Characterization of non-ionic surfactants in Winsor & Newton's water-mixable oil paints

Abstract

Water-mixable oil (WMO) paints are a relatively new addition to the artists’ palette that have increased in usage and popularity over the past few decades. Understanding the composition and properties of WMO paint is essential for the artists who use them and for the people who work with WMO paintings—particularly art conservators. However, the formulations that make WMO paints possible remain undisclosed by paint manufacturers.
In this research, fresh paint and naturally-aged paint-outs from Winsor & Newton’s Artisan WMO products were used to study the chemistry of WMO paints. The development of a unique high-pressure liquid chromatography mass spectrometry (HPLC-MS) method enabled the separation and detection of the non-ionic surfactants in the Artisan WMO paints. A data visualization strategy using contour plots was applied for interpretation of results, affording, for the first time, complete structural identifications of surfactant molecules in a WMO paint. The paints were found to contain a complex mixture of ethoxylated molecules, including polyoxyethylene (POE) fatty acid esters derived from the fatty acid mixture of linseed oil.
An exudate that formed from the cured paints was also analyzed. More-saturated fatty acid esters, POE oleate and POE stearate, were detected; however, more-reactive POE fatty acid esters were found absent, providing insight relating to the stability of surfactants in the Artisan WMO paints as well as potential curing reactions.
HPLC-MS studies were contextualized with physical observations and surface measurements that were performed on the cured samples. Tacky surfaces observed on the cured paint films showed significantly higher surface gloss and surface hydrophilicity; these observations were quantified using glossimetry and droplet contact angle measurements. These results provide insight into the unique chemistry present in WMO paints and provide a basis for future studies for art conservators and conservation scientists working with this new media.

Full text

Characterization of non-ionic surfactants in Winsor & Newton's water-mixable oil paints
Brynn N. Sundberg* and Anthony F. Lagalante
Department of Chemistry, Villanova University, 800 Lancaster Avenue, Villanova, PA 19085
*corresponding author, brynnsundberg@gmail.com
Highlights
 A HPLC-MS method was developed for surfactants in water-mixable oil (WMO) paints
 Polyoxyethylene fatty acid esters were identified in Winsor & Newton Artisan WMOs
 WMO exudate analysis showed 18:1- and 18:0-derived surfactants migrated out of paint
Highlights

Graphical Abstract

1. Abstract
Water-mixable oil (WMO) paints are a relatively new addition to the artists’ palette that
have increased in usage and popularity over the past few decades. Understanding the composition
and properties of WMO paint is essential for the artists who use them and for the people who work
with WMO paintings—particularly art conservators. However, the formulations that make WMO
paints possible remain undisclosed by paint manufacturers.
In this research, fresh paint and naturally-aged paint-outs from Winsor & Newton’s
Artisan WMO products were used to study the chemistry of WMO paints. The development of a
unique high-pressure liquid chromatography mass spectrometry (HPLC-MS) method enabled the
separation and detection of the non-ionic surfactants in the Artisan WMO paints. A data
visualization strategy using contour plots was applied for interpretation of results, affording, for
the first time, complete structural identifications of surfactant molecules in a WMO paint. The
paints were found to contain a complex mixture of ethoxylated molecules, including
polyoxyethylene (POE) fatty acid esters derived from the fatty acid mixture of linseed oil.
An exudate that formed from the cured paints was also analyzed. More-saturated fatty
acid esters, POE oleate and POE stearate, were detected; however, more-reactive POE fatty acid
esters were found absent, providing insight relating to the stability of surfactants in the Artisan
WMO paints as well as potential curing reactions.
HPLC-MS studies were contextualized with physical observations and surface
measurements that were performed on the cured samples. Tacky surfaces observed on the cured
paint films showed significantly higher surface gloss and surface hydrophilicity; these
observations were quantified using glossimetry and droplet contact angle measurements. These
results provide insight into the unique chemistry present in WMO paints and provide a basis for
future studies for art conservators and conservation scientists working with this new media.
2. Keywords
water-mixable oil paint, modern paints, conservation, liquid chromatography, mass spectrometry,
surfactants
3. Introduction
Water-miscible, or water-mixable, oils (WMOs) were introduced to the art supply market
in the 1990s, formulated in response to demands for less-toxic art materials [1,2]. The renowned
paint manufacturer Winsor & Newton released their line of Artisan Water-Mixable oil paints, oil
medium, linseed oil, stand oil, fast-drying medium, and impasto medium in 1997. Today, WMOs
are available in a large range of colors from several brands, including Winsor & Newton (Artisan),
Royal Talens (Cobra), Grumbacher (Max), Holbein (Aqua Duo), and Lukas (Berlin). Each brand
has its unique, proprietary means to create water-miscible products. However, there are many
advantages to uncovering and understanding the chemistry of WMOs. Artists can make informed
decisions about the use of WMO paints in their works, conservators can improve their treatment
protocols and therefore contribute to the growing body of knowledge on WMO paints, and
conservation scientists can develop models of the WMO system and make better hypotheses
around WMO behavior.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

The water-miscibility of WMOs may typically be attributed to the addition of surfactants
and/or the chemical modification of the vegetable oils into emulsifiers. The packaging of Artisan
products reads that a “modified” oil, either linseed or safflower, is used as a binder for tubed WMO
paints; the modified oil is also sold separately for use as a medium. Investigations into the
composition of Artisan and Max emulsifiers have accomplished partial identifications: studies
using pyrolysis gas chromatography mass spectrometry (Py-GC-MS) [3,4] and electrospray
ionization quadrupole time-of-flight mass spectrometry (ESI-QToF-MS) [5] have detected
ethoxylated molecules. Ethoxylation is the addition of ethylene oxide to a substrate molecule and
is a common means of creating non-ionic surfactants. To completely identify a polyoxyethylene
(POE) surfactant requires more information, including the identity of the substrate molecule and
other end-groups that surround the polymer chain and the dispersity of POE polymer chain. The
identities of the ethoxylates in any brand of WMO paint have not yet been determined.
The absence of information on WMO formulation has led to speculation on the subject,
relying primarily on a 1995 patent assigned to M. Grumbacher, Inc. on Water-Reducible Artists’
Oil Paints Compositions [6] to provide evidence of the formulation of WMO paints [3,4,7,8].
Evidence does not favor this assumption, as WMOs by different manufacturers vary greatly in
their working properties when wet and in their physical properties when dry [1,5], which implies
differences in formulation. Furthermore, the suggestions of the patent should not supplant
analytical verification; to date there is a significant lack of publications that include detailed
material characterization of these paints.
According to Chambers et al., the best WMO formulations are created with ethoxylated
emulsifiers derived from fatty acid esters, or their polyol derivatives, which contain at least one
unsaturated group [6]. The methodology is appealing, as phase-separation of conventional
surfactants from coatings and films is pervasive and problematic [9]. Non-chemically incorporated
water-soluble components remaining in the dry film would be susceptible to the flux of water,
which could result in exudation from the paint film. Surfactants with unsaturated alkyl chains in
theory should, through oxidative polymerization, form covalent cross-links with the drying oil.
Such cross-linking could prevent the emulsifier from phase-separating, help maintain the long-
term stability of surfactants in the film [10,11], and even reduce surfactant removal upon aqueous
exposure [9].
The study of surfactants in painted works of art has thus far focused on acrylic latex paints,
which contain surfactants to facilitate emulsion polymerization. As described previously,
surfactants tend to phase-separate after the film has dried, which is a now well-documented
problem in conservation of acrylic paintings [12,13]. Studies on behavior of surfactants in WMOs
are few, but the high gloss, film softness, and surface tack observed in such studies [2,5] indicate
that surfactant migration is likely also occurring in WMOs. In one study on based on the
Grumbacher formulation, GC-MS and singled-sided nuclear magnetic resonance (ssNMR)
measurements of linseed paints with and without POE sorbitol hexaoleate suggested that the
addition of the surfactant resulted in changes in the oil’s structural network and slowed the rate of
curing [7].
4. Research aim
For structural determination of surfactants in complex matrices, mass spectrometry-based
techniques are essential, and in the case of mixtures of surfactants, chromatography is also
necessary. This research has adapted and applied various methods from studies of complex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

