ELECTRONIC AND MAGNETIC PROPERTIES OF

HALF HEUSLER ALLOY NiCrSi

A PROJECT REPORT
SUBMITTED TO
DEPARTMENT OF PHYSICS
CENTRAL CAMPUS OF TECHNOLOGY
INSTITUTE OF SCIENCE AND TECHNOLOGY
TRIBHUVAN UNIVERSITY
NEPAL

FOR THE AWARD OF

BACHELOR OF SCIENCE ([Link].) IN PHYSICS

By
BISHARTHA MANANDHAR
SYMBOL NO: 500080090
T.U. REGISTRATION NO: 5­2­8­35­2017

APRIL,2022

i
 DECLARATION
This project work entitled “ELECTRONIC AND MAGNETIC PROPERTIES OF HALF
HEUSLER ALLOY NiCrSi” is being submitted to the Department of Physics, Central Cam­
pus of Technology, Institute of Science and Technology (IoST), Tribhuvan University (T.U.),
Nepal for the partial fulfillment of the requirement to the project work in Bachelor of Science
([Link].) degree in Physics. This project work is carried out by me under the supervision of
Prem Sagar Dahal and co­supervision of Sashi Nepal in the Department of Physics,Central
Campus of Technology, Institute of Science and Technology (IoST), Tribhuvan University
(T.U.), Nepal.
This work is original and has not been submitted earlier in part or full in this or any other form
to any university or institute, here or elsewhere, for the award of any degree.

.............................
Signature
Name of student : Bishartha Manandhar
Symbol No.500080090
T.U. Registration No. 5­2­8­35­2017

iii
 LETTER OF FORWARD
Date:

On the recommendation of PREM SAGAR DAHAL and SASHI NEPAL, this project
work is submitted by Bishartha Manandhar,Symbol No. ,T.U. Registration No, entitled ELEC­
TRONIC AND MAGNETIC PROPERTIES OF HALF HEUSLER ALLOY NiMnSb” is for­
warded by the Department of Physics, Central Campus of Technology, for the approval to the
Evaluation Committee, Institute of Science and Technology (IoST), Tribhuvan University
(T.U.), Nepal
He has fulfilled all the requirements laid down by the Institute of Science and Technology
(IoST),Tribhuvan University (T.U.), Nepal for the project work.

Sunil Kumar Bhagat

Associate Professor
Head of Department of Physics
Central Campus of Technology
Tribhuvan University

iv
 BOARD OF EXAMINATION AND CERTIFICATE OF
APPROVAL

This project work (PRO­406) entitled “ELECRTONIC AND MAGNETIC PROPERTIES

OF HALF HEUSLER ALLOY NiCrSi” by BISHARTHA MANANDHAR (,)under the
supervision of PREM SAGAR DAHAL and co­supervision of SASHI NEPAL in the De­
partment of Physics, Central Campus of Technology, Institute of Science and Technology
(IoST),Tribhuvan University (T.U.), is hereby submitted for the partial fulfillment of the
Bachelor of Science([Link].) degree in [Link] report has been accepted and forwarded
to the Controller of Examination, Institute of Science and Technology, Tribhuvan University,
Nepal for the legal procedure.

Prem Sagar Dahal Sashi Nepal

Department of Physics Department of Physics
Central Campus of Technology University of Delware
Tribhuvan University

Assistant Professor Dipendra Sah Hem Shrestha

External Supervisor Internal Supervisor
Department of Physics Department of Physics
Mechi Multiple Campus Central Campus of Technology
Tribhuvan University Tribhuvan University

Assoc. [Link] Kumar Bhagat

Head of Department
Department of Physics
Central Campus of Technology
Tribhuvan University

v
 ACKNOWLEDGEMENT

Firstly, I would like to dedicate this project work to my father who through the stars have been
guiding and blessing me to work harder day by [Link],I would like to thank my project
supervisor Mr. Prem Sagar Dahal at Central Campus of technology,Tribhuvan University and
Sashi Nepal at University of Delaware,USA. Their encouragement and assistance at all times
have been of great values.
I would also like to thank Assis. Prof Sunil Kumar Bhagat (H.O.D, Department of Physics,

CCT, TU) and whole department for their cooperation and support. I also express my deepest
gratitude to my teachers of CCT. I would also like to express my gratitude to open source de­
velopers who developed packages like Quantum ESPRESSO, gnu­plot, xcrysden e.t.c, blog­
gers who shares their ideas, tutorials on solving problems on physics, programming e.t.c.,
without which this work would not be possible.

My big thanks also goes to my seniors Bishnu Prasad Belbase (Purdue University), Sarita
lwagun (T.U), Roshan Shrestha (CRNS,France) and Dhurba Raj Jaishi University of Utah)
for their valuable piece of advice and discussion on various topics. My sincere thanks to my
friends whose presence helped me in one way or other.

Finally, I must express my very profound gratitude to my parents and family for providing
me with continuous support and encouragement throughout my study and through the
process of researching and writing this thesis.

vi
 ABSTRACT
In recent years, alot of interest have been shifted towards the spintronic [Link]
Heusler alloys have gained alot of popularity in spintronic application due to their half metal­
lic properties as shown by band structure [Link], we have studied the electronic
and magnetic properties of Half heusler alloy [Link] used plane wave pseudo­potential
method within DFT framework implemented using Quantum Espresso code. Our calculation
shows that NiMnSb exhibit strong half metallicity as indicated by avaiability of large amount
of states at fermi level below a gap for spin down electrons. The calculated total magnetic
moment is found to be 4.00 µ B/cell which is an integral value and found to be obeying slater­
pauling rule M = Nv for Half Heusler alloy.
Keywords:Half­heusler,Half­metallicity,Spintronic,bandstructure

vii
viii
 ACRONYMS AND ABBREVIATIONS

CP Carl Parrinello
DFT Density Functional Theory
DOS Density of State
ESPRESSO opEn­Source Package for Research in Electronic Structure,
Simulation, and Optimization.
GGA Generalized Gradient Approximation
HF Hatree­Fork
HK Hohenberg­kohn
KS Kohn­Sham
LSDA Local Spin Density Approximation
LDA Local Density Approximation
PBE Perdew­BurkeErnzerhof
PWscf Plane Wave self­consistent field
PDOS Projected Density of State
PAW Projector Augmented Wave
SCF Self Consistent Field
XC Exchange and Correlation

ix
List of Tables
1 Magnetic moments of individual constituent atoms of NiCrSi for primitive cell. 30

x
List of Figures
1 Schematic diagram of various possible structure of Heusler compounds . . . . . . . . . .2
2 Schematic representation of iterative solution of coupled single­particle Hartee equations . . 5
3 Schematic diagram showing algorithm of Self Consistent Field Calculation. . . . . . . . . 12
4 Input structure for calculation : the primitive cell of NiCrSi where Ni and Cr are anti­ferromagnetically
aligned with Si. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5 Plot of ecutw f c vs total energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
6 plot of number of k points along x,y or z direction vs total energy . . . . . . . . . . . . 22
7 Plot of lattice parameter vs energy . . . . . . . . . . . . . . . . . . . . . . . . . . 23
8 Energy diagram of different atoms . . . . . . . . . . . . . . . . . . . . . . . . . . 24
9 The first Brillouin zone for primtive cell of NiCrSi and path along high symmetric points. . . 25
10 Conventional unit cell of NiCrSi with 16 atoms/cell . . . . . . . . . . . . . . . . . . 25
11 Band structure of NiCrSi at lattice constant a=10.3883 a.u, for spin­up (majority) electrons. . 26
12 Band structure of NiCrSi at lattice constant a=10.3883 a.u for spin­down (minority) electrons. 27
13 Calculated spin resolved total density of states (DOS) of NiCrSi (for primitive cell). . . . . 28
14 Calculated Spin­Projected partial DOS of NiCrSi . . . . . . . . . . . . . . . . . . . . 30

xi
Contents
1 INTRODUCTION 1
1.1 General introduction: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Schrodinger equation and many­body Hamiltonian: . . . . . . . . 2
1.3 Hatree approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Hartee­fork Approximation: . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Density Functional Theory: . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5.1 Hohenberg­Kohn theorem: . . . . . . . . . . . . . . . . . . . . 7
1.5.2 Kohn­Sham approach: . . . . . . . . . . . . . . . . . . . . . . . 8
1.5.3 Kohn­Sham equations: . . . . . . . . . . . . . . . . . . . . . . 10
1.5.4 Exchange correlation functional: . . . . . . . . . . . . . . . . 12
1.6 Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.7 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7.1 General Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.7.2 Specific Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2 LITERATURE REVIEW 16

