Model-based analysis of photoinitiated coating degradation
Søren Kiil
Technical University of Denmark
Department of Chemical and Biochemical Engineering
Kgs. Lyngby, Denmark
Kurt Wood, Arkema
Coatings Research at DTU Chemical Engineering
CHEC Research Center
Research Topics (heavy duty coatings)
• Antifouling coatings for ships
• Anticorrosive coatings (incl. weathering)
• Blade coatings for wind turbines
• Intumescent coatings (fire protection)
2 DTU Chemical Engineering, Technical University of Denmark
People
• 2 professors (faculty)
• 3-5 Ph.D. students
• 2-4 M.Sc. Students
• Collaborators from coatings industry
Motivation and topics covered
• To obtain a ”complete” picture of UV-induced coating degradation
• Improve knowledge on interlayer adhesion loss
– problem between epoxy base coat and PU top coat
– loss of adhesion after few days of UV exposure
• Contents of presentation
3 DTU Chemical Engineering, Technical University of Denmark
• Contents of presentation
- mathematical model of coating degradation
- mechanisms and phenomena included
- verification
- high Tg coatings with narrow oxidation zones (2 µm)
- low Tg coatings with wide oxidation zones (250 µm)
- effect of light stabilizers
- future work
Thermoset coating exposed to UV radiation and humidity
Top view:
4 DTU Chemical Engineering, Technical University of Denmark
unexposed
exposed
Gloss loss not considered
Model contents
• Photoinitiated oxidation reactions
• Oxygen permeability
• Water absorption and diffusion
5 DTU Chemical Engineering, Technical University of Denmark
• Reduction of crosslink density (erosion)
• Development of an intrafilm oxidation zone
• Effect of light stabilizers (UV absorber and HALS)
For details on model assumptions see Kiil (2012), JCT Res.
Chemical degradation mechanism
• Case study with crosslinked epoxy-amine coatings
• Closed-loop chemical mechanism
6 DTU Chemical Engineering, Technical University of Denmark
• Reactions not included
Branching
Termination (not important here)
Adjustable parameters
• Rate constants for reaction (1) and/or (2)
(always included)
• Diffusion coefficient of oxygen in oxidation zone
(only when diffusion plays a role)
7 DTU Chemical Engineering, Technical University of Denmark
(only when diffusion plays a role)
• Stable conversion of crosslinks at coating surface when erosion is initiated (if erosion takes place)
• Rate constant of radical scavenger
(only when present)
Simulations and experimental data
ATR-FTIR
8 DTU Chemical Engineering, Technical University of Denmark
Clear coat, 50 °C, RH=75 %, Tg > 100 °C, UV=480 W/m2
NIST integrating sphere
(data from Nguyen et al., COSI, 2010)
Simulation of oxidation zone development
Clear coat, 50 °C, RH=75 %,
Tg > 100 °C, UV=480 W/m2
Oxygen diffusion very important
9 DTU Chemical Engineering, Technical University of Denmark
Oxidation zone thickness of about 2 µmin agreement with independent measurements by Monney et al. (1997, 1998, 1999)
Effect of nano-pigments on mass loss
50 °C, RH=75 %
Tg > 100 °C, UV=480 W/m2
MWCNT has a strong effect on stable
surface conversion
10 DTU Chemical Engineering, Technical University of Denmark
Nano-SiO2 has no effect
(same parameters used as for clear coat)
Experimental data from
Nguyen et al., NIST, (2010)
Effect of relative humidity, RH (50 °°°°C), on erosion rate
Transmission IR
11 DTU Chemical Engineering, Technical University of Denmark
RH=75 % RH=9 %
Experimental data from Rezig et al., NIST, (2006) (Tg=123 °C)
Simulations and experimental data (Mailhot et al., France, 2005)
0
0.2
0.4
0.6
0.8
1
yC
C (su
rfac
e v
alu
e)
0
0.002
0.004
0.006
0.008
0.01
∆O
D (
IR)
Simulation of yCC
Experimental data (∆OD)
0
0.08
0.16
0.24
0.32
0.4
0.48
0.56
yC
O
0
0.2
0.4
0.6
0.8
1
∆O
D (
167
5 c
m-1
)
Simulation
Experimental data (∆OD)
30 hours
ATR-IR data Micro-FTIR data
12 DTU Chemical Engineering, Technical University of Denmark
Clear coat, 60 °C, Tg=-50 °C, RH and UV unknown, no light stabilizers
0 40 80 120
Time (hours)
0
0.2
0.4
0.6
yC
O (
su
rfa
ce
va
lue
)
0
0.004
0.008
0.012
0.016
∆O
D (
IR)
0 0
Simulation of yCO
Experimental data (∆OD)
0 100 200 300
Position in coating (µm)
0
0.1
0.2
0.3
0.4
0.5
yC
O
0
0.2
0.4
0.6
0.8
∆O
D (
1675
cm
-1)
0 0
Simulation
Experimental data (∆OD)
110 hours
Effect of UV absorber on absorption and oxidation depth
Clear coat, 60 °C, Tg=-50 °C, 30 hours
Experimental data from Mailhot et al. (2004)
Without UV absorber the oxidation zone
thickness was 250 µm
0.10
0.2
0.4
0.6
0.8
1
El/E
O
Simulation
13 DTU Chemical Engineering, Technical University of Denmark
No oxygen concentration profile
Absorption of UV radiation limits
oxidation zone thickness
0 50 100 150
Position in coating (µm)
0
0.1
0.2
0.3
0.4
0.5
0.6
yC
O
0.05
0.06
0.07
0.08
0.09
0.1
∆O
D (
1675 c
m-1)
Simulation
Experimental data (∆OD)
Effect of molar absorptivity of UV absorber and rate constant of radical scavenger
Clear coat, 60 °C, Tg=-50 °C, 30 hours,
UV absorber present
0
0.1
0.2
0.3
0.4
0.5
0.6
yC
O
αUVA=600 m2/mol
αUVA=400 m2/mol
αUVA=200 m2/mol
14 DTU Chemical Engineering, Technical University of Denmark
0 50 100 150
Position in coating (µm)
0
0.1
0.2
0.3
0.4
0.5
yC
O
0
kRS=0
kRS=0.2 (m3/mol) s-1
kRS=0.5 (m3/mol) s-1
2 2R NO R R NOR⋅ + ⋅ →
2( )
R RS R R NOr k C C⋅ ⋅ ⋅− =
Reaction of radical scavenger (HALS):
UVA present
Conclusions
• Mathematical model developed for photoinitiated coating degradation
• The relative importance of the different rate phenomena quantified
• Model requires calibration of adjustable parameters for practical use
- from laboratory data to service life prediction
15 DTU Chemical Engineering, Technical University of Denmark
- from laboratory data to service life prediction
• Model cannot predict gloss loss and speciation of photoproducts
Goldschmidt and Streitberger (2007)
Future work
• Extend model to cyclic accelerated testing
- effects of rain erosion, time of wetness, UV radiation, and temperature
• Extend model to more complex epoxy-amine formulations
- including other pigments and additives
16 DTU Chemical Engineering, Technical University of Denmark
• Extend model to other coatings/binders
- requires a closed-loop mechanism and data for model calibration
• Use model for simulation of weather scenarios of commercial interest
• Detailed experimental data series required!
Thank you for your attention…
Financial support by the Hempel Foundation is gratefully acknowledged
Contact:Søren Kiil, [email protected]
17 DTU Chemical Engineering, Technical University of Denmark
Goldschmidt and Streitberger (2007)