effect.

The most iconic form of solar power, the solar cell, consumes
no fuel, has a wide power-handling capacity, has a high power-to-
weight ratio, produces no pollution, and is more and more often made
from inexpensive polycrystalline or amorphous materials, with even
polymer-based solar cells being introduced. The operating principle of
a solar cell was covered in Volume I, Chapter 4 (Figure 4.42), and in
Table 9.5 we present a short historical overview of some important
solar cell related milestones since the original observation of the
photoeffect by Antoine Becquerel in 1839. As with batteries, progress
in improving solar cell efficiency has been rather slow (there is no
Moore’s law curve for solar cells, either). Several important measures
are used to characterize solar cells, the most obvious one being the
total amount of electrical power produced for a given amount of solar
power input. Expressed as a percentage, this is known as the solar
conversion efficiency. Electrical power is the product of current and
voltage, so the maximum values for these measurements are
important as well: Jsc (short circuit current) and Voc (open circuit
voltage), respectively. Finally, in order to understand the underlying
physics, the quantum efficiency is used to compare the chance that
one photon (of a particular energy) will create one electron. A solar
cell must be capable of producing electricity for at least twenty years
without a significant decrease in efficiency, so lifespan is another
crucial characteristic. Because of the soaring oil prices, solar energy is
a hot topic again, as it was under President Carter (even Republicans
now concede that cheap oil is not a US birthright). The photovoltaic
cell is currently the fastest-growing form of alternative energy,
increasing by 50% a year, and where a kWh of PV electricity cost 50
cents in 1995, it fell to 20 cents in 2005. BBC Research, which charts
technology markets, expects the global solar market to grow to $32

earth’s surface, the average solar energy is ∼4 × 1024 J per year, and
billion by 2012, with thin-film solar cells expanding 45% a year. At the

the 2001 global energy use was ∼4 × 1020 J per year (increasing
∼2% annually). In the United States alone, the average power
requirement is 3.3 TW. With a 10% efficient solar cell, this means that
we would need 1.7% of the land area of the United States devoted to
PV to quench this thirst; this roughly corresponds to the area occupied
by the interstate highways. As we saw in Volume I, Chapter 4, a
traditional semiconductor solar cell is made from an n-type and a p-
type semiconductor put in contact with each other. The result is a
region at the interface, the p-n junction, where charge carriers are
depleted and/or accumulated on each side of the interface. The
transfer of charge carriers produces a potential barrier, typically 0.6–
0.7 V high. In the case of silicon (bandgap = 1.1 eV), sunlight provides
enough energy to promote electrons out of the valence band into the
conduction band, and when a load is placed across the cell, these
electrons flow from the p-type to the n-type material, move through
the external circuit, and then stream back into the p-type material,
where they recombine with the valence-band holes they left behind.
How sunlight creates an TABLE 9.5 Milestones in Solar Cell
Development 1839: Discovery of photovoltaic effect between two
electrodes in a liquid (Antoine Cesar Becquerel) 1876: Photovoltaic
effect in Se (Heinrich Hertz) 1883: First solar electric module (Charles
Fritts) 1906: Pochettino studies the photoconductivity of anthracene
1930: Research on Cu2O/Cu solar cell 1941: Patent on Si solar cell (R.
Ohl in Bell Labs) 1954: Crystalline Si solar cell (Bell Labs); 4%
efficiency 1958: Solar panels as backup and power assist in space
(Vanguard 1); 5 mW 1970: Zhores Alferov and his team in the USSR
created the first GaAs heterostructure solar cells 1973: Oil crisis—
solar energy is hot, subsequent Republican United States president
kills Carter’s efforts in favor of oil 1980: Solar cell using CdTe, CuInSe2
,TiO2 , etc. 1991: Liquid-junction solar cells with photosensitized
anode invented by Michael Grätzel and Brian O’Regan at the École
Polytechnique Fédérale de Lausanne, known as Grätzel cells 1995: Yu
and Hall make the first bulk polymer/polymer heterojunction
photovoltaic cell 2000: Peters/van Hall used oligomer-C60 diads/triads
as the active material in photovoltaic cells (see Figure 3.15, for an
example coming from the Stoddard-Zinc group) 2001: Schmidt-Mende
made a self-organized liquid crystalline solar cell of
hexabenzocoronene and perylene 498 From MEMS to Bio-NEMS:
Manufacturing Techniques and Applications electrical current in a
solar cell is detailed a bit further in Figure 9.23. If the incoming photon
energy Eγ < Eg , then the photon is not absorbed; with Eγ > Eg , the
photon is absorbed and the energy Eγ – Eg is lost via thermalization.
