and -XRD for the characterization and degradation of chrome yellow - - PowerPoint PPT Presentation

Synchrotron radiation-based µ-XANES and µ-XRD for the characterization and degradation of chrome yellow pigments: a focus on paintings by Vincent van Gogh

Letizia Monico

CNR-ISTM (Perugia, Italy) University of Antwerp (Belgium)

Darkening of chrome yellows in late 19th C paintings*

“[…] You were right to tell Tasset that the geranium lake should be included after all, he sent it, I’ve just checked — all the colours that Impressionism has made fashionable are unstable, all the more reason boldly to use them too raw, time will only soften them too much. So the whole order I made up, in other words the 3 chromes (the orange, the yellow, the lemon) the Prussian blue, the emerald, the madder lakes, the Veronese green, the orange lead, all of that is hardly found in the Dutch palette, Maris, Mauve and Israëls. […]” (letter n. 595, To Theo. Arles, 11 April 1888)

Sunflowers (1889, V. van Gogh; Van Gogh Museum, Amsterdam) Bank of the Seine (1887 V. van Gogh; Van Gogh Museum, Amsterdam, NL) Falling leaves (Les Alyscamps) (1888, V. van Gogh; Kröller-Müller Museum, Otterlo, NL)

* L. Monico et al., Anal. Chem. 83 (2011) 1224-1231; L. Monico et al., Anal. Chem. 86 (2014) 10804-10811.

What is changing? What can be done?

• Van Gogh was already aware of the instability of the chrome yellow pigments

Properties of lead chromate-based pigments

solubility

(1-x)PbCrO4∙xPbO PbCrO4 PbCr1-xSxO4

chrome orange [SO4

2-]>40%

Extenders

(commercial formulation)

Sulfates [BaSO4, CaSO4∙2H2O, KAl(SO4)2∙12H2O, PbSO4] Talc, kaolin, calcite Other chromate-based yellow pigments (CaCrO4 /BaCrO4) chrome yellows

Darkening of chromate-based pigments: interest in the painting conservation field

Evolution of the synthesis procedure

• controlled

experimental conditions [pH, temperature, presence of specific reagents (e.g., NH4HF2)]

• Coating methods (Sb-based compounds,

Al/Ti/Ce hydrous oxides, amorphous silica)

Until 1950

Lightfastness improvement

Late 20th c – early 21th c

(a) D. Bomford et al., in: “Art in the Making: Impressionism”, National Gallery Publications, London, 1990, p. 158;

• A. Burnstock et al., Z. Kunst technol. Konserv. 17 (2003) 74-84. (b)L. Zanella et al., J. Anal. Atom. Spectrom. 26 (2011) 1090-

1097; F. Casadio et al., Anal. Bioanal. Chem. 399 (2011) 2909-2920.

Analysis of several chromate samples taken from paintings and historical paint tubes. (a) Darkening of zinc chromate-based yellows.(b) Replacement with the more stable lead molybdate compounds. Keen interest of the darkening

• f the chrome yellow pigments

in the industrial field.

Reconsideration of the problem of the darkening of chromate-based yellow pigments (Pb-, Ba-, Sr-,Ca-, Zn/K-chromate) in the context of the conservation of paintings.

Aims and analyzed materials

1) Photochemical aging of late-19th century oil paint tubes

1 - What is the alteration mechanism of the chrome yellow pigments? 2 - What are the factors that induce the darkening of these compounds? 3 - How we can prevent/mitigate the degradation process on original paintings?

2) Study of a series of paintings by Vincent van Gogh and related micro- samples containing different types of chrome yellows

Degradation

• S-rich paint
• S-rich areas

How the sulfate anions influence the stability of chrome yellows?

Flemish Fauvist Rik Wouters (1882-1913) Elsens (Bruxelles)

UVA-Vis light Orthorhombic PbCr0.4S0.6O4 Monoclinic PbCr0.75S0.25O4

• XRD
• micro-Raman
• FTIR (transmission, ATR, reflection mode)
• UV-visible

(reflectance mode) and colorimetry

Similar information by employing portable instrumentations for non-invasive in situ measurements “Conventional-source” methods

• SR

µ-XRD (P06 and L beamline; DESY/HASYLAB, Hamburg)

