A-Si Materials and Solar Cells Research at Penn State

8/8/2019 A-Si Materials and Solar Cells Research at Penn State

1/29

Center for Thin Film DevicesCenter for Thin Film DevicesMaterial Research InstituteMaterial Research Institute

DOE Photovoltaics SubprogramPeer Review August 13-15, 2003

SiSi:H Materials and Solar Cells:H Materials and Solar Cells

Research at Penn StateResearch at Penn State

C. R. Wronski and R.W. Collins


2/29

Progress in Research on Si:HProgress in Research on Si:HMaterials and Solar CellsMaterials and Solar Cells

DevelopedReal-Time Spectroscopic Ellipsometry forin-situcharacterization of growth and evolution of microstructurein both films and solar cells. (Now being applied by others)

Developed new approaches for characterizing carrierrecombination and the multiple defect states in a-Si:H filmsand solar cells. (Past focus on just one defect state D0)

Developed deposition phase diagrams identifying themicrostructural transitions during growth from amorphousto mixed phase to a single microcrystalline phase. (Powerful

guide in material optimization)

A comprehensive understanding is being developed of Si:Hmaterial growth, microstructure, properties; mechanismslimiting p-i-n, n-i-p solar cell performance and stability.


3/29

Identified protocrystalline a-Si:H from its growth andmicrostructural evolution with thickness and substrate dependence.(Novel concept of a-Si:H deposited with hydrogen dilutionrepresenting the growth of such materials with outstanding properties)

Applied concept ofprotocrystallinity in optimizing intrinsic anddoped materials as well as solar cell structures. (Systematic approach)

Identified, separated and quantified carrier recombination in both p/iregions and bulk intrinsic layers of solar cells. (Not carried out in past)

Characterized carrier recombination in amorphous and mixed phase(a+c) materials with their effect on solar cell characteristics.(Importance not recognized in past)

Obtained the elusive direct correlations between light inducedchanges in a-Si:H films and corresponding solar cells.(Not establishedin past)

Addressed issues regarding nature, origins of different light induced

defects in a-Si:H and their dependence on microstructure. (Despiteextensive studies in past generally ignored)


4/29

Real Time Optics of Silicon Film PECVDReal Time Optics of Silicon Film PECVD

Developed at PSUrecently being appliedin other laboratories

Allows in situ

characterization ofgrowth (surfaceroughness)

microstructure, opticalproperties 1.5 to 4.5eV Acquisition time

~50ms allowsmonolayer growth to

be characterized

Suitable for analysis of inhomogeneous films with

micro/macro/geometric scale structure


5/29

Evolution of Surface Roughness duringEvolution of Surface Roughness during SiSi:H Film:H Film

Film Growth at Various R on R=0 SubstrateFilm Growth at Various R on R=0 Substrate

Dilution ratio in PECVD R=[H2]/[SiH4]H2 dilution extensively used in fabricationof Si:H materials and solar cells

Two microstructural/phasetransitions vs. thickness:

Roughening Transitiona(a+c)mixed phase

db = 3000 for R=15

db = 700 for R=20db = 200 for R=40

Smoothening Transition(a+c)mixed phase c :db > 7000 for R=15db = 3500 for R=20

db = 650 for R=40


6/29

Characterization of Recombination in Solar Cells FromCharacterization of Recombination in Solar Cells FromDark CurrentDark Current--Voltage CharacteristicsVoltage Characteristics

Voltage (V)

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2

C

urrentDensity(A/cm

2)

1e-10

1e-91e-8

1e-7

1e-6

1e-5

1e-4

1e-3

1e-2

1e-1

1.65eV

1.72eV1.86eV

Clear separation of p/iinterface recombination fromthat in the bulk of a-Si:H solar

cells has not allowed JD-V tobe used in characterizing gapstates.

Bulk recombination has beenidentified and quantified bysystematic reduction of p/i

contributions in cell structures Cell structures are studied in

which the two components ofcarrier recombination areclearly separated.

p(a-SiC:H)-i-n4000 R=10i-layers

Cody Gapof 200 p/iinterface layer

Information about the gapstates in the intrinsic layers can

be obtained directly from thebulk recombination.


7/29


8/29

Extended Phase Diagram:Extended Phase Diagram:SiSi:H Growth on:H Growth oncc--Si/oxide SubstratesSi/oxide Substrates

(1) a asurface roughening transition

(2) a (a+c)surface roughening transition

(3) (a+c) csurface smoothening transition

These transitions provide insights intomaterials and device optimization

Narrow window for protocrystalline

Si:H growth in a thick layer is

centered at R=10;

here the film surface is stablethroughout deposition.

