Progress in Organic Coatings 48 (2003) 8091

Oxidation of model compound emulsions for alkyd

paints under the influence of cobalt drier
Z. Okan Oyman, W. Ming , R. van der Linde
Laboratory of Coatings Technology, Department of Chemical Engineering and Chemistry, Eindhoven University of Technology,
P.O. Box 513, 5600 MB Eindhoven, The Netherlands
Received 17 January 2003; accepted 20 May 2003

Abstract
A systematic investigation on the oxidation of different model compounds (including methyl oleate (MO), ethyl linoleate (EL), and
methyl linolenate (MLn)) in emulsions under the influence of an emulsifiable cobalt drier was performed. Raman spectroscopy, attenuated
total reflectance (ATR)-FTIR, and 1 H NMR were used. Raman spectroscopy proved to be a powerful technique at following the oxidation
of these model compounds in emulsions not only after but also during the water evaporation. Raman and ATR-FTIR (D2 O was used instead
of H2 O) investigations clearly demonstrated that there was essentially no observable chemical change during the water evaporation. The
oxidation of model compounds in emulsions, after the water evaporation, was found to be similar to the oxidation of bulk/solvent-borne
model compounds. The presence of an emulsifier (sodium dodecyl sulfate) and water did not show any significant effects on the oxidation
of the model compounds. An increase in the intensity of conjugated double bonds and isolated trans-double bonds was first observed, and
followed by a decrease. More than 90% of double bonds disappeared for EL and MLn after 13 days of oxidation as examined by 1 H NMR,
which may be attributed to the addition of radicals to double bonds (propagation-like reaction) and the -scission of double-bond-containing
radicals into by-products. The number of double allylic group in the model compounds also showed a significant influence on the rate of
oxidation for emulsions after water evaporation.
2003 Elsevier B.V. All rights reserved.
Keywords: Cobalt drier; ATR-FTIR; Raman; Oxidation; Alkyd emulsion

1. Introduction
Currently for alkyd coatings, solvent-borne alkyd paints
still dominate the market, but due to more stringent environmental concerns, environmentally friendlier alternatives
have been sought after. Among them, alkyd emulsions have
shown a lot of success. However, not much has been published in terms of the oxidation of alkyd emulsions despite
numerous studies on the film formation and formulation of
alkyd emulsions [18]. One reason is the presence of water.
Most of the analytical techniques including FTIR fail to obtain sufficient chemical information when water is present.
A lot of efforts have been made to understand the oxidation of alkyd paints, but it is difficult to follow these
reactions in actual coatings. Model compounds for solventborne systems have been extensively investigated, including
methyl oleate (MO), ethyl linoleate (EL), methyl linolenate (MLn), methyl ricinoate (MR), and triglyceride oils;
Corresponding authors. Tel.: +31-40-2473066; fax: +31-40-2445619.
E-mail addresses: w.ming@tue.nl (W. Ming), r.van.der.linde@tue.nl
(R. van der Linde).

0300-9440/$ see front matter 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0300-9440(03)00119-X

in even simpler cases, dienes were also used such as 6,9pentadecadiene [9], 4,7-undecadiene [10], 3,6-nonadiene
[11], and 2,5-heptadiene [12]. Many techniques have been
used, such as FTIR [1316], NMR [17], Raman [1820],
SEC [13], MS [21], etc. The drying of alkyds usually involves evaporation of solvents (so-called physical drying,
when applicable) and oxidation (chemical drying) of unsaturated fatty acid ester side chains of alkyds. The oxidation
mechanism is illustrated in Scheme 1. The oxidation starts
by the abstraction of an allylic hydrogen atom from these
fatty acid ester side chains by existing radicals. The resulting
radical reacts with the molecular oxygen to form hydroperoxides, with the conjugation of double bonds taking place
simultaneously (Scheme 1(1.1)). These hydroperoxides are
decomposed into alkoxy (RO ) and peroxy (ROO ) radicals by metal-based catalysts (Scheme 1(1.2)). The radicals
may then recombine one another to form carboncarbon,
ether, and peroxy crosslinks that lead to the hardening of
alkyd paints (Scheme 1(1.3a)). The radicals can also initiate
the propagation of double bonds to form higher oligomers
(Scheme 1(1.3b)), especially when conjugated double bonds
are available. Besides these crosslinking reactions, radicals

