Solar cells for space:

III-V multi-junctions & Perovskites

Romain Cariou, Wilfried Favre & CEA-INES teams
Univ. Grenoble Alpes, CEA, LITEN, INES, Grenoble 38000, France
romain.cariou@cea.fr

05/01/2023
 OUTLINE

• III-V compounds

• III-V multi-junctions
• Basics & architecture
• Applications: CPV & Space

• Perovskite solar cells for space

• Perovskites versus III-V
• Materials & Devices
• Recent advances
 III-V COMPOUNDS

CONFIDENTIAL
 OUTLINE
III-V COMPOUNDS

• Combination of group III & V elements:

• III-V compounds
• Binary III-V: GaAs, InP, GaN, ...
• Ternary III-VxV1-x or IIIxIII1-vV : AlGaAs, GaInP, GaAsP, ...
• • Quaternary
III-V solar IIIcell
yIII1-yhistorical overview
VxV1-x : AlGaInP, AlGaInAs, GaInNAs, ...

• Architecture & working principle

• Applications
 III-V COMPOUNDS

• Combination of group III & V elements

• Most III-V monocrystals: ZnS structure

• Two interpenetrating face-centered cubic lattices
• Tetrahedral coordination, same as diamond cubic

Image from Wikipedia

 III-V COMPOUNDS

• Combination of group III & V elements

• Most III-V monocrystals: ZnS structure

• Wide range of crystal lattice & bandgap

Adapted from http://kmontgomery.net

 III-V COMPOUNDS

• III-V: a wide range of bandgaps

covering a large part of solar
spectrum

• Outstanding electrical/optical
properties

J.A. del Alamo, Nature 479, 317 (2011)

 III-V COMPOUNDS

• III-V: a wide range of bandgaps

covering a large part of solar
spectrum

• Outstanding electrical/optical
properties

• Less abondant & more

expensive than Si

Graph from Wikipedia, data from United Geological Survey – USGS, minerals.usgs.gov/minerals, 2013
 III-V MULTI-JUNCTION BASICS & ARCHITECTURE

CONFIDENTIAL
SINGLE JUNCTION GAAS SOLAR CELL
• Unlike Si, III-V direct semiconductor have
strong light absorption:
• 4µm of GaAs = 99% of 300-880nm AM0 photons

• Typical single junction GaAs cell:

emitter 1. n-GaAs absorber ~ 2 - 3µm
2. p-type emitter ~ 0.15µm
3. AlGaAs or AlInP window layer ~ 0.03µm

• High Eg (e.g. AlGaAs) or doping gradient (BSF)

layers on both side of pn junction
• Prevent minority carriers from reaching surfaces

• Photogenerated carriers travel a short distance

(drift/diffusion) to reach the pn junction
• III-V diffusion length ~ 1/100 Si diffusion length

F. Dimroth, Photovolt. Sol. Energy, 2017

 EFFICIENCY LIMIT OF SINGLE JUNCTION SOLAR CELLS

• Various losses limits

solar cell efficiency:
• E.g. Carnot cycle
(thermodynamics)

• Multi-junction solar
cells can reduce the
largest fraction of
losses:
• Thermalization
• Transmission

A. Polman et al.; Nature Materials, 11: 174-177, 2012

STANDARD III-V MULTI-JUNCTION SOLAR CELLS

• Standard III-V triple-junction:

• Stack of decreasing band gaps: GaInP/GaInAs/Ge
• Each subcell converts a given part of solar spectrum
STANDARD GAINP/GAINAS/GE MULTI-JUNCTION SOLAR CELLS

• Standard III-V triple-junction:

• Stack of decreasing band gaps: GaInP/GaInAs/Ge
• Each subcell converts a given part of solar spectrum

• Each subcell:
pn junction with
minority carrier
barrier layers

• Subcell thickness
typically in the µm
range (~ 150 nm
emitter)

• Tunnel junctions are

used to interconnect
subcells in series
 MUILTIPLE BANGAPS FOR BETTER SPECTRUM CONVERSION

• Stacking band gaps with decreasing energies

• Series connection: 𝑽𝒎𝒖𝒍𝒕𝒊𝒋𝒖𝒏𝒄𝒕𝒊𝒐𝒏 = σ𝒊 𝑽𝒔𝒖𝒃𝒄𝒆𝒍𝒍 & 𝑰𝒎𝒖𝒍𝒕𝒊−𝒋𝒖𝒏𝒄𝒕𝒊𝒐𝒏 = 𝑴𝒊𝒏(𝑰𝑺𝒖𝒃𝒄𝒆𝒍𝒍)

en.wikipedia.org/wiki/Multi-junction_solar_cell
 TUNNEL JUNCTIONS

• First TJ described by L. Esaki

in 1958:
• Made of Ge layers
• 3.3% degradation 50 years later !

