3.1. Performance in Solvent-Borne Alkyd Binders
The initial tests of drying activity was performed on four solvent-borne alkyd resins of medium (
S471 and
SP262) and long oil length (
S622 and
S252); all of them are commercial binders modified with semidrying soybean oil. Although manganese(III) acetylacetonate (Mn) is a dark-brown powder solid insoluble in organic solvents commonly used in paint-producing industry, the solubility problems were easily overcome by the use of dimethyl sulfoxide (DMSO) that was recently recognized as a promising candidate for replacement of volatile organic solvents in solvent-borne paints. It is a low volatile solvent classified as nontoxic with no risk for the human health [
28].
The activity of Mn was investigated on test coatings in the concentration range 0.1–0.01 wt.% of metal in dry matter content of alkyd binder. Set-to-touch time (τ
1), tack-free time (τ
2), dry-hard time (τ
3) and dry-through time (τ
4), estimated by Beck–Koller method, are summarized for formulations of
S471 and
SP262 in
Table 1 together with relative hardness determined by pendulum test 10 and 100 days after application. In the binder
S471, the title compound is highly active in whole concentration range, as the dry-hard time does not exceeds 9 h. Typical feature of the binder is a fast film formation due to absence of low volatile solvents in the binder. Therefore, according to the Beck–Koller method, the coatings can be considered as set-to-touch dry in few minutes after application on test substrate even in the absence of a drier. It should be noted that similar behavior was observed for resin
S622. Tack-free time of the formulations Mn/
S471 increases with decreasing concentration. However, similar dependency was not observed for parameters τ
2, τ
3, and τ
4, which is probably a result of better through drying of formulations with the lower drier content. The main advantage of Mn is unusually high activity at a considerably lower concentration than recommended for cobalt(II) 2-ethylhexanoate (Co). Therefore, Mn shows satisfactory performance in
S471 at concentration 0.01 wt.%, whereas Co is almost inactive at the same level as evident from high values of τ
2 (18.1 h). Furthermore, the coatings containing this commercial drier are not dry-through within 24 h.
Additional experiments on commercial manganese compounds revealed that Mn provides shorter tack-free times than manganese 2-ethylhexanoate, but their overall performance is similar as their dry-hard times and try-through times are comparable. Low activity of manganese(II) acetate tetrahydrate, predissolved in DMSO, is attributed to presence of aqua ligands in the coordination sphere of manganese, which considerably prolongs initial phase of the autoxidation process (
Table S1 in Supplementary Materials).
Promising drying activity of Mn was further evidenced in alkyd resin SP262. Therefore, dry-through time (τ4) shorter than 10 h is observed at the concentration range 0.1 to 0.03 wt.%. Commercial Co exhibits, at this concentration range, τ3 longer than 9 h and τ4 longer than 15 h.
Relative hardness of test coatings of Mn/
S471 and Mn/
SP262, measured 10 days after application (
Hrel;10d), varies around 20%, which is ~5% points lower value than observed for coatings cured with Co. Final relative hardness of the coatings, treated with Mn, varies around 35%. The coatings, cured with Co, are considerably harder due to higher density of cross-linking, which is probably caused by a higher long-term stability of the catalytically active species. Note that the lower final hardness is a common feature of the alkyd coatings treated with cobalt alternatives [
29]. Furthermore, the long-term stability of the active species, originating from cobalt carboxylates, has also some negative aspects, such as a tendency to early embrittlement and aging as the polymer degradation processes, which are also based on autoxidation and can be promoted by primary driers [
30].
The results of mechanical tests, performed on formulation of
S622 and
SP252, are summarized in
Table 2. In both alkyd resins, manganese-based drier performs well at concentration range 0.1 to 0.01 wt.% giving shorter τ
3 and τ
4 than appropriate coatings treated with Co. Considerably better drying performance is evidenced mainly at low metal concentration where formulations Co
/S262 and Co
/SP252 give very long touch-free times (τ
2).
The cured films of alkyd resins S622 and SP252 show generally a lower final hardness than aforementioned resins S471 and SP262. Such an effect occurs because of a different composition of the binder, due to a lower content of aromatic dibasic acid in the resins of long oil-length. Furthermore, the higher content of fatty acid chains results in a more effective film plasticization. The coatings cured with Mn show a lower final hardness then those treated with Co, which agrees with the trend observed for alkyds of medium oil length. It should be noted that dependence on the drier concentration is here more distinct due to a lower contribution of the phenylene building blocks on the polymer hardness.
Application of manganese-based driers in paint formulations can be limited with their dark coloration, which affects manly colorless lacquers and light-pigmented enamels [
31]. Although freshly prepared formulations of Mn are dark-brown colored at high concentrations, storage in closed vessels results in their discoloration in few days as documented on formulations of
S471 shown in
Figure 1a. Such color change is due to reduction of Mn
III to Mn
II, as proved by electron paramagnetic resonance (EPR). EPR spectrum of a sample measured after storage in a closed vessel for ten days exhibits a typical six-line pattern due to hyperfine interaction of unpaired electron with nucleus of
55Mn (
I = 5/2), which natural abundance is 100% (
Figure 1b). Calibration on freshly prepared formulation of manganese(II) acetate tetrahydrate in
S471 enabled to quantify amount of Mn
II in formulations of Mn/
S471. Freshly prepared sample of metal concentration 0.1 wt.% contains about 12% of Mn
II; 50% and 90% conversion of Mn
III to Mn
II was observed after one-day and seven-day storage, respectively. Our experiment with formulation of Mn/
S471 (0.03 wt.%), stored for one weak under inert atmosphere of nitrogen, revealed prolongation of tack-free time up to 2.7 h but overall performance seems to be similar to freshly prepared formulation as dry-hard time and dry-through time stay comparable.
