Basic Theories of Semiconductor

1
Basic Theories of Semiconductor

Electrochemistry
1.1. INTRODUCTION
Silicon, due to its importance in the electronics industry, is perhaps the most investi-
gated electrode material in the field of semiconductor electrochemistry. A wide range
of electrode phenomena, from electropolishing to formation of porous silicon as well
as from current multiplication to current oscillation, can occur on silicon electrodes.
As will be seen in later chapters, many details of these phenomena are not directly
explicable by the existing theories of semiconductor electrochemistry. This is perhaps
related to the fact that the basic theories deal with idealized situations, but the electrode
behavior of a specific system almost always deviates from the ideal situations in
different ways. Also, the rich details of each large-scale phenomenon, e.g., formation
of porous silicone are determined by a set of reactions that are intrinsically complex at
the atomic scale such that the rich details are lost when simplifications are made in a
theory. However, the basic concepts and theories can still be useful in describing the
electrode processes at certain temporal and spatial scales and the large-scale phe-
nomena can be understood by taking into account the different aspects and scales of
these processes.
This chapter provides an overview of the various aspects of semiconductor elec-
trochemistry, which will be useful in describing the specific phenomena on silicon elec-
trodes in later chapters. Also, this systematic description of the basic theories is intended
to be used as a reference base, on which the details of electrode behavior in different
situations are characterized and organized. To avoid unnecessary detail and use of
space, the descriptions in this chapter are limited to the basic concepts, the physical
schemes, and the essential quantities and equations that are required to understand
these concepts and schemes. More detailed treatment of the various aspects of semi-
conductor electrochemistry can be found in the literature.1,44,86,270,962
1
2 CHAPTER 1
1.2. ENERGETICS OF SEMICONDUCTOR/ELECTROLYTE INTERFACE
The energy spectrum of electrons in an ideal crystal consists of two different types
of energy bands: those with filled energy levels (allowed bands) and those with no
energy levels (band gaps). For a semiconductor the upper unfilled band is called the
conduction band and the lower almost filled band is called the valence band as shown
in Fig. 1.1. The energy levels in a semiconductor are characterized by the conduction
band edge and valence band edge and by the Fermi level The Fermi level
describes the equilibrium distribution of carriers in the bands and is the chemical poten-
tial of electrons in the semiconductor. The width of the band gap, depends
on the strength of the chemical bonds. For silicon at 300K the band gap is 1.12eV. The
effective band gap is reduced by heavy doping, larger than For silicon, a band
gap reduction of more than 100mV is associated with a dopant concentration of
The electronic conductivity of semiconductors, as expected from the band struc-

ture, can be generated by electrons of atoms of the basic substance in the crystal (in-
trinsic conductivity) as well as by electrons of impurity atoms or by the presence of
defects (extrinsic conductivity). In intrinsic semiconductors at the generation
of current carriers occurs as a result of the thermal excitation of some electrons from
the valence band to the conduction band, with the corresponding thermal rupture of
some chemical bonds. Simultaneously, an equal number of positively charged holes are
created in the valence band. In an electric field, these holes behave like particles pos-
sessing a positive charge equal in absolute value to the charge of the electron. For
extrinsic semiconductors, impurities and defects (which have energy levels located in
the band gap) are classified as either donors or acceptors as shown in Fig. 1.1. Donors,
usually located at energy levels slightly below the conduction band, give up excess
electrons to the conduction band, thereby creating electron conductivity (n-type semi-
conductors). Acceptors, located at energy levels slightly above the valence band,
capture valence electrons from atoms of the basic substance, producing hole conduc-
tivity (p-type semiconductors).
Similar to for the energy levels in semiconductors, the energy levels of elec-
trons in electrolytes associated with ions are characterized by the redox potential,
The redox potential describes the tendency of the species to give up or accept electrons
and can be considered as the effective Fermi level of the solution.
BASIC THEORIES OF SEMICONDUCTOR ELECTROCHEMISTRY 3
When a semiconductor is brought into contact with an electrolyte, equilibrium

is attained when the Fermi levels in the two phases become equal, that is,
In the case shown in Fig. 1.2 for an n-type semiconductor, where of the semi-
conductor is higher than that in solution, electrons will flow from the semiconductor
to the solution phase. The resulting excess charge in the solid semiconductor does not
reside at the surface as it would in a metal, but instead is distributed in a region near
the surface called the space charge region. The resulting electric field in the space
charge region is shown by a bending of the bands. In the case of Fig. 1.2 where the
semiconductor is positively charged with respect to the solution, the bands are bent
upward (with respect to the level in the bulk semiconductor). On the other hand, an
ionic layer on the solution side establishes an electric double layer between the solid
surface and solution.
Thus, several charged layers exist at the interface of a semiconductor and an elec-
trolyte. Figure 1.3 illustrates the different charged layers in the semiconductor/
electrolyte region. On the semiconductor side there is the space charge layer associated
with a band bending of The ionic layer on the solution side can be further divided.
The charged layer lies between the solid and the plane at the position of the closest
approach of mobile ions, called the outer Helmholtz plane (ohp). The Helmholtz layer
is formed by ions attracted to the electrode surface by the excess charge in the space
charge layer and also by the polar water molecules. In particular, the adsorption of
and is an important process that determines the potential drop in the Helmholtz
layer of many semiconductors. Between the outer Helmholtz plane and the solid surface
there is also a layer (not illustrated in the diagram) containing solvent molecules and
sometimes other specifically adsorbed species.44 The locus of the specifically adsorbed
ions is called the inner Helmholtz plane. The charge layer that extends from the outer
Helmholtz layer into the bulk, called the Gouy–Chapman layer, is a region of solution
with excess ions of one sign and its thickness depends on the electrolyte concentration.
In concentrated electrolytes (>0.1 M) the contribution of the Gouy–Chapman layer is
negligible and the potential drop on the solution side of the double layer can be
expressed by the potential drop in the Helmholtz layer,
4 CHAPTER 1
The typical thickness of each of these three layers is indicated in Fig. 1.3. The
actual thickness of these layers of the interface depends on the specific conditions in
the system such as doping level of the semiconductor, electrolyte concentration, and
bias condition. In many situations the presence of surface states due to adsorbates or
surface defects is also important in determining the distribution of charge in the deple-
tion region. Sometimes, electrode reaction processes can result in the formation of a
solid surface phase such as oxides. Depending on its thickness and degree of homo-
geneity, it can drastically affect the electrical properties of the interface. As will be seen
in the rest of this book, formation of oxide on silicon plays a critical role in many phe-
nomena observed on silicon electrodes.
1.2.1. Energy Levels in Semiconductor

For a semiconductor at T > 0 K a dynamic process exists: Electrons are constantly
thermally excited from the valence band to the conduction band, corresponding to
electron–hole generation. At the same time, some electrons are losing energy and falling
back to the valence band, corresponding to electron–hole recombination. At equilib-
rium, there is a constant density of electrons in the excited states, the rate of genera-
tion equaling the rate of recombination.
The distribution of the electrons among the allowed energy states in the semi-
conductor crystal at thermal equilibrium is described by the Fermi–Dirac distribution
function.45 It is denoted by which has the form
where is a reference energy called the Fermi energy. At meaning that

the probability of a state being occupied by an electron for the energy level at the Fermi
level is 1/2. For not too highly doped n-type material the Fermi level is well below the
conduction band such that and the Fermi function reduces to the simpler
Maxwell–Boltzmann distribution function:
The total density of electrons for not too heavily doped n-type material can then be
found by the product of the density of allowed states in the conduction band, g(E), and
the probability that these states are filled and integrating over the conduction band:
Similarly, for moderately doped p-type material, the density of holes in the valence
band is given by
where and are the effective densities of energy states at the bottom of the con-
duction band and at the top of the valence band, respectively. For large dopant con-
centrations or for silicon, these equations are no longer
valid as the Fermi–Dirac distribution cannot be approximated by the Maxwell–
Boltzmann distribution function and a different distribution function should be used.45
At very high doping levels or the semiconductor is degenerated
because the Fermi level is within the conduction or the valence band. As a result,
allowed states exist very near the Fermi level, just as in metal. Consequently, the elec-
tronic properties of heavily doped semiconductors become similar to those of metals.
Equations (1.2) and (1.3) essentially depict the chemical potential of the elec-
trons, and of the holes, respectively:
At equilibrium, Eqs. (1.4) and (1.5) become
where is the intrinsic carrier density. Equation (1.6) means that at thermal equilib-
rium the product of the electron and hole densities for a given semiconductor and tem-
perature is a constant. For silicon at room temperature, which
leads to
6 CHAPTER 1
For a doped silicon, depending on the dopant, one type of carrier is the majority
carrier as its concentration far exceeds the other, even at low doping levels. For
example, for donor concentration that is, which
is nearly 10 orders of magnitude below the majority carrier concentration.
In the neutral region of a semiconductor, the number of positive charges must be
exactly balanced by the number of negative charges. This means
assuming all the dopant atoms are ionized. Thus, when there is only one dopant, either
donor or acceptor, for n-type material and for p-type material
1.2.2. Energy Levels in Electrolyte