• Zigzag curve du to band-to-band tunneling

between highly doped p- and n-type layers

L. Esaki, Physical Review, 109: 603-604, 1958 | L. Esaki, Nature, 464: 31-31, 2010.
TYPICAL MULTI-JUNCTION CHARACTERISATIONS

Subcell measured under current limitation

conditions (light biasing of other subcells)
BEST THEORETICAL BANDGAPS CHOICE

Multi-terminal connection Series connection

S.P. Bremner et al., Progress in Photovoltaics: Research and Applications, 16: 225-233, 2008
RECORD III-V CELLS - CPV DESIGN
SPACE COMMERCIAL CELLS – SPACE DESIGN
44 MWp Touwsrivier / South Africa

CONCENTRATOR PHOTOVOLTAICS

CONFIDENTIAL
 CPV WORKING PRINCIPLE
Fresnel lens concentrator

• Substitution of flat solar cell surface by optical Light

Collection
elements
FLATCON®-Module developed at ISE
• Sunlight collection area > solar cell active area
Light
Acollection Conversion
Cgeo =
• Logarithmic increase of cell efficiency with Aconversion
concentration
• Splitting of quasi-fermi levels  Voltage increase
Mirror based concentrator
• Reduced use of expensive semiconductors

• Various optical systems:

• LCPV: Cgeo < 100 suns (can be silicon cells)
• HCPV: Cgeo > 300 suns (III-V cells)

Courtesy of Fraunhofer ISE

 CPV FACTS & FIGURES

• Cumulative installations (grid-connected): > 370 MWp

Suncore,
• Several power plants with capacity ≥ 30 MWp: Golmud, China
• Golmud, China, built by Suncore: 60 (2012) and 80 MWp (2013)
• Touwsrivier, South Africa, built by Soitec: 44 MWp (2014)

• Demonstrated reliability, with field data > 10 years

• Certified efficiency record:

• Solar cell: 47.1 % (NREL 6 junctions)
• Mini-module efficiency of 43.4% (ISE, full glass lens + wafer-bonded 4J)
• Module efficiency of 38.9 % by Soitec

• Field performance ratios of 74-80%, for power plant > 100 kWp

CURRENT STATUS OF CONCENTRATOR PHOTOVOLTAIC (CPV) TECHNOLOGY, Faunhofer ISE / NREl, 2017c
HOW IS CPV DOING ?
RIP ?with Si industry
Strong competition

300-1000 suns
<100 suns Concentrator Photovoltaics

Each time the cumulative

production doubled,
Si price went down by 24%
for the last 36 years

CURRENT STATUS OF CONCENTRATOR PHOTOVOLTAIC (CPV) TECHNOLOGY, Faunhofer ISE / NREL, 2017c
ISS
NASA

III-V SPACE PHOTOVOLTAICS

CONFIDENTIAL
 AM0 SPACE SOLAR SPECTRUM

• No atmosphere filtering: • 1-sun normalized intensity of 1366 W/m2

 closer to 5800K black body radiation • Higher intensity & blue rich compared
to AM1.5G

Solar Power Generation in Extreme Space Environments, NASA facts FS-2017-07-286-LaRC

 VARIOUS APPLICATION FIELDS

Courtesy of Claus Zimmermann - AIRBUS, REA workshop, Nov. 2018 Brussels

Van Allen radiation Belt

Inner H+: 10-60x102 km

Outer e- : 13-60x103 km
 SPACE ENVIRONMENTAL LOAD

Courtesy of Claus Zimmermann - AIRBUS, REA workshop, Nov. 2018 Brussels

 PEROVSKITE SOLAR CELLS FOR SPACE

Courtesy of Matthieu Manceau & Carla Costa

CONFIDENTIAL
III-V & PEROVSKITES SOLAR CELLS

Space standard solar cells: Alternative cell candidate for space (?) :
III-V/Ge space multi-junction Single junction Perovskite Solar Cells
(PSC)

