By Cynthia Templeman, Senior Engineer, Toyota Motor North America Research & Development
Automotive paint shops are typically among those using the most energy-intensive processes in a manufacturing plant. Studies by Argonne National Lab (ANL/ESD/10-6), Michigan Technological University (Automobile Assembly Painting: Energy and Environmental Impacts [John Sutherland]) and others estimate that 50 to 75 percent of the energy consumption and carbon dioxide emissions from “material transformations” for an assembly plant are directly related to paint, particularly for the painting, HVAC and heating. Besides energy and carbon dioxide emissions, paint shops also historically have used high amounts of organic solvents.
Toyota has systematically reduced the organic solvents used in its paint shops over the last 15 years by implementing 3-wet waterborne systems. However, even the most advanced waterborne systems still contain some organic solvents. Therefore, we continue to investigate alternative technologies to further reduce the environmental footprint. Figure 1 shows the reduction in organic solvent usage since 1999, as well as the potential for one such alternative technology – UV curing.
To investigate the potential of UV curing for automotive coatings, Toyota partnered with Dr. Alec Scranton and his group at the University of Iowa, particularly Beth Ficek Rundlett and Cindy Hoppe. Together, we studied how cationic UV curing could offer some benefits over the more common free radical chemistry. This work is covered by several US patents: US 8197911, 8993042 and 9274429.
Numerous studies were performed to probe the potential for cationic UV curing to be used for automotive applications. Experiments were designed to test and attempt to overcome the various drawbacks of the system. Materials were purchased and used as received.
Monomers and oligomers used: 3,4-epoxycyclohexylmethyl-3,4-epoxycyclohexane carboxylate, or “CDE” [Dow, Rahn, Synasia]; UVR6128 bis(3,4-epoxycyclohexylmethyl) adipate [Dow]; methyl 3,4-epoxycyclohexanecarboxylate [Dow]; 2-(3,4-epoxycyclohexyl)ethyltrimethoxysilane, or “ECHETS” [Sigma-Aldrich]; 3-ethyl-3-hydroxymetyloxetane [Toagosei]; Capa 3031 e-polycaprolactone [Perstorp]; limonene dioxide [Arkema]; 2-butoxymethyl-oxirane, or “BMO” [Sigma-Aldrich]; EPON 826 BPA resin [Hexion]; EPON 828 BPA resin [Hexion]; EPON 862 BPF resin [Hexion]; bis[2-(3,4-(epoxycyclohexyl)ethyl)]tetramethyldisiloxane [Sigma-Aldrich]; Vikoflex 7190 epoxidized linseed oil [Arkema]; SILMER EPC DI-50 vinyl cyclohexyl monoxide modified siloxane [Siltech]; SILMER EPC J10 vinyl cyclohexyl monoxide modified siloxane [Siltech]; Lite 2513 HP epoxidized cashew nutshell liquid [Cardolite].
Photoinitiators used: UVI6976 triarylsulfonium hexafluoroantimonate salts [Dow]; Irgacure 250 (4-methylphenyl)[4-(2-methylpropyl) phenyl]-iodonium hexafluorophosphate [Ciba]; (tolylcumyl) iodonium tetrakis (pentafluorophenyl) borate, or “IPB” [Rhodia]; diaryliodonium hexafluoroantimonate, or “IHA” [Sartomer].
Pigments and additives used: Monarch 880 carbon black [Cabot]; Monarch 1300 carbon black [Cabot]; TS6200 titanium dioxide [DuPont]; Tinuvin 123 decanedioic acid, bis(2,2,6,6-tetramethyl-1-(octyloxy)-4-piperidinyl) ester, reaction products with 1,1-dimethylethylhydroperoxide and octane hindered amine light stabilizer [Ciba]; Tinuvin 384 octyl 3-[3-(benzotriazol-2-yl)-5-tert-butyl-4-hydroxyphenyl]propanoate UV absorber [Ciba]; Lignostab 1198 4-hydroxy-2,2,6,6-tetramethylpiperidinoxyl radical light stabilizer [Ciba].