The chemical potential of electrons for a redox couple is given by the Nernst
equation
where is the standard redox potential, and and are the activities of the
oxidized and reduced species of the redox couple, respectively.
The redox potential is generally referred to the standard hydrogen potential
(SHE), which has an exactly defined energy, relative to the energy of the free
electron in vacuum or at infinity. Thus, electrode potentials of redox couples can be
expressed on the absolute energy scale according to
The negative sign of in the equation is due to the different signs in the conventional
and the absolute electron energy scales. has been found to be about –4.50 eV referred
to the vacuum level so that the electron energy of any redox couple is86
1.2.3. Distribution of Energy Levels in Electrolyte

The redox potential reflects the average energy levels at equilibrium of all the
individual redox species, both reduced and oxidized forms. It can be further divided
into two levels: the energy level of the reduced species, and the energy level of
the oxidized species, with is the most probable energy level
for the oxidizing species and Ered is the most probable energy level for the reducing
species. In a liquid electrolyte, the energy levels of individual ions tend to fluctuate due
to the solvation effect of the polar solvent molecules surrounding the ions. The dipoles
associated with the solvent molecules constantly move to/away and rotate around the
ions, causing thermal fluctuation in the polarization.
The energy level, and energy level, can be related to the redox energy
level, by a quantity called the reorientation energy which is
determined by the relaxation process involved in the regrouping and reorientation of
the solvation shell after electron transfer between the oxidized and reduced states. The
value of can be experimentally determined and is on the order of 0.5 to 1 eV.44,876 The
individual energy states are distributed over a certain energy range and can be described
by the density of occupied states and empty states Assuming a Gaussian type
of distribution and a harmonic oscillation for the solvation shell, and can be
described by
where and are normalizing factors. Figure 1.4 schematically illustrates the
energy levels of the reduced and oxidized species and their distribution functions.86
1.2.4. Energy Levels at Semiconductor/Electrolyte Interface

For a semiconductor/electrolyte interface, the equilibrium condition requires the
Fermi levels of the semiconductor and solution to be equal. To achieve equilibrium,
electrons cross the interface, which changes the charge distribution and the potential in
the interface region until
The band structure at equilibrium for the interface region is illustrated in Fig. 1.5.
The band bending is the potential drop in the space charge layer. The potential drop
on the solution side is measured by the Helmholtz potential. and are the
conduction band edge and the valence band edge at the surface, respectively. is the
measured potential. Normally, the band edge positions at the surface are equal for p-
and n-type materials because the same Helmholtz potential is expected for the two
8 CHAPTER 1
materials. The magnitude of the Helmholtz potential is determined by the adsorption/

desorption processes, which in many situations is a function of pH. By adjusting the
pH of the solution, can be made zero and the pH at is called the point of
zero charge (pzc).
The measured potential and thus and can be varied through external
polarization. is the applied potential when the electrode is externally polarized and
is the open-circuit potential without external polarization. When the semiconductor has
no excess charge, there is no space charge region and the bands are not bent. The elec-
trode potential under this condition is called the flatband potential The flatband
potential is an important quantity for a semiconductor electrode because it connects the
energy levels of the carriers in the semiconductor to those of the redox couple in the
electrolyte and it connects the parameters that can be experimentally determined to
those derived from solid-state physics and electrochemistry. It can generally be
expressed as
The value of the flatband potential of a semiconductor is quantitatively related to its

electron affinity n for silicon):880,953
where is defined as the energy of the conduction band at the point of zero charge
(pzc), that is, when the potential drop in the Helmholtz layer is zero. This means that
the electron affinity in the electrolyte equals the band edge plus the energy level of
the reference level on the absolute scale at the point of zero charge. Figure 1.6
schematically illustrates the relation among the different potentials and energy levels.
The conduction band edge is related to the flatband potential according to the follow-
ing equation
Therefore, by measuring the flatband potential at pzc, one can determine the energy
level of the semiconductor band in an electrolyte relative to the absolute scale or the
vacuum scale. The pzc of a silicon electrode in aqueous electrolyte is similar to that of
at about pH 2.2, since the silicon surface is generally covered with a thin layer
of oxide.716,903
1.3. POTENTIAL AND CHARGE DISTRIBUTION IN SPACE CHARGE LAYER
1.3.1. Carrier Density in Space Charge Region

The charge distribution within the space charge layer can be quantitatively
described by the Poisson equation:
where x is the distance from the surface and and are the donor and acceptor den-
sities, respectively. The electron and hole densities, n(x) and p(x), in the absence of a
current, follow the Boltzmann distribution:962
10 CHAPTER 1
where and are bulk electron and hole densities, respectively, and
At the surface where x = 0,
Three types of space charge layers, namely, depletion layer, accumulation layer, and
inversion layer, may occur in a semiconductor depending on the bias and equilibrium
conditions as shown in Fig. 1.7.
1.3.2. Depletion Layer

When the surface region is depleted of majority carriers, due to positive band
bending for n type or negative band banding for p type, a depletion layer is formed as
shown in Fig. 1.7(a). The depletion layer has no mobile carriers but has immobile
charges associated with the ionized dopants. The Poisson equation, Eq. (1.16) for
n-type material, can be reduced to
On integration with dV/dx (V = 0) = 0 one obtains

At the surface where the factor is negligible for a reasonable value

of From Gauss’s law we have
where is the total charge in the space charge layer per unit electrode surface area
and is the point in the semiconductor where the field is zero. Since the capacitance
of the space charge layer differentiating Eq. (1.23) with respect to an
expression for is obtained:
Equation (1.24) is the much-used Mott–Schottky equation, which relates the space charge
capacity to the surface barrier potential Two important parameters can be determined
by plotting versus the flatband potential at (where and the
density of charge in the space charge layer, that is, the doping concentration
1.3.3. Accumulation Layer and Inversion Layer

In the case of an accumulation layer, for example, when electrons are injected
into (instead of extracted from) an n-type sample, or holes are injected into (instead of
extracted from) on a p-type sample, the surface charge region has an excess of the
majority carriers. For n-type materials, according to Eq. (1.16),
where the first term represents the contribution of electrons in the conduction band and
the second represents the contribution of immobile positive ions. An accumulation layer
can be formed, for example, on an n-type material by applying a sufficiently large neg-
ative voltage on the semiconductor relative to the solution. If the band bending is such
that the Fermi energy moves into the band, a “degenerate surface” is formed, which
marks the transition from semiconducting behavior to metallic behavior. The thickness
of an accumulation layer is typically on the order of 100 Å. Such a layer as thin as
3 Å has been found at n-Si in acetonitrile.941
Under a depletion condition, an inversion layer is formed when near the surface
the minority carriers accumulate and are in equilibrium with those in the bulk (that is,
the consumption rate of the carriers in the electrochemical reactions is low). When an
inversion layer occurs, the density of minority carriers near the surface may exceed that
of the bulk majority carriers. Under such a condition, when n(x) < p(x) within the space
charge layer equation, the Poisson equation [Eq. (1.16)] becomes
where the first term represents the immobile ions and the second term represents the
density of minority carriers, which becomes comparable to the first term at relatively
12 CHAPTER 1
large positive potential values. The thickness of the inversion layer is on the same
order of magnitude as that of the accumulation layer, namely, about 100 Å. An inver-
sion layer does not usually occur at an electrode surface because the electrochemical
reactions generally consume quickly the minority carriers and prevent their accumula-
tion in the surface region. An inversion layer has been found to form on n-Si electrode
in a solution of containing redox couples of high potential
values.587
The capacitances for the accumulation layer and inversion layer are respectively
given by
The change of capacitance of the charged layer in the surface region of the semicon-
ductor with band bending is illustrated schematically in Fig. 1.8. For negative values
of the surface is in an accumulation and the capacitance is very large because the
thickness of the accumulation layer is thin. As becomes positive, the capacity
decreases as the thickness of the space charge layer increases. When an inversion layer
forms at a large the surface has an excess of minority carriers. In this case the mobile
minority carriers are mostly located near the surface so that the thickness is small com-
pared with the depletion layer thickness and the capacitance increases again. In the case
when minority carriers are consumed by electrode reactions and do not accumulate at
the surface, equilibrium is not established and a deep depletion layer is formed as shown
by the dashed line in Fig. 1.8.
1.3.4. Helmholtz Double Layer

