term is the kineticInenergy, the Kohn–Sham

the third term approachis thethe functional

Hartree (B.1)
energy F is and decomposed
the last one asisinthe
eqn 3.10. W
exchange and E with
the
correlation respect
decomposition
energy. toThe below, and
1index denoting
i runs are
2 over the equal
N to
wavefunctions zero
wavefunctions, (see
by as Exercise
for
usual. 3.4 f
convenience:
ibes the interaction
In this appendixof thewe electrons
go through with thethe nuclei, of
derivation thethe second
Kohn–Sham equations of i
The density, functional
n(r), is derivative):
Z
expressed in terms Z
of the (yet 2unknown) Z Kohn–Sham
Z therefore
gy, the third term3.2.
Section is The
the procedure
Hartree is energy
similarand to whattheX welast have one seen is the
in
r Appendix 1 A; n(r)n(r 0
)
dr E ⇤ E 0

Ecuaciones de Khon-Sham
wavefunctions,
we address i (r),
directly using
Fthe
[n] = eqn
case ofdrN 2.44:
n(r)V n (r)
electrons. (r) i (r) + drdr +E
n energy. The index i runs over N wavefunctions, as i
usual. = 2 0, 2 = 0. |r r0|
The Hohenberg–Kohn theorem (Section X 3.1.1)i states that1 the total energy, 2 E, in the
expressedelectronic
in terms of the (yet
ground state is a functional unknown)
n(r) =of the| electron Kohn–Sham
2
i (r)| . charge density, n(r), E = F [n]. (B.2)
i
In Se
ing eqn 2.44: theescribe This
Kohn–Sham isapproach
Here completely
el funcional dethe
the first laterm analogous
densidad
describes
functional en
F is the tointeraction
términos the de
decomposed calculation
laasof
energía
inthe of the
del
eqnelectrons
3.10. sistema
Wewithminima
report and th
the nuclei, m
The
theHohenberg–Kohn term theorem
is variables
the states
kinetic inthat
energy, the
the total
third energy,
termi is forE,
the reaches
Hartreeits minimum
energy and
thethe last o
Xof several
decomposition below, denoting the standard
wavefunctions calculus.
by However,
convenience: here minim
value
n(r)in=correspondence (r)|2 .withand
| iexchange
Z complicated the ground-state
Zcorrelation electron
energy. density.
TheZfunctions
Zindex
(B.2) i This
runs property
over provides
N wavefunctions, as
Xby the fact
⇤ is r
that
2 our 1 in terms0 n(r)n(r need 0 to satisfy the con
) the
us with
F [n]a=recipedr for
i n(r)VThe determining
density,
(r) the
n(r),
dr Kohn–Sham
(r) expressed(r) wavefunctions.
+ drdr In
of fact,
the +(yet
E minimum
unknown) Koh
xc [n].
In order
principle can be expressed
n
to
wavefunctions, deal
using e↵ectively
i
a ifunctional
(r), using 2 eqn
i
with
derivative,
2.44:2those constraints
as in eqn |r 3.11:r|0 it is convenient
heorem states that of the total energy,
Lagrange
i
multipliers E, reaches (Nocedal its minimumandXWright, 1999). (B.1) In this metho
with the ground-state
La the
densidad electron
electrónica density.
describes se thedefine This F
como: property provides with| the
n(r) = incorporates 2
Here first
newterm functional interaction
which =of0.the electrons
automatically i (r)| .
nuclei, the
thesecond
constraints:
(B.3)
i
rmining thetermKohn–Sham wavefunctions.
is the kinetic energy, the third term n
In isfact, the minimum
the Hartree energy and the last one is the
La energía
exchange and del estado
The
correlation base
Hohenberg–Kohn
energy.se ob8ene
The index minimizando
theorem
i runs states
over Nel funcional
that the
wavefunctions,total X
energy,
as usual.E,Areaches its m
ed usingThis
a functional
property is derivative,
entirely analogous as
L[ in1 ,to
eqn the
, 3.11:
minimum
2 terms ,
11 the . . . , principle
] = E[
22(yet unknown)
used 1 , in Appendix
]
2 Kohn–Sham ij
for
[h i ji
|
The density, value
n(r), isin correspondence
expressed in with of ground-state
the
deriving the Hartree–Fock equations. In order to make use of eqn B.3 we note that the electron density. This property
us with a recipe forapplies
determining the Kohn–Sham wavefunctions.ij In fact, the m