Additional losses occur through radiative recombination, trap states,
and Auger recombination. If the carrier lifetime is less than the
recombination lifetime then the charge is swept away by the inbuilt
electric field (E). This depends on carrier mobility, defect density,
carrier diffusion length, and exciton energy. The conventional
approach to PVs, as sketched in Figure 9.23, comes with several
disadvantages. First, the semiconductor bandgap dictates that only
photons with the bandgap energy or more contribute to producing a
current. Higher-energy photons, at the blue end of the spectrum, have
more than enough energy to cross the bandgap, but much of it is
wasted as heat. A cascade or tandem cell can achieve a higher total
conversion efficiency by capturing a larger portion of the solar
spectrum. In a typical multijunction cell, individual cells with different
bandgaps are stacked on top of one another. The cells are stacked in
such a way that sunlight falls first on the material having the largest
bandgap. Photons not absorbed in the first cell are transmitted to the
second cell, which then absorbs the higher-energy portion of the
remaining solar radiation while remaining transparent to the lower-
energy photons. These selective absorption processes continue
through to the final cell, which has the smallest bandgap (Figure
9.24). The ideal bandgap combination for a tandem solar cell is
typically 1.75 and 1.1 eV. Such bandgap combinations can be attained
using lattice-matched III–V semiconductor material combinations. The
best monolithic combination is represented by the GaInP/GaAs
tandem cell, possessing an AM 0 efficiency of 26.9%. The ideal
efficiency for an infinite stack is 86.8%, and for a GaInP/GaAs/Ge
triple-junction cell an efficiency of 32% has been measured.
Multijunction solar cells dramatically improve efficiency, but these
cells are still very high-cost and suitable only for large commercial
deployments (combinations include GaAs/InGaP, Ge/InGaAsN,
InP/InAlAs, InP/InGaAs, and GaN/InGaN). In general terms, the types of
cells suitable for rooftop deployment have not changed significantly in
efficiency, although costs have dropped somewhat because of
increased supply. A higher light intensity can be focused on the solar
cells by the use of mirror and lens systems, perhaps even a system
that tracks the sun. This is perhaps too expensive a solution for
rooftop implementation but could be useful in MEMS devices. Another
problem associated with conventional solar cells is that in order to
capture enough photons in the p-type layer, that part has to be quite
“thick.” However, a thick p-type layer also increases the chance that
electrons and holes will recombine before reaching the p-n junction.
These effects combine to produce an upper limit of about 20% on the
efficiency of silicon solar cells. The current problem with solar cells is
not really the efficiency (20%), but rather the cost of production.
Though it has come Conduction band Valence band E ermalization
Hole swept away Electron swept away Photon Eg Distance Energy Eγ
FIGURE 9.23 Energy diagram of a solar cell. Decreasing bandgap
Sunlight Solar cells FIGURE 9.24 Multijunction, cascade, or tandem
solar cell. Power and Brains in Miniature Devices 499 down by a factor
of 100 in fifty years, it still has to go down by a factor of 2–5 to be
competitive with fossil plants. The best known solar cells today
remain based on Si, and they are either monocrystalline,
polycrystalline, or hydrogenated amorphous (α-Si:H) based.
Polycrystalline films are poured and are cheaper and simpler to make
than monocrystalline silicon that is Czochralski grown (see Volume I,
Chapter 4). The efficiency of polycrystalline cells is, however,
somewhat less than that of monocrystalline cells. Hydrogenated
amorphous Si (α-Si:H) cells are relatively cheap, but they come with
the lowest efficiency (see Table 9.6). Although its semiconducting
properties are inferior to those of single-crystal Si, we learned in
Chapter 5 that α-Si:H, manufactured by plasma-enhanced chemical
vapor deposition (PECVD) from silane, can be doped both n- and p-
type (in contrast to α-Si) (see Table 5.8 for state-of-the-art parameters
of PECVD α-Si:H). Exposure of α-Si:H to light increases
photoconductivity by four to six orders of magnitude, and the
absorption coefficient of α-Si:H is more than an order of magnitude
larger than that of single-crystal Si near the maximum solar photon
energy region of 500 nm. Accordingly, the optimum thickness of the
active layer in an α-Si:H solar cell can measure 1 µm, much smaller
than that of singlecrystal Si solar cells. In Figure 5.15, we compared
polycrystalline-based solar cells with ultrathin amorphous Si on glass
and ultrathin amorphous Si on flexible substrates. Solar cell
technology represents the most MEMScompatible technology for
power integration. Lee et al.31 note that hydrogenated amorphous
silicon solar cells are a particularly attractive means to realize an
onboard MEMS power supply, and they implemented a high-voltage
integrated solar cell array31 as an electrostatic MEMS power supply.