The potential drop in the Helmholtz layer is defined as
where is the potential at the solid surface and is the potential at the ohp.
The charge on the solution side of the Helmholtz layer originates from the accu-
mulation of ions at the ohp, whereas on the solid side it can arise in three forms: an
accumulation of free charge, free charge trapped from the solid onto surface states, and
adsorbed ions.44 For nondegenerated semiconductors, is primarily determined by
adsorption/desorption processes between the surface and the electrolyte. The contribu-
tion from electron transfer between the surface and the bulk of the semiconductor is
negligible. This is because the amount of charge captured by the surface associated with
this transfer is on the order of or less which is very small in
comparison with the amount of charge (on the order of associated with the
adsorption/desorption processes on the surface. Thus, for non heavily doped semicon-
ductor the applied potential variation can be considered to drop entirely in the space
charge layer and the Mott–Schottky equation can be used for determination of the flat-
band potential. The contribution to the Helmholtz layer by charge transfer becomes
more important for degenerated semiconductors, which can accommodate more charge
in the bands near the surface.
When the charge on the solid surface is determined by the specific adsorption of
and which is the case for many semiconductors, is determined by the reaction
where is the adsorbed on the surface. At equilibrium the concentration of can

be expressed as
where A is a constant. Assuming the double layer capacitance, is independent of

we obtain
Equations (1.31) and (1.32) indicate that which changes linearly with will
vary slowly compared with which changes exponentially with As an
approximation we have
which shows that the Helmholtz potential varies by 60 mV per pH unit.

14 CHAPTER 1
The Helmholtz capacity of a metal electrode is typically on the order of

44 which is much greater than the value for a silicon electrode.848 It has been
reported that the Helmholtz layer capacitance of silicon in

94
KC1 aqueous solution is about In 0.1 M tetrabutylammonium perchlorate
486
(TBAP) in acetonitrile it is found to be about A larger value,
has been found for deep accumulation of a silicon in acetonitrile and the
value varies with the accumulated charge.941
1.3.5. Surface States

Surface states are the electronic energy levels that are located at the surface and
are capable of transferring charges with the solid and solutions. The surface states of a
clean surface, the dangling bonds resulting from the termination of the bulk structure
at the surface, are called intrinsic states. For silicon the dangling bonds on the surface
in solutions are generally terminated by hydrogen and other species so that there are
very few intrinsic surface states.
In addition to the intrinsic states, there are also Lewis sites and adsorbed elec-
troactive species. Lewis sites arise from the acid/base properties of the surface. A Lewis
acid site is a site attractive to electron pairs from an adsorbing molecule and a Lewis
basic site is able to donate electron pairs to an adsorbing molecule. For example,
a Lewis acid site M or a basic site N may adsorb a ion or a from water,
respectively:
Adsorption at acidic sites M causes the solution to become acidic and adsorption of
on Lewis basic sites causes the solution to become basic. Lewis sites are important in
two ways: they contribute to the Helmholtz double layer, and they result in chemical
adsorption and passivation of the intrinsic active surface sites. The surface of silicon is
dominated by basic Lewis sites as manifested by the strong hydrogen adsorption. But
the associated surface states are not active because they are located energetically in the
valence band.99
Adsorbed species other than hydrogen and hydroxyl ions that are able to give
up or accept electrons are also surface states. The reaction intermediates that are
able to act as donors or acceptors through charge transfer reactions can be viewed
as surface states. As will be described in more detail in the section on anodic beha-
vior, partially oxidized silicon atoms, i.e., the reaction intermediates,
act as transient surface states and play an important role in a range of electrode
processes.
Surface inhomogeneity such as vacancies, steps, kink sites, emergent disloca-
tions, and foreign elements in the lattice has a significant effect on adsorption, bonding
energy, and redox reactions. Such surface heterogeneity causes the energy levels of a
specific type of surface state to appear as bands because the exact chemical nature of
the surface defects will vary from point to point. In some cases electrode reaction can
result in the formation of bulk phases, for example, oxide films. The states and charges
that exist in the oxide and/or at the semiconductor/oxide interface are another form of
surface state (or interface state).
There are two kinds of surface states with respect to the energy levels of states
under filled or empty conditions: one kind has different energy levels when the surface
states are filled or emptied due to a reorganization effect, and the other, showing no
reorganization effect, has negligible change in their energy levels when the states are
filled or emptied. For example, according to Chazalviel and Rao,83 on a silicon surface
the surface states induced by deposited gold atoms belong to the second kind whereas
those associated with a surface oxide formed by the reaction between silicon and water
belong to the first kind.
It is generally difficult to quantitatively describe the surface states of a semicon-
ductor electrode considering the great diversity of surface states.486 In general, surface
states are energetically distributed in the band gap of the semiconductor with a density
expressed as and the associated charge density is a function of Under some
idealized conditions such as when the surface states are uniformly distributed in energy
or are localized at a single energy level, the charge density associated with surface states
can be quantitatively related to band bending For example, for a uniform dis-
tribution of acceptor surface states centering around an energy of (when the states
are filled to an energy there is no net surface charge), can be expressed as270
where is the total number of surface states per square centimeter. Note that when
In the case of acceptor states at a single energy level, we have
where is the degeneracy of the energy level.270

The capacitance of surface states, corresponding to the variation of the charge
stored in the surface states with is
The surface state capacitance is different from the space charge layer capacitance
and the Helmholtz layer capacitance in that there is in general no distance asso-
ciated with the surface state capacity.
Measurement of capacitance as a function of potential is most commonly used
for evaluation of the nature of surface states. For example, when the surface states are
energetically close to the conduction band they can be filled or emptied by electron
transfer with the conduction band:
16 CHAPTER 1
where is the empty state and the filled state. When the interaction with the solu-
tion can be neglected, the occupancy of the surface states is determined by the Fermi
level and the surface state capacitance can be expressed as808
where is the surface electron concentration, the frequency of the potential or current
modulation, and and the rate constants for the forward reaction and reverse reac-
tion described by Eq. (1.39). At low frequency where exhibits a
maximum as a function of potential
from which the density of surface states can be determined.
1.3.6. Fermi Level Pinning

The total charge on the solid side of the semiconductor/electrolyte interface
equals When surface states are present and their capacitance is significantly
larger than that of the space charge layer, that is, Fermi level pinning by the
surface states occurs.270 Figure 1.9 illustrates the effect of dopant concentration and
surface state density on the magnitude of surface state charges and space charge layer
charges. It can be seen that for a doping of a surface state density on the order
of which is about 1% of the surface atomic density, can result in
When the Fermi level is pinned, the change of potential will mainly occur in the
Helmholtz layer rather than in the space charge layer as in the case without Fermi level
pinning.
As a consequence of Fermi level pinning, the Schottky barrier height
at the semiconductor/electrolyte interface becomes constant with respect to redox
couples of different values. Fermi level pinning, which results in a constant barrier
height value, has been found to occur at the n-Si/acetonitrile interface containing dif-
ferent redox couples.486,935
1.3.7. Equivalent Circuit and Capacitance of Semiconductor/Electrolyte Interface

The electrical nature of the semiconductor/electrolyte interface region can be
described by an equivalent circuit of an array of resistors and capacitors arranged
according to the physical structure of the region. Figure 1.10 illustrates a typical equiv-
alent circuit of a semiconductor/electrolyte interface including the effect of the space
charge layer, Helmholtz layer, surface states, and electrolyte resistance. On application
of a voltage change of dV across the interface region, charge in the form of electrons,
holes, and ions will move in the corresponding layers, which can be characterized by
the capacitances of these layers.
The total potential change is divided between the space charge layer and the
Helmholtz layer:
By charge neutrality,
The overall capacity C is related to each of the capacities according to

