Determination of Activation Energy for

Encapsulant Browning of Photovoltaic Modules

by

Deepak Jain Veerendra Kumar

A Thesis Presented in Partial Fulfillment

of the Requirements for the Degree

Master of Science

Approved November 2016 by the

Graduate Supervisory Committee:

Govindasamy Tamizhmani, Chair

Devarajan Srinivasan

Bradley Rogers

ARIZONA STATE UNIVERSITY

December 2016

i

ABSTRACT

The primary goal of this thesis work is to determine the activation energy for encapsulant

browning reaction of photovoltaic (PV) modules using outdoor field degradation data and

indoor accelerated degradation data. For the outdoor field data, six PV modules fielded in

Arizona (hot climate) over 21 years and four PV modules fielded in New York (cold

climate) over 18 years have been analyzed. All the ten modules were manufactured by

the same manufacturer with glass/EVA/cell/EVA/back sheet construction. The activation

energy for the encapsulant browning is calculated using the degradation rates of short-

circuit current (Isc, the response parameter), weather data (temperature, humidity, and

UV, the stress parameters) and different empirical rate models such as Arrhenius, Peck,

Klinger and modified Peck models. For the indoor accelerated data, three sets of mini-

modules with the same construction/manufacturer as that of the outdoor fielded modules

were subjected indoor accelerated weathering stress and the test data were analyzed. The

indoor accelerated test was carried out in a weathering chamber at the chamber

temperature of 20°C, chamber relative humidity of 65%, and irradiance of 1 W/m2 at

340nm using a xenon arc lamp. Typically, to obtain activation energy, the test samples

are stressed at two (or more) temperatures in two (or more) chambers. However, in this

work, it has been attempted to do the acceleration testing of eight mini-modules at

multiple temperatures using a single chamber. Multiple temperatures in a single chamber

were obtained using thermal insulators on the back of the mini-modules. Depending on

the thickness of the thermal insulators with constant solar gain from the xenon lamp,

different temperatures on the test samples were achieved using a single weathering

chamber. The Isc loss and temperature of the mini-modules were continuously monitored

ii

using a data logger. Also, the mini-modules were taken out every two weeks and various

characterization tests such as IV, QE, UV fluorescence and reflectance were carried out.

Activation energy from the indoor accelerated tests was calculated using the short circuit

current degradation rate and operating temperatures of the mini-modules. The activation

energy for the encapsulant browning obtained from the outdoor field data and the indoor

accelerated data are compared and analyzed in this work.

iii

ACKNOWLEDGMENTS

I offer my utmost gratitude to Dr. Govindasamy Tamizhmani for accepting me in

his lab and giving me the chance to work with some of the best and world class

equipment in the industry now. Learning, as they say, is always an upward facing graph

and I consider myself lucky for working along with Dr. Mani and his insight and

expertise which have given me the opportunity to learn almost all almost aspects related

to the PV industry. I thank Dr. Michael Kempe, Dr. Dirk Jordan, and Dr. David Miller

from NREL for their valuable inputs in my work. I would also like to thank my

committee members, Dr. Rogers and Dr. Srinivasan, for their help and support along the

way.

I would like to also forward my gratitude to all the present and previous students

at ASU PRL for helping me at my hard time. I would also like to thank my colleagues who

worked with me especially Mr. Sai Tatapudi, our technical lab manager, for helping me out

with my experiments whenever I requested help. Besides the students with whom I have

been working with since I joined the lab, I would also like to give a special mention to all

the students who have been a part of ASU PRL. I thank Hamsini Gopalakrishnan for

helping with the formatting the document.

Finally, I would like to express my gratitude to my parents, Mr. Veerendra Kumar

and Mrs. Manju Jain, and my brother and sister-in-law, Mr. Pankaj Jain and Mrs. Neelam

Jain, my fiancée Trupti Jain for their love and support all throughout.

iv

TABLE OF CONTENTS

Page

LIST OF TABLES ............................................................................................................. vi

LIST OF FIGURES .......................................................................................................... vii

LIST OF ABBREVIATIONS ............................................................................................. x

CHAPTER Page

1 INTRODUCTION ...................................................................................................... 1

Background .......................................................................................................... 1

Why is Browning Important ................................................................................. 2

Scope of Work ...................................................................................................... 4

2 LITERATURE REVIEW ........................................................................................... 5

3 METHODOLOGY ..................................................................................................... 8

Module History .................................................................................................... 8

Visual Inspection ................................................................................................ 12

IV Measurements ............................................................................................... 13

Electroluminescence........................................................................................... 14

Dark IV ............................................................................................................... 14

Acceleration Factor ............................................................................................ 14

Acceleration Test Model .................................................................................... 15

3.7.1 Arrhenius Model ......................................................................................... 15

v

CHAPTER Page

3.7.2 Peck’s Model .............................................................................................. 15

3.7.3 Klingers Model ........................................................................................... 16

3.7.4 Modified Peck’s Model with UV and without Humidity ........................... 16

3.7.5 Modified Peck’s Model with UV and Humidity......................................... 16

Acceleration Testing of Mini-Modules .............................................................. 17

3.8.1 Cutting and Soldering ................................................................................. 18

3.8.2 Insulation..................................................................................................... 21

3.8.3 Data Logger ................................................................................................ 21

3.8.4 Installation of Mini-Modules in the Chamber ............................................ 21

3.8.5 Characterization Tests ................................................................................. 23

4 RESULTS AND DISCUSSION ............................................................................... 26

IV Measurements: .............................................................................................. 26

Determination of Activation energy using Field Test Data ............................... 29

4.2.1 Activation Energy Determination without UV Stress ................................ 29

4.2.2 Activation Energy Determination with UV Stress...................................... 32

Determination of Activation Energy using Accelerated Test ............................ 35

4.3.1 IV Measurements ........................................................................................ 35

5 CONCLUSION ......................................................................................................... 43

REFERENCES ................................................................................................................. 45

vi

LIST OF TABLES

Table Page

1 Nameplate Reading of the Modules ............................................................................... 12

2 Final Data Translated by PRL Template ....................................................................... 27

vii

LIST OF FIGURES

Figure Page

1: Failure Type Percentage for 1865 Modules in Arizona .................................................. 2

2: Performance Defect of Modules in Different Climatic Conditions ................................ 3

3: Power Degradation of Sun Power Modules for Different Climatic Conditions3 ............ 4