By interconnecting 100 individual solar cells in series, this team
measured an open circuit potential as high as 150 V under AM 1.5
conditions, a voltage high enough to drive onboard electrostatic
actuators. The conversion efficiency in that effort was only 0.2%,
however. Sakakibara et al.32 were able to generate more than 200 V
with a similar solar cell on an area of 1 cm2 and obtained a
conversion efficiency of 4.65%. In both cases, amorphous silicon was
used in a triplestacked PV structure generating up to 2.3 V per cell. To
obtain a very dense packing of array elements and to make the series
connection of the solar cells, the latter group used focused laser
beams for patterning electrodes and PV materials. Since solar light is
only intermittently available, electric storage elements need to be
implemented as well. Along this line, Kimura et al.33 fabricated a
miniature optoelectric transformer consisting of a p-n junction
photocell and a multilayer spiral coil transformer. III/V solar cells are
cells with a high efficiency, but they are made of very expensive
semiconducting materials. A single junction cell based on GaAs comes
with an efficiency of 25.1%. In 1970, Zhores Alferov and his team in
the USSR created the first GaAs heterostructure solar cells. Today, the
combination of GaAs with germanium and indium gallium phosphide
is the basis of a triple-junction solar cell that holds the record for solar
light conversion efficiency of over 32% and can operate also with light
as concentrated as 2000 Suns.* This type of solar cell powers the
robots Spirit and Opportunity, which are exploring the surface of Mars.
Newer solar cell materials include thin films of copper indium
diselenide (CuInSe2 ), copper indium gallium diselenide (CIGS), and
cadmium telluride (CdTe). Commercial leaders in this area are a firm
called First Solar, which uses cadmium telluride, and Miasolé, which
manufactures a thin-film PV cell based on an ultrathin layer of
photoactive CIGS on a stainless steel foil only 50 µm thick. The latter
can easily be used in PV modules or incorporated into building
materials such as membrane roofing. The efficiencies and bandgaps
of GaAs and two of * Where 1 Sun equals 1000 W/m² at the point of
arrival. One Sun is a unit of power flux. TABLE 9.6 Silicon-Based Solar
Cell Efficiency in % Type Efficiency in Laboratory Efficiency in
Production Monocrystalline 24 14–17 Polycrystalline 18 13–15
Amorphous 13 5–7 500 From MEMS to Bio-NEMS: Manufacturing
Techniques and Applications the newer solar materials are compared
with those of single-crystal and amorphous Si in Figure 9.25. One
development that made organic PVs an attractive option is the
availability of light-sensitive conjugated polymers: polymers with
alternating single and double carbon-carbon (sometimes
carbonnitrogen) bonds. In the 1970s, it was discovered that chemical
doping of conjugated polymers increased electronic conductivity
several orders of magnitude. Since that time, electronically
conducting materials based on conjugated polymers have found many
applications, including sensors, LEDs, and solar cells. Advantages of
polymer or organic PVs (OPVs) include the fact that they are easily
processed over large areas using spin-coating, doctor blade
techniques (wet-processing), evaporation through a mask (dry
processing), and printing (Figure 9.26). The usage of roll-to-roll
printing, especially, guarantees a favorable cost (see Figure 1.5). They
also have a low weight, exhibit mechanical flexibility and
transparency, and have a bandgap that can easily be tuned
chemically by incorporation of different functional groups. Efficiencies
stand at 5%, but there are no obstacles identified for 10%. The low
cost of plastic PV may enable solar power in applications where it was
previously uneconomic (e.g., solar tents, sails, packaging, etc.). For
rigid cells, ITO (Indium tin oxide)-coated glass usually serves as the
substrate for these organic cells. Figure 9.27 summarizes the best
research solar cell efficiencies for a broad range of materials from
universities, research institutes, and private companies. Besides PV
converters for solar light and laser 0.8 50 40 30 20 10 0 1.0 1.2 1.4
1.6 1.8 Bandgaps in eV CulnSe2 GaAs CdTe Si (amorph) Si Efficiency
in % FIGURE 9.25 Bandgaps and efficiencies of the most important
solar cell materials. The bars represent theoretical efficiencies, and
the smooth line represents practical efficiencies. FIGURE 9.26 Polymer
or organic photovoltaics. Efficiency (%) University of Maine Boeing
Boeing Boeing ARCO Boeing NREL Boeing Euro-CIS 1975 1980 1985
1990 1995 2000 NREL/ Spectrolab NREL NREL Japan Energy Spire No.
