PMMA Acrylic in a Stress-Response
Framework for PV Materials
Laura S. Bruckman
Materials Science
Case Western Reserve University
Cleveland OH 44106
Funding support from
Ohio Third Frontier
under Tech 11-060
& Tech 12-004
laura.bruckman@case.edu
Motivation: Lifetime & Degradation Science for Photovoltaics
Need scientific basis for PV module lifetimes
• They could last 50 year lifetimes
• They have 25 year product warranty period
Qualification testing of systems: infant mortality and manufacture defects
• It does not validate lifetime claims
Only recently, an international effort toward lifetime qualification testing
• NREL’s PVQATaskForce
• Avoid excessively high degradation rates
• Dramatically increase service lifetimes
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 2
Motivation: Lifetime & Degradation Science for Photovoltaics
Solar Durability and Lifetime Extension (SDLE) Center
• Determine at use conditions
• Materials, components, systems for PV
• System lifetime performance
Determine degradation modes, mechanisms and rates
• Under single, multi-factor, cyclic and sequential stresses
• Scientific underpinning of reliability and qualification standards
Quantitative degradation rate modeling
• Connects materials, components, system
• To overall degradation rate, linearity, reciprocity, additivity
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 3
L&DS: Stress Response Framework and Cross Correlation
Material Response Value
Accelerated
Real-time
Dose Steps, over time
System response at different stress rates
If system is only linear response
• Then increased stress rate
• Reduces time required for given net stress
If systems response is non-linear
• Then sub-, supra-linear response observed
• Photodarkening and photobleaching
Cross-correlation of stress/response
• Defines the system’s response function
• Captures linear and non-linear phenomena
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 4
Real World Stress Conditions
Stress Intensity –Irradiance
Graph of daily irradiance TMY3
𝜎j
Real world exposures have more than 1
stressor
𝑆=
(𝜎𝑖 ⨂𝜎𝑗 … ⨂𝜎𝑛 )𝑑𝑡
Graph is of daily Irradiance and Relative Humidity TMY3
Net Stress is a Stress Intensity over time
Stress is not consistent over time
Integration of net stress is “dose”
𝜎i
Real world conditions are multifactor
Cross correlation of stress
• Can handle the multifactor aspect
• Of real world conditions
TMY3 data from DView V 1.24 TMY3 data for Cleveland, Ohio National Renewable Energy Laboratory www.nrel.gov
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 5
Acrylic PMMA Case Study
Multiple grades of acrylic
• Ultra-violet transmitting (UVT)- no stabilizer package
• Multipurpose (MP)-contains some stabilizer package
Multiple stressors
• Multiple irradiance conditions
• UVA-340
• AM 1.5 Full Spectrum
Abengoa
SDLE Center, Roger H. French © 2012
Emcore
April 25, 2012, VuGraph 6
Case Study Approach: Lifetime & Degradation Science
Amass large volumes of data across:
• Stressors (single, multifactor, cyclic, sequential)
• Stress levels
• Degradation modes, mechanisms, quantitative rates
• Samples
Compare accelerated & real time test conditions
Cross-correlation
• Across two stressors and two stress levels
• Considering 4 degradation modes
Replex Plastics’ Mirror Augmented PV
The cross correlation studies:
• Basis of L&DS approach
• Provide a basis for making lifetime predictions
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 7
Stressors: QUV UVA-340 & Xe arc AM 1.5
QUV Accelerated Weathering Chamber
Fluorescent UVA-340
1.55 W/m2/nm at 340 nm
• Corresponding to 5X Suns
48.4 kW/m2 Newport Solar Simulator
• Corresponding to 50X Suns
Full Spectrum Xenon Arc AM 1.5
• With 13X concentrator
67°C sample temperature
SDLE Center, Roger H. French © 2012
50°C sample temperature
April 25, 2012, VuGraph 8
Average Induced Absorbance to Dose: Quantitative Metric
Average Induced Absorbance to Dose
• Change in optical absorbance per centimeter of material
Baseline to Dose Step
• Quantitative metric to monitor photodarkening and photobleaching
• Can be calculated on different dose basis (Full-Spectrum, UVA-340)
• Continuous degradation modes are predominantly observed
If successive dose steps, have the same value, then there is linear behavior
Two different exposures have the same value, then reciprocity is obeyed
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 9
Acrylic Grades: Multipurpose (MP) in QUV
2’
1
2
3
1
photodarkening
photobleaching
2
3
2’
1) Fundamental Absorption Edge
2) Stabilizer package bleaching
3) Visible region yellowing
Increasing
Dose
Initial photobleaching rate is high
• Region 2 and 2’ (1 dose)
Photobleaching rate decreases
• Dose 3 and 4
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 10
Acrylic Grades: Ultra-violet Transmitting (UVT) in QUV
1
1
Increasing
Dose
Increasing
Dose
3
3
Region 2 cannot be analyzed: no stabilizer
Fundamental Absorption Edge
Response Ratios UVT to MP
Region 1: 4.5
Region 3: 9.9
Degradation modes in acrylic
• Higher initial degradation rate
Region 3
• Similar degradation rates
• Have independent degradation rates
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 11
Acrylic Grades: Yellowness Index from QUV
UVT has a higher Yellowness Index
• Less stabilized acrylic
Yellowness Index relates to a higher optical absorbance (reduces lifetime)
Response Ratio UVT to MP: 11.7
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 12
Acrylic Grades: MP and UVT in the Newport Solar Simulator
1
Average Induced Absorbance Dose
2
• Rates are comparable within UVT and MP
3
2’
Response Ratios UVT to MP
Region 1: 1.9
Region 3: 6.1
photodarkening
photobleaching
MP grade
Fundamental Abs. Edge Degrades
• At almost constant rate in these grades
MP Grade Yellows At Lower Rate
• Attributable to stabilizers?
