Fahlman B.D. Materials Chemistry

Подождите немного. Документ загружается.

valence electron (i.e., possess a low work function). The holes that migrate from the

anode are blocked from further transport to the cathode by an organic layer of 2,9-

dimethyl-4,7-diphenyl-1,10-phenanthroline (BCP, Figure 4.79c), which has a rela-

tively low-lying HOMO.

In contrast, the anode is usually tin-doped In

(ITO, typically 1:9 Sn:In), since

this material is transparent and highly conductive (ca.1 10

1

[84]

The substitutional replacement of In

ions with Sn

ions results in n-doping of

the lattice, injecting electrons into the conduction band. The doping of Sn into the

lattice may also result in SnO sites; this introduces holes in the lattice that

reduces its conductivity. The conductivity of ITO is due to both Sn dopants and O

vacancies, as represented by the general formula In

2x

32x

[85]

Since the

organic hole-transport polymer is in direct contact with ITO in an OLED, the surface

properties of ITO is important for effective hole injection. It has been shown that

surface treatments such as UV ozone cleaning or Ar/O

plasma treatments, result in

an increase in the work function of the ITO surface as the Sn :In ratio is decreased

(and O concentration is incr eased).

The recombination of a hole and electron creates a quasiparticle known as an

exciton, which releases a photon of energy. Organic molecules in the emissive layer

facilitate exciton formati on – likely through hole-trapping, followed by Coulombic

attraction to the free electrons. A challenge in OLED design is to ensure that an

equal number of holes and electrons meet in the emissive layer. This is not trivial,

since holes migrate much slower than electrons in conductive organic molecules.

The organic molecules used in OLED s/PLEDs are p-conjugated, meaning that p

orbitals on adjacent (–C¼C–C¼)

units overlap, resulting in p valence and p

conduction bands. The observed color will depend on the HOMO–LUMO gap of

the molecule, which may be ﬁne-tuned by altering the length of the conjugation,

or nature of the molecular backbone (e.g., substitution of electron withdrawing/

donating groups, etc.).

A limitation of traditional OLEDs/PLEDs is their relatively poor quantum efﬁ-

ciency. An exciton may exist in either a single t (total spin, S ¼ 0) or triple t state

(S ¼ 1), with the triplet outweighing the singlet by a 3:1 ratio. The molecules in the

emissive layer are referred to as ﬂuorophores, which yield ﬂuorescent radiation only

when a singlet exciton is formed. Hence, the theoretical maximum efﬁciency that is

possible is only 25%. In an effort to improve the efﬁciency, a number of recent

studies have been devoted to using organometallic compounds (transition metals

with organic ligands), which allow for rapid intersystem crossing (ISC) from the

excited singlet to light-emitting triplet states (Figure 4.80). The process of collecting

the excitons in the lowest excited triplet state is referred to as triplet harvesting,

and results in a 100% quantum efﬁciency of the OLED.

[86]

This effect is most

pronounced for complexes of 4d and 5d transition metals with well-shielded valence

electrons, which exhibit large spin–orbit coupling

[87]

and concomi tant mixing of

singlet and triplet states.

[88]

328 4 Semiconductors

Figure 4.80. Illustration of triplet harvesting. In the absence of a triplet emitter, the triplet excitation

energy is converted to heat, losing 75% of the quantum efﬁciency. Also shown are a variety of

organometallic complexes, and their relative spin–orbit coupling values – directly proportional to their

use in phosphorescent OLEDs. Reproduced with permission from Yersin, H. Top. Curr. Chem. 2004, 241,

4.3. Light-Emitting Diodes: There is Life Outside of Silicon! 329

4.4. THERMOELECTRIC (TE ) MATER IALS

As our society attempts to wean itself from a dependence on fossil fuels, many

alternative energy sources are being investigated. One interesting potential is from

the conversion of waste heat (e.g., vehicle engine heat) into electricity. This conver-

sion is due to the thermoelectric effect, ﬁrst discovered for junctions (or thermocou-

ple) of two dissimilar metals. When metals are welded together, free electrons are

able to drift across the junction in a preferential direction, based on the different

atomic lattice structures of each metal. The migration of electrons effectively leaves

one metal electron-deﬁcient (positively charged), and the other metal negatively

charged.