HTL : Hole Transfer Layer

80 µm
1µm
Industrial cell R&D cell

ETL : Electron Transfer

Layer

4
STANDARD III-V VS PSC

III-V (industrial) Perovskite

(R&D)
Efficiency (AM1.5g) 39.2% 25.7%
Active layer thickness ~15µm <1µm
Specific power ~3W/g ~30W/g

D. Cardwell et al., 2017 IEEE 44th Photovolt. Spec. Conf. PVSC

Y. Hu et al., ACS Energy Lett. 6, 2917 (2021).

7
STANDARD III-V VS PSC

III-V (industrial) Perovskite

(R&D)
Efficiency (AM1.5g) 39.2% 25.7%
Active layer thickness ~15µm <1µm
Specific power ~3W/g ~30W/g
Complex Simpler (?)
Manufacturing High temp. >600°C Soft chemistry
<100°C
Cost €€€ €

7
STANDARD III-V VS PSC

III-V (industrial) Perovskite

(R&D)
Efficiency (AM1.5g) 39.2% 25.7%
Active layer thickness ~15µm <1µm
Specific power ~3W/g ~30W/g
Complex Simpler (?)
Manufacturing High temp. >600°C Soft chemistry
<100°C
Cost €€€ € (?)
Temperature & vacuum
✔ ?
resistance
UV resistance ✔ ?
Radiation hardness (e-,
✔ ?
p+)

7
STANDARD III-V VS PSC

III-V (industrial) Perovskite

(R&D)
Efficiency (AM1.5g) 39.2% 25.7%
Active layer thickness ~15µm <1µm
Specific power ~3W/g ~30W/g
Complex Simpler (?)
Manufacturing High temp. >600°C Soft chemistry
Challenges & advances of Pk cells <100°C
€€€
resistance against
Cost harsch space € (?)
Temperature & vacuum
resistance environment
✔ ? ?

UV resistance ✔ ?
Radiation hardness (e-,
✔ ?
p+)

7
 PEROVSKITES MATERIALS & DEVICES

Courtesy of Matthieu Manceau & Carla Costa

CONFIDENTIAL
 The LMPO lab
What are Perovskites ?

Rather “Old” materials

First described by mineralogist Gustav Rose in 1839 …

… the exact same year E. Becquerel reported on photoelectric effect 

ABX3 crystalline structure

A : a cation

B : a metal

X : an anion

N. Mercier (JPH 2018)

Parent Material = CaTiO3
 The LMPO lab
What are Perovskites ? The case of Perovskites for PV

Perovskites for PV : Metal Halides

Lead or Tin-based B = Pb2+ ; Sn2+

Organic or Inorganic cation A = Cs+, R-NH3+

Halides Anions X = I-, Br- (Cl-)

N. Mercier (JPH 2018)

 So far, to a very very large extent, Perovskites for PV are Lead-based Hybrid organic / inorganic ones

 Seems a lot of options with “A” versatility …

… but rather few cations lead to stable “3D” materials

 Why PK for PV Applications ?

Why are they good PV absorbers ?

Excellent combination of optical & electronic properties

Why are they very popular ?

Processing Ease & Versatility

& Because they can marry Silicon 

 Why PK for PV Applications ?

Interesting “Intrinsic” Properties for PV Devices :

“Direct” Bandgap
Steep optical absorption rise Charge
Generation
Low exciton binding energy (< kBT)

Low effective masses

No deep trap states
Charge
Transport
Low non-radiative recombination rates
(Long diffusion length) https://www.lmpv.nl/sq/

 Potential for high PV Efficiency

 Why PK for PV Applications ?

