Efficiency Solar Cells

Integrated, Flexible, High-eﬃciency Solar Cells:

Epitaxial Li-Oﬀ GaAs Solar Cells and Enabling Substrate Reuse

Jeramy D. Zimmermana, a a, b, c Kyusang Lee , and Stephen R. Forrest

a Department of Electrical Engineering and Computer Science, b Department of Physics, and c Department of Materials Science and Engineering, University of Michigan

Supported by Army Research Laboratory MAST Program Global Photonic Energy Corporaon Movaon

• Numerous uses for lightweight, high-efficiency, flexible photovoltaics: – Power at temporary off-grid locations. – Autonomous vehicles (e.g. UAVs). Rollable or foldable portable PV – Satellites. • No available technologies provide ARL MAST high-efficiency, high power density photovoltaics on lightweight flexible substrates. • III-V photovoltaics provide >28% power conversion efficiency (ηP). • Specific power densities of >6 W/g & 280 W/m2. NASA Helios • Substantial reduction in III-V PV cost structure: wafer reuse.

Power for UAVs 2 Power/Weight Tradeoﬀ

• Ideal cell: thin + high efficiency. GaAs InP • Active layers for GaAs cell is ~2 µm.

Trip. jun. • Lift-off uses back reflector à active layer Metam. thickness reduced by ~50%. Organic CIGS CdTe • GaAs has highest power:area and α-SI power:weight ratios Trip. Jun. achievable. C-Si Ge sub • Lift-off cells have higher power conversion efficiency than substrate cells.

3 Technologies & Goals

Epitaxial Li-Oﬀ (ELO)

• Light Weight & Flexible Thin-Film Solar Cell E. Yablonovitch et al, Appl. Phys. Le. 51, 2222 (1987). Epitaxial Protecon Layers

•Parent Wafer Reuse K. Lee et al, Appl. Phys. Le. 97, 101107 (2010)—UM. Cold Welding

• Simpliﬁed Transfer Process K. T. Shiu et al, Appl. Phys. Le. 95, 223503 (2009)—UM.

Mulple Growths on Single Wafer Low-Cost Solar-to-Electrical Energy Conversion

4 Non-Destrucve Wafer Reuse for Thin-Film PV Cells

Cold Welding

Growth Metallizaon Epitaxial Li 0ﬀ

Surface Cleaning GSMBE

Thin-Film Fabricaon

5 MBE Growth of Epi-layers

Molecular beam epitaxy used to grow:

p+-GaAs contact, 0.3 µm p-InGaP BSF, 75 nm

p-GaAs base, 3 µm Solar cell active layers n-GaAs emitter, 0.15 µm n-InGaAlP window, 25 nm n+-GaAs contact, 0.2 µm InGaP, 0.1 µm GaAs, 0.1 µm AlAs, 10 nm Sacrificial and release layers GaAs, 0.1-0.5 µm InGaP, 0.1 µm GaAs buffer, 0.2 µm Substrate + buffer layer

GaAs substrate, 350 µm

6 Cold Welding

• Au deposited on wafer and plastic handle. • Au surfaces bonded by applying pressure. – Metallic bonds formed at room temperature. – Adhesive-free bonding technology. • Simple transfer process.

Stamping head

wafer/epi layer/Au High Pressure

Kapton/Ir/Au

Stamping head

7 Transferring III-V PV Cells to Plasc

Epitaxial lift-off

GaAs Wafer

AlAs release layer GaAs Solar Cell Layer

Mylar/Kapton

GaAs Wafer • Wafer cold welded to Kapton®.

HF AlAs HF • Lift off performed with HF. 7 GaAs Solar Cell Layer – Etch selectivity is ~10 :1 (AlAs:GaAs). – Epitaxial lift-off enables wafer reuse. Mylar/Kapton – Efforts are focused on improving quality and speed of lift-off process. • Fabrication performed on Kapton® GaAs Solar Cell Layer substrate. Mylar/Kapton • Flexible without degradation to radius < 1 cm.

8 Acceleraon of ELO Using Strained Handles

Strain Control by Spuered Ir

Compressive Strain

7 mTorr Spuering Pressure

10 nm Ir 20 nm Ir

50 μm Kapton

8.5 mTorr Spuering Pressure Tensile Strain 7 nm Ir 28 nm Ir

Strain State Etch me (Ir thickness)

Substrate Neutral ~10 days 10nm AlAs Kapton/epi layer Compressive (21 nm) ~24 hrs Tensile (7 nm) <8 hrs 9 Inﬂuence of ELO on GaAs Wafer

Scanning Electron Microscopy Atomic Force Microscopy

1 μm RMS = 8 nm

O O

O Ga 3D Laser Microscopy 32 ��

10 Inﬂuence of ELO on GaAs Wafer

Scanning Electron Microscopy X-ray Photoelectron Spectroscopy

3d5/2 3d3/2

1 µm

As As2O5 Energy Dispersive Spectroscopy Original Wafer O O Ga As Atomic Concentraon Near Surface Original Wafer

As 49.5% Ga 48.3% ELO Processed Wafer O Ga As ELO Processed Wafer 3D Laser Microscopy As 41.2% Ga 33.2% O 25.6%