4: Gel Content and Cyasorb for Clear and Yellowed EVA5 ............................................... 6

5: 18years Old Arizona Module Front View ...................................................................... 8

6: 18years Old Arizona Module Rear View ....................................................................... 9

7: 21years Old New York Framed Module Front View ..................................................... 9

8: 21years Old New York Framed Module Rear View .................................................... 10

9: 21years Old New York Frameless Module Front View ............................................... 11

10: 21years Old New York Frameless Module Rear View .............................................. 12

11: Front View of the Mini-Module showing the Center Cells ........................................ 17

12: Rear View of the Mini-Module .................................................................................. 18

13: Cutting of the Back Sheet using the Blade ................................................................. 19

14: Rear View of the Mini-Module after Cutting the Back Sheet at the Required Position

........................................................................................................................................... 19

15: Soldering of the 22 AWG Wire to the Solder Bond of the Mini-Module .................. 20

16: Rear View of the Mini-Module after the Curing of Epoxy Resin .............................. 20

17: Mini-Module with no Insulation, Thin Insulation, and Thick Insulation ................... 21

18: Mini-Modules Installed in the Chamber using the Handi Clamps ............................. 22

19: Close up View of the Mini-Module Installed using the Handi Clamp ....................... 23

viii

Figure Page

20: Light Illuminated on Reference Cell and the Mini-Module using Indoor Solar

Simulator ........................................................................................................................... 24

21: Measurement of QE in the Exposed Part .................................................................... 25

22: Translated IV Curve (STC) for Arizona and New York Modules ............................. 26

23: Close up View of Translated IV Curve (STC) near Isc ............................................... 27

24: Annual Isc Degradation for AZ and NY ...................................................................... 28

25: Activation Energy for Peck’s Model with Different m Values Considering Ambient

Humidity ........................................................................................................................... 30

26: Activation Energy for Peck’s Model with Different m Values Considering Module

Humidity ........................................................................................................................... 31

27: Activation Energy for Klinger’s Model with Module and Ambient Humidity .......... 32

28: Peck’s Model with UV- Ambient Humidity for Exponent n Less than 1................... 33

29: Peck’s Model with UV- Ambient Humidity with Different m and n to get Activation

Energy Closer to 0.4 eV .................................................................................................... 33

30: Peck’s Model with UV- Module Humidity ................................................................ 34

31: Peck’s Model with UV- Module Humidity ................................................................ 35

32: Light IV for All the Cells of Mini-Module Before Stressing ..................................... 36

33: Light IV for All the Cells of Mini-Module after Stressing for Two Weeks ............... 37

34: Light IV for All the Cells of Mini-Module after Stressing for Four Weeks ............... 38

35: Comparison of IV Curve before and after Stressing for Mini-Module 4 Cell1 .......... 39

36: Percent Drop in Isc Per Hour after Four weeks of Accelerated Stress Testing ........... 40

ix

Figure Page

37: Percent Drop in Isc for Mini-Modules after Four Weeks of Stress Testing after

Removing the Non-Degraded Mini-Modules in the Increasing Order of Temperature ... 41

38: Activation Energy for Browning Considering Average Degradation of Each Mini-

Module .............................................................................................................................. 42

39: Activation energy for Browning Considering Degradation of Each Cell of Mini-

Module .............................................................................................................................. 42

x

LIST OF ABBREVIATIONS

PV - Photovoltaic

STC - Standard test conditions (Tmodule = 25°C; Irradiance = 1000 W/m2; AM = 1.5)

AM – Air Mass

AF – Acceleration Factor

Isc – Short-circuit

Voc – Open-circuit voltage

Imp – Maximum current

Vmp – Maximum voltage

Pm – Maximum power

MPPT – Maximum Power Point Tracking

DYI – Change in Yellowness Index

R - Gas constant 8.314 J.mol-1. K-1

Ir - Irradiance in W/m2

tf = time to failure

k = Boltzmann’s constant = 8.615 * 10-5 eV/K

T = Temperature in Kelvin

Tm = Temperature of the module (°C)

Ta = Ambient Temperature (°C)

m = Coefficient for RH

n = Coefficient for UV

E = Solar irradiance incident on module surface, (W/m2)

WS = Wind speed measured at standard 10-m height, (m/s)

xi

a = Empirically-determined coefficient

b = Empirically-determined coefficient

1

1 INTRODUCTION

Background

A PV module works based on the photoelectric effect where the incident photons excite

the atoms of the semiconductor material and an electron is generated. The generated

electrons are collected as electrical energy. It is vital for PV modules that the light

reaches the cells unobstructed; the medium should have maximum transmission. In a PV

module, the light passes through glass and the encapsulant. In the PV industry, the

dominant encapsulant used is based on a random copolymer consisting of about 67wt%

polyethylene and 33wt% polyvinyl acetate. Most PV module manufacturers warranty

80% power output after 25 years. The encapsulant degraded much sooner than 25 years

resulting in the deterioration of its properties especially in regions that receive high

irradiance.[1]

The main transmission losses originate from the encapsulant browning. Over time, the

encapsulant degrades turning brown in appearance in a process called encapsulant

yellowing or browning. The loss in transparency reduces the amount of light available for

the solar cell and in turn the power generated. The parameter most affected by browning

is short circuit current (Isc).

Crystalline silicon modules have glass-encapsulant-cell encapsulant-back sheet

construction and thin film have glass cell-glass construction. So browning predominantly

occurs in the first type as the encapsulant is present only in the former.

The lifetime of a PV module is 20-30 years, but it would be impractical to wait for

module failure to obtain valuable data. Accelerated tests, using intense test conditions

reduce the time to failure for the module. This enables researchers to obtain failure data

2

within months and analyze it. The stress factors for PV module acceleration testing are

temperature, humidity, UV dosage and voltage.

Why is Browning Important

Adam Alfred Suleske has done a performance evaluation on Photovoltaic modules in

desert climatic conditions. He has reported that 89.4% of 1865 modules (Figure 1) has a

browning defect for six models at APS star, Phoenix[2].

Figure 1: Failure Type Percentage for 1865 Modules in Arizona

Christopher Raupp of ASU PRL has reported that 21% of 39,431 modules (Figure 2)

from hot dry, temperate, cold dry and hot humid conditions have encapsulant

browning[3].

