Waterborne UV-Curable Coatings
Experimental
The model formulation used in this
study consisted of a novel light stabilizer (NLS) and a waterborne UV-curable
coating (Table 1). DOE was used to design test runs as well as to analyze test
data. DOE techniques allowed us to validate the data.
Optimal Results
Optimal results (Table 2) were used to
identify factors that reduce yellowing,
i.e., the concentration of NLS and the dry
film thickness (DFT). Analysis of variance
(Table 3), Parameter estimates (Table 4)
and effect tests (Table 5) were analyzed
from Table 2. The mathematical model
derived from the data indicates that two
factors, namely the NLS and the DFT,
exhibit some degree of influence on the
reduction of yellowing.
Novel Light Stabilizer Effects
The NLS was developed specifically for
waterborne UV-curable coatings. Test
data showed from its negative coefficient that the NLS in a waterborne, clear
UV-curable coating can get good performance. For the NLS, the effect of reducing yellowing during the curing process
was about 79% (Table 6).
Dry Film Thickness (DFT) Effects
Film thickness is the less significant factor in this study (Equation 1). Test data
showed from its positive coefficient that
the DFT of waterborne, clear UV-curable
coatings can’t affect good performance.
For DFT, the effect of reducing yellowing
during the curing process was only about
0.6% (Table 6).
The Optimal Prediction Profiler
The prediction profiler (Figure 1) shows
the effect that the NLS and the DFT have
on reducing yellowing after the curing
process. The data also suggested that by
increasing the NLS to 3.44% in a waterborne UV-curable coating and having a
coating thickness of 5 μm the best performance can be achieved [Delta Yellowness
Index ΔY= -0.76 (Figure 1)].
Response Surface Methodology (RSM)
Response surface methodology (RSM)
Table 2: Evaluation of optimal results.
Table 3: Analysis of variance (ANOVA).
Table 4: Parameter estimates.
Conc.(NLS) With waterborne UV-curable
coating w/w DFT (μm) ΔY ellowness Index (During Cure)
0 5 4. 32
0 10 5.05
0 20 6. 5
0 60 8. 5
1 5 1.59
1 10 1.64
1 20 1.71
1 60 1.83
3 5 0.2
3 10 0.35
3 20 0.48
3 60 0.59
5 5 0.06
5 10 0.10
5 20 0.18
5 60 0.27
Source DF Sum of Squares Mean Square F Ratio
Model 5 90.32985 18.0660 16.2928
Error 10 11.08835 1.1088 Prob > F
C. Total 15 101.41819 0.0002
Summary of Fit R2 = 0.89 R2 Adj =0.83, RMSE=1.05
Term Estimate Std. Error t Ratio Prob>|t|
Intercept 2.752804 0.558705 4.93 0.006
NLS(%)&RS -1.264535 0.147385 - 8. 58 <.0001
DFT((μm))&RS 0.032602 0.025871 1.26 0.2362
(NLS(%)-2.25)*(NLS(%)-2.25) 0.4509548 0.101358 4. 45 0.0012
(DFT((μm))- 23.75)*(NLS(%)-2.25) -0.010265 0.006343 -1.62 0.1366
(DFT( (μm))- 23.75)*(DFT( (μm))- 23.75) -0.000555 0.001111 -0.50 0.6284
Table 5: Effect tests.
Source Nparm DF Sum of Squares F Ratio Prob > F
NLS(%)&RS 1 1 81.625213 73.6135 <.0001
DFT((μm))&RS 1 1 1.760882 1.5880 0.2362
NLS(%)*NLS(%) 1 1 21.949115 19.7948 0.0012
DFT((μm))*NLS(%) 1 1 2.904596 2.6195 0.1366
DFT( (μm))*DFT( (μm)) 1 1 0.276431 0.2493 0.6284