Post on 22-Jun-2018
louise.hirst.ctr@nrl.navy.mil
Principles of photovoltaic energy conversion and pathways to high efficiency
Louise C. HirstNRC 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 Field demonstration
Theorist Test engineerExperimentalist
Founded by Thomas Edison in 1923
Key achievements:
First modern U.S. radar
First operational U.S. sonar
NAVSTAR GPS - based on the NRL TIMATIONprogram
First US Earth orbiting spacecraft - Vanguard I
Semi-Insulating Gallium Arsenide Crystals -Technique for growing high-purity single crystals
New capabilities
slide 1 of 26
louise.hirst.ctr@nrl.navy.mil
Opto-electronics and radiation effects
5000 ft2 class 100 cleanroom
Innovation for high specific power and power density for specialized applicationsOperation in extreme environments (e.g. space, underwater)Large scale power generation
Design, build, operate, and post-flight analysis
Satellite space experiments
Photovoltaics
Optical detectors for imager technologies
III-V device design an fabrication
Radiation effects
Night vision, thermal imaging, missile detection
Growth & characterization
Simulation capabilities
slide 2 of 26
louise.hirst.ctr@nrl.navy.mil
The Solar resource
0
100
200
300
400
500
600
700
1 2 3 4
Irra
dia
nce (
Wm
-2eV
-1)
E (eV)
The Sun is a blackbody with temperature ~6000 K
I(E) = 2Ac2h3
E2
exp(E/kT)-1
dE
1000 W.m-2 peak incident power
Solar energy use is innate
Directly as heat
Chemical, mechanical & electrical conversion
NASA 2009
Indrect an inefficient processesPhotovoltaics provides direct conversion
slide 3 of 26
louise.hirst.ctr@nrl.navy.mil
Incumbent PV technology
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
Extremely fast growing industry - 38 GW capacity installed in 2013
US residential installations outstripping non-residential
1/3 coming online without state incentives
1839 - Becquerel demonstration of photovoltaic effect
1954 - Bell Labs first ``high-power" Si solar cell (6%)
slide 4 of 26
louise.hirst.ctr@nrl.navy.mil
High efficiency PV
Key issue with multi-crystalline Si:
low power density (W/m2)
low specific power (W/kg)
Limited applications
Domestic power generation, suburban or rural residential installation
Portable
Weight/area are significant
Industrial architectures
High power density requirements
High efficiency PV provides solutions
Why is the efficiency of incumbenttechnology is fundamentally limited?
World record efficiency 44.7%
Mechanisms for achieving high efficiency
22% energy consumption - residential (2011)
Industrial, commercial and transport
slide 5 of 26
louise.hirst.ctr@nrl.navy.mil
Solar heat engine
TA
TA
T0
W
Q1
Q2
TS
QA QE
0
0.2
0.4
0.6
0.8
1
0 2000 4000 6000
(%
)
TA (K)
84.9% at TA = 2480 K
= ( (1 - TA4TS4 1 -( (T0TA
Heat engine - transfer energyhot source to cold sink
S
T
A
B
D
CP
V
AB
DC
A B C D
isotherm
adiabat
TA T0 T0
QA = ATS4 QE = ATA4Net energy flux Incident energy
=QA - QE
QA= 1 -
TA4
TS4
2nd law : S0
= 1 -T0TA
S = - + Q1TA
Q2T0
= WQ1
Q1 - Q2Q1
= = 1 -Q2Q1
Carnot engine is isoentropic
Absorber
slide 6 of 26
louise.hirst.ctr@nrl.navy.mil
Solar heat engine: invalid assumptions
How do current PV technologies deviate from the ideal?