3 MATERIALS AND METHODS 18

3.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.1 Implementation of DFT used in this work . . . . . . . . . . . 19
3.2.2 The Quantum ESSPRESSO code . . . . . . . . . . . . . . . . 19

4 RESULTS AND DISCUSSION 20

4.1 Pristine NiCrSi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.2 Convergence tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.2.1 Kinetic energy cutoff for wave function . . . . . . . . . . . . 20
4.2.2 Number of k­points in Monk­horst pack grid . . . . . . . . . 20
4.2.3 Lattice parameter optimization . . . . . . . . . . . . . . . . . 21
4.2.4 Electronic structure . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.2.5 Magnetic properties . . . . . . . . . . . . . . . . . . . . . . . . 28

5 Conclusion and Remarks 31

5.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.2 Novelty and National Prosperity aspect of Project work . . . . . . 31
5.3 Limitation of the work . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

xii
5.4 Recommendation for the further work . . . . . . . . . . . . . . . 32

xiii
CHAPTER 1

1 INTRODUCTION

1.1 General introduction:

In 1903, German mining engineer and chemist Friedrich Heusler studied a compound which
contains two parts copper, one part manganese and one part aluminium namely Cu2 MnAl
whose magnetic properties varies considerably with heat treatment and composition. They
were completely new at that time as they were Ferromagnetic although they were completely
made from non­ferromagnetic elements. Later on, a whole new class of compounds with the
similar lattices were grown and were named after him as Heusler compounds.

Heusler Compounds are exotic compounds which incorpoprates a huge number of magnetic
members exhibiting diverse magnetic phenomena like itinerant and localized magnetism, an­
tiferromagnetism , helimagnetism, Pauli paramagnetism or heavy­fermionic behavior. They
are generally represented by generic formula X2Y Z where X and Y are denoted transition
metal whereas Z is III­V group element. At first Heusler alloys studied were crystallizing
in the L21 structure which consists of 4 fcc sublattices and space group Fm3̄m. They are
called Full Heusler compounds having compound sequence of atoms as X­Y­X­Z. But there
also exists another compounds called Full Heusler compounds having same generic formula
X2Y Z but with the valency of X atoms less than that of Y atoms and with sequence of atoms
X­X­Y­Z. Later on it was found that it is possible to keep one of the sublattices unoccupied
giving rise to C1 b structure with generic formula XYZ called semi or half­Heusler compound.
Figure 1.1 shows various possible Heusler alloys Felser and Hirohata (2016) in which lattice
consist of 4 inter­penetrating fcc sub­lattices.

In 1983, during the study of half Heusler compound NiMnSb which crystallizes in c1b struc­
ture de Groot et al. discovered an exotic properties for first time called Half Metallicity. With
this discovery alot of attention was shifted toward the half­metallic Heusler compounds due
to its applications in the field of spintronic/magnetoelectronics. Usually compounds used to
show similar behavior either metallic or non­metallic at both the spin channel but in his new
discovery, he found that two spin channel shows completely different behavior. For majority
spin, the electronic band shows typical metallic behavior while for the minority spin channel,
electronic band exhibited semi­ conducting behavior with a gap at Fermi level.

In this section basic theoretical concept of many body hamiltonian and Density Functional

1
 Figure 1: Schematic diagram of various possible structure of Heusler compounds

Theory is outlined following ”Atomic and Electronic Structure of Solids” by [Link] Kaxi­
ras (2003), ”A Chemist’s Guide to Density Functional Theory” by [Link] and [Link]
and ”Density Functional Theory:A Practical Introduction” by David Sholl and [Link] Sholl
and Steckel (2011).Here we discuss about many body schrödinger equation, Born­Oppenheimer
approximation, Hatree­Fock approach, brief introduction to density functional theory and
exchange­correlation functional.

1.2 Schrodinger equation and many­body Hamiltonian:

Let us suppose that there is a well defined collection of atoms whose properties of material
are to be studied. Let us consider that system is composed of M nuclei and N electrons which
can be described by time­independent and non­relativistic schrödinger equation

Ĥi (⃗ ⃗1 , ..., R⃗M )ψi (⃗

x1 , ..., x⃗N , R ⃗1 , ..., RM ) = Ei ψi (⃗
x1 , ..., x⃗N , R ⃗1 , ..., R⃗M )
x1 , ..., x⃗N , R (1)

where Ĥ is many body Hamiltonian operator for system consist of N and M nuclei and is given
by,

h̄2 N 2 h̄2 M N M Z j e2 N N
e2 M M Z Z e2
i j
Ĥ = − ∑ i 2M j
∇ − ∑ ∇2j − ∑ ∑ |⃗r − R⃗ | ∑ ∑ |r⃗ j| ∑ ∑ |R⃗ j|
+ + (2)
2me i=1 j=1 i=1 j=1 i j i=1 j>1 i i=1 j>1 i

Here, in the equation, first and second terms define kinetic energy of N electrons and M
nuclei respectively and other three terms define the potential part of the Hamiltonian. Third
term represents attractive coulomb interaction between electron and nuclei. Fourth term and

2
fifth term represents electron­electron and nuclei­nuclei repulsive interaction respectively. me
is electronic mass, M j is mass of jth nucleus and ∇2 is Laplacian operator. ri j and Ri j is the
distance between ith and jth particle, i.e. ri j = |⃗ri − ⃗r j |. ψ (⃗ ⃗1 , R
x1 , x⃗2 , ......, x⃗N , R ⃗2 , ......., , R⃗N )
is wave function describing ith state of system, which depends on 3N spacial coordinates ⃗ri ,
and N spin coordinates αi of electrons, which are collectively termed ⃗xi and 3M spatial coor­
dinates of the nuclei, R
⃗ j .We know that, all the information we need to know about the system
is contained in wave function ψi . Lastly, Ei is the energy eigenvalue for ith state of system
described by ψi .
The key observation that one can make which can simply schrödinger equation is that atomic
nuclei is much heavier than the individual electrons. Each proton or neutron in nucleus is
about 1800 times heavier than electron. Thus, electrons moves much faster than nuclei. The
consequences of this fact is that we can consider, at least to a good approximation, electrons
is moving in the field of a set of fixed nuclei. This is known as Born­Oppenheimer approxi­
mation .But this allows us to solve schrödinger equation for electronic part only. Since, nuclei
are fixed, their kinetic energy is zero and the potential energy due to nuclei nuclei repulsive
interaction is just a [Link], Hamiltonian given in equation reduces to so called elec­
tronic Hamiltonian

ˆ = −h̄ N 2 N M Z j e2 N N
e2
Helec ∑ i ∑ ∑ r −⃗ R ∑ ∑ r⃗ j
∇ − +
2me i=1 i=1 j=1 i j i=1 j>1 i (3)
= T̂ + VNˆ e + Vˆe e
On solving schrödinger equation

ˆ ψelec = Eelec ψelec

Helec (4)

for electronic part only with electronic Hamiltonian Helec

ˆ , we get electronic energy (Eelec )
and electronic wave function ψelec which depends only on electronic coordinates and nuclear
coordinate do not appear explicitly in ψelec . The total energy Etot is the sum of electronic
energy Eelec and constant ­nuclei­nuclei repulsive interaction term Enuc .

