Photovoltaics: Present Status and Future Prospects

Vasilis Fthenakis

Center for Life Cycle Analysis, Columbia University, New York, NY 10027

Summary

This report provides an overview of recent developments in photovoltaics

(PV) efficiencies, production costs, and deployment and discusses how these
advances created the potential for photovoltaics to become a major
contributor in energy markets throughout the world. Specifically the report
deals with the following topics: a) PV market evolution; b) photovoltaic
materials and devices; c) PV manufacturing present status and prospects; d)
efficiency advancements for current and evolving technologies; e)
sustainability and life cycle assessments, and f) the intermittency challenge.
The technologies described in this report are crystalline silicon (c-Si),
amorphous silicon (a-Si), cadmium telluride (CdTe), copper indium gallium
diselenide (CIGS), copper zinc tin sulfide (CZTS), dye-sensitized solar cells,
and organic photovoltaics (OPV).

1
Table of Contents

1. Introduction
2. Photovoltaic Materials and Devices
3. PV Materials, Cells and Modules: Manufacturing Status
a. Solar Grade Silicon Production
i. Silane and chlorosilane production
ii. Siemens process
iii. Fluidized bed reactor processes
iv. Vapor to liquid deposition
v. Upgraded metallurgical silicon
b. Alternative Si-cell technologies
i. Kerfless wafers
ii. Thin-film silicon
c. Amorphous and Nanocrystalline Silicon (a-Si)
d. Cadmium Telluride (CdTe)
e. Copper Indium Gallium Selenide (CIGS)
i. Co-evaporation
ii. Reactive Sputtering/selenization
iii. Solution-based processes
f. Copper Zinc Tin Sulfide (CZTS)
g. Dye-sensitized solar cells
h. Organic photovoltaics (OPV)\
4. Future Outlook
5. Sustainability-Life Cycle Analysis
6. The Intermittency Challenge
7. Conclusions

2
 1. Introduction
Global power consumption currently stands at approximately 15 TW (1 TW =
1012 W), the vast majority of which is generated by the combustion of fossil
fuels. The associated release of CO2 from these anthropogenic sources has
dramatically altered the composition of the atmosphere and may
detrimentally impact global temperature, sea levels, and weather patterns.
Furthermore, the realization that fossil fuels are not inexhaustible and that
enhancing recovery of coal, oil and natural gas presents additional risks,
drive energy policy scenarios that are based on renewable forms of energy.
In most of the proposed scenarios, solar energy is the primary constituent as
it is the major energy source over large regions. Solar energy in addition to
maintaining life on the planet, is used on demand in three basic forms based
on anthropogenic processes: Electricity from the direct conversion of solar
energy using semiconductor materials (solar photovoltaics, PV); electricity
from captured thermal energy (concentrated solar power, CSP), and heat
from the sun (solar thermal). The later, in the form of water solar collectors is
cost-competitive with electric heaters in high solar irradiation regions all over
the world. Among the solar electric technologies, traditionally CSP were
considered as the most cost effective for central power generation in high
insolation
areas, while
PV were
considered
primarily for
small
dispersed
(mainly
residential)
applications. However, since about 2005, a large utility market has been
created for PV and the costs of large PV solar farms has been reduced to
below that of comparable size CSP plants.

3
Fig. 1. Recent growth of the PV Global Market; annual sales in GW
modules(source: http://solarcellcentral.com/markets_page.html)

Overall, the market for photovoltaics has been growing at an average rate of
45% per year over the past decade. The five year growth rate from 2007 to
2011 was approximately 70% per year, but this was slowed down to 15% in
2012 as the incentives in several European countries were reduced (Fig. 1).
While the growth numbers are very impressive, the 27 giga-watts installed in
2011 is just a fraction of one percent of the total amount of electricity that
was being generated by all sources, indicating that there is plenty of room
for further growth. 1,2

A synopsis of the market historical evolution is pertinent. Japan was the first
country to enact economic incentives for PV and led the global market in the
90s till the early 2000s in both production and installations. Then in 2003,
Germany enacted the Renewable Energy Standard (RPS) and created an
uncapped PV market that became a magnet for PV producers from all over
the world. The large market created in Germany is credited with enabling
the economics of scale in PV production and the manufacturing efficiencies
that reduced the cost of producing PV modules by a factor of three. In the
period 2005-2009 cadmium telluride (CdTe) thin film technology emerged
rapidly as the low-cost PV production leader. A single company First Solar
become the worlds higher volume PV manufacturer in 2009 capturing 13%
of the global supply, from a negligible position in 2003. However, the rapid
growth of CdTe PV was eclipsed by an increase of crystalline PV production
from China, which in 2010 became the worlds leading manufacturer. This
growth was spurred by strategic government investment, access to cheap
capital, low Chinese labor costs, and probable distortion of economics by
unconventional accounting and a low exchange rate. Protests by the

4
European and US manufacturers, resulted to the US Commerce Chamber
issuing a preliminary determination for a 31% tariff to be imposed on
Chinese imports into the US market to counter-balance the subsidies and
cost distortions of the Chinese imports, whereas proposed legislation in the
US Senate calls for a US-made requirement to projects eligible for the
business Investment Tax Credit (ITC) eligibility. Thus, projects using a large
fraction (e.g., >30%) of PV modules from China would be ineligible for the
30% ITC in the United States. Although the market growth and distribution of
market share is fluid, we believe that the market is poised for further
expansion in the next decade as PV has reached cost parity with peak rates
in south California, and South Italy and is expected to reach cost parity
throughout the U.S. and many other regions by 2030 (Fig. 2).

Fig. 2 Wholesale Levelized Cost of Electricity in the U.S.: Central (utility)

status and projections (Source: Lushetsky J., Solar Technologies Program, US-
DOE, 25th EUPV, Valencia, Spain, Sept. 2010.

This report provides an overview of PV technologies and innovations which

have the potential to significantly contribute to an expansion of the PV

5
markets in developed and developing countries alike. The descriptions of the
status and prospects of PV manufacturing are mainly based on a National
Science Foundation workshop held at the Colorado School of Mines.3

2. Photovoltaic Materials and Devices

Photovoltaics use semiconductor materials to generate electricity from solar

energy. A semiconductor is a solid, mostly crystalline, material such as
silicon, selenium or germanium. Semiconductors have electrical
conductivities greater than insulators but lower than metals which are good
conductors. At low temperatures they are insulators but at high
temperatures and/or when excited by sunlight, they conduct electrons. The
most commonly used semiconductor element is silicon belonging to column
IV of the periodic table which has four valence electrons and bond to four
other atoms of the same element to have a fully satisfied valence of eight
electrons. A number of compounds can be semiconductors having a valence
of eight electrons. Galium arsenide (GaAs), and cadmium telluride (CdTe),
are examples of IIIV, and IIVI compounds, respectively.
Semiconductors do not have free electrons to exhibit electrical conduction
at low temperatures; however at room temperature, they show a modest
electrical conductivity, which increases as temperature increases. The
conductivity behavior of semiconductors may be explained with a simple
model illustrating silicons atomic structure. Each silicon atom has four
nearest neighbors. It has 14 electrons in its atomic structure and four of
them are in the outermost orbit which is weekly-bound to the nucleus. These
four electrons are called valence electrons which form covalent bonds with
the four nearest-neighbor atoms (i.e., each valence electron of a silicon

6
atom is shared by one of its four nearest neighbors).

Fig.3 (a) Three-dimensional Silicon

Lattice
Fig 3(b) Silicon lattice with one Si atom
displaced by a pentavalent atom (e.g.,
As, P);
(c) Silicon Lattice with one Si atom displaced by a trivalent atom (e.g.,
Ga, B)
Source: Markvart T., Solar Electricity, UN Press4

The valence electrons thus help to bind one atom to the nearest neighbor,
resulting in these electrons being tightly bound to the nucleus. In chemically
pure or intrinsic silicon which contains no foreign atoms, this ideal situation
exists at 0 degrees Kelvin, and the crystal behaves as an electrical insulator
since no free electrons are available. When light is absorbed by the material,
the bond gets broken and the valence electron becomes a free electron
leaving behind a hole. The hole is an incomplete covalent bond. It takes
energy equal to band energy Eg, a characteristic parameter of a
semiconductor, to separate an electron from the bonding and make it a free

7
 electron, also referred to as conduction
electron (Fig. 4).

Figure 4. Schematic of energy gap in

semiconductor materials

For silicon, Eg is equal to 1.1 eV at room temperature. When a hole exists in

a bond, it is relatively easy for a valence electron in a neighboring atom to
leave its covalent bond to fill the hole, leaving a hole in its original position.
The hole therefore effectively moves opposite to that of the electron. The
movement of holes, behaving like positive charges, constitute an electrical
current. Small amounts of dopants, few parts per million, introduced into the
semiconductor structure can significantly increase the number of electrons
available for breaking away from their atoms.
When, a silicon atom is replaced by a trivalent atom like boron (B) or gallium
(Ga) which has three valence electrons, then bonding with only three nearest
neighbors is complete. This atom can capture an electron from a nearby
bond, establishing the fourth bond. The captured electron is not available for
conduction, but a hole created in the original position of the electron enables
conduction of electrons. This is called p-type doping or acceptor type
doping. The valence electrons are confined to the valence band; but when
light energy equal to or greater than Eg is incident, they move into the
conduction band, creating an equal number of holes in the valence band.
In Fig. 4, Ec represents the lowest energy that conduction electrons may
have and Ev represents the lowest energy that holes may have. Eg is the
separation of the conduction band edge from the valence band edge.

If a semiconductor of band gap energy Eg absorbs light energy (E=h),

greater than Eg, each absorbed photon raises one electron from valence

8
band to conduction band, creating one electron and one hole. Here h is
Plancks constant and is the frequency of the light wave. The carrier
generation rate per unit area of the semiconducting surface is therefore a
function of Eg; however, the reflection of incident light at the surface and
incomplete collection of charge carriers affect the generation rate.

The p-n junction

9
P-n junctions are formed by joining n-type and p-type semiconductor
materials, as shown below. Since the n-type region has excess electrons and
the p-type has excess holes, electrons diffuse from the n-type side to the p-
type side. Similarly, holes flow by diffusion from the p-type side to the n-type
side. If the electrons and holes were not charged, this diffusion process
would continue until the concentration of electrons and holes on the two
sides were the same, as happens if two gasses come into contact with each
other. However, in a p-n junction, when the electrons and holes move to the
other side of the junction, they
leave behind exposed charges on
dopant atom sites, which are
fixed in the crystal lattice and are
unable to move. On the n-type
side, positive ion cores are
exposed. On the p-type side,
negative ion cores are exposed. In summary, electrons move from the n-type
to the p-type and holes move from p-type to n-type and this movement
creates is a region where the negative electrons have become positive and
positive holes have become negative. This changing of charges creates a
barrier to further movement. An electric field forms between the positive
ion cores in the n-type material and negative ion cores in the p-type
material; this electric field pulls electrons and holes in opposite directions.
This region is called the "depletion region" since the electric field quickly
sweeps free carriers out, hence the region is depleted of free carriers (Fig. 5).
The width, W, of the depletion zone in p-n silicon devices is in the order of
100 nm. Due to , a "built in" potential is formed at the junction.

Principle of Operation of a P-N Junction Solar Cell

Solar cell is an electronic device which converts solar energy directly into
electrical energy through the photovoltaic effect. It is a typical
semiconductor p-n junction device. The principle of operation of a p-n-

10
junction solar cell are illustrated in Figures 6 and 7. When the light falls on
the device, the light photons of certain wavelengths are absorbed by the
semiconducting material (excites electrons which move from the valence to
the conduction band) and electrical charge carriers, electrons and holes, are
generated. These carriers diffuse to the junction where a strong electric field
exists, the electrons and holes are separated by this field and produce an
electric current in the external circuit, this current, called photo-current,
depends on the incident photon intensity and the nature of semiconductors
that constitute the junction device. Since the extra energy of photons is lost
into heat, only part of solar spectrum can be converted (Fig. 6).

Fig. 6 and 7. Photovoltaic Effect in a Solar Cell

Under illumination, a photo-generated current, Iph, flows as a reverse diode

current, which is linearly dependent on the intensity of incident light. Fig. 8
shows the current-voltage (I-V) curves, without and with incident light.

11
Fig. 8. I-V characteristics of a solar cell with external voltage and with
illumination

The resulting current is: I = -Iph + Io [ exp (qV/kT) -1 ] and is represented by

the lower curve in Fig. 8. Typically the I-V curves of solar cells are presented
only by the fourth quadrant (flipped up to show positive current) showing the
effect of both illumination and the diode, as shown in Fig. 9.

Fig. 9. Solar Cell I-V Curve

Two important parameters are the short-circuit current (Isc), and the open-
circuit voltage (Voc). If the terminals of the junction are shorted, no current

12
flows in the junction. When the junction is illuminated, a current called short-
circuit current flows in the junction from p- to n- side or otherwise depending
on which side of the junction, the light is incident. This photocurrent is due to
three contributions: holes generated in the n-region, electrons in the p-region
and electrons and holes in the depletion region before they recombine.

