253a: QFT1

Fall 2010
Matthew Schwartz

Lecture 10:
Spinors and the Dirac Equation
1 Introduction
From non-relativistic quantum mechanics, we already know that the electron has spin
usually write it as a doublet
 

| i = 

1
.
2

We
(1)

You probably learned that its dynamics in the non-relativistic limit are governed by the
Schrodinger-Pauli equation
 2




p
Bz
Bx iB y
K LK ) 1 0 2B
it | i = H | i = (
+ V (r) BB
| i
(2)
0 1
Bx + iB y
Bz
2m
e

where B = 2m is the Bohr magneton (the size of the electrons orbital magnetic moment) and
K = Kx Kp is thee angular momentum operator.
L
You may also have learned of a shorthand notation for this involving the Pauli matrices






0 1
0 i
1 0
1 =
, 2 =
, 3 =
(3)
1 0
i 0
0 1
which let us write the Schrodinger-Pauli equation more concisely


 2
p
K LK 122 2BBK K x
+ V (r) BB
it =
2m

(4)

This equation is written with the Pauli matrices combined into K = (1, 2, 3) to call attention to the fact that these is transform as the components of a 3-vector, just like the magnetic
field Bi. Thus (
K BK ) is rotationally invariant. This is non-trivial, and only works because
[i , j ] = 2iijkk

(5)

which are the same algebraic relations satisfied by infinitesimal rotations (we will review this
shortly). Another useful fact is
{i , j } i j + ji = 2i j

(6)

Keep in mind that i do not change under rotations they are always given by Eq. (3) in any
frame. is changing and Bi is changing, and these changes cancel when we write (
K BK ).
We could also have written down a rotationally invariant equation of motion for
1 t ii = 0

(7)

Since i transforms like a 3-vector and so does i, this equation is rotationally invariant. It
turns out it is Lorentz invariant too. In fact, this is just the Dirac equation!
If we write
= (122, 1, 2, 3)

(8)

= 0

(9)

Then it is

Section 2

which is nice and simple looking. Actually, this is the Dirac equation for a Weyl spinor, which is
not exactly the same as the equation commonly called the Dirac equation.
By the way, it does not follow that this equation is Lorentz invariant just because weve
written it as . For example,
( + m) = 0

(10)

is not Lorentz invariant. To understand these enigmatic transformation properties, we have to

know how to construct the Lorentz group out of the s and see how those act on . But
instead of just guessing the answer (or looking it up), were going to do something a little more
general, and motivate what kinds of transformation properties are possible.

2 Representations of the Lorentz Group

We have already argued that particles in our world should transform under unitary representations of the Poincare group. These are characterized by mass m and spin j. We motivated this
in our study of the spin 1 fields. Mass is Lorentz invariant, so it is an obvious quantum number.
Momentum is also conserved, but it is Lorentz covariant. If choose a frame in which the
momentum has some canonical form, for example p = (m, 0, 0, 0) for m > 0, then the particles
are characterized by the group that holds this momentum fixed. That is the little group, in this
case the group of 3D rotations, SO(3). The group of 3D rotations provides the second quantum
number, j. The way the polarizations transform under the full Lorentz group is then induced by
the transformations under SO(3) and the way the momentum transforms under boosts.
We also saw that to do field theory, we have to write down Largrangians involving fields.
These are things like V or or T which dont have to involve particles per se, although obviously in a physical theory they will. As we saw for spin 1, theres a lot a trouble that comes
from having to embed particles of fixed mass and spin into fields like V . For example, V has 4
degrees of freedom which describes spin 0 and spin 1, so the Lagrangian has to be carefully
chosen to make sure the physical theory never excites the spin 0 component. In addition, when
we want the particle to have m = 0 and spin 1, we needed to go further and make sure the longitudinal polarization is never produced. This led directly to charge conservation. The next logical step to make these embeddings a bit more systematic is to see what kinds of Lorentz1
invariant fields we can write down at all. This will reveal the existence of the spin 2 states, and
help us characterize their embeddings into fields.
A group is a set of elements {gi } and a rule gi gj = gk which tells how each pair of elements is multiplied to get a third. The rule defines the group, independent of any particular way
to write the group elements down as matrices. More precisely, the mathematical definition
requires the group elements to have inverses, there to be an identity element for which 1 gi =
gi, and for the rule to be associative (gi gj ) gk = gi (gj gk). A representation is a particular embedding of these gis into matrices. Often we talk about the vectors on which the
matrices act as being the representation, but technically the matrix embedding is the representation. Any group has the trivial representations r: gi 1. But more generally we care about
embeddings which are faithful, for which each element gets its own matrix. Recall that the
Lorentz group is the set of rotations and boosts which preserve the Minkoswki metric
T =

(11)

Lorentz transformation can act on space-time 4-vectors as

V V

(12)

Where is a combination of rotations and boosts. We saw that we could write a Lorentz transformation as the product of 3 rotations and 3 boosts:

Representations of the Lorentz Group

1
1
0
0
0
1
0 cos z sin z

cos

sin

1
y
y

=
0 sin z cosz 0

cosx
1
0
0
0
1
sin x
sin y
cos y

cosh y
sinh y
coshx sinhx
coshz
sinhx coshx

1
1

sinh y
1
1
cosh y
1
sinhz
1

sin x
cos x

sinhz

coshz

(13)

(14)

This is a particular representation of the Lorentz group. That is, it is one embedding of the
group into a set of matrices.
The group itself is a mathematical object independent of any particular representation. To
extract the group away from its representations, it is easiest to look at infinitesimal transformations. For vectors, the infinitesimal transformations V = V V are
V0 = iVi

(15)

Vi = iV0 ij k jVk

(16)

For any representation, we can always write these group elements as exponentials of the
infinitesimal generators
= exp(iiJi + iiKi) = 1 + i[iJi + iKi] + 

