Organic Solar Cells

1
Organic Solar Cells
Jung-Yong Lee for Peter Peumans Electrical
Engineering _at_ Stanford ppeumans_at_stanford.edu
Sponsors DOE, GCEP, NSF, CAMP
2
Transparent Contacts to Thin-Film Solar
Cells How Do You Make Something Transparent AND
Conductive AND Cheap?
3
Whats Wrong with ITO?
ITO
glass

• Performance
• Cost (10/m2)
• Brittle
• Compatibility with plastic substrates

4
Carbon Nanotubes Meshes
ITO
glass
CNT mesh
glass
5
Extraordinary Transmission in Subwavelength Metal
Gratings
40nm
400nm
100nm
Metal Ag
6
Extraordinary Transmission in Subwavelength Metal
Gratings
TM
450nm
TM
E
650nm
TM
850nm
Metal Ag
Color map E, Arrow Power Flow
7
Extraordinary Transmission in Subwavelength Metal
Gratings
TE
450nm
TE
E
650nm
TE
850nm
Metal Ag
Color map E, Arrow Power Flow
8
Extraordinary Transmission in Subwavelength Metal
Gratings
geometric aperture
Solar Irradiance (W/m2/um)
Transmission gt geometric aperture
Metal Ag
Color map E, Arrow Power Flow
9
Comparison
400nm period, 40nm linewidth
metal grating
20nm
10nm
glass
30nm
ITO
glass
CNT mesh
glass
10
J.-Y. Lee, S.T. Connor, Y. Cui and P. Peumans,
Nanoletters8, 689-692 (2008)
500nm
10?m
11
J.-Y. Lee, S.T. Connor, Y. Cui and P. Peumans,
Nanoletters8, 689-692 (2008)
200nm
500nm
_at_200C
10?m
12
A Solution-Processed Metal Grating
metal grating
glass
ITO
glass
CNT/metal NW mesh
glass
J.-Y. Lee, S.T. Connor, Y. Cui and P. Peumans,
Nanoletters8, 689-692 (2008)
13
Compatible with Flex
Flat 12.7 ohm/sq
20 of transmitted light is scattered
Radius 13.74 mm 12.5 ohm/sq
Radius 4 mm 12.6 ohm/sq
Radius 8.36 mm 12.5 ohm/sq
14
What Limits Performance?
What Limits Sheet Resistance?
Contact resistance is a performance limiter
Electroless plating annealing lowers sheet
resistance
10?m
15
Organic Solar Cell
Ag
BCP
PTCBI
CuPc
PEDOT
NWs
Glass
Glass/NW/PEDOT/CuPc 45nm/PTCBI 45nm/BCP 10nm/Ag
100nm
J.-Y. Lee, S.T. Connor, Y. Cui and P. Peumans,
Nanoletters8, 689-692 (2008)
16
Unconstrained 2-Junction Cell
Demonstrated with small molecules.
Ag
CuPc/C60
0.61Sun
AgNW
CuPc/C60
Semi-transparent cell
Glass/ITO
Top
Bottom
0.61 Sun
ITO/PEDOT/CuPc 40nm/C60 55nm/BCP 11nm/Ag
5nm/NWs/WO3 4nm/CuPc 40nm/C60 53nm/BCP 11nm/Ag
100nm
Lee, unpublished
17
Semi-transparent OPV
(a)
ITO/PEDOT/CuPc/C60/BCP/AgNWs
ITO/Cs2CO3/P3HTPCBM/AgNWs
18
Fully-Solution Processed Advantages

• No ITO
• No vacuum steps or switching between solution and
  vacuum
• Possible roll-to-roll printing fabrication
• Possible encapsulation during fabrication process
• Low cost

Source Deutsche Bank
18
19
Device Fabrication
Glass
Glass
Ag Nanowire Mesh
1?m
Ag Nanowire Mesh
PEDOT
PEDOT
P3HTPCBM
P3HTPCBM
Cs2CO3
Cs2CO3
Ag
Ag
1?m

• Every step from solution
• Possible encapsulation during fabrication

19
20
Device Characteristics
Nanowire Electrode
2.5 power conversion efficiency
20
21
Solution-Processed Graphene Transparent
Electrodes for Organic Solar Cells and
Light-emitting Diodes

• Junbo Wu1, Hector A. Becerril1, Mukul Agrawal1,
  Zunfeng Liu2,
• Yongsheng Chen2, Zhenan Bao1 and Peter Peumans1
• 1Stanford University, USA 2Nankai University,
  China
• Dec. 4th, 2008 _at_MRS

ppeumans_at_stanford.edu, junbowu_at_stanford.edu
22
Graphene Property

• Individual 2D-nanosheet of sp2-hybridized
  carbon.
• Extended delocalized electron system, and
  potential for ballistic charge transport.
• Insoluble, does not sublime.