Processing Interests for PV Devices :

Adjustable bandgap via versatile chemical composition (ABX3)

 X : Iodine / Bromine Mixtures

Eperon et al., Nat Rev Chem 2017, 1, 0095
 A : Cs / MA / FA Mixtures
 B : Sn / Pb Mixtures

Processing Versatility (Vacuum or Solvent-based)

Low / Moderate Temperature (usually < 150°C)

 A Decade Of Perovskite-based Solar Cells

2009 - 2021 : Publication & Efficiency Trends

Perovskite Solar Cells > 2000 papers per year since 2016

Certified Efficiency now at 25.7% in Single-Junctions

Highest Efficiency among all Thin-Films & Emerging PV

c-Si - MHPs tandems @ 29.5%

« Perovskite Solar Cells » in Web Of Sciences

5000

4000
Number of publications

3000
October 2021

2000

1000

0
09

10

11

12

13

14

15

16

17

18

19

20

21
20

20

20

20

20

20

20

20

20

20

20

20

20

Year
 ca 70 cm²
DEVICE AREA INCREASE – STATE OF THE ART

ca 12 cm²
Record Efficiency vs Device Active Area : PCE vs Area for PK & Other Technologies
~ 2% per order of magnitude in device surface from 0.1 to 10 cm² 28 mm²
~ 8% per order of magnitude in device surface beyond 10 cm²
Steeper decrease when compared to other PV technologies
No entry in the NREL Module Chart

Meredith et al., Nat. Comm. 2018, 9, 5261 Li et al., Nat. Rev. mater. 2018, 3, 18017
 PEROVSKITES RECENT ADVANCES

Courtesy of Matthieu Manceau & Carla Costa

CONFIDENTIAL
 SPECIFICITY OF SPACE ENVIRONMENT

Yongguang Tu et al. 2021 (DOI:

10.1002/adma.202006545)
2
48

PK SINGLE JUNCTION PERFORMANCE UNDER EXTREME TEMPERATURE

ELECTRONS IRRADIATIONS
BOL EOL
Pk studies : FAPbI3
Stack: ITO/TiO2/FAPbI3/Spiro-
OMeTAD/Ag Energy, Dose: 1MeV, 1x1016
e-/cm²

 SEM : no degradation of the morphology

 XRD : no degradation of the microstructure until 1x1015

e-/cm²

 Electricals properties no degradation up to 1x1015 e-

/cm², then loss of 10% of initial properties for 1x1016 e-
/cm²

1x1014e-/cm² = 18 years in
GEO
EOL (1x1012e- EOL (1x1014e- EOL (1x1015e- EOL (1x1016e- BOL/EOL :
BOL PCE (%) /cm²) /cm²) /cm²) /cm²) Beginning/End Of
Life
12,2% 13,3% 13,4% 12,4% 11,0%
DOI : 10.1109/PVSC.2017.8366410 (Jing-Shun Huand et al. 2018)
12
 PROTONS IRRADIATIONS

Pk studies CsFAMAPb(IBr)3
Stack: ITO/TiO2/Pk/P3BH/Au
Energy; Dose: 50keV ; 1x1015 p+/cm²

1x1014 p+/cm² = 10 years in

LEO
2x1015 p+/cm² = 18 years in
GEO

DOI : 10.1016/j.isci.2018.03.020 (Yu Miyazawa et al. 2018)

15
 PROTONS IRRADIATIONS

Pk studies CsFAMAPb(IBr)3 Pk studies : RbCsFAMAPb(Ibr) 3

Stack: ITO/TiO2/Pk/P3BH/Au Stack: ITO/TiO2/Pk/Spiro/Au
Energy; Dose: 50keV ; 1x1015 p+/cm² Energy; Dose: 50keV ; 1x1015 p+/cm²

Initial PCE = 21%

Jsc = 22,8mA/cm²
Voc = 1160mV
FF = 77%

1x1014 p+/cm² = 10 years in

LEO
2x1015 p+/cm² = 18 years in
GEO

 TR-PL : The carrier lifetime decreases at a fluence of 1 × 1015 p+/cm2

 Rate of non-radiative recombinations increases in the cell = appear of defects in the Pk
layer(Yu Miyazawa et al. 2018)
DOI : 10.1016/j.isci.2018.03.020 DOI : 10.1021/acs.jpclett.9b02665 (Shesaku Kanaya et al.