Speciation/high spatial resolution methods

• Capability to distinguish among different chrome yellow types (PbCr1-xSxO4, with 0≤x≤0.8)
• Information about the alteration products
• Information about the oxidation state and the distribution of a given element
• Additional analytical/morphological information at the nano-scale level

Analytical methods

• SR µ-XANES/µ-XRF at the Cr and S K-edges

(ID21 beamline; ESRF, Grenoble)

• Energy Electron Loss Spectroscopy (EELS)
• STEM-EDX
• Electron Paramagnetic Resonance (EPR)

Characterization and photochemical stability of different crystal forms of the chrome yellow pigment

In-house synthesized and commercial pigments

Pb(NO3 )2 + (1-x) K2CrO4 + xK2SO4 → PbCr1-xSxO4 ↓ + 2KNO3

1) Synthesis of PbCrO4 and PbCr1-xSxO4 (0.1≤ x ≤ 0.75) 3) Photochemical aging treatment

S1mono S3A S3B S3c S3D D1 D2 [SO4

2-]

0% 10% 25% 50% 75% In house-synthesized PbCr1-xSxO4 (CIBA and BASF) PbCrO4 +PbSO4 (1:2) S1ortho C PbCrO4 commercial

SOLARBOX 1500e system Cermax Xenon lamp

UVA-visible light different wavelength bands

• f the UV-visible light

2) Preparation of oil paint model samples

With increasing Cr content

• Shift of the diffraction peaks

toward lower Q values.

• Increasing
• f

lattice parameters.

Possibility to distinguish different types of the chrome yellow pigments also by means of IR and Raman spectroscopies [SO4

2-] 0%

10% 25% 50% 75% S1mono S3A S3B S3c S3D S1ortho

• rthorhombic
• rthorhombic

monoclinic monoclinic

Predominantly monoclinic Cr-rich nanorods

S Cr

100 nm

PbCr0.6S0.4O4 PbSO4 S Cr

200 nm

PbSO4

Predominantly

• rthorhombic S-rich

nanoparticles *L. Monico et al., Anal. Chem. 85 (2013) 851-859.

Characterization of different chrome yellow types*

16.20 16.74 17.28

(120) (111) (111)

D1

(020) (111)

Intensity

S1mono

Q (nm

• 1)

S3A S3B S3C S3D

• rthorhombic PbCr1-xSxO4

monoclinic PbCr1-xSxO4 monoclinic PbCrO4

(201)

PbSO4

• rthorhombic PbCrO4

(111)

S1ortho

2.48 2.49 2.50 S(VI)

Normalized fluorescence Energy (keV)

S3A S3B S3c S3D PbSO4 2.481

With increasing Cr content

• Gradual disappearance of the

pre-edge signal at 2.481 keV.

• Several post-edge features

become less clearly defined. Sulfate groups are more isolated

SR µ-XRD (P06 –DESY) S K-edge XANES (ID21–ERSF) STEM-EDX

[SO4

2-]

Orthorhombic phase and [SO4

2-]≥50 wt %

0% C monoclinic PbCrO4 D1 S1ortho 50% S1mono S3A S3B S3c S3D D2 0% 10% 25% 50% 75% Monoclinic PbCr1-xSxO4 PbCrO4 +PbSO4 (1:2) 0% 65%

commercial

• rthorhombic

monoclinic

UVA-Vis light

in-house synthesized Thin alteration layer (~3-4 µm thickness)

Spectroscopic measurements at high spatial resolution

SR-based µ-XANES and µ-XRF analysis at the Cr and S K-edges (ID21 beamline; ESRF, Grenoble)

* L. Monico et al., Anal. Chem. 83 (2011) 1214–1223; L. Monico et al., Anal. Chem. 85 (2013) 85 860-867.

Cr(VI)→ Cr(III)?

Artificially aged paint model samples*

PbCr0.4S0.6O4 monoclinic

• rthorhombic

PbCr0.75S0.25O4 historical

UVA-Vis light

Cr reference compounds: Cr K-edge XANES spectra

• Cr(VI) compounds

non-centrosymmetric tetrahedral coordination.

• Cr(III) compounds

centrosymmetric octahedral geometry.

• Pre-edge peak area proportional to the

amount of Cr(VI).

• Shift towards higher energies: increasing in

the valency of the absorbing atom and/or of the electronegativity of the nearest neighbour atoms.