surfactant mixtures, such as polysorbates [14] and ethoxylated castor oil [15], in order to perform
chemical characterization of Winsor & Newton’s Artisan WMO paints and media. Contour plots
are used to visualize and compare high-pressure liquid chromatography (HPLC) and ESI-MS
results, discerning differences between formulations of modified linseed paints, modified
safflower paints, and modified linseed medium. An exudate that formed from the Artisan paint
samples was also analyzed, providing insight into the implications of modified oils as vestigial
surfactants in artists’ oil paints. Chemical results are contextualized with studies on surface gloss,
surface hydrophilicity, and exudate solubility.
5. Materials and methods
5.1. Paint sample preparation
Winsor & Newton’s Artisan Water-Mixable Oil Colors in titanium white, alizarin crimson,
yellow ochre, phthalo green blue hue, and French ultramarine were studied as wet paints and as
dried films, as well as Winsor & Newton’s Artisan Water-Mixable Linseed Oil medium. Draw-
downs of the Artisan paints were prepared on glass slides and on biaxially-oriented polyethylene
terephthalate (Mylar) sheets. The draw-downs were prepared at an approximate wet thickness of
100 m to allow for sufficient drying and durable film formation. All paints were allowed to dry
for a minimum of three months before testing. Glass slides dried on laboratory benchtops exposed
to ambient fluorescent light and temperature for one month before relocating to dark storage in
microscope slide boxes kept at ambient temperature. Paint-outs on Mylar sheets were hung
vertically and remained exposed to ambient fluorescent light and temperature.
5.2. Gloss measurements
Gloss values were recorded using X-Rite SP62 colorimeter using Color iControl software.
Values were averaged between three equidistant 2-cm spots across the paint film.
5.3. Water droplet contact angle
Water droplet contact angle was measured by placing 3 L droplets of Millipore 18.2 MΩ
ultrapure-filtered water (DDI water) onto the surface of an Artisan WMO paint-out on a glass slide.
The droplet image was captured using an Olympus SZ40 zoom stereomicroscope (Waltham, MA)
mounted perpendicularly to a backlit slide stage (built in-house) as shown, and an AmScope
MU100 digital camera with AmScope software (Irvine, CA). Droplet contact angle was measured
from the images using ImageJ and the Low-Bond Axisymmetric Drop Shape Analysis plug-in
(LBA-DSA; downloaded at http://bigwww.epfl.ch/demo/dropanalysis/).
5.4. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR)
ATR-FTIR spectra were recorded on a Perkin-Elmer Spectrum One with a diamond ATR crystal
between 400 and 4000 cm-1.
5.5. Pyrolysis gas chromatography and mass spectrometry (Py-GC-MS)
Cured Artisan WMO paint films in titanium white and French ultramarine (0.4 to 0.6 mg)
were derivatized with 5 L of tetramethyl ammonium hydroxide (TMAH) for analysis. The
derivatized samples were pyrolyzed with a Shimadzu PY-3030D pyrolyzer unit and separated on
a Restek Rxi®-5Sil MS 5% diphenyl column (30 m, 0.25 µm film thickness) with a Shimadzu QP-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

2020 GC-MS. Analytes were fragmented via electron ionization (70 eV) and the quadrupole was
used in scanning mode.
5.6. High-pressure liquid chromatography tandem mass spectrometry (HPLC-MS/MS)
Fresh Winsor & Newton Artisan WMO linseed oil medium and uncured WMO paints were
prepared for HPLC-ESI-MS analysis by solid-phase extraction (SPE). A Supleco Supelclean™ 3
mL C18 SPE cartridge was rinsed with HPLC-grade methanol (Fisher) and conditioned with DDI
water. 5-7 mg of fresh Winsor & Newton WMO paint or medium was suspended in 1 mL of DDI
water and applied to the cartridge. Fractions were eluted in 20%, 40%, 60%, and 80% methanol in
water and 100% methanol. Unretained solids in the paint eluates were removed by centrifugation
at 8.5 x 103 RPM for 3 minutes. Paint film exudates were not prepared for analysis by SPE, as the
exudate had migrated out of the cured yellow ochre film to the reverse side of the slide and was
readily dissolved in methanol.
HPLC separation was performed using binary Shimadzu LC-20AD pumps and a SIL-20A
autosampler. An Agilent ZORBAX XDB-C8 column (2.1 x 50 mm, 5 m) and guard column were
kept at 25 °C using a Shimadzu CTO-20A column oven. Mobile phases containing 1 mM
ammonium formate (Sigma-Aldrich, 99.995+%) in DDI water or methanol were pumped at 1
mL/min and were programmed as shown in Table 1. Analytes were evaluated on an Applied
Biosystems (now SCIEX) 3200QTRAP mass spectrometer scanning at 1000 Da/s between 250
and 1200 m/z in linear ion trap mode with positive ESI polarity. Additional MS parameter settings
are listed in Table 3. Instrument control and mass spectra were obtained using Analyst v.1.6.2
software.
Table 1. Elution gradient and MS parameters for HPLC-ESI-MS of Artisan WMO paints.
Time (min)
% ammonium formate in methanol
Parameter
Setting
0
20
CUR
20 psi
9
70
IS
5500 V
20
95
DP
40 V
25
95
EP
10 V
34
20
CE
10 V
36
20
Temp
550 °C
GS1
60 psi
GS2
60 psi
5.7. Solubility tests
Solubility of the exudate produced by the WMO paints was evaluated using four solvents
common in the conservation practice of varying composition and polarity: mineral spirits
(HomeDepot), isopropanol (Sigma-Aldrich), ethanol (Mallinckrodt), acetone (Sigma-Aldrich),
saliva, and deionized water. One-cm areas with a uniform exudate layer were visualized with a
stereomicroscope and marked before cleaning (Leica). A cotton swab was rolled across the surface
until the exudate was considered to be removed by microscopic inspection; the number of rolls
was used as a metric for comparing solubility. Solubility tests were performed in triplicate for each
solvent.
6. Results and discussion
6.1. Properties of WMO sample sets
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