15
 Initial PCE = 18%
HIGH VACUUM Jsc = 22,4mA/cm²
Voc = 1050mV

Pk studies: CsFAMAPb(Ibr)3
Stack: ITO/SnO2/perovskite/Spiro
MeOTAD/Au
Tested pressures : 9x10-4 ; 2.2x10-4 ;
5x10-5 Pa
Characterisations : SEM, XRD, IV
Without encapsulation

Electricals properties : decrease of efficiency with diminution of Jsc and FF

 Migration of halogenated and organic atoms from Pk under vacuum and without
encapsulation

https://doi.org/10.1016/j.joule.2020.03.017 (Yan Jiang et al. 2020)

7
 Initial PCE = 18%
HIGH VACUUM Jsc = 22,4mA/cm²
Voc = 1050mV

Pk studies: CsFAMAPb(Ibr)3
Stack: ITO/SnO2/perovskite/Spiro
MeOTAD/Au
Tested pressures : 9x10-4 ; 2.2x10-4 ;
5x10-5 Pa
Characterisations : SEM, XRD, IV
Without encapsulation

Electricals properties : decrease of efficiency with diminution of Jsc and FF

 Migration of halogenated and organic atoms from Pk under vacuum and without
encapsulation

 SEM : phase segregation between Pk and

HTL from 2.2x10-4 Pa

https://doi.org/10.1016/j.joule.2020.03.017 (Yan Jiang et al. 2020)

7
 Initial PCE = 18%
HIGH VACUUM Jsc = 22,4mA/cm²
Voc = 1050mV

Pk studies: CsFAMAPb(Ibr)3
Stack: ITO/SnO2/perovskite/Spiro
MeOTAD/Au
Tested pressures : 9x10-4 ; 2.2x10-4 ;
5x10-5 Pa
Characterisations : SEM, XRD, IV
Without encapsulation

Electricals properties : decrease of efficiency with diminution of Jsc and FF

 Migration of halogenated and organic atoms from Pk under vacuum and without
encapsulation

 XRD : Decrease of the quantity of Pk phase (by an

increase of precursors and δ-phase)

https://doi.org/10.1016/j.joule.2020.03.017 (Yan Jiang et al. 2020)

7
TRACK RECORD OF PK CELLS TESTED IN STRATOSPHERIC / SPACE

Strastopheric Cells sent for Perovskites Perovskites Pk mini- Pk made in

balloon with tests on ISS cells on a cells on ISS modules on space ?
Pk cells rocket flight nanosat ?
2018 2019 2020 2021/0222 2022/2023 2030/2035
STRATOSPHERIC BALLOON – 32 KM

• Experimental proof: working devices in strato

conditions

• Identified challenges: extrême thermal cycling

with potential crystal phase transitions (at least
for MAPbI )
PK CELLS ON A ROCKET TO SPACE

• Confirmation of power
production in space (~ 240km)
with various Pk & organic cells

• Power production up to 14
mW/cm2

• Power production also in low

irradiance conditions (diffuse)
58

FLIGHT PROGRAM

• PCB designed for integration on side of a 3U nanosat

• QOSMOSYS, ZEUS-1 nanosat

• Thermal excursion in the range [-50; +60°C]

• Data monitoring for > 1 year @ 530 km

• Isc on shunted Pk array (right)
• Voc on Pk array (left)
• Temperature of the devices
• Sun direction & altitude

• Solar irradiance will be estimated with 0.5% precision

INTEREST MOON/MARS HUMAN BASE
• Typical target power ~ 1 MW
 only 12 kg of condensed Pk absorber

• With 3500$ lauching cost  0.1 $/W

(with 15% eff.)

• Space qualification standards (e.g.

irradiation sequences) may be adapated
to Pk cells

• Carefull selectrion of transparent

substrat, charge selecting layers

• Encapsulation in DC93500

• Substrate: FEP or transparent

polyimide for instance
Acknowledgements

THANKS FOR YOUR ATTENTION

Commissariat à l’énergie atomique et aux énergies alternatives INES Site

Alternative Energies and Atomic Energy Commission Institut National de l’Energie Solaire
17 av des martyrs 38000 GRENOBLE France National Solar Energy Institute
http://liten.cea.fr 50 avenue du lac Léman
73375 Le Bourget-du-Lac France
Établissement public à caractère industriel et commercial
+33 4 79 79 20 00
Public establishment with commercial and industrial character
RCS Paris B 775 685 019