6.00 6.03 6.06

PbCrO4 Cr2O3

Energy (keV) Normalized Fluorescence

Intense pre-edge peak 5.993 keV 1s →3d (dipole-forbidden) pre-edge peaks of low intensity 5.993 keV 1s→3d(eg) 5.990 keV 1s→3d(t2g) shift of the position of the absorption edge

6.00 6.02 6.04

Cr(III) Cr(VI)

unaged aged

Normalized Fluorescence

Energy (keV)

5.993

Historical sample and paint model S3D: XANES analysis

Reduction of the original Cr(VI) ~60-65% of Cr(III)-species at the surface

Aged

XANES spectra: 10.5 μm

UVA-visible light

Before After

XANES spectra: 8 μm

Yellow area Brown area

Sample A S3D (PbCr0.2S0.8O4) Only Cr(VI)

Reproduction of the same alteration process as

• bserved on the historical

sample A*

0 1 2 3 4 5 6 7 8 9 10 10 20 30 40 50 60 70 80 90 100

Depth (m) Cr relative abundance (%)

Cr(III) Cr(VI)

3 comp.: Cr(III) sulfate or acetate, Cr2O3∙2H2O and PbCrO4

2 comp.: Cr2O3∙2H2O and PbCrO4

200 μm 40 μm

All XANES spectra were fitted as a linear combination of a limited set of Cr-reference compound profiles

1 2 3 4 5 6 7 8 10 20 30 40 50 60 70 80 90 100

Depth (m) Cr relative abundance (%)

Cr(III) Cr(VI)

3 comp.: KCr(III) sulfate or Cr(III) acetyl-acetonate, Cr2O3∙2H2O and PbCrO4

2 comp.: Cr2O3∙2H2O and PbCrO4

Cross-sectioned samples

* L. Monico et al., Anal. Chem. 85 (2013) 85 860-867.

VL

Cr(III) distribution

Later al di stanc e, m ic romete r 100 200 300

Cr(VI) distribution

Later al di s tanc e, m i c rom eter 100 200 300

60-70% 30-40%

5.993 6.086

In line with the linear combination fitting of the XANES spectra, Cr(III) species are localized in the upper 3-4 µm of the paint (up to 60-70%).

Cr chemical state maps: historical sample A*

* L. Monico et al., Anal. Chem. 83 (2011) 1214–1223.

Map size (v×h): 42 ×300 μm2 pixel size (v×h): 0.25 × 1 μm2 dwell time: 100 ms/pixel

ID21 beamline

• D2 (mix PbCrO4:PbSO4=1:2): stable to the UVA-

visible light exposure.

• Monoclinic D1: similar SO4

2- amount of S3C but with

a response to light exposure that is comparable to that of the monoclinic S3B (SO4

2-~25 wt %)

Aged paint models: general observations*

[SO4

2-]

0% C Monoclinic PbCrO4 commercial D1 50% S1mono S3A S3B S3c S3D D2

Before After

0% 10% 25% 50% 75% Monoclinic PbCr1-xSxO4 PbCrO4 +PbSO4 (1:2) 65%

Before After

A B1 B2

~ 20-25% Cr(III) ~ 45-60% Cr(III)

18 36 54 72 30 40 50 60 70 80

S3D(m+o) S3C(m+o) D1(m) S3B (m)

[Cr(VI)]/[Crtotal] (%) [SO4

2-](%)

S1mono(m)

y=(83.1±0.8)+(-0.54±0.02)×x, R=0.95

* L. Monico et al., Anal. Chem. 85 (2013) 85 860-867.

Orthorhombic form less stable Degradation favored when SO4

2- anions

are included inside the crystalline structure

• No significant changes when SO4

2- <50% and when only the monoclinic phase is present. monoclinic

• rthorhombic

5.985 6.000 6.015

Normalized fluorescence Energy (keV)

Cr K-edge µ-XANES 5.993

20 40 60 80 100 120 10 20 30 40

red

>570

blue

335<<525

UV

240<<400

E

Aging time (hours)

UVA-Vis

>300 nm

Aging with different wavelength bands

200 400 600 800 1000

S3D powder (diluted with BaSO4)

240400 "UV" 335525 "blue"

300 "UVA-Vis"

Reflectance wavelength (nm)

570 "red"

Diffuse reflectance UV-Vis

Aging of a series of PbCr0.2S0.8O4 unaged red UV blue UVA-Vis

% Cr(III)

0 58 ± 3 51 ± 2 32 ± 2 0

• Wavelength bands for aging selected on the basis of the

UV-Vis spectrum of the S3D powder.