Illumination was done using either a 200 W Oriel mercury-xenon arc lamp with a wavelength range of 250 to 700 nm and typical intensity of 56 mW/cm2, or a by way of a Fusion UV Conveyor with an I-600 10-inch lamp equipped with an H-bulb. Table 1 shows the typical radiometry for the conveyor as measured by an EIT PowerPuck.
Test methods employed include both standard and proprietary protocols for determining the extent of curing that occurs in regions of a sample that were not exposed to UV radiation. They will be described as much as possible in conjunction with the results presented. In addition, differential scanning calorimetry was used to assess the activity of photoinitiators as well as the effect of different pigments on curing. Knoop Hardness was measured using a Tukon Microhardness tester with 25 gf x 11 seconds indentation. Visual and tactile assessments of samples and pencil hardness testing also were employed. Thickness was measured using a BYK Micro-TRI-Gloss µ or Fisherscope MMS.
Results and Discussion
Experiments were designed to understand and attempt to overcome drawbacks with respect to shadow zones and the incorporation of species that inhibit the cationic curing reaction but are necessary for consideration of automotive coatings.
Initial studies into shadow zone curing, done by the University of Iowa, involved illuminating cuvettes containing different monomers and photoinitiators from the bottom for a fixed time and measuring the amount of polymer grown “upwards” over time in the dark (after illumination ceased). Iowa calculated the distance to which the light penetrated as the “initial condition.” Figure 2 shows one example result using this test method.
Additional studies done with different materials and parameters showed that shadow zone curing follows Ficks second law of diffusion:
where x=shadow cure distance, D=diffusion coefficient, t=time
In addition, we found that the diffusion constant depends on both monomer diffusion into the polymer matrix and reactive diffusion, which depends upon the rate of propagation. Since diffusion and propagation rates generally increase with increasing temperature, we investigated the effect of temperature on diffusion into the shadow region of samples. Figure 3 confirms the temperature dependence of the shadow cure.
We also investigated the effect of formulation components, such as photoinitiator counter-ions. Figure 4 shows the shadow cure results for two systems with the same monomer and curing conditions. The two photoinitiators have similar structures and photolysis yields, but differ significantly in the size of the counter-ion. IPB’s counter-ion (tetrakis pentafluorophenylborate) is much larger than IHA’s (hexafluoroantimonate). It is known in the art that larger separation between the ion pair results in increased propagation rates due to increased mobility of the active cation. In our studies, we confirmed this effect: the system with the larger counter-ion (IPB) shows improved shadow cure results over the other (IHA).
From the results in Figures 2 through 4, it can be seen that polymerization can proceed to significant distances into a shadow zone under optimized conditions. These initial studies were done using a stationary lamp and monomer samples in cuvettes. Subsequent studies sought to advance understanding of the potential for shadow cure in more complex systems. Figure 5 shows the shadow-cured distance into a masked region using a UV conveyor with an H-bulb for illumination.
The data from Figure 5 confirm that significant shadow curing can occur under the right conditions. A follow-up experiment was conducted using different materials and a fixed line speed of 20 ft/min. Based on the data in Figure 5, the extent of shadow cure was expected to be around 27 mm if the materials were the same. Figure 6 shows the results of the follow-up experiment.
From the data shown in Figure 6, we can see that the majority of shadow cure occurred during the first 30 minutes of the test. We also can see that the bis-adipate, which has a similar structure to CDE but with a longer chain between the cycloaliphatic groups, showed slightly improved “final” shadow cure distance compared to the CDE alone or the CDE/BMO blend shown in Figure 5. This is likely due to the increased flexibility of the adipate compared to CDE and the more reactive epoxycyclohexyl groups compared to the glycidal groups on the BMO. Again, we can surmise that with the optimized conditions and materials, sufficient shadow cure is possible to polymerize regions of a coating that were in shadow during illumination.