Carolina State University Multijunction concentrators ˆree-junction (2-
terminal, monolithic) Two-junction (2-terminal, monolithic) Crystalline
Si cells Single crystal Multicrystalline ˆin film technologies
Cu(ln,Ga)Se2 CdTe Amorphous Si:H (stabilized) Emerging PV Dye cells
Organic cells (various technologies) Varian RCA Solarex UNSW UNSW
ARCO UNSW UNSW UNSW Spire Stanford Westinghouse UNSW
Georgia Tech Georgia Tech Sharp NREL Spectrolab NREL Masushita
MonosolarKodak Kodak AMETEK Photon Energy University So. Florida
NREL NREL NREL Cu(In,Ga)Se2 14x concentration NREL United Solar
RCA RCA RCA RCARCA RCA Spectrolab Solarex 12 8 4 0 16 20 24 28
32 36 University of Lausanne University of Lausanne Siemens 2005
Kodak UCSB Cambridge Groningen University Linz Berkeley Princeton
University Linz United Solar FIGURE 9.27 Research solar cell
efficiencies. (From Art Nozik, DOE Solar Energy Workshop, 2005.)
Power and Brains in Miniature Devices 501 light, microwaves could be
used to power microsystems. In the latter, extremely small receivers
and converters need to be constructed. MEMS and NEMS in Solar Cells
Introduction Micro- and nanotechnologies are crucial to increasing the
efficiency of existing solar cells, and NEMS is enabling totally new
solar cell designs as well. MEMS techniques allow for the fabrication of
lenses (including Fresnel lenses), mirrors, photonic crystals, and
cavities and the texturing of solar cell surfaces to enhance the
interaction of incoming photons with the solar cell active junctions. As
an example of a nanoenhancement of existing solar cells, consider
the use of quantum dots (QDs) to replace dyes in luminescent
concentrators. NEMS may make a much bigger impact, though, on
future solar cells. QDs, for example, exhibit optical properties that are
absent in the bulk material due to the confinement of electron-hole
pairs, in a region of a few nanometers and the wavelength at which a
QD absorbs or emits radiation can be adjusted at will (see Volume I,
Chapters 3 and 5). A first advantage of QDs, then, stems from the
fact that an optimum bandgap, corresponding to the highest possible
solar-electric energy conversion, can be achieved by using a mixture
of QDs of different sizes for harvesting the maximum amount of the
incident light (Eγ ). A second advantage stems from the increased
possibility of multiphoton absorption by nanoparticles. The formation
of multiple excitons per absorbed photon takes place only when the
energy of the absorbed photon is far greater than the semiconductor
bandgap (Eg ). Multiple photon absorption does not readily occur in
bulk semiconductors, where the excess energy quickly dissipates as
heat before it can create other electron-hole pairs. However, in
semiconducting QDs, the rate of energy dissipation is reduced
dramatically, and with the charge carriers confined within a
nanovolume, their interactions are enhanced, increasing the
probability for multiple excitons per incoming photon. A third
advantage of QDs is that in contrast to traditional crystalline
semiconductor materials that are rigid, QDs can be molded into a
variety of different forms, such as sheets or three-dimensional arrays.
Quantum dots can easily be combined with organic polymers or dyes
or made into porous films, and in colloidal form, they can be
processed to create junctions on inexpensive substrates such as
plastics or thin metal foils. A fourth advantage of QDs is that it is quite
easy to decouple the absorption and charge separation process.
