Dirac Equation in a Nutshell

Xiao-Bin Qiang1, 2 and Hai-Zhou Lu1, 2, ∗

1
Shenzhen Institute for Quantum Science and Engineering and Department of Physics,
Southern University of Science and Technology (SUSTech), Shenzhen 518055, China
2
Shenzhen Key Laboratory of Quantum Science and Engineering, Shenzhen 518055, China
(Dated: July 12, 2023)
This note is mainly based on the Refs. [1–5].

CONTENTS

I. Hamiltonian of Dirac Equation 1

II. Magnetic Field and Zeeman Effect 2

III. Spin-orbital Coupling 3

IV. Topological Insulator 5

References 5

I. HAMILTONIAN OF DIRAC EQUATION

Schrödinger equation describe the microscopic particles at low energy limit. If the speed of particles approach to
the speed of light c, the relativistic effect will be very obvious. However the Schrödinger equation does not satisfy
Lorentz covariant, it can not give rise to the relativistic energy-momentum relation. The Schrödinger equation is not
the correct description for relativistic particles. By introducing the concept of spinor, Dirac deduced his equation
which is Lorentz covariant. It is known as Dirac equation, one of the most important and most beautiful equations
in the history of science. For the free particles, the Dirac equation in natural units (c = ~ = 1) reads
(iγ µ ∂µ − m)ψ = 0,
where γ µ is the basis of Clifford algebra, ∂µ = (∂t , ∂x , ∂y , ∂z ) is the 4-momentum operator, ψ is the spinor wavefunction,
and m is the rest mass. The basis of the Clifford algebra γ µ satisfy the relation
{γ µ , γ ν } = γ µ γ ν + γ ν γ µ = 2η µν ,
where η µν is the Minkowski metric (or special relativity metric) with signature (+, −, −, −), which describing the
geometry of the 4D flat spacetime. From this algebra, we can obtain that
(γ 0 )2 = 1,
(γ i )2 = −1,
γ µ γ ν = −γ ν γ µ (µ 6= ν).

In general, we adopt the Dirac representation γ i = βαi , γ 0 = β, where αi = σx ⊗ σi and β = σz ⊗ σ0 are Dirac
metrics, and σi are Pauli metrics. In some cases, we also need the Hamiltonian of Dirac equation. Back to the Dirac
equation, we have
(iγ 0 ∂t + iγ i ∂i − m)ψ = 0,
⇒ iγ 0 ∂t ψ = −iγ i ∂i ψ + mψ,
⇒ iβ∂t ψ = −iβαi ∂i ψ + mψ,
⇒ i∂t ψ = −iαi ∂i ψ + mβψ.

∗ luhz@sustech.edu.cn
 2

Recalling the 3-momentum operator pi = −i∂i , the Hamiltonian of free Dirac equation can be written as

H = cα · p + mc2 β.
p
The eigenenergies of this Hamiltonian are E = ± m2 c4 + p2 c2 . These two energies correspond to the electron and
positron with spin up and spin down. This means eigen states are degenerate for different spin. The reason of this
comes from a truth that the free Dirac equation respects PT symmetry.

II. MAGNETIC FIELD AND ZEEMAN EFFECT

Under a magnetic field B = ∇ × A, the Dirac Hamiltonian can be written as

H = cα · (p + eA) + mc2 β,

where −e is the electron charge, α = (αx , αy , αz ), and A is the vector potential. The eigenequation is

[cα · (p + eA) + mc2 β]ψ = Eψ.

Here the spinor wavefunction is a 4 × 1 vector. To distinct two branch of energies, we can denote
 
φ
ψ= ,
χ

where φ and χ are 2 × 1 vectors. Then, we have

   
0 σ σ 0
c · (p + eA)ψ + mc2 0 ψ = Eψ,
σ 0 0 −σ0
  
(mc2 − E)σ0 cσ · (p + eA) φ
⇒ = 0,
cσ · (p + eA) −(mc2 + E)σ0 χ

where σ = (σx , σy , σz ). Write it as a set of equations

(
(mc2 − E)σ0 φ + cσ · (p + eA)χ = 0,
cσ · (p + eA)φ − (mc2 + E)σ0 χ = 0,
(
(E − mc2 )φ = cσ · (p + eA)χ,
⇒
(E + mc2 )χ = cσ · (p + eA)φ.