• Positive correlation between the darkening and the relative

amount of Cr(III)-species.

• Significant darkening for λ≤530 nm.

* L. Monico et al., Anal. Chem. 85 (2013) 85 860-867.

Effects of different white sources and monochromatic lights

• n

the darkening process

Aging experiments with commercial white sources

Different emission of the lamp in the region around the maximium absorption of the pigment

• Aging of a series of oil paints containing the photo-stable monoclinic PbCrO4 and the

light-sensitive PbCr0.2S0.8O4 The darkening of the surface depends

• n the employed illumination device

Exposure to white sources: Cr K-edge µ-XRF/µ-XANES

• PbCrO4: up to 10-15% of Cr(III)-amount, irrespective of the employed light sources (results not shown).
• PbCr0.2S0.8O4 : positive correlation between the Cr(III)-amount and the total color change (ΔE*) or the amount of

violet-blue-green radiation (400-530 nm) emitted by the source.

• Degradation not dependent on the radiant flux.

PbCr0.2S0.8O4

PbCr0.2S0.8O4

step sizes (h×v): 0.8×0.3 µm2; dwell time: 100 ms/pixel

ID21 beamline

Exposure to monochromatic lights of PbCr0.2S0.8O4

Study of the darkening response of the light-sensitive chrome yellow pigment towards exposure to selected wavelengths.

280 350 420 490 560 0.0 0.1 0.2 0.3 0.4

561 531 288 450 500 Absolute Irradiance (W/m²/nm) Wavelength (nm) 400 PbCr0.2S0.8O4 powder Reflectance

Wavelength (nm) FWHM (nm) Irradiance (W∙m-2) Total photon counts Aging time (hours) 288 15 4.5±0.2 6.04×1024 ~254 400 16 4.8±0.5 6.62×1024 ~190 450 13 3.8±0.2 6.5×1024 ~212 500 17 4.5±0.4 6.04×1024 ~150 531 16 3.7±0.1 6.4×1024 ~180 561 17 3.9±0.2 6.7×1024 ~180

PbCr0.2S0.8O4 The wavelengths employed for the aging experiments were selected on the basis of the UV-Vis spectrum of the PbCr0.2S0.8O4 powder.

Cr(III)-oxides Organo-Cr(III)-compounds

step sizes (h×v): 0.8×0.3 µm2 dwell time: 100 ms/pixel

Monochromatic lights: Cr K-edge µ-XRF/µ-XANES

PbCr0.2S0.8O4 Confirmation of the effects of violet-blue-green light

• The Cr(III)-amount depends on the wavelength.
• Two different positive correlations: indication of the

formation of various Cr(III)-compounds. EPR analysis. Clear evidence of the presence of Cr(V)-compounds.

• Their presence can justify the effect of the green light on the darkening process.
• Cr(V)-species may play a key role in driving the photo-reduction pathways, favoring the

formation of different Cr(III)-compounds.

Original paintings and micro-samples

Detection of different crystal forms of the chrome yellow pigment and nature and distribution of Cr-based alteration products

• n. 674, To Theo van Gogh.

Arles, 4 September 1888

• n. 593, To Theo van Gogh.

Arles, 5 April 1888

• n. 689, To Theo van Gogh. Arles,

September 1888

• n. 710, To Theo van Gogh. Arles, 22 October 1888
• n. 758, To Theo van Gogh. Arles, 14 -17 April 1889
• n. 863, To Theo van Gogh.

Saint-Rémy-de-Provence, 29 April 1890

Lemon

• n. 687, To Theo van Gogh.

Arles, 25 September 1888

Yellow-

• range

Orange

Order of different chrome yellows to Theo

1 2 3

• n. 622, To Emile Bernard.

Arles, 7 June 1888. Still life with coffee pot (F 410), Coll. Basil

• P. and Elise Goulandris, Lausanne

Row of cottages in Saintes-Maries (F 420), Private collection

Use of different types of chrome yellows

• n. 622, To Emile Bernard.

Arles, 7 June 1888. Sower with setting sun (F422), Kröller-Müller Museum, Otterlo

• n. 628 To Emile Bernard.

