This document discusses compound semiconductor solar cells, specifically focusing on gallium arsenide (GaAs) solar cells. It provides details on: 1) GaAs cells achieving higher efficiencies than silicon cells, up to 29% under concentrated light. 2) Reasons for the high efficiency of GaAs cells, including its ideal band gap of 1.42 eV and high absorption coefficient requiring only a few microns of thickness. 3) Types of GaAs solar cells including single homojunction, heterojunction, tandem, and cells grown on silicon/germanium substrates. The document outlines challenges with cost that must be addressed for terrestrial use of GaAs solar cells.

1. 1
Compound Semiconductor
Solar Cells
Nandita DasGupta
Dept. of Electrical Engg.
IIT Madras

2. 2
GaAs vs Si as a PV Material
• In 1989 experimental silicon cells reached efficiencies of
nearly 23% while GaAs cells reached efficiencies of
almost 26%
• Under concentrated light, GaAs cell efficiency = 29%
• Commercial GaAs solar cell efficiency = 20%
• Record shattering performance of the GM Sunraycer
using GaAs cell in the Solar Challenge car race across
Australia in 1987

3. 3
Reasons for high Efficiency
• GaAs band gap is 1.42 eV, nearly ideal for
single-junction solar cells.
• Gallium arsenide has high absorption
coefficient and requires a cell only a few
microns thick to absorb sunlight
(Crystalline silicon requires a cell
thickness >100 microns)
• Unlike silicon cells, GaAs cells are
relatively insensitive to heat. (Cell
temperatures are often quite high,
especially for concentrator applications)

4. 4
Other Positives
• Alloys made from gallium arsenide using
aluminum, phosphorus, antimony, or indium
have characteristics complementary to those of
gallium arsenide, allowing great flexibility in
high-efficiency cell design.
• Gallium arsenide is highly resistant to radiation
damage. This, along with its high efficiency,
makes GaAs very desirable for space
applications

5. 5
Types of GaAs Solar Cells
• Single homojunction
• Heterojunction
• Multiple heterojunction (Tandem)
• GaAs on Si/Ge
• Re-usable substrate

6. 6
Single Homojunction
• P-n junction created by MOCVD/MBE (Not by
Diffusion/Implantation!!)
• Single crystal made by LEC (Liquid
Encapsulated Czochralski) or by Gradient
Freeze technique
• Synthesis of GaAs is from elemental Ga and As
in a 2-zone furnace using a sealed boat
• As is at 600-620C while Ga is at 1240-1250C

7. 7
Problems in GaAs Homojunction
Solar cell
• GaAs has high Surface Recombination
velocity
• Unlike silicon, no native oxide is available
that can passivate the surface
• Passivation schemes add to process
complexity
• Remedy: Use Heterojunction cells

8. 8
III-V Heterojunction Solar Cell
• AlGaAs transparent Window on top
• Heavily doped AlGaAs reduces parasitic
resistance and increases Voc
• Lower surface recombination velocity – better
passivation
• GaP may also be used for window layer
N+AlGaAs~1019/cm3
NGaAs~1018/cm3
pGaAs~1018/cm3
p-GaAs

9. 9
Tandem Solar Cell – Basic Concept
• A stack of individual solar cells that absorb
different part of solar spectrum
• Radiation falls first on material with largest
Eg
• Photons not absorbed are transmitted to
second cell and so on
• Higher total conversion efficiency as large
portion of the solar spectrum is captured

10. 10
Structure of a Tandem Cell
• Current research efforts on III-V-PV cells with highest
efficiencies are focused on multi junction cells
• 4-, 5-, or even 6-junction solar cells are developed
• Aim : to attain a conversion efficiency ~ 50%.

11. 11
Salient features of SE4 Cell
• Focus towards the low band gap part of such a >
3 junction solar cell
• Problem is the realization of a subcell with an
absorber material in the band gap range of
around 1eV
• SE4 has successfully developed a low band gap
tandem solar cell with optimized band gaps of
0.75eV and 1.15eV
• InGaAs (Eg=0.75eV) and InGaAsP absorbers
connected by an InGaAs/GaAsSb Esaki-tunnel-
diode-like junction.