In order to obtain the effective equation near the energy of mc2 (positive branch), we express χ by φ, i.e.,
1
χ= cσ · (p + eA)φ.
E + mc2
Substitute it into the first equation, we have
1
(E − mc2 )φ = cσ · (p + eA) cσ · (p + eA)φ.
E + mc2
Now we shift the zero point of energy to mc2 by noticing the relation between E and non-relativistic (Schrödinger)
energy Enon :

E = Enon + mc2 ,

where mc2 is the rest energy. Then,

1
(Enon + mc2 − mc2 )φ = cσ · (p + eA) cσ · (p + eA)φ,
Enon + mc2 + mc2
1
⇒ Enon φ = cσ · (p + eA) cσ · (p + eA)φ.
2mc2 [1 + Enon /(2mc2 )]
 3

At the low-energy limit, Enon  2mc2 , thus,

1
cσ · (p + eA) cσ · (p + eA)
2mc2 [1 + Enon /(2mc2 )]
1
'cσ · (p + eA) [1 − Enon /(2mc2 )]cσ · (p + eA)
2mc2
1
' [σ · (p + eA)][σ · (p + eA)].
2m
By using the algebra of Pauli matrices (σ · A)(σ · B) = A · B + iσ · (A × B), we have
1
Enon φ = [σ · (p + eA)][σ · (p + eA)]φ
2m
1 
= (p + eA)2 + iσ · [(p + eA) × (p + eA)] φ
2m
1 
(p + eA)2 + e~σ · (∇ × A) φ

=
2m
 
1 2 e~
= (p + eA) + σ · B φ,
2m 2m
where we have used
[(p + eA) × (p + eA)]φ
=(p × p + ep × A + eA × p + e2 A × A)φ
= − ie~[∇ × (Aφ) + A × ∇φ]
= − ie~[∇φ × A + (∇ × A)φ + A × ∇φ]
= − ie~(∇ × A)φ.
Consequently, we obtain the effective Hamiltonian under a magnetic field in the low-energy limit
1 e~
Hnon = (p + eA)2 + σ · B.
2m 2m
The second term is exactly the Zeeman Hamiltonian
e~ e
Hz = σ · B = S · B,
2m m
where S = (~/2)σ is the spin operator, µB = e~/(2m) = 5.788 × 10−5 eV/T is the Bohr magneton.

III. SPIN-ORBITAL COUPLING

Considering a potential V (r), the Dirac equation reads

H = cα · p + mc2 β + V,
The eigenequation is
(cα · p + mc2 β + V )ψ = Eψ.
Similarly, we can denote
 
φ
ψ= .
χ
Thus, the eigenequation becomes
   
0 σ 2 σ0 0
c · pψ + mc ψ = (E − V )ψ,
σ 0 0 −σ0
  
(mc2 − E + V )σ0 cσ · p φ
⇒ = 0,
cσ · p −(mc2 + E − V )σ0 χ
 4

Write it as a set of equations

(
(mc2 − E + V )σ0 φ + cσ · pχ = 0,
cσ · pφ − (mc2 + E − V )σ0 χ = 0,
(
(E − V − mc2 )φ = cσ · pχ,
⇒
(E − V + mc2 )χ = cσ · pφ.

In order to obtain the effective equation near the energy of mc2 (positive branch), we express χ by φ, i.e.,
1
χ= cσ · pφ.
E + V + mc2
Substitute it into the first equation, we have
1
(E − V − mc2 )φ = cσ · p cσ · pφ.
E − V + mc2
Now we shift the zero point of energy to mc2 by using E = Enon + mc2 , i.e.,
1
(Enon + mc2 − V − mc2 )φ = cσ · p cσ · pφ,
Enon + mc2 − V + mc2
1
⇒ (Enon − V )φ = cσ · p cσ · pφ.
2mc2 [1 + (Enon − V )/(2mc2 )]

