Content uploaded by Jaap J. Boon
Author content
All content in this area was uploaded by Jaap J. Boon on Sep 27, 2015
Content may be subject to copyright.
161
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
Received November 2005
Analytical Imaging Studies of Cross-
Sections of Paintings Affected by Lead
Soap Aggregate Formation
Katrien Keune and Jaap J. Boon
Lead soap aggregates have been found in lead-containing oil paint layers in paintings from the thirteenth to the twentieth
century. They severely affect the stability of the paint layers and disturb the surface of the paintings. Paint cross-sections from five
paintings affected by lead soaps were selected to illustrate and investigate this degradation phenomenon with the analytical imaging
techniques of Fourier transform infrared spectroscopy, secondary ion mass spectrometry and scanning electron microscopy combined
with X-ray analysis. Examples are given of lead soaps forming in a mature paint system or, alternatively, in the early drying
stage of the oil; lead soaps forming from various types of lead-containing pigments or driers; lead soaps forming in multiple paint
layers; and lead-containing crystallization products inside aggregates. The phenomenon of lead soap aggregates is multifaceted, and
one general scenario describing the formation of lead soap aggregates cannot explain all aspects. However, the integration of the
chemical information and its distribution among the paint layers leads to the proposal that reactive free monocarboxylic fatty acids
play a key role in lead soap aggregate formation. The availability and release of these fatty acids depends on the original paint
composition, the build-up of the layers, and the conservation/environmental exposure history of the painting.
INTRODUCTION
Oil paint defects related to lead soaps were first
characterized in 1997 during research associated with
the restoration of The Anatomy Lesson of Dr Nicolaes
Tulp (MH inv. no. 146), a heavily restored painting by
Rembrandt van Rijn in the collection of the Royal
Cabinet of Paintings, the Mauritshuis, The Hague.
Several articles have, since then, been published on the
large lead soap aggregates present in the lead-containing
ground of this painting [1, 2], as well as those in paint
layers of other paintings [3–6]. Lead soaps are now
understood to be a widespread phenomenon in paint-
ings [7]. Such phenomena as metal soap aggregation
and protrusion in paint layers, increased brittleness of
paint layers, increased transparency of the paint, paint
loss and efflorescence can all be related to the existence
of growing lead soap masses in paintings. These defects
have been observed in oil paintings with lead-containing
paint layers irrespective of the type of support. The
painted works of art affected range from the thirteenth
to the twentieth century and come from a wide spread
of geographical locations. Not only lead soap aggregates,
but also zinc soap aggregates are frequently observed
in paintings. Zinc soap formation and aggregation was
first demonstrated in the nineteenth-century painting
Falling Leaves (Les Alyscamps) by Vincent van Gogh [8].
Numerous nineteenth-century paintings in various
collections have also been found to be affected by zinc
soap formation [9–11]. Besides the formation of lead
and zinc soaps, other metal soaps, such as copper [12]
and potassium soaps [13], have been observed to form
in oil paintings. This paper focuses on lead soaps and
how analytical imaging studies of paint cross-sections
from paintings can elucidate the processes involved
in lead soap aggregation and mineralization. For that
purpose, paint cross-sections have been selected from
five paintings that are seriously affected in different ways
by lead soap formation.
The process of lead soap aggregate formation in
paintings is multifaceted. It is therefore important to
obtain compositional and spatial distribution information
about the components and their behaviour from studies
162 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
of paintings to understand how the composition and
multilayered nature of the paints contribute to the
process. A combination of analytical imaging techniques
using light microscopy (LM), imaging specular reflection
Fourier transform infrared (imaging-FTIR) spectroscopy,
scanning electron microscopy combined with X-ray
analysis (SEM-EDX) and imaging secondary ion
mass spectrometry (SIMS) provide best insight into
the current appearance and composition of lead soap
aggregate-containing paint films [4, 14–18]. By LM,
lead soap aggregates look transparent to whitish opaque,
while they strongly fluoresce under ultraviolet (UV)
illumination. In some aggregates small orange particles
are observed, which have been identified as red lead by
reflection visible light imaging microspectroscopy and
Raman spectroscopy [14, 15]. SEM images demonstrate
the distribution of elements such as lead, carbon,
oxygen present inside lead soap aggregates and other
elements around them. This has made it possible to
propose mechanisms for the growth process of lead soap
aggregates at the expense of certain mineral phases in
the paint. Imaging-FTIR is a vital analytical method for
the identification of metal soaps in paint cross-sections
because it detects and localizes the metal–carboxylate
bond very well [15, 16]. SIMS has been applied to
localize, speciate and map the monocarboxylic fatty
acid, lead and lead soap composition inside metal soap
aggregates in more detail [15, 17, 18].
The sources of lead in lead soaps are various types of
pigments or driers, while the fatty acids derive from the
oil network. The major detectable components in the
lead soap aggregates, besides mineralized products, are
lead soaps of monocarboxylic fatty acids. This is deduced
from SIMS data on paint cross-sections [9, 15, 17, 18] and
FTIR, gas chromatography with mass spectroscopy (GC-
MS) and direct temperature-resolved mass spectroscopy
(DTMS) data on isolated lead soap aggregates [4, 19–21].
The presence of predominantly monocarboxylic fatty
acids inside aggregates is remarkable, as dicarboxylic fatty
acids (diacids), in contrast to monocarboxylic fatty acids,
are present in significant amounts in mature oil paints.
Approximately 7% of the fatty acid moieties in fresh
linseed oil are saturated monocarboxylic fatty acids, while
the diacid oxidation products are potentially derived
from the 93% of unsaturated fatty acid moieties. Lead
soap formation is vital for the stability of oil paintings
[22] and starts at an early stage in lead white paints as
shown by solid state nuclear magnetic resonance (NMR)
[23]. Most defects associated with lead soap aggregates
are noticed after the paint has aged, suggesting that these
form during the mature ionomeric phase of the paints,
and to a lesser degree during the drying process of the
oil. Since fatty acids have been shown to react rapidly
with various lead mineral phases to form lead soaps in
vitro [22], they are now thought to play a key role in the
process of lead soap aggregate formation in aging paint
layer systems. The formation and growth process of the
lead soap masses requires sufficient free monocarboxylic
fatty acids to react with lead mineral phases in paint
layers to form the soap masses or to migrate within the
(multi)layered paint system.
In this paper, six paint cross-sections illustrative of
different aspects of lead soap aggregates in paint layers
are presented. Comparative chemical microscopic data
are presented to trace the interactions between possible
reactive components in a mature lead white paint from
the seventeenth century, in a younger lead white paint
from the nineteenth century and in a multilayered
paint system with various lead-containing pigments
from the seventeenth century. Data on various forms
of remineralization inside lead soap aggregates are also
presented.
RESULTS AND DISCUSSION
Interaction between the reactive components in
paints forming lead soaps
The greyish lead white-containing ground of The
Anatomy Lesson of Dr Nicolaes Tulp by Rembrandt van
Rijn (1632, canvas, lined), exhibited in the Royal Cabinet
of Paintings, the Mauritshuis, The Hague (inv. no. 146),
is affected by extensive lead soap aggregate formation,
which leads to crater-like holes and protrusions with
diameters of approximately 100 µm on the surface of
the painting. The paint cross-section (MH146/B39)
representing the early stages of lead soap formation is
depicted in Figure 1a. More advanced stages of lead soap
aggregate formation in this painting have been presented
before [1, 2, 14–16, 18]. The paint cross-section shows
a layer build-up of paint on a so-called double ground,
which consists of a reddish-brown earth-pigmented
layer (1) below a greyish lead white-containing ground
(2). A flesh-toned paint layer (3), composed mainly of
finely grained lead white, is positioned on top of the
lead white-containing ground, resulting in a multilayered
system with two different types of lead white-containing
paint layers. This example was chosen because it could
give some insight into the questions of whether metal
soap formation is limited to a single layer, and whether
the nature of the lead white paint itself plays a role in the
formation of metal soaps.