12. 12
SE4.Cells

13. 13
The Bottomline is Cost
• The prohibitive cost is the largest barrier to the
success of GaAs cells for terrestrial use
• Two main ways to reduce the cost
• The first approach is to fabricate GaAs cells on
cheaper substrates like silicon or germanium,
• The other is to grow GaAs cells on a removable
GaAs substrate that can be re-used to produce
other cells.

14. 14
GaAs on Si/Ge
• Growing good-quality GaAs epitaxial layers
requires a substrate with a crystal structure
matching that of gallium arsenide and with
similar thermal properties
• The slight mismatch in crystal structures
between Si or Ge and GaAs causes
imperfections in the growing GaAs crystal
• One way around this problem is to grow a
relatively thick buffer layer of gallium arsenide
between the silicon and the active GaAs

15. 15
GaAs on Si/Ge
• The highest efficiencies of GaAs-on-Si cells
obtained in 1989 exceeded 22%.
• Several laboratories are working to improve our
understanding of the GaAs-on-Si material
system
• In February 2008, IMEC announced that it has
achieved a conversion efficiency of 24.7% for a
single-junction GaAs solar cell on a Ge substrate
• The record cell measures 0.25 cm2 and shows a
Voc of 999mV, Jsc of 29.7 mA/cm2 and a fill factor
of 83.2%

16. 16
Next target
• Improving the efficiency of the single-
junction GaAs cell is another step towards
the development of a triple-junction solar
cell
• Focus is on stacked cells consisting of top
cells with III-V materials and bottom cells
made from Ge
• Target : a conversion efficiency of >35%

17. 17
Re-usable substrate
• Thin films of single-crystal GaAs grown on thick,
reusable substrates of the same material
• These thin films can then be peeled off and
incorporated in a PV device
• Substrate can then be used to grow several
more thin films
• Photovoltaic cells made from this process, with
efficiency ~ 24%, were used to make an
experimental flat-plate module at a record
efficiency of > 20%.

18. 18

19. 19
Epitaxial Lift-Off (ELO) Process

20. 20
Problems in ELO
• Microscopic cleavage crack in the lifted-off
films
• Cracks increase with area – scaling up is a
problem
• Cracks are due to strain in the polymer
support layer
• Thin photoresist with silicon support layers
resulted in crack-free lift-off

21. 21
Tandem Cell on Quartz
ELO process has been used here to transfer the film onto quartz

22. 22
Bonded Tandem Cell
• The 2-junction GaInP/ GaAs solar cell is bonded
with another silicon solar cell
• The GaAs active layers were detached from the
substrate and placed on quartz

23. 23
Nitride-based solar Cells
New kid on the block?
• As III-nitride Indium Gallium nitride (InGaN) has a
band gap energy in the range 0.7–3.4 eV depending
on its composition, this material is very promising for
a multijunction tandem solar cell
• According to the theoretical estimation, a conversion
efficiency over 50% is expected when a tandem solar
cell composed of ten sub-cells with a band gap energy
from 0.7 to 2.4 eV is realized
• Such an InGaN multijunction tandem solar cell will
be the next generation of the InGaP/InGaAs/Ge
tandem solar cell.

24. 24
Nitride Solar Cells
• Technology challenges
– High-quality film growth and Mg doping for InGaN
films with a variety of band gap energy
– Metal Contacts
• n-GaN/p-InGaN heterojunction and p-GaN/i-
InGaN/n-GaN heterostructure solar cells have
been reported.
• Schottky InGaN on GaN solar cell instead of
using a p-n junction
• An InGaN homojunction solar cell has not been
reported yet.

