Solid State Communications 124 (2002) 443–447

www.elsevier.com/locate/ssc

Why tight-binding theory?

Walter A. Harrison*
Department of Applied Physics, Stanford University, Stanford, CA 94305, USA
Received 1 July 2002; received in revised form 9 August 2002; accepted 19 August 2002 by M. Cardona

Abstract
In the context of computational physics other methods are more accurate, but tight-binding theory allows very direct physical
interpretation and is simple enough to allow much more realistic treatments beyond the local density approximation. We
address several important questions of this last category: How does the gap enhancement from Coulomb correlations vary from
material to material? Should the enhanced gap be used for calculating the dielectric constant? For calculating the effective mass
in k-dot-p theory? How valid is the scissors approximation? How does one line up bands at an interface? How should we match
the envelope function at interfaces in effective-mass theory? Why can the resulting quantum-well states seem to violate the
uncertainty principle? How should f-shell electrons be treated when they are intermediate between band-like and core-like? The
answers to all of these questions are given and discussed.
q 2002 Elsevier Science Ltd. All rights reserved.
PACS: 71.10.–w; 71.15.Fv; 71.20.Mq
Keywords: A. Semiconductors; A. Surfaces and interfaces; C. Electronic band structure; C. Electron–electron interaction

1. Introduction solids, including semiconductors. This is the realm of

It is most appropriate to raise the question about tight-
binding theory in this volume honoring the years of
achievement of Professor Salim Ciraci. It was his thesis
work [1] on the bond-orbital model, in 1974 at Stanford
University, which marked the beginning of the renaissance
of tight-binding theory. This theory had its origins already in
the thesis of Felix Bloch [2] in 1928, but was generally
considered to be too naive to be worth considering. In a
similar way the nearly-free-electron theory of Bloch’s thesis
was not considered meaningful until the advent of
pseudopotential theory. In Professor Ciraci’s research it
became clear that tight-binding theory provided a clear
understanding of the electronic structure of the tetrahedrally
coordinated semiconductors, and a key to simple calcu-
lations of the properties of these systems. The subsequent
developments, including much of Ciraci’s work, have
utilized more accurate methods and more complete
calculations to obtain properties of the entire range of
* Tel.: þ1-650-723-4224; fax: þ 1-650-725-2189.
E-mail address: walt@stanford.edu (W.A. Harrison).
0038-1098/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 3 8 - 1 0 9 8 ( 0 2 ) 0 0 5 4 9 - 5
computational physics where the central goal is high
accuracy and as complete a solution as possible of the
quantum-mechanical problem of the many-electron ground
state, or of excited states. The starting point is generally a
local density approximation (LDA) which reduces the
problem to a one-electron context, though sometimes
corrections to the LDA are made afterward.
Tight-binding theory does not fare well in the context of
computational physics. It is based on the idea that one-
electron states of molecules and solids can be written as
combinations of a small number of states of the constituent
atoms. With such a limited basis it can never be very
accurate and its use usually involves assumptions and
approximations of the type that computational physicists
wish to avoid. On the other hand, it has the advantage that
the results and the theory itself are very easy to understand.
For that reason most results of detailed computational
treatments of electronic structure are interpreted in terms of
the concepts of tight-binding theory. Further, it has been
possible to use the fact that both tight-binding theory and
the opposite extreme, nearly-free-electron theory, are
meaningful to provide the central parameters for both
444 W.A. Harrison / Solid State Communications 124 (2002) 443–447
theories [3,4]. Then tight-binding theory becomes an
independent, though not so accurate, method for predictions
in all of the areas covered by LDA.
Another aspect, upon which we wish to focus here, is that
by simplifying the representation of the electronic states so
greatly it is sometimes possible to treat the many-electron
aspects more realistically than can be done conveniently in
the context of a full computational treatment of the
electronic states. We begin with a discussion of the familiar
enhancement of the band gaps over those from LDA.
2. Energy-gap enhancement
The correction to the band gap provides a very clear
illustration of the contrast between the two different
theoretical goals. In 1983 it was pointed out by Sham and
Schlüter [5] and by Perdew and Levy [6] that LDA, if
correctly carried out, leads to band gaps which are
significantly smaller than the experimental ones. This
came as a shock to many of us since computational
physicists at the time had been adjusting some parameter
in their LDA calculations to obtain the observed gap,
concealing this fundamental difficulty. From the papers by
Sham and Schlüter and by Perdew and Levy it was seen that
there was a discontinuity in the energy-dependent exchange
and correlation potential which enhanced the gap. It was
subsequently incorporated by computational physicists in
their calculations.
