Studies in Conservation

ISSN: 0039-3630 (Print) 2047-0584 (Online) Journal homepage: http://www.tandfonline.com/loi/ysic20

Nineteenth century chrome yellow and chrome

TM
deep from Winsor & Newton

Vanessa Otero, Joana V. Pinto, Leslie Carlyle, Márcia Vilarigues, Marine Cotte
& Maria João Melo

To cite this article: Vanessa Otero, Joana V. Pinto, Leslie Carlyle, Márcia Vilarigues, Marine
Cotte & Maria João Melo (2016): Nineteenth century chrome yellow and chrome deep from
TM
Winsor & Newton , Studies in Conservation, DOI: 10.1080/00393630.2015.1131478

To link to this article: http://dx.doi.org/10.1080/00393630.2015.1131478

Published online: 07 Apr 2016.

Submit your article to this journal

Article views: 3

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ysic20

Download by: [McMaster University] Date: 10 April 2016, At: 15:26

 Original research or treatment paper
Nineteenth century chrome yellow and
chrome deep from Winsor & NewtonTM
Vanessa Otero 1, Joana V. Pinto 2, Leslie Carlyle 1, Márcia Vilarigues 3, Marine
Cotte4,5 and Maria João Melo1
1
Department of Conservation and Restoration and LAQV-REQUIMTE, Faculty of Sciences and Technology,
Universidade NOVA de Lisboa, Monte da Caparica, Portugal, 2CENIMAT/I3N, Material Science Department,
Faculty of Sciences and Technology, Universidade NOVA de Lisboa, Monte da Caparica, Portugal, 3Department
of Conservation and Restoration and VICARTE, Faculty of Sciences and Technology, Universidade NOVA de
Lisboa, Monte da Caparica, Portugal, 4European Synchrotron Radiation Facility, Grenoble, France, 5Laboratoire
d’Archéologie Moléculaire et Structurale, Paris, France
Downloaded by [McMaster University] at 15:26 10 April 2016

The Winsor & NewtonTM (W&N) nineteenth century archive database includes digitised images of hand-
written instructions and workshop notes for the manufacture of their artists’ materials. For the first time,
all 183 production records for yellow lead chromate pigments were studied and evaluated. They
revealed that W&N produced essentially three pigment types: lemon/pale based on mixed crystals of
lead chromate and lead sulphate [Pb(Cr,S)O4]; middle on pure monoclinic lead chromate [PbCrO4]; and
deep that contains the latter admixed with basic lead chromate [Pb2CrO5]; accounting for 53, 22, and
21% of the production, respectively. Production records for primrose (4%) were also included since the
formulation results in mixed crystals with a high percentage of lead sulphate, which, according to the
literature, leaves it more prone to degradation. Each pigment type is characterised by only one or two
main synthetic pathways; process variations reveal a systematic and thorough search for a high-quality
durable product. A comparison of the chemical composition of pigment reconstructions with early W&N
oil paint tubes showed that their records entitled ‘pale’ and ‘lemon’ correlated with the pigment in their
tube labelled chrome yellow and, ‘middle’ and ‘deep’ with the label chrome deep. Lemon and middle
pigment formulations were made into oil paints to assess their relative photo-stability. The degradation
process was followed by colorimetry and was studied by synchrotron radiation-based techniques. Based
on the X-ray absorption spectroscopy data, the possibility for creating a stability index for chrome
yellows is discussed.
Keywords: Chrome yellow, Lead chromate, Reconstructions, Winsor & NewtonTM, μ-Fourier-transform infrared spectroscopy (μ-FTIR), Synchrotron
radiation (SR)-based techniques, X-ray absorption near-edge microspectroscopy (μ-XANES)

Introduction lead chromate pigments suffered similar degradation as

Winsor & Newton’s chrome yellow and chrome reported elsewhere (see ‘Degradation studies’ section),
deep and to compare reconstructions based on their formu-
Winsor & NewtonTM (W&N) sold yellow lead chromate lations with samples from early W&N tube paints,
paints under the trade names chrome yellow and chrome labelled chrome yellow and chrome deep.
deep listed from their earliest catalogue c. 18351 (Carlyle, Lead chromate was first discovered and synthesised
2001). Since their foundation in 1832 W&N traded on by L. N. Vauquelin at the beginning of the nineteenth
their reputation for high-quality durable products. This century (Vauquelin, 1809; Kϋhn & Curran, 1986).
study was carried out to evaluate whether their yellow The French chemist demonstrated that process modifi-
cations would influence the colour, documenting that
Correspondence to: Maria João Melo, Department of Conservation and different hues resulted according to the pH during man-
Restoration and LAQV-REQUIMTE, Faculty of Sciences and Technology, ufacture: an ‘orange yellow’ resulted in neutral con-
Universidade NOVA de Lisboa, 2829-516 Monte da Caparica, Portugal.
Email: [email protected] ditions ( pure lead chromate, PbCrO4), ‘a yellowish
1
W&N introduced chrome lemon in the latter part of the century, appearing red or sometimes a beautiful deep red’ in alkaline con-
for the first time in the W&N catalogue of 1886. In their catalogue of 1896, ditions (basic lead chromate, Pb2CrO5) and a ‘deep
chrome lemon is described as a combination of chromate and sulphate of
lead. lemon yellow’ in acidic solutions (mixed crystals of

© The International Institute for Conservation of Historic and Artistic Works 2016
Received June 2015; revised paper accepted September 2015
DOI 10.1080/00393630.2015.1131478 Studies in Conservation 2016 1
 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Figure 1 Acid–base equilibrium established in the range 1 < pH < 11 between chromate ion (CrO2−) and dichromate ion (Cr2 O2−
7 )
(Shriver & Atkins, 1999).

lead chromate and lead sulphate, Pb(Cr,S)O4)2 1846 to 1858 (Otero et al., 2012). The pigments
(Vauquelin, 1809; Harley, 2001, Burnstock et al., 2003). obtained were compared with early tube paints from
For their lead chromate, W&N maintained W&N as well as with samples from paintings by the
Vauquelin’s essential steps: a solution of a lead salt Portuguese modernist painter, Amadeo de Souza-
(nitrate or acetate) added to a dichromate solution Cardoso (1887–1918). Good correlation between the
(Cr2 O2−
7 ). As shown in Fig. 1, in solution the reconstructions, the tube paints, and the paint
Cr6+ species are in an acid–base equilibrium, which samples was found which supported the hypothesis
is shifted towards the chromate anion by adding that reconstructions from W&N’s production records
Pb2+, leading to the formation of insoluble chromates could serve as reference samples.
(Cotton & Geoffrey, 1988; Shriver & Atkins, 1999). In the present work, we broadened our study to
Downloaded by [McMaster University] at 15:26 10 April 2016

include all 183 entries in the W&N archive database

Degradation studies referring to yellow lead chromate pigments, aiming
Recently, by means of synchrotron radiation (SR) to identify their experiments, the rationale for choos-
based techniques and using artificially aged samples, ing certain formulations, as well as the pigment formu-
Monico et al. have studied the darkening of chrome lations which were specifically sold under the trade
yellow (observed in the paintings of Vincent van names chrome yellow and chrome deep. As will be
Gogh and his contemporaries) (Monico et al., 2011, shown, this enabled us to select the pigment formu-
2013; Tan et al., 2013). They established that degra- lations expected to most likely undergo degradation.
dation results from the reduction of the original oxi- In our study, we considered only yellow lead chromate
dation state Cr6+ to Cr3+ with the formation of production (which represents 61% of all of the yellow
Cr3+ oxo compounds such as the green Cr2O3 and chromate production records). The records for chro-
Cr2O3·2H2O (Monico et al., 2011). They also demon- mates based on zinc, barium, and strontium were not
strated that solid solutions of lead chromate and lead included nor were chrome orange and chrome red
sulphate, richer in lead sulphate, are more prone to entries. These will be discussed in future work.
degradation, especially under UV-VIS (λ ≥ 300 nm), Oil paints prepared with the reconstructed pigments
UV (240 ≤ λ ≤ 400 nm) or blue light (335 ≤ λ ≤ and linseed oil were air-dried and then irradiated in a
525 nm) (Monico et al., 2013b). Xenon-light apparatus (λirr > 320 nm). The degra-
dation process was studied by SR-based techniques,
W&N’s chrome yellow pigment formulations
including X-ray absorption near-edge microspectro-
The yellow lead chromate pigment reconstructions in
scopy (μ-XANES), X-ray fluorescence microspectro-
this study were based on production records found in
metry (μ-XRF), and Fourier transform infrared
‘The W&N archive database (Researcher Edition)’
microspectroscopy mapping (μ-FTIR). Evidence that
which provides efficient access to their nineteenth
contributes to a prediction of the sensitivity of
century artists’ materials and their commercial prep-
chrome yellow paints to photo degradation will be
aration (Clarke & Carlyle, 2005; Carlyle et al., 2011;
discussed.
Otero et al., 2012). Access to digitised images of their
extensive chrome yellow pigment formulations
prompted our research on the stability of their pigments. Results and discussion
In a previous publication, seeking to ascertain the W&N’s production records for yellow lead chromates
role of the various ingredients, our reconstructions of include the following names: primrose chrome, lemon
their chrome yellows were based on a selection of chrome, pale chrome, middle chrome, and deep
four of W&N’s production records3 dating from chrome.4 These names are further modified with adjec-
tives including, best, genuine, fine, super, pure,
2
Pigments commonly named chrome yellow are composed of pure lead no. 1. Despite this variety, their retail catalogues
chromate (PbCrO4) or mixed-crystals of lead chromate and lead sulphate
(Pb(Cr,S)O4), while those called chrome orange are composed of basic from the nineteenth century listed only three chrome
lead chromate (Pb2CrO5) (Eastaugh et al., 2004). yellows: chrome yellow, chrome deep, and chrome
3
The four production records enabled the production of ‘middle’ type com-
pounds. They were named P1–P4: two are single entries (L3a and P5a, see
4
Appendix 1) and the two other correspond to M1a and M2a (please see There are 20 records that present different names, which were included as
next section and Table 1 for details on the nomenclature). process variations in Appendix 1.