Graphite Crystal
Graphene

23
Solution-Processed Graphene

• Attack graphite with oxidizer
• Sonicate and wash to get GO sheets
• Redisperse in water and spin-coat
• Reduce processed GO films.

XPS Signal
Becerril, H.A. et. al., ACS Nano, 2, 464-470,
2008.
Hannes C. Schniepp, J. Phys. Chem. B, No. 17,
110, 2006.
24
Reduced Graphene Oxide (RGO) Film
Vacuum
CNT film
25
Organic Photovoltaic Cells on RGO

• RGO film 7nm, 95, 75k?-150k?
• Reduced Jsc and FF
• Series and shunt resistance

J. Wu, H.A. Becerril, et. al., Appl. Phys. Lett.,
263302, 2008 .
26
Organic Light-emitting Diodes on RGO

• RGO film 7nm, 82, 1.8k?
• Typical display brightness 100-300cd/m2
• EQE photon out/electron in

J. Wu, M. Agrawal, H.A. Becerril, et. al.,
manuscript in preparation.
27
Organic Solar Cells
Cu Phthalocyanine (CuPc)

• Abundant 100,000 tons/year
• Mature industry/markets
• Low-cost 1/g ? 17/m2
• Non-toxic
• Can be purified to high degree
• Stable
• Efficiency lt5

28
Highest Efficiency 6.5
6.5 Polymer Device
Kim et al., Science (2007)
29
The Internal Efficiency of Solar Cells is an
Inverse Function of Film Thickness
?ED
2.exciton diffusion
?CT
3.charge transfer
?CC
?A
4.charge collection
1.absorption
cathode
anode
acceptor
donor
30
The Internal Efficiency of Solar Cells is an
Inverse Function of Film Thickness
Exciton diffusion length LD ltlt Optical absorption
length 1/a
?ED
2.exciton diffusion
?CT
3.charge transfer
?CC
?A
0.9
4.charge collection
1.absorption
absorption
quantum efficiency
0.6
EQE?AIQE?A?ED
0.3
cathode
anode
0.0
400
600
800
wavelength(nm)
acceptor
donor
2xLD
EQE External Quantum Efficiency
LA
31
How can we maximize the optical absorption of a
thin-film?
32
Geometric Light Trapping
t100µm
50µm
10µm
t2µm with light trapping
2µm
Silicon

• Geometric light trapping makes films look up to
  4n2 times thicker
• Proof assumes geometric optics regime
• Surely, we must be able to do better in the
  subwavelength regime where coherence can be
  exploited?

T. Tiedje et. al., IEEE Trans. Electron. Dev.31,
711 (1984)
33
What are the limits to light trapping in the
sub-wavelength domain? Can you beat geometric
optics (4n2)?
34
Optical Antennas

• Radiation capture cross section (?/2)2
• Focusing of captured power into a very small
  volume

35
A Different Kind of Plasmon-Assisted Solar Cell

• Ag nanoparticles scatter incident light into
  waveguided/TIR modes
• Far field effect
• Equivalent to geometric approaches

36
Metal Nanostructures As Optical Antennas
X10-19
10nm Au sphere in CuPc
Absorption in organic
(W/m3)
Total absorption
Organic only
Absorbed Power (W)
Isolated Au NP
10nm Au sphere at the active interface
w/ AuNPs
10nm
X 105
4
15nm
3
PTCBI
w/o AuNPs
2
CuPc
1
Fujimori, Dinyari, Lee and Peumans, in press
37
How Does This Work?

• Note you cant do this in the geometric optics
  domain
• This is intrinsically a wave phenomenon

38
Electrostatically-Assisted Aerosol Deposition
Vacuum-deposited
10nm
EA-Aerosol (dry, clean, finely dispersed)
15nm
Vacuum-deposited
Pump
Chamber
Holder
VSG
Substrate
Grid
Mist
Corona discharger
Material solution
10kV
Nozzle
Electrode A
N2
Heater
Mass flow controller
Water
Ultrasonicator (2.4MHz)
Atomizer
Fujimori, Dinyari, Lee and Peumans, in press
39
Electrostatically-Assisted Aerosol Deposition
w/ discharge
w/o discharge
200
Dye on glass sub.
100
-
-
-
-
-
-
-
-
-
-
-
1?m
1?m
0nm
-
-
-
-
-
-
-
-
-
-
-
-
AuNR on TEM grid
-
-
-
-
-
-
-
-
1?m
1?m
-
-
-
-
-
-
40
Increase in Optical Absorption
Surfactant hexadecyl-trimethyl-ammoniumbromide
(CTAB)
41
Increased Cell Performance
S1
S1
20nm
S2
S3
0.1nM AuNS
S2
S3
S2
S1
0.1nM AuNS 2ppm Alq3
S3
AuNSs solution
Fujimori, Dinyari, Lee and Peumans, in press
42
Broadband Cavity Organic Solar Cell
40 increase in JSC