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
The backscattered electron (BSE) image visualizes a
different granularity in the two lead white-containing
layers 2 and 3 (Figure 1b). Layer 2 is loosely packed and
the size distribution of the lead white grains in this layer
is broad. Overall, layer 2 is a darker grey than layer 3,
which means that it has a lower BSE intensity, indicative
of a lower relative amount or density of a heavy metal, in
this case lead. Layer 3 is considered to be a single layer,
but the BSE image reveals less electron backscattering in
the lower part of the layer and more backscattering in
the upper part.
A comparable distribution in layer 3 is seen in the
specular reflection FTIR images of carbonates (C–O
stretch vibration of carbonates at 1395 cm-1) and metal
carboxylate (asymmetric COO- stretch vibration of
metal carboxylate at 1514 cm-1 ) (Figures 1c and 1d).
The carbonates representative for intact lead white
are located in the upper part of layer 3 and in ‘hot
spots’ corresponding to the position of the large lead
white particles in layer 2. The metal carboxylates are
predominant in the lower part of layer 3 and in several
areas in layer 2. FTIR spectra from the upper and lower
part of layer 3 also illustrate a difference in composition
(Figure 1e).
Figures 1f and 1g depict BSE images of the metal
soap-rich areas elucidated with imaging-FTIR at higher
IMAGING STUDIES OF CROSS-SECTIONS OF PAINTINGS AFFECTED BY LEAD SOAP AGGREGATE FORMATION 163
Figure 1 Paint cross-section MH146/B39 taken from The Anatomy Lesson of Dr Nicolaes Tulp by Rembrandt van Rijn
(1632): (a) LM image; (b) BSE image; (c) FTIR image of carbonates at 1395 cm-1; (d) FTIR image of lead carboxylates at
1514 cm-1; (e) FTIR spectra corresponding to pixels in the FTIR image; and (f) and (g) BSE images of details marked in (b).
164 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
magnification. The metal soap-rich areas contain lead
white particles with greyish pigment-reduced halos.
Arrow 1 in Figure 1f indicates a lead white particle
surrounded by a greyish lead soap region with several
tiny lead white particles. Arrow 2 points towards a large
lead white particle that is partially reacted away and
is surrounded by a greyish halo indicating lead soaps.
Figure 1g visualizes a lead soap-rich area: the relative
number of pigment particles is much reduced and there
is less electron backscattering in these areas. In Figure 1g
the boundary between layers 2 and 3 is hardly visible,
while in Figure 1f the boundary is more obvious.
The lead soap regions shown by FTIR in layer 2
correspond with pigment-free greyish halos around
lead white particles and the transparent greyish region
in the BSE image. This suggests a reaction of fatty acids
with lead white particles to form lead soaps. As most
of the mineral matter is absent in the saponified region
of layer 2 (see, for example, Figure 1g) and a few tiny
remaining particles are seen in the halos, it is believed
that primarily smaller lead white particles have been
completely converted to lead soaps. Dissolution of lead
white particles is also observed in paint cross-section
MH146/B37, with a similar layer structure, which
originates from a different location in the painting. This
paint cross-section consists of a large lead soap aggregate
positioned in layers 2 and 3 (see Figure 2 inset). A detail
of the BSE image of the top of the aggregate and part of
layer 3 is depicted in Figure 2. Tiny lead white particles
are visible around the aggregate. The number of particles
and their size decrease towards the rim of the aggregate.
No distinct edge is observed between the paint and
the aggregate. It is deduced from the BSE image that a
conversion of lead white particles in layer 2, as well as in
layer 3, into lead soaps is taking place.
The FTIR data and the BSE image indicate that
two-thirds of layer 3 – the bottom part – of paint
cross-section MH146/B39 is also affected by lead soap
formation. The question arises whether layer 3 is really
a single layer, especially because the few large vermilion
particles are observed in the top of this layer (Figure 1a).
However, the LM image shown reveals only the right-
hand side of the paint cross-section, while in the middle
and left-hand side of this sample the vermilion is more
homogeneously distributed (not shown). Furthermore,
a partially saponified lead white top layer (3) positioned
on the double ground (1 and 2) is also observed in paint
cross-section MH146/B37.
As this phenomenon is observed in two different
cross-sections, it is concluded that layer 3 in MH146/
B39 is indeed a single layer, which is partially saponified.
Layer 3 has a relatively high pigment/binder ratio, so the
layer is likely to be poorer in medium than layer 2. The
electron transparency in the BSE image indicates that
only the lower part of layer 3 is affected by lead soap
formation and that the mineral fraction is reduced. Since
only the bottom part of layer 3 contains lead soaps, it
is likely that free fatty acids migrated from the layer(s)
below, which then reacted with lead white to form lead
soaps. Another possibility would be that lead soaps from
the layer(s) below migrated to the bottom of layer 3, but
in this case the mineral fraction would not be reduced,
as lead soaps would not be reactive towards lead white. A
further argument for this proposal is the poorly defined
boundary between layers 2 and 3 (Figure 1f). It cannot
be excluded that layer 3 has expanded due to the lead
soap formation in this layer. Both lead white paint layers
(layers 2 and 3) are thus postulated to react with fatty
acids to form metal soaps, and metal soap formation
is not limited to a single layer nor dependent on the
granularity of the lead white, but is determined by the
availability of free monocarboxylic acids. The lead white
pigment in these paint cross-sections is, in the authors’
opinion, the main lead source for the formation of lead
soaps because of the relatively large amount of lead soaps
present. However, it cannot be excluded that a lead
drier, lead-containing oil or red lead, has initially played
a role in the formation of the lead soaps. The key to the
formation of excess lead soaps remains the availability of
free fatty acids.
The second case study illustrates the migration of
reactive free fatty acids towards the surface of lead-
Figure 2 Paint cross-section MH146/B37: detail of BSE image of top
side aggregate; visible light image (inset).
IMAGING STUDIES OF CROSS-SECTIONS OF PAINTINGS AFFECTED BY LEAD SOAP AGGREGATE FORMATION 165
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
containing pigment particles to form lead soaps. This
phenomenon was investigated in a lead–tin yellow type
I paint [24] with a study of the paint cross-section CIA
1577 RS06 taken from a triptych (c. 1480, oak panel)
painted by an unknown Northern European painter,
displayed in the Almshouse of St John the Baptist and
St John the Evangelist, Sherborne (Dorset, UK) and
presently owned by the National Trust. The detailed
study of this paint cross-section leads to better insight
into lead soap aggregate formation in lead–tin yellow
paints. Data on another cross-section from this painting
have been presented earlier [25].