2 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Figure 2 The final colour is modulated by pH during manufacture. On the left: the lemony mixed crystals of lead chromate and
lead sulphate that precipitate at low pH; centre: the yellow pure lead chromate monoclinic phase (PbCrO4) at neutral pH; on the
right: an orange/red basic lead chromate (Pb2CrO5) formed in basic media.

lemon. No explanation for this is offered here, but it is of the chromate solution was controlled from the
notable that their archive recipes for oils encompassed beginning by the quantity of Na2CO3 added, which
the same level of detail and variety which was also acts as a buffer (Otero et al., 2012). The chromate sol-
Downloaded by [McMaster University] at 15:26 10 April 2016

omitted in the trade catalogues (Carlyle et al., 2011). ution is the only common ingredient to all their man-
Our systematic evaluation of all 183 production ufacturing processes (see Table 1).
records for yellow lead chromates showed that there Lead nitrate (Pb(NO3)2) and lead subacetate
were in fact a limited number of variations for synthe- (Pb(Ac)2·2Pb(OH)2) were prepared by adding
sising the pigments (Table 1). Assuming that the lead(II) oxide (PbO) to an aqueous solution of nitric
higher the number of entries the more frequently the acid (HNO3) or lead acetate (Pb(CH3CO2)2·3H2O),
process was carried out, we calculated that the main respectively. As will be discussed in greater detail
synthetic pathways given in Table 1 were the most sig- below both lead sources were used to produce
nificant in W&N’s commercial production. The full lemon/pale and middle, but for the deep which
range of their process variations are represented in initially requires a higher pH, only lead subacetate
Appendix 1 along with details on the ingredients, the was used, Scheme 1.
order of addition, as well as the respective number of Sulphate and lead sources as pairs. As lemon, pale,
production records for each manufacturing process. and middle were precipitated at neutral to low pH
Table 1 and the tables in the appendix were designed values, the pairs lead nitrate/sodium sulphate
to highlight the pattern of experimentation made by (Na2SO4) and lead subacetate/sulphuric acid
W&N. The main chemical processes involved are rep- (H2SO4) were the most frequently found. The paired
resented in equations (1)–(4). Pb(NO3)2/Na2SO4 induces the mixed crystals for-
Table 1 shows that, basically, for producing the W&N mation in low pH conditions5 and was most used for
yellow chromates, a solution of lead nitrate or subace- the lemon/pale. Likewise, the paired
tate: [Pb(NO3)2 or Pb(Ac)2·2Pb(OH)2)] was added to a Pb(Ac)2·2Pb(OH)2/H2SO4, at a neutral pH, was
chromate solution (CrO42−). The yellow lead chromates used for the formation of the deep. For the very low
as produced, can be described chemically as: (i) mixed pH’s observed in some primrose formulations, the
crystals of lead chromate and lead sulphate (light more unusual pairing of Pb(NO3)2/H2SO4 was found.
lemony yellows) and (ii) pure lead chromate monoclinic The crystal structure modifiers, which transform the
phase (yellow) as seen in Fig. 2 and Scheme 1. A mixed- pure lead chromate monoclinic phase into a solid sol-
crystal is a solid solution, i.e., a crystalline structure in ution of lead chromate and lead sulphate
which one or more atoms are substituted by others that Pb(Cr,S)O4 (mixed crystals) were Na2SO4 and/or
can assume the same geometry without changing the H2SO4, but these are only effective at low pH. To
structure. The term mixed-phase pigment is also com- obtain lighter hues, a high amount of lead sulphate
monly used (Buxbaum, 1998; Erkens et al., 2001). In in the crystal structure, i.e. a high proportion of sul-
the deep hues, basic lead chromate, known as phoenico- phate over chromate in the mixed crystal is required.
chroite (orange/red) is also found. This was confirmed in the W&N formulations.
Other materials were added in solution. The main
additives, which may have acted as extenders, were
The main ingredients and their roles in W&N’s plaster of Paris (CaSO4·12H2O), which would be
formulations
The chromate solution was obtained by adding sodium 5
In L1, the lead sulphate pulp must have a low pH for the successful for-
carbonate (Na2CO3) to a dichromate solution. The pH mation of the mixed crystals.

Studies in Conservation 2016 3

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Table 1 W&N chrome yellow key ingredients, synthetic pathways and final pigment formulations
Downloaded by [McMaster University] at 15:26 10 April 2016

§
Codes refer to variations in the manufacturing processes, see Appendix 1, Tables A1–A5.
‡
Solubility product of PbCrO4, Ksp = 1.8 × 10−14; of PbCO3, Ksp = 7.4 × 10−14; of PbSO4, Ksp = 2.53 × 10−8; BaSO4, Ksp = 1.08 ×
10−10, of CaCO3, Ksp = 3.36 × 10−9 and of CaSO4·2H2O, Ksp = 3.14 × 10−5 (Chang, 1994; Lide, 2006).
#
L1 and M1a both produce PbCO3 as by-product.
ǂ
Acid is provided by the lead sulphate pulp which has an acidic pH.
₸
In M2a, the addition of lead subacetate may form a small amount of Pb2CrO5.
¥
In D2a, CaSO4·2H2O may disappear with the addition of sulphuric acid.

Scheme 1 The main ingredients for the production of (i) mixed crystals of lead chromate and lead sulphate and (ii) pure
monoclinic lead chromate crystals. The type of lead source greatly influences the final pH (and colour): lead subacetate
(Pb(Ac)2·2Pb(OH)2) is usually employed for neutral/close to neutral pH and lead nitrate (Pb(NO3)2) when acidic solutions are
required. Additives are not represented

4 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

previously converted into gypsum (CaSO4·2H2O), Table 1 and equation (1).8 As noted above, primrose
barytes (BaSO4), lead sulphate (PbSO4), and chalk was included in the study due to its higher sulphate
(CaCO3). Besides these, experiments with China clay content and reports that this would render it more
(Al2Si2O5(OH)4), terra alba (finely powdered prone to degradation (Monico et al., 2013b).
CaSO4·2H2O), alum (AlK(SO4)2·12H2O), magnesium
sulphate (MgSO4), and potassium sulphate (K2SO4) 4 + H2 SO4 + Pb(Ac)2 · 2Pb(OH)2
CrO2−
were also carried out. 1<pH<3
 Pb(Cr, S)O4 (1)

W&N’s experiments with chrome yellow: the This pigment class is characterised by a high amount
manufacturing processes of lead sulphate in the crystal structure as a conse-
W&N’s books reveal thorough and informed testing of quence of the addition of a higher quantity of sulphu-
the ingredients (sources for chromate, sulphate, and ric acid when compared to lemon/pale chrome. This
lead as well as additives) and their order of addition. led W&N to choose lead subacetate as the lead
Once the main ingredients (chromate, lead ion, and source (likely because, as was found during synthesis,
crystal structure modifiers) had been established it is much easier to control the precipitation pH).
(which was to pair the chromate with either This pair, Pb(Ac)2·2Pb(OH)2)/H2SO4, produces the
Pb(NO3)2/Na2SO4 or Pb(Ac)2·2Pb(OH)2)/H2SO4 as low pH (<3.5) necessary to obtain, with high yield,
explained above), their next most significant tests the lemony mixed crystals displaying a high sulphate
Downloaded by [McMaster University] at 15:26 10 April 2016

