Non-destructive elemental analysis on objects d

1
Non-destructive elemental analysis on objects
dart

• Andrea Denker
• Ionenstrahllabor
• Hahn-Meitner-Institut Berlin

Ion Beam Laboratory ISL
Paintings
Art Historical Motivation
Metal Objects
Analysis PIXE
Conclusions
Ionenstrahllabor, Hahn-Meitner-Institut Berlin
2
Ionenstrahllabor ISL ion beam laboratory

• laboratory for ion beam applications
• internal and external ( 70) users
• ion energy eV lt Eionlt 800 MeV
• research areas
• materials modificationion-solid-interaction
• medical applications
• materials analysis

3
ISL Accelerators Target Stations
14 dedicated target stations
4
ion solid interaction

• ion beam induced self organisation
• lamella structure on NiO/SiO2/Si-Wafer with
  increasing fluence (230 MeV Xe)
• incidence angle 75, height 1 µm, width
  100 nm(Bolse et al., Appl. Phys. A 77, 1115
  (2003))

5
materials modification
Ion Track Membrane
irradiation stopping drying etching
rinsing
result filter with defined structure,
cylindrical or hour-glass like pores with
huge aspect ratio collaboration with industry
Gore-Membrane
3 µm
6
eye tumour therapy

• in collaboration with Charité
• 1991 first contacts
• April 1995 financing granted
• Mai 1995 laying of the cornerstone
• August 1997 technical completion
• 13.06.1998 authorisation
• 22.06.1998 first patient
• today more than 750 patients
• 85 choroidal melanomas - 97.8 tumour
  control - 95.0 eye retention
• 6 iris melanomas no relapses
• 5 choroidal haemangioma no relapses

line ofvision
proton beam
7
Materials Analysis ERDA Principle
Elastic Recoil Detection Analysis

• irradiation of the sample with heavy high
  energetic ions at grazing incidence
• coincident measurement of energy and
  time-of-flight of the out-scattered atoms from
  the sample
• pulsed heavy-ion beam (e.g. Kr, Xe, Au) with ?
  1.7 MeV/u

8
Materials Analysis ERDA Example

• Ti/Al multilayeron steel, 5 double layers of
  150 nm Al and 100 nm Ti
• scatterplot time-of-flight versus energy
• 350 MeV 197Au

9
art historical motivation the issues
19th century or older?
massive or gilded?
age
manu-facturing
Archaeology/ Art History
provenience
authenticity
attribution toworkshops
Ming orforgery?
conservation/restoration
decrease of copperin silver coins
10
motivation 19th century painting

• many copies of Allori, even by himself
• this copyunknown artist19th century??
• quality of painting(brushstroke...) ? much
  older ?

Christofano Allori (1615) Pitti Palace, Firenze
trust collection, Berlin
11
motivation 19th century or much older?

• stylistic evaluation not sufficient
• looking for more solid arguments chronology of
  pigments
• collaboration with restorators/art historians

12
motivation info by element identification
indirect dating(pigments used)
techniquemassive or gilded?
age
manu-facturing
Archaeology/ Art History
provenience
authenticity
Ming orforgery?(colours)
attribution toworkshops(metals used)
conservation/restoration
determination of Cu content
13
motivation analytical techniques

• objects valuable, unique
• required non-destructive!
• object remains in normal atmosphere
• no sampling
• no traces
• Proton Induced X-Ray Emission (PIXE)
• X-Ray Fluorescence analysis (XRF)

sampling
14
PIXE Proton Induced X-Ray Emission

• excitation of characteristic X-rays by proton
  irradiation
• multi elemental analysis
• in vacuum from Na, in air from K no organic
  materials
• determination of elements, not type of binding
• lead tin yellow - chromium yellow
• Malachite(green) (CuCO3Cu(OH)2) - Azurite(blue)
  (2 CuCO3Cu(OH)2)
• non-destructive in air and low proton
  intensities
• standard technique proton energies up to 4 MeV
• Louvre museum has a dedicated 2 MV accelerator
  for ion beam analysis