3

Figure 2: Performance Defect of Modules in Different Climatic Conditions

In one of the reports published by Sun Power, the degradation of their modules was

primarily due to encapsulant browning as shown in Figure 3. A new encapsulant was

formulated to mitigate this effect, but couldn’t completely get rid of encapsulant

browning.[4]

4

Figure 3: Power Degradation of Sun Power Modules for Different Climatic

Conditions3

As early as 1994 William H Holley et.al, announced having several reports of EVA

browning in field exposed flat plate PV nodules by all major EVA manufacturers.[5]

Scope of Work

In this project, a total of ten modules were investigated, six framed modules from

Arizona and four modules from New York. Two modules from New York are framed,

and two are frameless. The modules were characterized by Light IV, Dark IV, Infra-Red

imaging (IR), Electroluminescence (EL), and spectral measurements. The main objective

is to find the activation energy for encapsulant browning using Arrhenius, Peck’s, and

Klinger’s model and to find the best model among them.

Acceleration test is done on mini-modules with similar construction in a UV weathering

chamber. Since encapsulant, browning affects Isc, the degradation in Isc rather than power

degradation is considered. With the activation energy and the most representative model,

degradation of Isc due to browning for any climatic condition can be predicted.

5

2 LITERATURE REVIEW

Over time, the encapsulant in PV modules discolors becoming yellow/brown. The

discoloration follows a sigmoidal pattern and occurs due to the formation of

chromophores. A chromophore is the part of a molecule which is responsible for its color.

The color arises when a molecule absorbs certain wavelengths of visible light and

transmits or reflects others.

In the photothermal degradation of polymers, a common factor is the generation of free

radicals that are produced at the “activation sites” of impurities, ultraviolet (UV)-

excitable chromophores, and metal trace or ions in the polymers. UV light below 360 nm

is typically responsible for generating free radicals in the presence of oxygen.

EVA is more prone to being damaged by UV light than IR radiation from the sun. To

reduce UV degradation, additives such as UV absorbers, UV light stabilizers and

antioxidants are included in the encapsulant.

UV absorber dissipates the incident UV light into reheat or reemits the energy as the

harmless energy of higher wavelengths. The UV stabilizer acts as a “free radical

scavenger” to neutralize the free radicals generated in the polymeric matrix. The

Antioxidants are added to the polymer to reduce the thermal oxidation during thermal

processing and decompose the hydroperoxides in the polymer.[6]

The EVA films laminated in the PV modules stored in the dark are not degraded showing

that there is no degradation without weathering. Upon weathering, it is found that the gel

content increased rapidly and the UV absorber (Cyasorb) concentration decreased

rapidly. Upon further degradation, the EVA lost UV absorber rapidly and color changed

from light yellow to yellow to brown color depending on the extent of degradation.[7]

6

Degradation of EVA films upon weathering leads to yellowing of EVA, an increased

degree of cross-linking, loss of UV absorber (Cyasorb) and production of acetic acid. The

yellowed EVA has a 90-92% gel content as compared to 80-85% gel content of clear

EVA and the cyasorb is also reduced from 80-87% to 60-67% of the initial concentration.

The increase in the gel content and the decrease in the cyasorb concentration appears to

be two critical factors responsible for EVA yellowing as shown in Figure 4

Figure 4: Gel Content and Cyasorb for Clear and Yellowed EVA5

The UV light affects the EVA encapsulation. Yellowness Index (YI) is used to describe

the encapsulation discoloration. The change in Yellowness index (DYI) follows an

Arrhenius equation with an activation energy of 0.93eV. The DYI is proportional to the

irradiance intensity. The DYI is proportional to the logarithm of the illumination time.[8]

DYI = 9.1×10−24×𝑒−9000

𝑅𝑇 ×Ir×log (t)

where

DYI – Change in Yellowness Index

7

R - Gas constant 8.314 J.mol-1.K-1

Ir - Irradiance in W/m-2

David Miller et.al have quantified the activation energy based on the loss in

transmittance. Six different encapsulants, five formulations of EVA and one

thermoplastic polyurethane were studied in this test. The test is done in two devices:

Ci5000 ATLAS chamber and the NREL “UV suitcase.” The Ci5000 used exposure

settings of 1 W/m-2 at 340nm, with the indoor chamber temperature of 60˚C, the black

panel temperature uncontrolled and the chamber controlled to 30% RH, resulting in a

sample temperature of ~63 ˚C. The UV suitcase was set at 1 W.m-2 nm-1 at 340nm, with

a chamber temperature of 60 ˚C with uncontrolled RH, resulting in a sample temperature

of ~55 ˚C. An activation energy on the order of ~60 kJ⋅mol-1 is estimated for the

degradation of EVA-A.[9]

In the study done by Sun Power [4], it is reported that UV Absorber degradation does not

directly cause encapsulant browning. It follows the two-step reaction. The first step is a

photobleaching of UV absorbers and the second step is a first order reaction to form

yellow chromophores. The activation temperature (𝐸𝑎𝑅⁄ ) of the secondary browning

reaction is approximately 4700K.

8

3 METHODOLOGY

Module History

Six modules (framed, A60) from Tempe, Arizona, two from Valhalla (framed, A120),

two from Yonkers (frameless, A120) New York are used for analysis. Arizona modules

are exposed for 21 years at 33.4° tilt angle whereas modules from New York are exposed

for 18 years at 44°. One module from each category is shown in Figure 5 to Figure 10.

The performance tests done are IV, EL and IR. Modules in Tempe and Valhalla have

Tedlar-Polyester-EVA (TPE) back sheet (white in color) and the modules in Yonkers

have Tedlar back sheet (blue in color). Tedlar has low water permeability and Polyester

acts as electric insulation. All the modules in Arizona and New York have Cerium oxide

in glass.

Figure 5: 18years Old Arizona Module Front View

9

Figure 6: 18years Old Arizona Module Rear View

Figure 7: 21years Old New York Framed Module Front View

10

Figure 8: 21years Old New York Framed Module Rear View

11

Figure 9: 21years Old New York Frameless Module Front View

12

Figure 10: 21years Old New York Frameless Module Rear View

Visual Inspection

Visual inspection was done according to PRL checklist to find all the possible defects in

the modules according to PRL visual inspection checklist. Encapsulant browning can be

seen clearly from the visual inspection. Modules in Arizona has 36cells whereas modules

in New York has 72 cells. The nameplate reading is shown in Table 1.