centred primarily on the additives: gypsum, barytes, to chromate ratio. Interestingly, no extenders were
lead sulphate, and calcium carbonate. All possible used, however some of W&N’s records indicate that
combinations were tested, but only the ones that MgSO4 and K2SO4 were experimented with as
afforded a good quality product were selected by additional sulphate sources.
W&N (surmised from the frequency of their appear-
ance in the production records). The diversity that Lemon and pale production records
we find in the fillers may reflect either a search for a In the books there are 46 production entries corre-
better product, a better process (high quality and sponding to lemon and 51 for pale chrome, accounting
lower price), or perhaps the ready availability of the for 53% of all yellow lead chromate production
raw materials. records. These can be reduced to four main synthesis
The role of the various additives was detailed in a methods for lemon and 11 for pale chrome.
previous publication (Otero et al., 2012). For this Variations of these main processes are depicted in
paper, it is important to note our finding that additives Appendix 1. Most of the production records cover
were not only important during the synthesis (enabling the period from 1846 to 1893.
better pH control), but were also selected by W&N to Since the reaction scheme and final products are
provide pigment stability (see below). Several com- very similar for both lemon and pale entries, we
ments may be found in the books relating to stability, grouped them into a single class (it is unclear why
for example in one process to produce chrome deep these records were given separate names). Five of the
(D1b, see Table 1) was the note: ‘Kept its colour in synthesis methods, including the ones with the most
oil, on palette, very well when examined Sept 1870’.6 entries, are common to lemon and pale. The main syn-
For the processes appearing most frequently, W&N thetic pathways are L1 and L2a, Table 1 and equation
was relying on barium sulphate as an extender which (2). L2a is only found in the lemon records.
we anticipate will increase stability in the case of the
mixed crystals, as discussed below. Only for the pure 4 + Na2 SO4 + Pb(NO3 )2
CrO2−
monoclinic lead chromate (middle chrome) did we 1<pH<6
 Pb(Cr, S)O4 (2)
find a preference for calcium carbonate (to shift the
pH to neutral values allowing the formation of lead
This process is characterised by the pair Pb(NO3)2/
chromate, Scheme 1).
Na2SO4. All manufacturing processes are based on
adding the sulphate ions to the chromate solution
Primrose: ‘Light in weight, rich & bright lemon tint’7
before the addition of the lead source, resulting in
There were only eight production records for primrose the formation of mixed crystals at low to neutral pH
chrome, which can be reduced to three synthesis (1 ≤ pH ≤ 6).
methods (Appendix 1). These cover the period from These pigments form a paler yellow when compared
1846 to 1880. The majority of these records date to with pure lead chromates (Fig. 2). In L1, the low pH is
the 1850s, and pertain to the process coded as PR1a,
8
Equations 1–4 are given to help the reader to interpret the role played by
6
One and a half years after the pigment manufacture on the 26/02/1869; the main ingredients and by pH. These chemical reactions are not pre-
W&N book P3, page 054A, line 01. sented as balanced chemical equations, and only the sources for chro-
7
W&N book P1, page 235A, line 04. mate, sulphate, and lead are depicted.

Studies in Conservation 2016 5

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Table 2 Lab* coordinates# and composition of pigments representative of the main synthetic pathways

PR1a L1 L2a M1a** M2a** D1b D2a

Final pH§ ≈3.5 ≈6 ≈1 ≈6.5 ≈7

L* 90.5 ± 0.3 86.7 ± 0.6 87.1 ± 0.5 84.8 ± 0.3 76.1 ± 0.7 73.7 ± 0.8 74.4 ± 1.0
a* 1.0 ± 0.5 12.5 ± 1.9 9.3 ± 0.3 19.8 ± 0.2 28.5 ± 1.1 31.3 ± 1.4 31.4 ± 0.4
b* 80.2 ± 1.5 89.7 ± 3.6 84.4 ± 1.1 81.1 ± 2.5 69.8 ± 2.7 67.0 ± 2.4 69.8 ± 2.3
Composition¥ Pb(Cr,S)O4 Pb(Cr,S)O4 Pb(Cr,S)O4 PbCrO4 PbCrO4 PbCrO4 PbCrO4
PbCO3 BaSO4 CaCO3 BaSO4 BaSO4 BaSO4
PbSO4 PbCO3 Pb2CrO5 Pb2CrO5 CaSO4·2H2O
Pb2CrO5

The Lab* coordinates presented are the mean value for two final products of PR1a, L2a, and D2a and three final products for L1,
M1a, M2a, and D1b.
§
Final pH was measured at the end of the precipitation.
**In our previous publication (Otero et al., 2012), M1a and M2a correspond to P1 and P3, respectively.
¥
Underline indicates the element present in higher proportion in the mixed crystal.

produced with the lead sulphate pulp and, in L2a, by Middle and deep production records
sulphuric acid (Table 1). In L1, sodium carbonate is Middle and deep chromes are found in 40 and 38
Downloaded by [McMaster University] at 15:26 10 April 2016

used as a buffer, and lead carbonate (Ksp = 7.4 × entries, respectively, representing 43% of all the
10−14) forms as by-product, since its solubility yellow lead chromate production records. These can
product is very similar to lead chromate (Ksp = 1.8 × be reduced to eight synthetic methods for middle
10−14) (Chang, 1994; Lide, 2006). The resulting and ten for deep. All variations are depicted in
mixed crystals present different proportions of lead Appendix 1. Most of the production records are
sulphate but the monoclinic structure of lead chromate found in books ranging from 1834 to 1882, with the
is preserved (Fig. 4). W&N described the synthesised earliest record with a clear date being December 1852.
L2a as displaying a ‘fine colour’ and in their notes
report that it was ‘sent to Raynolds Co, N.Y.’9; L1 is pH≈7
4 + Pb(NO3 )2  PbCrO4
CrO2− (3)
described as a ‘Bright lemon tint, (…): light in
weight’.10 4 + H2 SO4 +Pb(Ac)2 · 2Pb(OH)2
CrO2−
pH≈7
Barium sulphate and lead sulphate were added in  PbCrO4 + Pb2 CrO5 (4)
the two main processes (Table 1), acting as extenders
(had carbonates been used, they would have acted as In general, all of the middle and deep manufacturing
buffers not allowing the lower pH values necessary processes produce lead chromate at a neutral pH
for the formation of the mixed crystals). The addition (equations (3) and (4)). Middle can be obtained both
of these sulphate compounds also shifts the equili- with either lead nitrate or lead subacetate (equations
brium reaction towards the mixed crystals formation. (3) and (4)). A good example is the manufacturing
It is also important to note the process variations process M1a (Table 1) where calcium carbonate is
L2b and L3b 11 where no extenders are added (see mixed with the chromate solution followed by the
Appendix 1). As in L1 we see that in L3b lead carbon- addition of lead nitrate. In this case, lead carbonate
ate forms during synthesis, showing a spectral finger- also forms as by-product. In W&N’s own words,
print very similar to L1. There are a significant ‘When dry was very bright and rich in tint, not very
number of production records for these two variations: heavy, rather close fracture for middle. A good
12% for L3b and 5% for L2b. process’.12 The other main process for middle, M2a
In 16% of the manufacturing processes lead subace- (Table 1), is described as ‘Very good rich color in Oil
tate is used with sulphuric acid (as found in the pro- & Kept a bright full tint in Oil’.13 In this case, lead sub-
duction of primrose, see Appendix 1) which also acetate was used and in order to lower the pH, sulphu-
results in mixed crystals with a high amount of lead ric acid was added prior to precipitation, shifting the
sulphate in the crystal structure. The resulting pig- system towards the pure lead chromate formation
ments are very bright with a lemony hue. (although a small amount of basic lead chromate
may also be formed). The process M2a is very
similar to deep (D1b), the only difference being the
order of addition of the ingredients. For the deep
9
W&N books A5 and 4P, pages 028A and 102A, line 01, respectively. (equation (4)), lead subacetate is added to the
10
W&N book X6, page 240, line 11.
11
The manufacturing processes L3a and L3b for Lemon are the same as
12
P3a and P3b for Pale, respectively. The percentage values given corre- W&N book P2, page 386A, line 01.
13
spond to the sum of their production records. W&N P1, page 192A, line 01.