Isc is a function of cell design and is proportional to photon flux. If the load is
an open circuit, the corresponding Vout is called open-circuit voltage, Voc. Isc
and Voc are related to the output power at the optimum operating point, Pmax
as

Pmax = Voc x Isc x FF

where FF, the fill factor.

FF= Vm x Im/Voc x Isc

Vm and Im are the voltage and current at the optimum operating point.
Typically Vm is 75% to 90% of Voc; and Im is 85 % to 99 % of Isc for silicon.

The efficiency of the solar cell is typically measured as:

= Voc x Isc x FF/ Incident Solar Power

The efficiency will be high is Isc and Voc are high and so also FF. it means the
dark current must be low. A sharp corner in the I-V characteristic is an
indicator of high FF.

In practice, the conversion efficiency of a solar cell is measured under

Standard Test Conditions (STC), which include (i) solar irradiance intensity of
1000 watts/m2, (ii) AM 1.5 solar reference spectrum, and (iii) cell
temperature during measurement, of 250C.

13
The Voc is directly proportional to the band gap energy whereas the
generation rate of the current decreases with increasing energy gap, thus
that the cell efficiency reaches a maximum at certain range of values of Eg
that corresponds to a maximum voltage-current product. More specifically, a
change in energy gap influences the energy conversion of a solar cell in two
ways. First, a larger Eg reduces the reverse saturation current and increases
Voc which in turn tends to increase the efficiency. Secondly, a larger Eg means
that fewer photons can be absorbed because only those photons greater
than or equal to Eg are absorbed, which in turn decreases the efficiency. The
net result of these two opposing effects is shown in Fig.10. For single-junction
devices, the calculated efficiency has a maximum of about 30% at Eg equal
to about 1.5 eV and falls off on either side of this value.

The semiconducting compounds, cadmium telluride(CdTe), gallium arsenide

(GaAs) and indium phosphide(InP) have Eg around this value. In practice,
several semiconductors with energy band gaps in the range 1.0 to 1.7 eV are
used for maximum efficiency
devices/solar cells. These
also include monocrystalline
and multicrystalline silicon
(Eg = 1.12 eV), copper
indium diselenide(Eg = 1.05
eV), amorphous silicon (Eg =
1.7 eV), cadmium telluride
(Eg = 1.45 eV), gallium
arsenide (Eg = 1.25 eV).

Fig. 10. Estimated Efficiency versus Energy Gap of the Cell Material

14
Figure 11 displays the champion overall power conversion efficiencies for
laboratory solar cells.13, 14, 15
It is shown that the PV technologies are at
different points in their respective learning curves when compared to the
ShockleyQueisser (S-Q) limit of ~31% solar energy conversion efficiency for
single junction devices.17 As shown in Figure 10, the band gap energy
dependence of the ShockleyQueisser limit is flat around the maximum, thus
the theoretical efficiency limits for these technologies are only marginally
different and are not a deciding factor in their competitiveness with each
other. Nevertheless, crystalline silicon, which is the most mature technology,
having been benefitted by investments in IC applications, has reached record
efficiencies of ~24%, whereas the less researched CdTe and CIGS thin-film
technologies have reached record cell efficiencies of 17.3% and 20.4%
correspondingly. The module efficiencies lag those of the laboratory cells;
the highest commercial production module efficiencies are 20.1%, 12.5% and
11.5% correspondingly for mono-crystalline Si, CdTe and CIGS modules.

15
Fig. 11. Record efficiencies for different types of solar cells in the laboratory
(Source: NREL)

In most cases, the record cells have seen little or no improvement over the
past decade while economies of scale and advances in manufacturing
science and technology have fueled the growth of the PV market through
cost reduction and module performance improvements16.

Solar Cell Design Principles

Solar cell design involves specifying the parameters of a solar cell structure
to maximize efficiency. The theoretical efficiency for photovoltaic conversion
is in excess of 86.8%.6 However, the 86.8% figure uses detailed balance
calculations and does not describe device implementation. For silicon solar
cells, a more realistic efficiency under one sun operation is about 29%. 7 The
maximum efficiency measured for a silicon solar cell is currently 24.7%
under AM1.5G. The difference between the high theoretical efficiencies and
the efficiencies measured from terrestrial solar cells is due mainly to two
factors. The first is that the theoretical maximum efficiency predictions
assume that energy from each photon is optimally used, that there are no
unabsorbed photons and that each photon is absorbed in a material which
has a band gap equal to the photon energy. This is achieved in theory by
modeling an infinite stack of solar cells of different band gap materials, each
absorbing only the photons which correspond exactly to its band gap. The
second factor is that the high theoretical efficiency predictions assume a
high concentration ratio. Assuming that temperature and resistive effects do
not dominate in a concentrator solar cell, increasing the light intensity
proportionally increases the short-circuit current. Since the open-circuit
voltage (Voc) also depends on the short-circuit current, Voc increases
logarithmically with the level of the incident light. Furthermore, since the
maximum fill factor (FF) increases with Voc, the maximum possible FF also

16
increases with light concentration, allowing concentrators to achieve higher
efficiencies.

In designing single junction solar cells, the principles for maximizing cell
efficiency involve the following: a) increasing the amount of light collected by
the cell that is turned into carriers; b) increasing the collection of light-
generated carriers by the p-n junction; c) minimizing the forward bias dark
current; d) extracting the current from the cell without resistive losses.

3. PV Materials and Modules: Manufacturing Status

A. Crystalline Silicon
Silicon is the most developed and best understood semiconductor, benefiting
from decades of development by the integrated circuit (IC) industry. Multi-
crystalline silicon (mc-Si) photovoltaics have the greatest market share
followed by mono-crystalline silicon, followed by CdTe thin-film photovoltaics.
Although CdTe thin-film has the lowest module production cost, mc-Si has
higher efficiency, reducing, therefore, the cost requirement for the mounting
structure and installation, since those are proportional to the area required
for the installation. The mc-Si industry is looking forward to lower polysilicon
prices, and improvements in wire cutting technology that reduced wafer
thicknesses that could keep c-Si silicon competitive.. Several companies
have started producing solar grade silicon at less energy and at a lower cost
than the current production via themodified Siemens process, and R&D is
taking place for the production of kerfless wafers, and the development of
solar cells from depositing a thin-film silicon on substrates as amorphous
silicon is deposited today.

17
A.1 Production of Solar-grade Silicon
The solar grade silicon production is known to be the most energy intensive
stage in the life cycle of silicon PVs when the typical Siemens process is
used. This process was originally designed to produce the semiconductor
grade silicon, which is purer than the solar grade silicon. In order to reduce
the high electric energy requirement, modified version of Siemens process
which operates under a relaxed condition, have widely been used. In
addition, novel process schemes and equipment have begun emerging such
as fluidized bed reactor (FBR) and upgraded metal-grade (UMG) processes.
Figure 12 presents the currently available routes to produce solar grade
silicon; these are reviewed below.

Most SGS production is involving purification through gasification of MG

silicon to silane gases and deposition of Si by reducing the silane feedstock

18
gases. A new process that upgrades the purity of MG silicon (UMG-Si) has
also become available to commercial production.

A1.1 Silane and Chlorosilane Production

Most solar grade silicon (SGS) currently is obtained through silicon deposition
in reactors from mono- or trichloro- silane gas; the vast majority of the SGS is
produced in Siemens reactors with a small fraction of the production based
on the fluidized bed reactors (FBR). Silane gases are produced by gasifying
metal-grade silicon (MG-Si, >99%) that is acquired by reacting silica (SiO2)
with coke (C) in electric arc furnace. Trichloro-silane (SiHCl3, TCS) gas is
synthesized reacting MG-Si with HCl in a Fluidized Bed Reactor (FBR) at 200-
500 C: Si + 3HCl SiHCl3 + H2. In this process, a side reaction becomes
dominant as the temperature increases, i.e., Si + HCl SiCl4+ 2H2,
producing SiCl4, the byproduct of this process. Therefore, removing the
reaction heat is critical to improve the efficiency of TCS synthesis.7 In mono-
silane (MS, SiH4) gas production process, first, SiHCl3 is synthesized from
MG-Si, SiCl4, and H2 (Si + 3SiCl4 + 2H2 4SiHCl3), then converted to SiH4.
Mono-silane is also commercially produced from SiF4 through the following
process: SiF4 + NaAlH4 SiH4 + NaAlF4.7

19
 a Siemens Reactor (b) Fluidized Bed Reactor

Figure 13. CVD Reactors to Produce Polysilicon8

A1.2 Siemens Process

The typical Siemens process based on TCS deposits silicon on rods at 850 -
1100 C.8 The bell jar type inner wall of the Siemens reactor should be
cooled by water to prevent Si precipitation on the wall of the reactor, causing
significant heat losses (Figure 13 (a)). The electricity consumption ranges
60-150 kWh per kg of purified silicon.7 This process generates byproducts of
H2, HCl, SiH2Cl2 and SiCl4. The latter is recycled to SiHCl3 by reacting with
H2 in a FBR from the following reactions: SiCl4 + H2 = SiHCl3 + HCl,
3SiCl4+2H2+Si = 4SiHCl3. MS-based Siemens can lower the deposition
temperature to 650-800 C. 7 But, formation of unwanted fine silicon powder
7,9
causes product quality and process control issues.

Once the silicon rods reach 10-15 cm and no further room to grow inside the
Siemens reactor, the deposition operation should stop and the completed
rods be removed. This type of batch process, therefore, requires frequent
assembly/disassembly of the reactor, which considered labor intensive and
time consuming.7

A1.3 Fluidized Bed Reactor (FBR) Processes

Fluidized Bed Reactor method produces particle silicon compared with rod
shape silicon ingot produced in the Siemens process. Silane or
trichlorosilane gas is continuously supplied released into a superheated FBR
chamber that contains seed grains of silicon for continuous production
(Figure 13 (b)). The electric energy demand of producing SGS through FBR is
reported 10-30% of what is required for the Siemens process.7,9,10 The FBR
route is more energy efficient than the Siemens process for several reasons:

20
It does not waste energy as in the Siemens process, and produces more
silicon per cubic meter of reactor space. 11

This process is easily contaminated due to the large surface area of Si

particles than the bar Si obtained in the Siemens process, and to the fact
that those particles frequently contact with the reactor wall. The particle
size of Si should be constant for continuous production in FBR that seed
particles may be supplied to the reactor. The inner wall of the reactor is
coated with SiC as a barrier against contamination. But, in a high
temperature beyond 1000C, carbon can be released from the coating to
contaminate silicon particles.8 To avoid precipitation of silicon on the wall
and contamination, radiant heating such as high frequency, micro waves,
and infrared rays and so on is often used in an attempt to heat only Si
particles. 9 The large surface area of the Si particles also causes trapping of
hydrogen gases inside and their burst triggers generation of unwanted fine
silicon powders. An additional dehydrogenation process may be required at
1150 C. 9

The MEMC Inc. has been commercially producing SGS in a FBR process since
1987 from MS (SiH4) synthesized through reduction of SiF4. In addition, a
new FBR SGS process in REC uses SiH4 produced from disproportionation of
SiHCl3 (4SiHCl3 3SiCl4 + SiH4). It is known that this method reduces
electricity consumption to 20% of the Siemens process and the cost to less
than 70%. 9 MS-FBR has a lower deposition temperature compared with TCS-
FBR as MS deposits at 750C compared with TCS at 1050C.12 However, MS-
FBR tends to produce a large amount of unwanted fine silicon dust, which
largely forms at the empty space above the fluidized bed or inside gas
bubbles.7

Hydrogen reduction in FBR using TCS as a raw material also produces

purified particle silicon, which is used by Wacker. However, the details of
Wacker FBR process are not well known.12 In TCS-FBR, residual Cl could be

21
the major impurity, formed from dissociation ofSiCl2 [2, 3]. The density of
the latter is known to be lesser with a lower temperature and high pressure
reactor condition. 7

A1.4 Vapor to Liquid Deposition (VLD)

This is the process used by Tokuyama, Japan. It is similar to the Siemens

process but produces liquid silicon instead of solid silicon bar. The
precipitation rate is 10 times faster than the Siemens process but it is hard to
remove impurities. Thus, this process is suitable for SGS rather than
electronic grade silicon which requires a higher purity. 9

A1.5 Upgraded Metallurgical Grade Silicon

Solar-grade silicon based on metallurgical refining processes, often called

upgraded metallurgical-grade silicon (UMG-Si), is expected to play an
important role in reducing the energy and cost of silicon production for
photovoltaics. The leading company in this area is Norways Elkem which
reports the use of UMG-Si in solar cells with comparable efficiencies to
polysilicon (poly-Si) from the traditional Siemens process. Elkems process
starts from the production of metallurgical silicon in a reduction furnace with
advanced process controls, followed by purification stages involving slag
treatment, leaching and solidification, and subsequent post-treatment. The
purification stages are designed to remove impurities like boron and
phosphorous which can adversely affect the operation of the p-n junction in
solar cells. Elkem reports having reduced the concentration of B and P to
0.22 ppm and 0.62 ppm correspondingly, making their product comparable
to solar grade silicon from the modified Siemens process.