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

81

Scheme 1. Oxidation mechanism for unsaturated fatty acids [1729].

may undergo -scission reaction to generate by-products,

such as aldehydes, ketones, epoxides, and carboxylic acids
[927]. An example given in Scheme 1(1.4) shows the
-scission of alkoxy radicals can also lead to the disappearance of double bonds.
Although similar oxidation reactions to solvent-borne
alkyds are believed to occur in water-borne alkyds, there
are some important differences that must be considered. It
is generally recognized that water-borne alkyds dry slower
than their solvent-borne counterparts probably due to the
presence of water. Water is one of the strongest complexing agents with many metal ions including Co3+ . The
formed [Co(H2 O)6 ]3+ hydrate complex will cause the loss

of catalytic capability for cobalt. Therefore, driers used for

solvent-borne alkyds may not be suitable for alkyd emulsions. In order to minimize this effect in alkyd emulsions,
driers that are emulsifiable in water and soluble in white
spirit were developed [28]. In addition, higher drier concentrations are needed to get a satisfactory drying performance
[14].
Investigations on the oxidation mechanism of alkyd emulsions with water-emulsifiable driers are still rare [14,29] and
the oxidation mechanism needs to be further clarified. The
effect of certain additives for water-borne alkyds (emulsifiers and thickeners, etc.) on the oxidation is not yet known.
It has been shown that the oxidation rate increases as the

82

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

number of double allylic group (between two double bonds)

increases for solvent-borne alkyds [13,30]. No study has yet
shown the effect of unsaturation number on the oxidation
rate in emulsions.
In this study, three model compounds in emulsions were
investigated towards a better understanding of the oxidation
mechanism and kinetics of alkyd emulsions in the presence
of emulsifiable cobalt driers and surfactants. The oxidation
of model compounds in emulsions during and after the water evaporation was followed by Raman spectroscopy, attenuated total reflectance (ATR)-FTIR, and 1 H NMR. Special
attention was paid to the possible effects of water and emulsifier on the oxidation.

2. Experimental
2.1. Materials
MO (6070%, the remainder: saturated esters), EL
(6570%, the remainder: ethyl oleate), MLn (7080%,
the remainder: methyl linoleate and saturated esters), and
n-pentanol were obtained from Sigma-Aldrich Chemie B.V.
(Zwijndrecht, The Netherlands). Model compounds may
contain very small amounts of hydroperoxides [15]. Sodium
dodecyl sulfonate (SDS) was obtained from Merck B.V.
(Amsterdam, The Netherlands). A cobalt drier (Nuodex
Web-Co 6, a water emulsifiable drier containing 6 wt.%
metal content) was obtained from Sasol Servo B.V. (Delden,
The Netherlands). All chemicals were used without further
purification.
2.2. Emulsion preparation
Model compounds of 20% were emulsified in water containing 1% (w/w) of SDS and 0.01% (w/w) n-pentanol by
dropwise addition of model compounds to water (o/w) using Branson type ultrasound sonifier for 10 min. After cooling down 0.07% (metal/oil, w/w) drier was added to the
prepared emulsion. Particle size was measured on a Coulter particle size analyzer (model LS 230) in order to check
the emulsion stability. Typically, the average particle size
was around 0.9 m, and the change in particle size over
a period of 2 days was negligible within the experimental
error.
2.3. Characterizations
The oxidation of model compound emulsions was in situ
followed by time-resolved ATR-FTIR on a Bio-Rad Excalibur FTS3000MX infrared spectrophotometer (30 scans per
spectrum with a resolution of 4 cm1 ) equipped with an ATR
diamond unit (Golden Gate), and by Raman spectroscopy
on a Dilor-Jobin Yvon-Horiba confocal dispersive Raman
spectrometer (633 nm excitation wavelength, 1000 m pin-

hole, and 8 mW laser intensity) with an Olympus MX40 microscope. The prepared emulsions were deposited on the diamond crystal of ATR unit for infrared analysis and on glass
slides for Raman analysis. The thickness of wet films was
kept around 200 m. Samples were withdrawn from films after the water evaporation and dissolved in CDCl3 (5 mg/ml)
for 1 H NMR analysis on a Varian 400 MHz spectrometer.