6 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Figure 3 Infrared spectra of the pigment reconstructions for middle (M1a) and deep (D2a); (•) CaCO3, ( ) PbCO3, ( ) BaSO4, (♦)
CaSO4·2H2O.
Downloaded by [McMaster University] at 15:26 10 April 2016

chromate solution and precipitation occurs at a basic production records: four for pale and six for middle. In
pH prompting the formation of basic lead chromate; these experiments, the formation of the less stable
sulphuric acid is added afterwards in order to lower orthorhombic form of lead chromate (Gray, 1988;
the pH to neutral. This control of the pH enabled Crane et al., 2001) is more likely with the use of lead
W&N to switch from a yellow coloured pure lead chro- acetate ( pH ≈ 6). They also experimented with lead
mate to an orange/red basic lead chromate (Fig. 2). white in five records for deep and in two of these
Individual experiments with deep were conducted hydrochloric acid had been previously added. In
by mixing lead nitrate with the chromate solution 1846, they also tested the use of lead chloride for pale.
(methods D3 and D4, Appendix 1) and with dichro- We found China clay with terra alba in one pro-
mate without the addition of the sodium carbonate duction record for pale and another for deep, while
buffer (D6, D9, and D10a, Appendix 1). In the case as noted above magnesium sulphate was used in a
of methods D3 and D9, the addition of Plymouth manufacturing process for primrose and, potassium
lime (interpreted as CaO) replaces the carbonate, sulphate in another two records for primrose and for
introducing the alkalinity necessary to produce basic deep.
lead chromate.
For the extender, depending on the pH desired, Pigment characterisation
either calcium carbonate or barium sulphate was Colour coordinates and the composition of the recon-
used. It is interesting to note that in the main process structed pigments obtained from the processes
D2a, gypsum was added together with barium sul- depicted in Table 1 are presented in Table 2 (see also
phate, both acting as extenders. The colour obtained Appendices 2 and 3). The a* coordinate clearly
is described as ‘nice close & good fracture & texture shows the increase in red proceeding from primrose
& light in weight – tint good’.14 The addition of (a* ≈ 0) to pale/lemon (a* = 10–15) through to
gypsum with barium sulphate also appears in D1, middle/deep (a* = 20–30), (also evident in Fig. 2).
M3a, and for lemon/pale in the L4 synthesis method. Spectral data of the pigments’ chromophores are
described below and their infrared fingerprints are
shown in Figs. 3 and 4. Identification of additional
Other variations
material, both by-products and deliberate additions,
W&N also performed experiments with other ingredi-
is straightforward by combining infrared,15 Raman
ents which are detailed in Appendix 1. As noted above
they were essentially using a buffered chromate sol- spectroscopy, and X-ray diffraction. Their spectral
ution with sodium carbonate, but we found potassium 15
The characteristic bands of barium sulphate are: FTIR, asymmetric
−1
carbonate substituting the sodium carbonate in 11 pro- stretching of SO2− 4 at 1186, 1120, and 1085 cm , symmetric stretching
−1
of SO2−4 at 984 cm ; asymmetric bending of SO2−
4 at 637 and
duction records for middle and deep and, in four 611 cm−1, Raman symmetric stretching of SO2− 4 at 987 cm−1, bending
2− −1
of SO4 at 647, 616 cm ; for gypsum are: FTIR, stretching of OH at
others no buffer was used and dichromate appears 3550 and 3405 cm−1, bending of H2O at 1684 and 1620 cm−1, asymmetric
−1 −1
alone. In two records, alum was also experimented stretching of SO2− 2−
4 at 1140 and 1122 cm , bending of SO4 at 670 cm ,
−1
Raman, symmetric stretching of SO2− 4 at 1007 and 1132 cm ; for chalk
with as a buffer for middle together with calcium car- are: asymmetric stretching of CO2− −1
3 at 1408 cm , asymmetric bending
−1
bonate. Lead acetate was tested as a lead source in 10 of CO2− 3 at 875 and 713 cm , Raman, symmetric stretching at
1085 cm−1; for lead carbonate are: FTIR, asymmetric stretching at 1400
−1 −1
and 1051 cm , asymmetric bending at 678 cm , Raman, symmetric
14
W&N book P3, page 106A, line 01. stretching at 1052 cm−1 (Hummel, 2002).

Studies in Conservation 2016 7

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

of bands at 325, 337, 357, 375, and 399 cm−1, assigned

to the CrO2−
4 bending modes. Only the pure monoclinic
phase of lead chromate was found by XRD, the orthor-
hombic phase was not detected.
As may be seen in Figs. 3 and 5, in M1a pigments
the presence of calcium carbonate causes an
overlap between the CrO2− 4 asymmetric stretching
band and the CO2− 3 asymmetric bending band at
−1
875 cm (Hummel, 2002).
Where present in low amounts, basic lead chromate
is not easily detected by FTIR since its band at
849 cm−1 overlaps with the CrO2− 4 asymmetric stretch-
ing bands of lead chromate (Roncaglia et al., 1985).
However, in larger amounts it is possible to observe
a small shift to lower wavenumbers and the CrO2− 4
asymmetric stretching band becomes sharper. On the
other hand, basic lead chromate has a Raman spec-
trum distinct from lead chromate; it shows three very
strong bands at 825, 837, and 846 cm−1, assigned to
Downloaded by [McMaster University] at 15:26 10 April 2016

CrO2−4 symmetric stretching modes. In the bending

region, the Raman spectra display bands at 322, 341,
355, 379, and 399 cm−1, attributable to the
CrO2−4 -bending modes (Roncaglia et al., 1985). Its
presence is also easily detected by XRD; diffraction
peaks are indexed to the monoclinic phase of basic
lead chromate (whose mineral form is called
phoenicochroite).
The infrared spectrum of the D2a pigment recon-
struction shows a sharper CrO2−
4 asymmetric stretching
band due to the presence of basic lead chromate, con-
firmed by Raman and XRD. As mentioned, the identi-
fication of gypsum and barium sulphate is very
straightforward. As noted above, M2a and D1b
processes are very similar and consequently the
composition of the resulting pigments is close, see
Appendix 3.

Primrose, pale/lemon chrome

As previously described (Otero et al., 2012), mixed
crystals of lead chromate and sulphate can be detected
Figure 4 Infrared spectra of the pigment reconstructions for by FTIR (Fig. 4) and Raman mainly due to the SO2− 4
primrose (PR1a) and lemon/pale (L1 and L2a); ( ) PbCO3, ( )
bands, which are shifted from those of pure lead sul-
BaSO4.
phate; in FTIR: asymmetric stretching at 1169 and
1054 cm−1, symmetric stretching at 967 cm−1; asym-
assignments were made in accordance with the litera- metric bending at 630 and 590 cm−1; in Raman: sym-
ture as will be described below (Hummel, 2002; metric stretching at 976 cm−1, bending at 640, 604,
Frost, 2004; Nakamoto, 2009). 448, and 437 cm−1 (Hummel, 2002). Differences are
also found in the CrO2− 4 bands. With XRD, it is
Middle and deep chrome observed that as the content of sulphate ions is
Pure monoclinic lead chromate was detected for higher, there is a progressive shift of the lead chromate
all middle and deep pigments (Fig. 3). Their diffraction pattern to higher values (Crane et al.,
infrared spectra show an intense and broad band at 2001). The detection of the lead sulphate orthorhom-
854 cm−1 (shoulders at 831 and 820 cm−1) assigned to bic phase occurs when the SO2− 4 molar fraction is
the CrO2−
4 asymmetric stretching mode. Their Raman higher than 0.6. After this, the crystal structure
spectra present an intense band at 839 cm−1, assigned begins to change towards the orthorhombic phase of
to the CrO2−
4 symmetric stretching mode, and a series lead sulphate (Crane et al., 2001). In the reconstructed

8 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure 5 Infrared (left) and Raman (right) spectra of the W&N chrome yellow and chrome deep oil paint tubes belonging to
Columbano. In the inset, comparison with the matching pigment reconstructions, L2b and M1, for 1300–650 cm−1. Also shown is
the spectrum from Lefranc’s ‘Jaune de Chrome Foncé’ oil paint tube. ( ) BaSO4, ( ) MgCO3, ( ) zinc soap, (•) CaCO3, ( ) lead
soap, (★) PbCrO4, and ( ) Pb(Cr,S)O4.

pigments, a change in the crystal structure was only crystals obtained have a SO2− 4 molar fraction ≤0.6.
observed for primrose-like pigments (PR1a). For The diffraction patterns were indexed to a pure mono-
lemon/pale pigments, neither phase transition nor clinic phase of lead chromate. Again, the orthorhom-
the lead sulphate phase is detected and the mixed bic phase of lead chromate was never detected.

Studies in Conservation 2016 9

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure 6 Infrared, Raman, and FORS spectra of middle and lemon oil paints before and after 3000h of irradiation with a Xenon
lamp; (•) CaCO3, (♦) CaSO4·2H2O, (★) PbCrO4, and ( ) Pb(Cr,S)O4.

It is important to note that in the presence of lead straightforward due to the SO2−
4 symmetric stretching
and barium sulphate, an overlap of the infrared band and the shifted CrO24 band (Fig. 5).
SO2−
4 asymmetric stretching and bending bands of
the mixed crystals with those of lead and barium sul- Comparison with early oil paint tubes
phate occurs (Hummel, 2002) (Fig. 4). However, the Comparisons were made with two W&N oil paint
detection of the mixed crystals continues to be tubes labelled chrome deep and chrome yellow

10 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Table 3 Description of the historic oil paint tubes

Chrome Jaune de chrome

Name on tube label Chrome deep (W&N) Chrome deep (W&N) yellow (W&N) foncé (Lefranc)

Pigment formulation PbCrO4 PbCrO4 Pb(Cr,S)O4 PbCrO4

CaCO3 (>>) MgCO§3 CaCO3 (>>) MgCO3 BaSO4 Lead soaps
MgCO§3
Paint tube Private collection Columbano Amadeo*
provenance (@DCR/FCT)# (MNAC collection) (@DCR/FCT)
§
Possibly, an additive; see text below for more details.
#
Acquired to a private collector, now at the Department of Conservation and Restoration, FCT-UNL.
*For more details, please see Otero et al. (2012).