The paint of sample CIA 1577 RS06 is built up in
four layers, a glue–chalk ground (1), an imprimatura
layer (2) (not visible in this paint cross-section), a lead–
tin yellow type I paint (3) and a top layer of a copper
green glaze (4) (Figure 3a). The lead–tin yellow paint
layer consists of a large lead soap aggregate on the left
and a smaller intact lead–tin yellow pigment agglomerate
on the right. The BSE image illustrates the lead soap
mass, a large inhomogeneous light-greyish aggregate
with dark-grey ‘cracks’ (Figures 3b and 3c). Tiny highly
electron-backscattering particles are positioned around
the soap mass. SIMS images illustrate tin and lead
around the aggregate corresponding with the highly
backscattering particles and with the intact lead–tin
yellow pigment agglomerate on the right (Figures 3b,
3d and 3e). Lead is the only metal detected inside the
metal soap aggregate (Figure 3d). The morphology of
the lead soap aggregate and the lead–tin yellow particles
positioned around the aggregate suggest that the growing
lead soap aggregate has pushed residual lead–tin yellow
Figure 3 Paint cross-section CIA 1577 RS06 from the Sherborne Triptych (unknown northern painter, fifteenth century): (a) visible LM image; (b) and
(c) BSE images; (d)–(i) SIMS images: (d) lead (+; m/z 208), (e) tin (+; m/z 52), (f) copper (+; m/z 63), (g) sum of deprotonated palmitic and stearic acid
(-; m/z 255 + m/z 283), (h) palmitic acid lead soap (+; m/z 461–463) and (i) chloride (-; m/z 35).
166 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
particles aside (Figure 3c). Furthermore, the distribution
of the copper corresponding to the green glaze top layer
shows that this layer is pushed up due to the expanding
lead soap aggregate (Figure 3f). The monocarboxylic
fatty acids (palmitic and stearic acids) are mainly
located in the aggregate and to a lesser extent in the
porous intact lead–tin yellow particle (Figure 3g). A
similar distribution is found for the fatty acid lead soaps
(palmitic acid lead soap in Figure 3h). Although some
lead soaps are detected in the porous lead–tin yellow
particle, their presence did not lead to the destruction
of the particle by expansion. Chloride is found to be a
dominant feature in the lead soap aggregate, as well as in
the intact porous lead–tin yellow particle (Figure 3i).
These analytical observations allow the proposal
of a mechanism that relates the degradation of the
original lead–tin yellow pigment to the formation of
lead soap aggregates. A hypothesis for the mechanism
has been presented earlier [25]. An essential element
in the formation of lead soaps is the existence of
nonstoichiometric lead orthostannate phases in the
lead–tin yellow type I pigment as a consequence of the
production process of the pigment. The resulting mixed
phases could be lead oxides co-crystallizing with lead
orthostannate or may be present as mixed crystals of
lead orthostannate and lead orthoplumbate, which have
a very similar crystal structure [26]. It is proposed that
the free monocarboxylic fatty acids react with reactive
lead phases to form lead soaps, leaving the less reactive
lead orthostannate unaffected. The local formation of
a relatively high amount of lead soaps leads to volume
expansion of the lead–tin yellow pigment particle
and drives the residual lead orthostannate towards the
periphery of the growing soap aggregate. The volume
of the lead soap mass is much larger than the original
lead–tin yellow particle(s), as can be deduced from the
green glaze layer pushed upwards. The reactive lead
fraction in the porous lead–tin yellow particle on the
right in sample CIA 1577 RS06 must have been too low
to result in sufficient lead soaps for fragmentation of the
particle. As the lead soaps are formed inside the porous
lead–tin yellow particle by the reaction of a reactive lead
component and fatty acids, it is concluded that the fatty
acids have migrated towards the surface of the lead–tin
yellow particle. From these two case studies, it can be
deduced that the monocarboxylic fatty acids separate
from the mature ionomeric network.
The third case study shows that lead-containing
mineral pigments do not always react away to form
lead soap aggregate-containing paint. In a commercial
primed canvas found in the archive of Olana State
Historic Site, home of the American artist Frederic
Edwin Church (1826–1900), the lead white particles
in the ground remain intact while efflorescence and
lead soap aggregates are observed. Oil paint degradation
phenomena are often seen in paintings by F.E. Church
and other painters from the Hudson River School. They
are postulated to originate in the ground and strongly
affect the appearance of the paintings [27]. Commercially
primed canvas was imported from England and bought
locally in New York by Church and his colleague artists
[27]. The chemistry of the primed canvas has been the
subject of extensive studies by GC-MS, DTMS, FTIR
and SEM-EDX [19, 20].
A paint cross-section (PCC01) originating from the
primed canvas reveals the layer structure of the ground,
a large aggregate in the middle of the sample and a
very thin dark layer of dust on top of the cross-section
(Figure 4a). The BSE image reveals a calcium carbonate
layer (1), a mixture of calcium carbonate and lead white
(2) and a densely packed lead white layer (3) (Figure
4b). The open holes in the aggregate are artefacts due
to electron beam exposure. The BSE image shows the
aggregate as a heterogeneous greyish mass with a dark-
grey core and a lighter-grey rim without particles inside
(Figure 4f). The pigment particles around the aggregate
and in the rest of the paint sample appear to be intact;
the transparent grey dissolution halos observed in
Figure 1f are not found around the pigment grains here.
Imaging-FTIR visualizes the presence of carbonates
(1405 cm-1) in all three layers except in the aggregate
(Figure 4d). Metal carboxylates (1510 cm-1) occur in
sufficient concentration for FTIR analysis only in the
aggregate (Figure 4c). The FTIR spectrum of a pixel
in the aggregate has two sharp peaks of C–H stretch
vibrations at 2924 and 2856 cm-1 for aliphatic chains,
and the asymmetric and symmetric COO- stretches of
lead carboxylate at 1510 and 1415 cm-1 (Figure 4e). A
spectrum derived from a pixel directly adjacent to the
aggregate reveals only the intense peak of the C–O
stretch vibration of carbonates at 1405 cm-1 (Figure 4e).
It seems that lead soaps are concentrated primarily in
the aggregate, while the mineral fraction remains intact.
SIMS analysis resulted in a better understanding of the
organic composition in and around the aggregate. Prior
to the SIMS analysis, the cross-section was coated with
gold (thickness 2 nm) to improve the organic ion yields
[28]. Figures 4g–4k depict the positive secondary ion
images of lead (m/z 206–208), calcium (m/z 40) and
lead stearic acid salt (m/z 489–491) and the negative
secondary ion images corresponding to deprotonated
stearic acid (m/z 283) and deprotonated azelaic acid (m/
IMAGING STUDIES OF CROSS-SECTIONS OF PAINTINGS AFFECTED BY LEAD SOAP AGGREGATE FORMATION 167
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
Figure 4 Paint cross-section PCC01 originating from the primed canvas from the estate of F.E. Church (1826–1900): (a) LM image; (b) BSE image;
(c) FTIR image of lead carboxylates at 1510 cm-1; (d) FTIR image of carbonates at 1405 cm-1; (e) FTIR spectra corresponding to pixels in the FTIR image;
(f) BSE image; (g)–(k) SIMS images: (g) lead (+; m/z 206–208), (h) calcium (+; m/z 40), (i) lead stearic acid salt (+; m/z 489–491), (j) deprotonated stearic
acid (-; m/z 283) and (k) deprotonated azelaic acid (-; m/z 187). Images f–k show artefacts resulting from the sample preparation: a circular feature in the
centre of the paint sample (indicated with the dashed area), an area on the right covering some of layer 3 and a scratch running from middle top to middle
right centre.