3. Results and discussion

3.1. Raman spectroscopy
Raman spectroscopy is a very useful technique for the
study of the oxidation behavior of model compound emulsions. It is quite advantageous to use Raman to monitor the
variation of double bonds during the oxidation reactions
since, in comparison with IR, double bonds demonstrate
stronger Raman absorptions [19]. The cis/trans-double
bond isomers (Scheme 1) can be easily identified. In addition, because water peak does not interfere with the major
peaks of the model compounds, Raman spectroscopy allows us to take a closer look at what takes place during the
water evaporation, even when the majority of water is still
present.
3.1.1. Oxidation of model compound emulsions during
water evaporation
It is generally accepted that the oxidation starts only after
the water evaporation for alkyd emulsions because the oxidation needs plenty of oxygen [14]. However, no direct experimental evidence has been shown yet. Fig. 1a shows the
Raman spectrum of an EL emulsion at the very beginning
of water evaporation. Although the broad water peak overlapped slightly with the CH stretching at around 2900 cm1 ,
the peak at 3012 cm1 (corresponding to the antisymmetric stretch of cis-C=CH) was clearly visible and could be
used for further analysis. A decrease in the intensity of
3012 cm1 peak is an indication of the starting point of an
oxidation reaction. Fig. 1b shows a plot of the 3012 cm1
peak intensity (normalized against CH2 -scissor deformation at 1443 cm1 ) for three model compound emulsions as
a function of time during the water evaporation. Obviously,
the 3012 cm1 peak intensity remained essentially the same
for all three model compounds during the observation period
(about 40 min). There was hardly any observable change for
other double bond related peaks of model compounds during the water evaporation. By using Raman it was clearly
shown that, for the cobalt drier, essentially no oxidation took
place during water evaporation.
In fact, water has to evaporate in order for the cobalt
drier to have a catalytic effect. Fig. 2 shows a typical weight
loss curve for EL emulsion during the water evaporation.
The weight loss experiment of a 20% EL emulsion was performed on a balance. Emulsion of 50 l (about 200 m in
thickness) was applied on a glass slide at 20 C and 50% of

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

83

1655
1443

3012

1265
water

500

1000

(a)

1500
2000
Wavenumber (cm-1)

Normalized Intensity

MLn

3000

3500

MO

1.2
0.8
0.4
0
0

(b)

EL

2500

10

20

30

40

Time (min)

Fig. 1. (a) Raman spectrum of EL emulsion at the very beginning of water evaporation, and (b) variation of intensities of antisymmetric cis-C=CH stretch
peak at 3012 cm1 (Raman) for model compound emulsions during water evaporation (normalized against CH2 -scissor deformation at 1443 cm1 ) as a
function of time under the influence of 0.07% Nuodex Web-Co.

humidity. The weight of the wet film was recorded during

the water evaporation. In this particular case, about 50 min
were needed to reach a completion of the water evaporation. As a matter of fact, a few parameters govern the evaporation rate of water even when the water amount is the

Fig. 2. Weight loss of 20% EL emulsion during water evaporation (at

20 C and 50% humidity with 200 m wet film thickness).

same, including temperature, humidity, and sample thickness. Since similar conditions were used for this weight
loss experiment with the Raman investigations, it can be
concluded that essentially no oxidation takes place during
the entire period of water evaporation of model compound
emulsions.
Water reduces oxygen diffusion and therefore retards the
oxidation process [14]. The oxidation can be accelerated as
soon as water evaporates. The oxidation reactions of model
compounds, even with effective driers, is a slow process.
Even after the water evaporation, some time is needed for
oxygen to react with unsaturated fatty acids. The efficiency
of catalyst also plays an important role for accelerating
these oxidation reactions. Different driers may demonstrate
different catalytic behavior. In the case of cobalt driers,
catalytic capability becomes effective only after the water
evaporation.
In the following, we will be focusing on the oxidation of
model compounds after the water evaporation. Raman peak
assignments are given in Table 1.