18 CHAPTER 1
Using Eq. (1.42), we obtain
Thus, acts as a capacitor in series with the parallel combination of and In

many situations because of the presence of an oxide film on the surface, particularly
for silicon in non-HF electrolytes, an extra RC component is involved due to the charges
and states in the oxide and at the semiconductor/oxide interface. The characteristics of
these charges and states are discussed in Chapter 3 on silicon oxide.
In the absence of an interfacial phase such as an oxide, any potential drop is across
the space charge layer or/and the Helmholtz layer, that is, The poten-
tial distribution across the space charge layer and in the Helmholtz layer, in the absence
of surface states and interfacial layer when is determined by the doping level,
band bending, and the capacitance of the Helmholtz layer. The partitioning of the poten-
tial can be measured in terms of the potential in the space charge layer relative to the
848
total potential drop across the interface region by a partitioning coefficient,
where may vary from 0 to 1. The potential partitioning can also be measured by
the differential form in terms of the fraction of the applied potential change dV across
the space charge layer:
The two coefficients are different in that quantifies the potential partitioning of
the total potential drop while reflects the partitioning as a result of potential change.
Thus, characterizes the steady condition of the interface or the dc condition and
characterizes the transient condition or the ac condition. Figure 1.11, as calculated
by Oskam et al.,848 shows the change of these two coefficients as a function of applied
potential. It can be noted that under a negative potential corresponding to an accumu-
lation band bending, the coefficients become rather small indicating that becomes
comparable to CH and there is a significant drop of applied potential across the
Helmholtz layer. Because of the shift in potential drop from the space charge region to
the Helmholtz layer, there is a maximum band bending above which all the applied
potential drops in the Helmholtz layer. For the maximum band bending
in the accumulation regime is about 300 mV.
1.3.8. Flatband Potentials

Flatband potential is a very important parameter for characterization of a semi-
conductor/electrolyte interface as it correlates the band edges to the redox potentials
in the electrolyte. It is most commonly determined by measuring the capacitance as a
function of potential. Flatband potential can also be estimated by measurement of the

onset potential for photocurrent.
In an idealized case when the effect of the Helmholtz layer can be neglected, i.e.,
when there is a negligible amount of surface states, that is, The
total capacitance of the semiconductor/electrolyte interface described by Eq. (1.45)
becomes The interface capacitance as a function of the electrode potential then
follows the Mott–Schottky equation:
A plot of versus potential V is linear and thus the potential where the line
intersects the potential axis yields the value of and the slope yields the value of
doping level The resulting flatband potential can be used to determine the potential
of the conduction band edge at the surface
where Nc is the effective density of states in the conduction band. The Schottky barrier
height is
In practice, the determination of flatband potential according to the Mott–
Schottky equation can be affected by a number of factors, e.g., high doping concen-
tration (i.e., is not valid), the presence of a high density of surface states (i.e.,
is not valid), and the influence of diffusion-controlled processes.44,274,935 For
highly doped semiconductors, the effect of potential drop in the Helmholtz layer, even
without the influence of surface states, can be significant.274 Figure 1.12 indicates that
a significant fraction of the external potential variation can be dropped in the Helmholtz
layer for moderately and highly doped semiconductors especially near the flatband
potential at which the capacitance of the space charge layer is the largest. Because of
the contribution from the Helmholtz layer, the Mott–Schottky plot tends to be linear
20 CHAPTER 1
at relatively large bias but curved at small bias. The potential intersection of the
Mott–Schottky plot, considering the effect of the Helmholtz layer resulting from high
doping concentration, is given by
where a voltage shift from by an amount of is due to the potential

drop in the Helmholtz layer.273
When there is a high density of surface states, the Fermi level can become pinned
within the potential range in which the surface states lie. For a monoenergetic surface
state, the intersect of the Mott–Schottky plot is shifted from its true flatband potential
value as illustrated in Fig. 1.13.44 In this case the applied potential is dropped in the
Helmholtz layer and the capacitance does not change with applied potential. When the
surface states are fully occupied or fully empty, the capacitance starts to change again
with applied potential according to the Mott–Schottky relation. The slope of the central
region in Fig. 1.13 is related closely to the density of surface states. The higher the
slope is in the central region, the lower the density of states. An infinite density of states
will lead to a zero slope.44 Assume is the positive charge when the donor surface
states are fully empty and is the negative charge when the acceptor surface states
are fully occupied. The effect of the donor surface states on the flatband potential can
be described by and that of the acceptor surface states by
The presence of an oxide layer on the silicon surface can also cause the flatband
potential to shift. The potential drop across an oxide is on the order of
where and are the dielectric constants of silicon and silicon oxide, respectively, d
the oxide thickness, and the surface field.717 For a thin oxide film of about 20 Å,
and (the dielectric constant of native oxides is much larger than
that of thermal oxide due to inclusion of water molecules), Thus, the
presence of a thin oxide film such as the native oxide has in general a weak effect on
the determination of flatband potential when the oxide is not associated with a high
density of surface states.
1.4. KINETICS OF CHARGE TRANSFER
1.4.1. Basic Theories

Electrochemical reactions on a semiconductor electrode involve charge transfer
between the species in the solution and charge carriers in the semiconductor. The basic
assumption in the theories of the kinetics of charge transfer reactions is that the elec-
tron transfer is most probable when the energy levels of the initial and final states of
the system coincide.44,79 This requires that the vibration of molecules and their sur-
rounding solvation shell be slow compared with the actual electron transfer between
the electrode and redox system in the solution. Thus, the efficiency in the redox reac-
tion processes is primarily controlled by the energy overlap between the quantum states
in the energy bands of the semiconductor and the donor or acceptor levels in the reac-
tants in the electrolyte. Often, the overlap between the electronic states in the semi-
conductor and the levels in the electrolyte is unfavorable, and radical surface states that
located within the band gap become the most important for the charge transfer.
In an ideal case when surface states are absent and charge transfer proceeds
directly between the energy levels in the bands and in the solution, according to
Gerischer,79 an anodic current involving an electron transfer from a molecule in the
electrolyte to the electrode and a cathodic current involving an electron transfer from
the electrode to a molecule in the electrolyte are given by
22 CHAPTER 1
where is the number of molecules that reach the electrode surface, the transi-
tion coefficient which strongly depends on the distance of the reacting molecules from
the surface, and the densities of empty states and occupied states in the
electrode, respectively, and and the densities of occupied energy states
and empty states in the electrolyte defined by Eqs. (1.11) and (1.12). The anodic current,
consists of an electron transfer current into the conduction band, and corre-
sponding to a transfer into the valence band. Similarly, the cathodic current, consists
of the contribution from the conduction band, and from the valence band, The
contributions of conduction and valence bands to the anodic and cathodic currents are
illustrated in Fig. 1.14.
The magnitude of the currents depends on the overlap between the levels in the
semiconductor bands and those in the solution. For a material having a large band gap
the overlap generally involves only the conduction band or the valence
band, whereas for a small band gap there may be overlap of a redox couple with both
the conduction band and the valence band so that the contributions from both bands,
i.e., and are significant. In general, the electron transfer probability is mainly deter-
mined by the energy correlations between the band edges of the semiconductor, and
and principal energy levels of the redox couple, and If is closer to
than to electron transfer in the conduction band is the more likely process. In
the opposite case, electron transfer in the valence band is more likely.
In the case of the conduction band process, the density of empty states,
in the conduction band can be taken to be constant because only a few of them are occu-
pied so that using Eq. (1.11) the conduction band anodic current, can be expressed as
In the case of the valence band process, an electron transfer into the valence band
is only possible if holes are present at the surface. This means the density of empty
states equals the hole density at the surface, Thus, the valence band anodic
current is
Integrals (1.52) and (1.53) can be analytically solved only by making some assump-
tions. Assuming for the conduction band process and for the valence
band process because electron transfer occurs mainly with 1 kT of the band edges and
assuming is constant, integrals (1.52) and (1.53) can be approximated by the fol-
lowing equations for the anodic currents via the conduction and valence bands:
Similarly, one can obtain the approximate equations for the cathodic currents via the
conduction and valence bands:
For non-heavily doped semiconductors, the potential variation occurs only across the
space charge layer and the potential across the Helmholtz layer is constant. This means
that the exponential terms in Eqs. (1.54) to (1.57) are independent of applied potential,
that is, Thus, and are independent of the potential and can
be taken as constants: and The only quantities that are dependent on poten-
tial are and which are described by Eqs. (1.19) and (1.20). At equilibrium the
anodic and cathodic currents across each energy band must be equal, i.e.,
Also, the surface carrier concentrations at equilibrium can be described by and

