© Fraunhofer ISE

Is it possible to design accelerated servicelife tests for PV modules?

Michael Köhl

Fraunhofer Institute for Solar Energy Systems

Freiburg, Germany

Presented at the EMPA Workshop

Durability of Thin Film Solar Cells

April 2011

© Fraunhofer ISE

General methodology for service life assessment

Assumptions

1. Only long-term wear out degradation is considered

2. The primary degradation factors are due to weathering

3. The stress-levels depend on local climate and installation

4. The stress-levels depend on the micro-climate at the module

5. The test samples (PV-modules or components) have to be considered as a black-box

6. The modelling is based on investigation of the degradation kinetics of real state-of-the-art modules

7. A service life of 25 years is required

© Fraunhofer ISE

General methodology

Local climate

PV-modules

Micro-climateMaterials

ALT conditions

Service life

Modeling the Accelerated Life Test conditions based on realistic loads

Models

Degradation

Performance change

© Fraunhofer ISE

General methodology Step 1: Outdoor exposure and climate monitoring

City or reference:

Freiburg Germany

Desert

Sede Boqer

Israel

Alpes

Zugspitze

Germany

Tropical

Serpang

Indonesia

Maritime

Pozo Izquierdo

Gran Canaria

Monitoring degradation factors for modelling degradation

Measurement of module performance over time for validation of ALT

© Fraunhofer ISE

Monitoring in Cologne (one year)

0 50 100 150 200 250 300 3500

20

40

60

80

100

RH

[%]

Zeit [d]0 50 100 150 200 250 300 350

-20

-10

0

10

20

30

40

50

60 T Modul T ambient

Tem

pera

tur [

°C]

Zeit [d]

Ambient humidity ambient and module temperatures

© Fraunhofer ISE

Monitoring in the Alpes (one year)

0 50 100 150 200 250 300 3500

20

40

60

80

100

RH

[%]

Zeit [d]0 50 100 150 200 250 300 350

-30

-20

-10

0

10

20

30

40

50

60 T Modul T ambient

Tem

pera

tur [

°C]

Zeit [d]

Ambient humidity ambient and module temperatures

© Fraunhofer ISE

0 50 100 150 200 250 300 3500

20

40

60

80

100

RH

[%]

Zeit [d]0 50 100 150 200 250 300 350

-20

-10

0

10

20

30

40

50

60

70 T Modul T ambient

Tem

pera

tur [

°C]

Zeit [d]

Ambient humidity ambient and module temperatures

Monitoring in the desert (one year)

© Fraunhofer ISE

Ambient humidity ambient and module temperatures

0 50 100 150 200 250 300 3500

20

40

60

80

100

RH

[%]

Zeit [d]0 50 100 150 200 250 300 350

10

20

30

40

50

60

70 T Modul T ambient

Tem

pera

tur [

°C]

Zeit [d]

Monitoring in tropical Serpong (one year)

© Fraunhofer ISE

General methodology Step 2: Micro - climate

Micro-climate : = stresses for the material caused by interaction of materials and ambient climate

1. Module temperature modeled by solar irradiation, ambient temperature, wind speed

2. Module surface humidity modeled by module temperature, ambient temperature and humidity

3. UV-radiation modeled from solar irradiation and spectral transmittance of laminated materials

4. Temperature cycles of module temperature

5. Leakage current as function of voltage, module temperature and surface humidity

6. Salt concentration correlated with wetting/drying cycles

7. …….

© Fraunhofer ISE

Micro-climate of modules

Module – Temperature in the desert

Day

Night

© Fraunhofer ISE

-20 -10 0 10 20 30 40 50 60 700

200

400

600

800

1000

1200

1400

1600 tropisch arid gemäßigt alpin

freq

uen

cy [

h]

Temperature / °C

Outdoor weathering with temperature monitoring

Ambient temperature Average module temperatur e(c-Si)

Macro-climate => Micro-climate

-20 -10 0 10 20 30 40 50 60 700

200

400

600

800

1000

1200

1400

1600

freq

uen

cy [

h]

Temperature / °C

tropical arid moderate alpine marine

Micro-climate of modules

© Fraunhofer ISE

Histogram of measured module temperatures in Cadarache, F

-20 0 20 40 60 80

0

200

400

600

800

1000

1200

1400

1600

1800

2000

2200

coun

t

temperature in °C

Avancis Flexcell FirstSolar Schott Sharp UniSolar Wuerth SchottCSIRef

Cadarache thin film moduleshistogramm of measured module temperature

Special thanks to Antoine Guerin de Montgareuil,

CEA-INES, Cadarache, France

a-Si 1 CIS 1 c-Si

a-Si 3 CIS 2

a-Si 4 CdTe

What about Thin Film Modules?

