


L1 L2

dete

ctor

wid

th N

p ∙ Δ

p

pixel size Δp

Detector panel width limits object diameter:

High resolution: L1 small, biggest practical Dobj ~ 0.9 detector width


X-ray tomography: resolution

8

Finite width of X-ray source spot size Δs and detector pixel size Δp mix to effective beam path width:


target X-ra

y im

age

dete

ctor


L1 L2

spot size Δs pixel

size Δp

587 keV-setup: L1 1786 mm, L1+L2 2500 mm

resolution limit: 0.2 mm


X-ray tomography: shielding, mechanics

9

not to forget, since expensive:– shielding for radiation safety (keep distance to object and detector to

reduce ambient scattering)– mechanical precision, thermal stability in the order of spatial resolution


target X-ra

y im

age

dete

ctor



X-ray tomography: setups

10

#$

!

%"

bERLinPro Gun1.1 @XRAY-LAB, 300 keV

#

$!

"

%

VSR-1-cell @Fraunhofer EZRT, 9 MeV

#

$

!

"

%


X-ray tomography: workflow I

11

Tomographic Inversion (proprietary algorithm)X-ray pictures of ~103

different orientations

sing

le to

som

e ve

rtic

al s

egm

ents

cubic voxel volume representation in 216 intensity-scale values (.rek-file, ~ 10 GB)

bERLinPro Gun1.1 @Fraunhofer EZRT, 9 MeV


bERLinPro Gun1.1: 1.4-λ/2-cell @ 1.3 GHz, with tunable choke cell

12

bERLinPro Gun1.1 (@Fraunhofer EZRT, 9 MeV, threshold 17464) vs. design

( 1 : 5 )

A-A

Projektionsmethodeprojection

Maßstabscale

Dokumenten Nr.document no.

Indexrev.

geprüftchecked

freigegebenapproved

Datumdate

Name

Benennungtitle

Werkstoffmaterial

Index / rev.

erstelltgenerated

Datumdate

gezeichnetdrawn Änderungstext / details of revision

Gewichtweight

Version

Status

DIN ISO 2768DIN EN ISO 13920

Allgemeintoleranz /dimension without tolerance

Tolerierung / tolerance DIN ISO 8015

Kanten / edges DIN ISO 13715

Oberfläche / surface DIN EN ISO 1302

Weitergabe sowie Vervielfältigung dieser Unterlage,Verwertung und Mitteilung ihres Inhaltes ist nicht ge-stattet. Zuwiderhandlungen verpflichten zu Schaden-ersatz. Alle Rechte für den Fall der Patenterteilungoder GM-Eintragung vorbehalten. Without expressed permission it is forbidden to copythis document, forward it to others and communicateor use the contents within. Offenders are liable tothe payment of damages. All rights are reserved inthe event of the grant of a patent or the registra-tion of a utility model or design. Artikelnummer

part number

Kundenref.-Nr. / customer ref.-id

GUN Cavity Inner w Chokecell Tuner Bell

18 kg

01.04.15 Scherer

Fertigungsfreigabe1

A

19.06.15 Fertigungsfreigabe

P96030

Scherer

A

Schmitz

Pekeler

Z194200

GUN_

Cavity

_Inn

er_w

_Cho

kece

ll_Tu

ner_

Bello

_P96

030_

1942

00.id

w/

V25

Ursprung / origin

research

instruments

Fügeprozess / Joining process DIN EN ISO 4063:2011

A

A

123456789101112

123456789101112

B

D

E

G

A

B

C

D

E

F

G

H

A

C

F

H

A1

511

1:2

-BF

auf Ausrichtung achten!

1 x 10 mbar l s Heliumleckrate: --1-10 <

Qualitätsanforderung nach: DIN EN ISO 13919 Level-B

10

(505,9)

511

20

405,5`0,4

stro

ng h

alf-c

ell d

efor

mat

ion

due

com

plic

ated

tuni

ng

good coincidence of full cell,

waist and stiffening ringscathode channel and flange

zone blurred

(cf. Y. Tamashevich, SRF21)


X-ray tomography: workflow II

13



sing

le to

som

e ve

rtic

al s

egm

ents


Generic 1-cell 1.3 GHz @Fraunhofer EZRT, 9 MeV

bulk material volume construction by intensity-scale threshold (Volumegraphics©, GOM-Inspect©)


X-ray tomography: influence of threshold choice

14

Generic 1-cell 1.3 GHz @Fraunhofer EZRT, 9 MeV


30000

20000

10000

Threshold selection directly determines allocation of

material borders.

Choice happens by an educated guess!

(Approaches smarter than single constant value in

use, but they also are based on additional

assumptions.)


X-ray tomography: workflow III

15



sing

le to

som

e ve

rtic

al s

egm

ents


Example: overlapping interior welds of the

coupler port of the VSR-1-cell



defect and shape accuracy check unpurified bulk material’s surface description (.stl-file)


Issues with STL-files

16

– extremely big (~ 100 … 101 GB)– isolated volumes, isolated/partially connected surfaces– large surfaces not “watertight”– not managed by field solvers

unpurified bulk material’s surface description (.stl-file)


X-ray tomography: workflow IV

17



sing

le to

som

e ve

rtic

al s

egm

ents




conversion to NURBS-delimited volume representation (Geomagic Design X ©)


X-ray tomography: workflow V

18



sing

le to

som

e ve

rtic

al s

egm

ents




conversion to NURBS-delimited volume representation (Geomagic Design X ©)

volume description (.step-file)

electromagnetic field solver (CST Studio ©)

JACoW-IPAC2019-WEPRB017

measured warm, air, 40% rel.hum.: 1476.979 MHz

corrected εr = µr = 1 as simulated: 1477.439 MHz

Δf / f = 5.0 ·10 -4 = 92 µm / Dequator

Nice, but no statistics …


Benefit of high-energy X-rays: 3 x Gun 1.1

19

– forget 300 keV for everything but outer surfaces

– 600 keV-class guns resolve outer surfaces and some internal ones with noise

– 9 MeV generated X-rays will resolve most, but not all internals, noise depending on overall attenuation

300 keV DC gun(Ti/NbTi choke cell tuner not welded)

587 keV DC gun

9 MeV Linac

Hans-Walter GlockTHOTEV07 20

Example: VSR-1-cell, scanned with few degrees tilt. Artificially enhanced surface roughness in

the shadow area of the waveguide extension, also noise around the screw nuts.

Surface artifacts (though appearing rather authentic)


Conclusions - poor:

21

– Meaningful X-ray tomography of niobium cavities need highest energies available, even beyond most “industrial” demands.

– Intrinsic calibration by additional knowledge is necessary to adjust threshold values needed for material border definition.

– Data evaluation requires capable resources, experience or/and good luck.

Conclusions - optimistic:

– X-ray tomography gives access to internals of fully processed and hermetically closed cavities.

– It does a good job in integrally capturing cavity shapes down to ~ 0.2 mm spatial resolution.

– As any emerging non-destructive testing procedure it has the potential to gain reliability with increasing practice.

Industrial X-ray Tomography as a Tool for Shape and ...

Documents

Transcript of Industrial X-ray Tomography as a Tool for Shape and ...