15
PIXE (Ep 1 and 4 MeV) compared to XRF

• PIXE pros
• detection limit down to µg/g(determined by
  sample and element looked for)
• depth information by variation of proton energy
• low background
• PIXE cons
• object has to be transported to accelerator
• maximal analytical depthca. 70 µm

• XRF pros
• detection limit 10 µg/g(determined by sample
  and element looked for)
• portable equipment
• XRF cons
• high background
• maximal analytical depth150 µm

16
High Energy PIXE

• analytical depth up to several mm
• large range of protons

• large cross section for K lines of heavy
  elements, better separation, less absorption
• small energy loss Þ non-destructiveÞ cross
  sections constant over large depth

• depth dependant infoline ratios
• radiation from all material, also nuclear
  reactions Þ higher background

17
high-energy PIXE analysis up to several mm

• background bearable
• Pb glass ? Pb K linesbackground from sample
• 30 µm gold foil behind 3 mm Pb glass? gold
  signal evident

18
PIXE Set-up

• analysis of art objects proton beam in air
• small straggling large and bulky objects
• HPGe 180 eV _at_ 5.9 keV
• laser forpositioning
• proton intensitylt 0.1 pA
• measuring time 200 s
• video camerasurveys anddocuments beam spot

19
measurements standards

• measurements on standards with certified
  composition
• steel (BAM 290-1) 1.88 mm and 26 mm thick
• lead bronze (BAM 374) 1.52 mm and 23.5 mm thick
• brass (BAM 375) 1.84 mm and 25.5 mm thick
• Cu traces (BAM 376) 1.62 mm and 22.8 mm thick
• thick and thin samples from the same material
• energy loss in thin sample 9 MeV, i.e.
  exit energy 59 MeV 90 of X-rays
• thick samples protons stop
• in addition5 mm thick Cu/Ag alloys from ÖGUSSA
• analysing software GUPIX

20
measurements thin standards - results
0.3 0 58.3 38.2 0 2.9 K K lines
used 1.9 4.2 0 76.6 5.8 0.9
0 4.7 5.8 0.2 K Pb K lines
radioactive source in radiation monitor nearby
brass BAM 375 1.84 mm (values in
) Fe Ni Cu Zn Sn Pb 0.21 0.11 58.32 38.02
0.21 2.90 0.3 0 58.6 38.3 0.2 2.7 L L
lines used steel BAM 290-1 1.88 mm (values in
) V Cr Mn Fe Co Ni Cu Mo W
Pb 1.91 4.18 0.24 76.03 5.12 0.33 0.08 4.83
6.27 - 1.9 4.3 0 76.5 5.8 0.9
0 4.7 5.9 0 L
21
paintings the challenges

• possible structure of a painting
• support (canvas, wood)
• ground (chalk, gypsum)
• imprimatura
• underpainting
• paint layer
• highlight
• problem paintings are not technical samples
• layers vary in thickness
• layers interfere
• ? precise determination impossible
• but that is not needed!

22
depth depending information

• collaboration with the Kunsthistorische Museum
  Wien (Dr. M. Griesser, H. Musner)
• 12 paint mock-ups using techniques of Italian and
  Flemish artists, 16./17. century
• analysed using PIXE, estimation of sequence and
  thickness from Ka/Kb, La/Ka, La/Lb ratios
• preparation of cross sections and analysed using
  a microscope

23
depth depending information

• layer element depth cross
  section PIXE
• chalk ground Ca 110 µm no depth info.
• yellow ochre Fe 50-110 µm 70/110 µm
• Azurite Cu 10-100 µm 20/ 30 µm
• lead white Pb 5-10 µm 5/ 20 µmhighlight

• works well for easy sequences
• ambiguities possible
• ? interpretation only in collaboration with art
  historians

24
Allori copy

• analytical task identification of pigments
  (indirect dating by pigment chronology)
• paint layers gt 100 µm thick
• Þ high proton energies necessary
• in total 15 different spots analysed on the
  painting

