Principles of photovoltaic energy conversion
and pathways to high efficiency
Louise C. Hirst
NRC associate
U.S. Naval Research Laboratory
louise.hirst.ctr@nrl.navy.mil
U.S. Naval Research Laboratory
The Naval Research Laboratory provides:
Primary in-house research for the physical, engineering,
space, and environmental sciences
Fundamentals
New capabilities
Field demonstration
Experimentalist
Test engineer
Theorist
Founded by Thomas Edison in 1923
Key achievements:
First modern U.S. radar
First operational U.S. sonar
NAVSTAR GPS - based on the NRL TIMATION
program
First US Earth orbiting spacecraft - Vanguard I
Semi-Insulating Gallium Arsenide Crystals Technique for growing high-purity single crystals
louise.hirst.ctr@nrl.navy.mil
slide 1 of 26
Opto-electronics and radiation effects
III-V device design an fabrication
Simulation capabilities
Growth & characterization
5000 ft2 class 100 cleanroom
Photovoltaics
Innovation for high specific power and power
density for specialized applications
Operation in extreme environments (e.g. space,
underwater)
Large scale power generation
Satellite space experiments
Design, build, operate, and post-flight analysis
Radiation effects
Optical detectors for imager technologies
Night vision, thermal imaging, missile detection
louise.hirst.ctr@nrl.navy.mil
slide 2 of 26
The Solar resource
Irradiance (Wm-2eV-1)
700
The Sun is a blackbody with temperature ~6000 K
600
500
E2
dE
2ΩA
I(E) = 2 3
c h exp(E/kT)-1
400
300
1000 W.m-2 peak incident power
200
100
0
1
2
E (eV)
3
4
NASA 2009
Solar energy use is innate
Directly as heat
Chemical, mechanical & electrical
conversion
Indrect an inefficient processes
Photovoltaics provides direct conversion
louise.hirst.ctr@nrl.navy.mil
slide 3 of 26
Incumbent PV technology
1839 - Becquerel demonstration of photovoltaic effect
1954 - Bell Labs first ``high-power" Si solar cell (6%)
Extremely fast growing industry - 38 GW capacity installed in 2013
Si-wafer PV - 90% of 2013 production
Multi-crystalline Si - 55% of total production
Record conversion efficiency for this technology is 20.4%
Most installed system ~15%
Extremely cheap - cells produced $0.2/W
Achieving grid parity in many parts of the world
US residential installations outstripping non-residential
1/3 coming online without state incentives
louise.hirst.ctr@nrl.navy.mil
slide 4 of 26
High efficiency PV
Key issue with multi-crystalline Si:
Limited applications
Domestic power generation, suburban
or rural residential installation
22% energy consumption - residential (2011)
Industrial, commercial and transport
Weight/area are significant
Industrial architectures
High power density requirements
Portable
low power density (W/m2)
low specific power (W/kg)
High efficiency PV provides solutions
World record efficiency 44.7%
Why is the efficiency of incumbent
technology is fundamentally limited?
Mechanisms for achieving high efficiency
louise.hirst.ctr@nrl.navy.mil
slide 5 of 26
Solar heat engine
1
TS
QA
Absorber
QA = AσTS4 QE = AσTA4
TA4
QA - Q E
Net energy flux
=1- 4
=
Incident energy
TS
QA
0.8
η (%)
QE
TA
Q1
0.6
0.4
Heat engine - transfer energy
hot source to cold sink
0.2
W
0
0
Q2
T0
2000
4000
TA (K)
C
B
S
TA
TA
T0
T0
B
C
louise.hirst.ctr@nrl.navy.mil
D
6000
η=
2nd law : ΔS≥0
W
Q -Q
T
Q
= 1 2 = 1 - 2= 1 - 0
Q1
Q1
TA
Q1
A isotherm
B
P
( (( (
TA4
η= 1- 4
TS
1-
T0
TA
adiabat
D
A
T
A
Q1
Q
+ 2
TA T0
ΔS = -
D
C
V
84.9% at
TA = 2480 K
Carnot engine is isoentropic
slide 6 of 26
Solar heat engine: invalid assumptions
TS
QA
How do current PV technologies deviate from the ideal?