84.9% at TA = 2480 K
= ( (1 - TA4TS4 1 -( (T0TS
Mismatch between absorption and emission angles-the terrestrial absorber is not a perfect blackbody cavity
Some energy transfers from TS to T0 without beingabsorbed
Some heat dissipation - dewar flasks are not perfectthermal insulators
Non-isoentropic
S=0TA
TA
T0
W
Q1
Q2
TS
QA QE
S
T
A
B
D
CP
V
AB
DC
A B C D
isotherm
adiabat
TA T0 T0
slide 7 of 26
louise.hirst.ctr@nrl.navy.mil
Semiconductors as absorbers
Semiconductors have an bandgap
TL
CB
VB
ener
gy
absorption elasticscattering
time
thermalization radiativerec
Teh=TL
eh
atomic radius
Ener
gy
(eV
)
C: Eg = 5.5 eVSi: Eg = 1.1 eV
Ge: Eg = 0.7 eV
Sn: Eg = 0.1 eV
Pb: Eg = --
CB
VB
slide 8 of 26
louise.hirst.ctr@nrl.navy.mil
Detailed balance limiting efficiency
Detailed balance approach:
Generalized Planck equation
= e
2c2h3 exp(E-/kT) - 1
.dEE2
0
(E).n(E, TS, = 0, S).dE
0
(E).n(E, TA, A = eV, emit).dE - e
n(E, T, , ).dE =
Boltzmann approximation: (E-)/kT >> 1
Infinite carrier mobility
Unity absorptivity above Eg and zero below
All recombination radiative
Particle number conserved
Maximum power point opperation
0
0.2
0.4
0.6
0.8
1
0.5 1 1.5 2 2.5 3 3.5
PowerFr
acti
on o
f in
ciden
t so
lar
radia
tion
Eg (eV)
Voltage
32.5% for Eg = 1.35 (31% for 1.31 eV numerical)
Curr
ent
den
sity
A.m
-2
1
200
400
J = Jabs - Jemit
slide 9 of 26
louise.hirst.ctr@nrl.navy.mil
Intrinsic losses
0
0.2
0.4
0.6
0.8
1
0.5 1 1.5 2 2.5 3 3.5
Power
Boltzmann
Eg (eV)
Mismatch between absorption and emissionangles:
kTA. ln(emit/abs). Jopt
Frac
tion o
f in
ciden
t so
lar
radia
tion
slide 10 of 26
louise.hirst.ctr@nrl.navy.mil
Intrinsic losses
0
0.2
0.4
0.6
0.8
1
0.5 1 1.5 2 2.5 3 3.5
Power
Boltzmann
Eg (eV)
below Eg
Mismatch between absorption and emissionangles:
kTA. ln(emit/abs). Jopt
Frac
tion o
f in
ciden
t so
lar
radia
tion
Transmission of low energy photons
0
Eg
E.n(E, Ts, =0, S).dE
slide 11 of 26
louise.hirst.ctr@nrl.navy.mil
Intrinsic losses
0
0.2
0.4
0.6
0.8
1
0.5 1 1.5 2 2.5 3 3.5
Power
Boltzmann
Eg (eV)
below Eg
Mismatch between absorption and emissionangles:
kTA. ln(emit/abs). Jopt
Frac
tion o
f in
ciden
t so
lar
radia
tion
Transmission of low energy photons
0
Eg
E.n(E, Ts, =0, S).dE
thermalizationThermalization of high energy photons
Eg
E.n(E, Ts, =0, S).dE - Eg. Jabs
slide 12 of 26
louise.hirst.ctr@nrl.navy.mil
Intrinsic losses
0
0.2
0.4
0.6
0.8
1
0.5 1 1.5 2 2.5 3 3.5
Power
Boltzmann
Eg (eV)
below Eg
Mismatch between absorption and emissionangles:
kTA. ln(emit/abs). Jopt
Frac
tion o
f in
ciden
t so
lar
radia
tion
Transmission of low energy photons
0
Eg
E.n(E, Ts, =0, S).dE
thermalizationThermalization of high energy photons
Eg
E.n(E, Ts, =0, S).dE - Eg. Jabs
Carnot Emission
Eg(TA/TS). Jopt Eg. JemitCarnot emission
TA
T0
W
Q1
Q2
TS
QA QE
slide 13 of 26
louise.hirst.ctr@nrl.navy.mil
Pathways to high efficiency
Hot carriers
Target dominant intrsinsic loss mechanisms:
Solar concentration
0
0.2
0.4
0.6
0.8
1
0.5 1 1.5 2 2.5 3 3.5
Power
Boltzmann
Eg (eV)
below Eg
Frac
tion o
f in
ciden
t so
lar
radia
tion
thermalization
Carnot emission
Cirac et al., Science 337, 1072 (2012)
Furman et al., 35th IEEE PVSC, 475 (2010)
~100 nmField localization
10
100
1000|E|2
Focusing module
Boltzmann loss
Multiple absorbers
Thermalization & below Eg
Hot-carrier solar cells
Sequential absorption and MEG
Directional emission
Adams et al., 35th IEEE PVSC, 1 (2010)
slide 14 of 26
louise.hirst.ctr@nrl.navy.mil
Multi-junction solar cells
Optimal Eg and separately contacted junctions assumed
number of junctions
frac
tion o
f in
ciden
t so
lar
radia
tion
1 2 3 4 5 60.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Power
below Eg
thermalization
Boltzmann
emissionCarnot Increasing junction number increases efficiency through
a reduction in thermalization and below Eg losses
Bandgap optimization is a materials engineering challenge
Spectral conditions: terrestrial/spaceconcentrator/flat plate?