(Fig. 5) which had belonged to the Portuguese Table 4 Pigment formulation, Lab* and ΔE* for oil paints over
Xe irradiation
painter, Columbano Bordalo Pinheiro (1857–1929);
a tube of W&N chrome deep oil paint from a tirr
(hour) L* a* b* ΔE*
private collection16; as well as a ‘Jaune de chrome
Foncé’ from the French producer Lefranc, which Middle* 0 64.78 24.88 73.39
had been left in the studio of another Portuguese PbCrO4
Downloaded by [McMaster University] at 15:26 10 April 2016

1500 48.13 14.35 39.91 38.9

>> CaCO3,
painter, Amadeo de Souza-Cardoso (1887–1918) CaSO4·2H2O 3000 46.63 13.78 38.25 41.1
(Table 3). Lemon Pb(Cr,S)O§4 0 70.14 17.68 82.47
The oil paint tube labelled chrome yellow is com- 1500 59.90 7.25 56.86 29.5
posed of mixed crystals, a small amount of barium sul- 3000 57.87 5.84 54.94 32.4
phate, and magnesium carbonate (Fig. 5). The latter
*Pigment formulation representative for middle chrome, with a
compound, designated as spar (Eastaugh et al., high extender concentration.
2004), was not listed in any W&N chrome yellow §
Mixed-crystal with SO2−
4 molar fraction between 0.5 and 0.6.

pigment production entry, but was used in their oil

paint formulations. Although the quantity of barium
W&N had chosen to stop using sodium carbonate as
sulphate detected is significantly smaller than that
a buffer in their M1a syntheses to avoid the formation
added in the L2a method (Fig. 4), a good match was
of lead carbonate since at that time there was a wide-
found between the pigment in the tube paint and the
spread distrust of lead compounds in artists’ materials
L2b pigment fingerprint (Fig. 5). As stated above,
(Carlyle, 2001).
L2b is a process variation of L2a where no BaSO4 is
As described in our previous publication (Otero
added, nonetheless the chromophores are very
et al., 2012), the tube of Lefranc’s ‘Jaune de chrome
similar. This corroborates our assumption that
foncé’ (yellow chrome deep) is composed of pure
W&N’s lemon/pale manufacturing processes were
lead chromate (Fig. 5) obtained from adding lead to
used to produce their chrome yellow oil paint.
a chromate source not buffered by the addition of a
The oil paint tubes labelled chrome deep contain a
carbonate compound. There is only one record
high amount of calcium carbonate together with
where W&N experimented with this simple process.
pure lead chromate (Fig. 5) and magnesium carbonate.
That W&N chose more complex formulations incor-
The W&N oil paint tube from the private collection
porating significant proportions of additives that
also contained magnesium carbonate. As noted
were chemically compatible is easily understood
above, the presence of calcium carbonate is character-
when we consider a note in W&N’s archive referring
istic of W&N’s middle chrome formulations, which
to this pigment’s body and high tinting strength: ‘no
support our hypothesis that their middle chrome
colour affords perhaps a greater opportunity for adul-
recipes were used to produce their chrome deep
teration 17 than chrome yellow. When pure it is of such
colour. The pigment in the chrome deep tube paints
a powerful body and has such strong qualities for stain-
correlates best with an M1a production variation
ing that it may be adulterated to a great extend without
(Fig. 5) where sodium carbonate was not added. As
much injuring its appearance as a saleable article (…)’
lead carbonate was not detected in the tube, this vari-
(quote from book G3 dated post 1844 ( p. P018).
ation is likely since the lack of sodium carbonate
Considering the results summarised in Table 1 and
ensures that lead carbonate does not form. It is also
the comparison with the early paint tubes, we
possible that by the end of the nineteenth century,

17
In this context the use of the term ‘adulteration’ is more likely to mean
16
Dated, based on the appearance of the tubes and their labels, late nine- ‘extension’ since there is ample evidence that W&N were seeking a high
teenth century W&N oil paint tubes. quality durable pigment.

Studies in Conservation 2016 11

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure 7 Cr K-edge μ-XANES spectra of middle and lemon oil paint reconstructions, an early W&N chrome yellow oil paint tube,
before and after 1000 and 15000h of irradiation, and the reference compounds: lead chromate (monoclinic), phoenicochroite, and
chromium oxide.

propose that pale and lemon formulations were used to Appendix 2 for details). For the irradiation times
produce the chrome yellow hue and, middle plus deep studied, three were very lightfast and two samples
were used to obtain the chrome deep hue. revealed strong colour changes as depicted in Table 4.
In addition, based on these preliminary obser-
vations, we propose that calcium carbonate can be Multi-analytical approach
associated with W&N’s chrome deep (Table 1 and Infrared, Raman, and UV–VIS reflectance spectra,
Appendix 1). Since calcium carbonate was not over irradiation time, for middle and lemon oil
detected in Lefranc’s chrome deep oil paint tube, this paints are presented in Fig. 6 and in Table 4 are the
opens a question for further research to establish Lab* colour coordinates. The first evidence for
whether its presence in W&N’s pigment formulation pigment degradation appears as a variation in colour
(and confirmed in the two oil paint tubes) could be (Table 4). The greatest changes in ΔE* are observed
used as a specific marker for a W&N produced for the pure monoclinic formulation, middle, where
pigment. Such research would necessarily entail inves- the pigment is admixed with two fillers (gypsum and
tigations of chrome deep produced by other contem- calcium carbonate) (Fig. 6). On the other hand, the
porary colourmen. highest decrease in a*, a loss of red, is observed for
lemon. The appearance of a green degradation
Structure and reactivity: selected examples of product such as Cr2O3 could explain this higher loss
historically accurate reconstructions in the a* coordinate when compared with b*
For a preliminary assessment of the stability to light of (Newman, 1997). Although the higher ΔE* observed
the reconstructed pigment formulations, and taking for middle is unexpected considering previous work
into account that the nineteenth century W&N (Monico et al., 2013b), these results may be explained
chrome yellow pigments were mainly based on two by taking into account the absorption of light by the
different types of pigment (i) mixed crystals with vari- pigment in the paint matrix. In general, the chromo-
able ratio of sulphate to chromate and (ii) the pure lead phore is expected to absorb less in a solid solution of
chromate monoclinic phase, five chrome yellow oil lead chromate and lead sulphate, because it is more
paint reconstructions were artificially irradiated (see ‘diluted’. However, the enhancement by a filler such

12 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

bands are relevant for the lemon oil paint, for which
differences in the ratio intensities between the sulphate
and the chromate absorption bands are observed.
Higher irradiation times will be necessary to fully
assess the degradation mechanism.

Synchrotron studies
SR μ-FTIR, μ-XANES, and μ-XRF data are in agree-
ment with the photo degradation pattern described
above. As seen in Fig. 7, Cr K-edge μ-XANES spec-
trum of Cr6+ compounds such as lead and basic lead
chromates, presents a well-defined pre-edge peak
around 5.993 keV due to a dipole-forbidden transition
Cr(1s) → 3d(t2), related to their noncentrosymmetric
Figure 8 Normalised Cr K pre-edge intensity before and
tetrahedral structure. On the other hand,
after irradiation of oil paint samples: middle and lemon Cr3+ compounds show two very low intensity pre-
chrome oil paint reconstructions and an early W&N chrome edge peaks around 5.990 and 5.993 keV assigned to
yellow oil paint tube (★). the electron transitions 1s → 3d(t2g) and 1s →
3d(eg), related to their centrosymmetrical octahedral
Downloaded by [McMaster University] at 15:26 10 April 2016