chain ruleFof=standard
wavefunctions, i (r), using
calculus eqn also
0. principle can be expressed
2.44: to the case of functional derivatives
(B.3)derivative, as in eqn 3.11: (the
X using a functional
definition nof functional derivative
where the constants is given by
n(r) = 11 , | 12 eqn, 21
i (r)|
3.24).
2
. and we22apply
If the chain (B.2)
are unknown. rule Theto replac
the U8lizando
functional la regla
derivative de la
of cadena:
F with respect ito any of the wavefunctions, F ⇤
, we find:
y analogousThe sensible since
toHohenberg–Kohn
the minimum principle the term used in Appendix
in the squareA brackets for
= 0. is required
i to vanis
theorem states that the total energy, n E, reaches its minimum
ck equations.
valueIn constants
in order to makewith
correspondence use are
ij the of called
Fground-state
eqn
=
FB.3Lagrange
nwe note
electron
=
F that multipliers,
density.
, the This propertyand can be
provides(B.4)
considere
This property is⇤ entirely i
calculus also applies
us with variables.
to for
a recipe In ofLagrange’s
thedetermining
case functional
the
i n analogous
Kohn–Sham
⇤
imethod
derivativesn tothe
wavefunctions.
the minimum principle used in Append
constrained
(the In fact, the minimum minimization p
principle can be deriving
expressed the Hartree–Fock
using a functional equations.
derivative, In as order
in eqn to make
3.11: use of eqn B.3 we note
derivative isCombinando
given by eqns las
eqnchaindos
A.63.24).ecuaciones
andofIfA.4
rule standard anteriores:
we isapply replacedthe chain
calculus by
alsothe rule
applies following
toto the case unconstrained
of functional derivatminim
⇤
e of F with respect to any of theofwavefunctions,
definition functionalFderivative i ,iswegiven find: by eqn 3.24). If we(B.3) apply the chai
Ln = 0. L
the functional derivative = of F 0, withi respect
= 1, 2,to any of the = 0, i, j = 1,i⇤2.
wavefunctions, ,
F This property
F n is entirely
F analogous to thei minimum principle used in Appendix ij A for
⇤ = ⇤ = i , F F(B.4)
n F
n the Hartree–Fock
i deriving i n equations. In order to make use of eqn
⇤ = n
B.3 we note that the
⇤ = n i,
It can immediately be verified that the
chain rule of standard calculus also applies to the icase of functional
i functional
derivativesderivatives
(the w
definition ofLagrange multipliers
functional derivative yield
is given precisely
by eqn 3.24). If the constraints
we apply the chain in
ruleeqn
to A.4.
 final result
At would be the
this point it same if we weretocorrectly
isF convenient treating them as interdependent).
⇤request that the Kohn–Sham wavefunction
case
ntees the
Bythat derivatives
combining eqnswith
B.3inrespect
and
the orthonormality
density eqnB.4 to iisorthat
we see
= 0,B.2 i can be normalized
treated (B.5)
the Kohn–Sham
correctly asorbitals
if independent
to must
N (i.e.the
satisfy
electrons. the
the ⇤ constraints:
following
final
Ecuaciones
condition:
result would
de Khon-Sham
be the same i if we were correctly treating them as interdependent).
ned
hich optimization whatproblem defined by eqns hB.5
i | and
j i =equations.
Fthe Hartree–Fock ij , is formally identical
B.6
Byis reminiscent
combiningof eqns was
B.3found
andinB.4 Appendix
we seeA for
that the Kohn–Sham orbitals must satisfy the
hat wasAppendix B A, and hence the same procedure can be
⇤ = 0, (B.5)
this point it is convenient to request that the Kohn–Sham
solved in Appendix i wavefunctions satisfy
which
following condition:
e orthonormality guarantees
constraints: that the density in eqn B.2 is correctly normalized to N e
thiswhich
aim iswe reminiscent
Para
The introduce
resolver of what
el
constrained the
sistema was found in
| jLagrange F
de ijecuaciones
h ioptimization
Appendix A for the Hartree–Fock equations.
functional:
ui8lizamos
problem mul8plicadores
defined (B.6) B.5de
by eqns Lagrange
and B.6 is formally
