1 R. R. King, University Seminar, 2014 Novel Semiconductor Materials for High-Efficiency Multijunction Photovoltaics Richard R. King Spectrolab, Inc. Sylmar, CA
1 R. R. King, University Seminar, 2014
Novel Semiconductor Materials for High-Efficiency Multijunction Photovoltaics
Richard R. King
Spectrolab, Inc. Sylmar, CA
2 R. R. King, University Seminar, 2014
Novel Semiconductor Materials for High-Efficiency Multijunction Photovoltaics
Richard R. King
Spectrolab, Inc. Sylmar, CA
Seminar University of California, Santa Barbara
Jan. 24, 2014
3 R. R. King, UCSB Seminar, Jan. 24, 2014
• Manuel Romero, Sarah Kurtz, Dan Friedman, Daryl Myers, Tom Moriarty, Keith Emery – NREL • Angus Rockett – Univ. of Illinois • Gerald Siefer – CalLab, Fraunhofer ISE • Geoff Kinsey – Fraunhofer CSE • Rosina Bierbaum – Univ. of Michigan • Russ Jones, Jim Ermer, Chris Fetzer, Abdallah Zakaria, Xing-Quan Liu, Daniel Law, Philip Chiu, Shoghig Mesropian, Xiaogang Bai, Dimitri Krut, Kent Barbour, Mark Takahashi, Andrey Masalykin, John Frost, Nasser Karam ...and the entire multijunction solar cell team at Spectrolab
Acknowledgments
4 R. R. King, UCSB Seminar, Jan. 24, 2014
• Big Picture
• Solar cell efficiency limits
• Unifying behavior in semiconductor energy levels
• Electronic activity of defects in different semiconductor families
• Multijunction solar cells and concentrator photovoltaics (CPV)
Outline
0
100
200
300
400
500
600
700
0 0.5 1 1.5 2 2.5 3 3.5 4Photon Energy (eV)
Inte
nsity
per
Uni
t Pho
ton
Ener
gy(W
/m 2
. eV
)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Phot
on u
tiliz
atio
n ef
ficie
ncy
.
AM1.5D, ASTM G173-03, 1000 W/m2Utilization efficiency of photon energy 1-junction cell 3-junction cell 6-junction cell
5 R. R. King, UCSB Seminar, Jan. 24, 2014
Big Picture
6 R. R. King, UCSB Seminar, Jan. 24, 2014
IPCC (2001) scenarios to 2100 IPCC (2001) scenarios to 2100
1000 years of Earth temperature history… and 100 years of projection
Climate Change – Temperature Anomaly by Year
• Fossil fuels are contributing to global climate change at alarming rate
• Further, dependence on imported fuels has a high toll in terms of political stability and national security
Rosina Bierbaum, Univ. of Michigan Intergovernmental Panel on Climate Change (IPCC)
7 R. R. King, UCSB Seminar, Jan. 24, 2014
Cogentrix, Alamosa, CO – 30 MW, III-V multijunction cells
Courtesy Amonix
Concentrator Photovoltaics (CPV) using high-efficiency III-V
3-junction solar cells
• Photovoltaic solar electricity combined with...
– power storage in plug-in hybrid vehicles
– long-distance power transmission from sunny locales to high-demand areas
offer a major part of a solution to these problems
8 R. R. King, UCSB Seminar, Jan. 24, 2014
Solar cell efficiency limits
9 R. R. King, UCSB Seminar, Jan. 24, 2014
Detailed Balance Limit of Solar Cell Efficiency
• 30% efficient single-gap solar cell at one sun, for 1 e-/photon
• 44% ultimate efficiency for device with single cutoff energy
Shockley and Queisser (1961)
Single- junction solar cell
10 R. R. King, UCSB Seminar, Jan. 24, 2014
0
100
200
300
400
500
600
700
0 0.5 1 1.5 2 2.5 3 3.5 4Photon Energy (eV)
Inte
nsity
per
Uni
t Pho
ton
Ener
gy(W
/m 2
. eV
)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Phot
on u
tiliz
atio
n ef
ficie
ncy
.
AM1.5D, ASTM G173-03, 1000 W/m2Utilization efficiency of photon energy to bandgap Eg to Voc at 1000 suns to Voc at 1 sun
Photon Utilization Efficiency 1-Junction Solar Cells
11 R. R. King, UCSB Seminar, Jan. 24, 2014
• Assumptions in Shockley and Quiesser (1961)
• Viewed from a different angle, these limitations represent opportunities for higher efficiency devices
Assumption limiting solar cell efficiency Device principle overcoming this limitation Single band gap energy Multijunction solar cells
Quantum well, quantum dot solar cells One e--h+ pair per photon Down conversion
Multiple exciton generation Avalanche multiplication
Non-use of sub-band-gap photons Up conversion
Single population of each charge carrier type Hot carrier solar cells Intermediate-band solar cells Quantum well, quantum dot solar cells
One-sun incident intensity Concentrator solar cells
Assumptions → Opportunities
12 R. R. King, UCSB Seminar, Jan. 24, 2014
C. H. Henry (1980) Theoretical Efficiency of Multijunction Solar Cells
3-junction solar cell
Theo. efficiency at 1000 suns 1J: 37% 2J: 50% 3J: 56% 36J: 72%
13 R. R. King, UCSB Seminar, Jan. 24, 2014
0
100
200
300
400
500
600
700
0 0.5 1 1.5 2 2.5 3 3.5 4Photon Energy (eV)
Inte
nsity
per
Uni
t Pho
ton
Ener
gy(W
/m 2
. eV
)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Phot
on u
tiliz
atio
n ef
ficie
ncy
.