( (( (
QE
TA4
η= 1- 4
TS
TA
T
1- 0
TS
Mismatch between absorption and emission angles
-the terrestrial absorber is not a perfect blackbody cavity
Q1
W
Some energy transfers from TS to T0 without being
absorbed
Q2
Some heat dissipation - dewar flasks are not perfect
thermal insulators
T0
C
B
S
TA
TA
T0
T0
B
C
louise.hirst.ctr@nrl.navy.mil
A isotherm
B
P
adiabat
D
A
T
A
84.9% at
TA = 2480 K
D
Non-isoentropic
ΔS=0
C
V
D
slide 7 of 26
Energy (eV)
Semiconductors as absorbers
Entropy free work is generated by thermalization
Teh>TL
energy
C: Eg = 5.5 eV
Si: Eg = 1.1 eV
Ge: Eg = 0.7 eV
CB
Sn: Eg = 0.1 eV
Pb: Eg = --
Teh=TL
time
CB
Δμeh
VB
VB
atomic radius
<ps
elastic
absorption scattering
thermalization radiative
rec
Recitification: with charge separation, the chemical
potential becomes a voltage across the device
The Photovoltaic Effect
Eg
Semiconductors have an bandgap
V = eμ
Most often realized in
a pn junction structure
Some transmission permitted to prevent
total thermalization
louise.hirst.ctr@nrl.navy.mil
slide 8 of 26
Detailed balance limiting efficiency
Detailed balance approach:
J = Jabs - Jemit
∞
∞
- e α(E).n(E, TA, μA = eV, Ωemit).dE
0
Generalized Planck equation
E2
.dE
c2h3 exp(E-μ/kT) - 1
n(E, T, μ, Ω).dE = 2Ω
Particle number conserved
Unity absorptivity above Eg and zero below
Current density A.m-2
0
Fraction of incident solar radiation
= e α(E).n(E, TS, μ = 0, ΩS).dE
1
0.8
0.6
400
200
0.4
Voltage 1
0.2
Power
All recombination radiative
0
Infinite carrier mobility
Maximum power point opperation
Boltzmann approximation: (E-μ)/kT >> 1
louise.hirst.ctr@nrl.navy.mil
0.5
1
1.5
2
2.5
3
3.5
Eg (eV)
32.5% for Eg = 1.35
(31% for 1.31 eV numerical)
slide 9 of 26
Intrinsic losses
Mismatch between absorption and emission
angles:
kTA. ln(Ωemit/Ωabs). Jopt
Fraction of incident solar radiation
1
0.8
0.6
0.4
Boltzmann
0.2
Power
0
0.5
1
1.5
2
2.5
3
3.5
Eg (eV)
louise.hirst.ctr@nrl.navy.mil
slide 10 of 26
Intrinsic losses
Transmission of low energy photons
Eg
E.n(E, Ts, μ=0, ΩS).dE
0
Mismatch between absorption and emission
angles:
kTA. ln(Ωemit/Ωabs). Jopt
Fraction of incident solar radiation
1
0.8
0.6
below Eg
0.4
Boltzmann
0.2
Power
0
0.5
1
1.5
2
2.5
3
3.5
Eg (eV)
louise.hirst.ctr@nrl.navy.mil
slide 11 of 26
Intrinsic losses
Thermalization of high energy photons
∞
E.n(E, Ts, μ=0, ΩS).dE - Eg. Jabs
Eg
Transmission of low energy photons
Eg
E.n(E, Ts, μ=0, ΩS).dE
0
Mismatch between absorption and emission
angles:
kTA. ln(Ωemit/Ωabs). Jopt
Fraction of incident solar radiation
1
thermalization
0.8
0.6
below Eg
0.4
Boltzmann
0.2
Power
0
0.5
1
1.5
2
2.5
3
3.5
Eg (eV)
louise.hirst.ctr@nrl.navy.mil
slide 12 of 26
Intrinsic losses
Carnot
Emission
Eg(TA/TS). Jopt
Eg. Jemit
TS
Carnot
emission
QA
QE
1
Thermalization of high energy photons
∞
E.n(E, Ts, μ=0, ΩS).dE - Eg. Jabs
Eg
Transmission of low energy photons
Eg
E.n(E, Ts, μ=0, ΩS).dE
0
Mismatch between absorption and emission