It is significant that this description gives no idea how
large the correction is, one half an electron volt in silicon, or
whether it would be larger or smaller in germanium, or what
it might be in rock salt or other systems. The computational
physicist can tell you the answer. So can tight-binding
theory, but it can also allow you to understand why.
An LDA calculation of the electronic structure of a
silicon atom will give an energy 1p for the valence p-state,
approximately equal to the 8 eV energy required to remove
the corresponding electron from the free atom, or the energy
gained in adding an electron to the ion, the atom with one
electron removed. Clearly if we were instead to add an
electron to the neutral atom, we gain a very much smaller
electron affinity because of the extra Coulomb repulsion
from the additional electron in the atom. That repulsion is
about
U ¼ e2 =ra < 7 eV; ð1Þ
with ra an atomic radius of some 2 Å. Similarly if we have a
collection of well-separated silicon atoms, removing a p-
electron from one atom and placing it in a p-state of a distant
atom costs an extra energy U, though the eigenvalue 1p is the
same. It is also true that if these atoms are part of a solid with
dielectric constant, 1 < 12 for silicon, the potential in Eq.
(1) is reduced to e 2/(1r ) so the energy required exceeds the
difference in the LDA eigenvalues by U/1 < 7/12, about
one half electron volt. This not only gives us approximately
the right value, it tells us that it will be about the same for
GaAs with a dielectric constant of 1 ¼ 11 and similar
atomic radius, but smaller for germanium with 1 ¼ 16: The
enhancement will be much larger for rock salt with a
dielectric constant of some 1.7 or for an inert-gas solid with
1 near one. (For details, see Ref. 3, 341 ff.) How much use
was it, in contrast, to know that the enhancement arose from
a discontinuity in the energy-dependent exchange and
correlation potential?
3. The susceptibility and the effective mass
Should we use this enhanced gap in the expressions for
dielectric susceptibility and effective mass, both of which
contain the gap? This question was addressed recently by
Lew Yan Voon and Harrison [7]. The dielectric suscepti-
bility arises primarily from polarization of individual bonds,
admixture to each bond orbital of an antibonding orbital in
the same site. However, there is no extra Coulomb repulsion
when the admixture is from the same site, just as there is no
correction to the LDA eigenvalues in transfer of electrons
between orbitals in the same atom. The main term in the
dielectric susceptibility should therefore be calculated using
the LDA bands without enhanced gaps. It is not easy to
confirm this by comparison of susceptibilities from LDA
calculations and experimental values because of other
approximations, such as the neglect of local-field correc-
tions, in the calculations. The finding that the unenhanced
gap is appropriate for calculating susceptibilities would
suggest to us that the enhancement should not be included in
the band calculation itself, unless one is willing to restrict
discussion to quasiparticle excitations.
The shift in energy of excitation of U/1 would seem to be
independent of wavenumber, suggested also by the idea of a
discontinuity of the exchange and correlation potential in
the gap. Thus we might expect we could use a ‘scissors
operation’ in the gap to uniformly raise the energy of LDA
conduction bands to obtain the corresponding excitation
energies. On the other hand, the effective mass in the
conduction band can be obtained from k·p theory, with a
squared matrix element of the gradient operator between
valence and conduction-band states, divided by the band
gap. In order to learn which gap to use, Lew Yan Voon and
Harrison [7] returned to a simple tight-binding chain of
atoms, with s-like states forming a conduction band and p-
like states forming a valence band. When the band gaps are
small, so the k·p shifts in the mass are large, the dominant
terms arise from coupling between p-states and neighboring
s-states [3,4], a coupling which transfers electronic charge
from p-states on one atom to s-states on neighboring atoms
with fully occupied p-states. The coupling to s-states on the
same site is much smaller [3,4]. Thus the enhanced gap
should be used, as concluded by Cardona’s group [8] on the
 W.A. Harrison / Solid State Communications 124 (2002) 443–447
basis of experimental measurements of the effective mass.
Use of the enhanced gap is contrary to the use of the scissors
operation since it changes the band curvature rather than
shifting the band as a whole. Again, it appears best to keep
the LDA bands, and correct each property as needed.