Module temperature for one year

© Fraunhofer ISE

Physical modeling of module temperature for each of the different modutypes using David Faiman‘s approach (could be King, Fuentes……as well)

v10mod

UU

HTT amb

Tmod module temperature

Tamb ambient temperature

v wind velocity

H solar radiation

U0 , U1 = module dependent parameters

Neglected: IR-radiation exchange and natural convection

U1 U0a-Si 1 10,7 25,7a-Si 3 5,8 25,8a-Si 4 4,3 26,1

CIS 1 3,1 23,0CIS 2 4,1 25,0CdTe 5,4 23,4c-Si 6,2 30,0

The parameters U are module-specific but location independent

M.Koehl et.al.: Modelling of the nominal operating cell temperature based on outdoor weathering, Sol. Energy Mat. Sol. Cells (2011)

Macro – climate => Micro – climate

Irradiation, wind, ambient temperature => Tmod

© Fraunhofer ISE

0 20 40 60 800

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Freu

ency

Temperature in °C

a-Si 1 a-Si 2 a-Si 3 a-Si 4 CIS 1 CIS 2 CdTe c-Si

Histogram of simulated module temperatures for one year in the Negev

c-Si CIGS

Micro-climate of Modules

© Fraunhofer ISE

0 10 20 30 40 50 600

10

20

30

40

50

60

Daily module temperature difference /K

Gran Canaria 2011

0 10 20 30 40 50 600

10

20

30

40

50

60

Prob

abili

ty d

ensi

ty

Daily module temperature difference /K

Negev 2009

0 10 20 30 40 50 600

10

20

30

40

50

60

Daily module temperature difference /K

Cologne 2009

0 10 20 30 40 50 600

10

20

30

40

50

60

Prob

abili

ty d

ensi

ty

Daily module temperature difference /K

Indonesia 2007

Micro-climate of modules

Daily temperature cycling during one year

0 10 20 30 40 50 600

10

20

30

40

50

60

Prob

abili

ty d

ensi

tyDaily module temperature difference /K

Alpes 2008

Type approval testing:

200 cycles from -40 to 85°C

Max. changes 100 K/h

With current injection

© Fraunhofer ISE

Surface humidity – desert

Night-time = condensation

Day-time = Drying

Micro-climate of modules

© Fraunhofer ISE

00:00 04:00 08:00 12:00 16:00 20:00 24:000,1

1

10

100

1000

10000

I leak

age

/ nA

Time / hh:mm

M17 - sunny day M18 - sunny day M17 - rainy day M18 - rainy day

Micro-climate of modules

Leakage currents as source of Potential Induced Degradation

Leakage currents depend on the module temperature, the voltage and the surface humidity

© Fraunhofer ISE

Leakage current as function of potential, relative humidity and temperature

I =G/(1+ exp(-G*(rh-0,35)*1,75) *(G/f(0)-1))with G = 13μA

Humidity and potential-induced degradation (PID) of modules

0 100 200 300 400 500 6000

5

10

15

20

25

30

25-25 55-25 85-25 85-40 85-55 85-70 85-85

I/µA

U / V

T[°C] rh [%]

0

2

4

6

8

10

12

14

0 20 40 60 80 100

Leak

age

curr

ent /

µA

Rel. humidity / %

Hoffmann, S., M. Koehl, Effect of Humidity and Temperature on the Potential Induced Degradation, accepted by PIP, 2012

Micro-climate of modules

© Fraunhofer ISE

General Methodology Step 3: Time-transformation functions

Time-transformation functions

1. Module temperature: Arrhenius, Eyring

2. Module surface humidity impact: power law, TOW

3. UV-radiation: Dose-function, reciprocity?

4. Temperature cycles of module temperature: Coffin-Manson

5. Potential induced degradation

6. Salt concentration correlated with wetting/drying cycles: ?

7. …….

Modelling of the degradation processes as function of the degradation factors

© Fraunhofer ISE

Time-transformation functions

Changes of performance or degradation indicator P = ti *(

Temperature + A exp[-EA /RTi]

Humidity + B f(rh)i exp[-EB /RTi]

UV-radiation + C Ini exp[-EC /RTi]

Temperature cycles + D f(T)i exp[-ED /RTi]

Potential-induced Deg. + E f(P)i fp(rh)i exp[-EE /RTi]

Salt + F f(S)i fp (rh)i exp[-EF /RTi]...........)

