Nanostructured Organic-Inorganic Hybrid Photovoltaic Cells

Mike McGehee Stanford University

Why solar cells are likely to play a big role in reducing carbon emission

• We probably need to generate ~ 30 TW of power without emitting carbon in 2050.

• The sun gives us 120,000 TW.

• Most other renewables can’t provide 30 TW.

• Using solar cells instead of burning coal to generate electricity is a much easier way to reduce carbon emissions than replacing gasoline in vehicles.

• Solar cells are safe and have few non-desirable environmental impacts.

Silicon Solar Cells

Modules• 12 % efficiency

• $350/m2

• $3/Wp (from manufacturer)

• $6/Wp (installed)

Average cost of PV cell electricity: $.27/kW-hr

Today’s grid electricity: $0.06/kW-hr

Area and cost needed to power the country

(150 km)2 of Nevada coveredwith 15 % efficient solar cells couldprovide the whole country with electricity

At $350/m2, we would have to spend $8 trillion to do this.

Additional $ would be needed for storage technology.

J.A. Turner, Science 285 1999, p. 687.

Roll-to-Roll coating: A route to taking the costs below $50/m2

and keeping efficiency > 10 %.

P. Fairley, IEEE Spectrum. Jan. 2004 p.28

January MRS Bulletin

Bulk heterojunction PV cells made by casting blends

Polymer C60 derivative Sariciftci 3.5 %Polymer polymer Friend 1.9 %Polymer CdSe nanorods Alivisatos 1.7 %Polymer ZnO nanocrystal Janssen 1.6 %

Energy conversionefficiency (AM 1.5)

Alivisatos et al., Science 295 (2002) p. 2425

polymer/nanorod blendpolymer/C60 blend

Heeger et al. Science 270 (1995) p. 1789.

absorber electron acceptor

Polymer Electron acceptor

E

☺ fabrication is simple

not all excitons reach at interface

there are deadends

Processes in bulk heterojunction PV cells

1

234

4

PolymerElectron acceptor(e.g. C60or TiO2)

Electrode

Electrode

Desirable

1 Absorption

2 Exciton diffusion3 Forward electron transfer

4 Charge transport to electrodes

e-e-

h+h+5

6

Undesirable recombination events

Electron acceptor

Polymer Electrode

Electrode

5 Geminate recombination (1 ns)

6 Back electron transfer ( 1 µs ?)

competing

competing

Bulk heterojunctions made by filling nanoporous films

Light absorber Electron acceptor EfficiencyPolymer sintered TiO2 nanocrystals Carter 0.2 %Ru Dye sintered TiO2 nanocrystals Grätzel 4 %*

*with solid state hole transporter

titania nanocrystals

e-

h+E

EA IM OS

Modifying the interface can be used to slow down back electron transfer.

Ordered bulk heterojunctions

~ 2 absorption lengths(200 - 300 nm)

< 2 exciton diffusion lengths (~10-20 nm)

Transparentelectrode(e.g. ITO)

Reflectingelectrode

1st semiconductor(e.g. P3HT)

Electron accepting semiconductor(e.g. titania)

• Almost all excitons can be split• No deadends• Polymer chains can be aligned

• Easy to model• Semiconductors can be changed without changing the geometry.

Making the nanostructures

• Block copolymer/ titania self-assembly

• Block copolymer lithography and reactive ion etching

• Nanoimprint lithography

See Vignesh Gowrishankar’s poster this afternoon.

Mesoporous titania films

☺ Film thickness can be varied from 50 - 300 nm☺ Pore radius: 4 nm in the plane and 2-3 nm perpendicular to the plane☺ Film quality is very high

Pores are not straight

Adv.Func.Mat., 13 (2003) p. 301

Anodic Alumina

< 300 nm PMMA~ 1mm

PDMS

PDMS

TCO

PDMS

TCO

Stamping nanoporous titania films

Initial anodic alumina template

intermediate PMMA stamp

Final TiO2 structure

Melt infiltration

heat

conjugatedpolymer

mesoporousTiO2

ITO

Following spin-

casting, heating

and rinsing

Mesoporous TiO2 film

Adv.Func.Mat., 13 (2003) p. 301

33 % of the volume of the filmcan be filled in several min.

Measuring exciton diffusion lengths by analyzing PL quenching

θ Ι

θΕ

CCD Spectrometer

Lenses

Sample442 nm

quencher (TiO2, C60, etc.) non-quencher (glass, quartz, etc.)

thickness (0 – 60 nm)

450 500 550 600 650 700 750 800

0

10000

20000

30000

40000

50000

60000

PL (a

rb. u

nits

)Wavelength (nm)

thickness

0 10 20 30 40 50 60

0

1x106

2x106

3x106

4x106

5x106

6x106

PLGlass PLTiO2

PL (a

rb. u

nits

)

Polymer Thickness (nm)

MDMO-PPV: amorphous, light emitting polymer

Quenching of MDMO-PPV on TiO2

It looks like the diffusion length is 15 nm, but actually the PL is suppressed because the titania film reflects light and causes destructive interference to occur near the titania.