168 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
z 187). Lead is detected in low yields in layer 1, in higher
yields in layers 2 and 3 and in much higher yields in the
core of the aggregate (ignoring the artefacts). Calcium is
only detected in layer 2 and in low yields in layer 1. The
relatively low ion yields of lead and calcium in layers
1 and 2 can be explained by the suppression of ions
from inorganic elements due to the gold coating [28].
The image of calcium and lead is much more defined
and sharper. There is a relatively high concentration of
stearic acid in the core of the aggregate, corresponding
to the dark-grey area of the aggregate in the BSE image.
The distribution of stearic acid in layer 3 also correlates
with the higher yield of lead in this layer (palmitic acid
has a similar distribution but is not shown here). The
ion yields of lead stearate (m/z 489–491) are high in the
core of the aggregate and low in the rim. The lead soap
ion intensity is lower in layers 2 and 3, but corresponds
roughly with the distribution of stearic acid (the ions
of lead palmitate have a similar distribution but are not
shown here). The azelaic acid ion is detected in the rim
of the aggregate and corresponds with the lighter-grey
region of the aggregate visualized in the BSE image.
The presence of lead-bound palmitic, stearic and azelaic
acids inside the aggregate is confirmed by DTMS studies
on isolated aggregates from the primed canvas [19, 20,
29]. Secondary ions of azelaic acid as well as stearic acid
were detected outside the paint sample, especially on
the surface of the Technovit embedding resin (Heraeus
Kulzer, Germany) above the paint sample. Since the gold
coating was thermally deposited on the paint sample, it is
likely that a fraction of the fatty acids has evaporated and
was deposited above the paint sample. Taking this into
account, the fatty acids detected by SIMS in the paint
sample must be stabilized through binding to metals or
as esters, and are not easily released by evaporation.
All the imaging techniques, FTIR, SEM and SIMS,
reveal a relatively high concentration of lead and/or
lead soaps in the aggregates. Based on the FTIR and
SEM data, it seems that lead soaps are very concentrated
in the aggregate. SIMS results confirm that lead soaps
are present in the aggregate, but also in the top layer
and occasionally in layer 3. The concentration of lead
soaps in layers 2 and 3 is below the detection limit of
specular reflection FTIR. It is not certain whether the
distribution detected by SIMS is a representative quanti-
tative picture of the fatty acid distribution. However,
there is an interesting difference in the distribution of
monocarboxylic fatty acids, in the core, and diacids,
which are only detected in the rim of the aggregate. This
distribution suggests a phase separation during or after
the aggregation and expansion of the lead soap mass,
implying that monocarboxylic fatty acids and diacid-
forming fatty acids were present simultaneously.
The defects in paintings by F.E. Church were
observed by Church himself and were also partly
repaired by him [27]. This indicates that the oil paint
defects were appearing after a relatively short period of
time. Church thought at some point that they might be
connected with the application of additional drier and
oil on an already primed canvas, a common practice
at the time [30]. If lead acetate, a potential source of
reactive lead, was added as a drier, it is now no longer
detectable as such in the analytical imaging data. As the
lead white mineral fraction in the ground appears still
to be in good condition, lead driers such as lead acetate
are presumed to have been the main reactive metal
source. Lead acetate can react rapidly with fatty acids
to form lead soaps. Lead soaps in the presence of intact
lead white are, therefore, taken here as evidence for
excessive amounts of lead driers such a lead acetate. The
lead soaps formed are postulated to have been dispersed
in the paint system initially and to have subsequently
migrated towards the surface or aggregated. Local
high concentrations of lead soap in the paint film may
have acted as nucleation points for aggregate growth.
The separate phases of monocarboxylic fatty acid and
diacid metal soaps in the aggregate indicate that these
compounds were present and mobile at the same time.
It is very likely that the lead acetate reacted with the
unsaturated monocarboxylic fatty acids, still present in a
young paint film, some of which would react further by
oxidation to form lead diacids. This implies that the lead
soap formation occurred during the drying of the oil,
before a mature oil paint system had developed. At this
stage it is not clear whether the fatty acids are derived
from the oil of the primed canvas or from oil added
in a later stage by the artist himself [30]. This process
contrasts with the lead soaps in the paints presented in
the previous case studies, where monocarboxylic fatty
acids are separating from the network leaving a dispersed
metal coordinated ionic network in place.
Multilayered paint system with various lead-
containing pigments
The reactivity and distribution of various lead-containing
sources in a multilayered, multi-component paint system
was studied in a paint cross-section (T03237s9) from A
Lady of the Greville Family and her Son (1640, canvas) by
G. Jackson at Tate in London. This cross-section has an
interesting layer structure, because all layers include a
lead-containing pigment. The sample consists of seven
IMAGING STUDIES OF CROSS-SECTIONS OF PAINTINGS AFFECTED BY LEAD SOAP AGGREGATE FORMATION 169
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
layers (Figure 5a): an oil-chalk ground with some lead
white particles (1); a grey priming consisting of lead
white and charcoal black (2); a pale brown paint with
lead white (3); an orange/red opaque underpainting
with red lead, red lake, vermilion and a white opaque
lead soap aggregate (I) (4); a thin red lake glaze (5); a
dark-yellow/orange matrix with red lake, lead white,
chalk and a white opaque lead soap aggregate (II) (6);
and a bright-yellow highlight with lead–tin yellow and
lead white (7). The lead soap aggregate (II) pushes up
the surface paint in layer 7 (Figure 5a). Layers 4 and
5 represent the sitter’s red chair, while layers 6 and 7
correspond to a yellow button on the upholstery.
The large lead soap aggregates I and II visualized
in the BSE image exhibit high electron backscattering
and a lamellar structure (Figure 5b). Imaging-FTIR
confirms the presence of lead soap in aggregates I and
II (Figure 5c) and shows the presence of carbonates
inside the opaque aggregates (Figure 5d). The BSE and
FTIR images also reveal the presence of the preliminary
stages of formation of lead soap aggregates in layers 4, 5
and 6. In the LM, BSE and FTIR images, the smallest
lead soap-rich areas are indicated by X, while the larger
ones are indicated by Y (Figures 5a–5c). The smaller lead
soap-rich areas, X, embedded in layer 4 (Figures 5e and
f) are starting to push up the red lead paint. Note that
red lead particles seem to be present inside the lead soap
aggregates in the LM image, but this is not confirmed by
the BSE image. The observation of red lead inside the
aggregate is an optical effect, as the lead soap aggregate
is transparent. It is proposed that these masses, X, which
originate in layer 4 and expand into layer 6, develop into
lead soap aggregates like Y. The elemental maps of lead,
calcium, tin and aluminium support this interpretation
(Figure 5g). Lead (shown here as cyan) is detected inside
the lead soap aggregate, Y. Aggregate Y has pushed the
chalk (calcium, shown as white) of layer 6 aside and
interrupts the red lake glaze on an aluminium-based
substrate (aluminium, shown as yellow) of layer 5. The
elemental maps do not directly support a similar growth
mechanism for lead soap aggregate II. Aggregate II is
characterized by lead, but the chalk (calcium) encloses
the lead soap mass and the aluminium from the red lake
glaze is detected in layer 5 underneath the aggregate.