(17)
exp(i cigi)

cig

More generally, you could write any element g of any group G as g =

where
are
numbers and i are group generators. They are called generators because by multiplying by
numbers and exponentiating, you can get any element of the group.
The generators are in an algebra, because you can add and multiply them, while the group
elements are in a group because you can only multiply them. For example, the real numbers
form an algebra (there is a rule for addition and a rule for multiplication) but rotations are a
group (there is only one rule, multiplication). Groups that have an infinite number of elements
but a finite number of generators and do not have any weird singular points are called Lie
groups. Lie groups are critical to understanding the standard model, since the weak force is
described by the special unitary group SU(2) and the strong force by the group SU(3). For
QED, the only group we need is the Lorentz group, SO(3, 1). This is a special (determinant 1)
orthogonal (preserves a metric) group corresponding to a metric with (3, 1) signature (i.e. =
diag(1,-1,-1,-1)).
For G = SO(3), there are 6 generators: 3 rotations, i = Ji, and 3 boosts i+3 = Ki. For the
vector representation (above), the generators are

0
0
0

0 1
0
1
0
, J2 = V13 = i
, J3 = V23 = i

J1 = V12 = i

1 0
0
0 1
0
1
0
1 0

0 1
0
1
0
1
1 0

0
0
, K2 = V02 = i
, K3 = V03 = i

(18)
K1 = V01 = i

0
1
0
0
0
0
1
0

We can check that the generators in this representation satisfy some relations
[V01 , V12 ] = iV02 ,

[V12 , V23 ] = iV13 ,

(19)

which can be combined into the concise form

[V , V ] = i(V V V + V)

(20)

The generators in any other representation must satisfy these same relations. In fact, you can
even define the the Lorentz group is the set of transformations generated by these generators.

Section 3

For example, another representation of the generators is given by

J = i(x x )

(21)

These are the classical generators of angular momentum generalized to include time. You can
check that J satisfy the commutation relations of the Lorentz algebra.
Technically, we say the generators make up the Lorentz Algebra so(1,3) which generates the
Lorentz Group SO(1, 3). A common convention is to use lowercase letters for the names of algebras and uppercase letters for groups. As another technicality, we note that it is possible for two
different groups to have the same algebra. For example, the Proper Orthonchronous Lorentz
group and the Lorentz Group have the same algebra, but the Lorentz group has in addition the
discrete symmetries time-reversal (T) and parity-reversal (P). Its a small difference, but it is a
difference and the groups are not identical.

3 General representations
There is a very nice way to study the representations of the Lorentz group. Start with the rotation generators Ji and the boost generators K j . For example, you could take J1 = V23 , K1 = V01 ,
etc. as a particular representation of this basis. These satisfy
[Ji , J j ] = iijkJk

(22)

[Ji , K j ] = iijkKk

(23)

[Ki , K j ] = iijkJk

(24)

where ijk is defined by 123 = 1 and the rule that the sign flips when you swap any two indices.
For example, 213 = 1, 231 = 1, etc.. It is called the totally antisymmetric tensor and
comes up a lot in field theory. As is probably clear to you already, [Ji , J j ] = ii jkJk is the
algebra for rotations, so(3) so the Ji generate the subgroup of rotations in 3D.
Now take the linear combinations

which satisfy

1
Ji+ = (Ji + iKi)
2

(25)

1
Ji = (Ji iKi)
2

(26)

[Ji+, J j+] = iijkJk+

(27)

[Ji, J j] = iijkJk

(28)

[Ji+, J j] = 0

(29)

So we have found 2 commuting subgroups of the Lorentz group. The algebra of the J s is the
3D rotation algebra, so(3), more commonly called su(2). So we have shown that
so(1, 3) = su(2) su(2)

(30)

(technically, su(2) should really be sl(2, R) = so(1, 1) not su(2), but these algebras are the same
up to some is, and physicists usually just say su(2).).
The names so(n) and su(n) stand for special orthogonal algebra and special unitary
algebra. Orthogonal means it preserves a norm defined with transpose: V TV is invariant under
SO(n). Unitary means it preserves a norm defined with adjoint: V V is invariant under SU(n).
Special means the determinant is 1, which just normalizes the elements and gets rid of phases.
The decomposition so(3, 1) = su(2) su(2) makes studying the representations very easy. We
already know what the representations are of su(2), since this is the algebra of Pauli matrices,
which generates the 3D rotation group SO(3) (so(3) = su(2)). The representations are characterized by a quantum number j, and have 2j + 1 elements, labeled by m = jz, So representations of
the Lorentz group are characterized by two numbers A and B. The (A, B) representation has
(2A + 1)(2B + 1) degrees of freedom.

Spin

1
2

representation

K = JK + JK , where I now use the vector superscript to

The regular rotation generators are J
call attention the fact that the spins must be added vectorially (remember Clebsch-Gordon coefficients?). Since the 3D rotations, SO(3) is a subgroup of the Lorentz group, every representation of the Lorentz group will also be representations of SO(3). In fact, irreducible representations of the Lorentz group, which are characterized by two half-integers (A, B) produce many
representations of SO(3): with spins J = A + B, A + B 1,  , |A B |. For example,
+

rep of su(2) su(2) spins (SO(3) representations)

(A, B) = (0, 0)
J =0
1
1
(A, B) = ( , 0)
J=
2
2
1
1
(A, B) = (0, )
J=
2
2
1 1
(A, B) = ( , )
J = 1, 0
2 2
(A, B) = (1, 0)
J =1
(A, B) = (1, 1)
J = 2, 1, 0

(31)

We saw that the regular 4D vector representation A contains spins 1, 0 so now we understand
1 1
that it corresponds to the ( 2 , 2 ) representation of the Lorentz group. The general tensor repren n
sentations T 1 n corresponds to the ( 2 , 2 ). These are all irreducible representations of the
Lorentz group, but reducible representations of the SO(3) subgroup.