M M Zi Z j e2
Etot = Eelec + ∑ ∑
i=1 j>1 Ri j (5)
= Eelec + Enuc
Now moving forward we drop the subscript ”elec”. Although we have simplified our
problem by using Born­Oppenheimer approximation, we must not be fooled by clean hands
of schrödinger equation. Solving it for ψ (⃗xi ) is still an alarming task, because of nature of
electrons. ψ (⃗xi ) depends on each coordinates of N electrons. We should keep in mind that

3
there are larger numbers of electrons than nuclei. Even for single molecule of CO2 ,ψ is 66­
dimensional function 93 dimensions for each of 22 electrons).Also, ψ must change sign if
two electrons with same sign interchange their position. This is called exchange property
and is consequence of Pauli exclusion principal. Moreover, each electron is affected by other
electrons in the system which is called correlation property. These all facts gives us a sense,
why solving Schrödinger equation is formidable task.

1.3 Hatree approximation

We can further untangle our problem by mapping the many­body problem to equivalent one­
electron problem by considering system as a collection of classical ions and essentially single
quantum mechanical particle that reproduce the behavior of electrons in the system; this is
called single­particle picture. While developing one­electron picture of material, we will not
neglect the exchange and correlation effects of electrons, but take them into account in a av­
erage way.
The simplest approach is to assume many­body wave function ψH can be written as

ψH (⃗(xi ) = ϕ1 (x⃗1 )ϕ2 (x⃗2 )ϕ3 (x⃗3 ).......ϕN (x⃗N ) (6)

with index i running over all electrons, which would be appropriate if electron in system were
non­interacting. Wave function (ϕi (⃗xi )) represents the ­single­particle states (normalized to
unity) in which individual electrons would be found if this were realistic approximation. This
is known as Hatree approximation. The total energy of system is

H = ⟨ΨH |H|ΨH ⟩
(7)
 2 N M Z j e2 e2

−h̄ 1
= ∑⟨ϕi | ∇2 − ∑ ∑ |ϕi ⟩ + ∑ ⟨ϕi |⟨ϕ j | |ϕi ⟩|ϕ j ⟩
i 2me i=1 j>1 |⃗
ri − R ⃗ j| 2 i̸= j |⃗r − ⃗r′ |

Using variational principle, we get single­particle Hartee equations

h̄2 2 N M Z j e2
 
1
− ∇ −∑ ∑ 2
+ e ∑ ⟨ϕ j | |ϕ j ⟩ ϕi (⃗r) = εi ϕi (⃗r) (8)
2me i=1 j>1 |⃗
⃗ j|
ri − R i̸= j ⃗r − ⃗r′
where, constants εi are Lagrange multipliers introduced to take into account the normaliza­
tion of the single particle ϕi . Third term on L.H.S inside bracket is called Hatree potential
i.e. potential due to all other electrons. Hatree potential only includes coulomb replulsion
between electrons. If all other ϕ j (⃗r j ), i̸= j,are known then each ϕi (⃗ri ) can be determined by
solving corresponding single particle Schrödinger equation. This can be solved iteratively
and schematic representation of iterative solution is shown below.

4
 Figure 2: Schematic representation of iterative solution of coupled single­particle Hartee equations

1.4 Hartee­fork Approximation:

We must note that there is something fundamentally wrong with the Hartee product wave
function. Since, it does not change sign on interchange of electrons. Any good approxima­
tion to the true wavefunction should include spin explicitly and should be antisymmetric to
interchange of electrons. To incorporate the effect of fermionic nature of electrons in many­
body wave function ψ (⃗ri ), in Hartee­Fock scheme N electron wave function is approximated
by antisymmetrized product of N one electron wave function Xi (⃗xi ). This product is called
slater determinant, ϕSD , given as

X1 (⃗
x1 ) X2 (⃗
x1 ) ... XN (⃗
x1 )
1 X1 (⃗ x2 ) X2 (⃗x2 ) ... XN (⃗
x2 )
ψ ̸= ϕSD = √ .. .. .. (9)
N! . . ... .
X1 (⃗xN ) X2 (⃗xN ) ... XN (⃗xN )

Here, Xi (⃗xi ) are called spin orbitals and are composed of spatial orbitals ϕi (⃗r) and one of two
spin function α (s) or β (s).
Xi (⃗x) = Φ(r)σ (s); σ = α , β (10)

Spin functions have property of ortho­normality i.e ⟨α ||α ⟩ = ⟨β ||β ⟩ = 1 and ⟨α |||β ⟩ =
⟨β ||α ⟩ = 0. The spin orbitals are chosen to be orthonormal.
Z
Xi∗ (⃗x)Xi (⃗x)d⃗x =< Xi |X j >= δi j (11)

5
Where, δi j is Kronecker delta (δi j = 1 for i = j and δi j = 0 for i ̸= j).Here spin orbital carries
usual physical interpretation that |X(⃗x)|2 represents the probability of finding electron with
spin σ within the volume element dv.

Now, we use variational principle to find the best slater determinant that yields the lowest en­
ergy. Here, we vary spin orbitals Xi under constraints that they yields orthonormal such that
Slater determinant yields minimal energy. The expectation vlaue of the Hamilton operator
with Slater determinant gives Hartee­Fock energy,EHF .

N
1 N N
EHF = ⟨ϕSD |Ĥ|ϕSD ⟩ = ∑⟨i|ĥ|i⟩ + ∑ ∑ |⟨ii| j j⟩ − ⟨i j| ji⟩| (12)
i 2 i j
where,
−1 2 M
 
Zj
Z
⟨i|ĥ|i⟩ = Xi∗ (⃗
x1 ) ∇ −∑ X (⃗
x )d⃗
x (13)
2 j |⃗
r1 − R⃗ j| i 1 1
gives the contribution due to the kinetic energy and electron­nucleus attractive interaction and
1
ZZ
⟨ii| j j⟩ = x1 )|2
|Xi (⃗ x2 )|2 d⃗
|X j (⃗ x1 d⃗
x2 (14)
r12
1
ZZ
⟨i j| ji⟩ = x1 )X j∗ (⃗
Xi (⃗ x2 )Xi∗ (⃗
X j (⃗x1 ) x2 )d⃗
x1 d⃗
x2 (15)
r12
are coulomb and exchange integral respectively which represent the interactions between two
electrons.

1.5 Density Functional Theory:

Density functional theory is one of most successful quantum mechanical approach to the study
of materials. Nowadays, it is commonly used to calculate for example band structure of solids
in physics and binding energy of molecules in chemistry. The foundation of Density Func­
tional Theory as it is known today was laid in 1964 when Hohenberg and Walter Kohn first
published their original paper in Physical Review, which verified that electron density can
be used as fundamental variable instead of many­body wave function to extract information
about material. In 1965, Kohn and Sham, in their paper developed a technique of finding
right electron density that involves solving set of single­electron equations. This theory was
breakthrough in understanding material in fundamental way and it revealed a wide route for
research of complex materials such as carbon nano­tube, protein,etc.

6
1.5.1 Hohenberg­Kohn theorem:

At the center of Density Functional Theory, there rest two theorems by Hohenberg and Kohn.
The first Hohenberg­Kohn theorem states,

Theorem 1:”The ground state energy from Schrödinger’s equation is unique functional of the
electron density ρ (⃗r).”

In other words, the density ρ (⃗r) is uniquely defined for given external potential V(⃗r) for
electrons. In order to prove this, suppose there are two non­trivially different potentials V(⃗r)
and V’(⃗r), give rise to the same density n(⃗r).