structure (Pantelouris et al., 2004; Monico et al.,

as calcium carbonate, of the photo degradation rate 2011a).
has been already observed in paints (Ferreira et al., μXRF mapping acquired at different energies, 5.993
2010; Perthue et al., 2013). Calcium carbonate does and 6.08 keV for 1500 hours of irradiation, shows that
not display a sharp cut-off, and between 300 and the thickness of the degraded surface layer in middle
350 nm it both reflects and scatters light, therefore it and lemon oil paints is between 15 and 20 μm. No
has been proposed that both the scattered and reflected additional molecular information was obtained by
light in this region could be re-absorbed by the SR μ-FTIR for the irradiation time studied.
pigment/binder system, thus intensifying light absorp- However, XANES data provided a powerful insight
tion (Ferreira et al., 2010). by establishing a correlation between irradiation time
In order to quantitatively compare the photo-stab- and the ratio of Cr K pre-edge peak from samples of
ility of the lead chromate pigments it will be necessary all paints studied before and after irradiation (Figs. 7
to measure quantum yields of reaction: this is presently and 8). This ratio reports the rate of Cr3+ that is
in progress in an experimental set-up that includes a being formed from Cr6+ (Fig. 8), which is higher for
wider range of reconstructed pigments and additives the middle formulation, in agreement with the
as well as longer irradiation times. For the present observed colour changes (Table 4). As shown by
study, based on our ageing conditions, we conclude others, this may be the result of a conversion into
that colour changes are more accentuated in one Cr2O3, a thermodynamically more stable species
W&N’s middle pigment reconstruction. It is also (Tan et al., 2013).
worth noting that in both UV–VIS absorption A first assessment of the stability to light of a sample
spectra a shoulder with a maximum at 600 nm is from the early W&N chrome yellow oil paint tube
being formed, which is more pronounced for lemon. (Table 3 and Figs. 5 and 7) was also performed for
This adds a blue component to the paint that together the irradiation time 1500 hours. A ΔE* of 24.4 was
with the yellow produces a greenish tone, which could observed and as also noted for the lemon oil paint
not be directly observed in the Lab* data (Table 4). μ- reconstruction, changes in the sulphate and the chro-
Raman was performed on cross sections of the oil mate infrared absorption bands of the mixed crystals
paints before and after irradiation (t = 3000 hours), start to appear (Appendix 4). As may be seen in
however only the pigment spectral fingerprint was Fig. 8, the ratio in the irradiated/not irradiated
detected with no degradation products. On the other samples from the W&N chrome yellow oil paint tube
hand, the infrared spectra showed that the pigment is very close to the lemon oil paint reconstruction.
to oil ratio is changing and that during irradiation, This suggests that it will be possible to compare the be-
extenders may also suffer degradation being converted haviour of original materials to reference materials
into other compounds, yet to be identified18 (see produced through reconstructions, including stable
Fig. 6). Spectral changes in the pigment absorption reconstructed chrome yellow formulations as well as
those more prone to degradation. In future work, we
18 aim to plot the percentage of transformation of orig-
Analyzing the oil matrix or oil–pigment interactions is outside the scope
of preliminary work reported here. inal Cr6+ in Cr3+ and to correlate it with a

Studies in Conservation 2016 13

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

degradation index as well as with a colour alteration other documentary sources in an effort to continue
prediction (from low conversion to very high the assessment of the vulnerability and degradation
conversion). pathways of lead chromate based pigments.

Conclusions Acknowledgments
Access to the records from W&N’s nineteenth century The authors are grateful to Winsor & NewtonTM,
archive made it possible to carry out a comprehensive ColArt Fine Art & Graphics Ltd, in particular
study of their manufacturing processes for lead chro- Emma Pearce and Ian Garrett for making the
mate pigments. All 183 production records for archive project possible. The W&N archive project
W&N’s yellow lead chromate pigments could be was funded by the Netherlands Institute for
grouped into distinct pathways resulting in three Scientific Research (NWO) as part of the De
types of pigments (Table 1): lemon/pale, middle, and Mayerne Programme and in the UK, by a Resource
deep. Lemon/pale pigments are composed of mixed Enhancement Grant from the Arts and Humanities
crystals of lead chromate and lead sulphate Research Council (AHRC). The authors acknowledge
[Pb(Cr,S)O4] with a SO2− 4 molar fraction ≤ 0.6; the Portuguese Science Foundation (FCT-MEC) for
middle pigments are mainly pure lead chromate PhD grant SFRH/BD/74574/2010 and project
[PbCrO4], and deep have a mixture of the latter with PTDC/EAT-EAT/113612/2009. The ESRF is
basic lead chromate (Fig. 2). Few records for ‘prim- acknowledged for granting beamtime (experiments
rose’ pigment were found and those present mixed HG-28 and HG-53). We also thank the
Downloaded by [McMaster University] at 15:26 10 April 2016

crystals of lead chromate and lead sulphate with a REQUIMTE supporting project UID/QUI/50006/
higher SO2−4 molar fraction (>0.6). Their experimen- 2013 and all team members of the Modern Art
tation with formulation variations demonstrates a Center of Calouste Gulbenkian Foundation, in par-
thorough and systematic investigation into the chemi- ticular to director Isabel Carlos and curator Ana
cal processes during manufacture, a testament to Vasconcelos e Melo, for their collaboration.
W&N’s determination to produce the best quality
and most stable product possible.
We have shown that the pigments in early W&N’s References
oil paint tubes labelled chrome yellow and chrome Burnstock, A.R., Jones, C.G. & Cressey, G. 2003. Characterisation
deep were made in accordance to their production of Artists’ Chromium-Based Yellow Pigments. Zeitschrift für
Kunsttechnologie und Konservierung, 1: 74–84.
records entitled lemon/pale and middle/deep, respect-
Buxbaum, G. ed. 1998. Industrial Inorganic Pigments, 2nd ed.
ively. From all the additives that they tried and used, Weinheim, New York, Chichester, Brisbane, Singapore,
we draw attention to their adoption of barium sul- Toronto: Wiley-VCH, pp. 116–9.
Carlyle, L. 2001. The Artist’s Assistant: Oil Painting Instruction
phate, used mostly in their records for lemon Manuals and Handbooks in Britain, 1800–1900, with
chrome, and to calcium carbonate, which may be Reference to Selected Eighteenth-century Sources. London:
Archetype Publications. pp. 498, 504, 521, 533–6.
characteristic of their chrome deep. Carlyle, L. 2005. With contributions by M.S.tols Witlox, K.P.ilz,
Overall, the chrome yellow paint reconstructions M.C.havannes & B.B.aade. HART Report, De Mayerne
Programme 2002–2005, Unpublished Report.
which were irradiated displayed high stability to Carlyle, L., Melo, M.J., Alves, P., Vilarigues, M. & Otero, V. 2011. A
light, and only two samples, a middle and a lemon Question of Scale and Terminology, Extrapolating from Past
pigment, showed pronounced colour change Practices in Commercial Manufacture to Current Laboratory
Experience: The Winsor & Newton 19th Century Artists’
(Table 4). These two paint formulations were further Materials Archive Database. In: J.B. Ridgland ed. Preprints
studied by means of SR based techniques. of the ICOM-CC 16th Triennial Meeting, Lisbon. Almada:
Critério-Produção Gráfica, paper 0102.
Importantly, with XANES data we were able to Chang, R. 1994. Química, 5th ed. Portugal: McGraw-Hill, pp.
show a variation in the ratio of Cr3+/Cr6+ with 757–9.
Clarke, M. & Carlyle, L. 2005. Page-image Recipe Databases, A
irradiation time ( photochemical ageing) (Fig. 8). New Approach for Accessing Art Technological Manuscripts
This approach will be developed in the future using a and Rare Printed Sources: The Winsor & Newton Archive
Prototype. In: I.V.erger, I.P.reprints of the ICOM-CC 14th
wider set of irradiated reconstructions from historical Triennial Meeting, The Hague. London: James & James, vol.
recipes and samples from early oil paint in tubes, I, pp. 24–29.
which will enable us to propose a light susceptibility Cotton, F.A. & Geoffrey, W. 1988. Advanced Inorganic Chemistry,
5th ed. New York: John Wiley & Sons, pp. 736–52.
index for yellow lead chromates. Such an index will Crane, M.J., Leverett, P., Shaddick, L.R., Williams, P.A.,
allow the evaluation of the degradation state of a Kloprogge, J.T. & Frost, R.L. 2001. The PbCrO4–PbSO4
System and its Mineralogical Significance. Neues Jahrbuch für
chrome yellow, essential to assess its current state of Mineralogie-Monatshefte, 11: 505–19.
preservation. Eastaugh, N., Walsh, V., Chaplin, T. & Siddall, R. 2004. The
Pigment Compendium, a Dictionary of Historical Pigments, 1ª
We aim to continue the comparison between our ed. Oxford: Elsevier Butterworth-Heinemann, pp. 98–100.
reconstructions and samples from artists’ paints and Erkens, L.J.H., Hamers, H., Hermans, R.J.M., Claeys, E. & Bijnens,
M. 2001. Lead Chromates: A Review of the State of the Art in
paintings contemporary with the archive production 2000. Surface Coatings International Part B-Coatings
records, and to supplement this where possible with Transactions, 84(3): 169–76.