4. Matching bands and states at interfaces
In discussing the electronic structure of interfaces
between semiconductors, the question arises immediately
of how to set the energy scales, which are generally arbitrary
in a band calculation, relative to each other on the two sides.
Tight-binding theory would at first seem to answer the
question immediately since the band structures are obtained
relative to the same set of atomic energy levels for all
constituent atoms, giving a ‘natural’ scale for the bands.
However, more careful thought [9] indicates that this does
not set the correct scale at an interface. If the two hybrid
orbitals forming the interface bond between the two
semiconductors have different energy, a polar bond will be
formed. This produces a dipole which shifts the bands on the
two sides relative to each other. It is not difficult to see [9]
that this has the effect of reducing the difference in hybrid
energies by a factor of the reciprocal of the large
semiconductor dielectric constant. Thus the bands shift till
the hybrid energies are nearly the same on both sides. As a
consequence, we can make a table (Table 19-1 of Ref. 3) of
the valence-band maxima relative to the hybrid energy (the
average hybrid energy in the case of compounds such as
GaAs, equal to the average of the energy of the eight valence
orbitals for the system), and the offset of the valence bands
is obtained by subtracting these tabulated values. The
conduction bands are then given using the enhanced gap on
each side of the interface if we wish to discuss quasiparticle
excitations. This offset of the valence-band orbitals could
have been obtained from a full multilayer LDA band
calculation, carried out self-consistently with alternate
layers of the two semiconductors. If the layers were
sufficiently thick, the offset should be obtainable from the
resulting states.
The use of the effective mass concept is valid and useful
when we work with states near the band edge in a
semiconductor, and where the dynamics of wave packets
is accurately describable by an effective-mass Hamiltonian.
It has been widely used in treating quantum wells and other
systems with interfaces between epitaxially grown semi-
conductors. Some workers have questioned the validity of
effective-mass theory in quantum wells so that construction
of a packet within the well would require almost the entire
band. However, tight-binding theory shows that the
effective-mass bands remain valid, and the boundaries
only affect what wavenumbers enter, even for wells
containing as few as two atomic planes. The only possible
445
difficulty comes in the finding of realistic boundary
conditions.
The immediate suggestion for matching effective-mass
envelope functions at an interface would be to match c and
7c at the boundary, as in ordinary quantum theory.
However, we may see immediately that this is not correct
if the mass differs on the two sides [10]. With a current
operator given by 2ðe"=imp Þcp 7c the current is then not
continuous across the interface. Of course one can use
continuity of the current as one condition, but the second
remains uncertain, the most common choice currently
being to match c and (1/m p)7c, sometimes called the
BenDaniel – Duke [11] form or the Bastard [12] boundary
condition.
We can also use tight-binding theory to find the correct
form [13], though this is disappointing at first. We can
construct simple tight-binding s-bands for each side of an
interface with nearest-neighbor coupling Vsss on the two
sides to correspond to different effective masses. We can
couple them at the interface and see what matching
conditions on the envelope function would give the correct
result. The disappointment is that we find the c on the right
is given by a constant times c on the left, plus another
constant times 7c on the left. 7c is similarly written with
two more constants. The only way simple conditions will be
obtained is if we assume that the Vsss across the interface is
given by the geometric mean of the Vsss on the two sides
[13]. That is a reasonable choice and the conditions become
a continuous (1/m p)1/2c and a continuous (1/m p)1/27c [13,
14] We are not guaranteed that this is a correct matching, but
if there is a correct simple choice this is it, and other forms-
such as the BenDaniel – Duke, or Bastard, form-are certainly
incorrect at least in this case where it could be tested.
It is interesting to explore the consequences when the
effective-mass change is large, as for a very small mass
outside a quantum well. Then the envelope wavefunction for
the lowest quantum-well state becomes nearly constant
within the well and drops discontinuously to a very small
value outside. The state is analogous to the standing sound
wave in an open organ pipe. In this case the energy of the
confined state is very near zero in the effective-mass band
and there is no confinement energy as we might anticipate
from the uncertainty principle for a state localized in a
thickness L with a mass m p, an energy "2 k2 =2mp with k ¼
p=L: Of course there is no difficulty. There is plenty of
kinetic energy in the periodic part of the Bloch function but
we have been misled by applying a vague ‘principle’ to an
unfamiliar circumstance. We have confirmed this lack of
confinement energy for an easily soluble illustration, a
Kronig – Penney model for a finite chain, with strong
attractive delta functions [14]. The lowest state is a sum of
states m1=2 e2mlx2xj l with nearly equal coefficients. The
effective mass inside is large and free-electron outside, and
the envelope function is nearly constant inside. The apparent
violation of the uncertainty principle could be very useful in
some application and is missed completely if one uses an
446 W.A. Harrison / Solid State Communications 124 (2002) 443–447
incorrect matching condition, such as the BenDaniel – Duke
form.