© Fraunhofer ISE

Degradation factor:

Temperature

Moisture

UV-Radiation

T cycles

Potential I D

Salt

………

Simple deterministic model for aging processes: Time-transformation functionsChanges of property P after the testing time ti

Sample dependent degradatioprocess parameters

P = mi=1 {ti (

+ A exp[-EA /RTi]

+ B f(rh)i exp[-EB /RTi]

+ C Ini exp[-EC /RTi]

+ D f(T)i exp[-ED /RTi]

+ E f(P)i fp(rh)i exp[-EE /RTi]

+ F f(S)i fp (rh)i exp[-EF /RTi]

+…… X Ini f(X) i exp[-EX /RTi] …..)}

© Fraunhofer ISE

ALT – conditions for different locations: temperature impact

Accelerated life testingEquivalent lab tests

(same changes of performanceor degradation indicator as after service life)

by integration of the outdoor stresses

Difference in testing time between 8 and 20

0 50 100 150 200100

1000

10000

100000

8000 h

2100 h

4200 hTe

stin

g tim

e at

85°

C fo

r 25

year

s in

h

Activation energy in kJ/mol

South France Tropic Arid Moderate, urban Alpine

16000 h

Temperature stress

General methodology Step 4: Accelerated life testing conditions

© Fraunhofer ISE

Corresponding temperature testing times at 85°C for 25 a exposure in Cadarache, France

based on monitored module temperatures

50 100 150 200 250 3001

10

100

1000

10000

100000

log

test

ing

time

[h]

activation energy [kJ/mol]

Avancis Flexcell First Solar Schott Sharp Unisolar Wuerth SchottCSIRef

testing times @ 85°C for different thin film modules exposed in Cadarache

Different cell-types:

Factor 2 – 4 in testing time(depending on the degradation processes)

In the range of damp/heat

7 different module types

© Fraunhofer ISE

2025 50 75 100 125 15010

100

1000

10000equivalent testing times for five suns UV test for desert climate loads

60 °C 70 °C 80 °C 90 °C

activation energy [kJ/mol]

test

ing

time

[h]

1700 h

5700 h

Cumulated dose of UV- and solar radiation:

120 kWh/m² (about 8 x IEC)

Outdoor testing with radiation monitoring for one year in the desert

0

1

2

3

4

5

6

-10 0 10 20 30 40 50 60 700

20

40

60

80

100

120 Global UV

glob

al ir

radi

atio

n pe

r tem

p.-b

in[k

Wh/

m²a

]

Module temperature [°C]

UV

dose

pro

tem

p.-b

in [k

Wh/

m² a

]

UV = 5.X % of solar radiation

Reciprocity: p = 1

ttest = (Ii / Itest)p ti exp [-(Ea / R)(1/Ttest - 1/Ti)]

ALT – conditions for different locations: UV-radiation impact

© Fraunhofer ISE

Effective surface humidity

rheff =1/(1+ 1000/exp(rh*k))

20 30 40 50 60 70 80 90 100 110 120 130 140100

1000

100002000 h

test

ing

time

@85

°C/8

5h fo

r 25

a se

rvic

e lif

e [h

]

activiation energy [kJ/mol]

Tropic Arid Alpine

2000 h

Humidity

M. Koehl et. al., Solar Energy Materials & Solar Cells 99 (2012) 282–291

ALT – conditions for different locations: humidity impact

00,10,20,30,40,50,60,70,80,9

11,1

0 0,5 1

Rel. surface humidity

© Fraunhofer ISE

Simulated histograms of the relative humidity

Ambient humidity = partial pressure / saturation pressure (Tamb)

Surface humidity = partial pressure / saturation pressure (Tmodul)

-20 0 20 40 60 801

10

100

1000

Goodwin Creek 2007-01-01-bis-2007-12-31_ (type cSi)histogram of the periods with high moisture

frequ

ency

dis

tribu

tion

in h

ours

module temperature in °C

rheff (sambient) rh (surface) rheff (surface) = 85%

-20 0 20 40 60 801

10

100

1000

rheff (sambient) rh (surface) rheff (surface) = 85%

Desert Rock 2007-01-01-bis-2007-12-31_ (type TF)histogram of the periods with high moisture

module temperature in °C

Humidity dose: teff = t * rheff / 0.85

Eff. Humidity: rheff = 1/(1+ exp(-rh*k) *(1/f(0)-1))

