Future prospects of semiconductor materials for solar and

photoelectrochemical cells

W. Walukiewicz Electronic Materials ProgramMaterials Sciences Division

Lawrence Berkeley National Laboratory

This work was supported by the Director's Innovation Initiative, National Reconnaissance Office and by the Office of Science, U.S. Department of Energy under Contract No. DE-AC03-76SF00098.

Solar to Fuel – Future Challenges and Solutions

LBNL Workshop March 28 – 29, 2005

LBNL solar workshop

Collaborators

J. Wu, K. M. Yu, W. Shan, J. W. Ager, J. Beeman, E. E. Haller, M. Scarpulla, O. Dubon, and J. Denlinger

Lawrence Berkeley National Laboratory, University of California at Berkeley

W. Schaff and H. Lu, Cornell UniversityA. Ramdas and I. Miotkowski, Purdue UniversityP. Becla, Massachusetts Institute of Technology

LBNL solar workshop

Outline

High Efficiency Solar Cell Concepts New semiconductors for multijunction solar cells

GaxIn1-xN alloys

Intermediate band solar cell materials Highly mismatched alloys (HMAs) II-Ox-VI1-x HMAs as intermediate band materials

Group III-nitrides for photoelectrochemical cells Challenges and prospects

LBNL solar workshop

Solar CellsUltimate Efficiency Limits

Intrinsic efficiency limit for a solar cell using a single semiconducting material is 31%.

Light with energy below the bandgap of the semiconductor will not be absorbed

The excess photon energy above the bandgap is lost in the form of heat.

Single crystal GaAs cell: 25.1% AM1.5, 1x

Multijunction (MJ) tandem cell Maximum thermodynamically

achievable efficiencies are increased to 50%, 56%, and 72% for stacks of 2, 3, and 36 junctions with appropriately optimized energy gaps

AM

1.5

sol

ar fl

ux(1

021 p

hoto

ns/s

ec/m

2 /m

)

1 2 3 4Energy (eV)

1

2

3

4

5

Cell 1 (Eg1)

Cell 2 (Eg2)

Cell 3 (Eg3)

Eg1 > Eg2 > Eg3

LBNL solar workshop

Multijunction Solar CellsState-of-the art 3-junction GaInP/Ga(In)As/Ge solar cell: 36 % efficient

M. Yamaguchi et. al. – Space Power Workshop 2003

LBNL solar workshop

Direct bandgap tuning range of In1-xGaxNPotential material for MJ cells

The direct energy gap of In1-xGaxN covers most of the solar spectrum Multijunction solar cell based on this single ternary could be very efficient

LBNL/Cornell work: J. Wu et al. APL 80, 3967 (2002)

LBNL solar workshop

InGaN is radiation hardelectron, proton, and He+ irradiation

LBNL solar workshop

Surface Electron Accumulation

Surface/interface native defects (dangling bonds) are similar to radiation-induced defects

VB

CBEg = 0.7eV

Surface Bulk

EF

~1eV

EFS

High concentration of defects near surface – Fermi level pinning

-7.5

-7.0

-6.5

-6.0

-5.5

-5.0

-4.5

-4.0

0 0.2 0.4 0.6 0.8 1

ECBE

EVBE

Ban

d E

dge

(eV

, rel

ativ

e to

vac

uum

leve

l)

GaNE

g=3.4eV

InNE

g=0.7eV

GaAsE

g=1.4eV

GaInPE

g=1.9eV

X (Ga)

In1-x

GaxN

EFS

LBNL solar workshop

P-type Doping of InN

-1000

-500

0

500

1000

1500

0 50 100 150 200 250

Ene

rgy

(meV

)

Depth (Å)

CB

VB

EF

NA = 1019

ND = 1017

Eb(acc)

= 0

LBNL solar workshop

In1-xGaxN alloys as solar materials

Significant progress in achieving p-type doping Exceptional radiation hardness established

Surface electron accumulation in In-rich alloys Quality of InN/GaInN interfaces

LBNL solar workshop

Multijunction vs. Multiband

junction1

junction2

junction3

I

Multi-junction• Single gap (two bands) each junction• N junctions N absorptions• Efficiency~30-40%

Multi-band• Single junction (no lattice-mismatch)• N bands N·(N-1)/2 gaps

N·(N-1)/2 absorptions• Add one band add N absorptions

LBNL solar workshop

0

Ei

Eg

Ec

iv

cicv

VB

IB

CB

qV

Luque et. al. PRL, 78, 5014 (1997)

Theoretical efficiency of Intermediate band solar cells

Intermediate Band Solar Cells can be very efficient Max. efficiency for a 3-band cell=63% Max. efficiency for a 4-band cell=72% In theory, better performance than any other ideal structure of similar complexity

But NO multi-band materials realized to date

LBNL solar workshop

Highly Mismatched Alloys for Multiband Cells

Oxygen in II-VI compounds has the requisite electronegativity and atomic radius difference

XO = 3.44; RO = 0.073 nmXS = 2.58; RS = 0.11nmXSe = 2.55; RSe = 0.12 nmXTe = 2.1; RTe = 0.14

Oxygen level in ZnTe is 0.24 eV below the CB edge

Can this be used to form an intermediate band?