4 Spin

1
2

representation

All we know so far are the tensor representations, T 1 n which have only integer spins, so to
1
get representations containing spin 2 , we need something new.
1
1
1
There exist two spin 2 representations, ( 2 , 0) and (0, 2 ). What do these representations actually look like? We need to find 2x2 matrices that satisfy
[Ji+, J j+] = iijkJk+

(32)

[Ji, J j] = iijkJk

(33)

[Ji+, J j] = 0

(34)

But we already know such matrices: the Pauli matrices

[i , j ] = 2iij kk

(35)

i j
, ] = iij k k
2
2 2

(36)

{i , j } = i j + ji = 2ij

(37)

rescaling we find
[
Another useful fact is that

Thus we can set Ji+ = 2i . This is the 2 in ( 2 , 0). What about Ji? This should be the 0 in ( 2 , 0).
1
The obvious thing to do is just take the trivial representation Ji = 0. So the ( 2 , 0) representation is
1

( , 0): Ji+ = i , Ji = 0
(38)
2
2
1

Similarly, the (0, 2 ) representation is

1
(0, ):
2

Ji+ = 0,

Ji =

i
2

(39)

Section 4

K = JK + + JK
What does this mean for actual Lorentz transformations? Well, the rotations are J
K
and the boosts are K = i(J J+ ) so
1
( , 0):
2

Ji =

1
(0, ):
2

i
,
2

Ji =

i
,
2

Ki = i
Ki = i

i
2

i
2

(40)
(41)

Since the Pauli matrices are Hermetian i = i we see that the rotations are Hermetian but the
boosts are anti-Hermetian. This is the same as what we found for the vector representation.
Also notice that these two representations are complex conjugates of each other. In contrast, the
vector representation was real .
1
Explicitly, if L is a ( 2 , 0) spinor, known also as a left-handed Weyl spinor, then under rotations angles i and boost angles i
1

Similarly,

L e 2
1

Infinitesimally,

R e 2

(iii +ii)

1
1
L = (1 + iii + ii +  )L
2
2

(42)

(iii ii)

1
1
R = (1 + iii ii +  )R
2
2

(43)

1
L = (ii + i)iL
2

(44)

1
R = (ii i)iR
2

(45)

Note again the angles i and i are real numbers. Although we mapped Ji or Ji+ to 0, we still
have non-trivial action of all the Lorentz generators. So these are faithful irreducible representations of the Lorentz group. Similarly
1
L = ( ii + i)L i
2

(46)

i
R
= ( ii i)R
2

(47)

4.1 Unitary representations

We have just constructed two 2-dimensional representations of the Lorentz group. But these
representations are not unitary. Unitarity means = 1, which is necessary to have Lorentz
invariant matrix elements


(48)
h |i | |

Since the group element is the exponential of the generator = ei, unitarity requires that =
, that is, that be Hermetian.
Since su(2) is the special unitary algebra, all of its representations are unitary. So, the generK = JK i KK are Hermetian. Thus exp(i +i J+i +
ators for the su(2) su(2) decomposition J
i i
i
i
iJ) is unitary, for real + and . But this doesnt mean that the corresponding representations of the Lorentz group are unitary. We said that a Lorentz group element is
= exp(iiJi + iiKi)

(49)

where the i are the rotation angles and i the boosts angles. These are real numbers. They
i
are related to the angles for the J generators of su(2) su(2) by +
= i + ii and i = i ii.
So for a boost, the J+ and J generators get multiplied by imaginary angles, which makes the
transformation anti-unitary. Thus none of the representations of the Lorentz group generated
this way will be unitary. We have just proven that there are no finite dimensional unitary representations of the Lorentz group.

Spin

1
2

representation

To construct a unitary field theory, we will have to use the same trick we used for spin 1
particles. We will construct an infinite dimensional representation by having the basis depend
on the momentum i(p). For fixed momentum, say p = (m, 0, 0, 0) in the massive case, or p =
(E , 0, 0, E) in the massless case, the group reduces to SO(3) and ISO(2) respectively. These
groups do have unitary representations. For the case of spin 1, we were led to Lagrangians with
1
2
F
for spin 1, and gauge invariance and charge conservation if m = 0. For spin 2 , we will find
that unitary and Lorentz invariance requires fermions to anticommute.

4.2 Lorentz invariant Lagrangians

Having seen that we need infinite dimensional representations, we are now ready to talk
 about

1(x)
fields. These fields are spinor-valued functions of spacetime, which we write as L(x) =
2(x)
1
1
for the ( 2 , 0) representation, or R for the (0, 2 ) representation.
As in the spin-1 case, we would like first to write down a Lorentz invariant Lagrangian for
these fields with the right number of degrees of freedom, in this case 2. The simplest thing to do
would be to write down a Lagrangian with terms like
(L) L + m2(L) L

(50)

However, we can see from the above that this is not Lorentz invariant:
1
1
L L = (L) [(ii + i)iL] + [(L) ( ii + i)i]L
2
2

(51)

= iL iL  0

(52)

This is just the manifestation of the fact that the representation is not unitary because the
boost generators are anti-Hermetian.

If we allow ourselves two fields, L and R, we can write down terms like R
L. Under
infinitesimal Lorentz transformations,





1
1
(R
L) = R
( ii i)i L + R
(ii + i)i L = 0
(53)
2
2
Which is great. However, this term is not real. To make it real, we can just add the Hermetian
conjugate, so



(54)
m R
L + L R
is ok. This is a mass term we still have no dynamics.
What about kinetic terms? We could try

R
L + L R

(55)

which is both Lorentz invariant and real. But this is actually

not a very interesting Lagrangian.


1
We can always split up our field into components L = , where 1 and 2 are just regular
2

fields. Then we see that this is just the Lagrangian for a couple of scalars. So its not enough to
declare the Lorentz transformation properties of something, the Lagrangian has to force those
transformation properties. In the same way, a vector field is just 4 scalars until we contract it
with in the Lagrangian.
To proceed, lets look at
L iL

(56)

1
1
L iL = L i[(i j + j ) jL] + [L ( i j + j ) j ]iL
2
2

(57)

This transforms as

j
i
(i j + ji)L + j L (i j ji)L
2 L
2
= jL L jijkL kL

(58)
(59)

Section 5

Thus we have found that


 
L L , L iL = iL iL ,

iL L ij k jL kL

Which is exactly how a vector transforms

(V0, Vi) = (iVi , iV0 i jk jVk)

(60)

(61)

So V L = (L L , L KL) is an honest-to-goodness Lorentz 4-vector. Therefore,

L tL + L j jL
(62)
i
h
i
h
is Lorentz invariant. Note that t L L + j L jL is also Lorentz invariant, but not that

interesting because it is a total derivative.