© Fraunhofer ISE

-500 0 500 1000 1500 2000 2500 3000 3500 4000 45000,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

1,1

1 2 3 4 5 6 7

Nor

mal

ised

pow

er

Exposure /h

Damp-Heat at 85°C and 85% rel. humidity

-500 0 500 1000 1500 2000 2500 3000 3500 4000 45000,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

1,1

1 2 3 4 5 6 7

Nor

mal

ised

pow

er

Exposure /h

General methodology Step 5: Accelerated life testing

Damp-Heat at 90°C and 85% rel. humidity

Seven different commercial c-Si modules

M. Koehl et. al., PV reliability (Cluster II): Results of a German four-year joint project - Part I, results accelerated ageing tests and modelling of degradation, 25th EU-PVSEC (2010)

© Fraunhofer ISE

-500 0 500 1000 1500 2000 2500 3000 3500 4000 45000,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

1,1N

orm

alis

ed p

ow

er

Testing time /h

Damp-Heat-Prüfung 85%rh @ 85°C

Damp-heat testing at 85%rh and 85°C, module 1

Induction phase

Degradation phase

Degradation function

© Fraunhofer ISE

-500 0 500 1000 1500 2000 2500 3000 3500 4000 45000,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

1,1

No

rmal

ised

po

wer

Testing time /h

2300 h3200 h

Service lifetime @ 0.8 Pmpp

Damp-heat testing at 85%rh and 85°C, module 1 and module 2

© Fraunhofer ISE

-500 0 500 1000 1500 2000 2500 3000 3500 4000 45000,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

1,1

No

rmal

ised

po

wer

Exposure time /h

Damp-heat tests at 85%rh@85°C and 85%rh@90°C (large dots), Module 1 and module 2

© Fraunhofer ISE

-500 0 500 1000 1500 2000 2500 3000 3500 4000 45000,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

1,1

Module 1 65 kJ/mol Module 2 33 kJ/mol

No

rmal

ised

po

wer

Exposure time /h

a = exp [-(Ea / R)(1/T1 – 1/T2)]

Damp-heat tests at 85%rh@85°C and 85%rh@90°C (large dots), Module 1 and module 2

© Fraunhofer ISE

20 40 60 80 100 120100

1000

10000

100000

test

ing

tim

e @

85°C

[h

]

Activation energy [kJ/mol]

tropical arid alpine

Equivalent testing times (25 years at 85%rh @ 85°C

Module 1 would be suitable for all climates

Module 2 would survive in the mountains

Service lifetime

© Fraunhofer ISE

-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 150

5

10

15

20

25

30

35

40

45

50

55

60

65

70

Displacement on cell (cm)

Inte

nsity

(arb

. uni

ts)

Average Ref._I(1660) Average 4000h d.h._I(1660)

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4 Average Ref._I(2863)/I(2913) Average 4000h d.h._I(2863)/I(2913)

Comparison of Vinyl-Band (red) and fluorescence back-ground (black)initial data and after 4000h damp-heat testing

Polymer Analysis with Raman-Spectroscopy and cell analysis by electroluminescence

Elektroluminescence-image of degraded cells

Location on cell [cm]

Inte

nsi

ty [

ar. u

nit

s]

Peike et.al.: Non-destructive degradation analysis of encapsulants in PV modules by Raman Spectroscopy, Sol. En. Mat. Sol. Cells (20

General methodology Step 6: Analysis of materials degradation

© Fraunhofer ISE

0 1000 h

GlasEVA EVA SiBS

3 mm 0,5 mm

Numerical simulation of energy and mass transport

Water vapour permeation and -diffusion in the back-sheet and the Encapsulant during damp-heat testing (85%rh @85°C)

General methodology Step 7: Simulation of materials degradation

© Fraunhofer ISE

time/d Tropic (hot and humid)

Between cell and glass

Behind the cell

time [d]

Simulation with real climate data

Alps (cold and dry)

Water vapour permeation and - diffusion in the Back-sheet and in the Encapsulant during damp-heat testing (85%rh @85°C)

time [d]

Numerical simulation of energy and mass transport

Wat

er c

once

ntra

tion

[rel

. uni

ts]

© Fraunhofer ISE

What happens with PV – modules in operation?

Degradation processes are induced or caused by transport phenomena

Time constant energy mech. tension charge mass

about zero light

seconds vibrations

minutes heat thermo-mech

hours to years ions oxygen, water vapo

years pollutants (salt, etc.