If the two metals are held at different temperatures, a voltage will result that is

proportional to the DT. The ratio (DV / DT) is referred to as the Seebeck coefﬁcient,

a, related to the band structure of the materials involved. The a value for semicon-

ductors is at least two orders of magnitude larger for semiconductors relative to

metals, giving rise to much greater voltages. Contrary to the Seebeck effect, when a

current is passed through a closed circuit of the two metals, heat is produced at one

of the junctions and is absorbed at the other. Since the latter junction is cooled, this

Peltier effect has been largely exploited for thermoelectric cooling applications

(e.g., auto seat coolers, com puter component cooling).

A TE device consist s of a heat source and sink, joined together via n-type and

p-type semiconductor materials (Figure 4.81).

[89]

The ﬁgure of merit, ZT (Eq. 22), of

the device is useful to determine its suitability for power generation or refrigeration –

a means to describe the transport properties of the TE material. An effective TE

Figure 4.81. Schematic of a thermoelectric device, which may be used for power generation or

refrigeration.

330 4 Semiconductors

material should have a high Seebeck coefﬁcient (heat conversion efﬁciency) and

electrical conductivity, as well as low thermal conductivity to maintain thermal

isolation in the device. The thermal conductivit y is related to the transfer of heat

through a material through either electron transport or quantized lattice vibrations

(phonons). Hence, the ideal TE material has been described as a phonon-glass/

electron-crystal, having the electrical properties of a crystalline lattice, and thermal

properties of an amorphous/glass-like solid.

ZT =

ð22Þ

where a is the Seebeck coefﬁcient (mVK

1

; 1–10 for metals, 150–250 for

semiconductors); s is the electrical conductivity; and l is the thermal conductivity

(electronic + lattice terms).

Semiconductors are much more effective TE materials (greater ZT values) than

metals due to their signiﬁcantly greater Seebeck coefﬁcients and lower thermal

conductivities. A barrier toward higher ZT values for semiconductors is their

relatively small electrical conductivity, especially at low temperatures. In order

to overcome this limitation, the chemical composition of the semiconductor may

be ﬁne-tuned to yield a small bandgap material (6k

T; k

¼ Boltzmann cst;

i.e., 0.16 eV at 300 K), o r one that is sufﬁciently doped with an intermediate

concentration (ca.10

–10

) of electroni c/thermal carriers exhibiting high

mobility through the lattice.

The most widely studied TE material is Bi

, consisting of a hexagonal unit

cell with repeating...[Te–Bi–Te–Bi–Te...Te–Bi–Te–Bi–Te]...units (Figure 4.82).

Whereas Te–Bi layers are bound by strong covalent interactions, the bonding

between adjacent Te layers is through weak van der Waals interactions. This results

in bulk anisotropic electrical and thermal conductivity, being most pronounced

along planes that are perpendicular to the c-axis of the unit cell. To further improve

ZT values, Bi

crystals may be doped with n- or p-type dopants, with most

desired compositions of Bi

2.7

0.3

(n-type) and Bi

0.5

1.5

(p-type), yielding

ZT ¼ 1 at room temperature. This improvement results from a decrease in thermal

conductivity of the lattice, brough t about by phonon scattering by the dopant atoms

(i.e., perturbing the symmetry of the lattice, affecting the organized lattice vibration

modes).

The following list represents the primary materials classes that have been

designed in an effort to optimize ZT for thermoelectric applications.

[90]

The general

strategy is to dope the lattice with sufﬁcient carriers (n- or p-dopants), while also

interrupting the phonon transport through the solid through the introduction of large

interstitial atoms.

(i) Complex Solid-State Inorganic Lattices (e.g., CsBi

,Bi

2x

3y

ZrTe

,Ag

m+2n

(M ¼ Sb, Bi), AgPb

SbTe

, “Half-Heusler alloys”

MNiSn (M ¼ Zr, Hf, Ti), Zr

0.5

0.99

0.01

). For half-Heusler

alloys, the unit cell is a combination of a NaCl lattice of two metals, with the

third metal occupying tetrahedral interstitial sites. For instance, for TiN iSn,

4.4. Thermoelectric (TE) Materials 331

the Ti and Sn form the NaCl lattice, with Ni occupying 1/2 of the available

tetrahedral interstitial sites (i.e., Ti:Ni:Sn ¼ 4:4:4 atoms per unit cell). This

combination offers a great deal of control over electronic/thermal conductivity

of the solid. While the introduction of Sn (p-doping) increases electrical conduc-

tivity, heavy metal atoms such as Ti and Ni cause a decrease in thermal

conductivity through phonon scattering.