84

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

Table 1
Important Raman band assignments of model compounds [19,20,32]
Band
1265
1306
1320
1443
1597
1634

Assignment
w
w
w
s
w
s

1655 s
1675 s
1744 w
28552893 vssh
29262955 vssh
3012 s

Non-conjugated cis-C=CH symmetric rock

CH2 in phase methylene twist
Non-conjugated trans-C=CH symmetric rock
CH2 -scissor deformation; (CH2 )
Antisymmetric conjugated C=C stretching
Symmetric conjugated C=C stretching (cistrans
and transtrans) and/or terminal C=C stretching
Symmetric cis-C=C stretching (non-conjugated)
Isolated trans-C=C stretching
C=O stretching (ester)
CH2 stretching (CH2 asymmetric and CH3
asymmetric)
CH2 stretching (CH2 symmetric and CH3
symmetric)
Cis-C=CH antisymmetric stretching

3.1.2. Oxidation after the water evaporation

3.1.2.1. Oxidation of MO emulsion. Fig. 3 shows Raman spectra during the oxidation of the MO emulsion with
and without Nuodex Web-Co. When Nuodex Web-Co was
not added (Fig. 3a), no obvious change in Raman spectra
was detected up to 130 h, indicating that there was almost no oxidation for MO. However, in the presence of
Nuodex Web-Co observable changes could be identified
after 8 h (Fig. 3b), for instance, the decrease of the cis-C=C
peak (1655 cm1 ) and the cis-C=CH symmetric rock peak
(1265 cm1 ). The appearance of isolated trans-C=C peak
(1675 cm1 ) was also observed in the later stage. This is a
clear indication that cis-C=C was converted into trans-C=C
[31]. The appearance of a shoulder at 1634 cm1 was
noticeable after 8 h. Although this peak can be due to conjugated C=C and/or terminal C=C, the exact assignment
of this peak is still unclear. Since MO has only one double bond, the conjugation of double bonds is not possible
(MO does not have EL impurities, as checked by GCMS).
Although formation of terminal C=C was already reported
for EL [33], we do not have a direct evidence for the formation of terminal double bonds in the case of MO. After
130-h oxidation of MO, cis- and trans-C=C peaks were still
present due to the low oxidation rate of MO.
3.1.2.2. Oxidation of EL emulsion. Fig. 4 shows the Raman spectra during the oxidation of the EL emulsion with
and without Nuodex Web-Co. In the absence of the drier,
only very small changes were observed over a period of
130 h (Fig. 4a), for instance, the slight decrease of the
1265 cm1 peak. In contrast, in the presence of Nuodex
Web-Co, the changes of double bonds were very obvious
within 3 h (Fig. 4b). The non-conjugated cis-C=C peak
(1655 cm1 ) decreased a lot, due to the conversion to conjugated C=C including both cistrans and transtrans-C=C
(Scheme 1), which are difficult to be differentiated. The decrease of the cis-C=CH symmetric rock peak (1265 cm1 )

Fig. 3. Raman spectra during the oxidation of MO emulsion at 1200

1800 cm1 : (a) without Nuodex Web-Co and (b) with Nuodex Web-Co.

was also obvious. The 1634 cm1 peak may correspond to

the symmetric stretching of the conjugated double bonds,
and the 1597 cm1 peak to the antisymmetric stretching.
The intensity of the antisymmetric stretching is generally low in Raman [32]. After 24 h oxidation, a peak at
1675 cm1 (isolated trans-C=C) appeared due to the consumption of neighboring double bonds. During the oxidation, conjugated double bonds first increased in intensity,
until reaching a maximum level. Afterwards, the intensity
of the double bonds started decreasing due to the propagation (especially after the formation of conjugated double
bonds) and/or -scission (Scheme 1).
3.1.2.3. Oxidation of MLn emulsion. The Raman spectra
of MLn emulsion during the oxidation with and without
Nuodex Web-Co are shown in Fig. 5. As expected, the intensity of C=C absorption including stretch (15971680 cm1 )

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

85

Fig. 4. Raman spectra during the oxidation of EL emulsion at 12001800 cm1 : (a) without Nuodex Web-Co and (b) with Nuodex Web-Co.

and rock (1265 cm1 ) was much higher than EL and MO,
and CH2 scissor peak (1443 cm1 ) of alkyl chain was much
lower because of the presence of three non-conjugated double bonds. In the absence of Nuodex Web-Co, MLn demonstrated much faster oxidation than EL (Fig. 5a); MO showed
almost no oxidation within 130 h. The appearance of conjugated peaks could be seen within 24 h oxidation, as indicated by a shoulder at around 1634 cm1 . Within 130 h the
intensity of double bonds decreased substantially. The faster
oxidation is due to the presence of two double allylic groups
in MLn.