Decoupling charge separation from light absorption results in many
more engineering options to make the solar cell of the future. For
example, the charge separation may be achieved in a semiconductor
QD, while the light absorption may be in an organic dye adsorbed on
its surface. The broader quantum solar cell strategy is thus to
sensitize a first material with bandgap Eg with another material of
bandgap (Eq ) smaller than Eg (i.e., Eq < Eγ < Eg ), and this second
semiconductor material can be a dye (monomeric), a photoactive
biopolymer, a QD, or a quantum well. QDs have the advantage over
dyes as they degrade less in sunlight. In its simplest form, a quantum
well solar cell (QWSC) consists of a p-i-n solar cell with a multiple
quantum well (MQW) system added to the intrinsic (i) region. The
idea, as with a tandem solar cell, is that the quantum wells absorb
light of energy below the bulk bandgap, Eg . MEMS in Solar Cells
MEMS techniques allow for the fabrication of lenses (including Fresnel
lenses), mirrors, photonic crystals, cavities, and solar cell surface
textures to enhance the interaction of incoming photons with the
solar cell. A simple method to put MEMS to work in a conventional
solar cell is by surface structuring the cell so that the light hits the PV
several times, as illustrated in the pyramid structuring in
Figure 9.28.34 NEMS in Solar Cells We present examples of quantum
technology benefiting solar cells by highlighting nanoporous
semiconductors in the liquid-junction Grätzel solar cell, two-photon
excitation in hybrid organic-QD PV devices, and increased efficiency in
a QWSCs. In Volume I, Chapter 7, we learned that in liquidjunction
solar cells, to avoid corrosion of the cell, one chooses a redox system
where the maximum density 502 From MEMS to Bio-NEMS:
Manufacturing Techniques and Applications of states of the reductant
(the peak of the Gaussian in Volume I, Figure 7.62) is at the same
level as the valence band edge of the semiconductor.35 In this
configuration, empty levels (holes) in the valence band, generated by
shining light on the electrolyte/ semiconductor interface, react with
the reductant rather than with the semiconductor itself. In such liquid-
junction solar cells, appropriate reductants thus compete effectively
with the oxidation/corrosion of the semiconductor because they react
more avidly with the holes in the valence band than the material
itself.36 The maximum voltage generated by such a cell, in theory, is
simply the difference between the Fermi level of the semiconductor
and the redox potential of the electrolyte. The best known liquid-
junction solar cell today is the dye-sensitized Grätzel design.37–39
The Grätzel cell is significantly different from a conventional
semiconductor solar cell, in which the semiconductor acts as the
source of photoelectrons and also provides the potential barrier to
separate the charges to generate the current. In contrast, in the
Grätzel dye-sensitized solar cell, the semiconductor is used solely for
charge separation, and the photoelectrons are provided from a
separate photosensitive dye. As shown in Figure 9.29 the solar cell
consists of a TiO2 layer forming a nanoporous structure with a dye
(e.g., ruthenium-polypyridine) spread throughout its surface. The dye
molecules are small, and in order to capture a reasonable amount of
the incoming light, the nanoporous structure is used as a scaffold
holding large numbers of the molecules in a 3D matrix, vastly
increasing the number of molecules for a given surface area. The
charge separation is provided by the semiconductor-liquid junction
contact. The TiO2 layer sits on a transparent anode made of fluorine-
doped tin oxide (SnO2 :F) deposited on the side of the glass plate
facing the TiO2 layer. The electrolyte solution also holds a redox
couple (e.g., I – /I3 –) mediator, located in the space between the dye-
coated TiO2 and a cathode, typically a thin film of platinum metal.
Photons enter the cell through the transparent SnO2 :F window, and if
they have enough energy, they are absorbed by the dye, creating an
excited dye state. From this excited state an electron is “injected” into
the conduction band of the TiO2 . This way, the dye molecule is
oxidized, and it would decompose if it didn’t quickly react with iodide
in the electrolyte oxidizing it to form triiodide (dye regeneration
reaction): 3 2 I I e 3 − → + − − Reaction 9.6 The reaction with iodide
occurs very quickly compared with the recombination of the injected
electron with the oxidized dye molecule, effectively preventing short-
circuiting the solar cell. The injected electron then travels to the
cathode via the external circuit, and the triiodide recovers its missing
electron by diffusing through the solution to the Oxide Metal p-type
plated metal p + n + n ++ FIGURE 9.28 Surface structuring of a solar
cell surface.34 Conducting glass TiO2 Dye Injection S* Electrolyte Red
Mediator Interception Diffusion Maximum voltage Ox S° /S+ hv
Cathode Red = I– Ox = I3 – FIGURE 9.29 A liquid-junction solar cell:
the dye- sensitized Grätzel design. The S energy levels represent the
photosensitive dye levels. Power and Brains in Miniature Devices 503