Arles, 19 June 1888

Microsamples from original paintings*

6 samples: monoclinic PbCrO4 3 samples: similar to PbCr0.25S0.75O4 (monoclinic) 9 samples: similar to PbCr0.4S0.6O4 (monoclinic+orthorhombic)

22 cross-sections from (un)altered chrome yellow-based areas of paintings by V. van Gogh, P. Gauguin and P. Cézanne

12 samples: PbCr1-xS xO4 2 samples: monoclinic PbCrO4 and PbCr1-xS xO4 1 sample:

• rthorhombic

chrome yellow 1 sample: difference not detectable

• SR µ-XRD (P06-DESY), reflection mid-FTIR

and micro-Raman

Identification of different chrome yellow varieties

*L. Monico et al., Anal. Chem. 85 (2013) 851-859.

6.000 6.025 6.050

Energy (keV) Normalized Fluoresence

Bank of the Seine and Field with flowers near Arles*

Lateral distance, micrometer

C C Cr r r( ( (V V VI I I) ) ) C C Cr r r( ( (I I II I II I I) ) )

Lateral distance, micrometer

C C Cr r r( ( (V V VI I I) ) ) C C Cr r r( ( (I I II I II I I) ) )

Ba-L

Lateral distance, micrometer

Unaltered chrome yellow Altered layer

Field with flowers near Arles (1888, V. van Gogh; Van Gogh Museum, Amsterdam, NL) 6.000 6.025 6.050

Energy (keV) Normalized Fluoresence

Bank of the Seine (1887, V. van Gogh; Van Gogh Museum, Amsterdam, NL)

~ 60% Cr(III) Monoclinic PbCrO4 ~ 90% Cr(III)

Altered layer

S-poor PbCr1-xSxO4 (x~0.1)

Local presence of Cr(III)-secondary products: Cr(III) oxide/hydroxide, organo-Cr(III) compounds

varnish varnish

Cr K-edge µ-XRF/µ-XANES (ID21-ESRF)

Map size: 28×28 μm2 pixel size: 0.4 ×0.4 μm2 dwell time: 300 ms/pixel. Map size: 11.6×24 μm2; pixel size: 0.4 ×0.4 μm2; dwell time: 300 ms/pixel.

Only Cr(VI) Only Cr(VI)

Falling Leaves (Les Alyscamps)

Particles [Cr(III)]=~85-98% [S(VI)]=~80% Yellow [Cr(III)]=~0% [S(VI)]=~90%

Cr(III) sulfate –based particles

yellow [Cr(III)]=~95% [S(VI)]=~90% yellow

varnish

Falling leaves (Les Alyscamps) (1888, V. van Gogh; Kröller-Müller Museum, Otterlo, NL)

S-rich PbCr1-xSxO4 (x~0.5)

Cr K-edge* S K-edge* S K-edge* Cr K-edge*

particles

Energy: 6.086 keV map size (h×v): 420 ×87 μm2 pixel size (h×v): 1×0.5 μm2 dwell time: 150 ms/pixel *exp. time: 150 ms/pixel; pixel size Cr K-edge (h×v): 0.7×0.25 μm2; pixel size S K-edge (h×v): 0.7×0.5 μm2

• Cr(III)-sulfates present at the surface and inside the varnish.
• Varnish enriched of reduced sulfur-species.

• The multi-technique approach based on SR-based X-ray methods and laboratory spectroscopic

techniques allowed the degradation process of the lead chromate-based yellows and some of the factors activating this process to be elucidated.

• The darkening is ascribable to a photochemical reduction of the chromate ions to Cr(III)-
• compounds. Cr(V)-species might act as intermediates in the Cr(VI)→Cr(III) reduction process
• Key darkening factors: Sulfate amount and crystal structure of the pigment

Summary and conclusions

Orthorhombic solid solutions are more soluble and more light-sensitive than the monoclinic ones

• Smaller amount of reduced chromium were found when S-species are not included inside the

crystalline structure of the lead chromate-based pigment.

• It is possible to slow down the darkening of chrome yellows minimizing their exposure to

wavelengths shorter than about 530 nm.

• Direct evidence for the fact that the darkening of some chrome yellow paint areas in paintings

by V. van Gogh is ascribable to a reduction process of the original chromate. Identification of different chrome yellow types: lightfast PbCrO4 and light-sensitive PbCr1-xSxO4 (x~0.5) Reduced Cr found either as Cr(III)-rich particles or as a thin alteration layer