Similarly, everything is the same for R but with i replaced by i:

R
R = iL iL

(63)

R
iR = iR
R ij k jR
kR

(64)

And V R = (R
R , R
jR) transforms like a vector. So L tL L j jL is Lorentz
invariant.
Defining

= (1, K ),

= (1, K )

We can write all the Lorentz-invariant terms we have found as




L = iL L + iR
R + m R
L + L R

We added a factor of i in the kinetic term to make the Lagrangian real:





iL L = i L L = iL L

where we have used = and integrated by parts.

Theres an even shorter-hand way to write this. If we write our spinor as a doublet


L
=
R
Then, define



= R
L
And using the 4x4 matrices



=

Our Lagrangian becomes
L = i ( m)

(65)

(66)

(67)

(68)
(69)
(70)
(71)

Which is the conventional form of the Dirac Lagrangian. The equations of motion which
follow are
which is the Dirac Equation.

( m) = 0

(72)

5 Dirac matrices
Expanding them out, the Dirac matrices are




1
0 i
0 =
i =
1
i 0

(73)

Dirac matrices

Or,

They satisfy

2 =

0
1
, 1 =

1
1
1
0

0
i
0

0 i
0
, 3 =

1
i 0
i
0
1

0 =

0
0

1
0
0
1

(74)

0
0

(75)

{ , } = 2

(76)

In the same way that the algebra of the Lorentz group is more fundamental than any particular
representation, the algebra of the s is more fundamental than any particular representation of
them. We say the s form the Dirac algebra, which is a special case of a Clifford algebra.
This particular form of the Dirac matrices is known as the Weyl representation.
The Lorentz generators are
i
(77)
S = [ , ]
4
They satisfy the Lorentz algebra for any s satisfying the Clifford algebra. That is, you can
derive from { , } = 2 that
[S , S ] = i(S S S + S)

(78)

Note: this is different from the 4-vector representation, with the matrices V . We have
found 2 different 4-dimensional representations. In each case, the group element is determined
1 1
by 6 real angles (3 rotations and 3 boosts). There is the vector or ( 2 , 2 ) representation,
which is irreducible, which has Lorentz group element
V = exp(i V )
1

(79)

and the Dirac representation, ( 2 , 0) (0, 2 ), which is reducible and has Lorentz generators
s = exp(i S )

(80)

There are actually a number of Dirac representations, depending on the form of the matrices.
We will use two: the Weyl and Majorana representations.
In the Weyl representation, the Lorentz generators are




1
i i
k
Si j = ij k
(81)
, Ki = S0i =
i
k
2
2
Or, very explicitly

1
1
1
S12 =
1
2

i
S01 =
2

0 1

,
, S23 = 1 1 0

0 1
0 1
2
1
1 0
1 0

0 1
0 1
1

1 0
1
, S02 = 1 1 0
, S03 = i

0 1
0
1
1
2
2
1 0
1 0
1
0 1
i

1 0
S13 =
2

(82)

(83)

These are block diagonal. These are the same generators we used for the ( 2 , 0) and (0, 2 ) representations above. It makes it clear that the Dirac representation of the Lorentz group is
reducible, it is the sum of a left-handed and a right-handed spinor representation.

10

Section 6

Another important representation is the Majorana representation

!
!
!
!
0 2
i 3 0
0 2
i 1
0
0
1
2
3
=
, =
, =
=
2 0
0 i 3
2 0
0
i 1

(84)

In this basis the s are purely imaginary. The Majorana is another ( 2 , 0) (0, 2 ) representation
of the Lorentz group which is physically equivalent to the Weyl representation.
The Weyl spinors, L and R are more fundamental than Dirac spinors like because they
correspond to irreducible representations of the Lorentz group. But the electron is a Dirac
spinor. Thus to do QED, it is easiest just to stick to the s and to get used to manipulating
them. Eventually, when you do supersymmetry, or study the weak interactions, you will need to
use the L and R representations again.

6 Rotations
Now lets see what happens when we rotate by an angle in the z plane. We use
(z) = exp(izJz )

(85)

How do we exponentiate a matrix? The standard trick is to first diagonalizing it with a unitary
transformation, do the rotation, then transform back. This unitary transformation is like
choosing a direction, except it is purely mathematical as the direction must be complex!
First, for the vector representation

0
0

0 1
1
= U 1
U
(86)
J3 = V12 = i

1 0
1
0
0
So,

exp( iz)
U
V (z) = exp(izV12 ) = U 1
(87)

exp(iz)
0
V (2) = 1

(88)

That is, we rotate 360 degrees and were back to where we started.
For the spinor representation
s(z ) = exp(izS12 )

(89)

the 12 rotation is already diagonal:

So,

1
S12 =
2

s(z) = exp(izS12 ) =

1
1

1
1

exp( 2 z)

(90)

exp( 2 z)

exp( 2 z)

exp( 2 z)

s(2) =

1
1

(91)

(92)

11

Lorentz Invariants

Thus a 2 rotation does not bring us back where we started! If we rotate by 4 it would. So we
1
say spinors are spin 2 . What does this mean physically? I have no idea. There are plenty of
1
physical consequences of spinors being spin 2 , but this business of rotating by 2 is not actually
a physical thing you can do.
As an aside, note that the factor of 2 is determined by the normalization of the matrices,
which is set by the Lie Algebra. For each representation by itself, this would be arbitrary, but it
is important for expressions which combine the representations to be invariant.

7 Lorentz Invariants
The matrices themselves transform nicely under the Lorentz group.
s1 s = (V )

(93)

where the s are the Lorentz transformations acting on each individually, as a matrix, and
the V is the vector representation which mixes up the Lorentz indices. That is, writing out the
matrix indices , this means
(s1) (s) = (V )

(94)

where refers to which matrix, and and index the elements of that matrix.
Then the equation
{ , } = 2

(95)

+ = 2

(96)

i
S = [ , ]
4

(97)

really means
And the equation
Should really be written as

S
=

For an expression like


i 

4

V V = V V = V { , }V

(98)
(99)

To be invariant, it must be that in transforms in the vector representation.