Table 1 Nameplate Reading of the Modules

13

Place Serial No Model Isc Voc Imp Vmp FF Pm

Tempe, AZ 4958 A60 3.89 21.4 3.59 17.4 75 62.5

Tempe, AZ 4960 A60 3.86 21.5 3.62 17.3 75.5 62.5

Tempe, AZ 4961 A60 3.85 21.2 3.57 17.6 76.8 62.9

Tempe, AZ 4968 A60 3.87 21.4 3.6 17.3 75.2 62.5

Tempe, AZ A002 A60 3.71 21.2 3.4 17.4 75.2 59.2

Tempe, AZ A003 A60 3.67 21.2 3.39 17.2 74.9 58.2

Valhalla, NY 6917 A120 3.97 42.3 3.67 33.5 73.2 123.1

Valhalla, NY 6919 A120 3.90 42.5 3.63 34.0 74.5 123.4

Yonkers, NY 2312 A120 3.94 42.0 3.60 34.0 74.0 122.3

Yonkers, NY 2330 A120 3.86 42.7 3.56 35.9 77.5 120.6

IV Measurements

Modules were cooled to 15° C. and IV curves were taken continuously for every degree

20° and 30°C. These IV curves were used to find the temperature coefficient for voltage

and current. IV was taken using Day Star IV curve tracer. Polycrystalline silicon

reference cell was used to measure the irradiance to match the Arizona and New York

modules. Temperature of the module was measured using the thermocouple attached to

the center of the module. All IV curves were taken while the irradiance is above 1000

W/m2 and was achieved by keeping the manual two axis tracker normal to the incident

light. IV curves are translated using the PRL template.

14

Electroluminescence

Using EL imaging, it is possible to detect cell cracks in photovoltaic modules. Cell cracks

appear as dark lines on the solar cell in the EL image. When the electric current is passed

through PV module in forward bias, it emits light in the IR range (1150nm) which is

captured using the sensovation HR-830 pro camera. Current supplied was 1.33*Isc as per

convention. EL image was captured after the exposure time of 30 seconds in the dark

room to avoid noise in the image.[10], [11]

Dark IV

Dark IV was done by keeping the module in the dark room at 25°C. Current is made to

flow from positive to negative by using power source. The voltage and current were

simultaneously and continuously measured as the voltage is increased from zero to open

circuit voltage (Voc). Dark IV is a more effective way to determine series resistance and

shunt resistance than Light IV.[12]

Acceleration Factor

Isc degradation rate is considered instead of power degradation rate for all the ten modules

as the parameter which is affected by browning is short circuit current (Isc).[10]

Degradation rate is found by using the equation

Percent Isc Degradation rate per year = (Initial Isc – Final Isc) ∗ 100

Initial Isc ∗ Number of years

Temperature of the module is calculated by using the Sandia model. It is represented as

[13]

.( )a b WS

m aT E e T

where Tm is the Temperature of the module

15

E = Solar irradiance incident on module surface, (W/m2)

WS = Wind speed measured at standard 10-m height, (m/s)

a = Empirically-determined coefficient establishing the upper limit for module

temperature at low wind speeds and high solar irradiance

b = Empirically-determined coefficient establishing the rate at which module temperature

drops as wind speed increases

Acceleration Test Model

Activation energy is calculated using different acceleration models and acceleration

factor is calculated by

AF = Percent Degradation rate for site1

Percent Degradation rate for site2

3.7.1 Arrhenius Model

This is the most ubiquitous model used to find the degradation due to different modes. It

can be represented as

Acceleration Factor AF = 𝑒−Ea

𝐾 (

1

𝑇1−

1

𝑇2)

where

A is a constant

k = Boltzmann’s constant = 8.615 * 10-5 eV/K

T = Temperature in Kelvin

Ea = Activation energy

3.7.2 Peck’s Model

This model is initially used for finding the degradation of epoxy encapsulation in

hermetic packages[14]. It is represented as

16

Time to failure tf = A (%RH)𝑚 𝑒−Ea

𝑘𝑇

where

m is a discrete variable

%RH = Relative Humidity

3.7.3 Klingers Model

This model is used as an alternative for finding the degradation in hermetic packages. It is

represented as

Time to failure tf = A (%RH

100−𝑅𝐻%)𝑚 𝑒

−Ea

𝑘𝑇

where

m is a discrete variable

3.7.4 Modified Peck’s Model with UV and without Humidity

Pecks model is modified to include UV instead of RH. It is represented as

Time to failure tf = A (%UV)𝑛 𝑒−Ea

𝑘𝑇

where

n is a discrete variable

3.7.5 Modified Peck’s Model with UV and Humidity

In this model all the three stress factors temperature, humidity and UV are considered for

degradation.

Time to failure tf = A (%RH)𝑚(%UV)𝑛 𝑒−Ea

𝑘𝑇

where

m and n are discrete variables

17

Acceleration Testing of Mini-Modules

Eight mini-modules having a similar construction as A60 and A120 modules were chosen

for acceleration testing in Atlas Ci4000 chamber. All the mini-modules have Cerium

oxide in glass similar to Arizona and New York modules. The acceleration tests were

done at a chamber temperature of 20°C, chamber relative humidity of 65%, rack

temperature of 50°C and irradiance of 1 W/m2 at 340nm. It consists of 18 cells, nine cells

on each side. Six cells (3 cells at the center on each side) are considered for the study as

shown in Figure 11 and Figure 12

Figure 11: Front View of the Mini-Module showing the Center Cells

18

Figure 12: Rear View of the Mini-Module

3.8.1 Cutting and Soldering

For characterizing these cells, the back sheet was cut from the back as shown in Figure

13 and Figure 14. A 22 AWG wire of length 3inches was soldered to the solder bond of

these six cells as shown in Figure 15. To prevent moisture from entering the mini-module

through this contact, 5 minute epoxy resin was applied to the contact as shown in Figure

16 and some resin was poured on the other points on the wire to have equal weight

distribution and to ease the pressure on the contact. It was left for 24 hours to cure.