Synthesis Very low solid solubility limits of

O in II-VI compounds Nonequilibrium synthesis

required

-2.0

0.0

2.0

5.2 5.6 6.0 6.4 6.8

VBCB

Ene

rgy

(eV

)

Lattice parameter (Å)

EFS

EO

CdTe

CdTe

MnTe

MnTe

ZnTe

ZnTeZnSe

ZnSe

MgTe

MgTe

LBNL solar workshop

Zn1-yMnyOxTe1-x: Intermediate Band Material

K. M. Yu et. al., Phys. Rev. Lett., 91, 246403 (2003)

LBNL solar workshop

Zn0.88Mn0.12O0.03Te0.97: Intermediate Band Semiconductor

LBNL solar workshop

Photovoltaic action

LBNL solar workshop

How efficient can they be?Multi-band ZnMnOTe alloys

The location and the width of the intermediate band in ZnMnOxTe1-x is determined by the O content, x

Can be used to maximize the solar cell efficiency

30

35

40

45

50

55

60

2.1

2.2

2.3

2.4

2.5

2.6

2.7

0 0.01 0.02 0.03

max

imum

effi

cien

cy (%

)

optimal operation voltage (V

)

oxygen mole fraction

Zn0.88

Mn0.12

OxTe

1-x

EO=2.06eV

EM=2.32eV

C=3.5eV

Calculations based on the detailed balance model predict maximum efficiency of more than 55% in alloys with 2% of O

1.6

1.8

2.0

2.2

2.4

2.6

2.8

0.00 0.05 0.10 0.15 0.20

Zn0.88

Mn0.12

OxTe

1-x

E_E

+

0 0.01 0.02 0.03 0.04

ener

gy (e

V)

CLM

2x (eV2)

EM

=2.32eV

EO=2.06eV

O mole fraction, x

E_

E+

LBNL solar workshop

Intermediate band semiconductors Challenges an prospects

Synthesis of suitable materials with scalable epitaxial techniques (MBE growth of ZnOxSe1-x achieved)

N-type doping of intermediate band with group VII donors (Cl, Br)

Control of surface properties of the PLM synthesized materials

Other highly mismatched alloys: GaPyNxAs1-x-y

Fundamentals Nature of the intermediate band: localized vs. extended Carrier relaxation processes

LBNL solar workshop

Photoelectrochemical cells for hydrogen generation

Joel W. Ager, Alexis T. Bell,* Miquel Salmeron, Wladek WalukiewiczElectronic Materials ProgramMaterials Sciences Division

Lawrence Berkeley National Laboratory

*Chemical Sciences Division

InN support:FY03 LDRD, FY04 Director's Innovation Initiative, National Reconnaissance Office

LBNL solar workshopJ. A. Turner, Science 285, 687 (1999)

LBNL solar workshop

Photoelectrochemical H2 generation

1.       Absorption of light near the surface of the semiconductor creates electron-hole pairs.

2.       Holes (minority carriers) drift to the surface of the semiconductor (the photo anode) where they react with water to produce oxygen: 2h+ + H2O -> ½ O2 (g) + 2H+

3.       Electrons (majority carriers) are conducted to a metal electrode (typically Pt) where they combine with H+ ions in the electrolyte solution to make H2 :

2e- + 2H+ -> H2 (g)

4.       Transport of H+ from the anode to the cathode through the electrolyte completes the electrochemical circuit.

The overall reaction : 2h + H2O -> H2(g) + ½ O2 (g)

LBNL solar workshop

Why is it hard to do?

Oxides Stable but efficiency is

low (large gap) III-Vs

Efficiency is good but surfaces corrode

Approaches Dye sensitization

(lifetime issues) Surface catalysis

No practical PEC H2 production demonstrated

Efficiency and lifetimeAdapted from M. Grätzel, Nature 414, 388 (2001)

LBNL solar workshop

What are the fundamental issues?

Band structure engineering To match water redox potentials and achieve high solar

efficiency Fundamental understanding of the

electrode/electrolyte interface To accelerate water splitting reaction and reduce corrosion

LBNL solar workshop

Why use nitrides?Direct bandgap tuning range of InGaN

The direct energy gap of In1-xGaxN covers most of the solar spectrum Multijunction solar cell based on this single ternary could be very efficient

LBNL/Cornell work: J. Wu et al. APL 80, 3967 (2002)

LBNL solar workshop

III-Nitrides – tuning the band edges

Their conduction and valence band edges straddle the H+/H2 and O2/H2O redox potentials.

They can be made with the optimal bandgap of ~2.0 eV

Experimentally determined by our group

They have superior corrosion resistance compared to other semiconductors of similar energy gaps.

InGaN: J. Wu et al. APL 80, 3967 (2002)GaNAs: J. Wu et al., PRB

LBNL solar workshop

PhotocurrentIn0.37Ga0.63N

LBNL solar workshop

Surface modification conceptsCatalysis and corrosion inhibition

Catalysts can facilitate the oxidation of water on the anode and reduction of protons on the cathode

Candidate materials Anode – Pt, Pt/Ru alloys, RuO2, MoO3, ZrO2

Cathode – Porphyrins, phtalocyanins, ferrocenes Corrosion can be inhibited by an oxide coating

H2OHO O H H

e

H+ h

h + e

e

Catalyst Photoanode

O2

To cathode

LBNL solar workshop

Fundamental and Practical Issues

Synthesis of materials: MBE, MOCVD, PLM Charge transport and doping Evaluate photo cathode (p-type semiconductor

surface) vs. photo anode (n-type semiconductor surface) designs

Measurements of band offsets Fundamental studies in-situ and ex-situ of the

electrolyte-semiconductor interface Surface modification Kinetic H2 production and corrosion rates.