Predicve physico-chemical modeling of intrinsic degradaon ... - 04... · (93) The interface residual can be set up on a pairwise basis. From Figure 3 we 18. CSP Program Summit 2016

energy.gov/sunshotenergy.gov/sunshotCSPProgramSummit2016

energy.gov/sunshotenergy.gov/sunshotenergy.gov/sunshot

CSPProgramSummit2016

Predic'vephysico-chemicalmodelingofintrinsicdegrada'onmechanismsforadvancedreflectormaterials

RossLarsenSeniorScien<stNa<onalRenewableEnergyLaboratory

energy.gov/sunshotenergy.gov/sunshotCSPProgramSummit2016 2

Outline

•  Projectobjec<ves

•  Approach

•  Kine<csandtransport

•  Mechanicalproper<es

•  Agingofreflectorstacks

•  Summaryandoutlook


ProjectTeamandAcknowledgments

NREL:RossLarsenRayGroutChaoZhangAndrewFergusonRobertTirawatMaPhewGray(nowwithSkyFuel)VinceWheeler(nowatAustralianNa<onalUniversity)

Costsharepartners:DLR

PeterHellerFlorianSuPer

DOECSP

DE-FOA-0000861


ProjectObjec'ves

•  Developnovelmodelingcapabilityforunderstandingreflectordegrada<onthatcombinesareac<on-diffusion-op<csformalismwithexperimentstoiden<fyunderlyingratesandmechanismsfordegrada<on

•  Createanddisseminateatool,validatedbytes<ngagainstmul<pleexperiments,thatcanbeusedbytheCSPcommunitytoevaluatedegrada<onmechanisms

Time

Refle

ctance StackdesignsAandB

~30yr

polymer/glasssubstrate

reflec<nglayer

ARcoa<ng

protec<vecoa<ng


Micro-kinetic Description

Intrinsic material property characterization

Transport Optical Damage

Mechanical Property

Accelerated Aging

IndividualComponentLayer

Characterizesinglelayers

orasimplifiedstack

σ c = σ c (nC,nbl ,nOx,nH2O

,T )

∂nC(r,t)∂t

= DC∇2nC(r,t)− α(λ)Iλ (r,t)dλ nC(r,t)− n*(r,t)− nbl (r,t)⎡

⎣⎤⎦∫

∂n*(r,t)∂t

= D*∇2n*(r,t)+ α(λ)Iλ (r,t)dλ∫ nC(r,t)− n*(r,t)⎡

⎣⎤⎦ − kOxnOx(r,t) ⋅n*(r,t)

∂nbl (r,t)∂t

= +kOxnOx(r,t) ⋅n*(r,t)

∂nOx(r,t)∂t

= DOx∇2nOx(r,t)− kOxnOx(r,t) ⋅n*(r,t)+S(r,t)

Iλ (r,t) = Iλ (z = 0,t)exp[−α (λ) nC (z,t)−n* (z,t)⎡⎣ ⎤⎦dz

0

L

∫]

Simplifiedexamplephoto-kine<cscheme

αλ = αλ (nC,nbl ,nOx,nH2O

,T )

D = D(nC,nbl ,nOx,nH2O

,T )

kOx = kOx(nC,nbl ,T )

k = k(T,P)

nOx(0,t) = kPOx

ProjectApproach


ProjectApproach

•  Createandvalidatemodelsandsodwaretopredictandunderstanddegrada<onofadvancedreflectorsspanningdecades•  Op<cs,reac<on-diffusionkine<cssolver(feedback)•  Finiteelementmechanicalmodeling

Op<cs(fullHelmholtz)

Kine<cs/transport(splitoperator)

Mechanicalmodelingdelamina<on,cracking,etc.

(FEM)


Op'cssolver

•  Transfermatrixmethodfor1Dreflectorstacks•  Fullyvectorial(sandppolariza<on)•  Numericallyexactforpiecewiseconstantstacks•  Exactreflectancewithoutneedforupwinding,boundaryelements,etc.(3Dversion:boundaryelements)

0

0.2

0.4

0.6

0.8

1

500 1000 1500 2000 2500

Refle

ctan

ce

Wavelength (nm)

500 A Cu/1000 A Ag/1000 A Al_2O_3500 A Cu/2000 A Ag/1000 A Al_2O_3500 A Cu/1000 A Ag/6000 A Al_2O_3500 A Cu/2000 A Ag/6000 A Al_2O_3

Unpolarizedreflectancefromalumina-coatedsilverreflector


Kine'cs-transportsolver

TransientmeasurementofpermeabilityofPETfilmtowater

Delayedonsetconsistentwithnon-FickiandiffusionDualsorp<on:

!"!" = !!

!!"

11+ ! !!!!