19

Figure 13: Cutting of the Back Sheet using the Blade

Figure 14: Rear View of the Mini-Module after Cutting the Back Sheet at the

Required Position

20

Figure 15: Soldering of the 22 AWG Wire to the Solder Bond of the Mini-Module

Figure 16: Rear View of the Mini-Module after the Curing of Epoxy Resin

21

3.8.2 Insulation

Temperature of the module increases when the insulation is stuck to the backside of the

module [15]. To maintain the mini-modules at different temperatures, insulation with

different thickness was used. Two mini-modules without insulation, three mini-modules

with thin insulation (1/8th inch), and three mini-modules with thick insulation (1/2 inch)

were used as shown in Figure 17.

Figure 17: Mini-Module with no Insulation, Thin Insulation, and Thick Insulation

3.8.3 Data Logger

Data logger was used to measure the temperature of all the mini-modules. Two HOBO 4-

Channel Thermocouple Data Logger - UX120-014M were used for eight mini-modules.

Type-T Thermocouple was attached to the left center cell of the mini-module using the

thermal tape as shown in the figure.

3.8.4 Installation of Mini-Modules in the Chamber

Eight mini-modules were installed in the chamber all around the circumference of the

rack as shown in Figure 18 such that all mini-modules receive the same amount of UV.

Mini-Modules were clamped using the hand clamps (Figure 19) and were secured with

aluminum tape to protect from UV.

22

Figure 18: Mini-Modules Installed in the Chamber using the Handi Clamps

23

Figure 19: Close up View of the Mini-Module Installed using the Handi Clamp

3.8.5 Characterization Tests

Mini-Modules were removed every two weeks for doing characterization tests such as

indoor IV, QE, reflectance and UV fluorescence.

3.8.5.1 Indoor IV using Solar Simulator

IV curve of each cell was measured using indoor solar simulator at an irradiance of 1000

W/m2 using the reference cell as shown in Figure 20. All the mini-modules were kept at

the same position as shown in Figure 20 while taking IV curve to reduce the

measurement error.

24

Figure 20: Light Illuminated on Reference Cell and the Mini-Module using Indoor

Solar Simulator

3.8.5.2 Quantum Efficiency Measurements

A cell QE was performed using PV measurement’s Solar Panel Quantum Efficiency

Measurement System; model QEX12M to obtain QE losses in the shunted regions. QE

was done on each cell every two weeks similar to IV on the exposed region by using the

black shade with holes as shown in Figure 21.

25

Figure 21: Measurement of QE in the Exposed Part

26

4 RESULTS AND DISCUSSION

IV Measurements:

IV curve was taken for ten modules and was translated through ivpc3 software to check if

the modules are affected by defects other than browning as shown in Figure 22. Close up

view near Isc is shown in Figure 23. Only five modules (4958,4961,4968,2330 and 6917)

were considered for finding the acceleration factor as the IV curve of other modules are

inclined near Isc as it may have defects other than encapsulant browning.

Figure 22: Translated IV Curve (STC) for Arizona and New York Modules

0

0.5

1

1.5

2

2.5

3

3.5

4

0 5 10 15 20 25 30 35 40 45 50

Cu

rren

t (A

)

Voltage (V)

6919 6917 frameless 2330 frameless 2312

4960 4968 4958 4961

A002 A003

27

Figure 23: Close up View of Translated IV Curve (STC) near Isc

All the IV curves are translated through PRL template. Translated values are shown in

Table 2.

Table 2 Final Data Translated by PRL Template

Place Serial No Model Isc Voc Imp Vmp FF Pm

Tempe, AZ 4958 A60 3.54 21.6 3.24 17.2 73.0 55.8

Tempe, AZ 4960 A60 3.51 21.6 3.16 17.1 71.1 53.9

Tempe, AZ 4961 A60 3.53 21.6 3.27 17.2 73.6 56.1

Tempe, AZ 4968 A60 3.52 21.5 3.26 16.9 73.0 55.2

Tempe, AZ A002 A60 3.35 21.2 3.13 16.7 73.6 52.2

Tempe, AZ A003 A60 3.28 21.1 3.05 16.9 74.2 51.4

Valhalla, NY 6917 A120 3.51 42.8 3.23 33.9 72.7 109.4

3.3

3.35

3.4

3.45

3.5

3.55

3.6

0 5 10 15 20 25 30 35 40 45 50

Cu

rren

t (A

)

Voltage (V)

6919 6917 frameless 2330 frameless 2312

4960 4968 4958 4961

A002 A003

28

Valhalla, NY 6919 A120 3.49 43.1 3.18 34.2 72.2 108.8

Yonkers, NY 2312 A120 3.56 42.7 3.25 33.9 72.5 110.1

Yonkers, NY 2330 A120 3.48 42.9 3.18 34.4 73.3 109.4

Isc degradation for five modules is shown in Figure 24. It can be clearly seen that the

modules in New York are degraded more than the modules in Arizona. The median Isc

degradation rate per year for Arizona modules is 0.42 %, and modules in New York is

0.58%. The acceleration factor is calculated by taking the ratio of degradation rate in

Arizona to degradation rate in New York and is 0.7271.

Figure 24: Annual Isc Degradation for AZ and NY

0.43%0.40%

0.43% 0.42%

0.62%

0.54%0.58%

0.00%

0.10%

0.20%

0.30%

0.40%

0.50%

0.60%

0.70%

4958 4961 4968 Median of AZ 6917 2330 Median of NY

Arizona New York

% Is

c D

egra

dat

ion

per

yea

r

Percent Isc degradation per year for AZ and NY modules

29

Determination of Activation energy using Field Test Data

Activation energy is calculated using different models and the results are distributed in

two sections- one with UV stress and one without UV stress.

4.2.1 Activation Energy Determination without UV Stress

Arrhenius model, Peck’s model, Klinger’s model are discussed in this section where the

acceleration factor is correlated to either temperature or temperature and humidity but not

UV

4.2.1.1 Arrhenius Model

In this model, activation energy depends only on module temperature. The median

activation energy calculated for this model is -0.15 eV. The negative value clearly shows

that it doesn’t follow this model. Temperature in Arizona is greater than the temperature

in New York. But the Arizona modules have degraded less than the New York modules

which results in negative value

4.2.1.2 Peck’s Model with Ambient Humidity

Activation energy depends on module temperature and ambient humidity in this model.

The median activation energy calculated using different m values 2.5, 2.66, 2.8 and 3 is

shown in the Figure 25[16]. Ambient humidity is considered for the humidity at both

sites. Activation energy for all m values is high (0.83 to 1.03 eV) compared to the

reported value of 0.4 eV.