In the presence of Nuodex Web-Co, the same conversion

of the double bonds was reached within 24 h (Fig. 5b), instead of 130 h in the absence of the drier. MLn got oxidized
much more rapidly than MO and EL. The intensity of conjugated double bonds formed within 3 h was much higher
than that of EL during the same period. Increasing the number of double bonds in the fatty acid chain can increase the
oxidation rate of model compounds in emulsions. This trend
is similar to solvent-borne systems. Similar to EL, the intensity of conjugated and isolated C=C first increased and
then decreased due to the consumption of double bonds.

86

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

Fig. 5. Raman spectra during the oxidation of MLn emulsion at 12001800 cm1 : (a) without Nuodex Web-Co and (b) with Nuodex Web-Co.

3.2. ATR-FTIR
3.2.1. Oxidation of model compound emulsions during the
evaporation of deuterated water
Since the water peak overlapped with peaks of interest
in the model compounds, it was not possible to follow the
initial stages of oxidation by FTIR. Replacing H2 O with
D2 O could solve this problem. Fig. 6 gives IR spectra of
D2 O, H2 O, and EL; clearly it is possible to monitor the
3010 cm1 peak in the presence of D2 O.
Similar to the Raman study on the oxidation of model
compound emulsions described in Section 3.1.1, the first
50-min oxidation of the model compound D2 O emul-

sions was followed by ATR-FTIR. Fig. 7 gives the intensity of the 3010 cm1 peak (the symmetric stretching of
cis-C=CH) against an internal standard (OCO stretching at 1178 cm1 ) as a function of time. Obviously for
the three model compounds studied, essentially no change
in the 3010 cm1 peak intensity was observed. Therefore,
FTIR also confirmed that, for the cobalt drier, essentially
no oxidation took place when water is present.
3.2.2. Oxidation after water evaporation
Although water interfered with the peak of interest, water
evaporated almost completely after a certain period. After
the evaporation of water, the oxidation of model compounds

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

Absorbance (a.u.)

0.4

H2O

evaporation in the presence of Nuodex Web-Co over a period

of 50 h at 6501050 cm1 region. This region shows changes
in the wagging of cis- and trans-C=CH as reported previously with linseed oil [16]. The absorption at 660780 cm1
is probably due to the combination of CC alkane skeletal
vibration and the wagging of cis-C=CH. During the observation period, this peak became narrower due to the formation of isolated trans-C=C (970 cm1 ). As MO has only one
double bond, the conversion of cis- to trans-C=C was very
low within 50 h [31].

D2O

EL
0.2

3010

0
3500

3000

2500

2000

-1

Wavenumber (cm )

Fig. 6. Infrared spectra of H2 O, D2 O and EL (20003700 cm1 ).

Table 2
Important infrared band assignments of model compounds [16,19,32]
Band

Assignment

723
723 m
987 m
947 s and 987 s
970 s
1370 m
28552893 vssh

CC skeletal vibration of CH2 chain

Cis-C=CH planar bending
Transtrans conjugated C=CH wag
Cistrans conjugated C=CH wag
Isolated trans-C=CH wag
CH2 wagging
CH2 stretching (CH2 asymmetric
and CH3 asymmetric)
CH2 stretching (CH2 symmetric
and CH3 symmetric)
Cis-C=CH symmetric stretching
OH stretch (free-bonded)

29262955 vssh
3010 w
34513600 m (broad)

can be easily followed in situ by time-resolved ATR-FTIR.

We will demonstrate the oxidation of model compound
emulsions after the water evaporation. Assignments of
infrared peaks are given in Table 2.