Let E and E’ be the total energies and ψ and ψ ′ be the wave functions for the systems with
hamiltonian H and H’ respectively where hamiltonian H is defined with potential V(⃗r) and H’
is defined with potential V’(⃗r).
E = ⟨ψ |H|ψ ⟩ (16)

E ′ = ⟨ψ ′ |H|ψ ′ ⟩ (17)

Then, using variational principle, we have,

E < ⟨ψ ′ |H|ψ ′ > = ⟨ψ ′ |H +V ′ −V ′ |ψ ′ ⟩

= ⟨ψ ′ |H ′ +V −V ′ |ψ ′ ⟩
(18)
= ⟨ψ ′ |H ′ |ψ ′ ⟩ + ⟨ψ ′ |(V −V ′ )|ψ ′ ⟩
= E ′ + ⟨ψ ′ |(V −V ′ )|ψ ′ ⟩

Here, inequality is due to fact that two potentials are completely different in non­trivial way.
Similarly it can be shown that

E ′ ⟨E − ⟨ψ |(V −V ′ )|ψ > (19)

Adding equation (21) and (22), we get

(E + E ′ ) < (E + E ′ ) + ⟨ψ ′ |(V −V ′ )|ψ ′ ⟩ − −⟨ψ |(V −V ′ )|ψ > (20)

Here, last two terms gives

Z Z
′ ′ ′ ′ ′ ′
⟨ψ |(V −V )|ψ ⟩ − ⟨ψ |(V −V )|ψ ⟩ = ρ (⃗r)V (⃗r) −V (⃗r)d⃗r − ρ (⃗r)[V (⃗r) −V ′ (⃗r)]d⃗r

=0
(21)

7
because of our intital assumption that densities ρ (⃗r) and ρ ′ (⃗r) are same. This leads to contra­
dictory condition
(E + E ′ ) < (E + E ′ ) (22)

which suggests us that our initial assumption must be wrong. Hence, we conclude that there
must exist one­to­one correspondence between external potential V (⃗r) and electron density
n(⃗r). Also the wave function is determined by the external potential, and therefore wave
function must also be unique functional of electron [Link] we denote the hamiltonian as
sum of kinetic energy term (T) and electron­electron interaction term (U) other than V i.e. H
= T + U then there must exist a universal functional of density

F[ρ (⃗r)] = ⟨ψ |(T +U)|ψ ⟩ (23)

since the kinetic energy (T) and electron­electron interactions (U) terms are common to all
solids. This functional only depends on electron density which is uniquely determined by
external potential V which differs from system to system. So, we conclude from above con­
sideration that total energy of system can be written as functional of density and is given
by Z
E[ρ (⃗r)] = ⟨ψ |H|ψ ⟩ = F[rho(⃗r)] + V (⃗r)ρ (⃗rd⃗r (24)

Although, first Hk theorem proves that functional of electron density exists but it says noth­
ing about which is the actual electron density. Second HK theorem defines a very important
property of the functional.

Theorem 2 :”The electron density that minimizes the energy of the overall functional is the
true density corresponding to the full solution of the schrödinger equation.

We could use variational principle to vary the electron density until the energy from the func­
tional is minimum, only if we knew the true form of the functional. But the true form of
functional is not known, thus making this idea of no practical use.

1.5.2 Kohn­Sham approach:

In 1965, after Hohenberg and Kohn published their amazing work, Kohn and Sham published
second major paper of modern density functional theory. In their paper, Kohn and Sham rec­
ommended a way to cope up with unknown functional that we discussed in last section. Kohn
and Sham introduced a concept of representing the real system by fictional reference system
of no­interacting fermions, built from single particle orbitals such that major part of kinetic
energy can be calculated with good accuracy. The remainder is encompassed by non­classical

8
contribution to electron–electron repulsion. The central idea of this approach is to calculate as
many information as possible exactly, and leaving small amount of energy to be determined
by approximating the functional.
HK theorem allows to construct many body theorem using electron density as basic variable.
In this framework, we can write ground state energy as
Z
E0 [ρ ] = lim F[ρ (⃗r] + ρ (⃗r)V (⃗rd⃗r (25)
ρ→

Where, the universal functional F[ρ (⃗r] includes contributions of kinetic energy, classical
coulomb interaction and non­classical portion due to self interaction correction,exchange and
correlation effects,
F[ρ (⃗r)] = T [ρ (⃗r)] + J[ρ (⃗r)] + Encl [ρ (⃗r] (26)

Among these, only J[ρ (⃗r)] is known, while the explicit form of other two terms are not known.
Thomas­Fermi model is realization of this very idea but failed completely. The reason be­
hind this is that this model assumes very simple functional form of kinetic energy with its
dependence on ρ 5/3 dd⃗r. Thus, it seems very important to find different way to treat kinetic
R

energy with better control of accuracy and which was done by Kohn and Sham.
To understand the Kohn­Sham approach to cope with this problem, we first need to review
Hatree­method discussed in section 3.1.3. In Hf metod, the slater determinant ϕSD , con­
structed from N spin orbital, is used as approximation to true N­electron wave function. But it
can be shown that slater ϕSD is exact wave function of N non­interacting fermions moving in
the field of effective hartee potential VHF . This is valid if we are dealing with non­degenerate
states. For this type of wave function, kinetic energy can be expressed as
1 N
⟨X|∇2 |Xi ⟩ (27)
2∑
THF = −
i
Here, the spin orbitals Xi are chose such that the energy EHF achieve a minimum value under
the constraints that Xi remains orthogonal.

EHF = min ⟨ϕS D|T̂ + V̂ + Vˆee |ϕSD ⟩ (28)

ϕSD →N

Now, we know that slater determinant can be used as exact wave function for non­interacting
fermions, introducing local potential Vs (⃗r) it is possible to define a non­interacting reference
system with Hamiltonian given as
1 N 2 n
(29)
2∑
Hs = − ∇ + ∑ Vs (⃗ri )
i i

9
This Hamiltonian does not contains any electron­electron interaction terms, so it describes
non­interacting system. The corresponding ground state wave function is slater determinant
given by
ϕ1 (⃗x1 ) ϕ2 (⃗x1 ) ... ϕN (⃗
x1 )
1 ϕ1 (⃗ x2 ) ϕ2 (⃗ x2 ) ... ϕN (⃗
x2 )
θs = √ .. .. .. (30)
N! . . ... .
ϕ1 (⃗xN ) ϕ2 (⃗ xN ) ... ϕN (⃗
xN )
The spin orbitals can be obtained by solving

fˆKS ϕi = εi ϕi (31)

where, fˆKS are one­electron Kohn­Sham operators, defined as

1
fˆKS = − ∇2 −Vs (⃗r) (32)
2
Here, spin orbitals ϕi are called Kohn­Sham orbitals (or KS orbitals) to differentiate from its
HF counterpart
Now, VS can be chosen such that the density of non­interacting fermions in fictitious system
is exactly equal to ground state density of real system of interacting electrons.
N
ρS (⃗r) = ∑ ∑ |ϕi (⃗r, s)|2 = ρo (⃗r) (33)
i s

1.5.3 Kohn­Sham equations:

The sparkle of Kohn and Sham is in realizing that if it is not possible to calculate kinetic
energy accurately through explicit functional we should calculate as much as we can of the
true kinetic energy exactly and deal remainder in approximately. To obtain the exact kinetic
energy of the non­interacting reference system with same density as that of real interacting
system, kohn and sham proposed to use equation 30 as
1 N
⟨ϕi |∇2 |ϕi ⟩ (34)
2∑
TS = −
i

Undeniably, the kinetic energy of non­interacting reference system is not equal to true kinetic
energy of real interacting system. To account this, Kohn and Sham introduced the following
separation of functional F[ρ ]

f [ρ (⃗r)] = Ts [ρ (⃗r)] + J[ρ (⃗r)] + EXC [ρ (⃗r)] (35)

10
where EXC is called exchange­correlation energy defined as

EXC [ρ ] ≡ (T [ρ ] − Ts [ρ ] + (Eee [ρ ] − J[ρ ]) = Tc [ρ ] + Encl [ρ ] (36)