Obtained from the three

transitions detected

during Si:H film growth:

Microstructure and its evolution isstrongly dependent on substrate

TEM f Si H d i d i h R 20 Si O/SiO


9/29

TEM ofTEM ofSiSi:H deposited with R=20 on c:H deposited with R=20 on c--SiSi/SiO/SiO22

TEM fTEM f SiSi H d i d i h R 10 C (H d it d ith R 10 C ( ))


10/29

TEM ofTEM ofSiSi:H deposited with R=10 on Cr (:H deposited with R=10 on Cr (evapevap))

S i f f Si fi


11/29

Schematic of the structure of Si:H filmsSchematic of the structure of Si:H filmson aon a--Si:H (R=0)Si:H (R=0)

R

Despite evolutionary nature, protocrystalline a-Si:Hhas uniform properties over extended regions ofthickness

Great attention must be given to the transition a(a+c) and its thicknessdependence on R- films and cells

Phase diagrams are a powerful guide in optimizing deposition conditions forfast growth

T T iti Ph Di Eff t f f Pl P


12/29

Two-Transition Phase Diagram:Effect of rf Plasma Power

Phase diagrams depend on

deposition conditions otherthan R.

Identify effect of depositionparameters on a a anda (a+c) transitions;regimes of protocrystalline

Si:H growth. Large shifts in transitions

when the plasma power is

increased.

Phase diagrams are a powerful guide inoptimizing deposition conditions for fast growth.

O ti i ti P i i l f TO ti i ti P i i l f T St iSt i LL


13/29

Optimization Principle for TwoOptimization Principle for Two--Step iStep i--LayerLayerof aof a--SiSi:H p:H p--ii--n Solar Celln Solar Cell

Optimization principle Prepare interface and bulk

i-layers with the maximumR=[H2]/[SiH4] values possible

without crossing the a(a+c)transition for the desiredthickness concept ofprotocrystallinity is useful

Difficulties If the a(a+c) transition is

crossed accidentally in this

process, one must decrease Rbelow 10 (below protocrystallineregime) to suppress continuedgrowth of the microcrystallites

real time monitoring andcontrol are needed

Two step optimization of R=10 bulk i-layerswith R=40 p/i 200 layerImprovement:

Voc 0.86 to 0.92VAnnealed FF same 0.72DSS FF 0.60 to 0.66

P f f (P f f ( SiCSiC H)H) ii S l C llS l C ll


14/29

Performance of p(aPerformance of p(a--SiCSiC:H):H)--ii--n Solar Cellsn Solar Cellswith Onewith One--Step and TwoStep and Two--Step iStep i--LayersLayers

Summary of a detailedstudy based on phasediagrams on theoptimization of cellperformanceImprovement:

Voc 0.86 to 0.92Annealed FF same 0.72DSS FF 0.60 to 0.66

Note: Optimumperformance withR=40 adjacent to p-a-

SiC:H limited to 200thickness.

Nature of (a+Nature of (a+ c) phase and its effectc) phase and its effect


15/29

Nature of (a+Nature of (a+c) phase and its effectc) phase and its effecton solar cell performanceon solar cell performance

From RTSE and AFM for R=40 on R=0film onset ofc nucleation occurs atthickness of 200, with completecoalescence ofc nuclei within d=400.

Increase in recombination due to reductionin the mobility gaps in R=40 layer to1.62eV at 300 and 1.22eV at 400.

Presence of such an a(a+c) transitiongreatly increases carrier recombinationand has profound effect on cellcharacteristics.

Voltage (V)0.0 0.2 0.4 0.6 0.8 1.0

CurrentDensity(mA/cm

2)

0

2

4

6

8

10

12

R=40 p/i interface

layer thickness

200300400

Increase in d

4000 p(a-SiC:H)-i-nR=10 i-layer

Such phase transitions are evenmore critical in n-i-p structures

where the (a+c) phase is in directcontact with the p-layer.

P t t lli it C t li d tProtocrystallinity Concept applied to


16/29

Protocrystallinity Concept applied toProtocrystallinity Concept applied topp--Si:HSi:HContacts in nContacts in n--ii--p Cellsp Cells

Phase diagrams of p-Si:H layers onR=10 a-Si:H were used inoptimizing VOC in n-i-p cells.

The maximum VOC occurs with

R=150 and corresponds to aprotocrystalline layerterminated at200 or close to the (a+c) phase.

The lowest VOC is obtained with

R=200 where the layer has evolvedinto apurely c-Si:H phase because i/p recombination increasessignificantly.

The erroneous conclusions thathighest VOCs are obtained withc p-Si:H held for a long timeis due to characterizing

microstructure on layers>>200.