Exciton diffusion: Take home messages

1. Most diffusion lengths (Ld) in the literature for polymers are inaccurate because of interference effects.

2. Ld is probably < 10 nm because excitons find low energy sites and get stuck there.

3. Efficient collection of excitons requires that the light absorbing molecule to touch the electron acceptor or that resonant energy transfer be used to take the energy over the barrier.

Diffusion

Resonance Energy Transfer

Improving device efficiency with resonant energy transfer

SnO2:F

Ag

TiO2

D2

D1

D1 : P3HT

D2: PTPTB

Eg (D1) > Eg (D2)

Spectra of P3HT and PTPTB

400 500 600 700 800

Absorption of P3HT Photoluminescence of P3HT Absorption of PTPTB

Abs

orba

nce

Photolum

inescence

Wavelength (nm)

The overlap of the P3HT emission spectrum and PTPTB absorption spectrum promotes rapid Förster transfer.

Quenching of P3HT emission with TiO2 and PTPTB

550 600 650 700 750 800 850 9000

5000

10000

15000

20000

P3HT TiO2/P3HT PTPTB/P3HT TiO2/PTPTB/P3HT

PL

inte

nsity

Wavelength (nm)

43 % quenching

81 % quenching

11-nm P3HT

12.4 nm TiO2

5-nm PTPTB

11-nm P3HT

11-nm P3HT

5-nm PTPTB

11-nm P3HT

12.4 nm TiO2

81 % quenching

Design of PV cells

• We find that excitons in PTPTB are not split by electron transfer to titania.

• Excitons are split by hole transfer to the P3HT.

• Cells made with MDMO-PPV did not work well because the excitons were not split.

PV cell performance

0.0 0.2 0.4 0.6-1.5

-1.0

-0.5

0.0 TiO2/P3HT(50 nm)/Ag TiO2/PTPTB(5 nm)/P3HT(50 nm)/Ag

Voltage (V)

Cur

rent

den

sity

(mA

/cm

2 )

FTO/TiO2/P3HT/Ag Jsc: 0.46 mA/cm2; Voc: 0.64 V;FF: 0.63; ηp: 0.19%.

FTO/TiO2/PTPTB/P3HT/AgJsc: 1.44 mA/cm2; Voc: 0.605 V; FF 0.61; ηp: 0.53%.

Energy transfer triples the efficiency.

100 mW/cm2 AM 1.5 conditions

P3HT/nanoporous titania photovoltaic cells

ηpwr would be ~ 0.45% under actual sunlight

At 33 mW/cm2, 514 nm:ηpwr = 1.5 %FF = 0.51

EQE = 10 %

-0.2 0.0 0.2 0.4 0.6 0.8

-1.5

-1.0

-0.5

Cur

rent

Den

sity

(mA

/cm

2 ) Voltage (V)

Blue: Mesoporous titaniaGreen: Solid (nonporous) titania App. Phys. Lett. 83 (2003) 3380

Photocurrent is only generated at the top

0.1 0.2 0.3 0.40

1

2

3

4

5

Measured Predicted using Beer's Law

EQ

E a

t 514

nm

(%)

Optical Density of Infiltrated PolymerSnO2:F

Ag

SnO2:F

Ag

SnO2:F

Ag

SnO2:F

Ag

App. Phys. Lett. 83 (2003) 3380

The polymer chains do not pack well in the curved pores.

How high does the hole mobility have to be?If the films are 270-nm thick and tback transfer = 1 µs, then

µ needs to be approximately 10-2 cm2/Vs.

s s s ss s s s s s

s s s ss s s s s s

µFET as high as 0.1 cm2/Vs

µSCLC as high as 3 x 10-4 cm2/Vs

Hole-only P3HT diodes in anodic alumina

after Al anodization and pore widening to expose SnO2:F

following P3HT infiltration andtop electrode deposition

SnO2:F

Au/Ag

P3HT &Al2O3

We have grown 20-120 nm diameter pores

200 nmTopview SEM

Cross-Section SEM

Hole mobility versus pore diameter

Neat film ofP3HT

For 75-nm pores, µ is enhanced by 20x.

0 20 40 60 80 100 120 140

1E-4

1E-3

0.01

Hol

e M

obili

ty (c

m2 /V

s)

Pore Diameter [nm]

Conclusions from my research

• Energy transfer is important for harvesting excitons

•Even with a little bit of chain alignment in straight pores, the P3HT µ can be increased by 20X.

• We need a method for patterning 20 nm wide holes that are 200 nm deep in a suitable semiconductor.

Acknowledgements

Kevin Coakley, Bhavani Srinivasan, Yuxiang Liu, Shawn Scully, Chia Goh, Melissa Summers

Polymer synthesis: Katya Kadnikova, Jinsong Liu, Jean Fréchet (UC Berkeley)Mesoporous Titania: Karen Frindell, Galen Stucky (UC Santa Barbara)