B. Alternative Si-cell Technologies

22
 B.1 Kerfless Wafers

Improvements in wire-saw technology have enabled the reduction of wafer

thicknesses from ~400 m a decade ago to 140 m at present. However,
over than 50% of the silicon is lost as silicon kerf and recycling of this kerf is
difficult as it is contaminated with solvents and abrasives. Kerfless wafer
could potentially reduce Si use by 50% or more . Techniques for the direct
production of wafers from the melt were invented in the 1970s,20, 21 and after
decades of development they have now reached the market. The two closely
related techniques are edge-defined, film-fed growth (EFG) and string ribbon
silicon technologies. In the EFG process the Si wafers are pulled out from the
melt through a graphite die using capillary action. This process was
developed by ASE Americas22 and is now employed by Schott Solar. Ribbon
silicon is produced by pulling a pair of high-temperature strings through a
crucible of molten Si, and this technology has been promoted and
implemented by Evergreen Solar. These two growth techniques produce
vertical sheets of mc-Si approximately 300 m thick and up to a 100 mm
wide.
The quality of the material produced by these techniques is somewhat
inferior to standard block-cast mc-Si, but is continuously improving with
champion laboratory cells now reaching power conversion efficiencies over
18% and 14.5% efficient modules already on the market.14, 15 Passivation of
surface and bulk defects is critical to achieving high efficiency, and this is
usually achieved through the deposition of a hydrogen-rich silicon nitride
layer, which also serves as an anti-reflection (AR) coating.23 These
technologies are expected to become more cost-competitive as energy costs
continue to rise, but improvements in manufacturing are needed to compete
with conventional c-Si technology (Fig. 3). In particular, further reductions in
wafer thickness coupled to improvements in throughput are viewed as the
most important tasks for scaling these technologies to TW levels. Increasing

23
the grain size and crystal quality will also be important for further
improvements in efficiency.3

B2. Thin-film- Silicon

Silicon has an indirect band gap, and, therefore, relatively thick layers are
needed to allow for momentum extraction and energy absorption. It is
assumed that silicon must be thicker than 100 m to effectively absorb
light,24 however theoretical modeling has shown that ~40 m may be ideal
for obtaining maximum performance,25, 26 and that efficient cells could be
obtained with only 1 m of single crystal silicon with state of the art surface
passivation.27 If such thin layers could be produced using a kerfless process,
and therefore avoid the large silicon loss due to cutting, this would result in
an order of magnitude reduction in materials cost and energy with respect to
todays state-of-the-art wafers. Thin silicon is also amenable to use of
bifacial architectures,28, 29 which harvest light from both directions.
There are two general approaches to the fabrication of thin-film-Si; these are
a) heteroepitaxial growth followed by lift off or removal of a sacrificial
substrate30, and b) peeling ultrathin silicon layers off of silicon ingots using
techniques such as stress-induced liftoff.31 The startup company Silicon
Genesis has recently introduced a process where this is achieved through a
combination of ion implantation and thermal treatment, producing kerf-free
wafers as thin as 25 microns.32 Substantial challenges remain once thin-film-
Si is produced. Achieving high efficiency will require the use of the most
advanced technologies with respect to both surface passivation33, 34 and light
trapping.35, 36 However, results to date are encouraging. The University of
Stuttgart has fabricated a 16.7% solar cell from 45 m thick Si produced by
lift off, while 8.2% modules have been fabricated using 2 m polycrystalline
Si.13 Commercialization of such efforts will need to address the nontrivial

24
challenge of mechanically handling these very thin wafers while maintaining
high throughput and low cost.

C. Amorphous and Nanocrystalline Silicon (a/nc-Si)

Solar cells based on hydrogenated amorphous silicon (a-Si:H or a-Si) were

invented by Carlson and Wronski at RCA Labs in 1976.61 Manufacturing is
achieved by plasma-enhanced chemical vapor deposition (PECVD) using
mixtures of H2 and SiH4. Hot-wire chemical vapor deposition has been offered
as an alternative,62 but has yet to be implemented in large scale
manufacturing. The capability to dope a-Si in a controllable fashion is
relatively poor and attempts to do so leads to the creation of additional
recombination centers. Because of these effects p-i-n device structures are
almost always used.63 Benefiting from synergies with the IC industry a-Si
was rapidly commercialized and the first PV products appeared in the early
1980s. Early devices rapidly surpassed the 10% conversion plateau, but it
was quickly recognized that these devices suffered from light-induced
degradation through the now well-known Staebler-Wronski effect64; light
exposure leads to a reduction of the solar cell efficiency over months which
eventually stabilizes at efficiencies around 6-7%. Nevertheless, for decades
a-Si was by far the most successful thin film technology, achieving market
shares in excess of 10% before the phenomenal growth of CdTe PV
production.
Leading manufacturing companies include Sharp and Oerlikon. One of the
most attractive features of a-Si is that devices can be deposited at low
temperature (< 200 C), enabling the fabrication of lightweight, flexible
laminates on temperature sensitive substrates. This is a unique trait that
provides a competitive advantage in markets such as consumer products and
building integrated photovoltaics (BIPV). Though discovered much earlier, 65
another major change that has occurred over the past decade is the
integration of micro (c-Si) or nanocrystalline (nc-Si) into device structures.

25
The quality of PECVD deposited material is strongly influenced by the level of
silane dilution in hydrogen, and high H2 dilution levels (>90%) lead to the
formation of crystalline domains within the material. The primary advantage
of nc-Si is that it is much less susceptible to Staebler-Wronski degradation. 64
Another important feature is that a/nc-Si is amenable to the formation of
multi-junction devices; most commercial devices are based on either tandem
cells or even triple junction cells. A common configuration is the
micromorph tandem, which pairs an a-Si top cell with a nc-Si bottom cell.66
Solar cells with record efficiencies are based on triple junctions that employ
germanium alloys to further improve absorption in the red region of the solar
spectrum.67 A related success story has been the introduction of the a-Si/c-Si
heterojunction with intrinsic thin layer or HIT cell, which boasts 21%
68
conversion efficiency. In addition to boosting current the intrinsic a-Si:H
layers appear to be important for passivation of the underlying c-Si material.

D. Cadmium Telluride (CdTe)

The first reports of CdTe-related PV devices appeared in the 1960s40 but

efficiencies were very low till the early 90s.41 CdTe has a number of intrinsic
advantages as a light absorber. First, its band gap of 1.45 eV is ideally
positioned to harness solar radiation. Its high optical absorption coefficient
allows light to be fully captured using only two microns of material. CdTe
sublimes homogeneously and the compounds thermodynamic stability
makes it is nearly impossible to produce anything other than stoichiometric
CdTe.42 Thus, simple evaporation processes may be used for film deposition.
Close-space sublimation employs diffusion as the transport mechanism,41
while very high rates (>20 m/min) may be obtained using convective vapor
transport deposition.43
Standard CdTe-based devices employ a superstrate configuration: production
begins with a glass substrate followed by the successive deposition of the

26
transparent conducting oxide (TCO, SnO2:F), the n-type window layer (CdS),
the p-type CdTe absorber, and finally the back contact (ZnTe/Cu/C). CdTe PV
manufacturing is uniquely equipped to be integrated with the production of
float line glass44. Glass exits a float line at ~600 C, which happens to be an
optimal temperature for vapor-phase deposition of the SnO2:F (FTO), CdS,
and CdTe. Part of First Solars success has been due to their ability to
integrate these various process steps into an in-line manufacturing process
reducing the processing time from glass to a finished module down to 2.5
hours.
With low manufacturing costs established, the biggest opportunities for CdTe
lie in the improvement of device efficiency. Champion cells (Fig. 4) convert
just over 50% of their S-Q potential, while commercial modules are at ~11%
power conversion efficiency. Improving efficiency will require enhancements
in both current and voltage. The former is perhaps the most straightforward
route, as much of the blue region of the solar spectrum is absorbed in the
TCO and CdS layers that make up the front contact. Top laboratory cells have
replaced the FTO with advanced TCOs such as cadmium stannate45 and ITO.46
Likewise, the CdS window layer (2.6 eV) absorbs a significant fraction of the
blue light. Integration of advanced front contacts into manufacturing appears
to be the near term strategy. This will not be trivial because ITO is expensive
and cadmium stannate is a complex material.47 Furthermore it is not clear
what might be used to substitute for CdS though sulfides of zinc and indium
have attracted significant interest.48
The more daunting challenge is improving the voltage. The open-circuit
voltage (Voc) of champion CdTe cells is well below that of similar band-gap PV
materials. For example, the best Voc obtained in CdTe is 230 mV short of GaAs
devices, which has a similar band gap. Short carrier lifetimes are at the root
of this limitation. The combined effect of defects and grain boundaries limits
minority carrier lifetimes in polycrystalline CdTe to a few nS, even the best
devices. These lifetimes are very short compared to almost 1 s for epitaxial
CdTe49 or hundreds of ns for CIGS.50 Sites and Pan51 showed through

27
simulation that increasing the carrier lifetime or the use of a p-i-n device
structures may be two viable routes to increase the efficiency to above 20%.
The short term goal of commercial manufacturers is to raise module
efficiencies from current levels to >15% by 2014 through a combination of
process integration, research, and development. In the longer term they are
targeting 18% as an achievable goal for module efficiency.52
A number of fundamental questions remain unanswered for CdTe PV. At
present, the issue of extending carrier lifetime is partially addressed by
chemical passivation. Examples include the introduction of O2 during CdTe
growth,53 post-deposition CdCl2 treatments,54 and controlled diffusion of Cu
from the back contact.55 The empirical recipes associated with these
processes constitute the black art of CdTe manufacturing. Clearly a
preferable route would be to understand the nature of the defects states so
one could prevent their formation in the first place or develop alternative and
perhaps better passivation strategies. Fundamental research in
understanding these defects and how to passivate them would be
transformative leading to improvements in one of the most promising solar
cell technologies. Another fundamental question concerns the role of grain
boundaries in these devices.56 CdTe is an interesting and unusual material in
that solar cells based on polycrystalline CdTe outperform devices made using
single crystal CdTe. It is thought that grain boundaries can have both positive
and detrimental impacts on charge transport, but the current level of
understanding is not sufficient to suggest how one might engineer a desired
morphology. The use of p-i-n structures to create high-efficiency devices
requires deliberate control of the sample free carrier density, which is not yet
fully understood or achieved. A final area that deserves attention is the back
contact. It is difficult to contact CdTe because it has low conductivity.
Moreover, the back contact has been implicated as a potential contributor to
degradation.57, 58 The issues discussed above are non-trivial and will require
substantial investment and fundamental research to resolve.

28
Another issue to be mentioned with respect to large-scale CdTe
manufacturing is perceptions with respect to both cadmium toxicity and
tellurium availability. The toxicity issue is one of public perception which has
been amplified by competitive business interests. Cadmium is indeed a toxic
element, but the risk of exposure once incorporated into PV modules is
minimal. Extensive testing of broken and heated panels has shown that
emissions due to fire59 or leaching60 are negligible. To their credit, all CdTe
manufacturers are committed to 100% ownership of recycling, which in part
is related to the issue of Te availability discussed later in this paper. One also
notes that Cd will continue to be produced as a natural byproduct of Zn
mining. Perhaps the best argument for CdTe PV is that it serves as a means
to sequester this element in an environmentally beneficial manner. While
scientific arguments can be made that the toxicity of Cd is not a significant
issue, governmental policy in individual countries may dictate whether this
issue hampers the deployment of CdTe solar cell technology.

E. Copper Indium Gallium Selenide (CIGS)

Copper indium diselendide (CuInSe2 or CIS) CIS has a band gap of 1.05 eV
slightly off the maximum efficiency energy gap (fig 10). However the band
gap may be continuously engineered over a very broad range (1.05 2.5 eV)
74
by substituting either Ga for In or S for Se. CIGS is currently the efficiency
leader among thin film technologies cells with 20.3% efficiencies currently
reported in the laboratory75, 76. Commercial production of CIGS began in
2007, with a few companies operating facilities with 10-30 MW/year
capacities but the foretold large production has not happened yet as
reproducibility and production yield have proven to be challenging.
However, Solar Frontier in Japan recently expanded production to 900 MW.