14 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Ferreira, J.L., Melo, M.J. & Ramos, A.M. 2010. Poly(vinyl acetate) Artificial Aging of Model Samples of Co-Precipitates of
Paints in Works of Art: A Photochemical Approach. Part 1. Lead Chromate and Lead Sulfate. Analytical Chemistry, 85:
Polymer Degradation and Stability, 95: 453–61. 860–7.
Frost, R.L. 2004. Raman Microscopy of Selected Chromate Nakamoto, K. 2009. Infrared and Raman Spectra of Inorganic
Minerals. Journal of Raman Spectroscopy, 35(2): 153–8. and Coordination Compounds, Theory and Applications in
Gray, A. 1988. Lead Chrome Pigments. In: P.A. Lewis, ed. Pigments Inorganic Chemistry, 6th ed. Hoboken, NJ: Wiley-
Handbook. Properties and Economics, 2nd ed. New York: Interscience.
Wiley-Interscience Publication, vol. 1, pp. 315–25. Newman, R. 1997. Chromium Oxide Greens. Chromium Oxide
Harley, R.D. 2001. Artists’ Pigments c. 1600–1835. London: and Hydrated Chromium Oxide. In: E.W. Fitzhugh, ed.
Archetype Publications, pp. 100–02. Artists’ Pigments. A Handbook of their History and
Hummel, D.O. 2002. Atlas of Plastics Additives. Analysis by Characteristics, 1st ed. Cambridge: Cambridge University
Spectrometric Methods. Berlin: Springer-Verlag. Press and Washington: National Gallery of Art, vol. 3, pp.
Kϋhn, H. & Curran, M. 1986. Chrome Yellow and Other Chromate 273–93.
Pigments. In: R.L. Feller, ed. Artists’ Pigments. A Handbook of Otero, V., Carlyle, L., Vilarigues, M. & Melo, M. 2012. Chrome
their History and Characteristics, 1st ed. Cambridge: Yellow in Nineteenth Century Art: Historic Reconstructions
Cambridge University Press and Washington: National of an Artists’ Pigment. RSC Advances, 2(5): 1798–805.
Gallery of Art, vol. 1, pp. 187–218. Pantelouris, A., Modrow, H., Pantelouris, M., Hormes, J.,
Lide, D.R., ed. 2006. CRC Handbook of Chemistry and Physics, 87th Reinen, D. 2004. The influence of coordination geometry
ed. USA: CRC Press, pp. 8–121. and valency on the K-edge absorption near edge spectra of
Monico, L., Snickt, G., Janssens, K., Nolf, W., Miliani, C., selected chromium compounds, Chemical Physics 300,
Verbeeck, J., Tian, H., Tan, H., Dik, J., Radepont, M. & 13–22.
Cotte, M. 2011a. Degradation Process of Lead Chromate in Perthue, A., Thérias, S., Gardette, J.L., Boyer, D. & Boutinaud, P.
Paintings by Vincent van Gogh Studied by Means of 2013. Photodegradation of Down-converting Composites for
Synchrotron X-ray Spectromicroscopy and Related Application as Coatings for Organic Solar Cells. Poster pre-
Methods. 1. Artificially Aged Model Samples. Analytical sented at the International Symposium Frontiers in Polymer
Chemistry, 83: 1214–23. Science. 21–23 May, Sitges, Spain, [P3.111].
Monico, L., Snickt, G., Janssens, K., Nolf, W., Miliani, C., Roncaglia, D.I., Botto, I.L. & Baran, E.J. 1985. Vibrational
Downloaded by [McMaster University] at 15:26 10 April 2016

Verbeeck, J., Tian, H., Tan, H., Dik, J., Radepont, M., Spectrum of Pb2CrO5. Journal of Materials Science Letters,
Hendriks, E., Geldof, M. & Cotte, M. 2011b. Degradation 4(11): 1427–8.
Process of Lead Chromate in Paintings by Vincent van Gogh Shriver, D.F. & Atkins, P.W. 1999. Inorganic Chemistry, 3rd ed.
Studied by Means of Synchrotron X-ray Spectromicroscopy Oxford: OUP, p. 329.
and Related Methods. 2. Original Paint Layer Samples. Tan, H., Tian, H., Verbeeck, J., Monico, L., Janssens, K. &
Analytical Chemistry, 83: 1224–31. Tendeloo, G. 2013. Nanoscale Investigation of the
Monico, L., Janssens, K., Miliani, C., Brunetti, B.G., Vagnini, M., Degradation Mechanism of a Historical Chrome Yellow Paint
Vanmeert, F., Falkenberg, G., Abakumov, A., Lu, Y., Tian, H., by Quantitative Electron Energy Loss Spectroscopy Mapping
Verbeeck, J., Radepont, M., Cotte, M., Hendriks, E., Geldof, of Chromium Species. Angewandte Chemie International
M., Loeff, L., Salvant, J. & Menu, M. 2013a. Degradation Edition in English, 52(43): 11360–3.
Process of Lead Chromate in Paintings by Vincent van Gogh Vauquelin, L.N. 1809. Mémoire sur la meilleure méthode pour
Studied by Means of Spectromicroscopic décomposer le chrômate de fer, obtenir l’oxide de chrôme,
Methods. 3. Synthesis, Characterization, and Detection of préparer l’acide chrômique, et sur quelques combinaisons de
Different Crystal Forms of the Chrome Yellow Pigment. ce dernier. Annales de Chimie, 70: 70–94.
Analytical Chemistry, 85: 851–9. Zanella, L., Casadio, F., Gray, K.A., Warta, R., Ma, Q. & Gaillard,
Monico, L., Janssens, K., Miliani, C., Snickt, G., Brunetti, B.G., J. 2011. The Darkening of Zinc Yellow: XANES Speciation of
Guidi, M., Radepont, M. & Cotte, M. 2013b. Degradation Chromium in Artist’s Paints after Light and Chemical
Process of Lead Chromate in Paintings by Vincent van Gogh Exposures. Journal of Analytical Atomic Spectrometry, 26:
Studied by Means of Spectromicroscopic Methods. 4. 1090–97.

Studies in Conservation 2016 15

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Appendix 1: Full description of all manufacturing processes for each pigment type

Table A1 Codes for the ingredients used by W&N

Downloaded by [McMaster University] at 15:26 10 April 2016

Table A2 W&N primrose chrome (PR) manufacturing processes

Figure A1 Number of production records and date range for each of the primrose chrome manufacturing processes.

16 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Table A3 W&N lemon (L) and pale (P) chrome manufacturing processes
Downloaded by [McMaster University] at 15:26 10 April 2016

§
This order of addition of the ingredients corresponds to the majority of the productions records, however, there are also
production records where lead sulphate is added after precipitation.
#
Two production records vary the order of addition with the chromate solution added afterwards to the sodium sulphate and
barytes solution.

Studies in Conservation 2016 17

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure A2 Number of production records and date range for each of the lemon (L) and pale (P) chrome manufacturing
processes.

Table A4 W&N middle chrome (M) manufacturing processes

#
In three production records, barytes is added before precipitation.

18 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

List of suppliers
Laborspirit, Estrada de Pintéus 15, Fracção P, 2660-
194 Santo Antão do Tojal, Loures, Portugal, supplied
all chemicals from Sigma-Aldrich, Panreac, Riedel-de
Haёn, and Roig Farma; Kremer: Hauptstr. 41–47, DE
88317 Aichstetten, Germany; Sibelco: SCR-Sibelco
N.V. Plantin Moretuslei 1A, B-2018 Antwerp,
Belgium.

Figure A3 Number of production records and date range for Accelerated aging
each of the middle chrome manufacturing processes. Oil paints were prepared in 2011 by hand-grinding
with a glass muller on a glass slab. A mixture of
33.3% of untreated linseed oil* to pigment by weight
Appendix 2: Experimental section was used throughout (*HART project seed lot
Materials and methods “Electra” extracted in 2005, Carlyle, 2005). Paints
All reagents (excluding calcite and barytes) and sol- were applied onto glass slides with a palette knife.
vents were of analytical grade, and Millipore water The paints were irradiated for a period of 3000
was used. The summary of the conditions used for irradiance = 8630 MJ/m2).
Downloaded by [McMaster University] at 15:26 10 April 2016

hours (total The

the main synthetic pathways listed in Table 1 is pre- irradiation experiments were carried out in a
sented in Table A6. Solutions were always mixed CO.FO.ME.GRA accelerated aging apparatus
under agitation. All pigments were thoroughly (SolarBox 3000e) equipped with a Xenon-arc light
washed, dried and then ground to a powder. All source (λirr > 320 nm) with constant irradiation of
were fully characterised and at least twice synthesised. 800 W/m2 and black standard temperature of 50°C.

Table A5 W&N deep (D) chrome manufacturing processes

LW = lead white (2PbCO3·Pb(OH)2). HCl = hydrochloric acid. These are only used in deep chrome experimental processes.
#
§
In three production records, barytes is added before precipitation.

Studies in Conservation 2016 19

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Figure A4 Number of production records and date range for each of the deep chrome manufacturing processes.