5. Correlations in f-shell metals
The principal defect of the LDA arises from the
introduction of self-direct and self-exchange interactions,
as in Hartree– Fock Theory (e.g. Ref. 3, 593ff). In Hartree –
Fock these cancel exactly and provide a major simplification
in yielding the same Hamiltonian for every electron state. In
LDA this very large self-interaction is added to the much
smaller exchange between different orbitals and approxi-
mated by a function of electron density. When Coulomb
effects are large, this makes the LDA results very inaccurate.
Modifying the density functional, or adding gradient
corrections, does not even address the question. Making
self-interaction corrections [15] can rectify the errors in
principle, while retaining the more accurate aspects of the
LDA but we prefer to proceed with tight-binding theory
which makes it possible to omit the self interactions from
the outset.
A familiar case where Coulomb correlations become
dominant is directly describable in tight-binding theory. It is
a pair of atoms, such as lithium, in which a single valence s-
orbital of energy 1s on each atom is important. The two
orbitals are coupled by a Vsss which depends on the
internuclear distance d and if the two valence electrons are
placed on the same atom there is an excess energy.
U p ¼ U 2 e2 =d: ð2Þ
We have actually stripped this system down to the point that
we may find an exact solution of the correlated tight-binding
problem. There are only six two-electron states, spin-up and
spin-down on atom one, spin-up on atom one and spin-down
on atom two, etc. With reflection symmetry between atoms
we may reduce the solution of the six-by-six Hamiltonian to
quadratic equations, leading to a total ground-state energy of
(Ref. 3, p. 595)
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 spacing. In the rare earths and heavy actinides, the simple-
Up Up 2
1total ¼ 21s þ 2 þ4Vss2 s : ð3Þ
2 2
This leads to the LDA result, 21s þ U p =2 þ 2Vsss (note
Vsss , 0) at very small spacing where V2sss is much greater
than U p2, including the U p/2 from the 50% probability of
both bond electrons lying on the same atom. This result also
properly segregates the electrons as 21s at large spacing.
There is an additional repulsive interaction between the
atoms arising from the nonorthogonality of s-states on
neighboring atoms [3] which we need not consider here.
This two-atom result carries some important qualitative
and quantitative messages. We note that the force between
the two atoms, the negative of the derivative of energy with
respect to d, is reduced at all spacings, but only partially, and
it does not drop discontinuously. We would not do so badly
in estimating it from LDA and certainly it would be
inappropriate to decide that at some point the states were
‘core-like’ and to drop them from the calculation. We would
like to carry these answers over to important solid-state
cases, such as f-shell metals, where we know that electron
correlations become central.
The difficulty is that we cannot extend this exact solution
to the state of many electrons. On the other hand, it is often
fruitful even within LDA to simplify the electronic structure
to something like a two-level system. One step in the
calculation of the total energy in any system is the
summation of the energies of occupied electronic states.
This can also be done in a Friedel approximation in which
the density of states for the f-bands is replaced by a
rectangular distribution of 14 states per atom. Then the only
parameter needed is the width Wf in energy of the
distribution. If we are willing to make that approximation
in the context of LDA, the energy gain per atom due to
formation of f-bands for the Zf electrons per atom is
simply [3]
Wf
Etot ¼ 2Zf ð14 2 Zf Þ ð4Þ
28
with Wf playing the same role that Vsss played in our two-
level example. By comparing results for the two limits we
can see that the effect ofqffiffiffiffiffiffiffiffiffiffiffi
the Coulomb
ffi correlations is to
replace Wf in Eq. (4) by Uf2 þ Wf2 2 Uf ; where Uf is the
extra Coulomb energy from placing two electrons in an
atomic f-state rather than having one in an atomic s-state [3].
We have sacrificed some accuracy in the LDA aspects of the
electronic structure, but may have been able to make a very
realistic description of the central many-electron aspects.
We see that the correlations effectively broaden the bands,
but reduce their effect in holding the atoms together.