UV-absorbing or -blocking species
To investigate the potential to UV cure and UV “shadow cure” coatings that contain species that are more commonly found in automotive coatings, we incorporated different pigments and additives into model systems to gauge their effects. We selected a generic titanium dioxide pigment, as well as Tinuvin 384 UV absorber and hindered amine light stabilizers Tinuvin 123 and Lignostab 1198. Lignostab 1198 is not an automotive product but was tested as a model compound. Table 2 shows the results for curing pigmented films with variable concentrations of titanium dioxide using different illumination conditions. This testing was done on flat panels, and the checks for tackiness, hardness and thickness were done in the illuminated region (not a shadow region).
The results in Table 2 show that at low concentrations (1 to 3 percent by weight) the titanium dioxide did not inhibit the curing of the coating. Films were tack-free within the illumination time, and all had a high 3H hardness. With 6 percent by weight loading, tack-free times were longer, at 30 minutes, but the final films were equally hard as the films with lower loading. When exposure time is reduced for a fixed titanium dioxide concentration, tack-free times increase, as expected, but the films achieve higher final hardness. But allowing more of the polymerization to occur “in the dark,” it is surmised that less trapping of active centers occurs, allowing for a higher conversion and, therefore, an increased hardness (CDE monomer forms hard, rigid films).
The data in Table 3 show that curing films containing additives to improve weathering performance, such as UV absorbers and hindered amine light stabilizers (HALS), is more challenging than films containing, for example, titanium dioxide. By nature, the UV absorbers compete with the photoinitiator during the illumination time. However, despite this competition, the film with Tinuvin 384 UV absorber still achieved a tack-free surface within the illumination time and resulted in a hard film. However, incorporating HALS seems to have a bigger effect on the ability of the film to cure. With Tinuvin 123, we can see considerable inhibition – with 96 hours being required to reach a tack-free surface but the final film has similar hardness to those containing Tinuvin 384 or titanium dioxide. This suggests that, although the reaction rate is slower, final conversion is about the same. The film containing Lignostab 1198 did not cure at all.
As shown in Figure 7, Tinuvin 384 is a benzotriazole-based UV absorber and, upon excitation with UV irradiation, the phenolic hydrogen is abstracted and adds to the adjacent nitrogen. Therefore, the nitrogen is not available to interfere directly with the propagating active center, and curing is able to proceed normally. The Lignostab 1198 is a radical form of a piperidine derivative. When strong acids, such as a cationic active center, are present, the acid interacts directly with the nitrogen. This effectively prevents the active center from propagating. Therefore, as we saw in Table 3, the film containing Lignostab 1198 did not cure. Tinuvin 123 also is a piperidine derivative; however, it has a large hydrocarbon chain that provides steric hindrance to the nitrogen. Therefore, the cationic acid center has less interaction with the nitrogen than in the case of Lignostab 1198. For the film containing Tinuvin 123, we saw significant inhibition, but ultimately the film cured to the same hardness as previously tested films.
New methods for overcoming drawbacks
We considered several methods to overcome some of the drawbacks of UV curing, considering usage for automotive coatings. One method we developed avoids competition between the photoinitiator and other potential UV-absorbing species. Some examples of this method, called “multi-layer curing,” are shown in Figure 8.
From the examples in Figure 8, we can see that the cationic active centers have sufficient mobility to diffuse into layers of coating that were separately applied and did not contain photoinitiator. Although the “gouge-free” time was not recorded for the cases shown in Figures 8a and 8b, from 8c we can see that this propagation occurs within a reasonable amount of time.
Another method that was developed to overcome drawbacks of UV curing is termed “pre-activation of the photoinitiator.” In this method, photoinitiator is dissolved in an inert solvent, such as propylene carbonate, and that solution is then exposed to UV irradiation. Since there is no monomer present, no polymerization occurs in the solution. A portion of the pre-activated solution is then added to a monomer solution and mixed using, for example, a two-component paint system. The active centers become intimately mixed with the monomer and therefore, upon application to a substrate, the film does not need to be illuminated. This avoids the issue of shadow zones.