angles:
kTA. ln(Ωemit/Ωabs). Jopt
Fraction of incident solar radiation
TA
Q1
thermalization
W
0.8
Q2
T0
0.6
below Eg
0.4
Boltzmann
0.2
Power
0
0.5
1
1.5
2
2.5
3
3.5
Eg (eV)
louise.hirst.ctr@nrl.navy.mil
slide 13 of 26
Pathways to high efficiency
Target dominant intrsinsic loss mechanisms:
Carnot
Boltzmann loss
emission
1
Hot carriers
Focusing module
Field localization
~100 nm
|E|
2
1000
Directional
emission
100
10
Furman et al., 35th
IEEE PVSC, 475 (2010)
Adams et al., 35th
Ciracì et al., Science IEEE PVSC, 1 (2010)
337, 1072 (2012)
Thermalization & below Eg
Multiple absorbers
Hot-carrier solar cells
Fraction of incident solar radiation
Solar concentration
thermalization
0.8
0.6
below Eg
0.4
Boltzmann
0.2
Power
0
0.5
1
1.5
2
2.5
3
3.5
Eg (eV)
Sequential absorption and MEG
louise.hirst.ctr@nrl.navy.mil
slide 14 of 26
Multi-junction solar cells
fraction of incident solar radiation
1.0
Carnot
emission
0.9
Boltzmann
Bandgap optimization is a materials engineering challenge
0.8
0.7
thermalization
0.6
below Eg
0.5
Conditions:
Spectral conditions: terrestrial/space
concentrator/flat plate?
Stacked cell or monolithic growth?
Separately contacted or current matched?
0.4
Solutions: III-V alloys
0.3
0.2
Power
0.1
0.0
Increasing junction number increases efficiency through
a reduction in thermalization and below Eg losses
1
III
V
B
N
Al Si P
Ga Ge As
In Sb
2
3
4
5
6
number of junctions
Optimal Eg and separately contacted
junctions assumed
louise.hirst.ctr@nrl.navy.mil
slide 15 of 26
Multi-junction solar cells
AlP
2.5
GaP
Industry work horse, space applications:
Current matching
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Eg optimization
AlAs
2.0
Eg (eV)
AlSb
1.5
GaAs
Si
Emerging technologies for high efficiency:
InP
1.0
GaSb
Ge
0.5
InAs
0.0
5.4
5.6
5.8
6.0
6.2
lattice constant (Å)
louise.hirst.ctr@nrl.navy.mil
InSb
6.4
slide 16 of 26
Multi-junction solar cells
AlP
2.5
GaP
Industry work horse, space applications:
Current matching
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Eg optimization
AlAs
2.0
Eg (eV)
AlSb
1.5
GaAs
Si
Emerging technologies for high efficiency:
InP
Quantum wells
1.0
Reduce middle cell
Eg for current matching
GaSb
Ge
0.5
InAs
0.0
5.4
5.6
5.8
6.0
6.2
lattice constant (Å)
InSb
6.4
Felixble system for Eg
optimization
Ekins-Daukes et al., Appl. Phys. Lett., 75, p. 4195 (1999)
InGaAs GaAsP GaAs
louise.hirst.ctr@nrl.navy.mil
slide 17 of 26
Multi-junction solar cells
AlP
2.5
Industry work horse, space applications:
Current matching
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Eg optimization
AlAs
GaP
2.0
Eg (eV)
AlSb
1.5
GaAs
Si
Emerging technologies for high efficiency
InP
Quantum wells
1.0
Reduce middle cell
Eg for current matching
GaSb
Ge
0.5
InAs
0.0
5.4
5.6
5.8
6.0
6.2
lattice constant (Å)
InSb
6.4
Felixble system for Eg
optimization
Metamorphic growth
MM cell
InGaP
Confine defects to a buffer layer to
move lattice constant
InGaAs