Next consider
( s)(s)

(100)

For this to be invariant, we would need s = s1, that is, for the representation of the Lorentz
group to be unitary. The spinor representation, like the vector representation is not unitary,
because the boost generators are anti-Hermetian.
It is useful to study the properties of the Lorentz generators from the Dirac algebra itself,
without needing to choose a particular basis for the . First note that
{ , } = 2

02 = 1,

i2 = 1

(101)

So the eigenvalues of 0 are 1 and the eigenvalues of i are i. Thus if we diagonalize 0, we

would find it Hermetian, and if we diagonalize 1, 2, or 3 we would find they are anti-Hermetian. This is true in general.
0 = 0,
Then,

(S ) =

i
[ , ]
4

i = i

ih i ih i
, = ,
4
4

(102)

(103)

12

Section 7

Which implies

Sij
= Sij

S0i
= S0i

(104)

Again, we see that the rotations are unitary and the boosts are not. You can see this from the
explicit representations above. But because we showed it algebraically, it is true in ANY representation of the Dirac algebra.
Now, observe that one of the Dirac matrices is Hermetian, 0 (0 is the only Hermetian
Dirac matrix because the metric signature is (1,-1,-1-,1)). Moreover
0 i0 = i = i,

= 0 0

0S
0 = 0

0 0 0 = 0 = 0

i i
ih
ih i
, 0 = 0 0, 0 0 = [ , ] = S
4
4
4

0) = exp( i S ) = s1
(0s0) = 0exp(i S ) 0 = exp( i 0S

(105)
(106)
(107)
(108)

Then, finally,


0 ( s)0(s) = 0s1s = 0

(109)

(110)

which is Lorentz invariant.

We have just been re-deriving from the Dirac algebra point of view what we found by hand
from the Weyl point of view. The conjugate of is not but

The point is that transforms according to s1. Thus is Lorentz invariant.

We can also construct objects like like
,

all transform nicely under the Lorentz group.

Also
L = (i m)

(111)

(112)

is Lorentz invariant. We abbreviate this with

L = (i m)

(113)

(i m) = 0

(114)

(i m ) = 0

(115)

This is the Dirac Lagrangian.

The Dirac equation follows from this Lagrangian by the equations of motion

To be explicit, this is shorthand for

By multiplying on the right, we find

1
1
0 = (i + m)(i m) = ( { , } [ , ] m2)
2
2

(116)

= ( 2 + m2)

(117)

So satisfies the Klein-Gordon equation

( + m2) = 0

(118)

It is in this sense that people sometimes say the Dirac equation is the square-root of the KleinGordon equation.

13

Coupling to the photon

We can integrate the Lagrangian by parts to derive the equations of motion for :

So,

L = (i m) = i ( m)

(119)

i m = 0

(120)

This on the right is a little annoying, so we often hide it by writing

( i m) = 0

(121)

where the derivative acts to the left. This makes the conjugate equation look more like the original Dirac equation.

8 Coupling to the photon

Under a gauge transform transforms just like a scalar, since the gauge invariance has nothing
to do with spin. So,
ei

(122)

Then we use the same covariant derivative ieA as for a scalar. So

D = ( ieA )

(123)

(i + e A m) = 0

(124)

Then the Dirac equation becomes

Now we try to reproduce the Klein Gordon equation for a scalar field coupled to A .

[(i + eA )2 m2

(125)

Following the same route as before

0 = (i + eA + m)(i + eA m)

(126)


2

= (i + eA )(i + eA ) m
=

(127)


1
1
{i + eA , i + eA }{ , } + [i + eA , i + eA ][ , ] m2
4
4

(128)

Before the antisymmetric combination dropped out, but now we find

[i + eA , i + eA ] = e[i A iA ] = eiF

So we get


ie

2
(i + eA ) + F [ , ] m
4
2

(129)
(130)

Which contains an extra term compared to the Klein-Gordon equation.

i
What is this term? Well, recall that 4 [ , ] = S , our Lorentz generators. These have the
form (in the Weyl representation)




i i
1
k
(131)
, S0i =
Sij = ijk
i
k
2
2
And since
F0i = Ei ,
We get
(

e
( ieA ) + m
2
2

Fij = i jkBk

K + i EK )K
(B

K iEK )K
(B

(132)
!)

=0

(133)

14

Section 10

This corresponds to a magnetic dipole moment. With conventional normalization, the size of the
e
magnetic moment is B = 2m . Weve made a physical prediction: charged fermions should have
e
magnetic dipole moments with size given by exactly B .
This is pretty remarkable physical result. For a free spinor, we reproduce the equation of
motion of a scalar field. But when the spinor is coupled to the photon, we find an additional
interaction corresponds to a magnetic dipole moment. We can read off that the electron has spin
1
. Note: the coupling to the electric field is not an electric dipole moment that would not have
2
an i, but is simply the effect of a magnetic moment in a boosted frame.

9 Probability current
The Noether current associated with the global symmetry ei is
J =

(134)

This, like any Noether current, is conserved on the equations of motion even if we set A = 0.
Note that the zero component of this current is

Q = = L L + R
R

(135)

We originally hoped this would be Lorentz invariant, which it is not. Now we see that it transforms as the 0 component of a conserved current. We can interpret this as the probability density for a fermion. The expectation value of Q is electron-number, which is the number of electrons minus the number of positrons. The spatial components of J denote electron numberflow. This is the same thing as the charge current, which couples to A up to a factor of the
electric charge e.

10 Helicity eigenstates
1

Dirac spinors, what we have been using, are 4 component complex fields in the ( 2 , 0) (0, 2 )
representation of the Lorentz group. Lets return for a moment to thinking about the 2-component fields.
In the Weyl basis, the matrices have the form


0
=
(136)
0
and Lorentz generators are block diagonal

(ii + i)i
=

(ii i)i

(137)

We can write our spinor as the left and right handed parts
=
The Dirac equation is

meaning

L
R




(138)


(139)

(i + eA )R = mL

(140)

(i + e A )L = mR

(141)

m i D
iD m

L
R

So the electron mass mixes the left and right handed states.

