Alex Freundlich

1
Opportunities in Material Science and Device
Physics to Address Challenges for a Terawatt
Level Deployment of Solar Photovotaics

• Alex Freundlich
• Photovoltaics and Nanostructures laboratories
• Center for Advanced Materials, Physics and ECE
  department at University of Houston,

2
Current PV Landscape
3
Current PV Markets Shares and Cost
2006 revenue from PV installation was 13
billions PV equipment and materials sector
generated 3.7 billion in 2006
4
FAQ on PV

• Performance
• gt 20 commercially
• Stability
• gt 25 years
• energy payback lt 3 yrs
• Cost
• gt 4 /W (5/W)installed
• Growth of PV
• 30-40/yr, gt 1GW/yr since 2004
• 13 Billions (2006)
• DOE roadmap 30 GW installed US by 2030
• Storage for grid application gt20 of the energy
  mixd

5
Energy payback time
After Ansema et al. 2006
Installing 1 TW of Silicon PV 2- 3 TW
additional energy from existing fuel mix!
Need to develop technologies with low upfront
energetic cost
6
(No Transcript)
7
Working principle and Limiting factors of
semiconductor devices
100

hn
Converted (1 sun)
90
Non absorbed
80
Thermalized
70
60
efficiency ()
50
40
30
20
10
0

0
0.5
1
1.5
2
2.5
3
bandgap eV
8
Multijunction Cells
W/m2/nm)
Wavelength (nm)

• Junctions 1 2 3 6
• Efficiency (under 500X) 28 32 40 43
• Efficiency (1 sun AM0) 25 27 30 -

9
Champion Solar Cells
40
30
10
(No Transcript)
11
World Energy Supply
Electric Energy grows at faster pace than Total
Energy.
Source http//www.iea.org/Textbase/publications/i
ndex.asp
Alex Freundlich XVI
IMRC, Cancun, Oct 28 2007
12
World Energy Fuel Shares 2003
Total Energy 14.5 TWy
Electric Energy 2 TWy PV lt 0.1 of total
Source http//www.iea.org/Textbase/publications/i
ndex.asp
13
Energy Reserves Resources
RsvReserves ResResources
after N. Lewis http//nsl.caltech.edu/energy.html
Reserves( based on 1998 annual use )
Resources ( based on 1998 annual use )
oil 40-78 51-151 Gas
68-176 207-590 Coal 224 2160
14
Future Energy Fuel Shares
50
Business as usual CO2 gt750 ppm by 2050
CO2 lt 450 ppm implies gt 20 TWy Renewables
Renewable Nuclear Coal Oil Gas
Business as usual
40
30
Total Energy (TWy)
20
750 ppm
650 ppm
550 ppm
450 ppm
10
350 ppm
M. I. Hoffert et al., Nature 395, p. 881 (1998).
0
1990 2000 2010 2020 2030 2040 2050
2060 2070 2080 2090 2100
Year
Electricity?
Electricity
A considerable amount of carbon free energy
sources is required to maintain a sustainable
economic and population growth.
http//nsl.caltech.edu/energy.html
15
Future Electricity Needs
CO2 lt 450 ppm implies gt 20 TWy Renewables
Photovoltaics?
A. Feltrin and A. Freundlich, Proceedings of Rio
06 Global Climate change event, 2006
16
PV production

0.0002 TWy
17
Todays PV vs Future Electric Needs
18
Todays PV vs Future Electric Needs
Silicon?Thin Films? Conc. cells?

• Evaluate (near) long term challenges imposed by
  Material Reserves for TW level PV production

19
Assumptions

• Known Global Material Production and Planetary
  Reserves United States Geological Survey.
  (http//minerals.usgs.gov/minerals/pubs/commodity/
  )
• Best PV device efficiencies and designs
  
  M. A. Green et al., Progress
  in Photovoltaics 14, p. 45 (2006).
• Insolation 1460 kWh/m2 per year (4 hours AM1.5).
  
• No Recycling losses.
• 25 of reserves used in active PV devices.

20
World Solar Annual Radiation
Assuming 4 hours daily exposure to AM1.5 in
southern Europe 1460 kWh/m2 (Note 1 TWy 6
TWp)
21
Assumptions

• Known Global Material Production and Planetary
  Reserves United States Geological survey.
  (http//minerals.usgs.gov/minerals/pubs/commodity/
  )
• Best PV device structures and efficiencies
  
  e.g. M. A. Green et al.,
  Progress in Photovoltaics 14, p. 45 (2006).
• Insolation 1460 kWh/m2 per year (4 hours AM1.5).
• No Recycling losses. Yield 100!
• 25 of reserves used for active PV devices.