Stacked cell or monolithic growth?
Conditions:
Separately contacted or current matched?
BAlGaIn
NPAsSb
SiGe
III VSolutions: III-V alloys
slide 15 of 26
louise.hirst.ctr@nrl.navy.mil
Multi-junction solar cells
InSb
AlSb
InAs
GaSb
InPGaAs
AlAs
AlP
GaP
2.5
2.0
1.5
1.0
0.5
0.05.4 5.6 5.8 6.0 6.2 6.4
lattice constant ()
Eg (
eV
)
Ge
Si
Industry work horse, space applications:
Current matching
Eg optimization
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Emerging technologies for high efficiency:
slide 16 of 26
louise.hirst.ctr@nrl.navy.mil
Multi-junction solar cells
InSb
AlSb
InAs
GaSb
InPGaAs
AlAs
AlP
GaP
2.5
2.0
1.5
1.0
0.5
0.05.4 5.6 5.8 6.0 6.2 6.4
lattice constant ()
Eg (
eV
)
Ge
Si
InGaAs GaAsP GaAs
Industry work horse, space applications:
Quantum wells
Current matching
Eg optimization
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Reduce middle cellEg for current matching
Felixble system for Egoptimization
Emerging technologies for high efficiency:
Ekins-Daukes et al., Appl. Phys. Lett., 75, p. 4195 (1999)
slide 17 of 26
louise.hirst.ctr@nrl.navy.mil
Multi-junction solar cells
InSb
AlSb
InAs
GaSb
InPGaAs
AlAs
AlP
GaP
2.5
2.0
1.5
1.0
0.5
0.05.4 5.6 5.8 6.0 6.2 6.4
lattice constant ()
Eg (
eV
)
Ge
Si
Industry work horse, space applications:
Quantum wells
Current matching
Eg optimization
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Reduce middle cellEg for current matching
Felixble system for Egoptimization
Emerging technologies for high efficiency
Metamorphic growth
Confine defects to a buffer layer tomove lattice constant
Ge
buffer
InGaAs
InGaP
MM cell
slide 18 of 26
louise.hirst.ctr@nrl.navy.mil
Multi-junction solar cells
InSb
AlSb
InAs
GaSb
InPGaAs
AlAs
AlP
GaP
2.5
2.0
1.5
1.0
0.5
0.05.4 5.6 5.8 6.0 6.2 6.4
lattice constant ()
Eg (
eV
)
Ge
Si
Industry work horse, space applications:
Quantum wells
Current matching
Eg optimization
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Reduce middle cellEg for current matching
Felixble system for Egoptimization
Emerging technologies for high efficiency:
Metamorphic growth
Confine defects to a buffer layer tomove lattice constant
buffer
GaAs
InGaP
Ge
Ge
buffer
InGaAs
InGaP
InGaAs
MM cell
IMM cell
Geisz, et al., Appl. Phys. Lett, 93, p. 123505 (2008)
slide 19 of 26
louise.hirst.ctr@nrl.navy.mil
Multi-junction solar cells
InSb
AlSb
InAs
GaSb
InPGaAs
AlAs
AlP
GaP
2.5
2.0
1.5
1.0
0.5
0.05.4 5.6 5.8 6.0 6.2 6.4
lattice constant ()
Eg (
eV
)
Ge
Si
Industry work horse, space applications:
Quantum wells
Current matching
Eg optimization
InGaP/GaAs/Ge
EMCORE ZTJ - 29.5 %
Reduce middle cellEg for current matching
Felixble system for Egoptimization
Emerging technologies for high efficiency:
Metamorphic growth
Confine defects to a buffer layer tomove lattice constant
Bonded 4J Soitec - World record (44.7%, 297X)
Epitaxial lift-off & mechanical stacking
GaAs
InGaP
Ge
InP
InGaAs
InGaAsP
Dimmroth, et al. Prog. Photovolt., 22, p. 277 (2014)
slide 20 of 26
louise.hirst.ctr@nrl.navy.mil
NRL pathway to 50%
InSb
AlSb
InAs
GaSb
InPGaAs
AlAs
AlP
GaP
2.5
2.0
1.5
1.0
0.5
0.05.4 5.6 5.8 6.0 6.2 6.4
lattice constant ()
Eg (
eV
)
Ge
Si
3J on InP (1.74, 1.17, 0.7 eV)
InGaAs/InGaAs strain compensated QW
InAlAsSb quaternary - new material
InGaAsP
InAlAsSb
InP
InGaAs/InGaAs QW
AM1.5D, low AOD, 500X
-0.2
1.0
0.8
0.6
0.4
0.2
0.0
0 5 10 15 20 25Distance (nm)
Ener
gy
(eV
)
In0.6
6G
aAs
In0.3
9G
aAs
Eg
Simulate material properties:
Ternary end point lattice constants and bandalignements from experimentEstimate bowing parameter to interpolate
MBE growth for development:Immiscibility - kinetics and thermodynamicsTemperature - Growth and anneal
Characterization:Emission : PL Device: QE and DLTSAbsorption: PLE and transmission
slide 21 of 26
louise.hirst.ctr@nrl.navy.mil
Broader perspective
Cost - extremely expensive materials and fabrication methods
Still significant issues with MJ devices other than laboratory efficiency:
$/W values - difficult to compete with flate plate Si
Implement in highly focusing solar concentrator systems
Improved efficiency through Boltzmann loss reductionSeverly limits applications options - only suitable for desert power stations
Spectral sensivity - limits annual energy yieldsMaterials abundance
Substrate removal and reuse
Recycling
Ultimately limited by junction number - complexity is not free and only offers incrementalimprovement in efficiency
low power density (W/m2)
low specific power (W/kg)
Industrial, commercial, transport & portable applications
slide 22 of 26
louise.hirst.ctr@nrl.navy.mil
The hot-carrier solar cell
Fundamentally different heat engine
Steady-state hot-carrier population
Teh>TL
CB
VB
ener
gy
absorption elasticscattering
time
thermalization radiativerec
Teh=TL
eh
hot-carrierSJMJ
Change the rate balance betweenabsorption and thermalization
Carriers do not fully thermalization
Contact to the hot-carrier populationvia an energy selective contact
metal metalESC ESCabsorber
Eg
e
h
eh
TL TS > Teh> TL
eh
Cooling confined within an narrowenergy range is isoentropic
Carrier population thermally equilibrateswithout dissipating excess heat energy
Most like the Carnot engine
slide 23 of 26
louise.hirst.ctr@nrl.navy.mil
Making an HCSC
Hot-carrier absorber Energy selective extraction
Broadband absorberRestricted carrier-phonon interaction
Steady-state non-equilbrium hot carrier population
Achievable levels of solar illumination
Energy selective contact
High current, concentrator devices
Reduced range of energy states relative to absorber
Carrier transmission
Req
uir
emen
tsD
evic
e so
luti
on
s
Slow carrier cooling in QWsRecord low thermalization coefficient in InAlAs/InGaAs wells
in press IEEE J. Photovolt., 2014
E-field enhancement nanostructuresAbsorption in ultra-thin device
High carrier density -> hotter carriers
Recent progressResonant tunneling Dimmock et al., Prog. Photovolt, 22, p. 151, 2014
Semi-selective energy barrierHirst et al., Appl. Phys. Lett., 2014
Quaternary superlattice structureson InP for miniband formation
Pabs (W.cm-2)
Teh
(K)
300
320
340
360
20 40 60 80 100 120
InAlA
s
InGaAs
15 nm
InAlA
s
InP
slide 24 of 26
louise.hirst.ctr@nrl.navy.mil
NRL hot-carrier solar cell design
0
-100 0 100
0
200 nm
field enhancement in ultra-thin InGaAsquantum engineered PV converter
nanostructure
louise.hirst.ctr@nrl.navy.mil
The finish line
Pathways to high efficiency
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)
0
0.2
0.4
0.6
0.8
1
0.5 1 1.5 2 2.5 3 3.5
Power
Boltzmann
Eg (eV)
below Eg
Frac
tion o
f in
ciden
t so
lar
radia
tion
thermalization
Carnot emission
MJ - Leading high efficiency PV technology
InGaAsP
InAlAsSb
InP
InGaAs/InGaAs QW
Race to 50%
Metamorphic buffers
Quantum wells
ELO and bonding
What's there at the finish line?
Teh>TL
CB
VB
ener
gy
absorption elasticscattering
time
thermalization radiativerec
Teh=TL
eh
hot-carrierSJMJ
Ultimately limited by junction number
slide 26 of 26