The remaining part of true kinetic energy, TC , which is not covered by TS , and is added to
the non­classical electrostatic contributions. Thus, EXC contains both non­classical effects
of exchange and correlation, self interaction correction, which contributes to the potential
energy of the system and also the portion of kinetic energy which cannot be calculated ex­
[Link] for energy of non­interacting reference system is composed of two part: ki­
netic energy and energy due to interaction with external potential. From HK theorem, we
learned that the total energy must be the functional of density. Similarly, the interaction with
external potential is an explicit functional of density, ρ . Hence, T S must also be the func­
tional of density. But it is necessary to not that density does not appear in expression for TS
explicitly: KS orbital appear in equation 37 not the density.
To define VS such that it gives the slater determinant which is characterized by the same den­
sity as that of our real system, we can write the expression for energy of real interacting system
as
Z
E[ρ (⃗r)] = TS [ρ ] + J[ρ ] + EXC [ρ ] + VNe ρ (⃗r)d⃗r
(37)
1 ρ (⃗r)ρ (⃗r)
Z Z Z
= TS [ρ ] + d⃗r1 d⃗r2 + EXC [ρ ] + VNe ρ (⃗r)d⃗r
2 r1 2

Now we apply variational principle to minimize energy expression under usual constraint of
⟨ϕi ||ϕ j ⟩ = δi j . The resulting equations are
 Z M   
1 2 ρ (r2 ) ZA −1 2
− ∇ + d⃗r2 +VX C(⃗r1 ) − ∑ ϕi = ∇ +Ve f f (⃗r1 ) ϕi
2 r1 2 A r1 A 2 (38)
= εi ϕi

Comparing this equation with equation 35, we would see that the term Ve f f is identical to Vs .
Therefore, we have

M
ρ (⃗r2 ) ZA
Z
Vs (⃗r) ≡ Ve f f (⃗r) = d⃗r2 +VXC (⃗r1 ) − ∑ (39)
r12 A r1A

where VXC is potential due to the exchange–correlation EXC and is simply defined as functional
derivative pf EXC with respect to ρ , i.e.,
XC
VXC ≡ (40)
ρ

11
 Figure 3: Schematic diagram showing algorithm of Self Consistent Field Calculation.

As we already know about the contribution in equation 42, we can have VS which can used
in one­particle equations to determine KS orbitals and thus we can determine ground state
energy using equation 36 and the ground state using equation 40. But we should note that VS
already depends on density and thus on the KS [Link] one­particle KS equations must be
solved iteratively . Algorithm for solving KS equation can be illustrated as shown below.

1.5.4 Exchange correlation functional:

Kohn­Sham formalism allows us to calculate exactly the most contribution to electronic en­
ergy of atomic or molecular system, including the major portion of kinetic [Link] re­
maining unknown portion are collectively termed as exchange­correlation functional EXC .
Although, its explicit form is yet unknown, in order to apply the kohn­sham formalism the
explicit approximation to this approximation is to be found. The quality of density functional
approach revolves wholly on accuracy of the chosen approximation,EXC . Due to this, scien­
tist has long been trying to find better approximation for decades. Many approximation to
explicit form of exchange­correlation functional has been [Link] them, Local den­

12
sity approximation (LDA) and generalized gradient approximation (GGA) are mostly used.

Local density approximation [LDA]:

In this section, let us consider a model system of homogeneous electron gas in which electrons
are distributed evenly with uniform positive external potential, and overall charge neutrality
is [Link] all approximate exchange­correlation functional is based on this model
[Link] model assumes that the number of electrons N and volume of electron gas V ap­
proaches infinity, while the electron density, N
V = ρ , remains constant [Link],
this condition resembles the model of ideal metal with perfect crystal. Our model system
of homogeneous electron gas, although pretty far from realistic atomic or molecular system
which is characterized by varying electron density, has a impportant position in density func­
tional theory. This is because, explicit form of exchange­correlation functional is known
exactly or with very high accuracy only in this system. At the center of this model is the
approximation that we can write EXC as
Z
LDA
EXC [ρ ] = ρ (⃗r)εXC (ρ (⃗r))d⃗r (41)

Here,εXC (ρ (⃗r)) is exchange–correlation energy per particle of homogeneous electron gas of

density ρ (⃗r) and is weighted with the probability ρ (⃗r) that there is infact an electron in this
[Link] approximation which allows us to write EXC in this form is known as local density
[Link], εXC (ρ (⃗r)) can be written as

εXC (ρ (⃗r)) = εX (ρ (⃗r)) + εC (ρ (⃗r)) (42)

The exchange part, εX (ρ (⃗r)), represents the exchange energy of electrons in homogeneous
electron gas with certain density and has form given by
r
3 3 3ρ (⃗r)
εx (ρ (⃗r)) = − (43)
4 π
4
using equation 46 in equation 44 gives the ρ 3 dependence of the exchange [Link] ex­
change functional use here is generally known as Slater [Link] no such explicit form
of the correlation part εc is known. Build on highly accurate numerical quantum Monte­Carlo
simulation of the homogeneous electron gas, many author have presented analytical expres­
sion for ε . The most widely used representation for εc was developed by Vosko, Wilk and
Nusair in 1980 while the most recent one was developed by Perdew and Wang in 1992. Ap­
proximate functional can also be represented in a way known as unrestricted version in which
electron density ρ (⃗r) is regarded as the sum of ρϕ (⃗r) and ρϕ (⃗r) and is used as [Link]

13
it to unrestricted case, we will get what is known as Local spin­density [Link],
exchange­correlation functional is generally written as
Z
LSDA
EXC [ρ ] = ρ (⃗r)εXC (ρα (⃗r), ρβ (⃗r))d⃗r. (44)

Generalized gradient approximation [GGA]:

Though useful in some cases, Local density approximation is not very accurate for realistic
atomic and molecular system and due to which it is not quite useful in many application in
material science. Idea behind generalized gradient approximation is not only to use the infor­
mation about density ρ (⃗r) at particular point , but supplement the density with information
about gradient of charge density, ∇ρ (⃗r) to account the non­homogeneity of electron density
in real system. Now, we consider the expression with ρ and ρ ′ referring to α or β spin,

∇ρσ ∇ρσ′ ′
Z Z
GEA ρ ,ρ ′
Exc [ρα , ρβ ] = ρεxc (ρα , ρβ )d⃗r + ∑ Cxc (ρα , ρβ ) 2 2 + ... (45)
ρ ,ρ ′ ρσ 3 ρσ ′ 3
This form of functional is called the gradient expansion approximation (GEA) and it applies
to model in which the density is varying very [Link], if GEA is used to solve
real molecular problem, it does not give desired improved accuracy and frequently gives result
which is even worst than given by local density approximation. The reason for failure is that
the exchange­correlation hole associated with a functional 48 has lost many properties which
make the LDA hole physically significant. The problem is solved by enforcing the constraints
that are valid for true holes also for the ones beyond­LDA functionals. Thus those functionals
which includes gradient of charge density and where hole constraints have restored are col­
lectively called generalized gradient approximations(GGA).In general, these functionals
can be written as. Z
GGA
EXC [ρα , ρβ ] = f (ρα , ρβ , ∇ρα , ∇ρβ )d⃗r (46)

1.6 Rationale

In this present work, we studied the electronic and magnetic properties of Half heusler alloy
NiCrSi which are said to be half metallic in nature. This property is very interesting due to its
wide range of use in spintronic deices. Moreover,both structural and magnetic properties of
Half heusler alloys can be controlled by substitution of constituent elements of alloys. Such
controllability helps to minimise a stray field for the applications of an HDD read head, an
MRAM cell and others. These magnetic properties are important for spin injection, accumu­
lation, operation and detection in spintronic devices.

14
My research focus on the electronic and magnetic properties of NiCrSi. Here I try to re­
establish that NiCrSi is half metallic in nature using Density Functional Theory with the ap­
plication of Quantum Espresso. Quantum espresso is an open source computer codes based
on density functional theory, plane waves and pseudopotentials.

1.7 Objectives

1.7.1 General Objectives

1) To study the magnetic and electronic properties of the alloys.

2) To do the ground state calculations of the alloy.
3) To verify the slater­pauling behaviour.

1.7.2 Specific Objectives

1) To study electronic bandstructure through bandstructure and density of states.

2) To verify the band gap at Fermi level(E f )for only one spin sub­band i.e. half­metallicity.
3) To check if the alloy have direct or indirect band gap.

