CAMPUS DOE, Office of Basic Energy Sciences

Adhesion and Thermomechanical ReliabilityCenter for Advanced Molecular Photovoltaics

STAN FORD UN IVERSITY

e gy Sc e ces

Vit li B d Ch i B F d N J ff Y

Adhesion and Thermomechanical Reliability for PV Devices and Modules

Vitali Brand, Chris Bruner, Fernando Novoa, Jeff Yang (PhD students)

Monika Kummel (Post Doc)Reinhold H. Dauskardt (dauskardt@stanford.edu)

Collaborators:Craig Peters, Roman Gysel and Mike McGehee (Stanford)

Uraib Aboudi and Peter Peumans (Stanford)Uraib Aboudi and Peter Peumans (Stanford)David Miller, Michael Kempe and Sarah Kurtz (NREL)

William Hou and Yang Yang (UCLA)

Industry collaborators include Colin Reese, Shawn Scully and Juanita Kurtin (SpectraWatt), Steve Lin (Vitex), Matthew Robinson (Nanosolar), Christine McGuiness and Darin Laird

(Plextronics).

Solar OutlookThen To make a difference

1. Efficiency2 R li bilit

1. Cost2 R li bilit

Then… To make a difference…

2. Reliability3. Cost

2. Reliability3. Efficiency

• Engineer durable solar technologies with robust and predictable lifetimes. Start early in development – avoid roadblocks.y

• Leverage from reliability physics in microelectronics – thin-film metrologies, kinetic models, accelerated tests, life prediction.

• Are degradation processes coupled and how?

• Kinetic models for damage evolution - basis for life prediction and accelerated testing (T environment stress solar flux etc )accelerated testing (T, environment, stress, solar flux, etc.)

• Effective defensive strategies – e.g. transparent barriers with anti-reflective properties.

Degradation and Reliability of PV Devices and ModulesH2O, O2, H2

other active chemicalUV ExposureH2O, O2, H2

other active chemicalUV Exposure

Severe operating environments.

Exposure to thermal cycling, stress, moisture, chemically

other active chemical species

h t h i lcracking and

surface weathering

other active chemical species

h t h i lcracking and

surface weathering

, , yactive environmental species, and UV.

Uncertain degradation kinetics

photochemical reactions

cracking and debonding

defect evolution in

photochemical reactions

cracking and debonding

defect evolution in and reliability models.Substrate nanomaterial

layersSubstrate nanomaterial

layers

AM1 solar spectrum Activation Energies for Bond Rupture:VDW ~ 0.001 - .4 eV/bond H-bonds ~ .1 - .4 eV/bond

Ester Hydrolysis (100% RH)Sun in UV

Ester Hydrolysis (100% RH) E* ~ .81 eV/bond

SiO2 cracking in water E* ~ 1.39 eV/bond

Solar spectrum from B. Van Zeghbroeck, U. Colorado

Evolution of Defects and Device Reliabilityabsence of chemically active environmental species, damage propagates ify p , g p p g

2/ mJGG cpresence of chemical species and photons, damage propagates even if

environment and stress accelerates 2/ mJGG stress acceleratesdefect evolution

Role of coupled kinetic parameters:

/ mJGG c

H2O

h• mechanical stress

• temperatureH2O

• environmental species

• photons

Substrate(photochemical reactions)

Typical Film Adhesion Tests that Don’t Work WellIndenter

Interface Crack

IndenterPlastic zone

S b t t

Film• Indentation/Scratch Test

- complex stress and deformation fields- principally qualitative results Substrate

MP

p p y q- (nano) scratch test even less quantitative

Interface CrackFilm

• Peel/m-ELT Test- difficult to apply loads- plastic deformation of film

t t li ti i ELT Substrate

Film

- temperature complications in m-ELT

• Blister Test

PSubstrateFilm- compliant loading system

- environmental effects- etching/machining of cavity difficult

Cavity in Substrate

Major limitations: need detailed film properties, film stress relaxation and film plasticity principally qualitative results for all above methods!