Alex Freundlich XVI
IMRC, Cancun, Oct 28 2007
22
Existing PV Material production
For Si, Al, Ti, Zn, S, P etc. are abundant.

Mostly obtained as refinement by-products of zinc
and copper ores.

• Extracted from United States Geological survey.
  (http//minerals.usgs.gov/minerals/pubs/commodity/
  )
• ()data incomplete

23
Short term limitations
Assuming utilization of 100 of 2004 material
production capacity
?16.5
?18.4
?9.5
?10.4
?39.0, ( 200x AM1.5)
24
Planetary Material Reserves
For Si, Al, Ti, S, P reserves are abundant.

• Extracted from United States Geological Survey.
  (http//minerals.usgs.gov/minerals/pubs/commodity/
  )
• ()data incomplete

25
Silicon solar cells
World reserve potential
.
?20.3
Large silicon supply, but silver (contact
material) supply is limiting factor.
A. Freundlich, A. Freundlich, Renewable Energy,
2007 in press
26
Silicon solar cells
World reserve potential
.
?24.7
?20.3
Large silicon reserves, but silver (contact
material) supply is a limiting factor.
27
Thin film solar cells
Limited by materials used for electrodes
Limited by materials used for solar cells
World reserve potential
SnO2
28
Multi-junction concentrator cells
World reserve potential
III-V MJC 200x
?30.2
?39.0
29
Overcoming existing Material Scarcity challenges

• Cryst.-Si and polyc.-Si PV (?20-25) may be
  limited by Ag reserves, use of alt. electrodes
  for N-type Si (silicides) for gt2-3 TWy PV
• CdTe and CIGS (?15-20) likely will remain
  sub-ltltTW technologies (lt 20-100GWy) limited by Te
  and In reserves. Development of alternative
  calcogenide solar cells.
  
• By circumventing the use of ITO electrodes, a-Si
  and dye-sens. (? 10) cells have a good
  potential for multi-TWy PV production (?
  limited?)
• For III-V concentrators (?gt35) Ge reserves are a
  barrier for TW PV use GaAs substrate TWy. New
  technologies III-V/Si or lift-off cells should
  enable access to multi TWy PV, In-free
  multi-junctions and use of alternative electrodes
  for gt5 TWy PV.

30
Technologies beyond bulk p-n junction
31
(No Transcript)
32
Many emerging concepts(beyond tandem cells)

• Engineering incident sun spectrum and
  transparency losses
• Nano emitters (dot concentrator)
• Surface plasmonics
• Up converters
• Down converter
• Intermediate band solar cells and impact
  ionization cells
• Efficiency projections (detail energy balance
  projections)
• Inserting 0,1 and 2D semiconductor structures in
  solar cells
• Polymer and hybrid cells
• Nanotubes/dot polymers
• Exciton dissociation

33
absorption via surface plasmon excitation
Enhanced optical absorption via surface plasmon
excitation in metal nanoparticles
Schaadt, Feng, and Yu Appl. Phys. Lett. 86,
063106 2005
34
Engineering incident sun spectrum and
transparency losses
h?(E1) h?(E2) h?(Eg)
h?(2Eg) 2 h?(Eg)
up-conversion of sub-bandgap
down-conversion of sunlight
35
Example of down conversion Exciton
multiplication with colloidal dots
V. Klimov , Appl. Phys Lett, 89, 123118 2006
36
Quest for Ultimate theoretical efficiency
Intermediate band solar cell (Luque and Marti,
PRL 97) with multiple intermediate levels
A.Brown and M. Green J. Appl. Phys. 94(9),
6150-6158 (2003)

• Are there any practical material systems for
  IBSC?
• impurity/defect levels (issues extremely low
  density of states)
• quantum confined systems (issues carrier
  extraction)

37
Intermediate band cell with incomplete
absorption and collection
E1
E2
? ? n2 sin2(?c), ?S 2x10-5, n1-23
38
Quantum confined p-i-n solar cells
photo-absorption and carrier collection
39
Multiple intermediate levels/confined states
available in low dimensional structures
40
Example of calculated wire absorption
GaInAs
GaInAsN
GaAsN
A. Feltrin A. Alemu and A. Freundlich, Phys. Rev.
B 2006
41
IBSC Detailed energy balance absorption
Quantum confined solar cell
E a