• For very thin absorbers, cavity structures can be
  used to increase optical absorption across a wide
  spectral range
• Benefits are inversely proportional to spectral
  bandwidth

Agrawal and Peumans, Optical Express 2008
43
Cavities
Guided Resonances
Optical Antennas
44
Textured Thin-Film Solar Cells

Guided Resonances
From Rech and Wagner, Appl. Phys. A69, 155 (1999).
45
Geometric Light Trapping is Ineffective in Thin
Films
1.2µm-thick µc-Si cell
16
x10
8
Yablonovitch Limit
7
6
5
100 Increase
40 Increase
4
Short Circuit Current Density
3
Flat Cell
2
1
0
200
400
600
800
1000
1200
1400
Wavelength (nm)

• Flat 19 mA/cm2
• Textured 37 mA/cm2

• Flat 15 mA/cm2
• Textured 21 mA/cm2

J. Muller et al., Solar Energy77, 917 (2004)

• Geometric light trapping does not work well for
  thin-film solar cells
• Subwavelengthtextures dont couple efficiently
• Local photon density of states in thin slab lt
  bulk
• Transparent conductors / metal reflector are
  partially absorptive

46
Light Trapping in the Wave Domain
2um

• At ?1000nm, Labs213µm? 5 absorption single
  pass
• 2n2 enhancement in optical path length should
  yield 100 absorption
• 5 ltlt 2n2 25
• Can we do better?

P. Bermel, et al., Optics Express 15, 16986 (2007)
47
Geometric Light Trapping RCWA Study
Cross-section
Top view
Layer 1
Period
Layer 2
Layer 3

• Constraints Imposed
• First layer homogeneous Absorption of short
  wavelengths
• Symmetrical in two directions Polarization
  independence
• Volume of silicon constant Fair comparison with
  flat cell
• Periodicity Square and hexagonal lattices

48
Spectral Response
Agrawal and Peumans, submitted
49
Conclusions

• Metal optical antennas can increase local optical
  absorption in thin films
• One-dimensional tuned cavities can increase
  optical absorption in thin films
• Two-dimensional dielectric gratings achieve
  Yablonovitch limit in thin sub-wavelength
  structures
• The limits to light trapping in the wave domain
  are identical to those in the geometric domain
• This is also the case for surface
  plasmonpolaritons
• Wave optics (dielectric or surface
  plasmonpolaritons) does not allow you to beat
  4n2!
• But you can get very close to the limit

50
Spectral Selective Absorbers and Emitters for TPV
March 2009 _at_ IMEC, Belgium

• Nicholas Sergeant
• Peter Peumans
• Stanford Organic Electronics Lab
• Electrical Engineering Department
• Stanford University

51
Introduction

• Selective coatings are essential for the
  performance of CSP and TPV systems

Solar TPV
51
52
Aperiodic metallo-dielectric stacks

• The aperiodic multilayer stacks are composed of
  metals (W, Mo, Ta) and dielectrics (MgF2, MgO,
  TiO2, Al2O3) with a high refractive index
  contrast.
• Stacks have 5 to 20 layers of subwavelength
  thicknesses varying from 5 to 150 nm.

1 mm
Substrate
52
53
Absorber Optimization Problem

• Design an absorber stack (green) that has
• Maximized absorptivity in the solar spectrum
• Low parasitic IR emission

Solar Spectrum
BB radiation TltTsun
a
1
NIR ? FIR
Vis
l
53
54
TPV Absorber
Mo TiO2 MgF2 aperiodic stack (green) optimized
for absorption at 1500K. (Blackbody and Solar
Spectrum are normalised)
55
Emitter Optimization Problem

• Design an emitter that has
• an emissivity spectrum (green) peaked just above
  the bandgap energy Egap
• Low emissivity for other wavelengths

56
TPV Emitter
Mo MgO aperiodic stack optimized for emission
towards GaSb cell. Mo TiO2 MgF2 aperiodic stack
also optimized for emission towards GaSb cell.