AM1.5D, ASTM G173-03, 1000 W/m2Utilization efficiency of photon energy 1-junction cell 3-junction cell
Photon Utilization Efficiency 3-Junction Solar Cells
14 R. R. King, UCSB Seminar, Jan. 24, 2014
Unifying behavior in semiconductor energy levels
15 R. R. King, UCSB Seminar, Jan. 24, 2014
Unifying trends of defect energy levels in different semiconductors
• A number of phenomena have been observed indicating the tendency for some defect levels to form at a nearly constant energy in different semiconductors with respect to the vacuum level
Universal alignment of hydrogen levels in semiconductors Van de Walle and Neugebauer, Nature (2003)
Bulk reference level in Cu(GaxIn1-x)(SySe1-y)2 semiconductors Turcu, Kötschau, and Rau,, J. Appl. Phys. (2002)
Fermi-level stabilization energy Walukiewicz, Phys. Rev. B (1988)
• These observations indicate a common defect configuration at the atomic scale may be responsible for the near constancy of energy level in different semiconductors
• The position of common defect energies has the power to explain and predict doping properties and defect recombination activity – in the same semiconductor family ( e.g., CuInSe2 and Cu(GaIn)(SSe)2 ) – and in different families ( e.g., Cu(GaIn)(SSe)2 and GaInN )
16 R. R. King, UCSB Seminar, Jan. 24, 2014
Hydrogen energy levels in different host semiconductors
• Hydrogen energy level (red bars) is at nearly constant level with respect to vacuum level when incorporated into a wide variety of semiconductors
• Applies to group-IV, III-V, and II-VI semiconductors, and even to electrolytes Van de Walle and Neugebauer, Nature, 2003
17 R. R. King, UCSB Seminar, Jan. 24, 2014
Unifying trends of defect energy levels in different semiconductors
• Questions: → To what extent and in what circumstances do the energy levels
of certain defects mirror the constancy of hydrogen atom states in different semiconductors?
→ Does this behavior establish the tendency for some types of defects to form near a specified energy from the vacuum level?
18 R. R. King, UCSB Seminar, Jan. 24, 2014
Admittance Spectroscopy
• Changes in capacitance with frequency and temperature → extraction of defect energies and densities through admittance spectroscopy
• Leads to a clear picture of the evolution of spontaneously forming defect energies in Cu(GaxIn1-x)(SySe1-y)2 chalcopyrites with changing S and Ga composition
Turcu, Kötschau, Rau, JAP, 2002
19 R. R. King, UCSB Seminar, Jan. 24, 2014
Similarity of defect energy levels in different chalcopyrites
• Trap energies due to native defects in Cu(GaxIn1-x)(SySe1-y)2 are approximately constant with respect to group-VI composition in semiconductor
• Trap energies are closer to midgap for higher bandgap compositions → leads to higher recombination at higher Eg
→ challenge for finding high Eg top cell material for chalcopyrite-based multijunction cell Turcu, Kötschau, Rau, JAP, 2002
20 R. R. King, UCSB Seminar, Jan. 24, 2014
Unifying trends of defect energy levels in different semiconductors
• Questions: → To what extent and in what circumstances do the energy levels
of certain defects mirror the constancy of hydrogen atom states in different semiconductors?
→ Does this behavior establish the tendency for some types of defects to form near a specified energy from the vacuum level?
→ To what degree does the reference bulk defect energy level observed in I-III-VI chalcopyrites extend to other semiconductor families?
→ What does this behavior say about the recombination at defects in different host semiconductors, especially the low recombination activity observed in materials such as CuInSe2 and GaInN?
21 R. R. King, UCSB Seminar, Jan. 24, 2014
Fermi-level stabilization energy
• Fermi level stabilizes at nearly constant level with respect to vacuum level in a wide variety of semiconductors and semiconductor families → Both for damage induced by radiation and at semiconductor surfaces → Close to energy of hydrogen incorporated in semiconductors
• Fermi-level stabilization energy EFS is near midgap in GaAs and GaInP, but near conduction band in GaInN → reduces recombination activity of states at EFS as In content goes up
Walukiewicz, PRB, 1988; Li et al., PRB, 2005
22 R. R. King, UCSB Seminar, Jan. 24, 2014
Experimental and Theoretical Bandgap-Voltage Offset Woc
0.0
0.5
1.0
1.5
2.0
0.6 1 1.4 1.8 2.2Band Gap Eg (eV)
E g/q
, Voc
, and
(Eg/q
) - V
oc (
V)measured Vocmeas. Eg from EQEWoc = (Eg/q) - Vocradiative recomb. onlydetailed balance model
d-A
lGaI
nP
GaA
s1.
4 - e
V G
aInA
s
o-G
aInP
AlG
aInA
s
d-A
lGaI
nP d
-GaI
nP
d-A
lGaI
nP
0.97
-eV
GaI
nAs
GaI
nNA
s
1.10
-eV
GaI
nAs
1.24
-eV
GaI
nAs
1.30
-eV
GaI
nAs
Ge
(ind
irect
gap
)
AlG
aInA
s
Voc and band gap-voltage offset Woc = (Eg/q) - Voc
of solar cells with wide range of band gaps
Si
(indi
rect
gap
)
0.79
-eV
GaI
nAs
• Difference between bandgap and steady-state quasi-Fermi level splitting (open-circuit voltage) in solar cells is strikingly similar across wide range of III-V and group-IV semiconductors (as well as II-VI, I-III-VI, and other classes of semiconductors)
Ref.: R. R. King et al., Prog. in PV, doi: 10.1002/pip.1044 (2010)
23 R. R. King, UCSB Seminar, Jan. 24, 2014
Semiconductor Eqns. Formulated in Terms of Band Gap-Voltage Offset W
• Woc formulation has more physical basis, related to NC , NV rather than ni2
• Far more invariant with respect to Eg , good for multiple subcells in MJ cells
• Makes Woc a convenient indicator of solar cell quality – amount of SRH recombination vs. radiative recombination – for wide range of Eg
• Ko ≡ Jo / ni2 has nearly all band gap dependence taken out
kTqVi enpn /2=
= 2ln
inpn
qkTV
kTqWVC eNNpn /−=
( ) VqEW g −≡
=
pnNN
qkTW VCln
kTqW
VC
ph
i
oo
oceNN
JnJK /
2 =≡kTqVpho
oceJJ /−=
24 R. R. King, UCSB Seminar, Jan. 24, 2014
Unifying trends of defect energy levels in different semiconductors
• Each of these observations reveals a profound, unifying aspect of the fundamental nature of these materials
• Connections between defect energies in very different semiconductor systems
→ may help identify fundamental principles behind the remarkably low defect recombination activity in some types of semiconductors, such as CuInSe2 and GaInN
• Finding answers to these questions will help us understand at a deeper level why we observe the semiconductor properties that we do
25 R. R. King, UCSB Seminar, Jan. 24, 2014
Electronic activity of defects in different
semiconductor families
26 R. R. King, UCSB Seminar, Jan. 24, 2014
Electronic activity of defects in different semiconductor families
• Semiconductor growth and characterization
• Some interesting materials systems – Metamorphic III-Vs
– Dilute nitride GaInNAs(Sb)
– Polycrystalline I-III-VIs and related materials Chalcopyrites (CuvAg1-v)(AlxGayIn1-x-y)(SzSe1-z)2 e.g., CIGS
Kesterite Cu2ZnSn(S,Se)4
– Perovskites CH3NH3Pb(X)3 where X = Cl, Br, I
27 R. R. King, UCSB Seminar, Jan. 24, 2014
Organometallic Vapor-Phase Epitaxy (OMVPE)
As
Arsine, AsH3
H
Hydrogen, H2
In
Trimethylindium TMIn, In(CH3)3
P
Phosphine, PH3 Diethyltelluride
DETe, Te(C2H5)2
Te
Ga
Trimethylgallium or TMGa, Ga(CH3)3
C
Typical Precursor Chemicals:
H2 + TMGa
H2 AsH3/Ph3
Sealed “Bubbler” containing MO source precursor
Chemical Reactor Chamber Low Pressure & High Temperature
Exhaust Gasses
Point Of Use Gas Scrubber
Scrubbed Hydrogen
Crystal growth on substrates on heated rotating graphite
H2
28 R. R. King, UCSB Seminar, Jan. 24, 2014
Molecular Beam Epitaxy
29 R. R. King, UCSB Seminar, Jan. 24, 2014
Selenization and Sulfidation CIGS and CZTS
www.unk.edu/nss/chemistry.aspx?id=41606
www.beltfurnaces.com/efficiency_of_CIGS.html
www.prweb.com/releases/Smit_ovens/CIGS/prweb2039344.htm
auo.com/print.php?sn=192&lang=en-US
onlin
elib
rary
.wile
y.co
m/d
oi/1
0.10
02/c
phc.