30

Figure 25: Activation Energy for Peck’s Model with Different m Values Considering

Ambient Humidity

4.2.1.3 Peck’s Model with Module Humidity

Activation energy depends on module temperature and ambient humidity in this model.

The median Activation energy calculated for different m values is shown in Figure 26.

Activation energy considering module humidity is high compared to the previous Peck’s

model with ambient humidity. As the ratio of module humidity is greater than the ratio of

ambient humidity, the activation energy using this model is higher than the Peck’s model

with ambient humidity. The calculated activation energy using this model is more

(1.17eV to 1.43eV) than 1eV for all the m values.

0.830.89

0.95

1.03

0.00

0.20

0.40

0.60

0.80

1.00

1.20

Pecks m=2.5 Pecks m=2.66 Pecks m=2.8 Pecks m=3

Act

ivat

ion

en

ergy

(eV

)

31

Figure 26: Activation Energy for Peck’s Model with Different m Values Considering

Module Humidity

4.2.1.4 Klingers Model

The activation energy calculated using this model with ambient and module humidity

respectively is shown in Figure 27. Activation energy using Klinger model is 0.68 eV and

0.77eV and is near to 0.4 eV. As expected activation energy calculated using ambient

humidity is less compared to activation energy considering module humidity.

1.171.25

1.33

1.43

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

Pecks m=2.5 Pecks m=2.66 Pecks m=2.8 Pecks m=3

Act

ivat

ion

en

ergy

(eV

)

0.77

0.68

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

Klinger m=1 (module humidity) Klinger m=1 (ambient humidity)

Act

ivat

ion

en

ergy

(eV

)

32

Figure 27: Activation Energy for Klinger’s Model with Module and Ambient

Humidity

4.2.2 Activation Energy Determination with UV Stress

In this model acceleration factor is correlated to temperature, humidity, and UV. Previous

models without UV has been modified to include UV.

4.2.2.1 Modified Peck’s Model with UV and without Humidity

Activation energy calculated using this model is -0.89 eV. It is negative because the

temperature and UV in Arizona is high compared to New York and the degradation of

modules in Arizona is lower than the degradation of modules in New York.

4.2.2.2 Modified Peck’s Model with UV and Ambient Humidity

In this model all the three factors-temperature, ambient humidity and UV are considered

as stresses. m (RH exponent) is considered the same as in Peck’s model and n (UV

exponent) varies from 0.6 to 1. The introduction of UV resulted in a decrease in the

activation energy compared to Peck’s model with ambient humidity which can be seen in

Figure 29. Activation energy calculated with m = 2.5 and n = 1 is 0.52eV is close to 0.4

eV as mentioned in the literature. Pecks model with m=2.5,2.66,2.8 and 3 and different n

values are tried to get activation energy of 0.4 eV as shown in Figure 29. But they are not

considered as n cannot be greater than 1 as mentioned in the literature.

33

Figure 28: Peck’s Model with UV- Ambient Humidity for Exponent n Less than 1

Figure 29: Peck’s Model with UV- Ambient Humidity with Different m and n to get

Activation Energy Closer to 0.4 eV

4.2.2.3 Modified Peck’s Model with UV and Module Humidity

In this model, module humidity is considered as a stressor. m is taken as 2.5, 2.66,2.8 and

3 similar to the previous models. Activation energy calculated using this model is shown

0.630.60

0.570.54

0.52

0.690.66

0.630.60

0.58

0.740.72

0.680.65

0.63

0.820.79

0.760.73

0.70

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

0.6 0.7 0.8 0.9 1 0.6 0.7 0.8 0.9 1 0.6 0.7 0.8 0.9 1 0.6 0.7 0.8 0.9 1

m=2.5 m=2.66 m=2.8 m=3

Act

ivat

ion

en

ergy

(eV

)

m- RH coefficient,n - UV coefficient

0.45

0.37

0.3

0.44

0.36

0.3

0.45

0.38

0.310.34

0.49

0.42

0

0.1

0.2

0.3

0.4

0.5

0.6

1.25 1.5 1.75 1.5 1.75 2 1.5 1.75 2 2 1.75 2

m=2.5 m=2.66 m=2.7 m=2.8 m=3

Act

ivat

ion

en

ergy

(eV

)

m- RH coefficient,n - UV coefficient

34

in Figure 30. Activation energy calculated is more than 0.6 eV for all different

combination of m and n exponents. Even though the activation energy calculated with

m=2.5,2.66,2.8,3 and n=3 are closer to 0.4 eV, they are not considered as n is greater than

1 which is not possible.

Figure 30: Peck’s Model with UV- Module Humidity

0.90 0.87 0.84 0.82 0.79

0.98 0.95 0.92 0.89 0.86

1.05 1.02 0.99 0.96 0.93

1.16 1.12 1.09 1.06 1.03

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

0.6 0.7 0.8 0.9 1 0.6 0.7 0.8 0.9 1 0.6 0.7 0.8 0.9 1 0.6 0.7 0.8 0.9 1

m=2.5 m=2.66 m=2.8 m=3

Act

ivat

ion

en

ergy

(eV

)

m- RH exponent,n - UV exponent

35

Figure 31: Peck’s Model with UV- Module Humidity

Determination of Activation Energy using Accelerated Test

Mini-modules were stressed in an acceleration chamber and the characterization tests

were done every two weeks. Activation energy was calculated by considering the short

circuit degradation of cells at different temperatures.

4.3.1 IV Measurements

IV curve for the chosen cells before stressing of mini-module is shown in Figure 32. All

the IV curves are translated to STC. Isc of the cells ranges from 180 mA to 200 mA. On

analyzing the IV curve, it can be found that cells don’t have any major defects. IV curves

for the cells after stressing for two weeks and four weeks are shown in Figure 33 and

Figure 34 respectively. It can be seen that Isc has dropped after stressing.

0.29

0.360.38

0.43

0.51

0

0.1

0.2

0.3

0.4

0.5

0.6

3.00 3.00 3.00 3.00 3.00

m=2.5 m=2.66 m=2.7 m=2.8 m=3

Act

ivat

ion

en

ergy

(eV

)

m- RH exponent,n - UV exponent

36

Figure 32: Light IV for All the Cells of Mini-Module Before Stressing

0.00

0.05

0.10

0.15

0.20

0.25

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Cu

rren

t (A

)

Voltage (V)

37

Figure 33: Light IV for All the Cells of Mini-Module after Stressing for Two Weeks

0.00

0.05

0.10

0.15

0.20

0.25

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Cu

rren

t (A

)

Voltage(V)

38

Figure 34: Light IV for All the Cells of Mini-Module after Stressing for Four Weeks

IV curve for mini-module cell4 before stressing, after stressing for two weeks and four

weeks is shown in Figure 35. Isc has dropped from 185 mA to 159 mA after stressing for

four weeks.