11 Protecon Layers

Protection Layer Structure • Protection layers

… – Inserted between AlAs and In0.49Ga0.51P (0.1 µm) wafer or active layers. AlAs (10 nm) – Prevent HF from contacting the In0.49Ga0.51P (0.1 µm) wafer. GaAs (0.1 µm ) – InGaP removed with H PO :HCl or HCl:H O. In0.49Ga0.51P (0.1 µm) 3 4 2 GaAs wafer – GaAs removed with H3PO4:H2O2: H2O. Atomic Force Microscopy – Reduces RMS roughness to that of original wafer. New wafer 10 nm • As2O5 particles difficult to remove

RMS = 0.20 nm – Tri-layer necessary. Non-protected wafer – Over-etched to undercut particulates. RMS = 0.97 nm

Protected wafer K. Lee et al, APL 97, 101107 (2010). K. Lee et al, JAP 111, 033527 (2012). RMS = 0.21 nm

5 �m 12 Thermal Decomposion

Fresh wafer ELO processed

Protection layer RTA treated removed

• RTA in N2 at 600 °C for 1 min. • GaAs/InGaP bilayer protection removed with wet etches. • Surface returned to like-new condition. 13 Plasma Pre-clean

After ELO After plasma cleaning 3d laser microscope

AFM

RMS = 8 nm RMS = 0.2 nm

• SF6+Ar ICP plasma for 1 min. • GaAs/InGaP bilayer protection removed with wet etches. • Surface returned to like-new condition. 14 Surface Chemistry Comparison

X-ray Photoelectron Spectroscopy

Energy Dispersive Spectroscopy 50.0 48.3 49.5 49.3 41.2 25.6 33.2

2.2 0.7 O Ga As O Ga As O Ga As Surface Cleaned Wafer Original Wafer ELO Processed Wafer Aer ELO 15 Epitaxial Protecon Layer & Surface Cleaning

Original wafer Atomic Force Microscopy RHEED 10 nm AlAs 100 nm GaAs RMS = 0.2nm 100 nm InGaP Growth GaAs substrate Metallizaon & Cold Welding Epitaxial li oﬀ

RMS = 8 nm Film - Substrate

Surface Pre-cleaning Thin Thermal or Plasma Cleaning Surface Cleaning Protecon layer removal Regrowth

RMS = 0.2 nm Solar Cell Fabricaon 16 Growth Quality Comparison

Cross-Seconal Hall Eﬀect Measurement

Transmission Electron Microscopy Original Growth Grown on Fresh Wafer Doping concentraon 2 1.67 x 1018 /cm3 1950 cm /Vs (110) Regrowth 5 nm Doping concentraon 2 2 nm 1.78 x 1018 /cm3 1930 cm /Vs

1st Regrowth Photoluminescence Measurement

(110)

5 nm 2 nm

2nd Regrowth

(110)

5 nm 2 nm

17 Device Performance Comparison

Current Density-Voltage Curve n+ GaAs Ohmic AlInP window n-GaAs p-GaAs InGaP BSF

GaAs Substrate

External Quantum Eﬃciency Device Eﬃciency Parameters

JSC [mA VOC FF PCE -2 cm ] [V] [%] [%] Fresh 27.7±0.5 1.00 83.4 23.1±0.6 Wafer Reused 28.0±0.4 1.00 81.6 22.8±0.5 wafer

18 GaAs Thin-Film Solar Cells

Current Density-Voltage Curve

n+ GaAs Ohmic AlInP window n-GaAs p-GaAs InGaP BSF GaAs Ohmic Gold Kapton®

External Quantum Eﬃciency

ELO Processed GaAs Thin-Film Solar Cells Device Eﬃciency Parameters

JSC VOC FF PCE 23.1 mA/cm2 0.92 V 75.6% 16.1%

19 Cost Esmate: Substrate-Based III-V PV

1 m2 GaAs PV Cells: • Substrates (56 x 6”) $17,000 • Source material (4 μm GaAs growth) $80

• Liquid nitrogen for cryo panels $400

• Electricity (~120 kW-hr) $15

• Wafer processing $75 - Convenonal Au contacts $40 - AR coang (ZnS/MgF2)150nm $10 - Lithographic paerning $25 • Total $17,570 • Cost/Wp @ 23% $76

(assumpons: Standard MBE design, no large reducons in price for bulk quanes, and labor and capitol cost not included.)

20 Cost Esmate for Flexible III-V PV 1 m2 GaAs PV Cells: • Substrates (56 x 6”): 50 uses $340 • Source material (2 μm GaAs growth) $40 • Electricity (~120 kW-hr) $15 • Wafer processing $35 - Metal/Kapton® foil $10 - Ag/Cu contacts $10 - AR coang (ZnS/MgF2)150nm $5 - HF etchant (HF:AlAs >106:1), 1 L 10% $5 - Non-Lithographic paerning $5 • Total $430 • Cost/Wp @ 30% $1.43

(assumpon: No LN2—requires MBE redesign.)

21 Concentraon

• Grid parity possible by using concentrators. Diffuse light • 4x concentraon: Direct light Directlight ~$0.70/Wp. • 10x concentraon: ~$0.45/Wp. • Requires solar

MJ cell tracking. PV Cell CompoundReflector Parabolic Reflector

22 Summary

• Solar Cell Fabricaon – Via cold-welding & epitaxial li-oﬀ. – Expedited li-oﬀ process using strained handle. • Reuse of GaAs Wafers – Lace matched epitaxial protecon layer. – Thermal & plasma surface cleaning. – GaAs regrowth aer protecon layer removal.

– Idencal solar cell performance.

K. Lee et al, J. Appl. Phys. 111, 033527 (2012). [email protected] [email protected] 23