© Fraunhofer ISE

UV-source for combined UV and humidity tests

200-250 W/m²

2*3*1,6m²

Multiple stress testing

0 – 100°C

0 – 85%rF

Physical limits of combined testing

Testing with radiation requireslarge areas (expensive)

Concentration is needed foracceleration (4-5 X)

Highly efficient irradiation sourcesrequire lamp cooling

Water cooling limits the freezingat temperature cycling

Solar simulator (incl. VIS and NIR) causes heating of the samples

Cooling of the samples spoils high humidity by condensation at theheat exchanger

How to integrate mechanicalloads

General methodology Step 8: Service life testing

© Fraunhofer ISE

Summary

Time-series of climatic data ambient temperature and humidity, solar irradiation, wind speed

Modeling the module temperatures ambient temperature, solar irradiation, wind speed, module-specific coefficients (mounting situation might be considered)

Modeling the UV-radiation 5.5% of the solar radiation, module temperature

Modeling the effective surface humidity ambient temperature and humidity, module temperature

Modelling the micro-climatic stress conditions

© Fraunhofer ISE

Summary

Use a simple time-transformation function (Arrhenius based, eg)Time, module temperature and other degradation factors, but separately first

Modeling the module temperature stress as function of the material-specific activation energy,

(could be eventually included in damp-heat testing)

Modeling the UV-radiation impactas function of the material-specific activation energy (which is low, UV-dose more importan

Modeling the moisture test Higher test temperatures needed, as function of the material-specific activation energy,

Modelling the ALT conditions

© Fraunhofer ISE

Conclusions

One test:

Infant mortality, quality Ieststype approval testing acc. to IEC or UL

Enhanced stress testing:

Infant mortality, higher quality requirementsoffered by a number of test labs

Degradation over time:

Performance, materials or degradation indicator over time=> Changes of micro-climate (stress) because of material changesstability beyond infant mortality, induction periods, stress factor sensitivity

Single constant stress testing

Needed for service life testing:

Performance or degradation indicator over time until failure

© Fraunhofer ISE

Conclusions

Temperature cycling:

Thermo-mechanical stressNo scientific base for type approval testing acc. to IEC or UL

Which relaxation time at which temperature?

Temperature cycling with humidity:

Closer to reality, takes into account temperature dependence of water vapour permeation

Voltage cycling or UV-radiation cycling:

Dark periods allow recovery or diffusion of reactants

Single cyclic stress testing

Needed for service life testing:

Investigation of relaxation times and diffusion processesFrequency and amplitudes of dynamic mechanical testing

© Fraunhofer ISE

Conclusions

Reasons:

Material changes caused by a degradation process due to stress factor 1 might changethe micro-climate from stress factor 2

A combination of stress factors might cause new degradation processes(Photodegradation and hydrolysis, hydrolysis and corrosion)

Problems:

How to design life-tests for cdegradation changed micro-climatic stress?

How to define accelerated life tests with similar acceleration factors for all stress factors taking into account different time constants?

Multiple stress testing

Needed for service life testing:

A big number of unknown factors have to be determined:P = A ti exp[-EA /RTi] + B ti f(rh)i exp[-EB /RTi] + C ti Ini exp[-EC /RTi] + D ti f(T)i exp[-ED /RTi] + E ti f(P)i fp(rh)i exp[-EE /RTi] + Fpq ti f(Sp)i f (Sq)i exp[-Epq /RTi]

Experimental design for a respective number of tests at different stress levels

© Fraunhofer ISE

for your attentionThanksTo my colleagues

Daniel Philipp

Franz Brucker

Stefan Hoffmann

Philipp Huelsmann

Markus Heck

Stefan Brachmann

Karl-Anders Weiss

Stefan Wiesmeier

To our partners

TÜV Rheinland

Schott Solar

Solarfabrik

Solarwatt

Solarworld

Solon

© Fraunhofer ISE

Lugano

Switzerland

Mai 3 – 4 2012

Testing

Analysing

Simulating module - reliabilty

Meeting of the IEC TC82 WG2

After the Workshop in Stresa

Workshop on Reliability of PV-Modules

www.supsi.ch/go/pv-module-reliability

Organised by Fraunhofer ISE and SUPSI

© Fraunhofer ISE

Structure of the workshop

The topics will be presented by experts and further developed in small discussion groups.

Block I: Mechanics

Block II : PID -Humidity (Potential induced degradation)

Block III : UV –Humidity

Block IV : Failure modes and effects

.

Block V: Materials

Plenary discussion with presentation of discussion results

Optional: Visit of the outdoor exposure test site of ISAAC Supsi