v14.
9/is
suet
oc
30 R. R. King, UCSB Seminar, Jan. 24, 2014
Recombination activity at defects
• Characterization of energy levels, capture cross sections, density, and recombination activity of defects: – deep-level transient spectroscopy (DLTS) – admittance spectroscopy, both in the dark and with photoexcitation – time-resolved photoluminescence (TRPL) – among others
• Atomic reconstruction and bonding configuration of defects: – transmission electron microscopy (TEM) – atom probe tomography (APT) – atomic force microscopy (AFM) – Fourier transform infrared spectroscopy (FTIR) – other imaging and characterization tools
31 R. R. King, UCSB Seminar, Jan. 24, 2014
Metamorphic III-Vs
32 R. R. King, UCSB Seminar, Jan. 24, 2014
Terrestrial Conc. Cell Designs from 40% to 50%
(a) (b) (c) (d) (e) (f) (g) (h) (i) (j)3J Lattice-
Matched (LM)C3MJ+
3J Meta-morphic (MM)low mismatch
C4MJ
3J Meta-morphic (MM)high mismatch
3J InvertedMetamorphic
(IMM)
4J Meta-morphic (MM)high mismatch
4J Double-Grade InvertedMetamorphic
(MMX2)
5J Lattice-Matched (LM)
w. epitaxial Ge subcell
5J Lattice-Matched (LM)w. GaInNAsSb
subcell
5J Lattice-Matched (LM)
SemiconductorBonded (SBT)
6J Triple-GradeInverted
Metamorphic(MMX3)
MJ Cell 39.42% 40.00% 40.54% 43.26% 44.44% 47.87% 43.25% 47.43% 47.64% 50.91% EfficiencyChange 0.0% 1.5% 2.8% 9.7% 12.7% 21.4% 9.7% 20.3% 20.9% 29.2% in Power from C3MJ+ Efficiencies for AM1.5D, ASTM G173-03 spectrum, 50.0 W/cm2 (500 suns), 25°C
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Subc
ell B
and
Gap
s (e
V)
34%
36%
38%
40%
42%
44%
46%
48%
50%
52%
MJ
Cel
l Effi
cien
cy (
%)
C1 Eg
C2 Eg
C3 Eg
C4 Eg
C5 Eg
C6 Eg
MJ CellEfficiency
transparent buffer
1.35-eV GaInAs cell 2
1.83-eV GaInP cell 1
metal gridline
0.67-eV Ge cell 3and substrate
• Modeled production avg. efficiency of 40.0% at 500 suns (50.0 W/cm2)
33 R. R. King, UCSB Seminar, Jan. 24, 2014
• Low dislocation density in active cell layers in top portion of epilayer stack: ~ 2 x 105 cm-2 from EBIC and CL meas. • Dislocations confined to graded buffer layers in bottom portion of epilayer stack
GaInAs cap
GaInAs MC
GaInP TC
0.2 µm
Tunnel junction
Pre-grade buffer
Misfit dislocations
GaInAs graded buffer to 8%-In
0.2 µm
Ge substrate
Cross sectional TEM Ga0.44In0.56P/ Ga0.92In0.08As/ Ge
Cell
34 R. R. King, UCSB Seminar, Jan. 24, 2014
Line of 0%relaxation
Line of 100%relaxation
Qy
(Stra
in) Å
-1
Qx (Tilt) Å-1
Ge
Ga0.92In0.08As MC
GaInP TC
GradedBuffer
(115) glancing exit XRD
Line of 0%relaxation
Line of 100%relaxation
Qy
(Stra
in) Å
-1
Qx (Tilt) Å-1
Ge
Ga0.92In0.08As MC
GaInP TC
GradedBuffer
(115) glancing exit XRD
• GaInP/ 8%-In GaInAs/ Ge metamorphic (MM) cell structure
• Nearly 100% relaxed step-graded buffer → removes driving force for dislocations to propagate into active cell layers
• 56%-In GaInP top cell pseudomorphic with respect to GaInAs middle cell
High-Resolution XRD Reciprocal Space Map (RSM)
35 R. R. King, UCSB Seminar, Jan. 24, 2014
0
10
20
30
40
50
60
70
80
90
100
300 400 500 600 700 800 900 1000 1100 1200 1300 1400Wavelength (nm)
Inte
rnal
Qua
ntum
Effi
cien
cy (%
)
0.96
eV1.40
1.08
1.26
1.38
1.30
GaInAs single-junction solar cells
2.4% lattice mismatch
1.6% lattice mismatch
Internal QE of Metamorphic GaInAs Cells on Ge
Metamorphic = "changed form"
36 R. R. King, UCSB Seminar, Jan. 24, 2014
1 µm 20 µm
Cathodoluminescence (CL)
disloc. density = 4.4 x 106 cm-2
Plan-View Transmission Electron Microscopy (TEM)
disloc. density = 3.1 x 106 cm-2
23%-In GaInAs double heterostructure on Ge
Dislocation Imaging in 23%-In GaInAs
37 R. R. King, UCSB Seminar, Jan. 24, 2014
Ge or GaAs substrateGe or GaAs substrate
nucleationnucleation
buffer layerbuffer layerbuffer layerbuffer layer
emitter
1.39-eV GaInAsinverted LM subcell
base
emitter
1.39-eV GaInAsinverted LM subcell
base
contactcontact
metalmetal
50 µm 8e-9766-1
1.39-eV ILM subcell GaInAs comp. 2% In Latt. mismatch 0.1% Disloc. density 2.5 x 105 cm-2
50 µm 8e-9756-1
50 µm 8e-9760-1
50 µm 8e-9783-11
EBIC images and dislocation density of inverted metamorphic cell test structures
1.39-eV GaInAs
1.10-eV IMM subcell 23% In 1.6% 3.9 x 106 cm-2
0.97-eV IMM subcell 33% In 2.3% 5.0 x 106 cm-2