This has provided a novel view of the nature of f-shell
metals [3]. The electron states are separated into a simple-
metal part and an f-electron part, the latter given by Eq. (4)
with the modified Wf. The simple-metal part of the energy is
minimum at a large spacing and the f-electron part at a small
metal terms lead to a large spacing and, at that spacing, Wf is
so small compared to Uf that the f-levels are indeed core-
like, with the occupied levels well below the Fermi energy
and the empty levels Uf higher, well above the Fermi level.
The resulting states with local moments are a result of the
large spacing, not the cause of it. In the light actinides, Wf
becomes so large that the partially filled f-band contracts the
spacing just as the corresponding d-bands cause the high-
density structures of the transition metals, only slightly
weakened by the electron correlations. In between these
extremes, in plutonium for the actinides and cerium for the
rare earths, the two minima are distinct and competitive.
Both high-density and low-density structures exist depend-
ing upon pressure and temperature, and the properties of
each form follow from the density which occurs.
 W.A. Harrison / Solid State Communications 124 (2002) 443–447
6. Conclusion
We return to our starting question: why tight-binding
theory? Here it replaced the discontinuity in the energy-
dependent exchange and correlation potential by a U/1 from
which we could estimate the gap enhancement for any
system. It is approximately correct to use that enhanced gap
in calculating m p with the k·p formula, but the unenhanced
gap is better for calculating the dielectric susceptibility. It
aligned the two sets of bands at an interface by matching
hybrid energies. If the effective mass changes across an
interface, we should share the discontinuity as (1/m p)1/2
between the value and the slope, though this gives quantum-
well states which appear to violate the uncertainty principle.
The tight-binding view of the f-shell metals was simple
enough that we could make an exact calculation of the
effects of Coulomb correlations within that representation. It
gave us a new view of the nature of f-shell systems, quite
different from the prevalent view that electron localization
removes the f-shell electrons from the bonding role,
producing the large atomic volumes of the rare earths and
heavy actinides. We can even understand why imposing that
prevalent assumption on an LDA computational study will
give essentially the right volume, even if for the wrong
reason.
If it were possible to fully incorporate Coulomb
correlations in an otherwise LDA approach, it would
certainly lead to the correct volumes for f-shell metals, but
it is not clear it would tell us why-any more than experiment
does. It would also give the correct gap enhancement, m p,
susceptibility, and quantum-well states, but not the relations
between them. Perhaps the meaning of ‘understanding’ is a
447
point of view which contains the essential truth, but is
simple enough for our minds to encompass.
References
[1] W.A. Harrison, S. Ciraci, Phys. Rev. B 10 (1974) 1516.
[2] F. Bloch, Z. Physik 52 (1928) 555.
[3] W.A. Harrison, Elementary Electronic Structure, World
Scientific, Singapore, 1999.
[4] W.A. Harrison, Electronic Structure and the Properties of
Solids, W.H. Freeman, San Francisco, 1980, reprinted by
Dover, New York, 1989.
[5] L.J. Sham, M. Schlüter, Phys. Rev. Lett. 51 (1983) 1888.
[6] J.P. Perdew, M. Levy, Phys. Rev. Lett. 51 (1983) 1884.
[7] L.C. Lew Yan Voon, W.A. Harrison, submitted for publi-
cation.
[8] L.C. Lew Yan Voon, M. Willatzen, M. Cardona, N.E.
Christensen, Phys. Rev. B 53 (1996) 10703 and references
therein.
[9] W.A. Harrison, J. Tersoff, J. Vac. Sci. Technol. B4 (1986)
1068 discussed also in Ref. 3.
[10] W.A. Harrison, Phys. Rev. 123 (1961) 85.
[11] D.J. BenDaniel, C.B. Duke, Phys. Rev. 152 (1966) 683.
[12] G. Bastard, Wave Mechanics Applied to Semiconductor
Heterostructures, Halsted Press, New York, 1988.
[13] W.A. Harrison, A. Kozlov, Matching conditions in effective
mass theory, in: P. Jiang, H.-Z. Zheng (Eds.), Proceedings of
the 21st International Conference on the Physics of Semi-
conductors, Beijing, China, August, 1992, vol. 1, World
Scientific, Singapore, 1992, p. 341.
[14] L.C. Lew Yan Voon, W.A. Harrison, in preparation.
[15] R.D. Cowan, Phys. Rev. 163 (1967) 54 and a number of more
recent references given in Ref. 3, p. 598.