(ii) Crystal Structures with “Rattlers” (e.g., Rare-earth (e.g., La, Ce, Nd, Sm, Eu,

etc.) doped AB

-based (A ¼ Co, Fe, Ru, Os, Ir; B ¼ P, As, Sb), Figure 4.83),

[91]

or complex antimonide skutterudites (e.g., b-Zn

,Yb

MnSb

7x

[92]

or clathrates (e.g.,Ba

,E¼ Ge, Si)(see endnote

89). In these structures, dopant atoms are weakly bound to the cage, and “rattle”

in resp onse to increasing temperature. As the atom within the cages becomes

smaller/heavier, the amount of structural disorder will increase causing a larger

decrease in lattice thermal conductivity.

(iii) Oxides (e.g., NaCo

,Ca

[93]

0.02

0.98

O, b-SrRh

[94]

). These

structures consist of CoO

layers, which serve as effective electronic transport

Figure 4.82. Unit cell of Bi

– a widely studied thermoelectric material. The blue atoms are Bi, and

the pink atoms are Te. Reproduced with permission from Tritt, T. M.; Subramanian, M. A. MRS Bull.

332 4 Semiconductors

media within the latt ice. A variety of metal/metal oxide layers (e.g., Na, Ca, Sr,

Bi) are placed in between the metal oxide layers to serve as phonon scatterers

(Figure 4.84).

(iv) Nanoarchitectures (e.g., PbTeSeTe/ PbTe quantum dot superlattices, nanowires

(Bi

[95]

, PbSe

[96]

, etc.), nanoporous silicon,

[97]

and “nanoparticle-in-alloy”

materials

[98]

). The electrical and thermal conductivities of a bulk material are

strongly coupled, which makes it difﬁcult to increasingly improve ZT values.

However, in a solid with at least one dimension in the nanoregime, one may

Figure 4.83. Crystal structures of an unﬁlled IrSb

skutterudite (top – Sb atoms shown in black) and

Type I/Type II clathrates (bottom). In the skutterudite structure, a void is present in the center of the

unit cell, surrounded by 12 Sb atoms. For the clathrate structures, tetrahedrally bound framework atoms

(e.g., Ge, Sn, Si) are illustrated in blue, and guest atoms within the various cages are shown in orange

and pink. Reproduced with permission from (i) Nolas, G. S.; Poon, J.; Kanatzidis, M. MRS Bull. 2006, 31,

American Chemical Society.

4.4. Thermoelectric (TE) Materials 333

dramatically alter the carrier- transport properties by simply varying its dimen-

sions due the conﬁnement of the carriers inside the material.

[99]

We will discuss

quantum conﬁnement within nanostructures in more detail in Chapter 6 .

The comparison of ZT values for a host of the above bulk materials is displayed

in Figure 4.85, along with the variation of ZT with temperature. The search for

new TE materials continues to be an extremely active area of research; recent

work has resulted in suitable TE components at virtually all temperature regimes.

However, the search continue s for materials with ZT values in excess of 2.0. Though

most bulk materials exhibit ﬁgure-of-merits 1.5, certain types of nanomaterials

are the only candidates to date with ZT values >2.0 – yet another triumph of

nanotechnology!

[100]

IMPORTANT MATERIALS APPLICATIONS III: PHOTOVOLTAIC

(SOLAR) CELLS

The increasingly volatile prices and availability of gasoline and natural gas, and the

recent British Petroleum oil spill disaster has brought about a heightened awareness

of our dependence on nonrenewable resources. Currently, over 90% of the US

electricity supply comes from the combust ion of fossil fuels (i.e., coal, oil, and

natural gas) and nuclear power generation.

While the US coal reserve is estimated at 290 billion tons (enough to last another

230 years at current production levels), there is an increased concern about the

adverse environment al effects associated with its combustion, namely SO

,NO

and CO

emissions. With the energy needs of our world likely to double within our

Figure 4.84. Crystal structure schematic of CoO

-based TE oxides. Shown are (a) Na

CoO

, (b)

, and (c) Bi

. Reproduced with permission from Koumoto, K.; Terasaki, I.;

334 4 Semiconductors

generation, there continues to be active development of alternative renewable

energy sources. In an earlier chapter, we dealt with hydrogen-based fuel cells; we

now will describe another extr emely attractive option – harvest ing the unlimited

power from the sun.

The majority of commercial photovoltaic cells utilize silicon-based technology.