0.84-eV IMM subcell 44% In 3.1% 6.3 x 106 cm-2
Ge substrateGe substrate
nucleation and pre-grade buffernucleation and pre-grade buffer
transparent MM transparent MM graded buffer layersgraded buffer layers
transparent MM transparent MM graded buffer layersgraded buffer layers
emitter
1.10-eV GaInAsinverted MM subcell
base
emitter
1.10-eV GaInAsinverted MM subcell
base
contactcontact
metalmetal
1.10-eV GaInAs
Ge substrateGe substrate
nucleation and pre-grade buffernucleation and pre-grade buffer
transparent MM transparent MM graded buffer layersgraded buffer layers
transparent MM transparent MM graded buffer layersgraded buffer layers
emitteremitter
0.970.97--eV GaInAseV GaInAsinverted MM subcellinverted MM subcell
basebase
emitteremitter
0.970.97--eV GaInAseV GaInAsinverted MM subcellinverted MM subcell
basebase
contactcontactcontactcontact
metalmetal
0.97-eV GaInAs
Ge substrateGe substrate
nucleation and pre-grade buffernucleation and pre-grade buffer
transparent MM transparent MM graded buffer layersgraded buffer layers
transparent MM transparent MM graded buffer layersgraded buffer layers
emitteremitter
0.840.84--eV GaInAseV GaInAsinverted MM subcellinverted MM subcell
basebase
emitteremitter
0.840.84--eV GaInAseV GaInAsinverted MM subcellinverted MM subcell
basebase
contactcontactcontactcontact
metalmetal
0.84-eV GaInAs
Dislocations in Inverted Metamorphic Cells – EBIC
38 R. R. King, UCSB Seminar, Jan. 24, 2014
0
1
2
3
4
5
6
7
8
9
0 10 20 30 40 50In Composition for GaxIn1-xAs (%)
Dis
loca
tion
Den
sity
from
EB
IC (
106 c
m-2
)
and
Pho
ton
Inte
nsity
from
CL
(10 3
cps
)
0
5
10
15
20
25
30
35
40
45
50
Car
rier L
oss
(%)
Dislocation density from EBICOverall % photon intensity from CL% carrier loss at each dislocation from CL
-0.07 0.64 1.36 2.07 2.79 3.51Lattice Mismatch Relative to Ge (%)
Inverted metamorphic (MM)GaxIn1-xAs solar cells
Dislocations in Inverted Metamorphic Cells
Ref.: R. R. King et al., 23rd European Photovoltaic Solar Energy Conf., Valencia, Spain, Sep. 2008.
39 R. R. King, UCSB Seminar, Jan. 24, 2014
1
10
100
1000
10000
0 100 200 300 400 500
Time (ns)
Phot
olum
ines
cenc
e In
tens
ity (
arb.
uni
ts)
τeff = 47 ns
AlGaInP/ GaInP/ AlGaInP double heterostructure
1
10
100
1000
0 2000 4000 6000 8000 10000
Time (ns)
Phot
olum
ines
cenc
e In
tens
ity (
arb.
uni
ts)
τeff = 2450 ns
GaInP/ GaInAs/ GaInP double heterostructure
• Double heterostructures grown with AlGaInP/GaInP and GaInP/GaInAs interfaces, in stack similar to MJ cells
• TRPL measurements at NREL
• Minority-carrier lifetime up to 2450 ns in 1%-In GaInAs on Ge substrate
Time-Resolved Photoluminescence:
GaInP- and GaInAs-base DHs
40 R. R. King, UCSB Seminar, Jan. 24, 2014
0.1
1
10
100
1000
10000
0 5 10 15 20 25 30 35Indium Mole Fraction of GaInAs Lattice-Matched to Base (%)
τ eff
Mea
sure
d by
TR
PL (
ns)
Base MaterialRecent data nid-GaInAs, recent data p-GaInP (disordered)Previous data nid-GaInAs nid-GaInP (ordered) nid-GaInP (disordered)
Eg = 1.407 eV
1.813 eV
1.887 eV 1.311 eV
1.736 eV
1.807 eV
1.114 eV
1.619 eV
0.994 eV
1.529 eV
Time-Resolved PL of LM & MM Double Heterostructures
41 R. R. King, UCSB Seminar, Jan. 24, 2014
Sublattice ordering
42 R. R. King, UCSB Seminar, Jan. 24, 2014
P Ga In
[001]
[100] [010]
(111) planes
all In
all Ga
all In
Group-III Sublattice Ordering
Ga0.5In0.5P fully ordered
(order parameter η = 1)
CuPtB ordering on [111] or [111] planes
In practice:
η = 0.4-0.5 , Eg ≈ 1.8 eV for GaInP lattice
matched to GaAs
Eg(η) = Eg(0) - (0.471 eV)η2
43 R. R. King, UCSB Seminar, Jan. 24, 2014
Ga0.5In0.5P fully disordered
(order parameter η = 0)
No CuPtB ordering
In practice: η = 0.0-0.1 , Eg ≈ 1.9 eV
for GaInP lattice matched to GaAs
Bandgap difference with respect
to ordered GaInP occurs in conduction band → ∆Ec
P Ga In
[001]
[100] [010]
(111) planes
random Ga & In random
Ga & In
random Ga & In
Group-III Sublattice Ordering
44 R. R. King, UCSB Seminar, Jan. 24, 2014
[110] cross-section of disordered GaInP epilayer showing [110]-oriented P dimers of the β(2 x 4) reconstruction. The stresses caused in in the growing crystal by surface phosphorus atoms provide the thermodynamic driving force for ordering.