The paint-outs prepared at Villanova University are only the third Winsor & Newton
Artisan WMO samples to be analyzed. The earliest published results on the properties of Artisan
WMO samples were by the chief chemist at Winsor & Newton, Alun Foster, published in the
proceedings from the Modern Paints Uncovered Symposium in 2007 [2]. An additional set of
samples encompassing WMO products from Winsor & Newton, Lukas, and Royal Talens were
prepared by Pieter Keune and studied by Alma Jongstra at the University of Amsterdam [5].
Generalities in physical properties of the model samples studied here will be established by
comparison with those of the existing observations of Foster and Jongstra. A comparison of
selected results are shown in Table 2.
The Artisan WMO medium was studied concomitant with different pigment chemistries
from a selection of five popular colors: alizarin crimson, yellow ochre, phthalo green, French
ultramarine, and titanium white (which also contained zinc white). Paint-outs were prepared on
glass slides and laid flat on laboratory benchtops to dry. Although the paint films solidified, the
surfaces of the films, with the exception of the phthalo green films, remained tacky and soft, to
such a degree that it was difficult to determine whether the films had sufficiently cured. Dust
quickly became embedded into the tacky surfaces, prompting storage after one month. Two years
since the creation of the samples, the paints remain tacky and soft, never reaching the dust-free
and dry-hard film requirements specified by the ASTM for artists’ oil paints [16]. These
observations are consistent with the sample properties described by Foster and Jongstra. Moreover,
results from Foster’s comparison of film hardness after two years drying time showed that one of
the hardest WMO films, titanium white, was still much softer than the other oil and oil-resin
equivalents from Winsor & Newton [2]. It should be noted that the glass slides were not prepared
with any type of ground or priming layer, although the technique is commonly practiced by artists
who use oil-based media. The same follows for samples tested in the studies outlined by Foster
and Jongstra. Future studies must be done to investigate the effect of porous and non-porous
substrates and surrounding layers.
Tack, gloss, and film thickness appeared to be correlated. While titanium white and yellow
ochre films remained uniformly glossy, the French ultramarine, phthalo green, and alizarin
crimson slides developed a matte oval in the center surrounded by areas of higher gloss around the
edges of the glass slides (Figure 1). Areas of high tack were also relatively thicker and of
noticeably higher gloss, and areas of less tack were relatively thinner and matte. An exception was
observed for the titanium white films, which dried into hard, glossy films with very little tack. The
titanium white paint is the only WMO studied that contains zinc oxide, which is known to increase
the brittleness and crosslinking of oil paint films [17], which is likely related to the film hardness
observed. The phthalo green films also had little tack, but were soft and very matte.
No other trends with respect to pigment were discernable, despite the many hypotheses
surrounding the effects of pigment on the stability of modern oil paints. It has been observed that
oil paints containing ultramarine or iron oxide pigments show medium separation more often than
other pigments [18], but no difference in surface tack or exudation was observed for Artisan yellow
ochre or French ultramarine samples. Similarly, paints based on organic pigments are thought to
result in weaker, less-cured polymeric networks than metal-containing pigments [19], yet Artisan
alizarin crimson paint-outs appeared as firm as inorganic pigmented colors. Furthermore, pigment-
specific phenomena observed in WMO paints must also consider possible effects due to the
pigment-specific formulation [6] of emulsifier, stabilizer, and other ingredients, which has not yet
been studied.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Table 2. Studies of Winsor & Newton Artisan WMO paint films and reported properties.
Publication
Foster 2007
Jongstra 2018
Sundberg 2020
Hue
(pigment)
Titanium white (PW6 and
PW4)
Cadmium yellow hue light
(PY3 and PY65)
Titanium white (PW6 and
PW4), yellow ochre,
alizarin crimson, French
ultramarine, phthalo green
Test sample
support
Oil sketching paper, glass
plates
Coated contrast cards
Glass microscope slides
Film age
2 years
3 years
2-3 years
Surface tack
Significant portions
remained tacky (multiple
thicknesses)
Tackiness, especially in
thicker paint-outs (200 m)
Tackiness, especially in
thicker sections of paint-
outs (100 m)
Film
hardness
(ASTM
D3363)
HB (middle grade);
Traditional oils passed
highest hardness grade
Below 6B (failed lowest
hardness grade)
Not measured, although
subjectively softer
Adhesion
on canvas
Similar to traditional oils
Loss of bulk from impasto
films
Not measured
Other
observations
Ti white formed hardest
films; other colors softer
and tackier
Loss of bulk and cracking
of impasto samples
appears to originate from
bulk.
Ti white formed hard, brittle
films. Formation of exudate
on reverse of some colors.
Figure 1. Winsor & Newton Artisan WMO paints on glass slides 3 years after painting. On the
left, a matte area on an otherwise-glossy French ultramarine paint-out, and on the right, an
exudate migrated to the glass on the reverse of a phthalo green (blue hue) paint-out.
Variations in gloss—described in the physical observations above—were quantified using
glossimetry, the results of which are shown in Figure 2. The most noticeable and significant
differences amongst the paint films are the high gloss in the yellow ochre paints and the absence
of gloss in the phthalo green paints. The gloss of the alizarin crimson, French ultramarine, and
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

titanium white paints are not significantly different. The magnitudes of the standard deviations are
consistent with the variation due to the matte spots and glossy perimeters.
The surfaces of the paint films could be separated into two types: areas of high gloss and
high tack and areas of low gloss and low tack. The surface chemistry of the two surface types was
investigated by droplet contact angle measurements. Because the liquid droplet studied here is
water, a contact angle (θ) less than 90° is also indicative of surface hydrophilicity.
Areas of higher gloss showed significantly smaller contact angles, indicative of greater
hydrophilicity and lower surface energy in accordance with the interfacial behavior of surfactants.
Only phthalo green WMO paints, having significantly less gloss than the other films, did not show
a difference in contact angles. In a study presented by Foster, across five different hues, Artisan
WMOs were consistently glossier than colors from the other Winsor & Newton oil and oil-resin
paint lines [2]. According to contact angle results presented in Figure 2, a superficial surfactant
layer may be responsible for the difference.
Figure 2. Surface measurements of Artisan WMO paints. On the left, averages and deviations in
gloss measured across two-years-old Artisan WMO paints (n=9, nTiwhite=6). On the right, average
and deviation in droplet contact angles measured in areas of distinctly high gloss and low gloss
on each Artisan WMO paint (n=3). All paints had significantly smaller contact angles and higher
surface hydrophilicity in glossy areas, with the exception of phthalo green, which was uniformly
matte.
Interestingly, the glass slides of some colors developed a tacky substance on the reverse of
the glass, a phenomenon which had not been documented in the other WMO sample sets, although
exudates are not uncommon for modern oil paints [20]. For some alizarin crimson and yellow
ochre paint-outs, this exudate coated the entire reverse of the microscope slide, appearing under
magnification as small beaded droplets. For other paint-outs, including the French ultramarine,
yellow ochre, and phthalo green, the exudate extended approximately 0.5 cm in from the perimeter
of the slide, as shown in Figure 1. Titanium white paints did not produce any observable exudate.
It is possible the droplets on the backside of some paint-outs is due to evaporative transfer from
one slide to another under the reduced airflow of the slide boxes. However, for other slides, the
exudate certainly appears to be migrating from the paint layer on the front to the edges of the
slides’ backside. The presence of an exudate did not reduce paint film adhesion. In fact, high
adhesion to the glass was observed; the WMO films could not be cleanly separated from the slide.
The paint-outs prepared on Mylar sheets were stored hanging vertically and received more
light exposure than the paint-outs on glass slides. The resultant films on Mylar were harder, and
their surfaces were less tacky than the paint films on glass slides, although the overall relationship
0.00
10.00
20.00
30.00
40.00
50.00
60.00
Alizarin
crimson
French
ultramarine
Phthalo
green
Titanium
white
Yellow
Ochre
Gloss
70.00
75.00
80.00
85.00
90.00
95.00
100.00
105.00
110.00
115.00
120.00
Alizarin
crimson
French
ultramarine
Phthalo
green
Titanium
white
Yellow
ochre
Contact angle (θ)
High gloss Low gloss
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