Quantitative Adhesion/Cohesion and Debond KineticsP/2P/2

Pre-notch

Thin films

43

P/2P/2Pre-notch

Thin films

43

Al (100 nm)h

b

LL

a

Thin filmsh

b

LL

a

Thin films

P 0

P 0

FPB adhesionITO (150 nm )

PEDOT:PSS (50-100 nm)

( )Ca (7 nm)

P3HT/PCBM (~150 nm)cohesive

failure

Degradation Kinetics

a

P

Thin films

bh

a

P

Thin films

bh

DCB cohesion

Glass Substrate

10-5

10-4

(m/s

)

UV intensity2

J/m

2 )

Degradation Kinetics(temp/environment /UV effects)

Adhesion/Cohesion

DTS system and support contact dauskardt@stanford.edu

10-7

10-6

10

Rat

e, d

a/dt

No UV

1.2 mW/cm2

1

1.5

ergy

, G (J

10 10

10-9

10-8

k G

row

th R 0 mW/cm2

0 60.5

1

hesi

on E

n

threshold crucial

0 1 2 3 4 5 6 710-11

10-10

Cra

ck

Strain Energy Release Rate, G (J/m2)

0.6 mW/cm2

0

Coh

Solar Cells

for reliability

Inherent Solar Cell Thermomechanical Reliability, Gc

Grad students: Vitali Brand, Jeff Yang and Chris Bruner

Adhesion/Cohesion Sample Preparation

Glass Substrate

Thin films sandwiched between elastic substrates

P/2P/2Pre-notch

43

P/2P/2Pre-notch

43

Al (100 nm)Ca (7 nm)

Epoxy Bond (2 m) h

b

LL

a

Thin filmsh

b

LL

a

Thin films

FPB adhesion

ITO (150 nm )PEDOT:PSS (50-100 nm)

P3HT/PCBM (200 nm)P

Thin films

bh

0P

Thin films

bh

0

Gunes, et. Al. Chem. Rev. 2007.

Glass Substratea

P

a

P DCB cohesion

Fabricated 4-point bend adhesion and DCB cohesion test structures using standard epoxy bonding techniques.Si il t t l b t t h idSimilar transparent glass substrates on each side.

8

Adhesion/Cohesion of P3HT/PCBM Structures

6

m2 )

Std. 1:1 P3HT:PCBM

Thicker P3HT:PCBM

Al (100 nm)Ca (7 nm)

P3HT/PCBM (200 nm)cohesive4

rgy,

G (J

/m

compositionP3HT:PCBM1:0 3:1 1:3

G S

ITO (150 nm )PEDOT:PSS (50-100 nm)

P3HT/PCBM (200 nm) failure

2

ctur

e E

ner

Glass Substrate

0

Frac

Solar Cells

• XPS reveals similar debond path for DCB and 4-pt bend samples• C ~ 92%, S ~ 6%, O ~ 2%

Suggests cohesive failure in PCBM:P3HT layerXPS Spectra

• Suggests cohesive failure in PCBM:P3HT layer

Factors Effecting Cohesion of P3HT/PCBM Layers

• Heterojunction layer thickness 16

18

20

PAE/SiCN (Mixed-Mode)

J/m

2 )

• Heterojunction layer thickness– cohesion in polymer layers is sensitive

to layer thickness– plastic energy dissipation in organic layers 6

8

10

12

14

16( )

PAE/SiO2 (Mixed-Mode)

PAE/SiCN (Mode I)

e En

ergy

, Gc (

J

plastic energy dissipation in organic layers

0 50 100 150 200 2500

2

4

6

Frac

ture

Film Thickness, h (nm)

Kearney, Dauskardt, et.al. Acta Mat. 2008

• Composition of the heterojunction layer– limited bonding to fullerene – expect low cohesion– preliminary measurements indicate higher ratios

( )

p y gof P3HT to PCBM make stronger active layer

• AnnealingAnnealing– morphology of the P3HT:PCBM film

changes with annealing, expect effect of morphology on cohesion

Thermal Anneal (150C)before 30 min 2 h

TEM of P3HT:PCBM filmHeeger, et. Al. Adv. Func. Mat. 2005.