Degradation modes in acrylic
• Have independent degradation rates
SDLE Center, Roger H. French © 2012
1
3
April 25, 2012, VuGraph 13
UVT grade
Response Ratios: Comparison of Stressors
1
1
2
2’
3
3
MP Grade
QUV Stress caused more degradation
UVT Grade
• On an equivalent UVA 340-dose basis
QUV (4th dose-206 MJ dose, 39 days) and Newport (2nd dose 253 MJ dose, 6 days)
Higher photobleaching rate of MP in Region 2 and 2’
Higher photodarkening rate of UVT in Regions 1 and 3
A higher sample temperature of the QUV?
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 14
Conclusions: L&DS Case Study of PMMA Acrylic
PMMA abs. edge degradation
• Comparable response for
• 50 X suns, full spectrum AM 1.5
UVT has a higher response ratio
• Compared to MP
• For both stress conditions
QUV caused a higher response
• In both UVT & MP
• On a comparable dose basis than NSS
There is not one “acceleration factor”
for a material
• Irradiance
• Temperature
• Region of the spectrum
• Convolution of stress conditions
SDLE Center, Roger H. French © 2012
Population Based Studies
Sample Sets
Baseline, Steps 1,2,3,4…
Studies
System Level
Component Level
Materials
Degradation
Modes
Mechanisms
Quantitative Rates
Exposures
(Stresses)
Evaluations
(Responses)
Real Time
Accelerated
Sequential 
Quantitative Rates
Single, Multi-Factor
Cyclic, Sequential
Performance
Canary: Precursor
System Technology Model
Stress Response Cross Correlation
Lifetime Prediction by Modes
Translational Predictions
April 25, 2012, VuGraph 15
Backgrounder
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 16
Elements of the Lifetime & Degradation Science Methodology
In a stress & response framework
Population based studies
• Of systems, components, materials
Study protocols: exposure, evaluations
• Many stressor types, levels & cycles
• Exercising multiple degradation modes
Multiple responses measured per step
• Determine quantitative degradation rates
Cross-correlation of stress & response
• To produce system technology models
• Lifetime predictions by degradation modes
• Translational predictions for stress conditions
SDLE Center, Roger H. French © 2012
Population Based Studies
Sample Sets
Baseline, Steps 1,2,3,4…
Studies
System Level
Component Level
Materials
Degradation
Modes
Mechanisms
Quantitative Rates
Evaluations
(Responses)
Exposures
(Stresses)
Sequential 
Quantitative Rates
Real Time
Accelerated
Performance
Canary: Precursor
Single, Multi-Factor
Cyclic, Sequential
System Technology Model
Stress Response Cross Correlation
Lifetime Prediction by Modes
Translational Predictions
April 25, 2012, VuGraph 17
Acknowledgments
• Dr. Roger French
• French Group
Myles P. Murray
Yang Hu
Wei-Chun Lin
Esther Deena
• Funding support from Ohio Third Frontier under Tech 11-060 & Tech 12-004
• Replex Plastics
Lifetime &
Degradation
Science
LCOE
Lifetime
Performance
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 18
Lifetime
Cost
Conclusions
The lifetime of PV materials and modules need to be determined
Stress-Response Framework
• Allows for cross-correlation of responses to a stress or multiple stressors
UVT has a higher response ratio
• Compared to MP
• For both stress conditions
QUV caused a higher response
• In both UVT & MP
• On a comparable dose basis than NSS
There is not one acceleration factor for a material
• Irradiance
• Temperature
• Region of the spectrum
• Convolution of stress conditions
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 19
Response Ratios-Comparison of Stressors
Region 1 (275
Region 2 (298
Region 2’ (339
Region 3 (440
nm)
nm)
nm)
nm)
UVT
2.8
3.3
4.6
6.3
MP
1.2
1.6
0.86
3.8
Response Ratios of from QUV (step 4) to Newport solar simulator (step 1) for
each region for the two grades of PMMA of equivalent dose steps.