The lead source for the lead soaps in masses X and Y is
the red lead from layer 4 and the fatty acids are supplied
by the organic-rich layer 6. It is not clear what the lead
source of aggregate II is. A lead soap aggregate formed
from lead–tin yellow-containing oil paint would be
expected to have tin-containing particles enclosing the
aggregate [25]. The absence of these particles around
aggregate II (see tin distribution (red) in Figure 5g)
clarifies that the lead in aggregate II does not derive
from the lead–tin yellow in layer 7. It is possible that in
this case lead white or lead-containing drier present in
layer 6 is the lead source. However, if a lead drier plays a
role in this layer, the relatively large amount of lead soap
is difficult to explain. As aggregates X and Y indicate that
layer 4 is the lead source for the lead soap aggregates,
migration of lead soaps from layer 4 into layer 6 is a
plausible explanation. Migration of lead soaps into other
layers or to the surface of the painting is quite often
observed in paint systems [13, 31, 32].
The lead soaps in the lead–tin yellow layer (7) are
homogeneously distributed, and no aggregate formation
of lead soap is observed. Excess formation of lead soap
does not affect all lead-containing layers in this sample.
The presence of carbonates and the absence of lead
carboxylates in layers 2 and 3 indicate that the lead white
has remained unaffected (Figures 5c and 5d). This would
point to a limited availability of reactive fatty acids.
To conclude, the process of lead soap aggregate
formation is certainly not limited to a single layer, but
affects the whole multilayered system, although this is
expressed in different ways. In this example, lead oxides
appear to be more reactive than lead white towards
fatty acids. The fact that the lead white-containing
layers remain unaffected can be caused by a relatively
low concentration of organic compounds. It must be
mentioned that the relative amount of available reactive
free fatty acids also determines the formation of lead
soaps. The role of the amount and availability of fatty
acids can also be deduced from the formation of lead
soap aggregates in layer 6. This layer seems to be rich in
organic material and may be a good source of reactive
free fatty acids. The nature of such a red lake oil paint
determines the reactivity towards lead oxides in layers 4
and 7, as the fatty acids are not sufficiently ‘entrapped’ by
a metal ion in this layer.
Remineralization inside lead soap aggregates
Mineralization processes are observed inside the lead
soap aggregates, as was noted in the previous cross-
section. Often, lamellar bands are seen in large lead
soap aggregates, and the compound precipitating in
lead soap aggregates is a form of lead carbonate [4, 14].
However, in many examples orange particles, identified
as red lead, are also found to be associated with lead
soap aggregates [14, 15]. It is argued that these particles
inside or on the rim of the aggregates are formed
during their development and cannot be a residue of
170 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
Figure 5 Paint cross-section T03237s9 taken from A Lady of the Greville Family and her Son by G. Jackson (1640): (a) LM image; (b) BSE image; (c) FTIR
image of lead carboxylates at 1519 cm-1; (d) FTIR image of carbonates at 1400 cm-1; (e) BSE image, (f) light microscopic image; (g) overlay of elemental
distribution of lead (cyan), aluminium (yellow), calcium (white) and tin (red); (h) outline of paint cross-section indicating location of FTIR (green box), high
magnification BSE image (e) and LM image (f) (blue box) and elemental maps (red box).
IMAGING STUDIES OF CROSS-SECTIONS OF PAINTINGS AFFECTED BY LEAD SOAP AGGREGATE FORMATION 171
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
driers added to the lead white ground. This will be
illustrated with a paint cross-section (HSTB43/3) taken
from the herald (on the south-west side of the room)
painted by Christiaen van Couwenbergh (1651) as part
of the Oranjezaal of the Royal Palace Huis ten Bosch
(The Hague, The Netherlands). The paint cross-section
encloses a large aggregate with orange particles in the
lead white ground.
The sample HSTB 43/3 consists of four layers (Figure
6a): two cream-coloured ground layers containing lead
white and umber (1 and 2); a layer containing pure
lead white (3); and a thin layer with bone black, umber
and red ochre particles (4). Layer 2 contains a large
heterogeneous aggregate (diameter about 100 µm)
with many horizontally oriented orange particles of
red lead of about 3–4 µm diameter inside the partially
transparent and partially white and opaque mass. The
centre of the aggregate does not fluoresce, in contrast to
the left, right and upper sides near the rim (not shown).
Imaging-FTIR at 1516 cm-1 revealed metal soaps with
a high relative intensity inside the aggregate and with
much lower intensities in layers 2 and 3 (Figure 6b).
The areas where no metal soaps are detected correspond
to the large lead white particles. The FTIR peaks
representative of metal soaps are absent in layer 1. The
carbonates at 1408 cm-1 are very abundant in all the
layers and have a higher intensity in the centre of the
aggregate, showing the precipitated lead carbonates in
the centre (Figure 6c). The BSE image reveals a densely
packed paint with a wide size distribution of lead white
particles (Figure 6d). The boundaries between layers
1, 2 and 3 are hardly visible. Layer 1 has a higher BSE
intensity than layers 2 and 3 (ignoring the tiny dark
spots corresponding to the brown particles in layer 2).
The aggregate has a heterogeneous structure with fine
particulate lamellar bands in the centre (Figure 6e)
and multiple coarser-grained crystals. The rim shows a
lower BSE intensity than the centre. The bright highly
electron-backscattering particles inside the aggregate
correspond to the orange particles of red lead seen
in the visible light image. Their identification was
confirmed by Raman spectroscopy (courtesy of the
Scientific Department of the Metropolitan Museum,
New York). Some of the crystals observed in the visible
light image are positioned deeper in the lead soap mass.
The conditions for formation of these new crystal phases
in paintings are not completely understood. Red lead
is stable under relatively high pH and redox potential
(Eh) conditions (see pH-Eh diagram in [15]) and is
expected to form under similar conditions when lead
soap destabilizes releasing lead ions. Considering the
reactivity of free fatty acids towards red lead and lead
hydroxycarbonate, it has to be concluded that the fatty
acids remain sequestered when these mineral phases
develop inside a lead soap mass.
It is remarkable that the aggregate is present in a
relatively intact, densely packed lead white matrix,
with a relatively low concentration of lead soaps. This
layer seems to be relatively medium-poor, and no lead
compounds other than lead white are present. Lead
soap aggregates have generally not been observed in
the grounds of the paintings of the Oranjezaal, thus this
example is an exception [33]. It is thought that the lead
soap aggregate started as an accidental oil-rich inclusion
in the priming ground and developed further into a
metal soap mass in the paint when reactive fatty acids
were released by hydrolytic processes affecting the ester
bonds. A further initial presence of lead drier or leaded
oil cannot be excluded.
The relative number, particle size and position of
the orange particles of red lead are convincing evidence
that red lead was not originally present in the paint.
In the hypothetical case that red lead could have been
the starting material for lead soap formation, it is
expected that it would have reacted away [34]. Red lead
is not observed in the numerous cross-sections made
from 30 paintings in the Oranjezaal with similar lead
white-containing oil grounds. The red lead particles
are positioned in the aggregate between the lamellar
structure and the lead soap-rich rim. This indicates that
they were created during the formation of the lamellar
structure or at least during or after the growth of the
aggregate.
Chloride inside lead soap aggregates
The aggregates in sample HSTB 43/3 contain, besides
mineralization products, a relatively high concentration
of chlorine (assumed to be present as chloride), which
is visualized by SIMS. The advantage of SIMS applied
to paint cross-sections is that elements which cannot
be detected by EDX due to their low concentration
can often be detected by SIMS. The SIMS images in
Figures 6f–6h show the distribution of lead (+; m/z
208), chloride (-; m/z 35) and deprotonated palmitic
acid (-; m/z 255). The lead is homogeneously distributed
in the aggregate and in layers 1 and 2. Chloride and
deprotonated palmitic acid predominate in the aggregate.