15

Helicity eigenstates

In the absence of a mass, this implies

0 = i R = (E + K Kp )R

(142)

0 = i L = (E K Kp )L

(143)

So the left and right handed states are eigenstates of the operator K Kp with opposite eigenvalue.
This operator projects the spin on the momentum direction. We call spin projected on the
direction of motion is called the helicity, so the left and right handed states have opposite
helicity.
The fact that projection of spin on the direction of momentum is a good quantum number
for massless particles works for massless particles of any spin. For any spin, we will always find
KKp s = s Es, where JK are the rotation generators of spin s. For spin 1/2, JK = K2 . For phoJ
tons, the rotation generators are listed in section 2. For example, Jz = V23 has eigenvalues 1
with eigenstates (0, i, 1, 0) and (0, i, 1, 0). These are the states of circularly polarized light in
the z direction. They are helicity eigenstates. So massless particles always have two helicity
states. It is true for spin 1/2 and spin 1, as we have seen, it is true for gravitons (spin 2),
Rarita-Schwinger fields (spin 3/2) and spins s > 2 (although, as we have seen, it is impossible to
have interacting theories with massless fields of spin s > 2).
We have seen that the L and R states

do not mix under Lorentz Transformation

L and R each have two components on which the s act. These are the two spin
states of the electron both left and right handed spinors have 2 spin states.

L and R have opposite helicity.

Using
=
The Lagrangian

L
R

0 =

R
L

(144)

L = (i + e A m)

(145)

L = iL D L + iR
D R m(L R + R
L)

(146)

becomes

Which is what we derived in the beginning. Note that L and R by themselves must be massless. To write down a mass term, we need both a L and a R.
It is helpful to be able to project out the left or right handed Weyl spinors from a Dirac
spinor. We can do that with the 5 matrix
5 = i 0 1 2 3

(147)

in the Weyl representation

5

=
So
PR =

1 + 5
=
2

0
1


,


,

PL =

1 5
=
2

(148)

1
0

These are projection operators since PR2 = PR and PL2 = PL and



 

 

L
L
L
0
=
=
, PL
PR
R
R
R
0

(149)

(150)

These projectors are useful because they are basis-independent.

10.1 Spin, Helicity, Chirality

In a relativistic theory, spin can be a confusing subjet. There are actually three concepts associated with spin: spin, helicity and chirality. It is worth keeping these straight.

16

Section 11

K = K2
Spin is a vector quantity. We say spin up, or spin down, etc. It is the eigenvalue of S
for a Fermion. If there is no angular momentum, the spin and the rotation operators are idenK = JK . We also talk about spin s, as a scalar, which is the eigenvalue s(s + 1) of the opertical S
2
1
K
ator S . This is what we mean when we say spin 2
Helicity refers to the projection of spin on the direction of motion. Helicity eigenstates satK Kp
S
isfy E = . Helicity eigenstates exist for any spin. For spin 1, circularly polarized light are
the helicity eigenstates.
Chirality is a concept that only exists for Fermions. It refers to the representation of the
Lorentz group the fields transform under, i.e. L or R for a Dirac Fermion. These are eigenstates of 5 for which we use L and R as subscripts. We write 5 L = L and 5 R = R.
3
Chirality works for higher half-integer spins too. For example, a spin 2 field can be put in a
Dirac spinor with a index, . Then 5 = are the chirality eigenstates. Up until now
we have been saying right-handed and left-handed, which was shorthand for chirality.
For free massless spinors, the spin eigenstates are also helicity eigenstates and chirality eigenstates. The Hamiltonian for the massless Dirac equation commutes with the operators for chiK Kp
J
K . The QED interaction A = L L +
rality, E , the helicity, 5, and the spin operators, S
RAR preserves chirality. Helicity, on the other hand, is not necessarily preserved by QED: if
a left-chirality spinor reverses direction, its helicity flips. Thinking about the helicity of spinors
at high energy is therefore useful, while thinking about chirality is not so much, because it never
changes.
For massive spinors, the free Hamiltonian no longer commutes with the chirality operator
due to the m = mLR + RL term. Thus even under free evolution L will pick up a R
K commutes
component over time. However, for the free theory which is rotationally invariant, J
with the Hamiltonian. Thus, for a free particle, spin, momentum, and helicity are all conserved.
Helicity is therefore a good quantum number for a free massive theory. However, in the massive
case, when we go to the non-relativistic limit, it is often easier to talk about spin, the vector.
Projecting on the direction of motion doesnt make so much sense when the particle is nearly at
rest, or in a gas, say, when its direction of motion is constantly changing. The QED interactions
do not preserve spin, however only a strong magnetic field can flip an electrons spin. So as long
as magnetic fields are weak, spin is good quantum number.
In practice, we hardly ever talk about chirality. The word is basically reserved for chiral theories, which are theories that are not symmetric under L R, such as the theory of the weak
interactions. We often talk about helicity. In the high energy limit, helicity is used interchangeably with chirality. As a slight abuse of terminology, we say L and R are helicity eigenstates.
In the non-relativistic limit, helicity is only used for photons, when it is synonymous with polarization of circularly polarized light. For fermions, we use spin, the vector, as the useful quantity.
11 Solving the Dirac equation
Lets take a break from the Dirac equation for a moment, and recall how we discovered antiparticles for complex scalar fields. The Lagrangian was
1
L = [( + ieA )][( ieA )] + m2
2

(151)

L = ( + ieA )( ieA )] + m2

(152)

The equations of motion are

( ieA )2 + m2 = 0

(153)

( + ieA )2 + m2 = 0

(154)

So we see that and have opposite charge, and we interpret them as particle and antiparticle. Recall that when we quantized the field , it created a particle and destroyed and antiparticle, and vice-versa for . But at the classical level, we can just think of as particle and
as antiparticle.