Substrates include soda lime glass, metal foils, or high temperature

polyimide (PI). The latter has garnered substantial interest for applications

29
such as BIPV and small portable applications. In the case of deposition on
flexible substrates it is critical to match the coefficient of thermal expansion,
with highest efficiencies obtained on titanium and stainless steel foils.
Fabrication of the CIGS device begins with the deposition of a Mo back
contact followed by the p-type CIGSS absorber (1-3 m), a thin buffer layer
(50-100 nm), with doped ZnO serving as the transparent front contact. Many
different approach exist for the formation of the CIGSS absorber.77 At
present, the performance of commercial modules is only about 11%
although the efficiency record of lab cells is 20.3% and much of this
difference attributed to the quality of the absorber layer.78 The approaches to
CIGS fabrication may be classified into three basic categories: co-
evaporation, reactive sputtering/selenization of metal films, and non-vacuum
and solution printing techniques. Below is a description of the major
advantages and issues associated with each approach (Wolden et al, 2011).

Co-evaporation
Co-evaporation is practiced either in a single or three-stages. The former is
employed by Q-cells and Wurth Solar and has produced 12.7% efficient
modules(Q-cells). The three-stage co-evaporation is the process that has
produced the world-record cells79 but so far it has produced 12.2% efficient
modules (Global Solar, Yohkon). Co-evaporation alternates between copper-
rich and copper-poor conditions to produce the large grains and graded Ga/In
profiles characteristic of high efficiency material.80 There are a number of
important practical challenges involved in the manufacturing of CIGS solar
cells. Evaporation sources typically have a cosine flux distribution, and it is
difficult to introduce sharp changes in composition or maintain uniformity
over large areas under the diffuse conditions of high vacuum. In addition,
sources must be mounted in a top-down configuration in order for large glass
substrates to be supported and heated to 600 C. In situ diagnostics such as
thermometry and laser light scattering, which are critical for process control
in the batch process, are being adapted for use in the manufacturing

30
environment.81 Likewise, atomic absorption spectroscopy and X-ray
fluorescence are employed for controlling element flux and in-line detection
of film composition, respectively. Another challenge with co-evaporation is
that the relatively unreactive Se must always be supplied in great excess,
leading to practical concerns related to condensation and materials
management. Despite these challenges Q-Cells has announced 12-13%
mass-produced modules and a 14.2% champion module with this process.
Through systematic optimization and accompanying improvements in yield
this may turn out to be a viable large-scale manufacturing strategy.

Reactive Sputtering/Selenization
Another method for synthesizing CIGS films is selenization of a stack or alloy
of the constituent metal films predeposited on a substrate in a
predetermined stoichiometry. There are many variations of this approach but
the most common is a two-step process where the metals are sputtered
onto the substrate and then converted to CIGS through annealing in a
chalcogen-containing environment. Practitioners include Solar Frontier who
have recently announced a scaling up of their production to 900 Mw/yr,
Avancis, Stion and TSMC Solar. The highest module efficiency obtained by
the two-step process is 12.6% (Solar Frontier). There are many pathways
and intermediates involved in transforming the metal into the chalcopyrite,
requiring careful optimization of the time-temperature-reactant profiles
employed.78 This approach creates a copper-indium-gallium-selenide-sulfide
(CIGSS) film, with shifts the band gap towards the maximum efficiency and
also improves the interface with the window layer, which are also sulfides
(CdS, ZnS, In2S3).78

Non-vacuum and Solution-printing Processes

The third general approach to CIGS manufacturing has been to eliminate
vacuum processing. In general these are also two-step processes,
application of a coating followed by a high temperature step for annealing or

31
sintering. Ostensible advantages include reduced capital requirements,
improved materials utilization, potentially lower energy requirements, and
compatibility with roll to roll (R2R) processing.82 A general challenge with the
non-vacuum based approaches is the potential of contamination introduced
by either the compounds themselves or the solvents employed. As such, it
has been much harder to produce dense, homogenous absorber layers. It is
also more challenging to produce chemically-graded structures with this
technique. Record cell efficiencies trail co-evaporation and metal
selenization, but values up to 14% have been obtained by a number of
techniques.
The non-vacuum strategies may be further divided into electrodeposition,
particulate deposition, and solution processes. Electrodeposition has been
around for decades83 and achieved cell efficiencies as high as 13.8%,84 but
concerns about up scaling appear to have limited commercial interest. The
particulate route is currently the most actively pursued, with variations
employing particles composed of CIGS, metal, metal oxides, and/or metal
selenides. In all of these methods a coating of particles is first formed on
the substrate surface and reacted and/or sintered at high temperature to
form the final film. It was found that CIGS particles required excessive
temperature for sintering.85 Likewise, problems with handling and premature
oxidation have limited the utility of metal particles. The best results have
come using slurries containing mixtures of metal oxide or selenide powders. 86
This approach was pioneered by Kapur and co-workers at ISET,87, 88 and more
recently championed by Nanosolar. The latter has reported 14% efficient
cells,89 and has announced that 10-11% modules will be available in the near
future. Solution approaches have employed the use of soluble metal salts,
organometallics, and hydrazine-based compounds. Best results have been
obtained with the latter,90, 91 however the highly reactive and toxic nature of
hydrazine poses additional complications for manufacturing.

32
With five elements and numerous binary and ternary phases, the CIGSS
system presents much greater complexity than the PV technologies
described previously3. Extensive theoretical work has made great advances
in understanding the electronic structure and role of defects in this system.92
These studies have been aided by improvements in advanced
characterization techniques. Raman and time resolved photoluminescence
are becoming useful for identifying the presence of secondary phases and
certain defects.81 It is well-known that sodium plays a critical role in the
morphology and electronic properties of CIGS. When soda lime glass
substrates are used, sodium diffuses into the CIGS layer from the glass.
Once the importance of Na was realized, more controlled and systematic
approaches that employ sputtered layers of Na-containing material have
been developed to gain control over Na introduction into the CIGS. There is
also significant attention being paid to the window layers deposited on the
CIGS absorber. While CdS remains the leading choice, both indium and zinc
sulfides are being pursued and in some cases commercialized. Part of the
interest is due to the desire to remove Cd, but a second motivation is
improving the blue response of these devices. There are strong interactions
between the buffer and the underlying absorber, and simultaneous
optimization of these layers is required for best performance. Some concerns
remain about the use of ZnO as the front TCO, and its potential impact on
long-term device stability. Moisture exposure is particularly detrimental, both
to the TCO and the heterojunction itself. Encapsulation in glass partially
alleviates this effect, but further development of transparent ultra-barriers is
required to improve the long term stability of flexible CIGS solar cells.93
A longer-term concern is the availability and price of In. Recycling of
indium will alleviate constraints on CIGS long-term production, but research
is needed to develop technologies for efficient and low cost recycling of all
the elements from the CIGS modules. The possibility of substituting
indium/gallium with earth abundant alternatives such as zinc/tin is discussed
below.

33
F. Copper Zinc tin Sulfide Selenide (CZTSS) Thin Films

There is interest in using abundant materials in thin-film photovoltaics,

driven by concerns about the availability of In, Te, Ga and Ge in the current
2nd generation thin-film PV. Such materials include oxides and sulfides of
base metals (e.g., Fe, Cu) that have band gaps in the range of 1 2 eV. The
most successful system to date has been copper-zinc-tin-sulfide (selenide),
or CZTS.125 Pioneered by Katagiri,125 in the past decade champion CZTS
devices have gone from less than 3% to approaching 10% efficiency (Fig. 5).
These results have come with only a handful of publications. CZTS shares
great similarities with CIGS, including similar device structures and
fabrication techniques for the formation of the absorber layer. Initial studies
focused on sulfidization of metal layers,126 but more recently co-
evaporation127 and non-vacuum techniques128-130 have garnered significant
attention. The current efficiency champion includes Se and was derived from
hydrazine precursors.130 The similarities to CIGS may have accelerated CZTS
solar cells initial success, but these same similarities may become
limitations in the long run.3

G. Dye-sensitized Solar Cells (DSC)

Dye-sensitized solar cells are based on the photo electrochemical effect

discovered by Bequerel in 1839. A relatively new concept, DSC was
introduced by Grtzel and co-workers in 1991.99 A comprehensive review on
the complex chemistry and processes involved in this system was recently
published by Hagfelt and co-workers.100 This hybrid material is typically
composed of organometallic dye molecules adsorbed to a mesoporous titania
nanoparticle film, with the pore space filled by an electrolyte. In this
structure light is absorbed by the dyes, which then inject an electron into the
conduction band of a wide band gap semiconductor like TiO2. The electron is

34
transported by hopping through the nanoparticle network to the front
contact where it exits and performs useful work before returning to the
platinized back contact. Here the electron reduces a redox couple, which in
turn diffuses through the electrolyte and regenerates a dye molecule to
complete the cycle. Dye sensitization of oxides was well-known at the time,
and Grtzels key innovations were in creating a nanoparticle film with high
surface area to improve light harvesting and in choosing components with
appropriate kinetics for fast charge transfer and slow recombination.
Grtzels group rapidly optimized the device to over 10% within a few years
of its introduction.101 This brought the attention of industry and today a
number of small companies including G24i, Solarprint and Dyesol are
engaged. Most current products are directed at the consumer market; for
example, DSC on flexible substrates that replace rechargeable batteries for
portable electronics. A unique feature of DSC is that their performance
improves under diffuse and low light conditions,102 enabling their use indoors
and without direct solar exposure. Devices can be fabricated in a number of
colors and levels of transparency, which is an attractive feature for
architectural and BIPV applications. Manufacturing can also be done at low
temperature using flexible substrates.

It is noted that champion cell efficiency has been stagnant at ~11% for the
past 15 years (Figure 5). The three main components in a DSC, the Ru-based
dye, the photoanode, and the iodine-based redox couple, have also remained
unchanged. Further optimization of any one of these components individually
is not likely to yield significant improvements in efficiency. The recent review
by Hamann and co-workers103 provides an excellent overview of the
complexity of the issues involved. First, the leading dye does not capture
much light past 750 nm, and harvesting the red and near-infrared portion of
the spectrum is needed to increase current densities. Second, the I 3-/I- redox
couple is positioned with a 550 mV overpotential relative to dye
regeneration. An alternative redox couple could potentially allow the Voc to

35
be improved by up to 300 mV, but recombination rates are typically much
faster with non-iodine redox couples. A combination of these two changes
could elevate device performance to > 16%. However as cautioned by
Hamann and colleagues,103 this will most likely require simultaneous
optimization of both dye and electrolyte and perhaps the development of
new photoanodes with faster charge transport as well. While there are
photoanode designs based on wide band gap semiconductor nanowires that
attempt to improve efficiencies, six years after their first introduction, the
efficiencies remain low.

With respect to manufacturing numerous module fabrication strategies are

being pursued, which general can be divided into monolithic or sandwich
constructions. The former offers advantages with respect to materials cost,
while the latter may be more amenable to R2R processing. Substrates
include glass, metal, and polymer foils, with best performance being
obtained on glass. Critical issues include stability and the production of large
area modules. At present mini-modules with areas < 100 cm2 are used, with
resistance losses being a one of the major challenges. The stability of a DSC
module is strongly related to the device encapsulation. Accelerated lifetime
testing has indicated that carefully encapsulated glass-based DSC can last
for over 20 years, although current products using DSC on plastic substrates
have lifetimes of only a few years. For outdoor applications, the sealing
material must, for example, be mechanically and thermally stable, stable
under UV exposure, and chemically inert to the electrolyte. Moreover, it
should prevent mass transport between adjacent cells. The issue so
100
important that Hagfelt et al. suggested that the leading manufacturing
approach for DSC may be the one that provides the most functional
encapsulation method. Replacing the liquid electrolyte with a gel or solid
would greatly improve encapsulation issues, but these changes have
resulted in decreased efficiency. Elimination of glass, implementation of R2R

36
manufacturing methods, and increased lifetimes will be critical to economics,
particularly if device efficiency remains below 12%.

H. Organic PV (OPV)

Spurred by the elemental abundance of carbon, extensive efforts to develop

solar cells using organic semiconductors are underway. Brabec and
colleagues104 recently provided a comprehensive review of the developments
in OPV over the past decade and the challenges that lie ahead. Figure 5
charts the progress of champion cell efficiencies for the past 15 years. While
most technologies have been relatively stagnant in their champion
efficiency, organic PV has made great strides in the past decade, with
Heliatek and Konarka being the current champions, each with devices
certified at 8.3%.105, 106 Unfortunately Konarka run out of cash and went out
of business in June of 2012. Others leading the development of OPV for
small devices include Solamer and Plextronics who held the efficiency record
in recent years. OPV devices are comprised of a heterojunction between an
electron donor molecule (e.g., P3HT, poly(3-hexylthiophene) or CuPC, copper
phthalocyanine) and an electron acceptor molecule (e.g., C60 or its
derivatives such as PCBM, phenyl-C61-butyric acid methyl ester). 107 The
essentially limitless varieties of candidate organic semiconductor materials
may be categorized as either solution-processable (polymers, dendrimers,
oligomers, or small molecules) or vacuum deposited (small molecules or
oligomers). Although superficially similar to inorganic p-n junctions, the OPV
junction is fundamentally different. Instead of directly creating an e-/h+ pair,
photon absorption produces an exciton, an uncharged excited state that
must diffuse to a donor/acceptor interface in order to dissociate into a free
e-/h+ pair. In organic materials the exciton can typically only diffuse 5-10 nm
before decaying to the ground state, a problem that limits performance and
is typically referred to as the exciton bottleneck. There are two ways to deal
with this. One can make a multilayer device that uses very thin

37
donor/acceptor layers such that a majority of excitons can diffuse to a
heterojunction interface.107 This approach is commonly used in vacuum
deposited devices. Or one can reduce the distance the exciton has to diffuse
before reaching the heterojunction by mixing the donor and acceptor
materials on a nanometer length scales to form a single-layer
interpenetrating bicontinuous network called a bulk heterojunction. This
approach is commonly used in solution-processable materials.