Normalized Intentisty

3.2.2.1. Oxidation of MO emulsion. Fig. 8a shows infrared

spectra during the oxidation of MO emulsion after water

MLn
EL
MO

1.05

87

0.85
0.65

3.2.2.2. Oxidation of EL emulsion. Fig. 8b shows infrared

spectra (6501050 cm1 region) during the oxidation of the
EL emulsion in the presence of Nuodex Web-Co over a period of 50 h. The decrease of the cis-C=CH wagging peak
at 723 cm1 was more obvious within 50 h. At the very beginning of the oxidation, three small peaks at 947, 970, and
987 cm1 were shown (EL already contains a small amount
of hydroperoxides [15]). These three peaks corresponded to
CH wagging of either conjugated or isolated trans-C=CH
bonds. Among them, peaks at 947 and 987 cm1 were likely
due to conjugated double bonds while the peak at 970 cm1
was from isolated trans-double bonds (Table 2) [32]. As the
oxidation proceeded, the intensity of these peaks increased.
After reaching a maximum level, conjugated double bond
peaks at 947 and 987 cm1 started decreasing in intensity.
The isolated trans-C=CH (970 cm1 ) peak remained with
the highest intensity within the observation period. The results were in good agreement with Raman results described
in Section 3.1.2.
3.2.2.3. Oxidation of MLn emulsion. Fig. 8c shows infrared spectra during the oxidation of the MLn emulsion
with Nuodex Web-Co up to 50 h. The decrease of cis-C=CH
wagging peak was much faster than the other two model
compounds. In the region of 930995 cm1 , an increase of
peak intensity due to the formation of conjugated and isolated trans-double bonds, followed by a steady decrease of
conjugated double bonds, was again observed, very similar
to the oxidation of EL. The isolated trans-C=CH (970 cm1 )
wagging peak also remained with the highest intensity for
MLn within the observation period. The assignments of MLn
peaks during oxidation were more complicated than for EL
due to the third double bond that, together with the other
two, may form many isomers. The differentiation of these
isomers was not attempted.
3.3. 1 H NMR

0.45
0.25

10

20

30
time (min)

40

50

Fig. 7. Variation of intensities of symmetric cis-C=CH stretch peak at

3010 cm1 (ATR-FTIR) for model compound emulsions during D2 O
evaporation (normalized against OCO stretching at 1178 cm1 ) as a
function of time under the influence of 0.07% Nuodex Web-Co.

The oxidation of model compound emulsions in the presence of Nuodex Web-Co was also examined by 1 H NMR.
The intensity of C=CH (5.4 ppm) of 1-h and 13-day oxidized samples is shown in Fig. 9. Previously, Raman and
ATR-FTIR already showed the great loss of double bonds
over a period of 130-h oxidation. After the samples were oxidized for 13 days, >90% of double bonds disappeared for

88

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

Time (h)
1
3
10
15
20
50

Absorbance

0.08

50
20
15
10
3
1

0.06
Absorbance

0.10

970

0.08

723

0.06

0.04

0.04

0.02
0.02
640

680

720

760

800

930

960

990

-1

(a)

1020

-1

cm

cm

Time (h)
1
1.5
2.5
3.5
5.5
20
50

Absorbance

0.10

0.08

723

50
20
5.5
3.5
2.5
1.5
1

0.12
0.10

0.06

947

0.08
0.06

0.04

0.04
640

680

720

760

800

930

960

-1

(b)

723

0.10

970 987

0.20
0.18
0.16

947
50
20
3.5
3
2.5
1.5
1

0.14
Absorbance

0.12

1020

cm

1
1.5
2.5
3
3.5
20
50

0.14

990
-1

cm

Time (h)

Absorbance

970 987

0.14

Absorbance

0.12

0.08
0.06

0.12
0.10
0.08
0.06

0.04

0.04
0.02

0.02
640

(c)

680

720
-1

cm

760

800

930

960

990
cm

1020

-1

Fig. 8. ATR-FTIR spectra (6501050 cm1 ) during 50 h oxidation of model compound emulsions: (a) MO, (b) EL and (c) MLn.

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

MLn

EL

MO

89

1-h
13-day
6.0

5.7

5.4

5.1

4.8

4.5

6.0

5.7

5.4

ppm

5.1

4.8

4.5

6.0

5.7

5.4

5.1

4.8

4.5

ppm

ppm

Fig. 9. 1 H NMR intensity of C=CH (5.4 ppm) for 1-h and 13-day oxidized model compound (MO, EL, MLn) emulsions in the presence of Nuodex
Web-Co after water evaporation.