At the low-energy limit, Enon − V  2mc2 , thus,

1
cσ · p cσ · p
2mc2 [1+ (Enon − V )/(2mc2 )]
 
1 Enon − V
'cσ · p 2
1− cσ · p
2mc 2mc2
1 1
= (σ · p)2 − σ · p(Enon − V )σ · p
2m 4m2 c2
2
p 1
= − σ · p(Enon − V )σ · p.
2m 4m2 c2
In order to get rid of the Enon , we use the fact that V and σ commute. We can write

(Enon − V )σ · p = σ · p(Enon − V ) + σ · [Enon − V, p]

p2
=σ·p + σ · [p, V ].
2m
Now, we have

p2 p3
 
(σ · p)(σ · [p, V ])
(Enon − V )φ = − − φ,
2m 8m3 c2 4m2 c2
 2
p3

p (σ · p)(σ · [p, V ])
⇒ Enon φ = +V − 3 2
− φ
2m 8m c 4m2 c2
 2
p3

p p · [p, V ] iσ · p × [p, V ]
= +V − − − φ,
2m 8m3 c2 4m2 c2 4m2 c2

where we have used (σ · A)(σ · B) = A · B + iσ · (A × B). The first two terms are just the Hamiltonian for the
non-relativistic Schrödinger equation. The third term is the relativistic correction to the kinetic energy. The forth
term is Darwin term. To analyze this term, note first that

[p, V ]φ = p(V φ) − V pφ
= (pV )φ + V pφ − V pφ
= −i~(∇V )φ.
 5

Thus, the Hamiltonian of Darwin term is

p · [p, V ]
Hdw = −
4m2 c2
1
= − 2 2 (−i~∇) · [−i~(∇V )]
4m c
~2
= ∇2 V.
4m2 c2
We are mainly concerned with the fifth term, the spin-orbit interaction. Accordingly, the spin-orbital coupling
Hamiltonian is
iσ · p × [p, V ]
Hso = −
4m2 c2
i
= − 2 2 σ · p × [−i~(∇V )]
4m c
~
= σ · (∇V × p).
4m2 c2
If we consider the Coulomb potential in a Hydrogen atom
e2 1
V =− ,
4π0 r
we can have
e2 r
∇V = .
4π0 r3
Then, the Hamiltonian is given by
~
Hso = σ · (∇V × p)
4m2 c2  2 
1 e r
= S· ×p
2m2 c2 4π0 r3
e2 1
= S · L,
8π0 m c2 r3
2

where L = r × p is the orbital angular momentum. This is exactly the spin-orbital coupling in Hydrogen atom.

IV. TOPOLOGICAL INSULATOR

In the condensed matter system, some low energy excitations around the Fermi level, i.e., quasi-particles, could be
described by the Dirac Hamiltonian. The only difference is that positron is replaced by hole, speed of light c is replaced
by Fermi velocity v, and m represents the band gap. However due to the particle-hole symmetry of Dirac equation,
there is no topological distinction between insulating bulk-state and vacuum. To avoid this issue, we introduce a
quadratic term −bp2 to the band gap, where b is a parameter. Thus, the Hamiltonian of a topological insulator can
be written as
H = vα · p + (m − bp2 )β.
This model can account various low-energy phases of matter from 1D to 3D. If parameter satisfies mb > 0, this model
is topological non-trivial and can be an effective model for topological insulator. If mb < 0, this model describes a
trivial insulator. One need notice that this Hamiltonian still respects PT symmetry.

[1] P. A. M. Dirac, The quantum theory of the electron, Proceedings of the Royal Society of London. Series A, Containing
Papers of a Mathematical and Physical Character 117, 610 (1928).
[2] P. A. M. Dirac, The principles of quantum mechanics, 27 (Oxford university press, 1981).
[3] S.-Q. Shen, Topological Insulators: Dirac Equation in Condensed Matter (Springer, 2017).
[4] J. Sakurai and J. Napolitano, Modern Quantum Mechanics (Cambridge University Press, 2017) pp. 105–108.
[5] D. J. Griffiths and D. F. Schroeter, Introduction to quantum mechanics (Cambridge university press, 2018).