Академический Документы
Профессиональный Документы
Культура Документы
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
81
82
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.
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.
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
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
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.
86
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
Absorbance (a.u.)
0.4
H2O
D2O
EL
0.2
3010
0
3500
3000
2500
2000
-1
Wavenumber (cm )
Assignment
723
723 m
987 m
947 s and 987 s
970 s
1370 m
28552893 vssh
29262955 vssh
3010 w
34513600 m (broad)
Normalized Intentisty
MLn
EL
MO
1.05
87
0.85
0.65
0.45
0.25
10
20
30
time (min)
40
50
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
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.
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.
H H H H
3010 cm-1 (cis-C=CH)
0h
1h
15 h
2850
2900
2950
3000
3050
3100
-1
Wavenumber (cm )
90
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.
91