which are given by Eqs. (19) and (20):
where is the band bending at equilibrium. At potentials departing from equilibrium,

the surface carrier concentrations are described by
24 CHAPTER 1
The current equations (1.54) to (1.57) can now be expressed in simple forms as follows:
for anodic current,
and for cathodic current,
where is the overpotential, which is a measure of the departure of the

potential drop in the space charge region, from that at equilibrium, Figure 1.15
shows the schematic i–V curves representing Eqs. (1.63)–(1.66).962 Under an anodic
polarization which provides a positive increases exponentially with and is
constant. On the other hand, under a cathodic polarization, increases exponentially
with cathodic polarization but remains constant. The exponential increase for
which depends on the supply of holes, is limited for an n-type material by the current
of holes. Similarly, the exponential increase of the electron current for p-type
material is limited by the current of electrons.
The net current is
This equation can be further simplified by neglecting the contribution due to the minor-
ity carriers. Thus, for an n-type material Eq. (1.68) becomes
The current is negative when it is cathodic and positive when it is anodic. Analogously,
the current on a p-type material when the contribution of minority electrons is negli-
gible can be expressed as
This form resembles that for the Schottky barrier at a metal/semiconductor interface
that is used in the literature.964
1.4.2. Limitations of the Basic Theories

The kinetic theory described by the above equations is derived with several basic
assumptions:962
1. Electron transfer occurs directly between the levels in the bands and the levels
in solution
2. The energies of electrons and holes in the band taking part in the charge trans-
fer are confined in a narrow interval (on the order of kT) of the band edge at
the surface.
3. The position of the band edge relative to the energy levels of the redox system
is independent of the potential change in the interface region, i.e., all the
applied potential drops across the space charge layer in the semiconductor.
26 CHAPTER 1
In many circumstances these conditions are not met and the equations are not
directly applicable to electrode reaction kinetics.86,270,274 For example, when charge
transfer via surface states plays a significant role in the electrode processes, the first
assumption is violated. In this case as shown in Fig. 1.16a, the charge transfer process
is not directly between the levels in the bands and those in the solution. In the case of
a strong band bending the thickness of the space charge layer may become very small
so that electron tunneling through the space charge layer occurs as shown in Fig. 1.16b.
Thus, the second condition is violated because the electron tunnels from energy levels
that are distant from the band edges at the surface. Violation of the third assumption
may, for example, occur when the semiconductor is highly doped so that the capaci-
tance of the space charge layer near the flatband potential is comparable to that of the
Helmholtz layer and a significant fraction of the potential drops in the Helmholtz layer.
This may also occur when the density of surface states is high and the associated charge
is comparable to that of the Helmholtz layer.
Also, although Eqs. (1.69) and (1.70) resemble those for a Schottky barrier, there
are several important differences in the physical and chemical details: (1) charge trans-
fer between a semiconductor and a solution is a slow process, whereas that between a
metal and a semiconductor is fast; (2) the diffusion of redox species in the solution
toward the electrode surface is slow whereas that of charge carriers in metal is fast; (3)
the reduced and oxidized species of the redox couple as donors and acceptors can
change independently whereas the occupied and unoccupied states of the metal cannot
be changed artificially; (4) a Helmholtz layer is present between the semiconductor
electrode and the solution whereas no such layer exists at the metal/semiconductor
interface.
1.4.3. Limiting Current

The limiting current for an n-type semiconductor electrode under a reverse bias
consists of that from majority carrier current in the conduction band and minority car-
riers in the valence band. The relative contribution of the two currents is a function of
temperature. For silicon at room temperature the limiting current in the dark is domi-
nated by the transport of minority carriers.699
When the limiting current in the dark is determined by the minority carriers, under
a steady-state condition, in the absence of additional generation in the bulk for n-type
semiconductor the hole concentration and hole diffusion current can be given by the
following equations:
where is the diffusion current, the diffusion coefficient of holes, p the hole con-
centration, the hole concentration at equilibrium, and the hole lifetime. Using the
boundary conditions at and p = 0 at (boundary between the space
charge layer and bulk) and solving Eqs. (1.71) and (1.72), the limiting (or saturation)
hole current is obtained as
where is the diffusion length.

The limiting current described by Eq. (1.73) is the maximum hole current that
can flow from the semiconductor bulk to the space charge region under a steady-state
condition. It depends on the assumption that bulk thermal generation is the only source
of the minority carriers. In many systems there may be other sources of minority carrier
generation, for example, generation in the space charge region and through the surface
states.607,717,839 In some cases the measured limiting current can be much larger than the
true limiting current due to current multiplication which generates current via injection
of carriers into the band of majority carriers.
1.4.4. Breakdown
Breakdown of a semiconductor electrode occurs when the limiting current
at reverse bias sharply increases with increasing potential. At breakdown the
electrode loses its “insulating” character and becomes “conductive.” Two types of
breakdown may occur in a semiconductor at high field: Zener breakdown and avalanche
breakdown.45,964
In Zener breakdown, the field may be so high that it exerts sufficient forces on a
covalently bound electron to free it, which creates two carriers, an electron and a hole,
to conduct the current. In this breakdown process, shown in Fig. 1.17a, an electron
makes the transition, or tunnels, from the valence band to the conduction band without
the interaction of any particles. It is essentially a band-to-band tunneling process. In
28 CHAPTER 1
this case the width of the space charge layer at a strong reverse bias becomes very thin
so that the bound electrons in the valence band can tunnel through the band gap into
the conduction band. Since the probability of the process is a strong function of the
thickness of the barrier, tunneling is only significant in highly doped material in which
the depletion layer is narrow.
In the case of avalanche breakdown as shown in Fig. 1.17b, free carriers are able
to gain enough kinetic energy from the field in the space charge layer and break the
covalent bonds in the lattice at collision. In this process, each carrier colliding with
the lattice atoms creates two carriers, which can participate in further avalanching
collisions, leading to a sudden multiplication of carriers in the space charge region. In
the energy-band diagram of Fig. 1.17b, this process can be represented as the excita-
tion of carriers from the valence band to the conduction band by absorption of the
kinetic energy of the free carriers moving in the lattice under a high field.
The field required for breakdown to occur and the mode of breakdown depend
on doping level. As the dopant concentration increases, the width of the space charge
layer decreases and the probability of tunneling increases rapidly so that Zener break-
down becomes more likely than avalanche breakdown. Zener breakdown is, in general,
involved in the electrode processes on and materials under a reverse bias.
Interface tunneling (shown in Fig. 1.17c), can also generate a large current at a
reverse bias for a semiconductor electrode/electrolyte interface when the energetic posi-
tion of a redox couple is favorable relative to the bands. Interface tunneling has been
found to be significant for silicon electrodes in HF solutions, where the dissolution

intermediates are further oxidized by injecting electrons in the conduction band, result-
ing in current that is much larger than the minority dark limiting current.8,38
1.4.5. Potential Distribution