between thickness, gloss, and tack was consistent with the observation made of the glass slides.
Carlyle has noted that when oil-based samples are applied onto non-porous surfaces, such as Mylar
or glass, oil separation tends to occur on the surface of the support as well as the surface itself [21].
Interestingly, exudates were only observed from the paint films on glass slides, and not on Mylar.
While glass and Mylar substrates were selected to minimize the influences of other variables, the
substrate and other surrounding materials will have an effect on the equilibria established between
the WMO film and the surfactants it contains, but nature of the relationships and mechanisms at
work will require further study.
6.2. Characterizing WMO surfactants
6.2.1. ATR-FTIR
Despite the many advantages of ATR-FTIR, to be able to distinguish and identify
components in a mixture of related molecules is challenging for any non-separative technique. In
comparing spectra of a traditional linseed oil medium with the Artisan water-mixable linseed oil
and a polyethylene glycol methyl ether standard, the presence of a peak at 1100 cm-1—
corresponding to the carbon-oxygen stretch of the ethoxy moiety—stands out in the Artisan sample
(Figure S1). The ethoxylate peak is also present in the spectra of dried Artisan linseed oil and
WMO paint films; however, it is difficult to distinguish amongst the many peaks that correspond
to the pigments (Figure S2). The results indicate that ATR-FTIR would not be suitable for non-
invasive detection of a WMO medium, much less characterization of the individual surfactants an
a WMO paint.
6.2.2. Py-GC-MS
Pyrolysis of the dried WMO oil paints gave useful insight into the composition of the dried
polymeric matrix. Table S1 shows qualitative results from spectral searches performed on
chromatographic peaks.
Several different peaks containing polyoxyethylene were present in the chromatogram as
acetates and alcohols; their chains ranged between three and seven monomeric units. This is
consistent with the results of Schilling et. al., which suggested that the ethoxylates detectable by
Py-GC-MS may be used as characteristic molecules for identifying WMO paints [3]. However,
additional information about the surfactant, including the surfactant end-groups, were not able to
be determined using Py-GC-MS.
6.2.3. HPLC-ESI-MS and contour plots
MS with a soft ionization technique was used for analysis of the intact surfactant molecules,
and HPLC separation was necessary to isolate individual surfactants for ESI-MS detection. Two
HPLC methods were compared, one using a C18 column and a THF/water mobile phase based on
Nasioudis et al. [15], and the other using a C8 column with a methanol/water mobile phase.
Notably more hydrophobic species were observed using the C8 column, and its reduced retention
allowed use of more traditional chromatographic solvents with a broader working range. Elution
order remained the same. While C18 columns are used almost exclusively for the analysis of
traditional oil paint media, the C8 column proved advantageous for WMO analysis.
Despite the chromatographic gains, HPLC-MS chromatograms and individual mass spectra
were insufficient to interpret the data. The spectral complexity was due to overlapping m/z
distributions of several ethoxylated molecules in the sample. Consequently, distinguishing and
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

attributing the many masses observed in a mixture of ethoxylates was a challenge. Contour plots
compile hundreds of mass spectra into a single image, providing a more comprehensive, intuitive
visualization approach that was essential to interpretation of results without additional
chromatographic separation. Figure 3 shows a sample contour plot, which displays retention time
(tR) on the x axis, m/z on the y axis, and relative ion intensity in color temperature, as detected in
SPE fractions from the Artisan linseed oil medium. Figure 4 compares the ion cluster data from
the Artisan linseed oil medium with data from two Artisan WMO paints: one of modified linseed
oil, the other of modified safflower oil. The presence of ethoxylated compounds is evidenced in
the contour plots by vertical bands of ions regularly interspaced by the weight of an ethoxylate
unit (CH2CH2O, 44.05 m/z). From a contour plot, co-eluting ethoxylates are easily distinguished
by observing the contour traced by the elution in order of POE chain length.
The analytes observed in the contour plots were categorized into three groups according to
their mass and chromatographic behavior. The compounds that elute in the first 6 minutes were
designated Ion Cluster 1 (IC1). In IC1, different POE chain lengths elute according to molecular
size (the number of ethoxylate units) and affinity to the stationary phase (hydrophobicity). Ion
clusters 2 and 3 (IC2 and IC3), primarily observed in the 100% methanol fraction, consist of more
highly-retained molecules that elute predominantly by hydrophobicity; POE chain lengths elute
almost simultaneously.
Figure 3. Contour plot of components detected in the 100% methanol fraction of Winsor &
Newton’s Artisan Water-Mixable Linseed Oil medium. Mass-to-charge ratio is on the y axis,
retention time is on the x axis, and relative ion intensity is expressed on a scale of pink to white.
Three ion clusters (ICs) of interest are labeled. Ethoxylates are easily visualized by the vertical
bands of ions regularly interspaced by the weight of an ethoxylate unit. The contour of a band
distinguishes one ethoxylated molecule from another.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Figure 4. Comparisons of ion clusters 1, 2, and 3 from contour plots of Winsor & Newton’s Artisan
Water-Mixable products: modified safflower oil paint, modified linseed oil paint, and modified
linseed medium. IC1 species were isolated from the 80% methanol SPE fraction; IC2 and IC3
species were from the 100% methanol fraction. Relative intensity is comparable between ion
clusters of a single product (with allowances for differences in ionization efficiency), but not across
products.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Figure 5. On the left, a comparison of IC2 from the contour plots of an Artisan linseed WMO paint
(yellow ochre, in blue) and an Artisan safflower WMO paint (titanium white, in pink). On the right,
the same contour plots are traced and labeled with R groups. Fatty acid ethoxylates (FAEs,
bottom) are traced with solid lines, and fatty acid methyl ester ethoxylates (FAMEEs, top) are
traced with dashed lines. POE linolenate is present in the linseed WMO and absent in the safflower
WMO, but POE linolenate methyl ester is present in both linseed and safflower WMOs.
The identities of the early-eluting analytes in the 80% methanol fractions were not
ascertained, as there are many isomeric ethoxylated molecules manufactured and possibly
included in WMO formulations. High-resolution MS fragmentation methods combined with end-
group analysis could be used to identify these species [23].
Identifying the analytes found in the 100% methanol fractions was more straightforward
due to the unique patterns in elution and mass presented by long-chain fatty acids. IC2 shows
distributions of fatty acid ethoxylates that are consistent with the fatty acid profiles in linseed and
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