Coloration of transparent paint films treated with Mn and Co was evaluated by spectroscopy in visible spectral region on formulations of
S471. The measurements were done on coatings of 120-μm wet thickness in transmission mode in concentration range 0.1 to 0.01 wt.% of metal in dry matter content. We note that the coatings were cast from freshly prepared formulations and their color was evaluated 3 and 60 days after application. The colorimetric data, summarized in
Table 3, revealed a strong chromatic shift to yellow (see parameter b*) only for formulation of Mn/
S471 at metal concentration 0.1 wt.%. In the range 0.06–0.03 wt.%, Mn shows acceptable values of b* as they near the values obtained for formulations of Co/
S471 at commonly used concentrations (0.1–0.06% wt.%). Negligible effect on color was observed at 0.01 wt.% of Mn. Its yellowish color is mainly due coloration of the alkyd binder itself.
Note that the treatment of S471 with Co also causes chromatic shift of the transparent coatings to greenish yellow, even though the drier solutions in inert solvents are purple. Surprisingly, the cobalt-based drier does not compensate yellowish appearance of the binder, as one could suggest, probably due to a redox process or chemical changes in the coordination sphere of cobalt.
Only minor color changes were observed on samples of Mn
/S471 and Co
/S471 stored on a diffuse daylight for 60 days. They are attributed to yellowing of the alkyd binder upon storage on the diffuse daylight. We note that coloration of coatings (
S471) treated with manganese(II) 2-ethylhexanoate is very similar to Mn
/S471 (
Table S2 in Supplementary Materials). Only at concentration 0.1 wt.%, it shows lower yellowing index (b* = 0.70).
3.2. Performance in High-Solid Alkyd Binders
Promising activity of the title compound in solvent-borne alkyd formulations led us to investigate its drying performance in high-solid formulations. This type of binder is currently preferred due to a low content of volatile organic compounds, but is generally more sensitive to a proper choice of the primary drier owing to negligible effect of solvent evaporation on the film formation process. The tests were performed on alkyd resins
FP07 (solid content = 89%),
SP00 (solid content = 99.5%) and
TI870 (solid content = 98.9%). For our purposes, the latter two binders were diluted with dearomatized white spirit to 90% solid content in order to improve their film-forming properties. Drying times, estimated on test coatings of 38 and 76 µm wet thickness, are summarized in
Table 4. Two sets of experiments were conducted since the high-solid binders are generally cured less homogenously due to limited diffusion of air-oxygen. The chemical curing of high-solid binders has to be faster than in the case of solvent-borne formulations in order to reach satisfactory drying times [
32].
The binder FP07 is well cured with Mn in the concentration range 0.1 to 0.01 wt.% as evident from dry-through times (τ4) of the 76 µm films those are used for standard tests of drying activity. Optimal performance is observed at 0.03 wt.% and only mild overdose effect is evidenced at higher concentrations. The 76 µm layers are cure almost homogenously as evident from comparison with 38-µm films. The cobalt based drier Co shows optimal performance 0.06 wt.%, where 76 µm coatings become through dry within 17.4 h. At higher or lower concentrations, the 76 µm coatings are not through dried within 24 h. Considerably lower values of τ4 were obtained for 38 µm films in concentration range 0.1–0.03 wt.% (7.1–9.9 h). This observation reveals a front formation and a limited oxygen penetration into the whole volume of the 76 µm alkyd layers.
Coatings of the binder SP00 are well through-dried by Mn in the concentration 0.1–0.03 wt.% as the τ4 values of 76 µm films do not exceed 11 h. Cobalt based drier Co shows a pure through drying only at high concentrations (0.1–0.06 wt.%). Optimal dosage is observed at 0.03 wt.% where through drying is satisfactory and values of τ1–τ3 still very low.
The coatings of TI870 treated with Mn and Co cannot be considered as through-dried after 24 h of curing, as they stay soft for a long time. Therefore, dry-hard time (τ3) is used for comparison of a driers performance. The binder TI870 shows a very strong tendency to front-formation, as evidenced on coatings treated with Co (cf., τ3 for 38 and 76 µm films at 0.1 and 0.06 wt.%). In case of Mn, smaller differences in τ3 of 38 and 76 µm films is observed. Nevertheless, at high concentrations (0.1–0.06 wt.%), they are considerably higher than in case of the other high-solid binders under the study.