The statically charged region at the silicon/electrolyte interface is composed of a
space charge layer in the silicon and double layer in the electrolyte. Any potential
change across the interface must be accomondated by one of the two layers or shared
by them. The partition of potential under a reverse bias when the current is very small
is determined by the relative value of the capacitance at the space charge layer, the
surface state capacitance, and the Helmholtz layer capacitance. On the other hand, under
a forward bias the partition of potential depends not only on the relative values of the
capacitances, but also on the kinetics of charge transfer across each layer. When there
is an oxide present, the potential may further be partitioned among the oxide film and
the two double layers.
In general, in the absence of an oxide the partition of the applied potential
across the space charge layer and the Helmholtz layer depends on doping concentra-
tion and current range. There are also two different potential distributions depending
on whether it is under a forward bias or a reverse bias. Under a forward bias for an
anodic process on a p-type semiconductor electrode the current density can be described
as follows724:
where is the exchange current density, and are the surface concentration of holes
under current flow and at equilibrium, and is the overpotential for the anodic reac-
tion. The total applied voltage, under anodization is the sum of the potential drop
in the Helmholtz layer, in the space charge layer, and concentration overpoten-
tial in the electrolyte,
That is,
According to the calculation by Kang and Jorne,724 the fraction of the potential drop
across the space charge region depends on the doping level and exchange current
density. Under the same current density, decreases with increasing doping level.
When the doping level is the voltage drop in the space charge region is very small,
on the order of millivolts, and most of the potential drop occurs in the Helmholtz layer.
As the doping level decreases, a relatively larger fraction of the total overpotential
occurs in the space charge region as shown in Fig. 1.18. At a very low doping level,
when the concentration of holes is very small, the potential drop occurs almost entirely
in the space charge layer. The plot of logarithmic current versus potential then gives a
60 mV/decade relationship.
30 CHAPTER 1
At the limiting current in the dark under a reverse bias, all the change in poten-
tial occurs across the space charge layer. But at current below the dark limiting current
the applied potential can be distributed over both the Helmholtz layer and the space
charge layer.717 The anodic charge transfer at a reverse bias involves two processes that
are electrically in series:
1. Electrochemical charge transfer through the Helmholtz layer according to
where is the exchange current, the overpotential across the Helmholtz

layer, and the charge transfer coefficient.
2. Electron transfer from the surface to the bulk of the semiconductor through
the space charge region with limiting current of
The total overpotential is
This equation indicates that potential change occurs mainly in the Helmholtz layer at
a current significantly less than the limiting current whereas it occurs mainly in the
space charge layer at a current near or equal to the limiting current and the Helmholtz
potential is almost fixed.
1.4.6. Current Multiplication

Current multiplication is a phenomenon associated with the limiting current when
a reaction of several charge transfer steps occurs and only a part of these steps involves
the minority carriers. For example, for an anodic oxidation of a reducing species A to
its oxidized form on an n-type material, current multiplication occurs as follows:
where A is first oxidized by a hole to become a radical Because of the closer posi-
tion of the intermediates to the edge of the conduction band, further oxidation to
is accomplished by injecting (n–m–1) electrons into the conduction band. This results
in a current that is n/(m + 1) times larger than the limiting current by the minority
carriers. It is commonly called current doubling when n/(m + 1) is 2. Current multi-
plication with a multiplication factor of 2 to 4 has been observed during the anodic
dissolution of n-type silicon in fluoride-containing solutions.908,909
1.5. PHOTOEFFECTS
When a semiconductor/electrolyte interface is irradiated with a light of energy

greater than the band gap photons are absorbed, breaking the covalent bonds holding
the electrons at atomic sites in the lattice, and electron–hole pairs are generated. Some
of these electrons and holes, especially those formed in the bulk semiconductor beyond
the space charge region, recombine with the generation of heat or emission of light.
However, the field in the space charge layer causes the separation of electrons and holes.
Thus, in the case of Fig. 1.19 the holes arrive at the surface at an effective potential
equivalent to the valence band edge and cause the oxidation of the redox couple R/O
from R to O while the electrons move into the external circuit through the semicon-
ductor electrode lead. The flow of holes and electrons in opposite directions can be
measured as photocurrent.
1.5.1. Photocurrent
The attainable photocurrent under a given illumination intensity depends on the
relative depth of light penetration in the semiconductor, the diffusion length of the
32 CHAPTER 1
minority carriers and the width of the space charge layer. As illustrated in Fig. 1.20,
the depth of light penetration is on the order of where α is the light absorption
coefficient defined as follows962:
where I is the light intensity, the light intensity entering the electrode, the frequency
of light, and x the distance from the surface into the bulk. The number of absorbed
photons per unit time and unit volume is and the rate of carrier generation for
one photon to one electron–hole pair is described by
The penetration depth of light may vary in a wide range since is a function of light
frequency as shown in Fig. 1.21.45,161 For silicon, the absorption of photons having ener-
gies greater than the band gap is almost entirely due to the generation of holes and elec-
trons. The optical penetration depth, is less than 10 nm when silicon is illuminated
with high-energy ultraviolet light whereas it is with light having a wavelength
of
Figure 1.20 shows the two extreme cases of light absorption, and
in relation to the diffusion length and the width of the space charge layer. In
the region the holes generated in the depletion layer, are transported
by the electric field to the surface where they are consumed in the electrode reactions.
Outside the depletion layer at the photogenerated holes are transported by
diffusion. The distance these nonequilibrium carriers travel during their lifetime is on
the order of the diffusion length. Thus, the holes generated deeper in the bulk,
mainly recombine before they reach the surface.
In the case of strong light absorption so that (Fig. 1.20a), the region for
photo hole generation is within the depletion layer. The photocurrent is independent of
the potential and represents the maximum attainable. On the other hand, when light
absorption is weak (Fig. 1.20b), the photocurrent is proportional to and thus
depends on the potential which determines the width of the space charge layer.
Photocurrent can generally be considered to consist of two parts, that due to gen-
eration in the depletion layer, and that from generation in the bulk,
With the assumptions that (1) there is negligible recombination in the space charge layer
and at the surface and all the carriers generated in the space charge layer are driven by
the field to the surface, and (2) the electrode reactions are sufficiently fast, using Eq.
(1.83) the contribution due to photogeneration in the depletion layer can be described
by
The photocurrent due to generation in the bulk, can be described by

34 CHAPTER 1
which means that under a steady-state condition the change of hole concentration
caused by diffusion, recombination, and photogeneration is zero. Using the boundary
conditions and p = 0 at one obtains the solution for as
Combining Eq. (1.84), (1.85), and (1.86), one obtains the expression for the absolute
value of photocurrent:
Equation (1.89) can be further simplified for the two extreme cases shown in Fig. 1.20.
For in the case of weak absorption, Eq. (1.89) can be expanded and one obtains
For strong absorption, i.e., Eq. (1.89) becomes
i.e., all created carriers contribute to photocurrent and is independent of doping

concentration and potential.
The ratio of the number of electrons transferred through the external circuit to
the number of photons striking the semiconductor surface is defined as quantum yield
or quantum efficiency:
The net photocurrent and the quantum yield are a function of a number of competing
processes1154 as shown in Fig. 1.22. For an n-type semiconductor, the externally mea-
surable current i is the difference between the photocurrent and the forward current of
electrons. The electron current is decreased to zero under certain anodic bias. While
the flux of holes to the surface is exclusively controlled by the solid-state properties,
all the other reaction steps depend on the surface properties of the semiconductor. The
holes arriving at the surface can either (i) transfer to an electron donor in the solution,
(ii) be trapped at the surface states, or (iii) recombine with electrons in the conduction
band in the depletion region or at the surface. Process (iii) does not generate current in
the external circuit, whereas process (ii) produces only transient current charging up
the surface states. Only process (i) produces steady photocurrent. The measured pho-
tocurrent can therefore be different from the flux of holes to the surface due to these
processes.
All these processes are strongly dependent on the band bending. In the absence
of surface recombination and with a fast kinetics of electron transfer, the photocurrent
increases steeply when the depletion layer starts to form and a saturation current
is quickly reached. On the other hand, with sufficient surface recombination or in
cases of slow electron transfer reactions, the apparent onset of the photocurrent
can be shifted to higher bias and the saturation current is only reached at a larger band
bending.
1.5.2. Photopotential
Photopotential or photovoltage is the potential change in the space charge layer
resulting from charge separation of photogenerated excess carriers in the field of the
space charge layer under galvanostatic conditions.86,962 It depends on the light intensity
and the original band bending in the dark. The maximum photopotential is obtained
when the energy band becomes flat. Also, depending on whether it is in depletion or
accumulation at the semiconductor surface, the photopotential can be negative or pos-
itive with changing sign at the flatband potential.278 The photopotential developed in a
depletion layer is much larger than that in an accumulation layer.
When an electrode is illuminated its potential becomes
This change of potential by illumination, is defined as photopotential,