25
Allori copy

• Judiths cloth, white
• Pb, traces of Fe Þ lead white
• no Ti, Zn, Ba Þ modern pigments can be excluded
• Judiths cloth, yellow
• Fe, Pb, traces of Ca
• from Fe Ka/Kb iron on top of lead Þ most
  probably yellow ochre on lead white
• no Cd, Cr ... Þ modern pigments can be excluded
• Holofernes face, light brown
• some Fe, Pb Þ most probably ochre mixed with lead
  white
• stripe in girdle, blue
• Pb
• no Fe Prussian blue (after 1735) can be excluded
• no Co modern pigments excluded
• blue colour due to Indigo or Ultramarine
  (Na,Al-compound)

26
Flemish Painting

• adoration of the shepherds, 17. century, unknown
  artist, Gemäldegalerie Berlin
• at HMI investigated with neutron
  autoradiographyyields distribution for Cu, Mn,
  Hg..blind for Pb, Ca, Sn, Fe (chalk, ochre,
  Pb-Sn-yellow, Pb-white)
• aim of PIXEsupplemental information
• blue in sky Fe, Pb, Co, As, Cu ? Smalte
• blue in window Fe, Pb, Cu

27
Flemish Painting

• red coatFe, Pb, Hg ? Cinnabar
• yellow coatFe, Pb, Sn ? lead-tin-yellow
• lead on dark spots in large depth ? lead-white
  ground
• confirmation of neutron autoradiography results
• additional information for Fe and Pb
• identification of yellow lead-tin-yellow

28
Modigliani Portrait

• Portrait,attributed to Modigliani
• X-ray image shows second picture underneath
• one or two artists?

29
Modigliani Portrait

• broad palette of pigmentsFe, Zn, Cd, Ba, and Pb
  in various amountson some spots Cr, Se, Hg
• problem with Ti Ba La 4.47 keV Ti Ka 4.51
  keVBa Lb 4.83 keVTi Kb 4.93 keV
• Pb in largerdepth

30
Prussian Medal

• Prussian Medal, about 1790 Deutsches
  Historisches Museum, Berlin
• weighing impossible massive object? gilded?
• t 200s, Ip 0.1 pA
• result
• medal La/Ka 1.09 1 µm Au-foil La/Ka
  40, 75 Au 15 Ag 10 Cu

31
Wiener Pfennig Hoard of Tulln

• medieval silver/copper-coin, annual change of
  imprint, from 1110/20 until 1395
• first analysis in 19. centurymelting of coins
  and chemical analysis
• 1983 chronological ordering by Bernhard Koch
• Tulln 1990 10394 Wiener Pfennige (8724 g) in
  one ceramic vase

• strong corrosion all coins corroded to one
  block
• coins extracted
• restoring necessary, chemically cleaned with
  acids

32
Wiener Pfennig the analytical problem

• idea determine the copper content in the hoard
  of Tulln to verify assumption of decreasing
  silver contentproblem light element (Cu) in
  heavy matrix (Ag) ? strong absorption of
  Cu-signal
• nuclear reaction 63Cu(p,3n)61Zn 61Zn ? 61Cu ?
  61Ni67.4 keV g lineProton Inducedg-ray
  Emission

• 2 informations
• Cu X-ray 10 µm
• Cu g line 300µm

33
Wiener Pfennig, results from Cu Ka

• measured 330 coins from the hoard of Tulln
• calculation of concentration using X-rays
  (concentrations close to surface)
• result 93 Ag 5 Cu2 Pb
• contradictionto data of19. century

34
Wiener Pfennig, results from Cu g-line

• calculation of Cu content using g-line (larger
  analytical depth)
• result higher Cu concentrations than on surface
• yet larger statistical errors
• 293 coins of other findings

• grouping of coins with the same minting
  yearand origin proof of Cudepletion close to
  surface