Equipment and data acquisition The laser power at the surface of the samples was
Colorimetry. L*a*b* coordinates were measured with varied with the aid of a set of neutral density filters.
a portable spectrophotometer Data Color All samples were analysed using 15 seconds laser
Downloaded by [McMaster University] at 15:26 10 April 2016

International, using the CIELAB system, D65 illumi- exposure time for 5 scans.
nant and 10° observer. Calibration was performed μ-FTIR. Infrared analyses were carried out with a
with white ( porcelain) and black standards. Nicolet Nexus spectrophotometer coupled to a
FORS. Reflectance spectra were acquired with Continuμm microscope (15× objective) with a MCT-
Ocean Optics equipment composed by a single-beam A detector cooled by liquid nitrogen. The pigments
dispersive fibre optic spectrometer (model MAYA were prepared as KBr pellets, and spectra were col-
200 PRO) equipped with 2048-linear silicon CCD lected in transmission mode, with a resolution of
array detector (Sony ILX511B). The illumination is 4 cm−1 and 64 scans. For the micro-samples from oil
an Ocean Optics HL-200-HP with 20 W halogen paints and paintings, spectra were obtained in trans-
light source in a single optical path covering the mission mode, 4000–650 cm−1, with a resolution of
360–2400 nm range. Spectra were obtained with an 4 cm−1 and 128 scans, using a Thermo diamond
integration time of 8 ms and 15 scans to average. anvil compression cell. The spectra are shown here
The measuring head, in a 45°/45° (illumination/ as acquired, without corrections or any further manip-
acquisition angles) configuration, gives a diameter of ulations, except for the occasional removal of the CO2
analysis of about 2 mm. As reference a Spectralon® absorption at ca. 2300–2400 cm−1.
standard was used. SR-based techniques were performed at the X-ray
XRD. X-ray diffractograms were acquired with a microscopy beamline, ID21, at the European
PANalytical X’Pert PRO diffractometer equipped Synchrotron Radiation Facility (ESRF, Grenoble,
with a X’Celerator detector and using CuKa radiation France). The scanning X-ray microscope has a
in the 10 < 2θ < 80 range with a step width of 0.02°. tunable energy ranging from 2.1 to 9.1 keV and is opti-
XRD was performed at CENIMAT/i3N–Centro de mised for very low background and low detection
Investigação de Materiais/Instituto de limits. The fluorescence excitation is stimulated with
Nanoestruturas, Nanomodelação e Nanofabricação a highly monochromatic beam by means of a fixed-
of the Universidade NOVA de Lisboa. exit, double crystal Si(111) monochromator, located
SEM-EDX. SEM images were obtained using a FEI upstream of the microscope, and which provides an
Quanta 400 FEG ESEM, with a Schottky emitter field energy resolution of ΔE/E = 10−4.
emission gun, operating at low vacuum conditions and μ-XRF and μ-XANES experiments were performed
at 15 kV, equipped with an EDAX Genesis X4M in vacuum (10−5 mbar). A collimated beam of
detector. Images were acquired using a backscattered 200 μm diameter and a focussed beam of ca. 0.3 ×
(BSE) electron detector. SEM was performed within 0.8 μm2 (horizontal × vertical) obtained by means of
the Portuguese microscopy network REM, at a Fresnel zone plate were used for sample irradiation.
CEMUP-Centro de Materiais, Universidade do Porto. The XRF signal was collected in the horizontal plane
μ-Raman. Raman microscopy was carried out using perpendicular to the incident beam direction by using
a Labram 300 Jobin Yvon spectrometer, equipped a single energy-dispersive silicon drift detector (Xflash
with a HeNe laser 17 mW operating at 632.8 nm. 5100-Bruker)., which is coupled to state-of-the-art
Spectra were recorded as an extended scan. The laser digital pulse processing electronics. The energy resol-
beam was focussed with 50× Olympus objective lens. ution of this detection system is 130 eV at 6 keV. The

20 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Table A6 Summary of the conditions used for the main synthetic pathways

Synthesis

PR1a 708 ml of H2O

+
3.53 g of K2Cr2O7 dissolved in 40 ml of H2O
+
2.94 g of Na2CO3 dissolved in 20 ml of H2O
+
3.8 ml of H2SO4
+
100 ml of Pb(Ac)2·2Pb(OH)§2 (leave till next day)

L1 60 ml of H2O
+
2.942 g of K2Cr2O7
+
3.29 g of Na2CO3 (stir till no effervescence takes place; 1.5 h boiling; leave the solution to get cold)
+
4.71 g of Na2SO4 dissolved in 35.5 ml of H2O
+
3.53 g of PbSO4 in acidic solution (0.2 ml of H2SO4/10 ml of H2O)
+
27 ml of Pb(NO3)¥2 (stir for 12 hour)
Downloaded by [McMaster University] at 15:26 10 April 2016

L2a 250 ml of H2O

+
9.89 g of Na2SO4
+
60 ml of boiling H2O + 5.88 g of K2Cr2O7 + 4.94 g of Na2CO3 (boil and stir for 1/2 h; add water to fill 60 ml)
+
14.12 g of BaSO4 (natural source; Sibelco)
+
47 ml of Pb(NO3)¥2
+
3 ml of H2SO4 in 4 ml of H2O (stir for 12 hour; leave till next day)

M1a 100 ml of H2O

+
5.88 g of K2Cr2O7 dissolved in 50 ml of H2O
+
3.92 of Na2CO3 dissolved in 35 ml of H2O
+
7.84 g of CaCO3 (natural source; Kremer)
+
55 ml of Pb(NO3)¥2 (stir for 12 hour)

M2a 100 ml of H2O

+
2.65 g of K2Cr2O7 dissolved in 25 ml of boiling H2O
+
1.59 of Na2CO3 dissolved in 25 ml of boiling H2O
+
0.6 ml of H2SO4
+
18.8 ml of Pb(Ac)2·2Pb(OH)‡2
+
2.47 g of BaSO4 (natural source; Sibelco) (stir for 20 min, wash and leave to air dry until next day)

D1b 100 ml of H2O

+
2.65 g of K2Cr2O7 dissolved in 25 ml of boiling H2O
+
1.59 of Na2CO3 dissolved in 25 ml of boiling H2O
+
2.47 g of BaSO4 (natural source; Sibelco)
+
18.8 ml of Pb(Ac)2·2Pb(OH)‡2
+
0.6 ml of H2SO4 (stir for 20 min, wash and leave to air dry until next day)
Continued

Studies in Conservation 2016 21

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Table A6 Continued

Synthesis

D2a 100 ml of H2O

+
2.65 g of K2Cr2O7 dissolved in 25 ml of boiling H2O

+
0.884 of Na2CO3 dissolved in 25 ml of boiling H2O
+
18.8 ml of Pb(Ac)2·2Pb(OH)‡2
+
2.22 g of CaSO4·12H2O (Vaz Pereira) + 1.24 g of BaSO4 (natural source; Sibelco) sifted into 70 ml of H2O and stirred for 1 h
+
H2SO4 till neutral (stir for 12 hour, wash and leave to air dry until next day)
§
118 ml of boiling H2O + 12.95 g of Pb(CH3COO)2·2H2O + 12.95 g of PbO (boil and stir for 2 hours and then fill up to make 118 ml).
¥
150 ml of boiling H2O + 32.3 ml of HNO3+ 53.5 g of PbO (boil and stir for 2 hours and then fill up to make 150 ml).
‡
50 ml of boiling H2O + 5.24 g of Pb(CH3COO)2·2H2O + 5.24 g of PbO (boil and stir for 2 hours and then fill up to make 50 ml).

detection limit for elements ranging from Fe to P is The energy calibration was performed using a metallic
around 10 ppm. The sample surface was oriented Cr foil. μ-XRF and μ-XANES were performed on
vertically and at an angle of 69° relative to the cross-sections and surface sections for each irradiated
Downloaded by [McMaster University] at 15:26 10 April 2016

incident beam. During the μ-XANES energy scans, time. For each section, XANES spectrum was
the position of the primary beam was maintained acquired at 6–10 different points. Normalisation of
stable within 1 μm. μ-XANES spectra were acquired all XANES spectra was performed by means of
by scanning the primary energy around the Cr K- ATHENA software (Monico et al., 2011a; Zanella
edge (5.96–6.12 keV) with a step size of 0.2 eV. et al., 2011).

22 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Appendix 3: Characterisation of pigment reconstructions

Downloaded by [McMaster University] at 15:26 10 April 2016

Figure A5 (A) Raman, (B) diffractogram, (C) FORS, and (D) SEM image of PR1a pigment reconstruction.

Figure A6 (A) Raman, (B) diffractogram, (C) FORS, and (D) SEM image of L1 pigment reconstruction.

Studies in Conservation 2016 23

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure A7 (A) Raman, (B) diffractogram, (C) FORS, and (D) SEM image of L2a pigment reconstruction.

Figure A8 (A) Raman, (B) diffractogram, (C) FORS, and (D) SEM image of M1a pigment reconstruction.

24 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure A9 (A) Infrared, (B) Raman, (C) diffractogram, (D) FORS, and (E) SEM image of M2a pigment reconstruction.

Studies in Conservation 2016 25

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure A10 (A) Diffractogram, (B) FORS, and (C) SEM image of D1b pigment reconstruction.

Figure A11 (A) Raman, (B) diffractogram, (C) FORS, and (D) SEM image of D2a pigment reconstruction.

26 Studies in Conservation 2016

 Otero et al. Nineteenth century chrome yellow and chrome deep from Winsor & Newton TM

Appendix 4: FORS and infrared spectrum of the early W&N chrome yellow oil paint tube before and
after irradiation
Downloaded by [McMaster University] at 15:26 10 April 2016

Figure A12 FORS (left) and infrared spectrum (right) of the early W&N chrome yellow oil paint tube before and after irradiation;
( ) BaSO4, ( ) MgCO3, ( ) zinc soap and ( ) Pb(Cr,S)O4.

Studies in Conservation 2016 27