42
MQW solar cells relieves need for use of thick
base
A. Freundlich et al, US patent 5,851,310
43
Demonstrated QC(MQW) cell advantages
Improved IR and Near bandgap photo-conversion
Superior Radiation Resistance
R.Walters et al, WCPEC, Vienna 1998
A. Freundlich et al, WCPEC Hawaii 1994
44
Optimizing MQW device response
High below-GaAs bandgap relative QE for a
GaAsN/GaAs MQW device
45
Voc for 1-1.2 eV dilute nitride solar cells
46
Self-annihilation of electron-irradiation-induced
defects in InAsxP1-x /InP multiquantum well solar
cells
A. Khan, A. Freundlich, J. Gu, A. Gapud, M.
Imazumi,Y. Yamaguchi Appl. Phys. Lett. 90,
233111 2007
47
Efficiency of quantum-well triple junction
GaInP/GaAs/Ge
1 sun AM0 efficiency in excess of 35
A.Freundlich, US patent 6,147,296 Nov 2000
48
Conclusion

• TW level deployment of PV would on innovations in
  material science to develop higher efficiency
  devices based on materials that are, stable,
  abundant, non toxic and have a low upfront
  energetic cost (fast payback time) .
• These Goals are at reach, by developing new
  technologies, marshaling the excellent resources
  of organizations like the Academy of Medicine
  Engineering and Science of Texas and Research
  Societies, and developing the talents of a new
  generation of scientists and engineers.

49
Thanks You
50
Suns generosity complements the existing
electric grid
51
(No Transcript)
52
Quantum confined semiconductor solar cells

• 1984- Chaffin et al (US patent 87) use of quantum
  well in solar cells
• 1990- Barnham and Duggan 1st MQW quantum confined
  p-i-n solar cell proposal
• Since 1990 some hundreds of publications on
  quantum well and quantum dot p-i-n solar cells
  (only few on quantum wires)
• Most realizations rely on the introduction of
  periodical array of semiconductor of a narrow
  bandgap material within the intrinsic region of
  conventional p-i-n diode

53
Self Assembled Quantum dot solar cells
Carrier generation and collection issues,
efficiency decreases as number of well increases
Lopez et al, J. Solar Energy Engin. Vol 129, p322
54
Tradeoff in Quantum-Confined Systems

Increase number of wells to Maximize
Absorption and optimal carrier collection
(Minimise Recombination in wells)
55
Improving the device performance
56
Improving the device performance
1-Role of E-field across i-region 2-Role of
carrier escape sequence from wells
57
Self Assembled Quantum dot solar cells
Research Mostly driven in an attempt to develop
an intermediate band solar cell
Demonstration of 2 photon up-conversion
XD Wang et al, Appl.Phys. Lett 85, 1356 (2004)
Marti et al, PRL 247701, 2006
58
InAsP/InP MQW cells Fabricated by CBE
59
Effect of built-in built-in E -Field across
i-region
60
Excitonic absorption - Electric field (well
thickness trade-off)
Large electric field affects confinement and
absorption properties
C. Monier and A. Freundlich, J. Appl.Phys
85,1999, p2713
61
Qualitative evaluation of carrier collection
0
5
10
15
CURRENT (mA)
20
25
30
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
VOLTAGE (V)
62
Qualitative evaluation of the carrier collection
63

I Serdiukova et al, Appl.Phys.Lett 74, 1999,
64
E-field /i-region thickness

65
Criticality of E-Field across i-region (Effect on
Voc)

Need for high built in E-field, limit number of
wells High Voc afforded for Egt Ecritical
66
Criticality of E-Field across i-region (Effect on
Voc)

Best tradeoff (absorption vs. collection) E
Ecritical
BUT occasional Voc degradation!
67
Effect of carrier escape rates
68
Tunneling escape
Thermal escape
h?lum
Ebarr
H. Schneider and K. v. Klitzing, Phys. Rev. B38
1988, p6160
69
Extracting Escape times

• ? Light holes(Tunneling 10-15sec)
• ? Heavy holes ?? electrons (Thermal 10-12sec)
  (varies among samples)
• Role of Escape in Voc degradation ?

V0(Jsc)
70
Confined gap vs As composition and well thickness
As concentration and InAsxP1-x well thickness
used to obtain same EQW but different hole and
electron confinement depths
71
MQW region details of studied samples
The InP/InAsP 10 period MQW samples studied here
only differ by their MQW region geometry and As
composition.
The samples can be categorized in two sets
depending on the value of their Voc
72
Correlation between Voc and escape times
Higher Voc Faster electron escape
73
Photoluminescence vs temperature
LOWER Voc
HIGHER Voc
Ea extracted from PL potential depth of the
faster carrier
A. Alemu, et al, J. Appl. Phys.99, 2006, 084506
(2006)
74
Correlation PL activation vs ? calculations
Good correlation between calculated barrier depth
of the faster carrier and the measured PL
activation energy
75
Possible qualitative explanation

• Light holes always escape 1st gt negative charge
  imbalance
• If the carriers to follow are
• electrons gt decreased negative charge imbalance
  possibly leading to a lowering of hole barrier
  and ease of carrier extraction (higher Voc)
• heavy holes gt increased negative charge
  imbalance, even more difficult to extract
  electrons (lower Voc)

76
Summary E-Field

• Need for high E-field (EgtEc) to prevent Voc
  degradation
• In the vicinity of EEc , carriers escape
  sequence appears to play a major role in
  determining the performance of nanostructured
  p-i-n solar cells.