In the OPV device structure, the heterojunction active layer(s) is(are)

sandwiched between a set of contact electrodes, with buffer layers likely to
be present. In the bulk heterojunction approach an asymmetry in the device
must be imposed by either using electrodes of different work function
(typically a front TCO contact modified with a conducting polymer PEDOT-
PSS, poly(3,4-ethylenedioxythiophene) poly (styrenesulfonate) and a back
contact metal of Ca or Al) or by inserting a buffer layer that blocks carriers
from leaving one side of the device. An oxide buffer layer is commonly
inserted to block holes from leaving the device through the front TCO
contact, which inverts the direction of operation of the device and allows the
use of a high workfunction Ag back contact.108-110. In the vacuum deposited
multilayer approach, co-doping of buffer layers has been used to great effect
to produce a true p-i-n structure that obviates the need for a mismatch in the
contact work-functions.111

There are several challenges to improve the efficiency of organic solar

cells.104, 112 These are being addressed through the development of novel
donor and acceptor materials, new buffer layer and electrode geometries,
innovative processing, and through the use of tandem architectures. A key
issue is to significantly raise the short circuit currents (Jsc) to above 20
mA/cm2. Present values are typically 10-12 mA/cm2 with champion values
approaching 17 mA/cm2.104 The main problem is that leading photoactive
layers do not efficiently harness photons in the red and infrared region of the

38
solar spectrum. Significant efforts have been directed at developing
improved low band gap polymers.113, 114 Advanced photon management
strategies are also being pursued to increase optical density.115 A second
challenge is to increase the open circuit voltage. Key to this is achieving
optimal band alignment of the device structure and minimizing the band
offset between donor and acceptor molecules while retaining efficiency
113,116
charge transfer. It is predicted that the maximal Voc in a standard donor-
acceptor device is 0.6 V less than the bandgap energy/e. Thus the goal is for
a Voc of 0.8 0.9 V for low band gap absorbers with band gaps ~1.4 1.5 eV.
Vocs above 1 V have been achieved, but only with high band gap materials.
Third, the fill factors (FF) have to be increased beyond 0.7, which has been
achieve in only a few champion devices.104 Organic solar cells typically have
poor FF relative to conventional p-n junctions. This is due to high series
resistance and/or carrier recombination as the carrier mobilities in organic
thin films are lower than their inorganic counterparts.
Simultaneous achievement of Jsc = 20 mA/cm2, Voc = 0.8 V, and FF = 0.7,
leading to an efficiency of 11% has not been achieved yet. Doing this in a
single junction device will require simultaneous optimization of all the
materials and interfaces. A possibly faster route to this goal will involve the
use of tandem configurations.117 These have been demonstrated using the
bulk heterojunction approach and are being used to effectively boost
efficiency in the evaporated small molecule approach as implemented for
instance by Heliatek. Passing the psychological milestone of 10% efficiency
could bring organic solar cells within striking distance of the existing thin film
technologies, particularly because manufacturing costs are expected to be
low.104 Even with existing materials and devices, the energy payback time for
OPV has been estimated at 0.3 years.118 However, an efficiency of closer to
15% may be needed to achieve a true grid parity LCOE of ~$0.07/kWh.119
Much work still needs to be done to demonstrate acceptable performance in
120
large area modules. At present OPV submodule (200 cm2) efficiencies
from leading companies are approaching 4%.121 This value lags substantially

39
behind the 9.2% efficiency in comparable sized DSC modules.13 Also,
published champion OPV devices are fabricated on glass. To be economical,
the substrate will likely need to be a low cost flexible material that is suitable
for R2R processing.
Another important issue that has to be resolved is the stability of organic
solar cells. The chemical, physical and mechanical degradation that are
predominant in OPV materials and devices have been well discussed.122 The
list of failure mechanisms of OPV cells is long and certainly as extensive as
for any other photovoltaic technology. Major issues include photodegradation
and the sensitivity of OPV components to oxygen, requiring the use of
ultrabarriers for encapsulation. The current goal is to increase lifetimes from
3 to 10 years, which is expected to be sufficient for consumer applications. 104
Due to the flexibility of organic synthesis, it can be estimated that there are
on the order of 1013 different material combinations that could be employed.
Whether the right combination of properties (e.g., band gap, charge mobility,
exciton diffusion length, etc.) exists and how to identify them remain open
questions. Optimizing the photoactive organic layer may be best addressed
using a combinatorial approach. On the other hand, candidate structures
and trends may be identified using a rational method that combines
computational methods with targeted synthesis.

4. Future Outlook

As discussed earlier, there is a large gap between the module and the cell
efficiency for all currently commercial PV technologies which can likely be
narrowed with manufacturing R&D. For crystalline silicon modules,
efficiencies can be improved by maximizing the collection of current, since
little can be done to further improve the voltage. Related efforts include
improvements in front side texturing, integration of back side reflectors, and
the use of advanced AR coatings.37 Another strategy is to reduce the level of

40
shadowing associated with front contacts by either reducing the line-widths
or by completely eliminating them using back side contacting schemes such
as those employed in Sunpowers high efficiency modules.38 Also the
limitation of quantum efficiency in the blue region of the spectrum can be
addressed with selective emitter designs and improved passivation
strategies.39 Through a combination of these strategies it is expected that
micro-crystalline-Si modules will exceed 20% within the decade, and mono-
crystalline -Si modules would approach 25%.3 Currently the best efficiencies
in nono-crystalline Si modules are slightly above 20% for modules
manufactured by SunPower and Sanyo.

The next few years will prove critical to amorphous silicon technologys long-
term viability. While necessary to improve stability, the transition from a-Si to
nc-Si has come at an expense. Due to its relatively low absorption
coefficient,69 nc-Si based devices need to be up to five times thicker than a-Si
to collect sufficient light. This issue is exacerbated by the fact that the
deposition rates for nc-Si are much lower than those for a-Si.70 Combined
with the relatively low efficiencies, this has made manufacturing of a-Si/nc-Si
based solar cells prohibitively expensive when compared to alternative
technologies such as CdTe and recently with cheap imports of c-Si modules
from China. Efforts to improve deposition rates include: a) use of very high
frequency (VHF: 25 100 MHz) plasma sources; b) operation at higher
pressures and c) development of linear plasma sources to maintain large
area uniformity with VHF modulation.71-73 Barring a significant breakthrough,
a/nc-Si may need to focus on market sectors that benefit from its low
temperature, low weight capability. Another strategy might be to examine if
a-Si could be used as a route to form ut-Si, perhaps by coupling with rapid
thermal processing.3

The improvement of the CdTe PV technology is expected to continue as first

solar continues to invest in R&D and others are following their lead (e.g.,

41
PrimeStar, Abound). Through improvements in efficiency and further
integration of manufacturing CdTe manufacturing costs could drop to $0.5/Wp
within this decade.
However, the c-Si industry has shown that they can also rapidly reduce costs,
and it is important to note that silicon PV module prices does not have to be
reduced as much as CdTe, if their efficiency is high. As module costs drop
below $1/Wp they become increasingly less important, as balance of system
issues including inverters, racks, installation, and space become important
cost drivers. These costs drop with module efficiency, and thus silicon can
afford to remain competitive despite higher manufacturing costs than CdTe.
It is too early to project what may happen to CIGS, though it is plausible to
expect that it might end up in a similar region, most likely somewhere
between CdTe and c-Si.3
There have been numerous scientific breakthroughs in the area of 3rd
generation PV,132-135 but practical devices are far from the point of
commercialization.136, 137 In general 3rd generation PV rely on the use of
nanostructures such as quantum dots and nanowires to generate the desired
effects. A critical yet unresolved problem with devices that employ such
structures is that they will be dominated by interfaces, the anathema of PV
technology.3 Interfaces typically serve as either recombination centers or
barriers to charge transport, and the demonstrated pathway to high
efficiency has been through their elimination. Record heteroepitaxial multi-
junction cells are produced by molecular beam epitaxy.138 The detrimental
impact of interfaces is quite plainly seen by comparing the performance
within the silicon system (in terms of performance c-Si > mc-Si > nc-Si > a-
Si). Likewise, record CdTe and CIGS thin film devices are characterized by
their large grain size.139 Development of optimized manufacturing
techniques requires sophisticated modeling to understand how to maintain
uniformity with respect to both area and time. Accompanying this goal is the
development of in-line diagnostics for real time process control.

42
 Developments in intelligent and potentially self-correcting control of process
flow would help enable and accelerate throughput volume.

5. Sustainability Life Cycle Analysis

Photovoltaics do not need any fuel to produce electricity but energy is

needed for generating their materials, cells, modules, and systems. As in all
types of products and systems, a complete evaluation of their environmental
profile must be done under the framework of a life-cycle analysis.140 The life-
cycle of photovoltaics starts from the extraction of raw materials (cradle)
and ends with the disposal (grave) or the recycling and recovery (cradle)
of the PVs components (Figure 14).

M, Q M, Q M, Q M, Q M, Q M, Q

Raw Material Acquisition Decommis-sioning

Treatment /Disposal
Material Processing
Manufactur-ing Operation

M, Q
E E E E E E

Recycling

M, Q: Material and energy inputs

E: effluents (air, water, solids)
E
Figure 14: Flow of the life-cycle stages, energy, materials, and effluents for
PV systems

The mining of the raw materials, for example, quartz sand for silicon PVs,
copper-, zinc-, and aluminum-ores for mounting structures and thin-film
semiconductors, is followed by the multiple stages of separation and
purification. The silica in the quartz sand is reduced in an arc furnace to
metallurgical-grade silicon that must be purified further into solar-grade
silicon (i.e., 99.999% purity), requiring significant amounts of energy.
Metallurgical-grade cadmium, tellurium, indium, gallium, and selenium for

43
CdTe and CIGS PV primarily are obtained as byproducts of zinc-, copper-, and
lead-smelting, and then further purified to solar grades. The raw materials
include those for encapsulations and balance-of-system components, for
example, silica for glass, copper ore for cables, and iron- and zinc-ores for
mounting structures. The production of all these materials requires large
amounts of energy, as does the manufacture of the solar cells, modules,
electronics, and structures, their installation, operation, and eventually their
dismantling and recycling141 or disposal.

Thus, the energy payback time (EPBT) is defined as the period required for a
renewable energy system to generate the same amount of energy (in terms
of primary-energy equivalent) that was used to produce the system itself.

Energy Payback Time (EPBT) = (Emat+Emanuf+Etrans+Einst+EEOL) / (Eagen Eaoper)

where,
Emat : Primary energy demand to produce materials comprising PV system
Emanuf : Primary energy demand to manufacture PV system
Etrans : Primary energy demand to transport materials used during the life
cycle
Einst : Primary energy demand to install the system
EEOL : Primary energy demand for end-of-life management
Eagen : Annual electricity generation in primary energy terms
Eaoper : Annual energy demand for operation and maintenance in primary
energy terms

An indicator more commonly used for comparing different types of energy-

production technologies is the Energy Return on Energy Investment (EROI)
that quantifies the benefit that the user gets out of exploiting an energy
source. Usually, it is expressed as the dimensionless ratio of the energy
generated from a system over that energy, or its equivalent from some other
source, that is invested in extracting, growing, or processing a new unit of

44
the energy in question. Thus, EROI even can be used for fuel-based power
production that never pays back its energy requirement as it continuously
needs energy in the form of depletable fuel to operate. The traditional way
of calculating the EROI of PV is EROI = lifetime/ EPBT; thus, an EPBT of 1 year
and a life expectancy of 30 years corresponds to EROI of 30:1. The EROI of
conventional thermal generation from fossil fuels has been viewed as much
higher than those of photovoltaics; this recently was shown to be a
misconception fostered by using outdated data and a lack of consistency
among calculation methods. 142

Several published PV LCA studies give differing estimates of the EROI. Such
divergence reflects the varied assumptions about key parameters, like
product design, solar irradiation, performance ratio, and lifetime. These
assessments also deviate because of the different types of installation used,
such as ground mounts, rooftops, and faades. Also, assessments still are
calculated from outdated information in the literature collected from
antiquated PV systems. An example of such misrepresentation of the current
reality was apparent in a recent PE Magazine article wherein the author
stated that photovoltaic electricity generation cannot be an energy source
for the future because photovoltaics require more energy than they produce
during their lifetime. Statements to this effect were not uncommon in the
1970s based on some early PV prototypes. However, todays PVs return far
more energy than that embodied in the life cycle of a solar system; Fig.15
illustrates this historical trend to betterment.