0.00

0.05

0.10

0.15

0.20

0.25

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Cu

rren

t (A

)

Voltage (V)

39

Figure 35: Comparison of IV Curve before and after Stressing for Mini-Module 4

Cell1

Percentage Short circuit current degradation per hour after four weeks of stress testing of

the cells are shown in Figure 36. Temperature reached by the mini-modules vary from

77.6°C to 123.4°C depending on the type of insulation. As some mini-modules are

shaded in the ATLAS chamber, they are not considered for the analysis. All the cells in

mini-module 3 and Cell 4 of the mini-module 6 are the cells which were shaded.

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Cu

rren

t (A

)

Voltage (V)

IV curve beforestressing

IV curve afterstreesing for twoweeksIV curve afterstressing for fourweeks

40

Figure 36: Percent Drop in Isc Per Hour after Four weeks of Accelerated Stress

Testing

It can be clearly seen in Figure 36 that short circuit current of the mini-module 2,3 and 7

have increased. So only mini-modules 1,4,5, 6 and 8 are considered as shown in Figure

37.

-0.0040%

-0.0020%

0.0000%

0.0020%

0.0040%

0.0060%

0.0080%

0.0100%

0.0120%

0.0140%

1 (th

in) C

ell 1

1 (th

in) C

ell 2

1 (th

in) C

ell 4

1 (th

in) C

ell 5

1 (th

in) C

ell 6

2 (th

in) C

ell 1

2 (th

in)C

ell 4

2 (th

in) C

ell 6

3 C

ell 1 (Sh

aded

)

3 C

ell 2 (sh

ade

d)

4 (th

ick) Cell 1

4 (th

ick) Cell 2

4 (th

ick) Cell 3

5 (th

ick) Cell 1

5 (th

ick) Cell 4

5 (th

ick) Cell 5

6 (th

ick)Cell 1

6 (th

ick) Cell 2

6 (th

ick) Cell 3

6 C

ell 4 (Sh

aded

)

7 (n

o in

sulatio

n) C

ell 1

7 (n

o in

sulatio

n) C

ell 2

7 (n

o in

sulatio

n) C

ell 4

7 (n

o in

sulatio

n) C

ell 5

7 (n

o in

sulatio

n) C

ell 6

8 (n

o in

sulatio

n) C

ell 1

8 (n

o in

sulatio

n) C

ell 2

8 (n

o in

sulatio

n) C

ell 3

8 (n

o in

sulatio

n) C

ell 4

8 (n

o in

sulatio

n) C

ell 5

107.6 102.5 98.5 123.4 106.1 105.5 89.4 77.6

Per

cen

t d

rop

in Is

c p

er h

ou

r

Percent drop in Isc for minimodules after four weeks of stress testing

Temperature

41

Figure 37: Percent Drop in Isc for Mini-Modules after Four Weeks of Stress Testing

after Removing the Non-Degraded Mini-Modules in the Increasing Order of

Temperature

Using the degradation rate, Activation energy for browning is calculated by plotting

ln(Percent Isc degradation per hour) vs 1/T(1/K) as shown in Figure 38. The slope is –Ea/k

where k is Boltzmann’s constant. The slope is -4706 if the average for each mini-module

is considered and hence the activation energy is 0.41eV. The slope is -5050.4 if

individual cells are considered as shown in Figure 39 and the activation energy is 0.44eV.

The coefficient of Determination (R2 value) is 0.20 and is very low as small cells with

close to 200mA current are used. The quality of cells was also not good. In order to get

good fit (R2 value), it is recommended to use high quality and big cells with minimum

current of 0.5 A and experiment should be repeated for multiple temperatures to get

activation energy.

-0.0020%

0.0000%

0.0020%

0.0040%

0.0060%

0.0080%

0.0100%

0.0120%

0.0140%

mo

du

le 8 C

ell 1

mo

du

le 8 C

ell 2

mo

du

le 8 C

ell 3

mo

du

le 6 C

ell 1

mo

du

le 6 C

ell 2

mo

du

le 6 C

ell 3

mo

du

le 5 C

ell 1

mo

du

le 5 C

ell 4

mo

du

le 5 C

ell 5

mo

du

le 1 C

ell 1

mo

du

le 1 C

ell 2

mo

du

le 1 C

ell 4

mo

du

le 1 C

ell 6

mo

du

le 4 C

ell 1

mo

du

le 4 C

ell 2

mo

du

le 4 C

ell 3

77.6 105.5 106.1 107.6 123.4

Per

cen

t d

rop

in Is

c p

er h

ou

r

Temperature (°C)

42

Figure 38: Activation Energy for Browning Considering Average Degradation of

Each Mini-Module

Figure 39: Activation energy for Browning Considering Degradation of Each Cell of

Mini-Module

y = -4706.7x + 1.4589

-14

-12

-10

-8

-6

-4

-2

0

0.0025 0.00255 0.0026 0.00265 0.0027 0.00275 0.0028 0.00285 0.0029

ln(%

Isc

Deg

rad

atio

n/

ho

ur)

1/T (1/K)

Activation Energy = 0.40 eV

y = -5050.4x + 2.3292R² = 0.2013

-14

-12

-10

-8

-6

-4

-2

0

0.0025 0.00255 0.0026 0.00265 0.0027 0.00275 0.0028 0.00285 0.0029

ln(%

IscD

egra

dat

ion

/ho

ur)

1/T(1/K)

Activation Energy = 0.43 eV

43

5 CONCLUSION

The activation energy for encapsulant browning reaction of photovoltaic (PV) modules

has been determined using outdoor field degradation data and indoor accelerated

degradation data. All the outdoor field testing and indoor accelerated testing modules

were manufactured by the same manufacturer with glass/EVA/cell/EVA/back sheet

construction. In the outdoor method, the activation energy for the encapsulant browning

is calculated using the degradation rates of short-circuit current (Isc, the response

parameter), weather data (temperature, humidity, and UV, the stress parameters) and

different empirical rate models such as Arrhenius, Peck, Klinger and modified Peck

models. In the indoor method, the activation energy for the encapsulant browning is

calculated using the short circuit current degradation rate and operating temperatures of

the mini-modules. The activation energy for the encapsulant browning obtained from the

outdoor field data and the indoor accelerated data are compared and analyzed in this

work.