15
CHAPTER 2

2 LITERATURE REVIEW

With the advent of the discovery of new kinds of materials, scientists are quite excited to dis­
cover more such materials whose exotic properties are are beneficial to human beings to bring
advancement in technologies. One such material is heusler alloys which are categorized into
half and full heusler alloys depending upon their crystalline structure.

While studying half heusler NiMnSb which crystallizes in c1b structure, in 1983 de groot
et al. predicted the half metallicity in Heusler alloys which shows metallic nature in ma­
jority spin channel and semiconducting nature in minority spin channel with a gap at Fermi
level De Groot et al. (1983). This gap leads to 100 percent spin polarization at the Fermi
level and therefore allowing full spin­polarized current in these compounds maximizing the
efficiency of magnetoelectronic devices.

In 2008, V.A. Dinh et al. predicted that NiCrSi and NiMnSi are half metals with integer val­
ues of the total spin moment at their equilibrium lattice constant Dinh et al. (2009).

In 2008, Galanakis et al. studied the effect of defects and interface in NiMnSi and [Link]
showed that creation of Ni defects induce impurity states within the gap which can couple to
interface states and completely destroy the spin polarization of the current injected into the
[Link] showed that NiVSb, NiCrSb and NiMnSb are HMF while NiTiSb shows
possibility of HMf.

In 2011, M.P. Ghimire et al. calculated the electronic and magnetic properties of NiXSb
(X=Ti,V,Cr,Mn) in zinc blende structure. It showed that NiVSb, NiCrSb and NiMnSb are
HMF while NiTiSb shows the possibilty of HMF. The DOS of different systems under study
were characterized by exchange splitting of 3d states of X atoms which can leads to larger
localized spin moments at the Ti,V,Cr or Mn sites Ghimire et al. (2011).

In 2003, [Link] et al. studied the electronic structure, chemical bonding and magnetism
of series of half heusler compounds XMZ where X=Fe,Co,Mi,M=Ti,V,Nb,Cr,Mo,Mz and
Z=Sn,Sb. Their detail analysis of various indicators of chemical bonding suggested the co­
valent hybridization of the higher valent transition element X with the lower valent transition
element M is the key interaction responsible for the formation of d­d gap in these system sep­
arating the bonding states from the antibonding states Nanda and Dasgupta (2003).

In 2002, Galanakis et al. showed that Full heusler alloys compounds containing Co,Fe,Rh

16
and Ru follows slater pauling behaviour and the total spin­magnetic moment per unit cell (Mt )
scales with the total number of valence electrons (Zt ) following the rule Mt = Zt ­24 Galanakis
et al. (2002).

17
CHAPTER 3

3 MATERIALS AND METHODS

3.1 Materials
Based on crystal structure and experimentally determined lattice constant of the fabricated
NiCrSi, we first performed convergence tests on kinetic energy­cutoff of wave function and
number of k­points in Monk­Horst pack grid and then optimized the lattice parameter using
energy minimization method, the result of which is discussed in next chapter. The input struc­
ture for calculation is shown in figure3.1.
Using optimized lattice parameter we relaxed the coordinates until the force between atoms

Figure 4: Input structure for calculation : the primitive cell of NiCrSi where Ni and Cr are anti­
ferromagnetically aligned with Si.

is less than 10−3 Ry/a.u. Considering all these calculated quantities, we performed electronic
structure calculation using plane­wave pseudo­potential method within DFT implemented
through Quantum­ESPRESSO distribution. We used ultra­softpseudo­potential for Nickel,
Chromium and Silicon. Exchange­correlation functional is approximated using modified gen­
eralized gradient approximation (GGA) intended for solids and surface system. Kinetic en­
ergy cutoff for wave function is taken to be 110 Rydberg. Energy tolerance was set to 10−4
Rydberg and scf threshold was set to 10−8 . Brillouin zone sampling was done using Monk­
Horst pack grid of size 8x8x8. Density of sates was calculated by linear tetrahedral integration

18
method. Crystal structure is visualized using tool XCrySden ([Link]). Plotting
is performed using gnuplot program.

3.2 Methods

3.2.1 Implementation of DFT used in this work

This theoretical work has been carried out using computer codes based on density functional
theory (DFT). Several computer codes implementing DFT are available and each one has it’s
own advantages and disadvantages. As a user, one has to chose which implementation is best
suited to problem under investigation. This choice depends to a large part on the available
feature of the code and also on the choice of basis set. In this work, we implemented pseudo­
potential plane wave based method using Quantum ESPRESSO package v6.8.

3.2.2 The Quantum ESSPRESSO code

QUANTUM ESPRESSO is an integrated suite of open­source computer codes for electronic

structure calculations and materials modeling, based on density­functional theory, plane waves,
and pseudo­potentials (norm­conserving, ultrasoft, and projector­augmented wave). The acronym
ESPRESSO stands for opEn Source Package for Research in Electronic Structure, Simulation,
and Optimization. The main components of Quantum ESPRESSO package are

PWscf: PWscf [Plane­Wave self­consistent field] is one of core component of Quantum

ESPRESSO distribution. Pwscf performs different electronic­structure calculation within
Density Functional Theory[DFT], using plane­wave basis set and pseudo­potential to repre­
sent electron­ion interactions. In particular, it can calculate ground state energy, one­electron
KS orbitals, atomic forces and stress, molecular dynamics, structural optimization e.t.c.

CP: CP[Car­Parrinello] is another core component of Quantum ESPRESSO distribution.

CP can perform Car­Parrinello molecular dynamics, also including variable cell dynamics.

PostProc:The PostProc module contains a number of codes (bands.x, dos.x,Projwfc.x, plot­

band.x etc.) for post­processing and analysis of data files produced by PWscf and CP.

19
CHAPTER 4

4 RESULTS AND DISCUSSION

In this chapter, we will discuss the result of present [Link] chapter contains the discus­
sion of convergence tests, lattice parameter optimization as well as electronic and magnetic
properties of NiCrSi.

4.1 Pristine NiCrSi

As already discussed in the previous section, we performed the scf calculation followed by
band calculation and dos calculation using Quantum [Link] this section, we will dis­
cuss the results (convergence tests, lattice parameter optimization, electronic structure and
magnetic properties) obtained from the calculation performed on pristine NiCrSi.

4.2 Convergence tests

In an ideal condition, larger the kinetic energy cutoff for the wave function and denser the
k­point grid,we can get accurate results of the scf calculation but at the expense of the greater
computational power. So, aim of convergence test is to determine the suitable value of cutoff
energy [ecut_wfc] and density of k­point grids which can give us fairly accurate results with
less computational power.

4.2.1 Kinetic energy cutoff for wave function

Taking size of k­mesh grid as 8x8x8, for different kinetic energy cutoff for wave function
[ecut_wfc], we performed scf calculation and noted the corresponding total energy per unit
[Link] of total energy vs ecut_wfc is shown below. From the plot, it can be seen clearly
that change of energy is not very significant for ecut_wfc above 100 Rydberg. Hence, our
choice of cutoff energy of 110 Rydberg is reasonable.

4.2.2 Number of k­points in Monk­horst pack grid

Taking ecut_wfc of 80 Rydberg, for different number of k points along x, y or z axis we

performed scf calculation and noted down the corresponding total energy per unit cell. Figure
below shows the plot of number of k points vs total energy. From above plots in figure, it is
clear that change in energy is insignificant for number of k points greater than 7. Hence, our

20
 Figure 5: Plot of ecutw f c vs total energy

choice of k point grid of size 8x8x8 is reasonable. Using the optimized energy cutoff value
and k point grid, further calculations were done.

4.2.3 Lattice parameter optimization

We optimized the lattice parameter of primitive cell of NiCrSi using energy minimization
method. We chose the lattice parameter which gives minimum energy per unit cell. For this ,
we performed scf calculation using different lattice constant and plotted the graph for lattice
parameter vs total energy. We used the package provided by Quantum Espresso called ev.x
and obtained the lattice constant for which the energy is minimum. The graph is shown below.
Using the package, we calculated the lattice parameter corresponding to minimum energy to
be 5.49723Å [10.3883 a.u.] which is nearly equal to the experimental value.