Effect of Composition of BHJ on Cohesion

33

(J/m

2 ) Failure path 75% P3HTP3HT/PCBM

2

ergy

, Gc

glass

1

ure

Ene

Failure pathfor 100% P3HT

P3HT

0

Frac

t

glass

0 25 50 75 100Composition P3HT/PCBM (%)

Cohesion increases with P3HT due to increased network formation

Phase TopographyTopography

AFM of Failure Path Near Ca Interface

5 µmp g p y

Ra = 20 nmRq = 24 nm

Annealing effects

AlCa

0 5

ea g e ec sduring Al vapor deposition is a

possibility20 µmcohesive surfaces

0.5 µm

1 µmTopography

PEDOT:PSSITO

P3HT/PCBM

ITO

R = 25 nm

GlSubstrate

Glass Substrate

Ra = 25 nmRq = 30 nm

Small Molecule Solar Cell Thin Films

C60 CuPc BCPC60 CuPc BCP

1.E-041.E-03

m/s

)

10-3

m/s

) Gc ~ 1.1 J/m223oCGc 1.1 0.4 J m2Ryan Birringer, Uraib Aboudi and Peter Peumans

BCP (~ 20 nm)Ag (~ 100 nm)

1.E-071.E-061.E-05

eloc

ity, v

(m

10-7

10-5

Velo

city

, v (m 50% R.H.

c

CuPc (~ 12 nm)CuPc:C60 (~ 16 nm)

C60 (~ 38 nm)( )

1.E-101.E-091.E-08

Deb

ond

Ve

10-9

11

Deb

ond

V

Threshold ~

0.7 J/m2

Glass SubstrateITO 1.E-11

0 0.5 1 1.5

Debond Driving Force, G (J/m2)

D 10-11

Molecular Bond Rupture KineticsMolecular Bond Rupture Kinetics(Barrier Films)

Grad student: Fernando Novoa and Monika Kummel

environment and 2/JGG stress acceleratesdefect evolution

2/ mJGG c

Weathering Test of Polysiloxane Barrier

UV exposure: 28 mW/cm2 at 6 mm UV-257nm

120 min 15 min 5 min120 min…………….15 min……………………5 min

Environment and Stress Accelerates DamageSi l t d UV EDoes UV exposure accelerate decohesion in solar cells? Simulated UV Exposure

applied load

Does UV exposure accelerate decohesion in solar cells?

What are the kinetics?

O i

UV Exposuretransparent substrate

Substrate

OrganicLayer Environmental Chamber

(gaseous and aqueous)

Debond TipDebond

Substrate

h Debond Tiph Debond Tip Reactions

Activation Energies:VDW ~ 0.001 - .4 eV/bond H-bonds ~ .1 - .4 eV/bond

Ester Hydrolysis (100% RH)

Substrate

y y ( )E* ~ .81 eV/bond

SiO2 cracking in water E* ~ 1.39 eV/bond

Assessing UV and Environment on Debonding KineticsSimulated UV Exposure

Glass Substrate

ITO

ITO

polysiloxanebarrier

Glass Substrate

ITO

Po

ad, P automated load

relaxation debond

Debonding Kineticsexplore role of:

test dataDTS Delaminator v4.0

Loa

ngth

, a

dP/dtrelaxation debond growth analysis

compliance analysis

• UV flux

• humidity, O2, OH, …

Cra

ck L

en

Time (s)

da/dt • temperature

• mechanical loadingsensitivity to < 10-10 m/s

DTS system and support contact dauskardt@stanford.edu

UV Exposure (3 4 eV)

UV Effects on Molecular Bond Rupture

10 5

10-4(m

/s)

UV Exposure (3.4 eV)

UV intensity1 2 mW/cm2

10-6

10-5

da/d

t(

Ghv ~ 1.3 J/m21.2 mW/cm2

10-8

10-7

h R

ate,

No UV0 mW/cm2

10-9

10

Gro

wth

2

i h htip GGda0 6 W/ 2

10-11

10-10

Cra

ck

Ghv ~ 0.8 J/m2

sinh htipov

dt0.6 mW/cm2

0 1 2 3 4 5 6 7Strain Energy Release Rate, G (J/m2)

Modeling Bond Rupture Kinetics• Interaction of moisture with strained debond tip bonds h

)complex activated() tipdebond( *2 B

hBOnH

• Interaction of moisture with strained debond tip bonds

H2O

h

kTU

kTUfrate o

**

expexp Substrate

• Atomistic bond rupture models:

2sinh htip

o

GGv

dda

Si-O-Si + H2O

activation barrier

h

• Damage growth rate:

odt 2

Si-OH + HO-Si

h

attemptfrequencyrg

y

kTu

Nwfv o

o1exp2

Bond Rupture ParametersN - bonds per unit area fo - attempt frequencyuo - work of rupture u1 - energy barrier

G

q yEn

e kTNw

o p 1 gy - N uo - 2NkTw - crack width Bond Rupture Reaction

UV SourcesFluorescent Lamp

AM1 solar spectrum

Mercury LampMercury Lamp (Pen-Ray from uvp)

3.4eV 4.9eV4.1eV

Solar spectrum from B. Van Zeghbroeck, U. Colorado

UV exposure 4 9 eV

UV Effects on Molecular Bond Rupture

1E-5

s) 10-5

UV exposure 4.9 eVadhesive failure at 20°C 40%RH

adhesive

1E-6

da/d

t (m

/s

10-6

Gl S b t t

Simulated UV Exposure

1E-7

h R

ate,

d

6.6

2 mW/cm2

4.6 mW/cm210-7

Glass Substrate

polysiloxanebarrier

1E 9

1E-8

d G

row

th mW/cm2

No UV10-8

10 9

Glass Substrate

0 1 2 3 4 5 61E-10

1E-9

Deb

ond 10-9

10-10

Simulated UV Exposure

0 1 2 3 4 5 6Strain Energy Release Rate, G (J/m2)

Delamination of EVA-TPE Lamination• Poly-ethylene vinyl acetate (EVA) copolymer extensively used by solar modulePoly ethylene vinyl acetate (EVA) copolymer extensively used by solar module

manufacturers, particularly for laminating c-Si photovoltaic modules.• Good optical properties and high adhesive contact with glass cover and Si cells.• Inexpensive and relatively easy fabrication.p y y

Parreta Antonio, et al., Solar Energy Materials & Solar Cells, 2005

• Delamination can occur between EVA and the front surface of the solar cells.• More frequent and in hot and humid climates• More frequent and in hot and humid climates.• Exposure to atmospheric water and/or ultraviolet radiation leads to EVA

decomposition to produce acetic acid, lowering the pH and increasing corrosion.• EVA Tg ~ -15°C so lower temperatures may result in “ductile-to-brittle” transition inEVA Tg 15 C so lower temperatures may result in ductile to brittle transition in

adhesive/cohesive properties.

Tedlar (Polivinylfluoride) II

Delamination of EVA-TPE Laminationadhesion

EVA main l t

Tedlar (Polivinylfluoride)PET

EVA seedTPE

EVA

II

40

45

50

55

Rat

e, G

(J/m

2 )

I

adhesion

Glass Substrate

encapsulantEVA

I15

20

25

30

35

ergy

Rel

ease

R

II

EVA TPE- Bridging TPE- No Bridging0

5

10

15

Stra

in E

ne

II

m/s

)

G 46 J/ 2 G =377 J/m2

time dependant debondingg g g g

Location of Failure

10-6

10-5

Rat

e, d

a/dt

(m Gc=46 J/m2 Gc=377 J/m2

Interface “I” located between EVA and GlassInterface “II” inside the TPE multilayer

9

10-8

10-7

nd G

row

th R IV III

0 100 200 300 400

10-9

Deb

on

Strain Energy Release Rate, G (J/m2)

10°C 10%RH

Summary for PV Durability and ReliabilityWe want to engineer durable PV devices and modulesWe want to engineer durable PV devices and modules

with robust and predictable lifetimes.L f li bilit h i i i l t i• Leverage from reliability physics in microelectronics –mechanisms, kinetic models, accelerated tests and life prediction

• Develop metrologies to quantitatively characterize thermo-• Develop metrologies to quantitatively characterize thermo-mechanical properties (e.g. adhesion, cohesion), photochemical and environmental degradation processes

• Are degradation processes coupled and how?

• Kinetic models of damage evolution - basis for life prediction and accelerated testing (effect of operating temperature, environment, mechanical stress, solar flux, etc.)

• Effective transparent barriers with anti reflective properties and• Effective transparent barriers with anti-reflective properties and low cost.