The chloride is more abundant in the centre of the
aggregate, whereas the palmitic acid ion yields are higher
near the rim on the left-hand side. The secondary ion
yields of palmitic acid are lower in layers 1, 2 and 3. The
172 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
Figure 6 Paint cross-section HSTB43/3 originating from a herald (south-west corner of the Oranjezaal, Huis ten Bosch) by Christiaan van Couwenbergh
(1651): (a) LM image; (b) FTIR image of lead carboxylates at 1516 cm-1; (c) FTIR image of carbonates at 1408 cm-1; (d) and (e) BSE images; (f)–(h) SIMS
images: (f) lead (+; m/z 208), (g) deprotonated palmitic acids (-; m/z 255), (h) chloride (-; m/z 35); (i) partial positive SIMS spectrum.
IMAGING STUDIES OF CROSS-SECTIONS OF PAINTINGS AFFECTED BY LEAD SOAP AGGREGATE FORMATION 173
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
lead soaps of palmitic and stearic acids are not abundant
in the positive ion spectrum and their intensities are
too low to image. Figure 6i, which presents part of the
positive ion spectrum (range m/z 405–495) calculated
for the aggregate, shows the lead cluster ions of Pb2
(m/z 410–416), Pb2O (m/z 426–432) and Pb2O2H
(m/z 443–449). The lead palmitate ions expected at m/z
461–463 overlap with isotope peaks from Pb2ClO (m/z
461–469). Two maxima in the nominal ion peak of m/z
463 observed at 463.2 D and 462.9 D are representative
for PbOOC(CH2)14CH3 and Pb2OCl, respectively. The
lead stearate ions overlap with the ion pattern of an
unidentified compound. Various lead chloride clusters
were detected in the negative as well as in the positive
SIMS spectrum. These data suggest the existence of lead
chloride mineral phases in the aggregate. In earlier work,
lead hydroxychloride (Pb3Cl4(OH)2) was identified
by X-ray diffraction analysis in a large remineralized
aggregate of The Anatomy Lesson of Dr Nicolaes Tulp [2].
In cosmetic containers from ancient Egypt, phosgenite
(Pb2CO3Cl2) was identified together with lead soap [35].
The source of chloride could be indirectly the lead white
itself, because the production of lead white is known to
involve various procedures and purification methods
that could contribute chlorides [36]. The exact role of
chloride and its relevance is currently unknown but it is
worth reporting. Chloride may play an important role in
metal soap formation and mineralization processes, as it is
found in high abundance in the aggregates. It is possible
that chloride-containing lead soaps are formed, in which
the chloride (Cl-) is the counter ion for lead fatty acid
salts – Pb(FA)+, because lead is divalent (Pb2+). No direct
evidence for the presence of Pb(FA)Cl as a molecular
entity has yet been found. The general association of
chlorides with lead soaps is an indication that lead masses
in paintings should not be considered as pure phases, but
rather as reaction products that are influenced by fluxes
of materials in the paintings.
CONCLUSION
The paint cross-sections presented in this paper show
the different aspects of lead soaps in multilayered paint
systems. The lead soap aggregates manifest themselves in
various ways depending on the paint composition, the
build-up of the painting and the chemical history of the
paint layers over the centuries. The analytical imaging
techniques employed can reveal many aspects of the
formation process of lead soap aggregates in paintings
but the actual process of formation requires chemical
and physical experiments that model the processes.
The various case studies demonstrate that one general
scenario describing all aspects of the lead soap aggregates
in paintings cannot yet be given. The difficulty is that
although the formation of lead soaps is straightforward,
the conditions under which this is happening are so
multifaceted.
It is deduced from the case studies of samples
MH146 and CIA 1577 RS06 that the monocarboxylic
fatty acids separated from the mature ionomeric oil
network and migrated towards reactive lead compounds
to form lead soaps. In sample PCC01 it is postulated
that excess lead drier reacted with drying oil to form
lead soaps, as mono- and dicarboxylic fatty acid lead
soaps are detected together in separate phases in an
otherwise apparently unaffected oil-primed canvas. Since
paintings are multilayered, fatty acids and lead soaps can
migrate by diffusion into other layers, thus creating new
chemical conditions. Samples MH146 and T03237s9
demonstrate that lead soap formation is not limited to
a single layer. The monocarboxylic fatty acids in samples
from painting MH146 migrated to the medium-poorer
lead white layer, whereas the lead soap aggregates in
sample T03237s9 started at the interface of a lead- and
medium-rich layer and grew into the medium-rich layer.
The aspect of lead soap masses is further complicated
by mineralization processes where lead is sequestered
into new mineral phases involving lead oxides, lead
carbonates and lead chloride phases.
The purpose of this paper has been to demonstrate the
complexity of the lead soap aggregates in paintings and
to show the potential of analytical imaging techniques to
unravel this complexity as far as is possible with samples
from paintings with an unknown initial composition,
and environmental and restoration history.
ACKNOWLEDGEMENTS
The authors are very grateful to P. Noble, Royal Cabinet
of Paintings the Mauritshuis in The Hague; J. Zucker,
New York State Bureau of Historic Sites at Peebles
Island; L. Speleers, the Institute for Atomic and Molecular
Physics (AMOLF), Amsterdam; A. Burnstock, Courtauld
Institute of Art, London; and J. Townsend, Tate, London,
for providing the paint cross-sections and for fruitful
discussions. The research at AMOLF forms part of
the Foundation for Fundamental Research on Matter
(FOM) research programme no. 49, ‘Mass spectrometric
imaging and structural analysis of biomacromolecules’.
This work was part of the De Mayerne programme
funded by the Dutch Organisation for Scientific
Research (NWO) and FOM, a subsidiary of NWO.
174 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
REFERENCES
1 Middelkoop, N., Noble, P., Wadum, J., and Broos, B., Rembrandt
under the Scalpel: The Anatomy Lesson of Dr Nicolaes Tulp Dissected,
Six Art Promotion bv, Amsterdam/The Hague (1998).
2 Noble, P., Wadum, J., Groen, K., Heeren, R., and van den Berg,
K.J., ‘Aspects of 17th century binding medium: Inclusions in
Rembrandt’s Anatomy Lesson of Dr Nicolaes Tulp’, in Art et chimie:
La couleur, ed. J. Goupy and J.-P. Mohen, CNRS Éditions, Paris
(2000) 126–129.
3 Plahter, U., ‘Baburen re-examined’, in Conservare necesse est:
Festschrift til Leif Einar Plahter, IIC Nordic Group, Oslo (1999)
65–67.
4 Higgitt, C., Spring, M., and Saunders, D., ‘Pigment-medium
interactions in oil paint films containing red lead or lead-tin
yellow’, National Gallery Technical Bulletin 24 (2003) 75–96.
5 Noble, P., Boon, J.J., and Wadum, J., ‘Dissolution, aggregation
and protrusion. Lead soap formation in 17th century grounds
and paint layers’, ArtMatters 1 (2003) 46–61.
6 Howard, H., Pigments of English Medieval Wall Painting, Archetype
Publications, London (2003) 138–140.