hich guarantees
Derivation of the Kohn–Sham
to the one
that the density
the orthonormality
i=
that
,
At this point it is convenient to request⇤ that
inX
constraints: was solved in
= 0, the Kohn–Sham wavefunctions satisfy
Appendix
eqn B.2 is correctlyi normalized to N electrons. A, and hence the same
(B.5)
procedur
which is equations
followed.
he constrained optimization
reminiscent L=F To
problem
of
thisdefined
what
aim we
was
by [h
ij
found
introduce
eqns
the one that was solved in AppendixijA, and hence the same procedure can be
i|| j
h iB.5
in
and
ji i =B.6
Appendix
the ijis,Lagrange
] ,A foridentical
ijformally the
functional:
Hartree–Fock (B.7)
(B.6)
equations.
Atwhich
lowed. this
To this guarantees
pointaim that the
it introduce
we is convenient theLagrange
density
to requestin eqnthat
functional: B.2X is correctly
the Kohn–Sham normalized to N electrons.
wavefunctions satisfy
Cumpliendo con las condiciones L= de F ortonormalidad [h iand | j iB.6 is ij ]formally
,
theThe
Lagrange constrained optimization
multipliers,
orthonormality and
constraints:X weproblem search defined
for the by eqns
extrema
ij
ij B.5
of this new identical
functional.
to the one thatLwas = F solved ijin[h Appendix i| j i ij ] , A, and hence the same procedure can be
(B.7)
followed.
In this Toappendix
this aimwewegointroduce
ij
through h
thethe | i
iderivation
j
Lagrange = ,
ijof functional:
the Kohn–Sham equations of (B.6)
with Lij the Lagrange multipliers, FX X and we search for the extrema of this new fu
Section
th ij the Lagrange Resolviendo
3.2.
multipliers, tenemos
The procedure and we is que:
similar
search for to what
the extrema we of havethisseennew in Appendix A; therefore
functional.
which guarantees
We obtain:that
= 0 the density
! in eqn
= B.2 is correctly
. normalized to N electrons. (B.8)
e obtain: we address directly ⇤ the case L of =N F electrons.
⇤ ij [h i ij | j i j ij ] , X (B.7)
The constrained ioptimization X L j F
! problem i defined by! eqns
that B.5 and B.6 is E, formally
in the identical
L F ij
The Hohenberg–Kohn theorem (Section 3.1.1)
= 0 states the total = energy, .
⇤ = 0 ⇤ = j⇤
ij j . ⇤
(B.8) ij j
nalto the
with one
electronic
ij the
derivatives that was
ground
Lagrange
F/
i statesolved
⇤ is a in
multipliers,
can Appendix
functional
i
be andofiwe
evaluated A,
thesearchelectronand byhence
forcharge the
idensity,
the extrema
inserting samej this
n(r),
of
eqnprocedure
E =new F [n].
B.1 can be
functional.
inside
In the Kohn–Sham ⇤i
approach theevaluated
functional F is decomposed as ininside eqn 3.10. We report
We obtain:
he followed.
functional derivatives
To this aim F/we can be
i introduce the by
Lagrange inserting
⇤ usando eqn B.1
functional:
n B.4: Podemos
The evaluar
functional
the decomposition below, denoting las L derivadas
derivatives the F/
wavefunctions Fi can X
by be i
el funcional
evaluated
for convenience: by inserting eqn B
⇤ = 0 X! ⇤ = ij j. (B.8)
eqn ZB.4:
⇢Z X Z
i ZZ Z Z j
r2 F [n]⇢=Z dr n(r)V (r) L =1F dr Z ⇤i (r)
r2 [h0 )i | 1 i
⇤Z0 n(r)n(r ij i (r)+ i+Ej [n] drdr ij
0 iZ]0 n(r)n(r0 )
, + Exc (B.7)
[n]. B.1 inside
n n (r) + F/ ⇢
2
i (r) +
The functional
n
dr n(r)V
derivatives
2 i2
1 Z
drdr
i can
ij be
|r2 r0 | 0 n(r)n(r
evaluated
2 xc Z )(r)
by inserting
|r r | 0 eqn 0
i (r)eqn
X+ B.4: r
dr n(r)V(r) n (r) + drdr 1 + E 0
xc n(r)n(r
[n] i)(r)
= Obteniendo:
n (r). i + 2 dr n(r)V (r)
n |r + r 0 | drdr
(B.9) (B.1) + Exc [n] i
with ij the
ij
Here the
j Lagrange 2 multipliers,
⇢ Z n and we search Z Z for the 2 extrema of this
|r new
r 0 | functional.
r2 first term describes the interaction of the 1 electrons0 with the nuclei, the second
j 0
X n(r)n(r )
We obtain: (r) + energy, the
term is the ikinetic drthird
n(r)V n (r)
term is+ the Hartree drdr energy and 0the + Exc
last one[n] is the i (r)
y using
ij thej (r).definition
2 of=functionaln ij L j (r).