35
Medieval brooches

• 630 - 660 used by noble ladies for
  coats iron with silver and gold
  coloured decoration, diameter about 5 cm
• 1940 - 1973 excavation of the grave
  field in Eltville (646 graves)
• one of the largest in Germany
• 1955 - 1975 restoration of the
  objects technique of the time
  stabilisation and corrosion
  protection by plastic cover
  (about 1mm thick)
• today removal of plastic impossible (destruction
  risk)

36
Medieval brooches

• range of 4 MeV protons in plastic 0.25 mm
• use of 68 MeV protons
• energy loss in 1 mm plastic (r 1 g/cm3) 1
  MeV
• small lateral straggling
• transmission after 1 mm
• Fe Ka 6.4 keV 20 Cu Ka 8.0 keV 46
  Pb La 10.5 keV 69 Ag Ka 22.0 keV 93
• absorption no principal problem
• only qualitative analysis

protons
X-rays
37
Medieval brooches

• today 19 brooches excavated 3 lost (World War
  II)
• identification of metals on 8 brooches
• doses made of brass
• confirmation inlay made of iron
• decoration made of silver and Cu/Zn/Ag alloy
• ? sophisticated objects, for noble people,yet
  productionwithout gold

38
Egyptian coffin mask

• gilded wood
• queen Satdjehuti 1600 B.C.
• excavated about 100 years ago, private ownership
• status excellent - repaired?
• sensitive object, difficult to position

• investigated on 21 points
• on nearly all pointsgt 92 Au, ca. 6 Ag, 1
  Cu, (river gold) at least 1.5 µm
• thickness and composition Þ original gilding
• mask now in the Staatlichen Sammlung ägyptischer
  Kunst, München

39
Egyptian scarab
appeared during World War II? is it real?

• investigated on 3 points
• ca. 95 Au, ca. 5 Ag, (river gold), massive
• additional synchrotron XRD on the eyes ? no
  glass
• ? most probably true ancient object

40
Chinese bowl manufacturing date?
report 1 (Japan) 500 years old 1 Mio.
20 cm
report 2 (Berlin) 100 years old max. 25 000

• both reports based on art historical expertise
• indirect dating identification of pigments
  (Cr in green after 1850)

41
Chinese bowl

• porcelain extremely sensitive
• high-energy protons small risk of damage due
  to low proton intensity and small dE/dx
• on bowl Pb (flux) and Cu (pigment)

• modern porcelain Cr (pigment)

• green colour no information
• yellow colour measured modern pigment
  detected
• Þ report 2 could beconfirmed

42
Conclusion I

• all these results could only be obtained thanks
  to fruitful collaborations with the following
  people
• Ägyptisches Museum und Papyrussammlung Berlin,
  Prof. D. Wildung
• Fachhochschule für Technik und Wirtschaft
  Berlin, K. Ebert
• Skulpturensammlung Berlin, Dr. K.
  Achilles-Syndram
• Rathgenforschungslabor Berlin, Dr. C.
  Goedicke, Prof. J. Riederer
• Universität of Marburg M.A. M. Blaich
• Kunsthistorische Museum Wien Dr. M.
  Griesser, H. Musner, R. Denk, H. Winter

43
Conclusion II

• the features of high-energy PIXE are
• elements heavier than Ca are detected
• non-destructive
• unique for thick layers
• allows analysis of sensitive and bulky objects
• fast method qualitative results after a few
  minutes
• more than 2000 objects analyzed
• coins
• brooches
• sculptures
• paintings
• ..
• fascinating interdisciplinary field

Thank you for your attention!
44
The Result of the Evaluation Process

• evaluation and recommendation of the referees
• In order to fulfil its mission, a facility with
  the necessary excellent performance was very
  successfully created.
• We are particularly worried about the
  understaffing of ISL and, therefore, recommend an
  increase of staff there, to allow them to
  increase the available beam time in order to
  respond to a growing demand in the field of
  materials modification.

Helmholtz Senate (Sep. 04)close down ISL until
end of 2007
HMI supervisory board (Nov. 04) close down ISL
until end of 2006 at the moment negotiations to
save eye tumour therapy