safflower oils, most easily observed by the overlap of two color-adjusted contour plots in Figure
5. The 18-carbon fatty acid ethoxylates elute in order of decreasing polarity, from most
unsaturated to most saturated, while POE palmitate elutes between POE linoleate and POE oleate.
For the WMO paints, IC2 also contains fatty acid methyl ester ethoxylates, which followed the
same elution order. Interestingly, the safflower WMO paint contains the POE methyl linolenate
species at ion intensities that are inconsistent with a modified safflower oil. Monoglycerides, as
might be found in a traditional vegetable oil binder, were not detected among the fatty acid
ethoxylates. Masses indicative of POE triglycerides were also absent from WMO contour plots.
Masses of POE diglycerides were not distinguishable from those of POE diesters, although the
absence of glycerol or glycerides renders the POE diglycerides less likely.
Further verification of the identities of the POE derivatives was sought through MS
fragmentation studies. IC3 is most prominent in the 100% methanol fraction of the Artisan linseed
oil, although traces are also visible in the safflower oil paint. Two co-eluting ions from IC3, with
masses one ethoxylate monomer apart (953.4 m/z and 997.4 m/z), were selected for product ion
analysis to confirm their structures. The product ion spectra are very similar: both precursor ions
produced one major product ion of [M+H]+ = 305.5 m/z and minor low molecular-weight
fragments consistent with an unsaturated alkyl chain (Figure 6). In IC2, the ethoxylate ions
corresponded to the loss of the 305.5 m/z fragment from the IC3 ions, thus appearing to be
structurally related. Fragmentation of prominent 648.4 m/z and 692.4 m/z ions in IC2 resulted in
analogous spectra with the same major product ion, 305.5 m/z. The patterns observed suggest that
the 953.4, 997.4, 648.4, and 692.4 m/z ions all have the same alkyl moiety, which produces the
305.5 m/z fragment, while respective POE chains are lost as neutrals. The characteristic 305.5 m/z
fragment structure in Figure 7 was proposed by Nasioudis et al. for major product ions observed
from ammonium adducts of fatty acid ethoxylates [15]. The IC3 masses and fragmentation
patterns are consistent with that of POE(9) and POE(10) dilinolenate (diesters), also shown in
Figure 8. The molecular ion masses and fragmentation of the 648.4 m/z and 692.4 m/z ions in IC2
are consistent with that of the POE(8) and POE(9) linolenate (monoesters). Other possible
molecules contained in WMOs, determined from molecular ion masses, are listed in Table 3.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Figure 6. Product ion spectra from ion cluster 3 ([M+NH4]+=997.4 m/z, left) and ion cluster 2
([M+NH4]+=692.4 m/z, right) in Winsor & Newton’s Artisan Water-Mixable Linseed Oil medium,
confirming their identities as POE(10) dilinolenate and POE(9) linolenate, respectively.
Figure 7. Fragmentation of the POE(10) dilinolenate ion (997.4 m/z, upper left) and the POE(9)
linolenate ion (692.4 m/z, upper right). Both result in the characteristic POE linolenate product
ion (30.55 m/z, bottom center).
Direct ethoxylation of fatty acids with conventional catalysts yields diesters and
polyethylene glycols in addition to the surface-active monoester. The diester is formed through
transesterification between monoesters or through nucleophilic substitution at both terminal
+EPI (692.50) CE (15) CES (5): 10 MCA scans from Sa... Max. 3.1e7 cps.
50 100 150 200 250 300 350 400 450 500 550 600 650 700
m/z, Da
0.0
2.0e6
4.0e6
6.0e6
8.0e6
1.0e7
1.2e7
1.4e7
1.6e7
1.8e7
2.0e7
2.2e7
2.4e7
2.6e7
2.8e7
3.0e7
3.1e7
Intensity, cps
692.4
305.3
675.4
149.3 177.2 306.8
135.2 159.1 371.2
199.2
136.3 437.3
349.4 657.4
249.2 481.2 569.4 613.489.2 680.8
+EPI (996.60) CE (60) CES (15): 6 MCA scans from Sa... Max. 1.6e7 cps.
100 200 300 400 500 600 700 800 900 1000
m/z, Da
0.0
1.0e6
2.0e6
3.0e6
4.0e6
5.0e6
6.0e6
7.0e6
8.0e6
9.0e6
1.0e7
1.1e7
1.2e7
1.3e7
1.4e7
1.5e7
1.6e7
Intensity, cps
305.8
997.4
78.7
149.4
80.9
287.8
134.8 160.1
120.7
213.8 243.7 349.9
190.4
116.8 251.6 393.8 676.1
481.9 645.2 719.0 909.7
+EPI (692.50) CE (15) CES (5): 10 MCA scans from Sa... Max. 3.1e7 cps.
50 100 150 200 250 300 350 400 450 500 550 600 650 700
m/z, Da
0.0
2.0e6
4.0e6
6.0e6
8.0e6
1.0e7
1.2e7
1.4e7
1.6e7
1.8e7
2.0e7
2.2e7
2.4e7
2.6e7
2.8e7
3.0e7
3.1e7
Intensity, cps
692.4
305.3
675.4
149.3 177.2 306.8
135.2 159.1 371.2
199.2
136.3 437.3
349.4 657.4
249.2 481.2 569.4 613.489.2 680.8
+EPI (996.60) CE (60) CES (15): 6 MCA scans from Sa... Max. 1.6e7 cps.
100 200 300 400 500 600 700 800 900 1000
m/z, Da
0.0
1.0e6
2.0e6
3.0e6
4.0e6
5.0e6
6.0e6
7.0e6
8.0e6
9.0e6
1.0e7
1.1e7
1.2e7
1.3e7
1.4e7
1.5e7
1.6e7
Intensity, cps
305.8
997.4
78.7
149.4
80.9
287.8
134.8 160.1
120.7
213.8 243.7 349.9
190.4
116.8 251.6 393.8 676.1
481.9 645.2 719.0 909.7
+EPI (692.50) CE (15) CES (5): 10 MCA scans from Sa... Max. 3.1e7 cps.
50 100 150 200 250 300 350 400 450 500 550 600 650 700
m/z, Da
0.0
2.0e6
4.0e6
6.0e6
8.0e6
1.0e7
1.2e7
1.4e7
1.6e7
1.8e7
2.0e7
2.2e7
2.4e7
2.6e7
2.8e7
3.0e7
3.1e7
Intensity, cps
692.4
305.3
675.4
149.3 177.2 306.8
135.2 159.1 371.2
199.2
136.3 437.3
349.4 657.4
249.2 481.2 569.4 613.489.2 680.8
+EPI (996.60) CE (60) CES (15): 6 MCA scans from Sa... Max. 1.6e7 cps.
100 200 300 400 500 600 700 800 900 1000
m/z, Da
0.0
1.0e6
2.0e6
3.0e6
4.0e6
5.0e6
6.0e6
7.0e6
8.0e6
9.0e6
1.0e7
1.1e7
1.2e7
1.3e7
1.4e7
1.5e7
1.6e7
Intensity, cps
305.8
997.4
78.7
149.4
80.9
287.8
134.8 160.1
120.7
213.8 243.7 349.9
190.4
116.8 251.6 393.8 676.1
481.9 645.2 719.0 909.7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