derivative in eqn 3.24,2F this X equation |r can r be
| (B.9)
exchange
jwritten rapidly as: X and correlation energy. The index i runs over N wavefunctions, as usual.
j
⇤ =0 ! ⇤ = ij j . (B.8)
The= density,ij n(r), (r). is expressed in terms of the (yet
j unknown) Kohn–Sham (B.9)
 j Z j 0
i
X
i
wavefunctions,
r 2
By using (r),
the using
n(r definition
) eqnE 2.44: of functional derivative in equation
eqn 3.24, this equation
e The
definition +
functional V n
of
(r) + functional
i
dr 0
derivatives 0
+ derivative
F/
xc
⇤ i (r)
canX= in
be
eqn
evaluated
ij j (r).3.24, by
this
(B.10)
inserting eqn
can
B.1
be
inside
By using
2 the definition
rewritten |r r of
rapidly | as: functional
n i derivativej in eqn 3.24, this equation can be
pidly as:
eqnrewritten
B.4: rapidly as: n(r) =
i
| i (r)| 2
. (B.2)
 = ij j (r). (B.9)
elation itspotential
the indefinition
j
ground state, ofis functional
Eslightly
= F [n], more
using complicated
the
derivativeelectron density.
in eqn but The also
3.24, inthis
remainingthat case isonlyc
question
equation
y is therefore
Byneeded;
using thehowi.e. to theactually
definition ofdetermine
functionalthe
correlation electron in
potential
derivative density.
ateqna 3.24,
given thispoint is determined
equation can be
rapidly as:
It turns out
rewritten that as:
rapidly the ground-state density, n0 , is precisely the function that minimizes
at the same
the total point.
energy, E = F [n]. This property is called the ‘Hohenberg–Kohn variational
2
 2 Z Z 0as follows
0 X Kohn, X 1964):
rprinciple’ r can be expressed
and
+ Vn (r) +0 dr n(r )
0 n(r )
+ EE(Hohenberg
xc
xc (r) =
and
ij j (r). (B.10)
+ Vn (r)2 + dr |r 0 r+ 0| n
i
i (r) = j ij j (r).
2
.4 In order to derive the Kohn–Sham equations |r r | F [n] n = 0.
j
it is useful to introduce (3.11)
Here we Dado
can el funcional
recognize
he functional derivative (Parr and Yang, F
the de la función
Hartree g(r),
potential,n una
n0
VH función
(r), and arbitraria
the h(r)
exchange y
1989). Given the functional F ofun parámetro
and correlation
an
), anrecognize realVε,
potential,
arbitrary xc la
(r),
the derivada
function Hartreefuncional
in the third
h(r), de F con
and fourth
potential,
and a respecto
terms
real V inside a g,
(r),
parameter the δF/δg,
andsquare
✏,esthe
the la función
brackets,
exchange que sa8sface
respectively.
functional and corre
derivativ
This
in property is analogous to the variational H
principle that led us to write the Hartree–
As la siguiente propiedad:
the case of the Hartree–Fock equations in Appendix A, the matrix ij of the
ect to
VxcLagrange
(r), g,
Fock in F/
the
equations g,
multipliersis
third the
in Chapter
canfunction
andbefourth
2. which
As interms
diagonalized satisfies
that(see inside
case, eqn statingthe
A.15). following
thethat
Bysquare property:
brackets,
aintroducing
functional derivative
‘rotated’ respec
case must
of bethe zero
wavefunctions, leads
ZHartree–Fock
i , asto in
aneqn
equation for
we the
equations
A.16, wavefunctions,
finally in
obtain Appendix i (r), which
the Kohn–Sham A, the can be used
matrix
equations of to
ij
construct
Section 3.2:the density,  as F
in eqn 2.44. d In fact, if we require these wavefunctions to be
multipliers can
orthonormal (i.e. h(r)dr be1satisfy
to diagonalized
2
(r)
eqn = then
2.56) (see
F [g(r)the eqn
+ ✏ h(r)]A.15). . By
Hohenberg–Kohn introducing
variational principle ‘ro
(3
r +gVn (r) + Vd✏ H (r) + Vxc (r) i (r) = " i i (r).
✏=0
(B.11)
ions,leadsi ,to:as in eqn  2 A.16, we finally obtain the Kohn–Sham equati
1 2
2:
, consider  the following r + Vn (r) + VH (r) + Vxc (r) i (r) = "i i (r),
2 functional:
(3.12)
1 2
r + nuclear
where the external Vn (r) + VHZ(r)V+
potential, VxcHartree
n , the (r) potential,
i (r) = "ViH , iand (r).the kinetic
energy, r 22 /2, are identical I[g]to =those dr g 2 (r).
in eqn 2.48. The extra term, Vxc , is given by:

Exc [n]
Vxc (r) = , (3.13)
the definition in eqn 3.24 we find: n n(r)

and is called the exchange and correlation potential. The steps required to obtain
eqn 3.12 from eqn 3.11 are described in Appendix B. The set of equations given
by eqn 3.12 are called Kohn–Sham equations and form the basis of the Kohn–Sham
theory (Kohn and Sham, 1965). As we will see in the remainder of this book, this
old density within a desired tolerance, at which point we say that we have ‘achieved
self-consistency’. This procedure is illustrated in Figure 3.3.

Fig. 3.3 Schematic flow-chart for finding self-consistent solutions of the Kohn–Sham
equations (eqns 3.27–3.32). The equality sign in the conditional symbol means that the
electron density at this iteration and the density at the previous iteration di↵er by less
than a desired tolerance. While conceptually it makes sense to compare densities in order
Intercambio y correlación
ectron gas it is possible to calculate the exchange energy exactly, and it is pos
determine the correlation energy using numerical techniques. The following sec

Gas de electrones
e devoted to these two aspects.
3.1
libres
Exchange and correlation energies of the electron gas
El gas de electrones libres es el modelo más simple de electrones en
he freeun sólido. En
electron gaseste modelo
is the se supone
simplest modelque los electrones
of electrons in a no interactúan
solid. In this model
entre sí, que el potencial debido a los núcleos es simplemente una
sumedconstante
that electrons do notfijarse
(que puede interact
en with
cero each other, that the
por conveniencia) potential
y que los N due to
uclei iselectrones
simply a están contenidos
constant (which encanuna
be caja grande
set to zero de
forvolumen V. . Bajo
convenience) and that
estosare
electrons supuestos,
containedlos in
estados
a largepropios
box ofy volume
los valores
V . propios en unidades
Under these assumptions
de Hartree
genstates están dados
and eigenvalues in por:
Hartree units are given by (Kittel, 1976):
1 |k|2
k (r) = p exp(ik · r) and "k = . (3
V 2
hese solutions represent stationary waves with wavevectors k (see Section 9.2).
genvalue of the highest occupied state is the Fermi energy, ✏F , and the correspon
avevector is the Fermi wavevector, kF , so that ✏F = kF2 /2. The beauty of the
ectron gas model is that all its physical properties depend on one single param
e electron density, n = N/V . For example, the Fermi wavevector is related to
nsity by:
1
2
kF = (3⇡ n) 3 . (3
s shown in Exercise 3.3, the exchange energy, EX , of the electron gas can be obta
 Aproximación de la densidad local (LDA)
Se basa en un sistema
hipoté0co eléctricamente
neutro de volumen infinito
caracterizado por un
numero infinito de
electrones interactuantes
moviéndose por un espacio
con carga posi0va y
uniformemente distribuida.
La densidad electrónica es
constante, ρ. Tenemos por
tanto un gas uniforme de
electrones.
El modelo LDA asume la siguiente expresión para la energía Exc:

Donde 𝜺 es la energía de correlación y de intercambio por

par;cula para el gas uniforme de electrones de densidad ρ.
La energía ahora la dividimos en una contribución de intercambio
y otra de correlación:
 Pseudopotenciales

pseudopotencial:
all-electron: φps(r) → función de onda Vps(r) → pseudopotencial
φ(r) → función de onda
V (r) → potencial completo • core: no par8cipan de manera ac8va en el enlace,
• valencia: determinan las propiedades del sistema.
en donde V es sus8tuido por Vps, cuya φps reproduce φ ∀ r > rc, reduciendo de
esta manera la expansión en funciones base de φ.
 Teoría de pseudopotenciales
Construcción de un pseudopotencial
Construcción de un pseudopotencial
Consideremos
Consideremos que H que
es elH es el Hamiltoniano
Hamiltoniano original conoriginal condefunciones
funciones de core
core y de valencia,
φc y φyv de y Ïv respectivamente, entonces:
valencia, Ïc entonces:
respec8vamente,
ÿ
v = Ïv +
Ïps –vc Ïc ’ –vc = ÈÏc |Ïps
v Í.
c