77
(No Transcript)
78
Boosting QE III-V (dilute nitride QWs)
C. Skierbiszewski et al., Appl. Phys. Lett., 76,
2409 (2000)

• Increased of me ? Increased absorption
  coefficient.
• Increased number of confined states.
• More localized e- wavefunction ?decrease barrier
  width.

79
Beyond identical periodic structures Thermally
assisted resonant tunneling
80
Resonant tunneling in MQW cells
81
Zero band offset for holes
Alemu and Freundlich, unpublished
Possibility of single carrier (electron) resonant
tunneling
82
MQW cell Modeling
P. Renaud et al WCPEC 1, Hawaii 1994, I Sediukova
PhD dissertation 1998 Calculation of Jsc (MQW)
following the approach of Leavitt and Bradshaw,
Phys Rev B 1990, Appl Phys. Lett.1990
83
Resonant tunneling in MQW cells
Alemu and Freundlich, presentation LL7.10, Symp
LL this meeting
84
Conclusion
Many attractive attributes of quantum confined
solar cells (QCSC) have already been identified
(defect tolerance and radiation hardness). In a
near term these devices may be used in boosting
multijunction solar cell efficiencies booster for
. Shallow well QCSC (lt300 meV) a promise for near
term realization (high efficiency tandem cell
boosters) Deep well devices have the strongest
potential for practical demonstration of many
emerging concepts (IBSC cells). In quantum
confined solar cells, beside the need for a
strong sunlight absorption, a rapid escape of
photo-generated carrier from well potentials
appears as being critical to attain decent
performance without a severe degradation of
device IV characteristics. Improvement of
fundamental body of knowledge and innovative
designs for better sunlight harvest and carrier
collection are necessary to fully realize the
potential of these devices.
85
(No Transcript)
86
Attributes of dilute nitrides
Reduction of the bandgap
Enhancement of effective masses
Higher Absorption Coefficient
87
Aknowledgements Current Graduate Students
Chandani Rajapaksha, Gokul Radhakrishnan, Wenkai
Zhu Former Students Leknath Bhusal, Stephane
Bouget, Aris Fotkatzikis, Fred Newman, Inna
Serdiuokova, Nasr Meldelci, Veronique Rossignol,
Current and Former Collaborators at UH
Andenet Alemu, Jose Antonio Coaquira, Andrea
Feltrin, Carlos Gramajo, Dr. Cedric Monier, Dr.
Marie Amandine Pinault, Philippe Renaud, Susan
Street, Mauro Vilela, Lance Williams, Current
and Former sponsor AFRL, AFOSR, NASA, NRL, State
of Texas
88
Absorption coefficient modeled following an
Elliot-like formalism
R.J. Elliot, Phys. Rev. 108, 1384 (1957)
89
Challenges a) Current matching between
sub-cells b) 1 eV material lattice matched to
GaAs/Ge

• Possible candidates III-V Dilute Nitrides
• Large Eg reduction upon addition of N
• in III-V semiconductors.
• 1.0 1.25 eV GaInAsN can be grown
• lattice matched on GaAs.

• Problems with bulk-like dilute nitrides
• Poor minority carrier properties.
• Doping issues.

90
Voc for 1-1.2 eV dilute nitride solar cells
91
I-V Characteristics and RTA

• Limited current output
• Current output similar to the one generally
  observed for bulk like alloys of similar bandgap.

Freundlich et al, J. Crys. Growth, V 301-302, pp
993-996 (2007).
92
I-V Characteristics and RTA

• Limited current output
• Current output similar to the one generally
  observed for bulk like alloys of similar bandgap.

• Open circuit voltage of about 0.6 V.
• Exceeds those reported for bulk like devices of
  similar bandgap (by 0.2-0.5 volts)

93
Voc for 1-1.2 eV dilute nitride solar cells
94
Voc for 1-1.2 eV dilute nitride solar cells
95
Summary and Conclusions

• GaAsN/GaAs MQW solar cells fabricated by CBE
• QE analysis validates the expected strong
  photo-conversion properties of III V dilute
  nitrides quantum-well based devises.
• As-grown devices show degraded current output
  similar to bulk InGaAsN of similar bandgap, Jsc
  improvemes upon RTA.
• Voc of about 0.6 volt exceeds those reported for
  bulk InGaAsN device of similar bandgap.