7 Noble, P., and Boon, J.J., ‘Metal soap degradation of oil
paintings: aggregates, increased transparency and efflorescence’,
in AIC Paintings Working Group Meeting 2006: Postprints, ed. H.
Mar Parkin, Thomson-Shore, Dexter (2007) Vol. 19, 1–15.
8 van der Weerd, J., Gelddof, M., Struik van der Loeff, L., Heeren,
R., and Boon, J., ‘Zinc soap aggregate formation in Falling
leaves (Les Alyscamps) by Vincent van Gogh’, Zeitschrift für
Kunsttechnologie und Konservierung 17 (2003) 407–416.
9 Keune, K., Binding Medium, Pigments and Metal Soaps Characterised
and Localised in Paint Cross-sections, PhD thesis, MOLART
Report 11, AMOLF, Amsterdam, University of Amsterdam
(2005) 144–150, www.amolf.nl/publications/theses/ (accessed
14 July 2007).
10 Beltinger, K., Englisch, G., Gros, D., Herm, C., and Stoll, A., ‘A
technical study of Ferdinand Hodler’s painting technique: Work
in progress’, in ICOM Committee for Conservation, 13th Triennial
Meeting, Rio de Janeiro, ed. R. Vontobel, James & James, London
(2002) Vol. I 388–393.
11 Osmond, G., Keune, K., and Boon, J.J., ‘A study of zinc soap
aggregates in a late nineteenth century painting by R.G. Rivers
at the Queensland Art Gallery’, AICCM Bulletin 29 (2004–
2005) 37–46.
12 Gunn, M., Chottard, G., Rivière, E., Girerd, J.-J., and Chottard,
J.-C., ‘Chemical reaction between copper pigments and
oleoresinous media’, Studies in Conservation 47 (2002) 12–23.
13 Spring, M., Higgitt, C., and Saunders, D., ‘Investigation of
pigment-medium interaction processes in oil paint containing
degraded smalt’, National Gallery Technical Bulletin 26 (2005)
56–69.
14 van der Weerd, J., Boon, J.J., Geldof, M., Heeren, R.M.A.,
and Noble, P., ‘Chemical changes in old master paintings:
Dissolution, metal soap formation and remineralization
processes in lead pigmented paint layers of 17th century
paintings’, Zeitschrift für Kunsttechnologie und Konservierung 16
(2002) 36–51.
15 Boon, J., van der Weerd, J., Keune, K., Noble, P., and Wadum,
J., ‘Mechanical and chemical changes in old master paintings:
Dissolution, metal soap formation and remineralization
processes in lead pigmented ground/intermediate paint layers
of 17th century paintings’, in ICOM Committee for Conservation,
13th Triennial Meeting, Rio de Janeiro, ed. R. Vontobel, James &
James, London (2002) Vol. I 401–406.
16 Heeren, R.M.A., Boon, J.J., Noble, P., and Wadum, J., ‘Integrating
imaging FTIR and secondary ion mass spectrometry for the
analysis of embedded paint cross-sections’, in ICOM Committee
for Conservation, 12th Triennial Meeting, Lyon, ed. J. Bridgland,
James & James, London (1999) Vol. I 228–233.
17 Keune, K., Noble, P., and Boon, J.J., ‘Chemical changes in
lead-pigmented oil paints: On the early stage of formation of
protrusions’, in Proceedings of ART 2002, the 7th International
Conference on Non-destructive Testing and Microanalysis for the
Diagnostics and Conservation of the Cultural and Environmental
Heritage, ed. R. van Grieken, K. Janssens, L. Van’t dack, and
G. Meersman, University of Antwerp, Antwerp (2002)
unpaginated.
18 Boon, J., van der Weerd, J., Keune, K., and Noble, P., ‘Chemical
changes in old master paintings: Dissolution, metal soap
formation and remineralization processes in lead pigmented
paint layers of 17th century painting’, in Deterioration of Artists’
Paints: Effects and Analysis: Extended Abstracts, UKIC (2001)
19–20.
19 van den Berg, J.D.J., Analytical Chemical Studies on Traditional
Linseed Oil Paints, PhD thesis, MOLART Report 6, AMOLF,
Amsterdam, University of Amsterdam (2002) 195–221, www.
amolf.nl/publications/theses/ (accessed 14 July 2007).
20 van der Weerd, J., Microspectroscopic Analysis of Traditional Oil
Paint, PhD thesis, MOLART Report 7, AMOLF, Amsterdam,
University of Amsterdam (2002) 123–145, www.amolf.nl/
publications/theses/ (accessed 14 July 2007).
21 Plater, M.J., De Silva, B., Gelbrich, T., Hursthouse, M.B.,
Higgitt, C.L., and Saunders, D.R., ‘The characterisation of lead
fatty acid soaps in “protrusions” in aged traditional oil paint’,
Polyhydron 22 (2003) 3171–3179.
22 Boon, J.J., Hoogland, F., and Keune, K., ‘Chemical processes in
aged oil paints affecting metal soap migration and aggregation’,
in AIC Paintings Working Group Meeting 2006: Postprints, ed. H.
Mar Parkin, Thomson-Shore, Dexter (2007) Vol. 19, 16–23.
23 Verhoeven, M.A., Carlyle, L., Reedijk, J., and Haasnoot, J.G.,
‘Exploring the application of solid-state nuclear magnetic
resonance to the study of the deterioration of paintings’, in
Reporting Highlights of the Mayerne Programme, MOLART report
13, ed. J.J. Boon and E.S.B. Ferreira, NWO, The Hague (2006)
33–42.
24 Kühn, H., ‘Lead-tin yellow’, in Artists’ Pigments, Vol. 2, ed. A.
Roy, National Gallery of Art, Washington (1993) 85.
25 Boon, J.J., Gore, E., Keune, K., and Burnstock, A., ‘Image
analytical studies of lead soap aggregates and their relationship
to lead and tin in 15th century lead tin yellow paints from the
Sherborne Triptych’, in Infrared and Raman Users Group (IRUG)
meeting, 29 March–1 April, ed. M. Picollo, Il Prato, Padova (2004)
66–74.
IMAGING STUDIES OF CROSS-SECTIONS OF PAINTINGS AFFECTED BY LEAD SOAP AGGREGATE FORMATION 175
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
26 Eastaugh, N.J., Lead Tin Yellow: Its History, Manufacture, Colour
and Structure, unpublished PhD thesis, University of London,
Courtauld Institute of Art (1988).
27 Zucker, J., ‘From the ground up: The ground in 19th-
century American pictures’, Journal of the American Institute of
Conservation 38 (1999) 3–20.
28 Keune, K., and Boon, J.J., ‘Enhancement of the static-SIMS
secondary ion yields of lipid moieties by ultrathin gold coating
of aged oil paint surfaces’, Surface and Interface Analysis 36 (2004)
1620–1628.
29 Ferreira, E., unpublished results, FOM/AMOLF, The Nether-
lands (2005).
30 Zucker, J., and Boon, J.J., ‘Opaque to transparent: Paint film
defects in the work of Frederic Church and the Hudson
River School’, in AIC Paintings Working Group Meeting 2006:
Postprints, ed. H. Mar Parkin, Thomson-Shore, Dexter (2007)
Vol. 19, 33–41.
31 van Loon, A., and Boon, J.J., ‘The whitening of oil paint films
containing bone black’, in ICOM Committee for Conservation,
14th Triennial Meeting, The Hague, ed. I. Verger, James & James/
Earthscan, London (2005) Vol. I 511–518.