Surface Reconstruction of GaInP
45 R. R. King, UCSB Seminar, Jan. 24, 2014
0
20
40
60
80
100
120
140
160
0 2 4 6 8 10 12 14Relative Omega, referenced to 1%-In GaInAs Peak (degrees)
1/2(
115)
XR
D In
tens
ity d
ue to
Gro
up-II
I O
rder
ing
in G
aInP
(co
unts
/s)
GaInP Ordering State andLattice Match to GaInAsordered, GaInP LM to 1%-In GaInAspartially disordered, "disordered, "ordered, GaInP LM to 8%-In GaInAsdisordered, "
-2 0 2
½(1
15) X
RD
Inte
nsity
Due
to G
roup
-III
Subl
attic
e O
rder
ing
in G
aInP
(co
unts
/s)
1.813 eV
1.867 eV 1.887 eV1.736 eV
1.807
-2 0 2 -2 0 2
Direct Meas. of GaInP Ordering from ½(115) XRD Peak
46 R. R. King, UCSB Seminar, Jan. 24, 2014
Dilute nitride GaInNAs(Sb)
47 R. R. King, UCSB Seminar, Jan. 24, 2014
Terrestrial Conc. Cell Designs from 40% to 50%
(a) (b) (c) (d) (e) (f) (g) (h) (i) (j)3J Lattice-
Matched (LM)C3MJ+
3J Meta-morphic (MM)low mismatch
C4MJ
3J Meta-morphic (MM)high mismatch
3J InvertedMetamorphic
(IMM)
4J Meta-morphic (MM)high mismatch
4J Double-Grade InvertedMetamorphic
(MMX2)
5J Lattice-Matched (LM)
w. epitaxial Ge subcell
5J Lattice-Matched (LM)w. GaInNAsSb
subcell
5J Lattice-Matched (LM)
SemiconductorBonded (SBT)
6J Triple-GradeInverted
Metamorphic(MMX3)
MJ Cell 39.42% 40.00% 40.54% 43.26% 44.44% 47.87% 43.25% 47.43% 47.64% 50.91% EfficiencyChange 0.0% 1.5% 2.8% 9.7% 12.7% 21.4% 9.7% 20.3% 20.9% 29.2% in Power from C3MJ+ Efficiencies for AM1.5D, ASTM G173-03 spectrum, 50.0 W/cm2 (500 suns), 25°C
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Subc
ell B
and
Gap
s (e
V)
34%
36%
38%
40%
42%
44%
46%
48%
50%
52%
MJ
Cel
l Effi
cien
cy (
%)
C1 Eg
C2 Eg
C3 Eg
C4 Eg
C5 Eg
C6 Eg
MJ CellEfficiency
0.67-eV Ge cell 5and substrate
1.12-eV GaInNAsSb cell 4
1.40-eV GaInAs cell 3
1.71-eV AlGaInAs cell 22.00-eV AlGaInP cell 1
metal gridline
• Modeled production avg. efficiency of 47.4% at 500 suns (50.0 W/cm2)
48 R. R. King, UCSB Seminar, Jan. 24, 2014
Dilute nitride GaInNAs(Sb)
• Two defect levels consistently seen in GaInNAs by DLTS
• Electron trap stays at 0.9-1.0 eV above valence band as N is added and Eg decreases
• Electron trap becomes quite shallow for N compositions near 1 eV
• 2nd defect level follows lowering of conduction band as N is added → stays near midgap and acts as effective recombination center
Ptak et al., NREL PRM, 2003
49 R. R. King, UCSB Seminar, Jan. 24, 2014
Dilute nitride GaInNAs(Sb)
• Presence of H greatly decreases formation energy of (N-H-VGa)2- defect complex
• Acts as acceptor
• May be one of the main fundamental defects in GaInNAs(Sb)
• Positron annihilation spectroscopy shows increased VGa in MBE-grown nitride in the presence of H
Ptak et al., NREL PRM, 2003
50 R. R. King, UCSB Seminar, Jan. 24, 2014
Dilute nitride GaInNAs(Sb)
• Quantum efficiencies well above 90% and comparable to those for GaAs can be achieved for a range of N compositions and bandgaps in GaInNAs(Sb) grown by MBE
• These current densities are high enough to contribute usefully to high-efficiency multijunction solar cells
Ptak et al., JAP, 2005
51 R. R. King, UCSB Seminar, Jan. 24, 2014
Dilute nitride GaInNAs(Sb)
• Dilute nitride GaInNAs grown by MOVPE can have high measured quantum efficiencies
• Annealing and thermal history have a strong influence on the quality of dilute nitride GaInNAs(Sb)
Volz et al., JCG, 2008
52 R. R. King, UCSB Seminar, Jan. 24, 2014
• FTIR spectra of Ga0.975In0.025N0.002As0.998 (2.5% In) • N primarily bound to Ga in Ga4N configuration before anneal (467 cm-1) • Higher fraction of N bound to In in Ga3InN configuration after anneal → greater mass of In causes lower vibrational frequency signal (457 cm-1) • Evidence for H forming bonds with N in GaInNAs – change upon annealing may be due to change from NH to NH2 defect complex
Dilute nitride GaInNAs(Sb)
• 0.2% N
• 2.5% In
Kurtz et al., APL, 2001
53 R. R. King, UCSB Seminar, Jan. 24, 2014
Dilute nitride GaInNAs(Sb)
• Probability of N finding In-N nearest neighbor environment much higher for greater In compositions in GaInNAs
• Change in N bonding environment also thought to be cause of blueshift in fundamental bandgap of GaInNAs upon annealing → ~20 mV per additional indium atom in bonding configuration
Volz et al., JCG, 2008
54 R. R. King, UCSB Seminar, Jan. 24, 2014
Dilute nitride GaInNAs(Sb)
• Chain-like columnar nitrogen ordering in the [001] direction causes local strain fluctuations as seen in strain sensitive (202) dark field TEM images
• Nitrogen chains are broken up resulting in more homogeneous strain in GaInNAs with increasing anneal schedule in images (a), (b), (c)
Volz et al., JCG, 2008
55 R. R. King, UCSB Seminar, Jan. 24, 2014
Chalcopyrites
Cu(In,Ga)Se2 (“CIGS”) and more
56 R. R. King, UCSB Seminar, Jan. 24, 2014
The CIGS Device
Light absorbing layer: a Cu(In,Ga)Se2 alloy Remarkably low recombination at extended defects Alternate window (emitter) layers available Back contact: nearly always Mo for stability
Angus Rockett – U. Illinois
57 R. R. King, UCSB Seminar, Jan. 24, 2014
TEM image Schematic
Devices are thought to be limited by recombination in the depletion region, not by heterojunction recombination.
• What is the major recombination center?
• What do grain boundaries do?
• Why does CuGaSe2 not work well?
• Why do some growth processes work better than others?