96
(No Transcript)
97
(No Transcript)
98
Thank you!
99
(No Transcript)
100
Figure 15 Open circuit voltage vs. bandgap for
conventional InGaAsN solar cells reported in the
literature and this work (semi-full circle).
101
GaInP/GaAs/Ge present status
Pros ? demonstrated efficiencies gt 30 AM0, ?
lattice matching to GaAs/Ge substrates ?
Industrially mature technology ? GaInP
radiation tolerance Cons ? GaAs bandgap is
too large for optimal operation with a top
cell at 1.85 eV
Efficiency and Radiation response Limited by GaAs
cell performance
102
(No Transcript)
103
Qualitative evaluation of the carrier collection
0
5
10
CURRENT (Am-2)
15
20
25
30
-0.4
-0.2
0.0
0.2
0.4
0.6
0.8
1.0
VOLTAGE (V)
104
Qualitative evaluation of carrier collection
External Quantum efficiency saturates at about
-0.9 volt, suggesting incomplete carrier
collection. Also improvement of a 1.37 eV range
valley in QE (GaAs N in barriers?).

• Thermal escape dominates tunneling escape. Slow
  escape

105
Qualitative evaluation of carrier collection
106
Qualitative evaluation of the carrier collection

• External Quantum efficiency saturates at about
  -0.9 volt, suggesting incomplete carrier
  collection

• Improvement of a 1.37 eV range deep in QE (GaAs
  N in barriers)

107
Non-intentional nitrogen doping with N plasma
source shutter closed.
108
Characterization techniques.

• In-situ
• RHEED
• Plasma optical spectroscopy

• Ex-situ structural and optical properties
• Low temperature PL
• HRXRD
• Photoreflectance

• Ex-situ Device characteristics
• Current voltage measurements (dark and
  illuminated at AM0, under 1 sun)
• Spectral response (with and without bias)

109
resonant tunneling in MQW cells
In conventional III-Vs simultaneous hole and
electron resonnant tunneling difficult to realize
110
Effect of residual H
Surface reconstruction affected by the presence
of excess hydrogen
111
Issues with QWs

• The photocurrent of multiquantum well solar
  cells increases with the number of wells.
• The number of wells that can be incorporated in
  a given cell (given i-region) is limited.
• The size of the i-region.
• The thickness of the well barriers. Small QW
  barriers lead to the formation of minibands that
  result in a reduction of the cell voltage.

112
Characterization techniques.

• In-situ
• RHEED
• Plasma optical spectroscopy

Particular attention minimize the effect of
residual hydrogen.
113
(No Transcript)
114
III. Conclusions

• Alternative approach to bulk III-V dilute
  nitride-based solar cells.
• GaAsN QW good candidates to replace InGaAs QWs in
  MQW-based solar cells.
• Preliminary set of GaAsN/GaAs MQW devices were
  fabricated by CBE.
• Limited as-grown device current output similar
  to bulk InGaAsN of similar bandgap. Current
  improved upon RTA.
• Voc of about 0.6 volt exceed those reported for
  InGaAsN device of similar bandgap (Eg1.1 eV)

115
Motivation Multi (4) junction tandem ?(AM0) gt40
µ
µ
,
116
General working principles
BC,LUMO
Efn
Eg
hn
Ef
Efp
BV,HOMO
Light Absorption (the drive out of equilibrium)
Quasi equilibrium (lifetime gtgt Thermalisat.
time)
Quasi equilibrium
Preferential collection at contacts (transit
time ltlt lifetime)

• Work available/absorbed photon qV DEf
• hn gt Eg gt DEf
• Name of the game Control of the path of the
  return to equilibrium
• Goal the radiative limit (DEf Eg)

117
Fabrication issues with Qdots
Issues with fabrication
XD Wang et al, Appl.Phys. Lett 85, 1356 (2004)
118
Limits of semiconductor devices
119
Multi-quantum well solar cell
120
High energy photons down conversion (single
junction device with efficiencygt42)
adapted after Klimov et al

• Recently down conversion of 1 high energy photon
  into 7 e-h pair demonstrated at Los Alamos (V.
  Klimov et al Nanoletters 2005)
• Could be combined with UHs superlattice cells
  designed for high efficiency IR PV (Bhusal,
  Freundlich,J.Appl Phys. 2007)

121
Absorption of dot wire well crystal
Qdot Qwire Qwell