cathode, where it is reduced back to iodide (redox regeneration
reaction): I e I 3 − + →− − 2 3 Reaction 9.7 The electron injection
process does not introduce a hole in the TiO2 , only an extra electron,
and recombination of this electron directly from the TiO2 to the
electrolyte is not possible because of differences in energy levels (see
Volume I, Chapter 7: if there are no overlapping energy levels, no
reaction can occur). This is a very important property of the dye-
sensitized Grätzel design: the electron-hole recombination that
plagues the efficiency of traditional cells does not exist here. The
potential used for external work in this solar cell is given by the
following equation: V E V ext F redox = − (9.1) The quantum
efficiency of a Grätzel-cell is very high because the huge surface area
of the TiO2 nanostructure makes the chance that a photon will be
absorbed very high, and the dyes are very effective at converting
photons to electrons. Most of the losses are due to conduction losses
in the TiO2 and the clear anode, and the optical losses stem from the
front electrode as well. The overall quantum efficiency is about 90%,
with the “lost” 10% being largely accounted for by the optical losses
in the anode. Although the dye is highly efficient in turning photons
into electrons, only those electrons with enough energy to cross the
TiO2 bandgap result in current being produced. The bandgap is larger
than that of silicon, which means that fewer of the photons in sunlight
are usable for power generation. In addition, the electrolyte limits the
speed at which the dye molecules can regain their electrons and
become available for photoexcitation again. These factors limit the
current generated by a liquid-junction solar cell; for comparison, a
traditional siliconbased solar cell offers about 35 mA/cm², whereas a
Grätzel cell offers about 20 mA/cm². The closest competitors in price-
performance terms are the various thin-film approaches, which are
currently somewhat more developed commercially. The dyes used in
early experimental Grätzel cells were sensitive only in the high-
frequency end of the solar spectrum (ultraviolet and blue). Newer
versions have a much wider frequency response; notably, the
triscarboxy-terpyridine ruthenium complex [Ru(2,2’,2”-(COOH)3 -
terpy)(NCS)3 ] is efficient right into the low-frequency range of red
and infrared (IR) light. The wide spectral response results in the dye
having a deep brown-black color, and it is referred to simply as black
dye. 39 These polymers also are also sensitive to infrared are called
thermovoltaics, and they generate electricity from both light and
heat. Current Grätzel cells demonstrate a conversion efficiency of
about 11%, whereas common low-cost commercial panels operate
between 12% and 15%. Flexible thin-film cells are typically around
8%. Perhaps the major disadvantage of a liquid-junction solar cell is
the use of liquid electrolytes, which brings leakage and corrosion
problems, as well as temperature stability problems. At low
temperatures, the electrolyte can freeze, ending power production
and potentially leading to physical damage. Higher temperatures
cause the liquid to expand, making sealing the cells a serious
problem. To avoid the above problems of liquid-junction solar cells,
several research groups aim to produce a hybrid organic-QD PV
device as the solid analog to the Grätzel cell. Researchers at the
University of Toronto mixed lead sulfide (PbS) QDs with the polymer
poly(2-methoxy-5-[2′-ethylhexyloxyp-phenylenevinylene]) (MEH-PPV).

into the infrared: on its own the polymer absorbs between ∼400 and
Wrapping MEH-PPV around the QDs shifted the polymer’s absorption

∼600 nm, and the QDs have absorption peaks that can be tuned from
∼800 to ∼2000 nm (see Figure 9.30).40 The best efficiency registered
thus far with this approach is 1.6%.41 In 2000, Nozik from the
National Renewable Energy Laboratory in Golden, Colorado, predicted
that theoretically, QDs could increase the efficiency of solar cells
through multiple exciton generation.42 “We have shown that solar
cells based on QDs theoretically could convert more than 65 percent
of the sun’s energy into electricity, approximately doubling the
efficiency of solar cells,” said Nozik.42 In 2004, Schaller and Klimov
from the Los Alamos National Laboratory in New Mexico were the first
to demonstrate this phenomenon experimentally, using QDs made of
lead selenide.43 Multiple excitons 504 From MEMS to Bio-NEMS:
Manufacturing Techniques and Applications start to form
simultaneously as soon as the photon energy reaches twice the
bandgap, and a group of researchers led by Nozik demonstrated that
the absorption of a single photon by 2.9 nm PbSe QDs yielded not one
exciton but three when the energy of the photon absorbed is four
times that of the bandgap43 (a quantum yield of 300%). Quantum
dots made of PbS also showed the same phenomenon. Quantum
wells, as we saw in Volume I, Chapters 3 and 5, are ultrathin layers
(nanostructures) of a narrower bandgap material (e.g., a GaAs well)
sandwiched between regions of wider bandgap (e.g., AlGaAs barrier).
They are made possible by the development of modern crystal growth
techniques such as