When sunlight comes in contact with a p–n diode, the absorbed energy causes

promotion of electrons from valence to conduction bands, generating additional

electron-hole pairs (EHPs). Since the n-side is much thinner than the p-side, most of

the photons are absorbed within the depletion region and p-side. The most noticeable

result of this excitation is an effective increase in the number of electrons in the

conduction band of p-Si (or holes in n-Si, referred to as the minority-carrier

concentration). As a result, electrons in the p region will diffuse into the depletion

region, where the junction potential (Figure 4.86a, E

) propels them back into n-Si

(vice versa for holes). It is important to note that only EHPs that are generated within

Figure 4.85. The temperature dependence of ZT values for various TE materials. TAGS refers to

(GeTe)

0.85

(AgSbTe

)

0.35

. Reproduced with permission from Tritt, T. M.; Subramanian, M. A. MRS

4.4. Thermoelectric (TE) Materials 335

Figure 4.86. Schematic of (a) a single-junction and (b) multi-junction photovoltaic cell. The single-

junction cell is reproduced with permission from Kasap, S. O. Principles of Electronic Materials and

336 4 Semiconductors

the depletion region or within a certain distance from this region (Figure 4.86a –

electrons within diffusion length L

, or holes within diffusion leng th L

) may

contribute to photocurrent. If the minority carriers are further away from the deple-

tion zone, recombination will occur rather than migration to their respective electro-

des (i.e., ele ctrons in n-Si ﬂowing through the external circuit to neutralize the

excess holes in p-Si).

In the design of solar cells, a small bandgap semiconductor is desirable since

it requires less energy to promote electrons from valence to conduction bands.

However, though the resultant photovoltage of the solar cell is directly proportional

to the bandgap, higher energy photons would be converted to heat rather than

electrical energy. The use of a higher bandgap material would result in a lower

photocurrent since only high-energy photons (E  E

) would be absorbed. The

bandgaps of Si and GaAs (1.1 eV (1,127 nm) and 1.4 eV (886 nm), respectively)

are able to absorb a broad range of electromagnetic radiation; however, the efﬁ-

ciencies of Si or GaAs based solar cells are only 20–25%. Thin-ﬁlm solar cel ls that

utilize amorphous silicon (a-Si) are also currently under investigation. Since Si

atoms do not form a continuous crystalline array, surface sites contain signiﬁcant

numbers of dangling bonds, which are easily passivated through reaction with

hydrogen (yielding hydrogenated amorphous silicon, a-Si:H). The larger bandgap

of a-Si (1.7 eV) is more effective at capturing broadband solar energy; however, the

best efﬁciencies to date are less than 10% for this technology.

In order to capture a larger range of photon energies, multijunction photocells

have rece ntly been introduced. This method consists of a stack of semiconductor

layers with decreasing bandgaps (Figure 4.86b). Top layers have higher bandga ps,

and are able to absorb higher energy photons; lower layers of the cell absorb the

transmitted lower energy wavelengths. The multijunction solar cell shown in Fig-

ure 4.86b has resulted in an efﬁciency of 34%, the highest value reported for all

photovoltaic cells to date.

[101]

It has been proposed that replacement of GaAs with a

material with an E

of 1.25 eV would result in a higher overall efﬁciency, since more

photons would be collected by the second layer and fewer would be transmitted to

the bottom Ge substrate. With additional tweaking such as altering the number of

layers, individual layer thicknesses/stoichiometries, it is expec ted that multijunction

photocells with efﬁciencies of 45–50% and higher will soon be realized.

While it is important to control the stoichiometry of each layer to adjust their

bandgaps, it is equally important to have as few interfacial mismatches as possibl e.

That is, each layer must be epitaxially grown to ensure that the lattice constants are

perfectly matched. It has been shown that a lattice mismatch of only ca. 0.01% is

enough to cause signiﬁcant electron-hole recombinations, resulting in lower cell

efﬁciency. CVD is the method of choice for the fabrication of these multilayer

devices; as you might expect, cells of this variety are relatively quite expensive.

In an effort to decrease the cost of productio n for solar cells, new architectures

known as dye-sensitized solar cells (DSC) are of increasing interest.

[102]

Instead of

light being absorbed by inorganic semiconductor laye rs, absorption occurs in an

organic-based ﬁlm that comprises light-harvesting dye molecules bound to a TiO

4.4. Thermoelectric (TE) Materials 337