Limitations on 1 Sun VLimitations on 1 Sun V imposed byimposed by


17/29

Limitations on 1 Sun VLimitations on 1 Sun VOCOC imposed byimposed byp/i Interface Recombinationp/i Interface Recombination

Systematic increases in 1 sun Vocfound with decrease in p/i interfacerecombination for protocrystalline

Si:H and a-SiC:H p-contacts. Very large increase in such

recombination in n-i-p cells with p-

Si:H occur when (a+c) phase at ornear i-layer (R=200).

The p/i recombination for the R=150p-Si:H is significantly lower than

the lowest achieved with a-SiC:Htwo step processes

Voltage (V)

0.5 0.6 0.7 0.8 0.9 1.0 1.1

CurrentD

ensity(A/cm

2)

1e-7

1e-6

1e-5

1e-4

1e-3

1e-2

1e-1

R=100R=150R=200

4000A n-i-p cellR=10 Bulk i-layer

This explains why the highest valuesof VOC are with p-a-Si:H cells

Defect states in the intrinsic layers of aDefect states in the intrinsic layers of a SiSi:H:H


18/29

Voltage (V)

0.2 0.4 0.6 0.8

CurrentDensity(A

/cm

2)

10-10

10-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2

0.4 m0.8 m1.5 m

i-layer thickness

Thickness (m)

0.4 0.8 1.2 1.6

Voc

(V)

0.85

0.86

0.87

0.88

0.89

p-i-n cell

Non-uniform distributionsof defect states across solarcells predicted by theDefect Pool Model haveoften been reported.

No evidence is found fromJD-V characteristics whose

bulk contributions areclearly identified.

Dependence on i-layerthickness, equivalence ofp-i-n and n-i-p structures.

Recombination consistentwith spatially uniform

defect states and those incorresponding films.

The ability to characterize bulk recombination

enables its limitations on different solar cellparameters to be quantified (e.g. VOC).

Defect states in the intrinsic layers of aDefect states in the intrinsic layers of a--SiSi:H:Hsolar cells with low p/i interface recombinationsolar cells with low p/i interface recombination

Light induced defects in the intrinsic layers of aLight induced defects in the intrinsic layers of a--SiSi:H cells:H cells


19/29

Light induced defects in the intrinsic layers of aLight induced defects in the intrinsic layers of a SiSi:H cells:H cells

Voltage (V)

0.2 0.4 0.6 0.8 1.0

CurrentDensity(A/cm

2)

10-11

10-10

10-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2

Annealed State1 Sun 30 mins

1 Sun 9 hrs1 Sun 100 hrs (DSS)

p-i-n cell with 4000 R=10 i-layer

The kinetics of the light

induced changes in JD-Vcharacteristics are similar tothose in FF of cells and the products in correspondingfilms.

JD-V characteristics

offer an new probefor investigating thenature and densities

of defect states inintrinsic layers ofsolar cells.

Increase inrecombination withintroduction of light

induced defects

Kinetics of light induced changes in FFKinetics of light induced changes in FF


20/29

Kinetics of light induced changes in FFKinetics of light induced changes in FF

Extensively used in

characterizing lightinduced defects insolar cells.

Find kinetics in cellshaving i-layers withdifferent

microstructureclearly point tocreation ofmultiple

defects.1 Sun Illumination Time Hours0.01 0.1 1 10 100 1000

FillFactor

0.52

0.54

0.560.58

0.60

0.62

0.64

0.66

0.68

0.700.72 R=0 25oC

R=10 25oC(Protocrystalline)

ND0 = 8-10x1016cm-3

4000p(a-SiC:H)-i-n

Multiple defects confirmed with the lack of correlation inthe FF degradation with ND0 (as measured with ESR),|(E)|, and presence of fast and slow states.

Direct correlation of recombination inDirect correlation of recombination in


21/29

Direct correlation of recombination inDirect correlation of recombination infilms and cellsfilms and cells

1/ cm-2V

0 4x106 8x106 12x106 16x106 20x10

FillFactor

0.52

0.54

0.56

0.580.60

0.62

0.64

0.66

0.68

0.70

0.72

R=0 25oC

R=0 75oC

R=10 25oC

R=10 75oC

Time

1/ cm-2V

1.0x106 2.0x106 3.0x106 4.0x106 5.0x106

FillF

actor

0.56

0.58

0.60

0.62

0.64

0.66

0.68

0.70

0.72

0.74

d=2000d=4000d=7000

Time

R=10 i-layer

4000 p-i-n

The elusive correlations betweenlight induced changes in thin filmsand those in solar cells have been

established. Because the nature and densities of

the different defect states in films and

cell i-layers are not yet known it is notpossible to directly correlate them.