buffer
Ge
louise.hirst.ctr@nrl.navy.mil
slide 18 of 26
Multi-junction solar cells
AlP
2.5
Industry work horse, space applications:
Current matching
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Eg optimization
AlAs
GaP
2.0
Eg (eV)
AlSb
1.5
GaAs
Si
Emerging technologies for high efficiency:
InP
Quantum wells
1.0
Reduce middle cell
Eg for current matching
GaSb
Ge
0.5
InAs
0.0
5.4
5.6
5.8
6.0
6.2
lattice constant (Å)
InSb
6.4
IMM cell
MM cell
InGaP
InGaAs
buffer
Ge
louise.hirst.ctr@nrl.navy.mil
Ge
InGaP
GaAs
buffer
InGaAs
Felixble system for Eg
optimization
Metamorphic growth
Confine defects to a buffer layer to
move lattice constant
Geisz, et al., Appl. Phys. Lett, 93, p. 123505 (2008)
slide 19 of 26
Multi-junction solar cells
AlP
2.5
Industry work horse, space applications:
Current matching
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Eg optimization
AlAs
GaP
2.0
Eg (eV)
AlSb
1.5
GaAs
Si
Emerging technologies for high efficiency:
InP
Quantum wells
1.0
Reduce middle cell
Eg for current matching
GaSb
Ge
0.5
InAs
0.0
5.4
5.6
5.8
6.0
6.2
lattice constant (Å)
Ge
InGaP
GaAs
InGaAsP
InGaAs
InP
InSb
6.4
Felixble system for Eg
optimization
Metamorphic growth
Confine defects to a buffer layer to
move lattice constant
Bonded 4J Soitec - World record (44.7%, 297X)
Epitaxial lift-off & mechanical stacking
Dimmroth, et al. Prog. Photovolt., 22, p. 277 (2014)
louise.hirst.ctr@nrl.navy.mil
slide 20 of 26
NRL pathway to 50%
1.0
AlAs
AM1.5D, low AOD, 500X
2.0
Eg (eV)
AlSb
1.5
GaAs
Si
InGaAs/InGaAs strain
compensated QW
InP
0.6
0.4
0.2
0.0
-0.2
0
1.0
GaSb
Ge
0.5
InAs
0.0
5.4
0.8
5.6
5.8
6.0
6.2
lattice constant (Å)
InAlAsSb
InGaAsP
InGaAs/InGaAs QW
InSb
6.4
In0.39GaAs
GaP
3J on InP (1.74, 1.17, 0.7 eV)
In0.66GaAs
2.5
Energy (eV)
AlP
Eg
5 10 15 20
Distance (nm)
25
InAlAsSb quaternary - new material
Simulate material properties:
Ternary end point lattice constants and band
alignements from experiment
Estimate bowing parameter to interpolate
MBE growth for development:
Immiscibility - kinetics and thermodynamics
Temperature - Growth and anneal
Characterization:
Emission : PL
Device: QE and DLTS
Absorption: PLE and transmission
InP
louise.hirst.ctr@nrl.navy.mil
slide 21 of 26
Broader perspective
Still significant issues with MJ devices other than laboratory efficiency:
Cost - extremely expensive materials and fabrication methods
Implement in highly focusing solar concentrator systems
Improved efficiency through Boltzmann loss reduction
Severly limits applications options - only suitable for desert power stations
$/W values - difficult to compete with flate plate Si
Industrial, commercial, transport & portable applications
low power density (W/m2)
low specific power (W/kg)
Materials abundance
Substrate removal and reuse
Spectral sensivity - limits annual energy yields
Recycling
Ultimately limited by junction number - complexity is not free and only offers incremental