17

Solving the Dirac equation

How do we know if we have a particle or an antiparticle? There is an easy way to answer

this. The free equations for and are the same:
( + m2) = ( + m2) = 0

(155)

These have plane wave solutions

= peipx

(156)

In the rest frame, p20 = m2, so p0 = m. The solution with p0 = m is confusing. It is a legitimate solution to the equation of motion, but it says that these particles are going backward in
time! But note that
= peip0 t

= peip0 t

(157)

So we can just as easily interpret these solutions as anti-particles going forward in time. Obviously this interpretation is easier to swallow, but Feynman spent some time showing that there
really is no physically distinguishable difference.
Now back to spinors. The Dirac equation is
(i + e A m) = 0

(158)

( i eA m) = 0

(159)

So is a particle with mass m and opposite charge to : the positron.

11.1 free field solutions

Since spinors satisfy the Klein-Gordon equation ( + m2) = 0 in addition to the Dirac equation, they have plane wave solutions
Z
d3 p
s(x) =
us(p)eipx
(160)
(2)3
with p2 = m2. These are like the solutions A (x) = (p)eipx for spin-1 plane waves. In fact
there are solutions to the equations of motion p2 = m2 with p0 < 0 as well as p0 > 0. We already
know that the negative frequency solutions are not a problem since they correspond to antiparticles with positive frequency. Rather than work also with we simply write the other solutions
as
Z
d3 p
vs(p)eipx
(161)
s(x) =
(2)3
p
with p0 = Kp 2 + m2 > 0 for both. Then we can think of us and vs are the polarizations for particles and anti-particles, respectively. Note that in QM notation, this is consistent with , us
and vs being kets (column vectors).
Now lets use the Dirac equation. In the Weyl basis, (i m) becomes




m p
m p
us(p) =
vs(p) = 0
(162)
p m
p m
In the rest frame, p = (m, 0, 0, 0), the equations of motion reduce to


1 1
1 1




1 1
vs = 0
us =
1 1

(163)

So solutions are constants

us =

s
s


,

vs =

s
s

(164)

18

Section 11

For any two-component spinors s and s. Explicitly, a basis of solutions is

1
0
1
0
0
1
0
1

u1 =
1 , u2 = 0 , v1 = 1 , v2 = 0
0
1
0
1

(165)

The Dirac spinor is a complex 4-component object, with 8 degrees of freedom. The equations of
motion reduce it to 4 degrees of freedom: spin up and spin down for particle and anti-particle.
Now lets boost in the z-direction. Peskin and Schroeder do the actual boost. But well just
solve the equations again in the boosted frame and match the normalization. If p = (E , 0, 0, pz)
then




E pz
0
E + pz
0

p =
, p =
(166)
0
E + pz
0
E pz

Let a = E pz and b = E + pz , then m2 = (E pz)(E + pz)=ab.

Then the Dirac equation is

ab 0
a2
0

b2
0 ab 0
(i m) = 2
(167)
u (p) = 0
b
0 ab 0
0
a2
0 ab
The solutions are



a1
a 0
b2 0 b

us = u (p) =
(168)
b1 =  b 0 
0 a
a2

which are easy to check. Note that in the rest frame pz = 0, a2 = b2 = m, and reduces to s
above. So the solutions in the pz frame are

0
E pz
s

0
E + pz
!

E + pz
0

s
0
E pz

us(p) =

Similarly,

vs(p) =

Using

p =

0
E pz

0
E + pz

p =

(169)

E + pz
0

0
E pz

0
E pz

,
0
E + pz

(170)

0
E + pz

0
E pz

(171)

We can write more generally

us(p) =

p s

,
p s

vs(p) =

p s

p s

(172)

where the square root of a matrix means diagonalize it and then take the square root. In practice, we will always pick p along the z axis, so we dont really need to know how to make sense

of p . But this notation is useful because it makes Lorentz invariance manifest.

In the massless limit, the solutions are

!
!
0
0
0
0

s
0 2E s

0
2E
, vs(p) =

!
!
us(p) =
(173)

2E 0

2E 0
s
s
0

19

Solving the Dirac equation

More explicitly, the 4 solutions are

0
0

us(p) = 2E
1 , 2E
0

0
1
,
0
0

0
0

vs(p) = 2E
1 , 2E
0

0
1

0
0

(174)

For Weyl spinors, there are only 4 degrees of freedom off-shell, so there can only be 2 on-shell
once we use the equations of motion. Recalling that the Dirac equation splits up into separate
equations for L and R, we see that there is only one particle and one antiparticle solution in
the top two rows, and one particle and one antiparticle solution in the bottom two rows. Thus
on shell, the 2 degrees of freedom Weyl spinors have are particle and antiparticle, for the same
helicity.

11.2 normalization and spin sums

We have chosen a normalization here. To figure out what the normalization is, lets compute
some scalar quantity (a norm). The obvious one is
!
!


0
1
s
s

= us (p)0us (p) =
(175)
1 0
p s
p s
=2

s

E pz
0
0
E + pz



E + pz
0
0
E pz

s s

= 2mss

(176)
(177)

This is the (conventional) normalization for the inner product .

For m = 0 this way of writing the normalization convention is not useful. Instead we can calculate
!
!

s
s

us(p)us(p) =
= 2Es s = 2Ess
(178)
p s
p s

R 3
These are the same 2E factors which help make Lorentz invariance manifest for
d p integrals.
We can also compute the outer product
2
X

us(p)us(p) = p + m

(179)

s=1

where the sum is over the spins. Both sidesPof this


equation are matrices. Think of this equation
as a completeness relation, something like
|s
s|. For the antiparticles,
s
2
X

s=1

It is also true that

2
X

s=1

vs(p)vs(p) = p m

usvs =

2
X

(180)

usvs = 0

(181)

s=1

You should verify these relations on your own.

To keep straight the inner and outer products, it may be helpful to compare to spin 1 particles. We have found
hs|si:
X
s

|sihs|:

3
X
i=1

[i(p)]j(p) = ij

[i(p)]i (p) =

p p
m2

us(p)us (p) = 2mss

2
X

s=1

us(p)us(p) = p + m

(182)
(183)

20

Section 12

So when we sum Lorentz indices or internal spinor indices, we use an inner product and get a
number. When we sum over polarizations/spins, we get a matrix.