The Real Device
Angus Rockett – U. Illinois
58 R. R. King, UCSB Seminar, Jan. 24, 2014
c
a
In
SeYet:
• Extended defects inactive
• Polar surfaces most stable
• Hole mobility phonon limited for p to >1019 cm-3
• Polycrystalline devices work better than single crystals
• Disordering energy is low so there are many point defects
• A polar compound so charged surfaces could be a problem
Chalcopyrite CIGS
Angus Rockett – U. Illinois
59 R. R. King, UCSB Seminar, Jan. 24, 2014
• Layers facet spontaneously into polar (112) type planes
• Smooth facets alternate with rough facets
• Indexing surface planes shows smooth planes are metal terminated
(220)/(204) epitaxial layer AFM image
Red: metal terminated Blue: Se terminated
(112
) A (112)B
Conclusion: Somehow the polar surfaces are stabilized, giving a very strong preference for these.
(220)/(204) Oriented CIGS
Angus Rockett – U. Illinois
60 R. R. King, UCSB Seminar, Jan. 24, 2014
CIGS solar cells • Are heterojunction devices with a very strongly inverted
junction (Cd doping overwhelms Fermi level pinning).
• Do not mind grain boundaries because they are highly faceted to extremely passive (112) surfaces.
• Heterojunction is made to these surfaces regardless of grain orientation.
• Point defects control doping in the bulk and are very consistent.
• Edge dislocations do not matter because they turn into (112) surfaces.
Conclusions of all of this…
Angus Rockett – U. Illinois
61 R. R. King, UCSB Seminar, Jan. 24, 2014
Perovskites
62 R. R. King, UCSB Seminar, Jan. 24, 2014
Perovskite solar cells
• Present perovskite-based solar cells evolved from dye-sensitized solar cell (DSSC) technology
• Perovskite CH3NH3PbClxI3-x absorbers have very high absorption coefficients allowing thin, practical layers to be used
• Very simple processing
63 R. R. King, UCSB Seminar, Jan. 24, 2014
Perovskite solar cells
Snaith, JPCL, 2013
• Much of electron and hole transport can take place through the perovskite light absorber material itself, rather than through the porous TiO2 scaffold used in DSSCs
• Perovskite cells work even better with insulating porous Al2O3 scaffold, avoiding voltage loss of 0.2-0.3 V from lower conduction band of TiO2
• External quantum efficiency (incident photon-to-electron conversion efficiency, or IPCE) of two types of perovskite absorbers, with bandgaps of ~2.2 and 1.55 eV
• >12% 1-sun eff. using porous Al2O3 scaffold, 15% with vapor-deposited perovskite
64 R. R. King, UCSB Seminar, Jan. 24, 2014
Perovskite solar cells
Snaith, JPCL, 2013
• High Eg of perovskites (2.2 eV for CH3NH3PbBr3 to 1.55 eV for CH3NH3PbI3 ) a good match for top cell of flat-plate, one-sun multijunction with silicon, Cu(GaxIn1-x)(SySe1-y)2 or kesterite Cu2ZnSn(S,Se)4 bottom cell
65 R. R. King, UCSB Seminar, Jan. 24, 2014
Multijunction solar cells and concentrator
photovoltaics (CPV)
66 R. R. King, UCSB Seminar, Jan. 24, 2014
Lattice-Matched and Metamorphic Cell Structure
67 R. R. King, UCSB Seminar, Jan. 24, 2014
Record efficiency III-V multijunction solar cells
68 R. R. King, UCSB Seminar, Jan. 24, 2014 Concentrator cell light I-V and efficiency independently verified by J. Kiehl, T. Moriarty, K. Emery – NREL
• First solar cell of any type to reach over 40% efficiency
Spectrolab Metamorphic GaInP/ GaInAs/ Ge Cell
Voc = 2.911 V Jsc = 3.832 A/cm2
FF = 87.50% Vmp = 2.589 V
Efficiency = 40.7% ± 2.4%
240 suns (24.0 W/cm2) intensity 0.2669 cm2 designated area 25 ± 1°C, AM1.5D, low-AOD spectrum Ref.: R. R. King et al., "40% efficient
metamorphic GaInP / GaInAs / Ge multijunction solar cells," Appl. Phys. Lett., 90, 183516, 4 May 2007.
Record 40.7%-Efficient Concentrator Solar Cell
• Efficiencies have now reached 41.6% for both metamorphic and lattice-matched 3-junction cells
69 R. R. King, UCSB Seminar, Jan. 24, 2014
External QE of LM and MM 3-Junction Cells
70 R. R. King, UCSB Seminar, Jan. 24, 2014
Record efficiency III-V multijunction solar cells
Solar Junction 3-junction cell with dilute nitride bottom cell
71 R. R. King, UCSB Seminar, Jan. 24, 2014
3-junction record efficiency cell with dilute nitride bottom cell
Solar Junction 44.0% eff. AM1.5D under concentration NREL confirmed
~1-eV GaInNAs(Sb) cell 3
1.42-eV GaAs cell 2
1.89-eV GaInP cell 1
metal gridline
www.semiconductor-today.com/news_items/2012/OCT/SOLARJUNCTION_151012.html
72 R. R. King, UCSB Seminar, Jan. 24, 2014
Record efficiency III-V multijunction solar cells
Soitec/Fraunhofer ISE 4-junction semiconductor bonded cell
73 R. R. King, UCSB Seminar, Jan. 24, 2014
4-junction record efficiency semiconductor bonded cell
Fraunhofer ISE/ Soitec 44.7% eff. AM1.5D under concentration Fraunhofer ISE confirmed
GaInAs cell 4
GaInPAs cell 3
GaAs cell 2(Al)GaInP cell 1
metal gridline
bondedinterface
www.ise.fraunhofer.de/en/press-and-media/press-releases/presseinformationen-2013/world-record-solar-cell-with-44.7-efficiency
74 R. R. King, UCSB Seminar, Jan. 24, 2014
Record Efficiency One-Sun SBT 5-Junction Cell
D. Law, P. Chiu et al., to be published
38.8%
Spectrolab 5-junction semiconductor bonded cell, 1-sun, AM1.5G
75 R. R. King, UCSB Seminar, Jan. 24, 2014
5-junction record 1-sun efficiency semiconductor
bonded cell
Spectrolab 38.8% eff. AM1.5G, 1-sun cell NREL confirmed
0.75-eV GaInAs cell 5
1.1-eV GaInPAs cell 4
1.7-eV AlGaInAs cell 2
1.4-eV GaInAs cell 3
2.0-eV AlGaInP cell 1
metal gridline
semiconductorbondedinterface
76 R. R. King, UCSB Seminar, Jan. 24, 2014
InP growth substrate GaAs or Ge growth substrate
1.4-eV GaInAs cell 3
1.7-eV AlGaInAs cell 2
2.0-eV AlGaInP cell 1 0.75-eV GaInAs cell 5
1.1-eV GaInPAs cell 4
GaAs or Ge growth substrate
1.4-eV GaInAs cell 3
1.7-eV AlGaInAs cell 2
2.0-eV AlGaInP cell 1
1.4-eV GaInAs cell 3
1.7-eV AlGaInAs cell 2
2.0-eV AlGaInP cell 1
GaAs or Ge growth substrate
1.4-eV GaInAs cell 3
1.7-eV AlGaInAs cell 2
2.0-eV AlGaInP cell 1
GaAs or Ge growth substrate
semi-conductor
bonded interface
metal gridline
0.75-eV GaInAs cell 5
1.1-eV GaInPAs cell 4
1.4-eV GaInAs cell 3
1.7-eV AlGaInAs cell 2
2.0-eV AlGaInP cell 1
semi-conductor
bonded interface
metal gridline
– Both high-bandgap and low-bandgap cell sets use high-quality, lattice-matched materials
– Atomically abrupt semiconductor bonded interface
– Both small-lattice (GaAs) and large-lattice (InP) growth substrates can be reused after substrate removal
• Direct semiconductor bonding for multijunction solar cells
Semiconductor-Bonded Technology (SBT) Terrestrial Concentrator Cell
77 R. R. King, UCSB Seminar, Jan. 24, 2014
• Fabricated “preliminary” SBT AM1.5G cell under IR&D leveraging knowledge from SBT space cells.