Aplicando H se obtiene,
Aplicando H se ob8ene, ÿ
H |Ïps
v Í = Áv |Ïv Í + –vc Ác |Ïc Í ,
c
ÿ
= Áv |Ïps
v Í+ –vc (Ác ≠ Áv ) |Ïc Í ,
c
ÿ
= Áv |Ïps
v Í+ (Ác ≠ Áv ) |Ïc Í ÈÏc |Ïps
v Í,
c
C D
ÿ
= Áv |Ïps
v Í + (Ác ≠ Áv ) |Ïc Í ÈÏc | |Ïps
v Í,
c

en donde Ác y Áv son los eigenvalores de core y de valencia, respec-

tivamente. 57
enOmar
donde εc y εv son| los
De la Peña-Seaman eigenvalores de
IFUAP core
Física y de valencia,
del Estado respec8vamente
Sólido ≠ Maestría (Física)
/82
57/82
 C D
ÿ
TeoríaTeoría
Hde|Ïpseudopotenciales
ps
Í = Á |Ï v Í+
ps
vde pseudopotenciales
v (Ác ≠ Áv ) |Ïc ÍÈÏc | |Ïps
v Í,
De laConstrucción
expresión anterior,
de un pseudopotencial
Construcción de un pseudopotencial c
De la expresión anterior,
De la expresión anterior,
podemos relacionar, C C DD
C ÿ ÿ D
= Ív=|ÏÁvps|Ï Í+ (Á(Ác c≠≠ÁÁvv)) |Ï
v v+
ps ps ps ps
H |Ïps
HÍ |Ï
vÿ v Á Í |ÏccÍÈÏ
ÍÈÏc |c | |Ï|Ï
v Í, Í,
v
+ (Á ) |Ïps = Áv |Ïps
c
H ≠
podemos relacionar,
podemos relacionar, v Á c |Ïc ÍÈÏ |
c c v Í v Í,
podemos relacionar,
Cc D
C ÿ D
H+
H+
ÿ (Áv ≠ Ác ) |Ïc ÍÈÏc | |Ïps
(Ácv ≠ Ác
∆ H ps |Ïv Í =
= Áv |Ïps
v Í ps ps ps
v Í,
ps Áv |Ïv Í ,
) |Ï ÍÈÏ | |Ï Í = Á |Ï Í ,
ps
c c vps v
psv
∆ H |Ï Í = ps
en donde vemos que los pseudo≠orbitales ps satisfacen una ecuación
c ps v Ávps|Ïv Í ,
ps ps Ïv
en donde vemosenque
donde
los vemos que los pseudo≠orbitales
∆ φpsHsa8sfacen =satisfacen
ps |Ïv ÍÏvuna
ps
Áv |Ïuna ,ecuación
v Í8po
tipo Schrödinger, con: con:
pseudo−orbitales
tipo Schrödinger,
ecuación Schrödinger, con:
en donde vemos que los pseudo≠orbitales Ïv satisfacen una ecuación
ps
ps = H + VR = T + (V + VR ) = T + Vps ,
Hps =Hcon:
tipo Schrödinger, H + VR = T + (V + VR ) = T + Vps ,
donde Vps representa un pseudopotencial el cual es dependiente de
Hps = sig.+ VR = T + (V + VR ) = T + Vps ,
donde VpsÁv representa
mediante la H un pseudopotencial el cual es dependiente
relación,
ÿel cual es dependiente de εv mediante la sig.
de
donde Vpsdonde
representa un pseudopotencial
Vla representa unVRpseudopotencial elc |cual es dependiente de
v mediante ps sig. relación,
Árelación, = (Áv ≠ Ác ) |Ïc ÍÈÏ .
Áv mediante la sig. relación,ÿFísica del Estado Sólido ≠ Maestría (Física)
c 58
/82
= (Á(Á v ≠ Ác ) |Ïc ÍÈÏc | .
Omar De la Peña-Seaman | IFUAP 58/82
ÿ
VVRR = v ≠ Ác ) |Ïc ÍÈÏc | .
cc 58 5
/82
Omar De la Peña-Seaman
Omar | IFUAP
De la Peña-Seaman | IFUAP Física
Física del Estado
del Estado Sólido
Sólido ≠ Maestría Maestría
≠ (Física) (Física) 58/82 5
 Propiedades del pseudopotencial
Teoría de pseudopotenciales
Propiedades del pseudopotencial