25. 25
InGaN-based Solar Cell designed at
Berkley
• Top cell made from InGaN and the
bottom cell bulk or thin film silicon
• Normally a tunnel junction has to be
provided between the sub-cells of a
tandem cell adding to process
complexity.
• However, for a particular InGaN
composition (In0.45Ga0.55N) Ec in
InGaN aligns with Ev in Si, allowing
the electrons in the n-InGaN to
automatically go to the valence band
of the p-Si.
• This provides a naturally occurring
low-resistance junction without any
need for heavy doping
• As Si substrate itself can be used to
take the back contact, the processing
can be simpler.

26. Thin film solar cells
• Low active material cost
• Fewer processing step – higher degree of
automation
• Independence from shortage in solar
grade silicon supply
• Easier monolithic integration of cells to
form modules
• Use of flexible/ already present surface

27. Evolution of CdS/CdTe Solar Cells
• Eg of CdTe =1.5 eV, almost perfectly matched
for electrical conversion of solar spectrum
• Simple heterojunction using n-CdS (Eg = 2.5
eV)as window layer and p-CdTe as absorbing
layer with metal contacts
• Viable thin-film cells with =10% by 1981
• Methods for thin film deposition –
electrodeposition, screen printing and close
spaced sublimation (CSS)
• Activation step: Annealing in CdCl2 at 400-500
to increase 

28. CdS/CdTe cells with TCO
• Scale-up to make
modules
• Use of TCO to meet
higher current
requirement in
modules
• Low resistive Tin
Oxide instead of
metal contacts

29. Realization of higher efficiency
• Buffer layer of high-
resistivity TCO
• Thinning of CdS to
allow more light
•  > 15% achieved in
1991
• Ting L Chu – key
technologist of CdTe
cells

30. Critical Issues
•   16.5% has been realized
• To improve it further
– Optimize/eliminate “Activation” step
– Increase doping in CdTe
– Improve contact to CdTe
•   20% is expected
• Limitation : Availability of Cd and Te

31. First Solar – the Big Player
• Grid parity of $1/watt – holy grail of PV
• First Solar’s manufacturing cost
=$1.14/watt and selling cost $2.45/watt
•  = 10.6% for 75W module (2008)
• Target – lower the manufacturing cost to
$0.7/watt and increase  to 12%
• In 5 years thin film PV could compete with
coal!!!

32. Secret of cost reduction
• Three key issues
• Very thin Active element – low material
cost
• Glass substrate – large area panels
• Time of manufacturing – 2.5 hours – 1/10th
of the time for Si solar cells
• Ease of p-n junction formation – p-CdTe
and n-CdS due to non-stoichiometry

33. Key Process Steps
• Elemental vapour deposition process in four
chambers
• Glass sheet coated with tin oxide heated to
600C in the first chamber
• CdS vapour formed by heating solid CdS to 700
C in the second chamber
• Submicron deposit of CdS film on glass
• CdTe deposition using a similar process in third
chamber
• Rapid cooling to 300 C using gust of nitrogen
to strengthen the material

34. CIS Solar Cells
• CIS is a more
complex system
• Needs tighter
process control
• Also offers more
flexibility
• Highest efficiency of
all thin film solar
cells  = 20%

35. CdTe vs CIS
• CdTe – simple process technology leads
to low manufacturing cost
• CIS – higher efficiency but complicated
processing
• Critical issue in CIS technology –
deposition of CdS buffer layer by wet
chemical process

36. Current Manufacturing Status

37. Solar powered Building
1306 CIS modules integrated into the façade of “Schapfenmuhle” in
Ulm, Germany with nominal installed power of 98kW

38. 38
References
• www.osti.gov/bridge/servlets/purl/10160785.../
10160785.pdf
• Wanlas et al, 5th International Conference on
InP and related materials, France, 1993
• Y.Yazawa et al, 26th PVSC, Annaheim, CA,
1997
• Y.Yablonovitch, Final technical report on ELO,
NREL.
• L. Hsu and W. Walukiewicz, Journal of Appl.
Phys., 104,024507, 2008