• Spectrolab’s SBT terrestrial cell achieved efficiency of 37.8%, 1-sun, AM1.5G (then world record) in April 2013.
• Recently achieved new world record efficiency of 38.8%, 1-sun, AM1.5G in August 2013
• Highest efficiency 1-sun terrestrial solar cell of any type.
• Expect Eff. > 47% at moderate concentrations
Record Efficiency = 38.83%
Record Efficiency One-Sun SBT 5-Junction Cell
Chiu et al., PVSC, 2013
78 R. R. King, UCSB Seminar, Jan. 24, 2014
Terrestrial Conc. Cell Designs from 40% to 50%
(a) (b) (c) (d) (e) (f) (g) (h) (i) (j)3J Lattice-
Matched (LM)C3MJ+
3J Meta-morphic (MM)low mismatch
C4MJ
3J Meta-morphic (MM)high mismatch
3J InvertedMetamorphic
(IMM)
4J Meta-morphic (MM)high mismatch
4J Double-Grade InvertedMetamorphic
(MMX2)
5J Lattice-Matched (LM)
w. epitaxial Ge subcell
5J Lattice-Matched (LM)w. GaInNAsSb
subcell
5J Lattice-Matched (LM)
SemiconductorBonded (SBT)
6J Triple-GradeInverted
Metamorphic(MMX3)
MJ Cell 39.42% 40.00% 40.54% 43.26% 44.44% 47.87% 43.25% 47.43% 47.64% 50.91% EfficiencyChange 0.0% 1.5% 2.8% 9.7% 12.7% 21.4% 9.7% 20.3% 20.9% 29.2% in Power from C3MJ+ Efficiencies for AM1.5D, ASTM G173-03 spectrum, 50.0 W/cm2 (500 suns), 25°C
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Subc
ell B
and
Gap
s (e
V)
34%
36%
38%
40%
42%
44%
46%
48%
50%
52%
MJ
Cel
l Effi
cien
cy (
%)
C1 Eg
C2 Eg
C3 Eg
C4 Eg
C5 Eg
C6 Eg
MJ CellEfficiency
0.67-eV Ge cell 5and substrate
1.12-eV GaInNAsSb cell 4
1.40-eV GaInAs cell 3
1.71-eV AlGaInAs cell 22.00-eV AlGaInP cell 1
metal gridline
• Modeled production avg. efficiency of 47.4% at 500 suns (50.0 W/cm2)
79 R. R. King, UCSB Seminar, Jan. 24, 2014
5-junction nitride solar cell measured light I-V
0
2
4
6
8
10
12
14
0 1 2 3 4 5
Voltage (V)
Cur
rent
Den
sity
(m
A/c
m2 )
4J, no nitride, EQE Jsc, no AR 20.2%
4J, no nitride, IQE Jsc 29.8% eff.
5J nitride, EQE Jsc, no AR 22.2%
5J nitride, IQE Jsc 32.4% eff.
AM0 efficiency projected from QP cells, no AR
• Addition of dilute nitride GaInNAs cell to 5-junction stack adds ~400 mV open-circuit voltage to cell at one sun
80 R. R. King, UCSB Seminar, Jan. 24, 2014
5-junction nitride solar cell measured quantum efficiency
0
10
20
30
40
50
60
70
80
90
100
300 500 700 900 1100 1300 1500 1700 1900Wavelength (nm)
Qua
ntum
Eff
icie
ncy
(%)
12.312.0 12.1 13.3 17.6 mA/cm2
IQE
81 R. R. King, UCSB Seminar, Jan. 24, 2014
Terrestrial Conc. Cell Designs from 40% to 50%
(a) (b) (c) (d) (e) (f) (g) (h) (i) (j)3J Lattice-
Matched (LM)C3MJ+
3J Meta-morphic (MM)low mismatch
C4MJ
3J Meta-morphic (MM)high mismatch
3J InvertedMetamorphic
(IMM)
4J Meta-morphic (MM)high mismatch
4J Double-Grade InvertedMetamorphic
(MMX2)
5J Lattice-Matched (LM)
w. epitaxial Ge subcell
5J Lattice-Matched (LM)w. GaInNAsSb
subcell
5J Lattice-Matched (LM)
SemiconductorBonded (SBT)
6J Triple-GradeInverted
Metamorphic(MMX3)
MJ Cell 39.42% 40.00% 40.54% 43.26% 44.44% 47.87% 43.25% 47.43% 47.64% 50.91% EfficiencyChange 0.0% 1.5% 2.8% 9.7% 12.7% 21.4% 9.7% 20.3% 20.9% 29.2% in Power from C3MJ+ Efficiencies for AM1.5D, ASTM G173-03 spectrum, 50.0 W/cm2 (500 suns), 25°C
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Subc
ell B
and
Gap
s (e
V)
34%
36%
38%
40%
42%
44%
46%
48%
50%
52%
MJ
Cel
l Effi
cien
cy (
%)
C1 Eg
C2 Eg
C3 Eg
C4 Eg
C5 Eg
C6 Eg
MJ CellEfficiency
transparent buffer
0.70-eV GaInAs cell 6
0.97-eV GaInAs cell 5transparent buffer
1.20-eV GaInAs cell 4transparent buffer
1.77-eV AlGaAs cell 21.465-eV AlGaAs cell 3
2.00-eV AlGaInP cell 1
metal gridline
Ge or GaAs
growth substrate
• Modeled production avg. efficiency of 50.9% at 500 suns (50.0 W/cm2)
82 R. R. King, UCSB Seminar, Jan. 24, 2014
0
100
200
300
400
500
600
700
0 0.5 1 1.5 2 2.5 3 3.5 4Photon Energy (eV)
Inte
nsity
per
Uni
t Pho
ton
Ener
gy(W
/m 2
. eV
)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Phot
on u
tiliz
atio
n ef
ficie
ncy
.