12 Spin and statistics a hint

Recall that we couldnt write down a mass term L L for just a left handed spinor because

So,

1
L = (i j + j ) jL
2

(184)

1
L = ( i j + j )L j
2

(185)

L L = jL jL  0

(186)

This means that L L is not boost invariant. We were able to write down a kinetic term for L,
but one might hope that there should also be some kind of bi-linear Lorentz invariant quantity
we can construct out of a (Weyl) spinor which would be a candidate for a probability in the
quantum theory. It turns out that there is such a thing:
T
L
2 L

(187)

is Lorentz invariant.
To see the Lorentz invariance, recall that for the Pauli matrices, 1 and 3 are real, and 2 is
imaginary.






1 0
0 i
0 1
(188)
, 3 =
, 2 =
1 =
0 1
i 0
1 0
So,
1 = 1,

2 = 2,

3 = 3

(189)

1T = 1,

2T = 2,

3T = 3

(190)

This implies 1T2 = 12 = 21, 3T2 = 32 = 23 and Tj 2 = 22 = 2 j . I.e.

Then

Tj 2 = 2 j

(191)

1
1
T
T T
T
(L
2) = (i j + j )L
j 2 = ( i j j ) L
2 j
2
2

(192)

Which cancels the transformation property of L. Thus

T
L
2 L

is Lorentz invariant.
Now, since 2 =

this is just



T
L2 L = 1 2
i

i
i



(193)

1
2

= i(1 2 2 1)

(194)

So, we have shown

We often write this as

(195)

1 2 2 1 is Lorentz invariant

1 2 2 1 = ,

0 1
1 0

(196)

which avoids picking a 2.

Theres only one problem: if the fermion components commute this is zero! For this to make
any sense, fermion components cant be regular numbers, they must be anticommuting numbers.
Such things are called Grassman numbers and satisfy a Grassman algebra. This is an example
of the spin-statistics theorem. I think its the simplest way to see the connection between
anti-symmetrization and Lorentz invariance.

21

Majorana fermions

You know about spin-statistics already from quantum mechanics. Say we have two lefthanded Weyl spinors and . Then T2 = i(1 2 2 1) is Lorentz invariant. This may
look more familiar if we use arrows for the and states:
1
|i = (|i|i |i|i)
2

(197)

This kind of two-particle wavefunction is automatically antisymmetric to the exchange of particles. This is the Pauli exclusion principle. Here we see that it is intimately related to Lorentz
invariance.
This isnt anything close to a derivation of the spin-statistics theorem. We just showed that
there is a Lorentz-invariant bi-linear we can construct from a Weyl fermion. In the next lecture,
we will show that the S-matrix is not Lorentz invariant unless Fermions anti-commute, which is
a much more convincing argument that the spin-statistics theorem must hold.

13 Majorana fermions
If we allow fermions to be Grassman numbers, then we can write down a Lagrangian for a single
Weyl spinor with a mass term
m
T

L
2 L)
(198)
L = iL L + i (L 2 L
2
These kinds of mass terms are called Majorana masses. Note that this mass term breaks the
symmetry under ei, since
T
T i
T
L
2 L L
e 2eiL = e2iL
2 L

(199)

So a particle with a Majorana mass cannot be charged.

The equation of motion for the Majorana fermion is

L + m2 L
=0

(200)

which follows from the Lagrangian above.

If we have a Majorana fermion, we can still use the Dirac algebra to describe it, but we have
to put it into a 4-component spinor


L
=
(201)

i2 L

This transforms like a Dirac spinor because 2 L

transforms like R. Then

m
m
T

= i (L 2 L
L
2 L)
2
2
which is the Majorana mass term.
Since (in the Weyl basis), using 22 = 1,


 

 
( i)( i)22 L
0 2
L
L

i2 = i
=
=
=

2 0
i2 L
i2 L
( i)( i)( 22)L

(202)

(203)

we can see that a Majorana fermion is one satisfying

= c i2

(204)

We call c the charge conjugate fermion. A Majorana fermion is its own charge conjugate.
Since it is real, it is also its own antiparticle.
Finally, note that in the Weyl basis 2 is imaginary and 0, 1, and 3 are real. Of course,
we we could just as well have taken 3 imaginary and 2 real, but its conventional to pick out
2. We can also define a new representation of the matrices by = 2 2. This satisfies the
Dirac Algebra because 22 = 1. Now define
c = i2

= i2 c

(205)

22

Section 13

If we take the Dirac equation

(i + e A m) = 0

(206)

and take the complex conjugate we get

( i + e A m) = 0

(207)

2( i + e A m)2 c = 0

(208)

(209)

(i eA m)c = 0

So c has the opposite charge from , which is another reason that Majorana fermions cant be
charged.

13.1 charge conjugation

The operation of charge conjugation
C:

c = i2

(210)

can be applied to any spinor, Dirac or Majorana. Let us see how the free Dirac spinors transform under charge conjugation. Recall that in the rest frame,




s
s
us =
, vs =
(211)
s
s
where and are constants, for spin up and spin down. Then,
|ic =

|ic =

1
0


  
  

1 c
0 i
1
0
= i2
=
=
= i|i
0
i 0
0
i

  
   
0 c
0 c
0 i
0
i
= i2
=
=
= i|i
1
1
i 0
1
0
c

(212)
(213)

So charge conjugation flips the spin. Thus

(usp)c =

s
s

(214)

This will be relevant to studying charge-conjugation invariance of general Lagrangians in QED

and beyond.

13.2 summary
In summary,
We have seen three types of spinors

Dirac Spinors: massive, left AND right handed

Weyl Spinors: massless, left OR right handed

Majorana spinors: real constrained Dirac spinors.

For a little physics on top of this algebra, theres a particle in nature called the neutrino which
is not charged. So it can be a Majorana or a Dirac fermion. In fact, were not sure what it is,
but were trying hard to find out. Weyl spinors are also important. They play a key role in the
theory of the Weak interactions. Weyl spinors are also critical for supersymmetry and string
theory. But for QED, we can just stick with Dirac spinors.