To resolve these inconsistencies in the estimates, the International Energy

Agency (IEA) PVPS Task 12 published the Methodology Guidelines on Life
Cycle Assessment of Photovoltaic Electricity for conducting balanced,
transparent, and accurate LCAs . 143 The results, presented in Fig. 16, were
obtained following these Guidelines and the associated Life Cycle Inventory

45
(LCI) data144; they represent consensus between LCA experts from the ten
Task 12 member countries.

50

mono-c-Si
multi-c-Si
5 CdTe
mono-c-Si
multi-c-Si
CdTe

0.5
1960 1970 1980 1990 2000 2010 2020

Fig. 15. EPBTs of various PV systems were reduced from about 40 yrs to 0.5
yrs from 1970 to 2010

Figure 16 plots the Energy payback times (EPBTs) of three major types of
commercial PV modules , i.e., mono-Si, multi-Si, and cadmium telluride
(CdTe).
1.8
1.6
1.4
1.2
1 BOS
0.8 Frame
0.6 Module

0.4
0.2
0
mono-Si PV 14% multi-Si 13.2% CdTe PV 10.9%

46
Figure 16: Energy payback times (EPBTs) of rooftop mounted PV systems for
U.S- and European- production and installation under average U.S. irradiation
of 1800 kWh/m2/yr (4.9 kWh/m2/day), a performance ratio of 0.75, and the
module efficiencies shown above. Data adapted from de Wild Scholten
(2009) and Fthenakis et al. (2009); note that module efficiencies have
increased since 2009, and, correspondingly, EPBTs have decreased.

These results are based on detailed process data obtained through

collaborations with thirteen European- and US-PV manufacturers. 145,146
The EPBT for the same type of systems installed in the US-SW were
decreased in proportion to the solar irradiation ratio (1800/2380) between
the US average- and SW-solar conditions. Thus, for SW irradiation, the EPBTs
for the three PV technologies shown in Fig. 16 are 1.3-, 1.3- and 0.5-years
and their corresponding EROIs are 23:1, 23:1 and 60:1.

Fig.17. Energy PaybackTimes for different insolation levels in the United

States. This Solar Resource map was produced by B. Roberts, National
Renewable Energy Laboratory for the U.S. Department of Energy; the colors
show annual averages of daily insolation for the south facing latitude-tilt

47
plane. The EPBTs were estimated by V. Fthenakis, Brookhaven National
Laboratory.

6. The Intermittency Challenge

The short-term variability of solar power recently garnered much attention

because the installed capacity is increasing very rapidly, and the technology
is on its way to become a significant part of the generator portfolio (power
capacity) of several countries, notably Germany (12%, 2010), and Spain
(4.3%, 2010) (German Energy Ministry, 2011; Red Electrica De Espana,
2010). German Energy Ministry. (2011). Erneuerbare Energien 2010.
Vierteljahrshefte zur Wirtschaftsforschung (Vol. 76). doi:10.3790/vjh.76.1.35
Red Electrica De Espana. (2010). Red Electrica De Espana. Retrieved from
http://www.ree.es/

Variability can be smoothened by aggregating solar generation over large

geographical regions and by combining solar and wind generation, since in
many places one compliments the other (e.g., more wind at night than
during the day and more in the winter than in the summer). This has been
147
illustrated for New York State and the ERCOT grid system.148

7. Conclusions

Although recent market uncertainties have slowed down PV production in

2012, the long-term potential of the technology is clearly upward, and it
would accelerate when electricity production by PV reaches parity with
conventional electric power generation technologies in large regions of the
world. It is expected that crystalline silicon will continue dominating the

48
market for the foreseeable future. However, significant opportunities remain
for thin-film photovoltaics, mainly through continued increase in module
efficiencies. If investments in CdTe and CIGS thin-film PV technologies
continue, it is not unreasonable to expect that the module efficiencies in
commercial production will reach those of the high efficiency crystalline-
silicon PV modules. Pending further improvements in efficiency, CdTes
market share could evolve to anywhere between 10 25%. After decades of
being the leading thin film technology, the prospects of amorphous silicon
making a major contribution to the utility sector appear severely constrained
by high manufacturing costs and low efficiency, with the exit of UniSolar from
the market being the evidence of the technology lacking competitive merit.
However, the a-Si/nc-Si (micromorph) architecture is still holding up. CIGS
has gained a foothold in PV manufacturing at ~100 MW/year and higher
(Solar Frontier) and the demonstrated potential in efficiency provides reason
for guarded optimism. The next five years will be critical for CIGS to
demonstrate robust manufacturing to allow it to break out and join CdTe as a
major player in the global PV market. Although encouraging industrial
progress has been made, DSC and OPV are still limited at present by low
efficiency and stability. Konarka, the US front-runner on OPV recently exited
the market as their production costs were relatively high.

Concerning the future, the future of solar PV still looks bright. The industry
has consistently been able to lower the cost of solar panels. This author
believes that this trend can be maintained, and if subsidies are continued for
the next 10 years, there is a real prospect for solar to become cost
competitive without any subsidies. What is needed more than anything else
at this critical conjuncture is the maintenance of federal tax incentives and
state renewable portfolio standard incentives or the enactment of feed-in-
tariffs following Germanys example. With solar replacing the current fossil-
fuel and nuclear based power infrastructure, the environmental benefits,
which are not accounted in the current cost structures, will be enormous.

49
References
[1] Zweibel K., Mason J. and Fthenakis V. Solar Grand Plan: Solar as a Solution, Sun&Wind
Energy, 4 (2008) 112-117.
[2] Fthenakis V. Mason J. and Zweibel K., The Technical, Geographical and Economic
Feasibility for Solar Energy to Supply the Energy Needs of the United States, Energy
Policy, 37, 387-399, 2009Fthenakis
[3] Wolden C, Kurtin J., Baxter J., Repins I., Shaheen S., Torvik J., Rockett A., Fthenakis V.,
Aydil E. , Photovoltaic Manufacturing: Present Status and Future Prospects, J. Vac. Sci.
Technolo. A 29(3), 030801-16, 2011.
[4] Markvart T., Solar Electricity, 2nd edition, Wiley, 2000
[5] Honsberg, C. B., R. Corkish, and S. P. Bremner, "A New Generalized Detailed Balance
Formulation to Calculate Solar Cell Efficiency Limits", 17th European Photovoltaic Solar
Energy Conference, pp. 22-26, 2001.
[6] Swanson, R., "Approaching the 29% limit efficiency of silicon solar cells", Thirty-First IEEE
Photovoltaic Specialists Conference, Lake buena Vista, FL, USA, 01/2005, pp. 889-94,
2005.
[7] Kim, H.Y., Preparation of Polysilicon for Solar Cells. Korean Chemical
Engineering Research, 2008. 46(1): p. 37-49.
[8] Filtvedt, W.O., et al., Development of fluidized bed reactors for silicon
production. Solar Energy Materials & Solar Cells, 2010. 94: p. 1980-
1995.
[9] Solar&Energy. Polysilicon production method. 2011 [cited 2011
November 8]; Available from:
http://www.solarnenergy.com/eng/info/column.php.
[10] Alsema, E. and M.J. de Wild-Scholten, Reduction of the Environmental
Impacts in Crystalline Silicon Module Manufacturing, in 22nd European
Solar Energy Conference. 2007: Milan, Italy.
[11] REC. REC's Lluidized Bed Reactor (FBR) Process. 2011 [cited 2011
November 8]; Available from: http://www.recgroup.com/en/tech/FBR/.

50
[12] Frost & Sullivan, Assessment of Polysilicon Global Market and it's
Fabrication Technology Landscape. 2010.
http://www.scribd.com/doc/56217734/Frost-Sullivan-Polysilicon-2010.
[13] M. A. Green, K. Emery, Y. Hishikawa, and W. Warta, "Solar cell efficiency
tables (version 36)," Prog. Photovolt: Res. Appl. 18, 346-352, (2010).
[14] D. S. Kim, V. Yelundur, K. Nakayashiki, B. Rounsaville, V.
Meemongkolkiat, A. M. Gabor, and A. Rohatgi, "Ribbon Si solar cells
with efficiencies over 18% by hydrogenation of defects," Solar Energy
Mater. Solar Cells 90, 1227-1240, (2006).
[15] S. Solar, "EFG-Solar Cell Product Specification Sheet," in May 1st, ed,
2010.
[16] M. A. Green, "Third generation photovoltaics: Ultra-high conversion
efficiency at low cost," Prog. Photovolt: Res. Appl. 9, 123-135, (2001).
[17] W. Shockley and H. J. Queisser, "Detailed balance limit of efficiency of
p-n junction solar cells," J. Appl. Phys. 32, 510-519, (1961).
[18] G. del Coso, I. Tobias, C. Canizo, and A. Luque, "Temperature
homogeneity of polysilicon rods in a Siemens reactor," J. Cryst. Growth
299, 165-170, (2007).
[19] S. Pizzini, "Towards solar grade silicon: Challenges and benefits for low
cost photovoltaics," Solar Energy Mater. Solar Cells 94, 1528-1533,
(2010).
[20] T. F. Ciszek, "Edge-defined, film-fed growth (EFG) of silicon ribbons,"
Mater. Res. Bull. 7, 731-737, (1972).
[21] T. F. Ciszek, J. L. Hurd, and M. Schietzelt, "Filament materials for edge-
supported pulling of silicon sheet crystals," J. Electrochem. Soc. 129,
2838-2843, (1982).
[22] J. P. Kalejs, "Modeling contributions in commercialization of silicon
ribbon growth from the melt," J. Cryst. Growth 230, 10-21, (2001).
[23] A. G. Aberle, "Overview on SiN surface passivation of crystalline silicon
solar cells," Solar Energy Mater. Solar Cells 65, 239-248, (2001).

51
[24] G. P. Smestad, Optoelectronics of Solar Cells. Bellingham, WA: SPIE
Press, 2002.
[25] M. Wolf, "Proceedings of the 14th IEEE PVSC," San Diego, 1980, p. 674.
[26] M. J. Kerr, J. Schmidt, and A. Cuevas, "Lifetime and efficiency limits of
crystalline silicon solar cells," in Proceedings of the 29th IEEE PVSC,
New Orleans, 2002, p. 250.
[27] M. A. Green, "Limiting efficiency of bulk and thin-film silicon solar cells
in the presence of surface recombination," Progress Photovolt: Res.
Appl. 7, 327-330, (1999).
[28] A. Luque, A. Cuevas, and J. M. Ruiz, "DOUBLE-SIDED N+-P-N+ SOLAR-
CELL FOR BIFACIAL CONCENTRATION," Solar Cells 2, 151-166, (1980).
[29] K. A. Munzer, K. T. Holdermann, R. E. Schlosser, and S. Sterk, "Thin
monocrystalline silicon solar cells," IEEE Trans. Electron Dev. 46, 2055-
2061, (1999).
[30] V. Depauw, I. Gordon, G. Beaucarne, J. Poortmans, R. Mertens, and J. P.
Celis, "Large-area monocrystalline silicon thin films by annealing of
macroporous arrays: Understanding and tackling defects in the
material," J. Appl. Phys. 106, 033516, (2009).
[31] F. Dross, J. Robbelein, B. Vandevelde, E. Van Kerschaver, I. Gordon, G.
Beaucarne, and J. Poortmans, "Stress-induced large-area lift-off of
crystalline Si films," Appl. Phys. a-Mater.Sci. Process. 89, 149-152,
(2007).
[32] C. Podewils, "A proton axe for thin silicon," Photon International May,
116-120, (2009).
[33] M. J. Kerr and A. Cuevas, "Very low bulk and surface recombination in
oxidized silicon wafers," Semicond. Science Technol. 17, 35-38, (2002).
[34] J. Schmidt, A. Merkle, R. Brendel, B. Hoex, M. C. M. van de Sanden, and
W. M. M. Kessels, "Surface passivation of high-efficiency silicon solar
cells by atomic-layer-deposited Al2O3," Prog. Photovolt: Res. Appl. 16,
461-466, (2008).