hydroxy groups of the polyoxyethylene, and remains in very high proportions at equilibrium, with
a molar ratio between monoesters, the diester, and polyethylene glycol at approximately 2:1:1 [24].
Table 3. Molecular ions identified in an Artisan Water-Mixable Linseed oil medium, linseed paint
(yellow ochre), and safflower paint (titanium white), including number of ethoxylate units (n) and,
when applicable, the fatty acid carbon count and unsaturation.
Possible identity
[M+NH4]+
Linseed
oil
Linseed
paint
Safflower
paint
1
Polyethylene glycol (n=8-16)
388.5-740.9
X
2
POE sorbitan (n=9-14)
578.9-799.1
X
3
POE linolenate (n=4-17, 18:3)
473.0-1046.2
X
X
4
POE linoleate (n=4-16, 18:2)
475.0-1003.2
X
X
X
5
POE palmitate (n=4-14, 16:0)
451.0-891.2
X
X
X
6
POE oleate (n=4-16, 18:1)
477.0-1006.3
X
X
X
7
POE stearate (n=5-13, 18:0)
523.0-876.9
X
X
X
8
POE linolenate (n=3-16, 18:3) methyl ester
443.3-1015.9
X
X
9
POE linoleate (n=4-15, 18:2) methyl ester
489.5-974.0
X
X
10
POE palmitate (n=4-15, 16:0) methyl ester
465.3-949.9
X
X
11
POE oleate (n=4-15, 18:1) methyl ester
491.3-975.9
X
X
12
POE stearate (n=6-13, 18:0) methyl ester
581.5-889.9
X
X
13
POE sorbitan oleate (n=8-12, 18:1)
799.1-975.5
X
14
POE dilinolenate (n=4-14, 18:3-18:3)
733.4-1173.7
X
15
POE linolenate-linoleate diester (n=6-13,
18:3-18:2)
823.5-1131.5
X
16
POE dilinoleate or POE linolenate-oleate
(n=6-12, 18:1-18:1 or 18:3-18:1)
825.3-1089.7
X
17
POE linolenate-stearate or POE linoleate-
oleate (n=8-10, 18:3-18:0 or 18:2-18:1)
915.5-1003.6
X
X
The majority of ions detected by HPLC-ESI-MS were ethoxylated molecules. Preferential
ionization and detection of surfactant molecules is a commonly-observed phenomenon in ESI, as
surfactants outcompete other analytes at interfaces [25]. Less volatile analytes, such as free fatty
acids, which are presumed to be present in an oil paint, were not observed by this technique.
Another explanation for the absence of other organic components may be due to the upper and
lower mass cutoffs of the mass analyzer. However, ions with masses greater than 1300 m/z were
also not detected in direct injection ESI-QToF analyses by Jongstra, whose mass measurements
extended to 2000 m/z [5]. Successful detection of the unmodified fatty acids and fatty acid
glycerides in WMO samples by HPLC-MS would require additional sample preparation steps, a
different ionization method, or another approach altogether.
The presence of fatty acid ethoxylates as “modified” oils is consistent with literature on the
emulsification and stability of surface coatings. Fatty acid ethoxylates are widely manufactured to
be used as lubricants, emulsifiers, and viscosity modifiers [24].
Having identified specific surfactants in the Artisan WMOs, existing surfactant literature
can be utilized to inform our understanding of the paints. Of particular interest are studies regarding
the degradation of alkyl ester ethoxylates, which is possible through two pathways: hydrolysis and
autoxidation. Alkyl esters are labile to hydrolysis, particularly in high alkaline or acid conditions
[26]. Another potential cause of alkyl ethoxylate degradation is the autoxidation of the POE chain
and the unsaturations of the alkyl chain, both of which can undergo radical reactions leading to
scission [27]. As with paint films, exposure to light contributes to the initiation of radical pathways.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

For example, mixtures containing polysorbate 80, a sorbitan fatty acid ethoxylate, produced up to
1000-fold increases in peroxide content when incubated in the presence of light [28]. Although
hydrolytic and autoxidative products of fatty acid ethoxylates were detected in the WMO paints,
they could not be conclusively assigned to those degradation phenomena. The effect of the
additional reactive potential created by the addition of POE on oil paint curing is a topic that will
require further study.
6.3. Analysis of an exudate
The aggregation of an exudate on the reverse of the WMO paint-out slides was a
phenomenon not previously documented for WMOs. Although the exudate may not be
representative of the WMO paint behavior, paint medium separation has been frequently observed.
The exudate composition was investigated by HPLC-ESI-MS. The contour plot of the exudate
showed many similarities with the contour plots from the wet paints, most notably, the presence
of fatty acid ethoxylates. Proposed exudate components are listed in Table 4, and overlapping IC2
contour plots of the wet Artisan linseed oil paint and the exudate from a dried film are shown in
Figure 8. The overlapping plots clearly show that the most reactive of the fatty acid ethoxylates,
that is, the fatty acids with higher degrees of unsaturation, are absent from the exudate. The contour
plots indicate that the surfactants that exuded from the paint were those unable or least likely to
cross-link into the film, while the more reactive fatty acid ethoxylates and fatty acid methyl ester
ethoxylates remained within the film. More polar, unidentified ethoxylates (IC1) were also
measured in the WMO exudate. Polyethylene glycol could be an intentional addition to the paints,
a byproduct of fatty acid ethoxylate synthesis, a product of autoxidation, or some combination of
the three.
Table 4. Molecular ions identified in a yellow ochre Artisan WMO paint and its exudate, including
number of ethoxylate units (n) and, when applicable, the fatty acid carbon count and unsaturation.
Possible identity
[M+NH4]+
Linseed
paint
Exudate
1
Polyethylene glycol (n=7-14)
344.4-652.8
X
2
POE linolenate (n=5-12, 18:3)
517.1-825.5
X
3
POE linoleate (n=5-12, 18:2)
519.0-827.6
X
4
POE palmitate (n=5-12, 16:0)
494.9-803.2
X
X
5
POE oleate (n=5-12, 18:1)
521.2-829.9
X
X
6
POE stearate (n=7-9, 18:0)
611.0-699.6
X
X
7
POE linolenate (n=5-12, 18:3) methyl ester
531.4-839.7
X
8
POE linoleate (n=5-12, 18:2) methyl ester
533.5-841.9
X
9
POE palmitate (n=5-12, 16:0) methyl ester
509.4-817.8
X
X
10
POE oleate (n=5-12, 18:1) methyl ester
535.4-931.8
X
X
11
POE stearate (n=7-9, 18:0), methyl ester
625.6-713.7
X
X
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Figure 8. On the left, a comparison of IC2 from the contour plots of an Artisan linseed WMO paint
(yellow ochre, in blue) and its exudate (in pink). On the right, the same contour plots are traced
and labeled with shared R groups. Fatty acid ethoxylates (FAEs, bottom) are traced with solid
lines, and fatty acid methyl ester ethoxylates (FAMEEs, top) are traced with dashed lines. POE
linoleic and POE linolenic derivatives are missing from the exudate contour plot; only the less-
reactive POE palmitic, POE stearic, and POE oleic derivates are found in the exudate.
The formation of exudates in modern oil paints is not uncommon and is currently
hypothesized to be due to the slow aggregation of polar molecules in an otherwise non-polar oil
environment [19,20]. While the species detected in the WMO exudate also contained polar
molecules, the polarity of the WMO film environment is presumed to be more polar than a
traditional oil film, and thus a different explanation may be required. The formation of this exudate
may simply be due to the interfacial preferences of non-ionic surfactants, more similar to the
behavior of surfactants in acrylic films. As noted earlier, the formation of an exudate may also be
related to the non-porous sample substrate used, although exudates were observed only with
certain colors and only on glass, not Mylar. In the cases that the exudate did not form, more
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