From the results in Figure 9, we can see that the pre-activated photoinitiator remains active for extended periods of time after illumination. This could potentially mean that illumination in an assembly shop may not even be necessary. Another concept for pre-activating the photoinitiator would be to illuminate it in the piping, before it is mixed with the monomer and other components, before application. This also would eliminate the need to illuminate the film after application. One consideration for this method is the necessity to control humidity. The University of Iowa found that spraying the pre-activated solution over a monomer film did not result in polymerization as expected, based on the results in Figure 9. It is hypothesized that the active centers were inactivated by the high humidity (65 to 70 percent RH) in the lab during atomization.
Inhibition by Alkaline Species
We sought to understand more about the effect of humidity and alkaline species on cationic UV curing. In one experiment, we attempted to block the humidity from entering into a film by incorporating hydrophobic monomers, in a method similar to that published by Dr. Zhigang Chen (formerly of North Dakota State University) in RadTech 2008 proceedings. We prepared two panels each of five different formulations, cured them and then stored the panels in either a desiccator or in a controlled humidity environment (20°C, 50% RH) for one week. Knoop hardness was measured on the panels after one week of storage. The results are shown in Figure 10.
From the results in Figure 10, we can see that some of the reactive diluents added to the CDE seemed to have a humidity blocking effect. Samples B and E showed similar hardness after both high and low humidity storage conditions. Samples A, C and D all resulted in untestable, peeled films from the high humidity conditions. So, we can understand that some components can be added to a formulation to make it more resistant to humidity.
We also studied the method of using the humidity within the film. Many groups are incorporating sol-gel precursors into UV-curable formulations as a way to add a means of crosslinking in shadow zones, or possibly to add new functionality to materials. We looked at these materials as a means of overcoming inhibition by humidity. A series of formulations were prepared with varying amounts of the sol-gel precursor (SGP) ECHETS, and films were monitored for hardness initially and after one week. Again, we stored films in either a desiccator or in 50 percent RH controlled environment. Figure 11 shows the results.
From the data in Figure 11, we can see that the sol-gel precursor effectively protects the film from inhibition due to humidity. There is no difference in hardness seen between samples that were stored in high humidity compared to samples stored in low humidity.
Finally, we studied the effect of pigment surface treatment on cationic UV curing. Two carbon black pigments (Monarch 880 and Monarch 1300) were incorporated into formulations. One of these has an acidic surface treatment, and the other has an alkaline surface treatment. We used photo-DSC (TA Instruments Q100) to grasp their effect on curing. The results are shown in Figure 12.
From the results in Figure 12, we can see that both carbon black pigments have an effect on peak maximum time (longer induction times than the formulation without pigment), but the alkaline pigment has almost twice as long peak maximum time compared to the acidic pigment. Similarly, from the total energy evolved, we can see that the acidic pigment formulation evolves almost the same amount of energy as the formulation without pigment, indicating that the conversion is similar in these systems. However, the formulation with the alkaline pigment evolves less than half of the total energy compared to the other two systems. This indicates that conversion is much less in that formulation.
We presented our findings related to cationic UV curing with the consideration of automotive coatings. UV curing presents an opportunity for large reductions in both process length and energy consumption in an automotive paint shop. In some formulations, organic solvent usage can be significantly reduced or even eliminated.
We found that the long-lived active centers through which polymerization occurs in cationic UV systems have the potential to cure significant distances into regions that were not previously illuminated with UV light, particularly if the formulation is optimized. We also found that typically additives for improving weathering resistance of coatings can hinder the cationic polymerization process, so care must be taken during formulation to select the most appropriate materials. We presented some new methods that we explored to overcome some of the drawbacks of UV curing, such as employing multi-layer curing or pre-activating the photoinitiator. Finally, we investigated the effect of humidity or other alkaline species on cationic UV curing and presented some approaches to mitigate the inhibitory effects.
I would like to thank the University of Iowa, Dr. Alec Scranton, Dr. Beth Ficek Rundlett and Dr. Cindy Hoppe for your hard work on this project.
I would also like to thank Jason Stark of Toyota North America R&D for your support on this project.