52
[35] D. Shir, J. Yoon, D. Chanda, J. H. Ryu, and J. A. Rogers, "Performance of
ultrathin silicon solar microcells with nanostructures of relief formed by
soft Imprint lithography for broad band absorption enhancement,"
Nano Letters 10, 3041-3046, (2010).
[36] S. Pillai and M. A. Green, "Plasmonics for photovoltaic applications,"
Solar Energy Mater. Solar Cells 94, 1481-1486, (2010).
[37] P. C. Rowlette and C. A. Wolden, "Digital control of SiO2-TiO2 mixed-
metal oxides by pulsed PECVD," ACS Appl. Mater. Interfaces 1, 2586-
2591, (2009).
[38] S. Eglash, "Competition improves silicon-based solar cells," Laser
Focus World 45, 39-44, (2009).
[39] P. Engelharet, S. Hermann, T. Neubere, H. Plagwitz, R. Grischke, R.
Meyd, U. Klug, A. Schoonderbeek, U. Stute, and R. Brendel, "Laser
ablation of SiO2 for locally contacted si solar cells with ultra-short
pulses," Prog. Photovolt. 15, 521-527, (2007).
[40] D. A. Cusano, "CdTe solar cells and photovoltaic heterojunctions in II-VI
compounds," Solid-State Electron. 6, 217-218, (1963).
[41] C. Ferekides and J. Britt, "Thin film CdS/CdTe solar cell with 15.8%
efficiency," Appl. Phys. Lett. 62, 2851-2852, (1993).
[42] R. F. Brebrick and A. J. Strauss, "Partial pressures and gibbs free energy
of formation for congruently subliming CdTe," J. Phys. Chem. Solid 25,
1441-1445, (1964).
[43] J. M. Kestner, S. McElvain, S. Kelly, L. M. Woods, T. R. Ohno, and C. A.
Wolden, "An experimental and modeling analysis of vapor transport
deposition of cadmium telluride," Solar Energy Mat. Solar Cells 83, 55-
65 (2004).
[44] P. V. Meyers and S. P. Albright, "Technical and economic opportunities
for CdTe PV at the turn of the millennium," Progress Photovolt: Res.
Appl. 8, 161-169, (2000).
[45] X. Wu, "High-efficiency polycrystalline CdTe thin-film solar cells," Solar
Energy 77, 803-814, (2004).

53
[46] A. D. Compaan, A. Gupta, S. Lee, S. Wang, and J. Drayton, "High
efficiency, magnetron sputtered CdS/CdTe solar cells," Solar Energy 77,
815-822, (2004).
[47] X. Wu, S. Asher, D. H. Levi, D. E. King, Y. Yan, T. A. Gessert, and P.
Sheldon, "Interdiffusion of CdS and Zn2SnO4 layers and its application
in CdS/CdTe polycrystalline thin-film solar cells," J. Appl. Phys. 89,
4564-4569, (2001).
[48] N. Naghavi, D. Abou-Ras, N. Allsop, N. Barreau, S. Bucheler, A. Ennaoui,
C. H. Fischer, C. Guillen, D. Hariskos, J. Herrero, R. Klenk, K. Kushiya, D.
Lincot, R. Menner, T. Nakada, C. Platzer-Bjorkman, S. Spiering, A. N.
Tiwari, and T. Torndahl, "Buffer layers and transparent conducting
oxides for chalcopyrite Cu(In,Ga)(S,Se)(2) based thin film
photovoltaics: present status and current developments," Prog.
Photovoltaics 18, 411-433, (2010).
[49] M. Carmody, S. Mallick, J. Margetis, R. Kodama, T. Biegala, D. Xu, P.
Bechmann, J. W. Garland, and S. Sivananthan, "Single-crystal II-VI on Si
single-junction and tandem solar cells," Appl. Phys. Lett. 96, 153502-3,
(2010).
[50] W. K. Metzger, I. L. Repins, and M. A. Contreras, "Long lifetimes in high-
efficiency Cu(In,Ga)Se2 solar cells," Appl. Phys. Lett. 93, 022110-3,
(2008).
[51] J. Sites and J. Pan, "Strategies to increase CdTe solar-cell voltage," Thin
Solid Films 515, 6099-6102, (2007).
[52] M. Beck, "Personal Communication," N. P. Workshop, Ed., ed, 2010.
[53] D. Cahen and R. Noufi, "SURFACE PASSIVATION OF POLYCRYSTALLINE,
CHALCOGENIDE BASED PHOTOVOLTAIC CELLS," Solar Cells 30, 53-59,
(1991).
[54] H. R. Moutinho, M. M. Al-Jassim, D. H. Levi, P. C. Dippo, and L. L.
Kazmerski, "Effects of CdCl2 treatment on the recrystallization and
electro-optical properties of CdTe thin films," J. Vac. Sci. Technol. A 16,
1251-1257, (1998).

54
[55] D. Grecu and A. D. Compaan, "Photoluminescence study of Cu diffusion
and electromigration in CdTe," Appl. Phys. Lett. 75, 361-363, (1999).
[56] Y. Yan, D. Albin, and M. M. Al-Jassim, "Do grain boundaries assist S
diffusion in polycrystalline CdS/CdTe heterojunctions?," Appl. Phys.
Lett. 78, 171-173, (2001).
[57] S. Erra, C. Shivakumar, H. Zhao, K. Barri, D. L. Morel, and C. S.
Ferekides, "An effective method of Cu incorporation in CdTe solar cells
for improved stability," Thin Solid Films 515, 5833-5836, (2007).
[58] T. A. Gessert, S. Asher, S. Johnston, M. Young, P. Dippo, and C. Corwine,
"Analysis of CdS/CdTe devices incorporating a ZnTe:Cu/Ti Contact,"
Thin Solid Films 515, 6103-6106, (2007).
[59] V. M. Fthenakis, M. Fuhrmann, J. Heiser, A. Lanzirotti, J. Fitts, and W.
Wang, "Emissions and encapsulation of cadmium in CdTe PV modules
during fires," Progress Photovolt: Res. Appl. 13, 713-723, (2005).
[60] H. Steinberger, "Health, safety and environmental risks from the
operation of CdTe and CIS thin-film modules," Progress Photovolt: Res.
Appl. 6, 99-103, (1998).
[61] D. E. Carlson and C. R. Wronski, "Amorphous silicon solar cell," Appl.
Phys. Lett. 28, 671-673, (1976).
[62] A. H. Mahan, Y. Xu, E. Iwaniczko, D. L. Williamson, B. P. Nelson, and Q.
Wang, "Amorphous silicon films and solar cells deposited by HWCVD at
ultra-high deposition rates," J. Non-Crystal. Solids 299-302, 2-8,
(2002).
[63] A. V. Shah, H. Schade, M. Vanecek, J. Meier, E. Vallat-Sauvain, N.
Wyrsch, U. Kroll, C. Droz, and J. Bailat, "Thin-film silicon solar cell
technology," Prog. Photovolt: Res. Appl. 12, 113-142, (2004).
[64] D. L. Staebler and C. R. Wronski, "Reversible conductivity charge in
discharge-produced amorphous Si," Appl. Phys. Lett. 31, 292-294,
(1977).

55
[65] S. Veprek and V. Marecek, "The preparation of thin layers of Ge and Si
by chemical hydrogen plasma transport," Solid-State Electronics 11,
683-684, (1968).
[66] A. Shah, J. Meier, E. Vallat-Sauvain, C. Droz, U. Kroll, N. Wyrsch, J.
Guillet, and U. Graf, "Microcrystalline silicon and micromorph tandem
solar cells," Thin Solid Films 403-404, 179-187, (2002).
[67] J. Yang, A. Banerjee, and S. Guha, "Triple-junction amorphous silicon
alloy solar cell with 14.6% initial and 13.0% stable conversion
efficiencies," Appl. Phys. Lett. 70, 2975-2977, (1997).
[68] M. Taguchi, H. Sakata, Y. Yoshimine, E. Maruyama, A. Terakawa, M.
Tanaka, and S. Kiyama, "An approach for the higher efficiency in the
HIT cells," in Conference Record of the 31st IEEE Photovoltaic
Specialists Conference, 2005, pp. 866-871.
[69] A. Poruba, A. Fejfar, Z. Remes, J. Springer, M. Vanecek, J. Kocka, J.
Meier, P. Torres, and A. Shah, "Optical absorption and light scattering in
microcrystalline silicon thin films and solar cells," J. Appl. Phys. 88,
148-160, (2000).
[70] T. Roschek, T. Repmann, J. Muller, B. Rech, and H. Wagner,
"Comprehensive study of microcrystalline silicon solar cells deposited
at high rate using 13.56 MHz plasma-enhanced chemical vapor
deposition," J. Vac. Sci. Technol. A 20, 492-498, (2002).
[71] M. Fukawa, S. Suzuki, L. H. Guo, M. Kondo, and A. Matsuda, "High rate
growth of microcrystalline silicon using a high-pressure depletion
method with VHF plasma," Solar Energy Mater. Solar Cells 66, 217-223,
(2001).
[72] Y. Mai, S. Klein, R. Carius, J. Wolff, A. Lambertz, F. Finger, and X. Geng,
"Microcrystalline silicon solar cells deposited at high rates," J. Appl.
Phys. 97, 2005).
[73] J. Rudiger, H. Brechtel, A. Kottwitz, J. Kuske, and U. Stephan, "VHF
plasma processing for in-line deposition systems," Thin Solid Films
427, 16-20, (2003).

56
[74] L. L. Kazmerski, F. R. White, and G. K. Morgan, "Thin-film CuInSe2/CdS
heterojunction solar cells," Appl. Phys. Lett. 29, 268-270, (1976).
[75] J. H. L. Stolt, J. Kessler, M. Ruckh, K.-O. Velthaus, ans H.-W. Schock,
"ZnO/CdS/CuInSe2 thin-film solar cells with improved performance,"
Appl. Phys. Lett. 62, 597-599, (1993).
[76] A. M. Gabor, J. R. Tuttle, D. S. Albin, M. A. Contreras, R. Noufi, and A. M.
Herman, "High-efficiency CuInxGa1-xSe2 solar cells made from (InxGa1-
) Se3 precursor films," Appl. Phys. Lett. 65, 198-200, (1994).
x 2

[77] A. N. Tiwari, D. Lincot, and M. Contreras, "The time for CIGS," Progress
Photovolt: Res. Appl. 18, 389-389, (2010).
[78] S. Niki, M. Contreras, I. Repins, M. Powalla, K. Kushiya, S. Ishizuka, and
K. Matsubara, "CIGS absorbers and processes," Prog. Photovolt: Res.
Appl. 18, 453-466, (2010).
[79] M. A. Contreras, B. Egaas, K. Ramanathan, J. Hiltner, F. Hasoon, and R.
Noufi, "Progress toward 20% Efficiency in Cu(In,Ga)Se2 polyscrystalline
Thin film solar cell," Prog. Photovolt: Res. Appl. 7, 311, (1999).
[80] K. Ramanathan, G. Teeter, J. C. Keane, and R. Noufi, "Properties of high-
efficiency CuInGaSe2 thin film solar cells," Thin Solid Films 480, 499-
502, (2005).
[81] R. Scheer, A. Prez-Rodrguez, and W. K. Metzger, "Advanced diagnostic
and control methods of processes and layers in CIGS solar cells and
modules," Prog. Photovolt: Res. Appl. 18, 467-480, (2010).
[82] C. J. Hibberd, E. Chassaing, W. Liu, D. B. Mitzi, D. Lincot, and A. N.
Tiwari, "Non-vacuum methods for formation of Cu(In, Ga)(Se, S)2 thin
film photovoltaic absorbers," Prog. Photovolt: Res. Appl. 18, 434-452,
(2010).
[83] R. N. Bhattacharya, "SOLUTION GROWTH AND ELECTRODEPOSITED
CULNSE2 THIN-FILMS," J. Electrochem. Soc. 130, 2040-2042, (1983).
[84] D. Lincot, J. F. Guillemoles, S. Taunier, D. Guimard, J. Sicx-Kurdi, A.
Chaumont, O. Roussel, O. Ramdani, C. Hubert, J. P. Fauvarque, N.
Bodereau, L. Parissi, P. Panheleux, P. Fanouillere, N. Naghavi, P. P.

57
 Grand, M. Benfarah, P. Mogensen, and O. Kerrec, "Chalcopyrite thin film
solar cells by electrodeposition," Solar Energy 77, 725-737, (2004).
[85] M. G. Panthani, V. Akhavan, B. Goodfellow, J. P. Schmidtke, L. Dunn, A.
Dodabalapur, P. F. Barbara, and B. A. Korgel, "Synthesis of CuInS2,
CuInSe2, and Cu(InxGa1-x)Se2 (CIGS) Nanocrystal "Inks" for Printable
Photovoltaics," J. Am. Chem. Soc. 130, 16770-16777, (2008).
[86] M. Kaelin, D. Rudmann, F. Kurdesau, T. Meyer, H. Zogg, and A. N. Tiwari,
"CIS and CIGS layers from selenized nanoparticle precursors," Thin
Solid Films 431-432, 58-62, (2003).
[87] V. K. Kapur, B. M. Basol, and E. S. Tseng, "LOW-COST METHODS FOR
THE PRODUCTION OF SEMICONDUCTOR-FILMS FOR CUINSE2/CDS
SOLAR-CELLS," Solar Cells 21, 65-72, (1987).
[88] V. K. Kapur, A. Bansal, P. Le, and O. I. Asensio, "Non-vacuum processing
of CuIn1-xGaxSe2 solar cells on rigid and flexible substrates using
nanoparticle precursor inks," Thin Solid Films 431-432, 53-57, (2003).
[89] J. v. Duren, D. Jackrel, F. Jacob, C. Leidholm, A. Pudov, M. Robinson, and
Y. Roussillon, "The next generation in thinfilm photovoltaic process
technology," in Conference Record of the Seventeenth International
Photovoltaic Science and Engineering Conference, Fukuoka, Japan,
2007.
[90] D. B. Mitzi, "Solution Processing of Chalcogenide Semiconductors via
Dimensional Reduction," Adv. Mater. 21, 3141-3158, (2009).
[91] W. Liu, D. B. Mitzi, M. Yuan, A. J. Kellock, S. J. Chey, and O. Gunawan,
"12% Efficiency CuIn(Se,S)2 Photovoltaic Device Prepared Using a
Hydrazine Solution Process," Chem. Mater. 22, 1010-1014, (2010).
[92] S. Siebentritt, M. Igalson, C. Persson, and S. Lany, "The electronic
structure of chalcopyritesbands, point defects and grain boundaries,"
Prog. Photovolt: Res. Appl. 18, 390-410, (2010).
[93] G. L. Graff, R. E. Williford, and P. E. Burrows, "Mechanisms of vapor
permeation through multilayer barrier films: Lag time versus
equilibrium permeation," J. Appl. Phys. 96, 1840-1849, (2004).