EL and MLn. But MO showed only a slight decease. It is already known that for MO hydrogen abstraction takes place
from mono-allylic groups, while hydrogen is abstracted from
double allylic groups in EL and MLn. The latter is energetically more favorable. From this point of view, MO oxidized
much slower than EL and MLn.
The above studies showed that, in a short period of time,
the intensity of all double bond related peaks for EL and
MLn decreased tremendously. The fast disappearance of
double bonds triggered us to further examine the oxidation mechanism. Muizebelt et al. [13,21,25] suggested that
for EL the recombination of radicals is the main route for
the formation of higher oligomers, while for MR the addition of radicals to double bonds plays a major role for
the consumption of the double bonds. Other reactions like
-scission were already known to decrease the unsaturation, but the loss of unsaturation would not be so great.
The recombination of free radicals does not affect the double bonds. In view of the enormous unsaturation loss in
this study, it appears that even for EL and MLn (containing only non-conjugated double bonds) the addition of free
radicals to double bonds (propagation-like reaction) also
plays an important role in the disappearance of the double bonds. It is known that, during/after the formation of
hydroperoxide, the non-conjugated double bonds convert
into conjugated bonds that render a structure similar to MR
(Scheme 1(1.3b)). Therefore, it is reasonable to attribute the
great loss of double bonds, even in the case of model compounds containing non-conjugated double bonds, to the direct addition of free radicals (R , RO , ROO ) to the double
bonds.

(3012 cm1 , Fig. 1a) by Raman. FTIR and Raman results

were in good agreement. Here we only demonstrated the
results on the basis of ATR-FTIR measurements.
The kinetic plots (natural logarithm of the peak height
(normalized values) versus time) of the oxidation of three
model compounds are shown in Fig. 11. The slopes indicated different oxidation rates. The oxidation rates of
model compounds were similar to the reported values for
solvent-borne/bulk systems [13]. Therefore, the oxidation
kinetics of model compound emulsions after water evaporation appeared to be similar to the solvent-borne/bulk
systems. The presence of SDS did not show any obvious
effect on the oxidation of model compounds. It can also be
seen in Fig. 11 that the oxidation rate of model compounds
in the presence of Nuodex Web-Co also increased with the
increasing number of allylic groups between two double
bonds. It means that, besides the drier capability, the unsaturation number of the alkyd resin in emulsion is also very
important to get an optimum drying performance.

H H H H
3010 cm-1 (cis-C=CH)
0h

1h
15 h

3.4. Oxidation kinetics of model compound emulsions

2800

2850

2900

2950

3000

3050

3100

-1

The oxidation kinetics of model compounds could be

followed in situ via either the decrease of the cis-C=CH
symmetric stretch (3010 cm1 ) by ATR-FTIR in Fig. 10 or
the decrease of the antisymmetric stretching of cis-C=CH

Wavenumber (cm )

Fig. 10. ATR-FTIR spectra of EL emulsion showing the variation of

cis-C=CH symmetric stretch peak during the oxidation under the influence
of Nuodex Web-Co.

90

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

y = -0.049x + 4.61
R2 = 0.99

4.5

Ln((A/A0)*100)

4
y = -0.37x + 4.92
R2 = 0.98

3.5
MO
EL
MLn

3
2.5
2

y = -0.55x + 4.91
R2 = 0.99

1.5
1
0

time(h)

Fig. 11. The oxidation kinetics of model compound emulsions obtained by the decrease of cis-C=CH symmetric stretch peak at 3010 cm1 (after water
evaporation) under the influence of Nuodex Web-Co (pseudo-first-order rates calculated from the graph are indicated).

4. Conclusions
Raman spectroscopy has been demonstrated to be a
suitable technique for the study of the oxidation of model
compounds (MO, EL, and MLn) in emulsions. From Raman and ATR-FTIR investigation, it has been clearly shown
that essentially no oxidation took place during the water
evaporation of three model compound emulsions. After the
water evaporation, the oxidation reactions started in a similar manner to the bulk/solvent-borne systems. In the case of
EL and MLn, as the oxidation proceeded, the cis-C=C converted into trans-C=C and conjugated C=C, and their intensities reached a maximum level. These double bonds started
to disappear afterwards, and conjugated double bonds decreased much faster than isolated trans-C=C. This is due to
the direct addition of radicals (R , RO , ROO ) to the double bonds (conjugated double bonds in particular; they were
formed during the oxidation). The chain propagation (like
step-growth polymerization) might be the main reason of
the great loss of double bonds. The number of double allylic
group in the model compounds also showed a significant influence on the rate of oxidation: the greater the unsaturation
number, the faster the oxidation. The kinetic study revealed
that the presence of water and surfactant in the emulsion
did not show observable effects on the oxidation of model
compounds. It can be concluded that the oxidation of model
compounds in emulsions takes place in a comparable way
to their counterparts in solvent-borne/bulk systems.