Can howeverrelate them through theirrole as carrier recombination centers,

Nr, where 1/Nr, (1-FF) Nr

Linear relationships are obtainedbetween 1/ and FF for cells havingdifferent thickness, different i-layersand at different temperatures.


22/29

Distinctly different light induced defectDistinctly different light induced defect


23/29

Distinctly different light induced defectDistinctly different light induced defectstates at and below midgap in astates at and below midgap in a--Si:HSi:H

Ener eV

0.9 1.0 1.1 1.2 1.3 1.4

NDS

(E)/N

AS(E

)

0

510

15

20

25

30

35CBA

Evolution of kN(E) fromd[(h)]/dE in degraded state(DS) normalized to annealed state

(AS) Distinctly different states created

around 1.0 and 1.2eV from EC

Improved microstructure(C R=0 20/s to B R=0 1.5/s toA R=10 )

Systematic suppression of1.2eV defects

Evolutions of defects and their temperature

dependence self-consistent with changes in FF and Only in protocrystalline a-Si:H is the defect state at1.0eV dominant

Evolutions of defects selfEvolutions of defects self--consistent with changes in FF andconsistent with changes in FF and


24/29

gg

(a) R=10

Ener eV

0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6

[d

(E)/dE]/[dAS

(E)/dE

]

0

5

10

15

20

25

25oC

75oC

Ener eV

0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6

[d(E)/dE]/[dAS

(E)/dE]

0

1

2

3

4

5

6

7

8

75oC

25oC

30 Hours of 1 Sun Illumination

ETL 20/s

1 sun degraded steady state

Degraded States after 1 sunillumination at 25, 75oC kN(E) SpectraNormalized to AS

In R=10, suppression of defects at

75oC, particularly at 1.2eV, consistentwith corresponding higher FF,

No change in spectrum ofR=0 20/sconsistent with virtually the same degradation kinetics at bothtemperatures.

Presence of multiple defect statesand their dependence onmicrostructure must be taken intoaccount in characterizing stability

of solar cell materials and SWE.

ConclusionsConclusions


25/29


RTSE is a unique and powerfultechnique for development ofphase diagrams.

Deposition phase diagrams extremely usefulin optimizationof Si:H materials for solar cells.

Concept ofprotocrystallinity shown to be useful in:Improvementof n, i, and p Si:H layersSystematic improvement of cell structures

Controllingdeleterious effects of a (a+c) transition onsolar cell characteristics

Overcoming erroneous conclusions drawn from

characterizing films thicker than the layers used in solarcells.



26/29

Co c us o sCo c us o s

New approach for characterizing (h) spectrahas:

Offered a more reliable method for evaluatingmaterials for solar cells.

Identified evolution ofdistinctly differentlight

induced defects. Points to a reason for the discrepancies between

stabilities claimed forfilms and those found in

correspondingsolar cells.



27/29

Identifying p/i recombination in solar cells key tocharacterizing i-layers and their contributions to cellcharacteristics.

JD-V characteristics used as a new probe for characterizingrecombination and defect states in intrinsic layers of solar cells. Presence ofspatially uniform densities of defects in the i-layers

in conflict with Defect Pool Model but allows correlations withcorresponding films.

Elusive direct correlations between recombination in thinfilms and their solar cells established.

Same creation and annealing kinetics of fast and slowstates established for FF and .

For thefirst time distinctly different light induced defect states

centered around 1.0 and 1.2 eV have been clearly identified.



28/29

The results on the two distinctly different light induced defectstates in a-Si:H are significant in that: They are consistentwith the discrepancies between changes in

dangling bond densities, ND0, and those in FF, . Their evolution is consistentwith that of fast and slow states.

They are consistentwith changes in stability of a-Si:H with

different microstructure. They show improved stability ofprotocrystalline a-Si:H is

accompanied bysuppression of the 1.2 eV defect state.

Theypoint outthe serious limitations of the commonly usedmethodologies of assessing the stability of solar cell materials.

The established presence and distinctly different evolutions of thetwo light induced states are not consistentwith a variety of

explanations proposed for the origin of the Staebler-Wronski effect.

Results?Results?


29/29

Results?Results?

12 Peer Reviewed Publications12 Peer Reviewed PublicationsScores of Conference Papers,Scores of Conference Papers,

Posters, and PresentationsPosters, and Presentations

... and 10... and 10 InvitedInvited Talks atTalks at

International ConferencesInternational Conferencesand Workshopsand Workshops

Documents

A-Si Materials and Solar Cells Research at Penn State