El pseudopotencial,
El pseudopotencial,
ÿ
Vps = V + VR = V + (Áv ≠ Ác ) |Ïc ÍÈÏc | ,
c

Posee las siguientes

posee las propiedades:
siguientes propiedades:
• Vps æ V ’ r > rc , debido a que las funciones de core son cero
en esa región,
• la construcción del pseudopotencial es lineal, en el sentido de que
la contribución de cada átomo es aditiva y separada,
• debido a que la contribución al potencial de core es repulsiva, Vps
es generalmente más suave que V ,
• lo anterior da como resultado una expansión reducida en funciones
base para los pseudo≠orbitales.

59
Omar De la Peña-Seaman | IFUAP Física del Estado Sólido ≠ Maestría (Física) 59/82
/82
Teoría de pseudopotenciales
Clasificación y caracterís0cas de los
Clasificación y características de los pseudopotenciales
pseudopotenciales
Existen básicamente dos tendencias para la generación de pseudopoten-
ciales, las cuales difieren del método de construcción y características
de los mismos,
Existen básicamente dos tendencias para la generación de pseudopoten- ciales, las cuales
difieren del método de que
• Pseudopotenciales construcción
conservany caracterís8cas
la norma, de los mismos,
• Pseudopotenciales ultrasuaves.
• Pseudopotenciales que conservan la norma,
Ambos esquemas tratan de cumplir con las siguientes metas:
• Pseudopotenciales ultrasuaves.
(1) Deben ser tan suaves como sea posible, tal que la expansión de
los pseudo≠orbitales
Ambos esquemas tratan de de cumplir
valencia con
use las
un siguientes
número demetas:funciones base
lo másser
1. Deben pequeño posible.
tan suaves como sea posible, tal que la expansión de los pseudo−orbitales
de valencia
(2) Deben seruselounmás
número de funciones
transferible basetal
posible, lo más
que pequeño posible.
si un pseudo es
2. Deben ser lo
generado más
para transferible
una posible,
configuración tal que
atómica, éstesi sea
un pseudo es generado
aún válido para para una
configuración atómica, éste
otras configuraciones, seaejemplo
por aún válido para otras configuraciones, por ejemplo en
en sólidos.
sólidos. q
(3) La pseudo≠densidad de carga ( v ÈÏps v Í) debe
|Ïps reproducir
3. La pseudo−densidad de carga v debe reproducir
la densidad de carga de valencia de manera tan precisa como sea
4.la densidad de carga de valencia de manera tan precisa como sea posible.
posible.
60
Omar De la Peña-Seaman | IFUAP Física del Estado Sólido ≠ Maestría (Física) 60/82
/82
Pseudpotenciales que Conservan la
Norma
Local

No Local
 Pseudpotenciales ultrasuaves
Teoría de pseudopotenciales
Pseudopotenciales ultrasuaves
El esquema de pseudopotenciales ultrasuaves también requiere que:

Ïps = Ï & Vps = V ’ r > rc ,

Ïps ”= Ï & Vps ”= V ’ r < rc ,

en donde se ha removido la constricción de conservación de norma,

para permitir que el pseudopotencial sea tan suave como sea posible.

Ventajas Desventajas
• se reduce enormemente la ex- • para r < rc , Ïps no tienen
pansión en funciones base, la misma norma que Ï
• no hay restricción en que tan (all≠electron),
grande puede ser rc . • la densidad no se puede obtener
q ú
mediante Ï Ï,
• los pseudopotenciales se vuel-
ven menos transferibles,
65
Omar De la Peña-Seaman | IFUAP Física del Estado Sólido ≠ Maestría (Física) 65/82
/82
 Conjunto base de Ondas Planas
• Se asume que el sistema se encuentra dentro de una celda unitaria
definida por los vectores unitarios

• El volumen de la celda unitaria es:

 Condiciones
Periódicas

Donde
Enteros