Acknowledgements
Authors would like to acknowledge the financial support from SENTER/Innovation Oriented Research Programme (IOP) on Heavy Metals (IOP Zware Metalen No.

IZW99241b), sponsored by the Netherlands Ministry of

Economic Affairs.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]

P.K. Weissenborn, et al., J. Coat. Technol. 72 (906) (2000) 6574.

G. Ostberg, M. Hulden, Prog. Org. Coat. 24 (1994) 281297.
R. Hurley, F. Buono, J. Coat. Technol. 54 (694) (1982) 5561.
P.K. Weissenborn, et al., Prog. Org. Coat. 40 (14) (2000) 253266.
E. Ciampi, et al., Langmuir 16 (3) (2000) 10571065.
C. Wang, F.N. Jones, J. Appl. Polym. Sci. 78 (9) (2000) 16981706.
J. Beetsma, Pigment Resin Technol. 27 (11) (1998) 1219.
J. Beetsma, A. Hofland, Surf. Coat. Int. 81 (10) (1998) 491.
R.A. Hancock, N.J. Leeves, et al., Prog. Org. Coat. 17 (3) (1989)
349358.
H.A.J. Carless, R.J. Batten, J. Chem. Soc., Perkin Trans. I (1987)
1999.
J.C. Hubert, R.A.M. Venderbosch, W.J. Muizebelt, R.P. Klaasen,
K.H. Zabel, Prog. Org. Coat. 31 (4) (1997) 331340.
E.N. Frankel, R.F. Garwood, J.R. Vinson, B.S.L. Weedon, J. Chem.
Soc., Perkin Trans. I (1982) 27072717.
W.J. Muizebelt, J.C. Hubert, M.W.F. Nielen, R.P. Klaasen, K.H.
Zabel, Prog. Org. Coat. 40 (14) (2000) 121130.
J.H. Hartsthon, J. Coat. Technol. 54 (687) (1982) 53.
S.T. Warzeska, R. van Gorkum, E. Bouwmann, et al., Prog. Org.
Coat. 44 (2002) 243248.
J. Mallgol, et al., Prog. Org. Coat. 39 (13) (2000) 107113.
W.J. Muizebelt, J.C. Hubert, R.A.M. Venderbosch, R.P. Klaasen,
K.H. Zabel, J. Coat. Technol. 69 (869) (1997) 5964.
J.E.D. Davies, et al., J. Chem. Soc., Perkin Trans. 2 (11) (1972)
15571561.
J.K. Agbenyega, M. Claybourn, G. Ellis, Spectochim. Acta A
47 (910) (1991) 13751388.
G. Ellis, M. Claybourn, Spectochim. Acta A 46 (2) (1990) 227241.
W.J. Muizebelt, W.F. Nielen, J. Mass. Spec. 31 (1996) 545554.
J. Mallgol, L. Gonon, et al., Prog. Org. Coat. 41 (13) (2001)
171176.
E.N. Frankel, W.E. Neff, E. Selke, D.D. Brooks, Lipids 22 (1987)
322327.
E.N. Frankel, W.E. Neff, E. Selke, D.D. Brooks, Lipids 23 (1988)
295298.

Z. Okan Oyman et al. / Progress in Organic Coatings 48 (2003) 8091

[25] W.J. Muizebelt, et al., J. Coat. Technol. 70 (876) (1998) 8393.
[26] N.A. Porter, Acc. Chem. Res. 19 (1986) 262.
[27] N.A. Porter, S.E. Caldwell, K.A. Mills, Lipids 30 (4) (1995) 277
290.
[28] J. Bieleman, Additives for Coatings, Wiley/VCH, Weinheim,
2000.
[29] J. Mallgol, et al., Polym. Mater. Sci. Eng. 85 (2001) 412413.

91

[30] V.A. Roginskii, Kinet. Catal. 31 (3) (1990) 475481.

[31] N.A. Porter, D.G. Wujek, J. Am. Chem. Soc. 106 (1984) 2626.
[32] B. Colthup, et al., The Handbook of Infrared and Raman
Characteristic Frequencies of Organic Molecules, Academic Press,
New York, 1993, p. 82.
[33] W.J. Muizebelt, J.W. van Velde, F.G.H. van Wijk, Adv. Org. Coat.
Sci. Technol. Ser. 13 (1991) 57.