Key conclusions from this study are:

1. The annual short-circuit current degradation rate due to browning for modules

fielded in New York (cold climate with about 50%RH) is, surprisingly, more than

the annual short-circuit current degradation rate for modules fielded in Arizona

(hot climate with about 30%RH).

2. Activation energy found using Arrhenius model and modified Peck’s model with

UV and without humidity are negative as both temperature and UV are high in

Arizona compared to New York and the browning causing Isc degradation

doesn’t follow these models.

44

3. Activation energy found using Peck’s model with module humidity is higher than

Peck’s model with ambient humidity as the ratio of module humidity in Arizona

to New York is greater than the ratio of ambient humidity in Arizona to New

York. Both the models gave activation energies ranging between 0.83eV and

1.43eV.

4. Activation energy found using Modified Peck’s model with ambient humidity and

UV and exponentials m = 2.5 and n = 0.6, 0.7, 0.8, 0.9 and 1 have an activation

energy close to 0.4 eV.

5. UV from sunlight causes browning (reaction) whereas oxygen diffusion through

the back sheet causes bleaching (counter reaction). For the future recommended

work, the oxygen diffusion needs to be considered in the empirical models along

with temperature, humidity, and UV.

6. For the fielded modules, the degradation rate was calculated based on the

nameplate rating. If there is any measurement uncertainty in the nameplate rating,

the degradation rate calculation may not be very accurate. So, it is recommended

to use statistically significant number of samples and remove any obvious outliers

for the future evaluations.

45

REFERENCES

[1] P. Klemchuk, M. Ezrin, G. Lavigne, W. Holley, J. Galica, and S. Agro, “Investigation of

the degradation and stabilization of EVA-based encapsulant in field-aged solar energy modules,”

Polym. Degrad. Stab., vol. 55, no. 3, pp. 347–365, 1997.

[2] A. A. Suleske, “Performance Degradation of Grid-Tied Photovoltaic Modules in a Desert

Climatic Condition,” Arizona State University, 2010.

[3] S. Tatapudi, C. Libby, C. Raupp, D. Srinivasan, J. Kuitche, B. Bicer, and G. TamizhMani,

“Defect and Safety Inspection of 6 PV Technologies from 56,000 Modules Representing 257,000

Modules in 4 Climatic Regions of the United States,” Photovolt. Spec. Conf., pp. 2–5, 2016.

[4] Z. Campeau, M. Anderson, E. Hasselbrink, D. Kavulak, Y.-C. Shen, R. Lacerda, A. Terao,

S. Caldwell, Z. Defreitas, and L. Leonard, “SunPower ® Module Degradation Rate”, Internet

Publication from SunPower Corporation, 2013, downloaded in September 2016

from https://us.sunpower.com/sites/sunpower/files/media-library/white-papers/wp-sunpower-

module-40-year-useful-life.pdf

[5] H. T. William H Holley, Susan C.Argo, James P.Galica, Lynee A. Thoma, Robert

S.Yongensen, Dr.Myer Ezrin, Dr. Peter Klemchuk, Gary Lavigne, “Investigation into the causes of

browning in EVA encapsulated flat plate PV modules,” IEEE, vol. 1, pp. 893–896, 1995.

[6] F. Pern, “Ethylene vinyl acetate (EVA) encapsulants for photovoltaic modules:

Degradation and discoloration mechanisms and formulation modifications for improved,” Die

Angew. Makromol. Chemie, vol. 252, pp. 195–216, 1997.

[7] F. J. Pern and A. W. Czanderna, “Characterization of ethylene vinyl acetate (EVA)

encapsulant: Effects of thermal processing and weathering degradation on its discoloration,” Sol.

Energy Mater. Sol. Cells, vol. 25, no. 1–2, pp. 3–23, 1992.

[8] B. Nehme, N. K. M’Sirdi, T. Akiki, and A. Naamane, “Contribution to the modeling of

ageing effects in PV cells and modules,” Energy Procedia, vol. 62, pp. 565–575, 2014.

[9] D. C. Miller, E. Annigoni, A. Ballion, J. G. Bokria, L. S. Bruckman, D. M. Burns, X. Chen,

L. Elliott, J. Feng, H. Roger, S. Fowler, X. Gu, P. L. Hacke, C. C. Honeker, M. D. Kempe, M.

Köhl, N. H. Phillips, K. P. Scott, et al., “Degradation in PV Encapsulation Transmittance : An

Interlaboratory Study Towards a Climate-Specific Test,” Photovolt. Spec. Conf., pp. 1–6.

[10] M. Chicca, “Characterization and Analysis of Long Term Field Aged Photovoltaic

Modules and Encapsulant Materials,” Arizona State University, 2015.

46

[11] M. Köntges, S. Kurtz, C. Packard, U. Jahn, K. A. Berger, K. Kato, T. Friesen, H. Liu, and

M. Van Iseghem, Review of Failures of Photovoltaic Modules. 2014.

[12] D. L. King, R. Hansen, J. A. Kratochvil, and M. A. Quintana, “Dark Current-Voltage

Measurements on Photovoltaic modules as a diagnostic or manufacturing tool,” Photovolt. Spec.

Conf., pp. 1125–1128, 1997.

[13] D. L. King, W. E. Boyson, and J. A. Kratochvil, “Photovoltaic array performance model,”

Sandia Rep. No. 2004-3535, pp. 1–43, 2004.

[14] O. Hallberg, “Humidity Accelerations for Testing Standards,” 3rdEuropean Symposium on

Reliability of Electron Devices, Failure Physics and Analysis, Schwabisch Gmund, Germany, 5-8

October 1992.

[15] A. Pavgi, “Temperature Coefficients and Thermal Uniformity Mapping of PV Modules

and Plants,” Arizona State University, 2016.

[16] R. Laronde, A. Charki, and D. Bigaud, “Lifetime Estimation of a Photovoltaic Module

Subjected to Corrosion Due to Damp Heat Testing,” J. Sol. Energy Eng. Asme, vol. 135, pp. 1–8,

2013.