58
[94] C. Wadia, A. P. Alivisatos, and D. M. Kammen, "Materials availability
expands the opportunity for large-scale photovoltaics deployment,"
Environ. Sci. Technol. 43, 2072-2077, (2009).
[95] V. Fthenakis, "Sustainability of photovoltaics: The case for thin-film
solar cells," Renew. Sustain. Energy Rev. 13, 2746-2750, (2009).
[96] K. Zweibel, "The impact of tellurium supply on cadmium telluride
photovoltaics," Science 328, 699-701, (2010).
[97] K. Zweibel, "Thin film PV manufacturing: Materials costs and their
optimization," Solar Energy Mater. Solar Cells 63, 375-386, (2000).
[98] M. Raugei, S. Bargigli, and S. Ulgiati, "Life cycle assessment and
energy pay-back time of advanced photovoltaic modules: CdTe and CIS
compared to poly-Si," Energy 32, 1310-1318, (2007).
[99] B. O'Regan and M. Gratzel, "A low-cost, high-efficiency solar cell based
on dye-sensitized colloidal TiO2 films," Nature 353, 737-740, (1991).
[100] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, and H. Pettersson, "Dye-
sensitized Solar Cells," Chem. Rev. 110, 6595-6663, (2010).
[101] M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E.
Mueller, P. Liska, N. Vlachopoulos, and M. Graetzel, "Conversion of light
to electricity by cis-X2bis(2,2'-bipyridyl-4,4'-dicarboxylate)ruthenium(II)
charge-transfer sensitizers (X = Cl-, Br-, I-, CN-, and SCN-) on
nanocrystalline titanium dioxide electrodes," J. Am. Chem. Soc. 115,
6382-6390, (1993).
[102] G. E. Tulloch, "Light and energy--dye solar cells for the 21st
century," J. Photochem. Photobiol. A: Chem. 164, 209-219, (2004).
[103] T. W. Hamann, R. A. Jensen, A. B. F. Martinson, H. Van Ryswyk,
and J. T. Hupp, "Advancing beyond current generation dye-sensitized
solar cells," Energy Environ. Sci. 1, 66-78, (2008).
[104] C. J. Brabec, S. Gowrisanker, J. J. M. Halls, D. Laird, S. J. Jia, and S.
P. Williams, "Polymer-Fullerene Bulk-Heterojunction Solar Cells," Adv.
Mater. 22, 3839-3856, (2010).
[105] Heliatek. (12/1/10). Available: http://www.heliatek.com/news-19

59
[106] Konarka. (12/1/10). Available:
http://www.konarka.com/index.php/newsroom/press-release-list/
[107] C. W. Tang, "Two-layer organic photovoltaic cell," Appl. Phys. Lett.
48, 183-185, (1986).
[108] M. S. White, D. C. Olson, S. E. Shaheen, N. Kopidakis, and D. S.
Ginley, "Inverted bulk-heterojunction organic photovoltaic device using
a solution-derived ZnO underlayer," Appl. Phys. Lett. 89, 143517-3,
(2006).
[109] M. D. Irwin, D. B. Buchholz, A. W. Hains, R. P. H. Chang, and T. J.
Marks, "p-type semiconducting nickel oxide as an efficiency-enhancing
anode interfacial layer in polymer bulk-heterojunction solar cells,"
Proc. Nat. Acad. Sci. 105, 2783-2787, (2008).
[110] C. Tao, S. Ruan, G. Xie, X. Kong, L. Shen, F. Meng, C. Liu, X.
Zhang, W. Dong, and W. Chen, "Role of tungsten oxide in inverted
polymer solar cells," Appl. Phys. Lett. 94, 043311-3, (2009).
[111] M. Riede and et al., "Small-molecule solar cellsstatus and
perspectives," Nanotechnology 19, 424001, (2008).
[112] H. Spanggaard and F. C. Krebs, "A brief history of the
development of organic and polymeric photovoltaics," Solar Energy
Mater. Solar Cells 83, 125-146, (2004).
[113] E. Bundgaard and F. C. Krebs, "Low band gap polymers for
organic photovoltaics," Solar Energy Mater. Solar Cells 91, 954-985,
(2007).
[114] R. Kroon, M. Lenes, J. C. Hummelen, P. W. M. Blom, and B. d. Boer,
"Small Bandgap Polymers for Organic Solar Cells (Polymer Material
Development in the Last 5 Years," Polym. Rev. 48, 531-581 (2008).
[115] M. J. Currie, J. K. Mapel, T. D. Heidel, S. Goffri, and M. A. Baldo,
"High-efficiency organic solar concentrators for photovoltaics," Science
321, 226-228, (2008).
[116] D. Veldman, S. C. J. Meskers, and R. A. J. Janssen, "The Energy of
Charge-Transfer States in Electron DonorAcceptor Blends: Insight into

60
 the Energy Losses in Organic Solar Cells," Adv. Funct. Mater. 19, 1939-
1948, (2009).
[117] T. Ameri, G. Dennler, C. Lungenschmied, and C. J. Brabec,
"Organic tandem solar cells: A review," Energy Environ. Sci. 2, 347-
363, (2009).
[118] A. Anctil, C. Babbit, B. Landi, and R. P. Raffaelle, "Life-Cycle
Assessment of Organic Solar Cell Technology," in 35th IEEE
Photovoltaic Specialist Conference, Honolulu, Hawaii, 2010.
[119] J. Kalowekamo and E. Baker, "Estimating the manufacturing cost
of purely organic solar cells," Solar Energy 83, 1224-1231, (2009).
[120] R. Tipnis, J. Bernkopf, S. J. Jia, J. Krieg, S. Li, M. Storch, and D.
Laird, "Large-area organic photovoltaic module-Fabrication and
performance," Solar Energy Materials and Solar Cells 93, 442-446,
(2009).
[121] V. Shrotriya, "Organic photovoltaics: Polymer power," Nat.
Photon. 3, 447-449, (2009).
[122] M. Jrgensen, K. Norrman, and F. C. Krebs, "Stability/degradation
of polymer solar cells," Solar Energy Mater. Solar Cells 92, 686-714,
(2008).
[123] A. O. Musa, T. Akomolafe, and M. J. Carter, "Production of cuprous
oxide, a solar cell material, by thermal oxidation and a study of its
physical and electrical properties," Solar Energy Mater. Solar Cells 51,
305-316, (1998).
[124] A. Ennaoui, S. Fiechter, C. Pettenkofer, N. Alonsovante, K. Buker,
M. Bronold, C. Hopfner, and H. Tributsch, Solar Energy Mater. Solar
Cells 29, 289-370, (1993).
[125] H. Katagiri, "Cu2ZnSnS4 thin film solar cells," Thin Solid Films
480, 426-432, (2005).
[126] H. Katagiri, K. Saitoh, T. Washio, H. Shinohara, T. Kurumadani,
and S. Miyajima, "Development of thin film solar cell based on

61
 Cu2ZnSnS4 thin films," Solar Energy Mater. Solar Cells 65, 141-148,
(2001).
[127] K. Oishi, G. Saito, K. Ebina, M. Nagahashi, K. Jimbo, W. S. Maw, H.
Katagiri, M. Yamazaki, H. Araki, and A. Takeuchi, "Growth of Cu2ZnSnS4
thin films on Si (100) substrates by multisource evaporation," Thin
Solid Films 517, 1449-1452, (2008).
[128] Q. J. Guo, H. W. Hillhouse, and R. Agrawal, "Synthesis of
Cu2ZnSnS4 Nanocrystal Ink and Its Use for Solar Cells," J. Am. Chem.
Soc. 131, 11672-+, (2009).
[129] C. Steinhagen, M. G. Panthani, V. Akhavan, B. Goodfellow, B. Koo,
and B. A. Korgel, "Synthesis of Cu2ZnSnS4 Nanocrystals for Use in Low-
Cost Photovoltaics," J. Am. Chem. Soc. 131, 12554-+, (2009).
[130] T. K. Todorov, K. B. Reuter, and D. B. Mitzi, "High-efficiency solar
cell with earth-abundant liquid-processed absorber," Adv. Mater. 22,
E156-E159, (2010).
[131] "Basic Research Needs for Solar Energy Utilization," Department
of Energy Report2005.
[132] R. D. Schaller and V. I. Klimov, "High efficiency carrier
multiplication in PbSe nanocrystals: Implications for solar energy
conversion," Phys.Rev. Lett. 92, 186601, (2004).
[133] A. J. Nozik, "Multiple exciton generation in semiconductor
quantum dots," Chem. Phys. Lett. 457, 3-11, (2008).
[134] W. A. Tisdale, K. J. Williams, B. A. Timp, D. J. Norris, E. S. Aydil,
and X.-Y. Zhu, "Hot-electron transfer from semiconductor nanocrystals,"
Science 328, 1543-1547, (2010).
[135] J. B. Sambur, T. Novet, and B. A. Parkinson, "Multiple exciton
collection in a sensitized photovoltaic system," Science 330, 63-66,
(2010).
[136] J. M. Luther, M. Law, M. C. Beard, Q. Song, M. O. Reese, R. J.
Ellingson, and A. J. Nozik, "Schottky solar cells based on colloidal
nanocrystal films," Nano Letters 8, 3488-3492, (2008).

62
[137] Y. Wu, C. Wadia, W. L. Ma, B. Sadtler, and A. P. Alivisatos,
"Synthesis and photovoltaic application of copper(I) sulfide
nanocrystals," Nano Letters 8, 2551-2555, (2008).
[138] J. F. Geisz, S. Kurtz, M. W. Wanlass, J. S. Ward, A. Duda, D. J.
Friedman, J. M. Olson, W. E. McMahon, T. E. Moriarty, and J. T. Kiehl,
"High-efficiency GaInP/GaAs/InGaAs triple-junction solar cells grown
inverted with a metamorphic bottom junction," Appl. Phys. Lett. 91,
2007).
[139] J. D. Beach and B. E. McCandless, "Materials challenges for CdTe
and CuInSe2 photovoltaics," MRS Bulletin 32, 225-229, (2007).
[140] Fthenakis V.M., Sustainability of photovoltaics: The case for thin-
film solar cells, Renewable and Sustainable Energy Reviews, 13, 2746-
2750, 2009.
[141] Fthenakis V., End-of Life Management and Recycling of PV
Modules, Energy Policy, 28, 1051-1058, 2000.
[142] Fthenakis V., Energy Return on Investment (ROI) tracks efficiency
gains, Solar Today, 24-26, June 2012.
[143] Fthenakis V.M. et al., Methodology Guidelines on Life Cycle
Assessment of Photovoltaics Electricity, PVPS Task 12, International
Energy Agency. 2009.
144] Fthenakis V.M. et al., Life Cycle Inventories and Analyses of
Photovoltaic Systems: 2011 Status, PVPS Task 12, 2nd Draft,
International Energy Agency, July 2011
[145] Fthenakis V.M., Kim H.C. and Alsema E., Emissions from photovoltaic
life cycles, Environ. Sci. Technol., 42 (6), 21682174, 2008.[
[146] Fthenakis V., Raugei M., Held M., Kim H. C., Krones J. An update of
Energy Payback Times and Greenhouse Gas emissions in the Life Cycle
of Photovoltaics, Proceedings 24th European Photovoltaic Solar Energy
Conference, Hamburg, Germany, 21-25 September 2009.
[147] Nikolakakis T. and Fthenakis V. The Optimum Mix of Electricity
from Wind- and Solar-Sources in Conventional Power Systems:

63
 Evaluating the Case for New York State, Energy Policy, 39(11), 6972-
6980, 2011.
[148] Denholm, P., Margolis, R., 2007. Evaluating the limits of solar
photovoltaics (PV) in traditional electric power systems. Energy Policy
35, 28522861.

64