Relative hardness of the test coatings treated with Mn and Co are summarized in
Table 5. As expected, the values are considerably lower than observed for solvent-borne systems, which is given by lower molecular weights of the fresh binders. Films of the high-solid binders, treated with Mn, show sufficient hardness already after 10 days of curing. The values of
Hrel;10d is comparable with appropriate systems treated with Co (
Table 5). Although the hardness development continues in both cases, the systems containing Co show higher final hardness (
Hrel;100d). The discrepancy is much stronger in case of the binder
TI870, which is probably caused by different fatty acid pattern.
TI870 is produced from tall oil fatty acids while the other alkyd resins, under the study, are modified with vegetable oils or soybean oil.
3.3. Kinetics of Autoxidation Process in Alkyd Film
The curing process in the alkyd resin
S471 was followed by time-resolved infrared spectroscopy. Such experimental method enables to follow autoxidation of unsaturated fatty acid chains of the alkyd resin through development of characteristic bands at 3008 and 989 cm
−1 (
Figure 2 and
Figure 3). The band at 3008 cm
−1 is assigned to stretching of C–H groups on the isolated double bonds in
cis-configuration [ν
a(
cis-C=
C–H)]. During the autoxidation process, the band decreases in intensity as peroxidation proceeds and isolated double bonds are converted to conjugated system of double bonds [
1,
6]. The appearance of conjugated double bonds can be followed on band at 989 cm
−1 that is assigned to C–H wagging [ω(
cis-trans-C=
C–H)] [
6].
The C–H stretching band at 3008 cm
−1 is very suitable for the investigation of autoxidation kinetics as it follows consumption of the active substrate (
Figure 2a). Based on our previous studies [
6,
30], the autoxidation of a thin layer of alkyd resins behaves as a reaction of pseudo-first-order until ~50% conversion, which is related with sufficient mobility of the system. In later phases of the process, the system becomes solid and the autoxidation slows down more rapidly than the pseudo-first order reactions as it no longer satisfy the rule of a stirred liquid, which becomes evident as deviation from linearity in the logarithmic plots.
Our kinetic experiments were performed on thin films of the test formulations, which enable the elimination of the effect of air-oxygen diffusion. The estimated rate coefficients (
kmax), induction times (
tind) and reaction half-life (
t1/2) for formulations
S471/Mn are summarized in
Table 6. In the concentration range 0.1–0.01 wt.%, the rate coefficient increases with increasing concentration, as evident from slopes of the time-development of [ν
a(
cis-C=
C–H)] in the logarithmic scale, whereas induction time decreases (
Figure 2b). The values of
kmax are enough high and comparable with Co. High curing power at low concentration are apparently caused by short induction times. Therefore, at 0.01 wt.%, Mn shows still satisfactory value of
tind (2.2 h) while initiation of the autoxidation by Co is tedious at this concentration (
tind = 6.9 h). The estimated half-lives of the autoxidation well correlate with the drying times established by mechanical tests. Nevertheless, the overdose effect mentioned at 0.1 wt.% was not captured by kinetic measurements owing to the use of thinner alkyd layers. Such discrepancy supports the aforementioned explanation that the overdose is caused by thin polymeric film on the coating surface that restricts the oxygen diffusion into whole volume of the coating.
Development of the bands at 989 cm
−1 evidences appearance of the conjugated double bond systems upon the autoxidation. The band reaches the maximum, at
tconj, and then a decrease in intensity is observed (
Figure 3). It is due to high reactivity of the conjugated double bonds and their participation in alkyd cross-linking [
1]. The value of
tconj increases with decreased concentration and further documents a relatively fast process even at metal content 0.01 wt.%.
3.4. Thickness Effect
Effect of film thickness on rate of the autoxidation process was estimated by attenuated total reflectance infrared spectroscopy (ATR-IR). As recently demonstrated on high-solid formulations, the ATR sampling method enables to follow the autoxidation process on the interface coating/ATR crystal. Series of the measurements on samples of different film thickness is very suitable to approach curing process in the different parts of a thick alkyd layer and to distinguish inhomogeneity caused by limited air-oxygen diffusion [
32].
Thickness effect was evaluated in formulation of Mn/S471 at metal concentration 0.03 wt.%, which was chosen based on optimal performance in mechanical tests and aforementioned kinetic experiments. Only one formulation was chosen for the scrutiny, as the evaluation by ATR-IR method is highly time-consuming.
Development of the characteristic band of ν
a(
cis-C=
C–H) in time, measured for coatings of 5, 25, 50, 100, 150, and 200 µm wet thickness is given in
Figure 4. The 25 and 50 µm coatings show a very similar development to the coating of 5 µm wet thickness, which was used for aforementioned kinetic measurements. It implies homogenous film drying and negligible effect of the air-oxygen diffusion in these films. In the case of 100 µm coating, a strong deceleration of the autoxidation is observed at ~40% conversion (
t ≈ 1.5 h). Nevertheless, the process reaccelerates after next ~1.5 h, which is a typical feature of the front-forming drying [
32]. Coatings of 150 and 200 µm wet thickness are not through dried within 24 h as evident from a negligible consumption of the isolated double bonds on the interface sample/ATR crystal. Such observation reveals a very slow movement of the front towards the bottom and its low permeability for the air-oxygen.