21
 Figure 6: plot of number of k points along x,y or z direction vs total energy

4.2.4 Electronic structure

Band structure and density of states are indispensable tools to get the insight of electronic
structure of materials. When atoms are isolated electrons occupy specific and discrete or­
bitals such as 1s, 2p, 3d, e.t.c When isolated atoms are brought together to form a solid, the
interactions between them perturbs the initial atomic energy levels Kittel and McEuen (2018).
The core electrons are unaffected by the interactions, but valence electrons being closer to
neighboring ions are mostly affected. The valence electrons meet each other in solid trying
to occupy same energy level, however , the core electrons remains stuck to nuclei in deep
potential. For example, in N atoms system, there are N 3d orbitals trying to occupy 3d en­
ergy level. Eventually, they settle down by sharing 3d energy level together and hence energy
bands is formed as shown in [Link] energy band consists of N energy levels from each
atom separated by indistinguishable energy difference and thus forms an energy continuum.

22
 Figure 7: Plot of lattice parameter vs energy

Due to periodic potential of ion core and response of electrons in semiconductor and insulator
to the potential, another interesting thing is happening which is the formation of band [Link]
understand this let us treat electrons in solid as nearly free electron. Thus, the situation not
very different from the case for free electron except when some electrons meet the Brillouin
zone boundaries with the wave vector k = nπ
a [a being the periodicity of lattice]. When this
happens the Bragg’s condition is satisfied and they are diffracted. This electron diffraction is
the underlying principle of XRD (X­ray diffraction) and cause of band gap in the semicon­
ductor and insulator.
Classification of materials as conductor, semiconductor and insulator is done on the basis of
band width and its relative position with respect to the fermi level. We know that the fermi
level corresponds to energy of highest occupied state. In conductor, the valence bands be­
comes wider due to large free electrons and thus overlap each other and placing the fermi
level in conduction band. In insulator and semiconductor, the fermi level lies at the top of
valence band or within the band gap which is relatively wider for insulator and narrower
for the semiconductor. Electronic structure can also be characterized by density of states

23
 Figure 8: Energy diagram of different atoms

(DOS). Density of states is also defined as number of electronic states per unit energy range.
Now, lets discuss about the electronic structure of the half Heusler alloy NiCrSi. We used
lattice parameter a = 10.3883 a.u. for primitive cell of NiCrSi and performed spin polar­
ized calculation using Quantum Espresso code. We calculated the band structure along path
Γ − X −W − K − Γ − L −U −W − L − K of Brillouin zone as shown in figure 8.
The obtained band structure of NiCrSi for spin–up and spin­down electrons is shown in fig­
ure 9 and 10 respectively. The band structure for spin up channel has metallic interaction at
Fermi level, thus implying the strong metallic nature of spin­up electrons. The band structure
for spin down channel exhibit indirect band gap near the fermi level. Minimum conduction
band is at point X and maximum of valence band is at symmetric point Γ. We measured the
band gap to be . The Valence band maximum (VBM) lies slightly above Fermi level, thus
indicating a very small amount of density of states at Fermi level. This indicates that we may
have 100% spin polarization at Fermi level, as required for half­metal. Hence, the compound
NiCrSi might exhibit half­metallic properties Dinh et al. (2009). The nature of obtained band
structure agrees well with that of literature.
From figure 11, DOS for majority spin­up electrons shows metallic nature and that for spin
down shows a band gap at Fermi level. Figure 12 shows the spin­resolved partial density
of states (pdos) for constituent elements. From plot shown in figure 11, we can see that the
nature of DOS is almost symmetrical for both spin. The detailed investigation of this alloy
reveals that almost all Ni­4d states are concentrated at lower energies in VB and occupied
but the Cr­3d are distributed both in the region of lower and higher energies and are strongly

24
Figure 9: The first Brillouin zone for primtive cell of NiCrSi and path along high symmetric points.

Figure 10: Conventional unit cell of NiCrSi with 16 atoms/cell

25
 4

0
Energy [eV]

-2

-4

-6

-8

-10
Γ X W K Γ L W U X

Figure 11: Band structure of NiCrSi at lattice constant a=10.3883 a.u, for spin­up (majority) electrons.

[Link] conduction band is originated from Cr 3d minority spin states. The 3d major­
ity/minority spin states of Cr are dominant in VB/CB. The Si­p states have a low intensity of
DOS. A few Cr 3d minority spins occupy the states ­1.0 eV in VB.
The large exchange spitting of Cr­3d states leads to the localized spin magnetic moment at Cr
site. Our calculation predicts larger spin moment at Cr site than at Ni site i.e. the spin mo­
ment is mainly carried out by the low valent transition metal atom Cr. Ni carries a small spin
moment of ∼0.05­0.15 due to the hybridization with the Cr­3d states in each alloys. The sp
atoms shows a negligibly small spin magnetic moment which is anti­parallel to the moment
of Ni and Cr Dinh et al. (2010). In both spin channels of NiCrSi, a low lying band ranging
µ ­11.0 to ­7.7 eV arises due to s electrons of sp element(Si) which remains isolated from rest
of the bands which is not shown in the figure.
The hybridization of Cr 3d and 4d(5d) states can open a minority spin gap at E f resulting in
the semi­conducting nature of the minority spin band. In the minority spin channel, the next
triple degenerated band t2g at Γ point is due to the sp­p states. It is followed by the doubly­
degenerated(eg ) and triple degenerated (t2g ) band formed by Ni and Cr bonding admixture.
The anti­bonding hybrids are separated from bonding hybrids by an indirect gap along Γ − X.
In this way we have exactly nine occupied minority bands. In majority state of this system,

26
 4

0
Energy [eV]

-2

-4

-6

-8

-10
Γ X W K Γ L W U X

Figure 12: Band structure of NiCrSi at lattice constant a=10.3883 a.u for spin­down (minority) elec­
trons.

27
 Spin Up
10 Spin Down

Density of states
5

-5

-10
-5 -4 -3 -2 -1 0 1 2 3
E-Ef [eV]

Figure 13: Calculated spin resolved total density of states (DOS) of NiCrSi (for primitive cell).

there is a series of bands from ∼ −6.0 to 0.0 eV where Ni d state contribute significantly
followed by a small contribution from Cr­d and Si­p [Link] contrast to minority states, we
have 11 electrons in majority bands. The two remaining electrons are found in anti­bonding
states Singh et al. (2012).

4.2.5 Magnetic properties

The magnetic property of materials can be traced back to magnetic moment of constituent
atoms or ions, which has three main sources, viz. spin of electrons, their orbital angular mo­
mentum about a nucleus and the chance in the orbital moment induced by applied magnetic
field. Magnetic moment depends on how the spin and angular momentum of electrons on
the electronic shell combine to give rise to total spin of atom or ions. This behavior is gov­
erned by quantum mechanical rules, according to which spin and angular momentum vectors
combine, supplemented by Hund’s rule that specify which states are preferred energetically.
Also, Pauli exclusion principle is satisfied. State of atom can be specified by three quantum
numbers, viz. total spin S, the orbital angular momentum L and total angular momentum J.
Electronic shells are filled according to Hund’s rule obeying Pauli exclusion principle. Atoms

28
with completely filled electronic shell has zero magnetic moment and that with partially­filled
shell has finite moment. Schematic representation of filling of electronic shell of Ni, Cr and
Si is shown in figure below.
It is evident that isolated atoms can have finite moment due to electronic spin [Link] when
atoms are brought together to form solid, valence orbital of atoms hybridize with other close
atoms forming bonding and anti­bonding states which can be filled with electrons in energet­
ically favorable configuration producing different magnetic behavior in solids. For example;
valence orbital of two atoms with finite magnetic moment can hybridize to form bonding and
anti­bonding states which can be filled with pair of electrons with opposite spins, thus leading
to no magnetic behavior in material due to electronic spin alone.
But some materials have spontaneous magnetic moment, meaning there is finite magnetic

moment even with no magnetic field. The existence of such spontaneous moment suggests
us there is certain regular arrangement of electronic spins and magnetic moments. Differ­
ent arrangement leads to different types of magnetic substance such as ferromagnet, anti­
ferromagnet, ferri­magnets e.t.c. Inverse Heusler alloys exhibit wide range of magnetic phe­
nomenon, as already discussed in chapter 1. The total magnetic moment is given by M=N,
where N↑ ­ N↓ is the number of unpaired electron and µ B is Bohr’s magneton.N can be found
by counting the electrons in the spin­up and spin­down sub­bands Poon (2001).