thorough drying was suspected, which may have impeded the mobility of molecules out of the
films. Surfactant aggregation at the surface and formation of an exudate may have negative
consequences for adhesion between paint layers, although no problems with adhesion were
observed in this sample set.
6.4. Solubility of fatty acid ethoxylates
As the exudate was most pronounced on the recto side of the painted slide, it was easily
visualized in raking light for solubility and removal trials. To investigate removal efficacy of
solvents, a selection of six solvents common in the conservation practice and of differing
composition and polarities (mineral spirits, isopropanol, ethanol, acetone, saliva, and distilled
water) were used to conduct trials with a cotton swab repeatedly rolled over the exudate. It was
apparent that removal of the exudate from a glass surface was more complicated than anticipated,
as swab rolls could initially remove a large portion of the exudate; however, a distinct residue
remained on the glass that could only be disrupted by the application of force and swabbing as
opposed to rolling. The polar organic solvents (isopropanol, ethanol, acetone) were most effective
at exudate removal in comparison to aqueous or aliphatic solvents. Additionally, solvents in the
wet swab were delicately brought to the surface of the yellow ochre painted film to investigate
colorant removal (Table 5). The use of isopropanol, ethanol, and acetone immediately and
significantly removed ochre from the paint film, while saliva, water, and mineral spirits did not.
This perhaps suggests that the physical action of solvent swabbing, rather than the chemical
process of solvent dissolution, may be explored as a means for disruption of the exudate during
cleaning. For instance, gelled aliphatic solvents (e.g. ShellSol D60) may be an initial low-polarity
starting point for cleaning investigations, as a gel would remove displaced exudate and the longer
working time with the aliphatic solvent would be advantageous.
Interestingly, the absence of water-sensitivity may differentiate the Artisan WMO samples
from the fragile, water-soluble, medium-rich skins observed on some modern oil paints and the
sensitive surfactant layer of acrylic paints.
Table 5. Solubility of the WMO exudate in different solvents and sensitivity of the WMO paint
surface to those solvents.
Solvent
Paint sensitivity
Mineral spirits
Low
Isopropanol
High
Ethanol
High
Acetone
High
Saliva
Medium
Deionized water
Medium
7. Conclusions
The HPLC-ESI-MS method, combined with contour plot visualization, was shown here to
be an effective means of simplifying a complex analytical problem and was successfully
implemented to characterize non-ionic surfactants in the Artisan WMO paints. Identification of
surfactants in the paints allowed connections to be made with existing literature regarding their
manufacture and stability. Furthermore, the ability to distinguish specific fatty acid ethoxylates
enabled comparisons between a WMO paint and its exudate, which gave insight into the chemical
reactions taking place inside the paint film upon curing.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