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 20
Motivation: Lifetime & Degradation Science for Photovoltaics
2010 Science For Energy Technology Workshop
• Convened by U. S. DOE, Basic Energy Sciences
Science challenges across 9 areas of energy
PV prioritized research directions
• Photovoltaic module lifetime and degradation science
• Fundamental properties of photovoltaic interfaces
• Advanced photovoltaic analysis & computational modeling for scale-up
Qualification testing of systems
• Is not sufficient for reliability & lifetime
• To avoid excessively high degradation rates
• Dramatically reduced service lifetimes
Determine degradation mechanisms and rates
• Scientific underpinning of reliability and qualification standards
Quantitative degradation rate modeling
• Connects materials, components, system
• To overall degradation rate, linearity, reciprocity
• And system lifetime performance
http://www.er.doe.gov/bes/reports/files/SETF_rpt.pdf
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 21
Overview of the SDLE Methodology
Does Current Qualification Testing
Advance PV Reliability & Lifetime?
• Provides no lifetime validation
Only pass/fail results
Structured approach: study protocols
• For exposures and evaluations
Development of new study protocols
• For lifetime qualification standards
• Testing doesn’t produce scientific insights
Into degradation mechanisms and rates
• Reliability testing not founded
In the physics of failure
• Qual. testing doesn’t advance PV research
Lifetime and degradation science
• A scientific foundation for advancing PV lifetime
• Quantitative degradation mechanisms & rates
• Real time and accelerated testing
• Single and multi-factor studies
Material, component & system approach
• Causation of failure arises in
complete systems and their interactions
• Must study component and material degradation Engineering epidemiology
• For lifetime prediction
To identify degradation mechanisms
e.g. PV Power degradation rates
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 22
SDLE Exposure Capabilities
Outdoor Testing
• SDLE SunFarm
Irradiance
Temperature and Humidity
Yang Hu et. al. in Silicon Poster Session
Indoor Testing
• Irradiance
Newport Solar Simulator 1-50x Suns
SpectroLab XT 30 CW 100-1200x Suns
Irradiance, Temperature, & Humidity
• Q-Labs QUV/Spray Weathering Chamber
• Q-Labs Q-Sun Weathering
• CSZ ETC 100 cu. ft. with 1 Sun Light Rack
Temperature & Humidity
• CSZ ETC 8 cu. ft.
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 23
SDLE Evaluation Capabilities
Optical Properties
• Cary 6000i with DRA-1800
• Filmetrics PARTS-UV
• HunterLabs UltraScan Pro
• Woollam VUV-Vase
Spectroscopic Ellipsometer
Light Scattering Properties
• ScatterMaster 3D
• Wyatt Dawn Helios II
Surface and Bulk Properties
• Kruss DSA Surface Chemistry
• Olympus Confocal Microscope
• PHI VersaProbe XPS Microprobe
Mechanical Properties
• Nanoindentation
• Peel Strength
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 24
Stress Response Framework for Lifetime & Degradation Science
System response to applied stress
Study response of system
• At many stress levels
• With many applied stresses (multi-factor)
If system is only linear response
• Then increased stress rate
• Reduces time required for given net stress
SDLE Center, Roger H. French © 2012
If systems response is non-linear
• Then sub-, supra-linear response observed
Cross-correlation of stress/response
• Defines the system’s response function
• Captures linear and non-linear phenomena
As a matrix transform of stress to response
April 25, 2012, VuGraph 25
Responses: Materials change as a function of stressors
Various levels of stress
• Stress intensity  and integrated stress S
Such as solar irradiance and solar dose
Measure responses
• Optical responses: abs/cm, index of refraction,
YI, haze …
• Mechanical: embrittlement, crazing, haze …
Cross-correlation defines transferability
• Reciprocity between irradiance and time
• Linearity over dose or time
• Additivity over wavelength
Determine range of applicability of
materials, components, systems
• Not just applicability to one set of stresses
SDLE Center, Roger H. French © 2012
Response a function of
• Stress level 
• Net stress S
𝑅 = 𝑓 𝜎𝑆 = 𝑓 𝜎 𝜎𝑑𝑡
Materials show many responses
1) Fundamental Absorption Edge
2) Light Stabilizer Package
3) Visible Yellowing
Expansion to multiple stresses
gives:
𝑅 𝑆𝑖 , 𝑆𝑗 , … 𝑆𝑛
= 𝑓(𝜎𝑖 , 𝜎𝑗 , … 𝜎𝑛 ) (𝜎𝑖 ⨂𝜎𝑗 … ⨂𝜎𝑛 )𝑑𝑡
April 25, 2012, VuGraph 26
Case Study: Mirror Augmented PV (MAPV) with Acrylic Mirrors
MAPV Value Proposition
• Increased irradiance on PV module
• Decreases LCOE
• Mirrors cost less than modules
Components in Acrylic:
• 1) Base resin
Residual monomer
• 2) Light stabilizer package
UV absorbers, radical scavenger
PV Module
Acrylic Back Surface mirrors
• Use acrylic transmission
• To reduce UV and IR load on module
Durability Questions
• Are acrylic mirrors durable for 25 years?