32 Skaliks, A., Blooming – Auswandern von Bindmittelbestandteilen
aus ölhaltigen Farbsystemen, Diplomarbeit, Fachhochschule Köln
(1999) [in German].
33 Speleers, L., private restorer, Amsterdam, personal communi-
cation (2005).
34 Saunders, D., Spring, M., and Higgitt, C., ‘Colour change
in red lead containing paint films’, in ICOM Committee
for Conservation, 13th Triennial Meeting, Rio de Janeiro, ed. R.
Vontobel, James & James, London (2002) Vol. I 455–463.
35 Cotte, M., Dumas, P., Richard, G., Breniaux, R., and Walter, P.,
‘New insight on ancient cosmetic preparation by synchrotron-
based infrared microscopy’, Analytical Chimica Acta 553 (2005)
105–110.
36 Carlyle, L. (ICN, The Netherlands), and Keune, K. (FOM/
AMOLF, The Netherlands), unpublished SIMS studies on
Dutch processed lead white.
AUTHORS
KATRIEN KEUNE graduated in chemistry from the
University of Amsterdam, The Netherlands (2000). She
subsequently joined the De Mayerne project team in
the Molecular Paintings Research group of Professor J.J.
Boon. She explored various analytical non-destructive
imaging techniques to obtain detailed molecular
information about paintings. Part of this work was
submitted as a PhD thesis entitled ‘Binding medium,
pigments and metal soaps characterised and localised in
paint cross-sections’ to the University of Amsterdam in
2005. Keune is a postgraduate research fellow (NWO
Talent-fellowship) at the Scientific Department of the
National Gallery in London (2006–2007). Address:
Kortenaerstraat 82, 2315 TP, Leiden, The Netherlands. Email:
katrien.keune@gmail.com
JAAP J. BOON was trained in geology and chemistry at the
Universities of Amsterdam, Utrecht and Delft Technical
University (1970–1974). After postdoctoral studies in
biology and mass spectrometry, he became research
associate at FOM/AMOLF in 1983. Boon masterminded
and coordinated the NWO priority project MOLART
(Molecular aspects of ageing in art) (1995–2002).
He was involved in seven new interdisplinary studies
developed in the framework of the NWO De Mayerne
programme (2002–2006). Boon is professor of analytical
mass spectrometry at the University of Amsterdam and
head of the Molecular Material Science of Art Group of
MOLART at AMOLF. Address: FOM Institute for Atomic
and Molecular Physics, Kruislaan 407, 1098 SJ, Amsterdam,
The Netherlands. Email: boon@amolf.nl
Résumé — Des agrégats de savon de plomb ont été trouvés dans des couches de peinture à l’huile contenant du plomb de
tableaux datant du XIIIe au XXe siècle. Ils affectent sérieusement la stabilité des couches picturales et perturbent la surface des
peintures. Des coupes stratigraphiques provenant de cinq peintures affectées par les savons de plomb ont été sélectionnées pour
illustrer et étudier ce phénomène de dégradation au moyen de la spectroscopie infrarouge à transformée de Fourier, la spectrométrie
de masse des ions secondaires et la microscopie électronique à balayage couplée à la spectrométrie des rayons X. On donne des
exemples de savons de plomb se formant : au sein d’un système de peinture mature ou, au contraire, lors des premières phases de
séchage de l’huile; à partir de différents pigments ou siccatifs à base de plomb; et dans des couches multiples. On traite aussi de
la cristallisation de produits contenant du plomb à l’intérieur des agrégats. Le phénomène de la formation d’agrégats de savons
de plomb a plusieurs facettes et un seul scénario décrivant la formation de ces agrégats ne peut en expliquer tous les aspects.
Cependant, la détermination des composés chimiques et leur distribution parmi les couches picturales conduit à proposer un schéma
de réaction dans lequel les acides gras monocarboxyliques libres jouent un rôle clé. La disponibilité et la libération de ces acides gras
dépend de la composition de la peinture originale, de la superposition des couches et des conditions environnementales auxquelles
les peintures ont été exposées au cours des ans.
176 K. KEUNE AND J. J. BOON
STUDIES IN CONSERVATION 52 (2007) PAGES 161–176
Zusammenfassung — Aggregate von Bleiseifen wurden in bleihaltigen Malschichten in Gemälden des 13. bis 20. Jahrhunderts
gefunden. Diese beeinflussen die Stabilität und die Oberflächenerscheinung der Malschichten gravierend. Malschichtquerschliffe
von fünf von der Bildung von Bleiseifen betroffenen Gemälden wurden zur Illustration und zur Analyse dieses Phänomens
ausgewählt und mit Hilfe von Fouriertransforminfrarotspektroskopie, Sekundärionenmassenspektrometrie und mit
Röntgenmikroanalyse kombinierter Rasterelektronenmikroskopie untersucht. Es werden Beispiele genannt, bei denen sich die
Metallseife entweder in gealtertem Malschichten gebildet hat oder in einem frühen Stadium des Trocknungsprozesses des Öls.
Die Bleiseifen bilden sich aus verschiedenen Typen bleihaltiger Pigmente und Trocknungsbeschleuniger. Bleiseifen bilden sich in
vielschichtigen Malschichten. Im Inneren der Aggregate befinden sich Kristallisationskeime. Das Phänomen ist vielschichtig und
ein einheitliches Szenario der Bildung der Bleiseifen kann nicht alle Aspekte beschrieben. Indessen kann man aus der Einbindung
aller chemischer Informationen und der Verteilung in den Malschichten annehmen, daß freie Monocarboxylfettsäuren eine
bedeutende Rolle bei der Bildung der Seifenaggregate spielen. Die Verfügbarkeit und die Freisetzung dieser Fettsäuren hängen
von der Zusammensetzung der originalen Malschicht, der Architektur der Schichten und der Konservierung / Geschichte der
Umgebungsbedingungen ab.
Resumen — Agregados de plomo saponificado han sido identificados en capas pictóricas al óleo que contienen pigmentos de
plomo desde el siglo XIII al XX. Estos afectan severamente a la estabilidad de las capas de pintura y estropean las superficies
de los cuadros. Estratigrafías tomadas de cinco obras afectadas por jabones de plomo se seleccionaron para ilustrar e investigar
este fenómeno de degradación con las siguientes técnicas analíticas: infrarroja por transformada de Fourier, espectrometría iónica
de masas secundaria y microscopía electrónica de barrido combinada a análisis por rayos X. Se aportan ejemplos de jabones
de plomo formados en sistemas ya curados de pintura o, alternativamente, en fases iniciales de secado del aceite; en jabones
de plomo formados a partir de diversos tipos de pigmentos de plomo o secativos; en jabones formados en capas múltiples de
pintura; y en productos de cristalización de contenido en plomo en el interior de los agregados. El fenómeno de los agregados
de plomo saponificado tiene múltiples facetas y un escenario que describa estos agregados ciertamente no puede explicar todos
los aspectos. Sin embargo, la integración de la información química y su distribución entre las capas pictóricas lleva a proponer
que los ácidos grasos monocarboxílicos reactivos juegan un papel clave en la formación de agregados de plomo saponificado. La
disponibilidad y neutralización de estos ácidos grasos depende de la composición de la pintura original, de la estructuración de las
capas y de la exposición ambiental/conservación de la pintura en el pasado.