The current measured at is the sum of electron and hole currents:
At the open-circuit potential condition the net current is zero, that is, and the
photopotential is called the open-circuit photopotential, When there is very little
surface recombination and the electrode reaction is sufficiently fast so that the con-
centration of the reagents at the surface remains equal to their equilibrium values, the
relation between open-circuit photopotential and light intensity for an n-type material
407
can be expressed for by
36 CHAPTER 1
where is the bulk minority carrier density and is the increased carrier density due
to light excitation. This equation indicates that large photovoltages are already obtained
at rather low intensities of light, because the carrier density created by light excitation
can easily exceed the minority carrier density. Since is a constant for a given
material and is proportional to the light intensity the above equation can be
reduced to
where b is a positive constant. The same expression as Eq. (1.95) can be derived for
962
It can be seen that is always negative because when an n-type semi-
conductor is illuminated, the band bending decreases. At relatively low intensities
of illumination, i.e., the absolute value of the photopotential increases
linearly with increasing whereas at high intensities, i.e., the change is
logarithmic.
When the photoeffect is limited mainly by the bulk recombination process, i.e.,
the diffusion of the minority carriers, the photopotential can be quantitatively related
to the minority diffusion coefficient, and the diffusion length, at a photocurrent
of 275:
When the redox species in the solution do not interact with the semiconductor
surface and the band edge at the surface is fixed with respect to the redox potential,
the changes in the redox potential in the solutions will result in identical
change in the photovoltage on an illuminated electrode. The presence of a redox
couple with more negative than for n-type material or more positive than
for p type results in an interface with no photovoltage. For more positive than
for n type or more negative than for p type, the photoeffects are also expected to
be minimal. Only for situated within the band gap can a large photovoltage be
expected.
For a semiconductor that is at equilibrium with a redox couple in the solution,
the band bending is determined by the redox potential. The maximum attainable pho-
tovoltage is the size of band bending, which is ideally related to the barrier height,
Under an equilibrium condition, generally does not
move into the valence band to form an inversion layer and thus the maximum band
bending cannot exceed that is, about the band gap.275 Thus, theoretically, the
maximum attainable photovoltage is close to the band gap when the redox potential is
near the valence band edge at the surface.
As a semiconductor/electrolyte interface for photoenergy conversion, the objec-
tive is to position close to for n-type material for photoanodes, or close to
for p type for photocathodes to achieve the largest band bending in the dark and thus
the maximum photovoltages. The maximum open-circuit photovoltage is equal to the
amount of band bending or the barrier height, which can be given as270
The theoretical maximum value is, however, not practically achievable due to many
kinetic processes at the semiconductor/electrolyte interface such as surface recombi-
nation and majority current. As a result, for a given silicon/electrolyte interface, con-
stant photovoltage over a wide range of redox potentials can occur, that is, band bending
is independent of the redox potential value. Thus, photovoltages in practical systems
are often not indicative of the barrier height.
1.5.3. Efficiency of Energy Conversion

The power efficiency of a photoelectrode can be characterized by the open-circuit
potential (OCP) and short-circuit current:675,812
where is the photovoltage at the photocurrent At OCP where is the maximum,

but is zero so that is zero. Large is obtained when both and are large.
In the application for photo power conversion devices, the objective is to maximize
The maximum theoretical solar energy conversion efficiency for a silicon elec-
trode is found to be about 25%. Systems with efficiency up to 16% have been realized
experimentally on silicon/liquid junctions.275,917 The primary source of inefficiency is
the presence of surface states at the silicon/electrolyte interface that act as recombina-
tion centers.548,622
1.5.4. Surface Recombination

Surface states, like the impurity levels in the bulk, may take part in the recombi-
nation of electrons and holes, which significantly affects the photoelectrochemical
processes.158,182,962 As described earlier, there are many different kinds of surface states
in terms of their physical nature and energetic distribution in the band gap. Surface
recombination may thus have very different rates depending on the nature of the surface
states. Also, of the total number of surface states, only some of them participate in the
processes of recombination.962 For highly doped material under a large band bending
(deep depletion), surface recombination by tunneling from the surface states to the
bands may also be significant.88
For a simple case where the surface recombination involves monoenergetic
surface states the basic processes are illustrated in Fig. 1.23.962 represents the flux
due to capture of electrons from the conduction band, the emission of electrons into
the conduction band, the capture of holes from the valence band, and the emis-
sion of holes into the valence band. The recombination rates of electrons, and of
holes, can then be expressed as
At equilibrium and the recombination rates are zero. Equation (1.98)

can be further expressed as
38 CHAPTER 1
where and are the capture coefficients for electrons and holes at the surface,
is the concentration of surface states, and are the concentra-
tions of electrons and holes at the surface with and the concentrations at equilib-
rium, and are the concentrations of captured electrons and holes in the surface
states, and is the degree of filling of the surface states. In the steady state,
and Eqs. (1.99) and (1.100) become
where For a nondegenerated material, suppose that (1) quasineutral-

ity is fulfilled at the boundary of the space charge layer, so that the concentra-
tions of excess electrons and holes are equal and (2) the departure from
equilibrium is insignificant, that is, then the surface recombination
velocity, which is defined as can be obtained from Eq. (1.101) as
The recombination velocity, which characterizes the recombination process, may vary
over a wide range, from 1 to at room temperature.962 Surface recombina-
tion centers that can be described by the one discrete recombination center model have
been found to exist in different silicon/electrolyte systems.61,182,278,808 The states that
can exchange charge carriers with only one of the bands are traps for electrons or
holes. Surface states that contribute to the interface capacitance but do not act as the
recombination center have been found on silicon electrodes in both fluoride and non-
fluoride solutions.93,808
1.6. OPEN-CIRCUIT POTENTIAL
OCP, or the rest potential of an electrode, is the potential of a freestanding elec-

trode without electrical connection to any other conducting materials. Thus, at OCP
there is no net current flow in or out of the electrode. OCP of an electrode is deter-
mined by the kinetic state of the electrode. It is the most easily measurable electro-
chemical parameter and at the same time is the most convoluted quantity as it is
determined by all the kinetic factors in the system. The electrode at OCP can be at an
equilibrium state or a nonequilibrium state depending on the nature of the particular
electrode/electrolyte system and the reference time scale.
If no net chemical change occurs in the system during the time of measurement,
the electrode at OCP can be regarded as at an equilibrium condition. This occurs, for
example, as shown in Fig. 1.24a, where the Fermi level of the semiconductor has the
same value as the reversible potential of a single redox couple in which the forward
current equals the reverse current In many situations and almost always for silicon,
the electrode/electrolyte interface at OCP is at a nonequilibrium condition and a net
chemical change in the system occurs when two or more redox couples are present in
the electrolyte. OCP in these situations is a mixed potential. This is shown in Fig. 1.24b,
where OCP is established with two redox couples, and with
but In this case, a net chemical change is
involved since and and a net amount of reduced species 1 and
oxidized species 2 is produced for a given time peorid.
When one of the redox couples is associated with the dissolution of the electrode,
OCP is also called the corrosion potential and the net dis-
solution rate at OCP is the corrosion current, This is generally the case
with a silicon electrode at OCP in aqueous solutions because the thermodynamic poten-
40 CHAPTER 1
tial of silicon is much more negative than that of water. OCP of silicon in water is gen-
erally a corrosion potential. The rate of corrosion, that is, the dissolution rate at OCP,
can be measured in terms of corrosion current. It can be estimated by determining
the polarization resistance at OCP, which was originally formulated by Stern and
Geary1148,1149 based on the mixed potential concept for an electrode under an activa-
tion-controlled condition. In the case where the corrosion rate of the semiconductor is
limited by the minority carriers, the corrosion current simply equals the limiting current
as illustrated in Fig. 1.25.962
As can be seen in Fig. 1.25, the corrosion potential may vary not only with the
value of but also with the slopes of the anodic and cathodic curves as well as the
limiting current values. Since the limiting current is sensitive to the lighting condition,
OCP or corrosion potential varies with the background photonflux. As stated above,
the corrosion potential is determined by the anodic current and cathodic current. Any
changes, such as surface preparation, solution composition and concentration, pH,
temperature, time, convection, aeration, background lighting, and so on, that affect the
anodic and the cathodic reactions will affect the value of the corrosion potential. Thus,
the corrosion potential as well as the corrosion current can vary greatly, depending on
the specific conditions.
The change of the corrosion potential in either the anodic or the cathodic direc-
tion may correspond to a decrease or increase in the corrosion current. The variation
of the corrosion potential and corrosion currents under various conditions can be gen-
eralized using schematic polarization curves in Fig. 1.26.970 The corrosion potential of
an active electrode in a solution is and are the corrosion
potentials under changed conditions.
In the case where the anodic dissolution is inhibited, e.g., by surface adsorption
of a chemical species, the anodic curve becomes This will result in a more positive
corrosion potential (from to if the cathodic reaction remains unchanged.
In such a situation the corrosion current is reduced with a more positive potential rela-
tive to the original value. On the other hand, if the anodic dissolution kinetics remains
unchanged but the rate of the cathodic reactions is changed from curve to curve
the potential also becomes more positive (from to However, in this case the
corrosion current is increased with a more positive potential.
The anodic curve becomes in Fig. 1.26 when the surface is passivated. If the
cathodic reaction is unchanged, the corrosion potential of the electrode becomes
which is more positive than The corrosion current is generally much smaller than
that at
The corrosion potential values that are much more positive than are usually
associated with both the passivation of the electrode surface and the presence of
oxidizing agents in the solution. The cathodic and anodic polarization curves in such a
solution are illustrated by and the coupling of which yields the corrosion poten-
tial The dissolution rate in this situation is usually very low. However, if the
surface is not passivated, the dissolution rate can be very high in the presence of oxi-
dizing agents, which is the case for etching of semiconductors in solutions containing
oxidizing agents. This can be appreciated by coupling curves and in Fig. 1.26.
A decrease in potential can be caused by either faster anodic dissolution kinetics
or a slower cathodic reaction. For example, a decrease in corrosion potential due to
deaeration of a solution is usually observed. Deaeration removes the dissolved oxygen
and thus reduces the cathodic reaction rate.
42 CHAPTER 1
1.7 EXPERIMENTAL TECHNIQUES
Electrochemical techniques generally involve the measurement of the electrical