29
 Figure 14: Calculated Spin­Projected partial DOS of NiCrSi

Z Ef
N= |D↑ (E) − D↓ (E)|dE (47)
−∞
where D↑ and D↓ are density of states of spin­up and spin­down respectively.
Our calculation on primitive cell of NiCrSi shows that ground state magnetic structure is

ferrimagnetic. Magnetic moments of Ni and Cr at A and B are anti­ferromagnetically aligned

with that of Si at D site. Here C site is vacant. Calculated total magentic moment of NiCrSi is
2.00 µ B/cell, which agrees with values from the theoretical (2.0 µ B/cell) and experimental
(1.95µ B/cell) calculation in [Link] moment of individual atoms are tabulated in
table 1.

Table 1: Magnetic moments of individual constituent atoms of NiCrSi for primitive cell.
Constituent atoms of NiCrSi Magnetic moment(µ B/atom)
Ni(A) 0.2351
Cr(B) 1.7325
Si(D) ­0.0712

30
CHAPTER 5

5 Conclusion and Remarks

5.1 Conclusion

In this work, using first principles calculation, we have studied the electronic and magnetic
properties of pristine Half Heusler alloy NiCrSi . Spin polarized first principle calculation has
been performed using Quantum ESPRESSO package withing DFT framework. Exchange­
correlation functional was approximated using generalized gradient approximation (GGA)
intended for solids. We used plane wave pseudo­potential method for the calculation.
NiCrSi crystallizes in Half Heusler structure with space group f4̄3m. Optimized lattice param­
eter using energy minimization method is 10.3883 a.u. Our ground state calculation shows
that NiCrSi is a perfect half­metal. Electronic structure of NiCrSi reveals metallic nature for
spin­up electrons while insulating nature for spin­ down electrons. There is a gap at fermi­
level for spin­down electrons. The gap width for spin­down states was found to be 0.112
eV. The origin of gap can be considered mainly due to covalent band gap and d­d band gap.
Ground state magnetic structure of NiCrSi is found to be ferri­magnetic with magnetic mo­
ment of Ni at A site and Cr at B site aligned anti­ferromagnetically with Si at C site. The
magnetic moment of Si was found to be very small compared to Ni and Cr. Calculated total
magnetic moment of NiCrSi is 2 µ B /cell. It is found that magnetic moment of constituent
atom has site dependence i.e magnetic behavior of atom depended on neighboring atoms.

5.2 Novelty and National Prosperity aspect of Project work

This project presents the electronic and magnetic properties of Half Heusler alloy [Link]
project revises the past work done on the relative topic And presents the results accordingly.
Comparison of analysis with past work shows similarity in the electronic and magnetic prop­
erties. Apart from past work, this paper presents Magnetic moments of constituent atoms.
Half heusler alloys are the possible candidate for the manufacture of Spintronic as well as
magnetoelectronics devices among HMF with some of them exhibiting the half­metallicity.
Heusler alloys incorporates a diverse magnetic phenomena like itinerant and localized mnag­
netism, antiferromagnetism, helimagnetism, pauli paramagnetism or heavy­fermionic behav­
ior and due to low curie temperature, they can be used in production of thermoelectric devices
also.
Furthermore, these properties can be utilized can achieved 100 percent spin polarization in­

31
duced by spontaneous [Link] possess a bandgap (δ ) at the Fermi level (EF ) for
only one spin sub­band, thus achieving 100 percent spin polarisation at EF . Heusler alloys
have been used in developments of HDD/Sensors, MRam/Spin­RAM, Racetrack memories
and even experiments are being conducted to exploit the use of spin in the production of quan­
tum computer.
Nepal being a poor country can gain alot from the possible advancement in spintronic as well
as in magnetoelectronic sector. With the high rate of data transfer which can further enhanced
our technological advancement in the field of data storage and processing, country like Nepal
can gain technological advancement at low cost. The processing speed that we enjoy now can
be doubled or even tripled at the same rate of cost. Similarly, low curie temperature can en­
hanced us to develop thermoelectric devices which can help us to cut out the loss of energy in
the form of heat as they can convert heat into electricty. Thermoelectric devices will be used
extensively for personal electronic applications. These applications extends from wearable
items, such as watches or cloths, to biomedical applications such as drug delivery and vital
signs monitoring. Current technology can provide devices for micro­thermoelectric applica­
tions, such as cooling of integrated circuit. If such properties can be haervested then country
like Nepal can gain alot.

5.3 Limitation of the work

The major limitation of our work is the computational power that i had which was much lower
than the required. It took us a long time to do simple calculations and as a result, my work took
alot of time due to the volume of the calculation. Similarly, Another limitation is the software
that I had used Quantum Espresso which is based on the plane wave method which is quiet
troublesome in case of Solids if we don’t give the proper selection of pseudo­potentials.

5.4 Recommendation for the further work

We can further extend our work by studying the thermoelectric,optical, response and trans­
port properties of NiCrSi as well as other materials. Similarly we can do doping at different
amount to see the effect of it in the electronic and magnetic properties of NiCrSi. Moreover
other calculation and studies such as Hubbard U (DFT+U), Berry’s phase polarization, Non­
collinear magnetism and spin­orbit coupling can also be done to extend our work further.

32
References
De Groot, R., Mueller, F., Van Engen, P., and Buschow, K. (1983). New class of materials:
half­metallic ferromagnets. Physical Review Letters, 50(25):2024.

Dinh, V. A., Sato, K., and Katayama­Yoshida, H. (2009). First principle materials de­
sign of half­metallic ferromagnetic half­heusler alloys. IEEE transactions on magnetics,
45(6):2663–2666.

Dinh, V. A., Sato, K., and Katayama­Yoshida, H. (2010). Half­metallicity and high­t c fer­
romagnetism in si­containing half­heusler alloys. Journal of superconductivity and novel
magnetism, 23(1):79–82.

Felser, C. and Hirohata, A. (2016). Properties, growth, applications.

Galanakis, I., Dederichs, P., and Papanikolaou, N. (2002). Slater­pauling behavior and origin
of the half­metallicity of the full­heusler alloys. Physical Review B, 66(17):174429.

Ghimire, M., Sinha, T., Thapa, R., et al. (2011). First principles study of the electronic and
magnetic properties of semi­heusler alloys nixsb (x= ti, v, cr and mn). Journal of alloys
and compounds, 509(41):9742–9752.

Kaxiras, E. (2003). Atomic and electronic structure of solids.

Kittel, C. and McEuen, P. (2018). Kittel’s Introduction to Solid State Physics. John Wiley &
Sons.

Nanda, B. and Dasgupta, I. (2003). Electronic structure and magnetism in half­heusler com­
pounds. Journal of Physics: Condensed Matter, 15(43):7307.

Poon, G. J. (2001). Electronic and thermoelectric properties of half­heusler alloys. In Semi­

conductors and semimetals, volume 70, pages 37–75. Elsevier.

Sholl, D. and Steckel, J. A. (2011). Density functional theory: a practical introduction. John
Wiley & Sons.

Singh, M., Saini, H. S., Kumar, S., and Kashyap, M. K. (2012). Effect of substituting sp­
element on half metallic ferromagnetism in nicrsi heusler alloy. Computational materials
science, 53(1):431–435.

33