• Do increased module stressors decrease
service life
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 27
Stressors: Inducing Responses in Materials
Compare 2 irradiance tools
• Full spectrum Xe AM1.5D
• Fluorescent UVA 340
48.4 kW/m2 Newport Solar Simulator
• Diverging Xenon Arc AM1.5
With 13X Concentrator
• Exposures at 48.4 kW/m2
QUV at 1.55W/m2 at 340 nm
Fluorescent weathering chamber
Outfitted with UVA 340 Lamps
Exposures at 1.55 W/m2 @ 340 nm
Bench top
Beam dump
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 28
Compare MP Responses: For 3 Degradation Mechanisms
0.008
Incremental absorbance to dose
Agreement in IAD rate
For Mechanism 1
2) Stabilizer package bleaching
• Photobleaching Rates Comparable
• Rates Determine Package Lifetime
3) Visible region yellowing
• AM1.5 Yields Greater YI Yellowing
• Spectral Differences AM1.5 vs UVA340
IAD of MP acrylic in AM 1.5
0.006
0.006
1
0.004
0.004
0.002
0.002
3
0
0
270
370
470
570
670
2
-0.002
-0.002
Produce different amounts of yellowing
2’
Degradation Mechanisms & Rates
• To Estimate Lifetime Performance
SDLE Center, Roger H. French © 2012
Wavelength (nm)
-0.004
-0.004
April 25, 2012, VuGraph 29
ΔAbs/cm per MJ/m2 UVA-340 dose AM 1.5
For AM1.5 and UVA-340 Experiments
1) Base resin photodegradation
IAD of MP acrylic in UVA 340
ΔAbs/cm per MJ/m2 UVA-340 dose
Abs
GJ
Incrementa l
per
Dose
2
cm
m
Absi 1  
Absi  

cm
cm

Dosei 1  Dosei
0.008
Comparison of 2 Different Types of Acrylic
IAD Results for Multipurpose Acrylic
• Two different stress conditions
Newport Solar Simulator
QUV Weathering Chamber
• QUV 206 MJ dose
• NSS 253 MJ dose
SDLE Center, Roger H. French © 2012
IAD Results for Ultra-violet
Transmitting Acrylic
• Two different stress conditions
Newport Solar Simulator
QUV Weathering Chamber
• QUV 206 MJ dose
• NSS 253 MJ dose
April 25, 2012, VuGraph 30
Examples of Lifetime Prediction Metrics for MAPV
Lifetime Prediction: After 12 years
• Optical absorbance will increase like UVT
• Due to consumption of stabilizers
PV technologies rely on 25 year
warranted lifetimes
Leading to 7% loss in transmitted
flux
• after 76 MJ/m2 flux
• 8 years 1 sun exposure
• Ignorance of degradation pathways is
hazardous
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 31
Conclusions
Lifetime and Degradation Science
• Furthers the understanding of PV reliability
• Determine degradation mechanisms and rates
• Beneficial for achieving 25 year lifetimes in PV modules
Lifetime &
Degradation
Science
Current qualification testing
• Doesn’t advance PV research
LCOE
SDLE Center
• Significant Outdoor and Indoor Testing Capabilities
• Optical, Light Scattering, Surface & Bulk, Mechanical
Lifetime
Performance
Evaluation Capabilities
Stress/Response Framework
• Cross-correlation of stress/response defines the system’s response function
Acrylic Case Study
• Compare 2 different types of acrylic in 2 different types of stresses
Acknowledgements
• Funding support from Ohio Third Frontier under Tech 11-060 & Tech 12-004
• Replex Plastics
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 32
Lifetime
Cost
Stress & response framework
Population based studies
Of systems, components, materials
Study protocols :exposure, evaluations
Many stressor types, levels & cycles
Exercising multiple degradation modes
Multiple responses measured at each step
Determine quantitative degradation rates
Cross-correlation of stress & response
To produce system technology models
Lifetime predictions by degradation modes
Translational predictions for stress conditions
SDLE Center, Roger H. French © 2012
April 25, 2012, VuGraph 33