signal of the semiconductor/electrolyte interface since the elemental step in an elec-
trochemical reaction is charge transfer across the charged layers at the interface. In
these techniques, the electrode surface is typically perturbed by a signal of potential,
current, or light, and the corresponding potential, current, or light, along with the phys-
ical and chemical changes that take place on the surface, is measured. The physical and
chemical nature of the interface layers and kinetics of the electrode reactions can then
be analyzed. The theories and methodologies employed in electrochemical techniques
are well documented.44,932,962–964
The properties of silicon electrodes has been investigated using a wide range of
experimental techniques. Specific examples are the determination of:
Steady-state or dynamic potential and current relationships to provide infor-

mation on the energetic dependence of electrode reaction rate2,21,34,271,939
Current or potential versus time to provide information concerning stability of
reactions and transition or transformation of surface condition74,77,286,291,860,874
Capacitance–potential relationship to reveal information on the energetic
position of semiconductor bands and surface states, especially the flatband
potential61,93,475,479,716,841,942
Impedance–frequency relationship to yield information on the resistive and
capacitive elements of the current path9,139,486,951
Current on rotating electrode to give information regarding mass transport
in electrolyte14,34,939 and on rotating ring-disk electrode for reaction
intermediates21,695
Short-circuit current of a p–n junction to discriminate between a conduction
and a valence band process34,969
Photocurrent to provide information on quantum efficiency and types of
carriers11,73,74,873
Impedance under illumination to generate information on recombination
processes93,808,932
Photocurrent with subband light to provide information on surface states400,935
Optoelectric impedance (photopotential or current response to modulated illu-
mination) or electro-optic impedance (photoresponse to potential modulation)
for information on the contribution of minority, majority, and injected carriers
in recombination and multiplication626,932
Microwave conductivity for instantaneous minority concentration and flat-
band potential597,616
Optical second-harmonic generation by pulsed laser for structural information
on interface606
Defletometric technique for electrolyte concentration gradien near the solid
electrode surface855,862
Bending beam method for stress of oxide film on the surface1058
Micropatterned samples for determination of carrier distribution and lateral
mobility across the electrode or near the surface585,587
In addition to the electrochemical techniques, many insitu and exsitu surface analyti-
cal techniques are used in studies of silicon electrodes, such as ellipsometry for
determining thin surface film thickness,98,200,240,404 infrared spectroscopy for surface
adsorption,215,227,395,409 XPS126,260,424 for surface composition, SEM12,247 and STM223,234 for
surface morphology, TXRF for surface concentration of metals,135 and AFM for surface
morphology.120,135
One particularly important aspect in electrochemical measurements is surface
treatment (or surface preparation). The electrochemical behavior of silicon can be
strongly affected by surface treatment procedures.87,161,600,717 The most important goal
of surface preparation is to produce a chemically clean and physically homogeneous
surface that is reproducible and relatively stable so that representative and reproducible
results can be generated. Since silicon is usually covered with a native oxide of various
thicknesses and chemical compositions in the air or in aqueous solutions, most surface
treatments are designed to remove this initial surface oxide film. The most common
procedure used in surface treatment of a silicon electrode includes a dip of the elec-
trode in a HF solution. A very common treatment of silicon electrodes for electro-
chemical experiments employs concentrated HF solutions such as 48% HF at room
temperature with dipping time ranging from 10 to 60s.161 Also, addition of oxidizing
agents such as to the HF solutions is used to enhance the etching rate. The surface
after a dip in HF solutions is typically terminated with hydrogen (see Chapter 2 for
details). As will be seen in the rest of this book, the surface condition resulting from
preparation by a specific cleaning solution may still vary to a wide extent due to the
many poorly controllable material and procedure factors.

Basic Theories of Semiconductor

Uploaded by

Copyright:

Available Formats

Basic Theories of Semiconductor

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Basic Theories of Semiconductor

Uploaded by

Copyright:

Available Formats

1

Basic Theories of Semiconductor

1.2. ENERGETICS OF SEMICONDUCTOR/ELECTROLYTE INTERFACE

The electronic conductivity of semiconductors, as expected from the band struc-

When a semiconductor is brought into contact with an electrolyte, equilibrium

1.2.1. Energy Levels in Semiconductor

where is a reference energy called the Fermi energy. At meaning that

At equilibrium, Eqs. (1.4) and (1.5) become

1.2.2. Energy Levels in Electrolyte

1.2.3. Distribution of Energy Levels in Electrolyte

1.2.4. Energy Levels at Semiconductor/Electrolyte Interface

materials. The magnitude of the Helmholtz potential is determined by the adsorption/

The value of the flatband potential of a semiconductor is quantitatively related to its

1.3. POTENTIAL AND CHARGE DISTRIBUTION IN SPACE CHARGE LAYER

1.3.1. Carrier Density in Space Charge Region

1.3.2. Depletion Layer

On integration with dV/dx (V = 0) = 0 one obtains

At the surface where the factor is negligible for a reasonable value

1.3.3. Accumulation Layer and Inversion Layer

1.3.4. Helmholtz Double Layer

where is the adsorbed on the surface. At equilibrium the concentration of can

where A is a constant. Assuming the double layer capacitance, is independent of

which shows that the Helmholtz potential varies by 60 mV per pH unit.

The Helmholtz capacity of a metal electrode is typically on the order of

reported that the Helmholtz layer capacitance of silicon in

1.3.5. Surface States

In the case of acceptor states at a single energy level, we have

where is the degeneracy of the energy level.270

from which the density of surface states can be determined.

1.3.6. Fermi Level Pinning

1.3.7. Equivalent Circuit and Capacitance of Semiconductor/Electrolyte Interface

The overall capacity C is related to each of the capacities according to

Using Eq. (1.42), we obtain

Thus, acts as a capacitor in series with the parallel combination of and In

1.3.8. Flatband Potentials

function of potential. Flatband potential can also be estimated by measurement of the

where a voltage shift from by an amount of is due to the potential

1.4. KINETICS OF CHARGE TRANSFER

1.4.1. Basic Theories

Also, the surface carrier concentrations at equilibrium can be described by and

where is the band bending at equilibrium. At potentials departing from equilibrium,

and for cathodic current,

where is the overpotential, which is a measure of the departure of the

1.4.2. Limitations of the Basic Theories

1.4.3. Limiting Current

where is the diffusion length.

found to be significant for silicon electrodes in HF solutions, where the dissolution

1.4.5. Potential Distribution

where is the exchange current, the overpotential across the Helmholtz

The total overpotential is

1.4.6. Current Multiplication

When a semiconductor/electrolyte interface is irradiated with a light of energy

The photocurrent due to generation in the bulk, can be described by

For strong absorption, i.e., Eq. (1.89) becomes

i.e., all created carriers contribute to photocurrent and is independent of doping

This change of potential by illumination, is defined as photopotential,

1.5.3. Efficiency of Energy Conversion

where is the photovoltage at the photocurrent At OCP where is the maximum,

1.5.4. Surface Recombination

At equilibrium and the recombination rates are zero. Equation (1.98)