AM1.5D, ASTM G173-03, 1000 W/m2Utilization efficiency of photon energy 1-junction cell 3-junction cell
Photon Utilization Efficiency 3-Junction Solar Cells
83 R. R. King, UCSB Seminar, Jan. 24, 2014
0
100
200
300
400
500
600
700
0 0.5 1 1.5 2 2.5 3 3.5 4Photon Energy (eV)
Inte
nsity
per
Uni
t Pho
ton
Ener
gy(W
/m 2
. eV
)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Phot
on u
tiliz
atio
n ef
ficie
ncy
.
AM1.5D, ASTM G173-03, 1000 W/m2Utilization efficiency of photon energy 1-junction cell 3-junction cell 6-junction cell
Photon Utilization Efficiency 6-Junction Solar Cells
84 R. R. King, UCSB Seminar, Jan. 24, 2014
0
20
40
60
80
100
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Photon Energy (eV)
Qua
ntum
Effi
cien
cy (
%)
1.91-eV GaInP Cell 1 EQE
Quantum efficiency of subcells in 6-junction cell
85 R. R. King, UCSB Seminar, Jan. 24, 2014
0
20
40
60
80
100
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Photon Energy (eV)
Qua
ntum
Effi
cien
cy (
%)
1.91-eV GaInP Cell 1 EQE1.81-eV GaInP Cell 2 EQE
Quantum efficiency of subcells in 6-junction cell
86 R. R. King, UCSB Seminar, Jan. 24, 2014
0
20
40
60
80
100
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Photon Energy (eV)
Qua
ntum
Effi
cien
cy (
%)
1.91-eV GaInP Cell 1 EQE1.81-eV GaInP Cell 2 EQE1.57-eV AlGaInAs Cell 3 EQE
Quantum efficiency of subcells in 6-junction cell
87 R. R. King, UCSB Seminar, Jan. 24, 2014
0
20
40
60
80
100
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Photon Energy (eV)
Qua
ntum
Effi
cien
cy (
%)
1.91-eV GaInP Cell 1 EQE1.81-eV GaInP Cell 2 EQE1.57-eV AlGaInAs Cell 3 EQE1.39-eV GaInAs Cell 4 EQE
Quantum efficiency of subcells in 6-junction cell
88 R. R. King, UCSB Seminar, Jan. 24, 2014
0
20
40
60
80
100
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Photon Energy (eV)
Qua
ntum
Effi
cien
cy (
%)
1.91-eV GaInP Cell 1 EQE1.81-eV GaInP Cell 2 EQE1.57-eV AlGaInAs Cell 3 EQE1.39-eV GaInAs Cell 4 EQE1.05-eV GaInNAs Cell 5 EQE
Quantum efficiency of subcells in 6-junction cell
89 R. R. King, UCSB Seminar, Jan. 24, 2014
0
20
40
60
80
100
0.5 1.0 1.5 2.0 2.5 3.0 3.5
Photon Energy (eV)
Qua
ntum
Effi
cien
cy (
%)
1.91-eV GaInP Cell 1 EQE1.81-eV GaInP Cell 2 EQE1.57-eV AlGaInAs Cell 3 EQE1.39-eV GaInAs Cell 4 EQE1.05-eV GaInNAs Cell 5 EQE0.67-eV Ge Cell 6 EQE6J, cumulative EQE
Quantum efficiency of subcells in 6-junction cell
90 R. R. King, UCSB Seminar, Jan. 24, 2014
MJ Solar Cell
Bypass Diode
Metallized Substrate
• Individual cells mounted on individual substrates, each with a bypass diode
• Suitable for point focus modules
Concentrator Cell Receivers
91 R. R. King, UCSB Seminar, Jan. 24, 2014
Case Study 1: Soitec Concentrix™ Technology
Courtesy Soitec
92 R. R. King, UCSB Seminar, Jan. 24, 2014
Soitec Concentrix™ Technology
• Concentration ratio ~500
• High efficiency cells based on III-V materials
• Fresnel lens as primary optics
Courtesy Soitec
93 R. R. King, UCSB Seminar, Jan. 24, 2014
Case Study 2: Solar Systems
Courtesy Solar Systems Pty Ltd, Australia
Dense array, active cooling
94 R. R. King, UCSB Seminar, Jan. 24, 2014
Solar Systems
Courtesy Solar Systems Pty Ltd, Australia, Photo: Pierre Verlinden
95 R. R. King, UCSB Seminar, Jan. 24, 2014
• Fossil fuels contributing to global climate change at alarming rate
• Multijunction cells break Shockley-Queisser single-junction efficiency limits
• Wide range of semiconductor bandgaps needed with low recombination
• Unifying behavior in semiconductor energy levels offers deeper experimental and theoretical understanding of universal patterns in formation of defect energy levels across broad classes of semiconductors
• Will provide framework to understand and better use inherently low defect recombination activity in certain polar covalent semiconductors, in families represented by CuInSe2, GaInN, and Cu2ZnSn(S,Se)4
• Novel semiconductor materials enable a zoo of new multijunction solar cells with efficiencies ranging over 50%
• Path to 50% efficiency promises to open wide geographic regions for cost-effective photovoltaics
• Efficiency advantage of 4, 5, and 6J cells outweighs the effect of variable spectrum on current balance
• New understanding of defect structure in semiconductor families such as chalcopyrite, kesterite, and perovskite materials will enable advances in wide-bandgap top cells for flat-plate multijunction cells, bringing together the high efficiency of multijunctions with thin-film technology for low-cost solar electricity
Summary and Future Prospects
96 R. R. King, University Seminar, 2014