improvement in efficiency
louise.hirst.ctr@nrl.navy.mil
slide 22 of 26
The hot-carrier solar cell
Carriers do not fully thermalization
Change the rate balance between
absorption and thermalization
Steady-state hot-carrier population
Teh>TL
energy
Fundamentally different heat engine
Teh=TL
time
CB
Δμeh
VB
SJ
MJ
hot-carrier
Contact to the hot-carrier population
via an energy selective contact
Cooling confined within an ∞ narrow
energy range is isoentropic
Carrier population thermally equilibrates
without dissipating excess heat energy
elastic
absorption scattering
thermalization radiative
rec
TL
εe
TS > Teh> TL
εeh
Most like the Carnot engine
Δμeh
Δμ
Eg
εh
metal
louise.hirst.ctr@nrl.navy.mil
ESC
absorber
ESC
metal
slide 23 of 26
Making an HCSC
Requirements
Hot-carrier absorber
Broadband absorber
Restricted carrier-phonon interaction
Steady-state non-equilbrium
hot carrier population
Achievable levels of solar illumination
Energy selective contact
Reduced range of energy states
relative to absorber
Carrier transmission
High current, concentrator devices
Recent progress
Slow carrier cooling in QWs
Record low thermalization
coefficient in InAlAs/InGaAs wells
Resonant tunneling
Dimmock et al., Prog. Photovolt, 22, p. 151, 2014
in press IEEE J. Photovolt., 2014
InAlAs
340
320
InAlAs
InP
15 nm
360
Teh (K)
Device solutions
Energy selective extraction
300
20
40
60
Pabs
80
(W.cm-2)
100
120
InGaAs
E-field enhancement nanostructures
Absorption in ultra-thin device
High carrier density -> hotter carriers
louise.hirst.ctr@nrl.navy.mil
Semi-selective energy barrier
Hirst et al., Appl. Phys. Lett., 2014
Quaternary superlattice structures
on InP for miniband formation
slide 24 of 26
NRL hot-carrier solar cell design
Ultra-thin (active region < 100 nm)
incident
sunlight
nanostructure <λ
Reduction thermalization quantum
well hot-carrier absorber
Superlattice energy selective contact
Plamonic waveguiding nanostrucutre
200 nm
Cost
Material abundance
0
0
field enhancement in ultra-thin InGaAs
quantum engineered PV converter
-100 0 100
Spectral sensivity
No complexity limit
CMP tomorrow 2:30pm in room 103 Nielsen Hall
louise.hirst.ctr@nrl.navy.mil
slide 25 of 26
The finish line
Carnot
emission
1
Fraction of incident solar radiation
Multi-crystalline Si - 55% of total production
Extremely cheap - cells produced $0.2/W
Key issue with multi-crystalline Si:
low power density (W/m2)
low specific power (W/kg)
Pathways to high efficiency
MJ - Leading high efficiency PV technology
thermalization
0.8
0.6
below Eg
0.4
Boltzmann
0.2
Power
0
0.5
Race to 50%
InAlAsSb
InGaAsP
2
2.5
Ultimately limited by junction number
Teh>TL
3
3.5
Teh=TL
time
CB
Δμeh
InGaAs/InGaAs QW
What's there at the finish line?
1.5
Eg (eV)
energy
Metamorphic buffers
Quantum wells
ELO and bonding
1
VB
InP
hot-carrier
SJ
MJ
elastic
absorption scattering
thermalization radiative
rec
louise.hirst.ctr@nrl.navy.mil
slide 26 of 26