The results of the compositional analyses on the Artisan WMO products provide a
foundation for understanding the macroscopic phenomena observed in WMO sample sets. The
clear variation in surface gloss and the persistent tackiness on the surfaces of the film likely result
from surfactants that have migrated to the boundaries of the paint films. The attribution of
differences in gloss to differences in surfactant aggregation was corroborated by data from contact
angle measurements, which showed significant differences in surface hydrophilicity between areas
of high and low gloss. More conclusive evidence of surfactant migration was found in the fatty
acid ethoxylates and other POE surfactants exuding from the WMO paint-outs. While the
exudation of polyethylene glycol, fatty acid ethoxylates, and other POE surfactants is of concern,
it has not yet been observed in other studies. Again, more complex, layered paint models will
change whether such outcomes are observed, but the tendency for medium separation seems to be
inherent to the less-cured WMO paints. Initial solubility results indicated that if such an exudate
was to be removed from a painted surface, polar solvents should be avoided.
The major physical observations made of this set of Artisan WMO paint samples were
consistent with observations from other WMO studies, which all show a tendency towards soft
films with tacky, glossy surfaces. Although a more complex sample stratigraphy would likely
affect these results, the tendency to develop tacky surfaces is so far shown to be intrinsic to the
Artisan WMO medium. The tacky surfaces of WMOs will certainly expedite the necessity for
cleaning embedded dust and soils. Future studies of the composition, properties, and mechanism
of formation of the surface layer will determine whether it can be conceptualized as a medium skin
(as in modern oil paints), as a surfactant layer (as in acrylic paints), or as something else entirely.
HPLC-ESI-MS results indicate that the medium and surfactant are one and the same, and perhaps
the results of both areas of research will inform our understanding of the tacky surfaces on WMO
films.
Studies on increasingly complex model systems will be necessary to place the phenomena
observed in isolated WMO systems within the context of a painting. The effect of substrate type
and porosity, ground, additional paint layers, and varnish will be an essential continuation of this
research. Additional studies will be necessary to understand the optimal storage conditions for
WMO films, especially given the presence of a polar, mobile fraction of surfactants in a paint film
already rendered hydrophilic by the addition of ethoxylates.
8. Acknowledgements
The authors gratefully acknowledge Alan Owens and Bryan Evans at Shimadzu for the Py-
GC-MS measurements. We thank Kristen deGhetaldi and Brian Baade for comments during the
preparation of this manuscript. This research was partially supported by a Villanova University
graduate summer research fellowship to BNS.
9. References
[1] S. Dye, Water Soluble Oils, North Light Books, Cincinnati, 2001.
[2] A. Foster, The performance and properties of artisan water mixable oil colour compared
with other oil-based paints by Winsor & Newton, in: Mod. Paint. Uncovered Proc. from
Mod. Paint. Uncovered Symp., 2007: pp. 53–57.
[3] M.R. Schilling, J. Mazurek, J.S. Learner, T.J.S. Learner, J.S. Learner, T.J.S. Learner,
Studies of Modern Oil-Based Artists` Paint Media by Gas Chromatography/Mass
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Spectrometry, in: T.J.S. Learner, P. Smithen, J.W. Krueger, M.R. Schilling (Eds.), Mod.
Paint. Uncovered - A Symp. Organ. by Getty Conserv. Institute, Tate, Natl. Gall. Art, Getty
Conservation Institute, Los Angeles, 2006: pp. 129–139.
doi:10.1017/CBO9781107415324.004.
[4] R.E. Hodgkins, B. Berrie, S. Lomax, Multi-analytical approach to characterization and
comparison of Grumbacher Pre-tested and Max water mixable oils, Technart. (2015).
[5] A. Jongstra, K.J. van den Berg, E. Hendriks, Research into water-miscible oil paints by
Royal Talens, Winsor & Newton and Lukas, University of Amsterdam, 2018.
[6] G. Chambers, W.M. Hoyte, G.J. Stegmeir, Water Reducible Artists’ Oil Paints
Compositions, USP5312482, 1994.
[7] N.A. Udell, R.E. Hodgkins, B.H. Berrie, T. Meldrum, Physical and chemical properties of
traditional and water-mixable oil paints assessed using single-sided NMR, Microchem. J.
133 (2017) 31–36. doi:10.1016/j.microc.2017.03.013.
[8] J.M. Myers, M. Cushman, Analysis of Cleaning Efficacy and Clearance of Silicone-Based
Pickering-Type Emulsions Used in the Cleaning of Water-Miscible Oil Paints, in: K.J. van
den Berg, I. Bonaduce, A. Burnstock, B. Ormsby, M. Scharff, L. Carlyle, G. Heydenreich,
K. Keune (Eds.), Conserv. Mod. Oil Paint., Springer International Publishing, Cham, 2019:
pp. 565–574.
[9] K. Holmberg, Applications of surfactants in paints, in: D.R. Karsa (Ed.), Surfactants Polym.
Coatings, Inks, Adhes., CRC Press LLC, Boca Raton, FL, USA, 2003: pp. 152–179.
[10] K. Holmberg, Polymerizable surfactants, Prog. Org. Coatings. 20 (1992) 325–337.
doi:10.1016/0033-0655(92)80022-O.
[11] K. Holmberg, Unsaturated monoethanolamide ethoxylates as paint surfactants, in: H.J.
Jacobasch (Ed.), Prog. Colloid Polym. Sci., Steinkopff, 1996: pp. 69–74.
doi:10.1007/BFb0114446.
[12] B. Ormsby, T. Learner, The effects of wet surface cleaning treatments on acrylic emulsion
artists ’ paints – a review of recent scientific research, Rev. Conserv. 10 (2009) 29–41.
[13] C.E. Dillon, A.F. Lagalante, R.C. Wolbers, Acrylic emulsion paint films: The effect of
solution pH, conductivity, and ionic strength on film swelling and surfactant removal, Stud.
Conserv. 59 (2014) 52–62. doi:10.1179/2047058412Y.0000000076.
[14] D. Hewitt, M. Alvarez, K. Robinson, J. Ji, Y.J. Wang, Y.H. Kao, T. Zhang, Mixed-mode
and reversed-phase liquid chromatography-tandem mass spectrometry methodologies to
study composition and base hydrolysis of polysorbate 20 and 80, J. Chromatogr. A. 1218
(2011) 2138–2145. doi:10.1016/j.chroma.2010.09.057.
[15] A. Nasioudis, J.W. Van Velde, R.M.A. Heeren, O.F. Van den Brink, Detailed molecular
characterization of castor oil ethoxylates by liquid chromatography multistage mass
spectrometry, J. Chromatogr. A. 1218 (2011) 7166–7172.
doi:10.1016/j.chroma.2011.08.032.
[16] ASTM, Standard Specification for Artists ’ Oil , Resin-Oil , and Alkyd Paints 1, Astm.
(2015) 1–9. doi:10.1520/D4302-05R10.
[17] M.F. Mecklenburg, Determining the Acceptable Ranges of Relative Humidity And
Temperature in Museums and Galleries: Part 2, Structural Response to Temperature, 2007.
[18] J. Lee, B. Ormsby, A. Burnstock, K.J. van den Berg, Modern Oil Paintings in Tate’s
Collection: A Review of Analytical Findings and Reflections on Water-Sensitivity, in: K.J.
van den Berg, I. Bonaduce, A. Burnstock, B. Ormsby, M. Scharff, L. Carlyle, G.
Heydenreich, K. Keune (Eds.), Conserv. Mod. Oil Paint., Springer International Publishing,
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Cham, 2019: pp. 495–522.
[19] I. Bonaduce, C. Duce, A. Lluveras-Tenorio, J. Lee, B. Ormsby, A. Burnstock, K. Jan van
den Berg, Conservation Issues of Modern Oil Paintings: A Molecular Model on Paint
Curing, Acc. Chem. Res. 52 (2019) 3397–3406. doi:10.1021/acs.accounts.9b00296.
[20] J. de Groot, L. Steyn, M. Stols-Witlox, K.J. van den Berg, Decision-Making Processes
Regarding the Treatment of Modern Oil Paintings (1950s–Present) Exhibiting Paint
Dripping and Oil Exudates, in: K.J. van den Berg, I. Bonaduce, A. Burnstock, B. Ormsby,
M. Scharff, L. Carlyle, G. Heydenreich, K. Keune (Eds.), Conserv. Mod. Oil Paint.,
Springer International Publishing, Cham, 2019: pp. 373–381.
[21] L. Carlyle, M. Witlox, Historically Accurate Reconstructions of Artists’ Oil Painting
Materials, Tate Pap. 7 (2007).
[22] S. Digney-Peer, A. Burnstock, T. Leamer, H. Khanjian, F. Hoogland, J. Boon, The
Migration of Surfactants in Acrylic Emulsion Paint Films, Stud. Conserv. 49 (2004) 202–
207. doi:10.1179/sic.2004.49.s2.044.
[23] F.G. Hoogland, J.J. Boon, Development of MALDI-MS and nano-ESI-MS methodology
for the full identification of poly(ethylene glycol) additives in artists’ acrylic paints, Int. J.
Mass Spectrom. 284 (2009) 66–71. doi:10.1016/j.ijms.2009.03.001.
[24] K. Kosswig, Polyoxyethylene Esters of Fatty Acids, in: N.M. van Os (Ed.), Nonion.
Surfactants Org. Chem. Surfactant Sci. Ser. ; V. 72, Marcel Dekker, Inc., New York, 1998:
pp. 123–146.
[25] R. Zhang, Y. Wang, L. Tan, H.Y. Zhang, M. Yang, Analysis of polysorbate 80 and its
related compounds by RP-HPLC with ELSD and MS detection, J. Chromatogr. Sci. 50
(2012) 598–607. doi:10.1093/chromsci/bms035.
[26] M.F. Cox, U. Weerasooriya, Methyl Ester Ethoxylates, J. Am. Oil Chem. Soc. 74 (1997)
847–859. doi:10.1007/s11746-997-0228-4.
[27] E. Santacesaria, D. Gelosa, M. Di Serio, R. Tesser, Thermal stability of nonionic
polyoxyalkylene surfactants, J. Appl. Polym. Sci. 42 (1991) 2053–2061.
doi:10.1002/app.1991.070420733.
[28] E. Ha, W. Wang, Y.J. Wang, Peroxide formation in polysorbate 80 and protein stability, J.
Pharm. Sci. 91 (2002) 2252–2264. doi:10.1002/jps.10216.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

Supplementary Material
Click here to access/download
Supplementary Material
Supplemental info.docx