(1+ !!!!)!

!"!"

Non-linearityintransportandsorp<onrequiresnovelsolverforboundaries

Figure 3: Four domains sharing a singe interface in the structured finite vol-ume formulation. Many domains meeting at a single interface adds little or nodi�culty to the use of the proposed interface residual method.

To handle an arbitrary number of subdomains, we first recall that we canwrite a set of equations of the form (50) for the ith subdomain,

"A

(i)II

A

(i)Ig

0 I

# u

(i)I

u

(i)g

!=

f

(i)I

u

(i)ng

!. (50)

So we have the interior nodes in terms of the ghost nodes,

u

(i)I

= �A

(i)�1II

A

(i)Ig

u

g

+ A

(i)�1II

f

(i)I

. (93)

The interface residual can be set up on a pairwise basis. From Figure 3 we

18


Kine'cs-transportsolver

TransientmeasurementofpermeabilityofPETfilmtowater

Delayedonsetconsistentwithnon-FickiandiffusionDualsorp<on:

!"!" = !!

!!"

11+ ! !!!!

(1+ !!!!)!

!"!"

Wehaveextractedpermeabilityasafunc<onofRH,TandPox,TforPET,PMMA,withandwithoutoxidecoa<ngs

P= !! + !!! !!!/!

energy.gov/sunshotenergy.gov/sunshotCSPProgramSummit2016 10CSPProgramSummit2016

0

20

40

60

80

100

120

140

0 5 10 15 20 25 30

Stre

ss (M

Pa)

Strain (%)

0h UV Exposure, Direction 1, Sample 1 0h UV Exposure, Direction 1, Sample 2 0h UV Exposure, Direction 2, Sample 1 0h UV Exposure, Direction 2, Sample 2 68h UV Exposure, Direction 1, Sample 1 68h UV Exposure, Direction 1, Sample 2 68h UV Exposure, Direction 2, Sample 1 68h UV Exposure, Direction 2, Sample 2 78h UV Exposure, Direction 1, Sample 1 78h UV Exposure, Direction 1, Sample 2 78h UV Exposure, Direction 2, Sample 1 78h UV Exposure, Direction 2, Sample 2

Mechanicalproper'es:UVdegrada'onofpoly(ethyleneterephthalate)films

2.0

1.5

1.0

0.5

0.0

Abso

rban

ce

500450400350300250

Wavelength (nm)

0 hrs (un-aged) 0.5 hrs 1 hr 2 hrs 3 hrs 4 hrs 5 hrs 6 hrs 7 hrs 8 hrs 9 hrs 10 hrs 12 hrs 14 hrs 16 hrs 20 hrs 32 hrs 37.5 hrs 48 hrs 55.25 hrs 80 hrs 96 hrs 125.5 hrs 160 hrs 191.25 hrs 229.25 hrs 302.25 hrs

10XUVA/UVBunderambientcondi<ons

Agedfilmsshowminimalchangeinelas<cconstants,drama<creduc<onin

ul<matefailurestrain


Mechanicalproper'es:FractureofbriIlecoa'ngsonstrainedPETfilms

11

•  Tensile properties of TiO2/PET thin film system

0

20

40

60

80

100

120

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09

Coating thickness of 100 nm




Pure substrate

Stre

ss (M

Pa)

Strain

y = 0.0083x + 3728

3550

3600

3650

3700

3750

3800

3850

3900

3950

0 100 200 300 400 500 600 700

Elastic modulus

Linear data fitting

Elas

tic M

odul

us (M

Pa)

Coating thickness (nm)

(a) (b)

Ø  Tensile load is transferred through the polymer substrate.

Ø  Coated sample shows an earlier yielding.



12

•  In situ tensile test shows progressive cracking behavior

50 µm

2.5%

9.9%

12.8%

18.2%

Tensile direction

Parallel crack

Buckle



13

•  Tape test shows delamination of TiO2 for UV aged sample



14

•  In situ tensile test shows progressive cracking behavior

50 µm

2.5%

9.9%

12.8%

18.2%

Tensile direction

Parallel crack

Buckle Howdofractureanddelamina<onchangewithaging?

Isthereanunderlyingkine<cmodel?


Agingofsimplifiedreflectorstacks

15


Agingofsimplifiedreflectorstacks:Un-aged

16

100

80

60

40

20

0

Refle

ctan

ce (

%)

15001000500wavelength (nm)

Silver (Reflectance)Literature

Adachi Johnson Schulz

Measured

Front-surface Al standard Predicted from VASE-derived Optical Constants

Stored in Air Spot 1 Spot 2 Spot 3

Cau<on:thereisnotnecessarily“canonical”silver

Ourmodelsnowhave“thin”and“thick”NRELsilverop<calconstants


Agingofsimplifiedreflectorstacks:Un-coated

17

Degradesamplesundercondi<onsof45C,40%RH2Xsunlight(ordark)HemisphericalR

0

10

20

30

40

50

60

70

80

90

100

250 750 1250 1750 2250

Refle

ctan

ce

Wavelength(nm)

5592-28-2,Cu(10)/Ag(500)

5592-28-2-1

5592-28-2-130hours

5592-28-2-160hours

5592-28-2-2

5592-28-2-230hours

5592-28-2-260hours

5592-28-2-290hours

5592-28-2-3

5592-28-2-330hours

5592-28-2-360hours

5592-28-2-390hours

5592-28-2-4

5592-28-2-430hours

5592-28-2-460hours

5592-28-2-490hours

Substrates/Condi<on1-Glass/Light2-Glass/Dark3-3milPET/Light4-3milPET/Dark

Silverdegradesdifferentlyinthedarkthanunderillumina<on


Agingofsimplifiedreflectorstacks:Un-coated

18

Degradesamplesundercondi<onsof45C,40%RH2Xsunlight(ordark)HemisphericalR

Silverdegradesdifferentlyinthedarkthanunderillumina<onReflectancechangesdifferentlywithAgthickness:thinfilminterference

0

10

20

30

40

50

60

70

80

90

100

250 750 1250 1750 2250

Refle

ctan

ce

Wavelength(nm)

5592-30-1,Cu(10)/Ag(100) 5592-30-1-1

5592-30-1-130hours

5592-30-1-160hours

5592-30-1-190hours

5592-30-1-1120hours

5592-30-1-2

5592-30-1-230hours

5592-30-1-260hours

5592-30-1-290hours

5592-30-1-3

5592-30-1-330hours

5592-30-1-360hours

5592-30-1-390hours

5592-30-1-3120hours

5592-30-1-4

5592-30-1-430hours

5592-30-1-460hours

5592-30-1-490hours

Substrate/Condi<on1-Glass/Light2-Glass/Dark3-3milPET/Light4-3milPET/Dark


Agingofsimplifiedreflectorstacks

19

Iden<fywhatisbeingformedwithacombina<onof:•  Variableanglespectroscopicellipsometry•  RutherfordbackscaPering•  X-rayphotoelectronspectroscopy•  Specularityandroughnessmeasurements

Iden<fyrelevant<mescalesandplausiblekine<cmodels

•  Forwardpredictchemicalchangesandcomparetoexp’tOngoingworktocorrelatechemistry/materialiden<tywithop<cal/mechanicalproper<estobuildpredic<vemodel


Silvercorrosionkine'cs:Insightsfromtheliterature

20

•  Quan<ta<veextrac<onofsilversulfida<onkine<cswithreac%on/advec%onmodelofdegrada<onofsilverindenudertube

FIG. 6. Photographs of Ag foils after 48 h exposure to corrosion-chamber air. Foil leading edge on the left. Flow rates through tube (cm3(STP) s-l): (a) 51.8, (b) 27.3, (c) 14.3,

(d) 7.67, (e) 3.83.

1187

2Ag+H2S+½O2->Ag2S+H2O

InputgasmixtureFixedflowrate

Exi<nggasmixture

Volpeetal.,CorrosionScience10,1179-1196(1989)



21

•  Quan<ta<veextrac<onofsilversulfida<onkine<cswithreac%on/advec%onmodelofdegrada<onofsilverindenudertube

Flowrates:3.83,7.67,14.3,27.3,51.8cm3(STP)/s

Volpeetal.,CorrosionScience10,1179-1196(1989)

2Ag+H2S+½O2->Ag2S+H2O



•  Lightdependence:arewephotolyzingO3?•  ArewemakingAg2O,AgO

•  Evolu<onofAgxOtoAg2O(aging)asmoreoxideforms?

•  Areweformingaheterogeneousmixedlayerofoxidesandsulfides?Chlorides?


SummaryandOutlook

23

•  Centralhypotheses•  Fundamentalchemicaleventscanbemeasuredandmechanismsinferred

•  Predic<vemodelscanbeconstructedbasedonthesemeasurements

•  Measurementsshowcomplexdegrada<onkine<csevenforsimplifiedgeometriesandmaterials•  Q:Isfullchemistryneeded?Canonegetawaywithasimplifiedpseudo-chemistry?

•  APen<ontoroughening/lossofspecularity•  Combinedop<cs/kine<cs/transportcodetobemade

availabletothecommunity

Documents

Predicve physico-chemical modeling of intrinsic degradaon ... - 04... · (93) The interface residual can be set up on a pairwise basis. From Figure 3 we 18. CSP Program Summit 2016