Quantum Field Theory PDF

L AGRANGIAN Q UANTUM F IELD T HEORY
IN M OMENTUM P ICTURE .
F REE F IELDS
L AGRANGIAN Q UANTUM F IELD T HEORY
IN M OMENTUM P ICTURE .
F REE F IELDS
B OZHIDAR Z. ILIEV
Nova Science Publishers, Inc.

New York

c 2008 by Nova Science Publishers, Inc.
All rights reserved. No part of this book may be reproduced, stored in a retrieval system or
transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical
photocopying, recording or otherwise without the written permission of the Publisher.
For permission to use material from this book please contact us:
Telephone 631-231-7269; Fax 631-231-8175
Web Site: http://www.novapublishers.com
NOTICE TO THE READER
The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or
implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of
information contained in this book. The Publisher shall not be liable for any special, consequential,
or exemplary damages resulting, in whole or in part, from the readers use of, or reliance upon, this
material.
Independent verification should be sought for any data, advice or recommendations contained in
this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage
to persons or property arising from any methods, products, instructions, ideas or otherwise
contained in this publication.
This publication is designed to provide accurate and authoritative information with regard to the
subject matter cover herein. It is sold with the clear understanding that the Publisher is not engaged
in rendering legal or any other professional services. If legal, medical or any other expert assistance
is required, the services of a competent person should be sought. FROM A DECLARATION OF
PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR
ASSOCIATION AND A COMMITTEE OF PUBLISHERS.
Library of Congress Cataloging-in-Publication Data

Iliev, Bozhidar Z. (Bozhidar Zakhariev)
Lagrangian quantum field theory in momentum picture : free fields / Bozhidar Z. Iliev.
p. cm.
Includes bibliographical references and index.
ISBN 978-1-60692-490-7
1. Quantum field theory. 2. Lagrangian functions. 3. Momentum (Mechanics) I. Title.
QC174.45.I45 2008
530.143dc22 2007049407
Published by Nova Science Publishers, Inc. New York

Contents
List of conventions xi
Preface xiii
Acknowledgments xv
I Lagrangian Formalism and the Momentum Picture 1

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Lagrangian Formalism . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3 Derivatives relative to operators and Lagrangian formalism . . . . . . . . . 4
4 Heisenberg Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
5 The Momentum Picture of Motion . . . . . . . . . . . . . . . . . . . . . . 8
6 General aspects of Lagrangian formalism in momentum picture . . . . . . . 12
7 On the momentum representation and particle interpretation . . . . . . . . 17
8 The momentum picture as 4-dimensional analogue of the Schrodinger one . 19
9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
II Free Scalar Fields 23

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
II.A Neutral scalar fields 25

2 Description of free neutral scalar field in momentum picture . . . . . . . . 25
3 Analysis of the Klein-Gordon Equation . . . . . . . . . . . . . . . . . . . 29
4 Frequency decompositions and their physical meaning . . . . . . . . . . . 32
5 The Klein-Gordon equations in terms of creation and annihilation operators 33
6 Commutation Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
7 Vacuum and Normal Ordering . . . . . . . . . . . . . . . . . . . . . . . . 42
8 State vectors and transitions between them . . . . . . . . . . . . . . . . . . 44
II.B Arbitrary scalar fields 47

9 Description of Free Scalar Field . . . . . . . . . . . . . . . . . . . . . . . 47
10 Analysis of the Field Equations . . . . . . . . . . . . . . . . . . . . . . . . 54
11 Frequency decompositions and their physical meaning . . . . . . . . . . . 56
12 The field equations in terms of creation and annihilation operators . . . . . 57
13 The charge and orbital angular momentum operators . . . . . . . . . . . . 61
v
vi Contents
16 State Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
17 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
III Free Spinor Fields 81

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2 Description of free spinor field in momentum picture . . . . . . . . . . . . 82
3 Analysis of the Dirac Equation(s) . . . . . . . . . . . . . . . . . . . . . . . 87
4 Frequency Decompositions . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5 The dynamical variables in terms of creation and annihilation operators . . 100
7 Anticommutation Relations . . . . . . . . . . . . . . . . . . . . . . . . . . 115
9 State Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
10 On the choice of Lagrangian and its consequences . . . . . . . . . . . . . . 131
11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
IV Free Vector Fields 145

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
2 Description of free vector field in momentum picture . . . . . . . . . . . . 147
3 Analysis of the Field Equations . . . . . . . . . . . . . . . . . . . . . . . . 153
4 Frequency decompositions and creation and annihilation operators . . . . . 158
5 The dynamical variables in terms of creation and annihilation operators . . 163
9 State Vectors and Particle Interpretation . . . . . . . . . . . . . . . . . . . 186
10 The massless case and electromagnetic field in Lorenz gauge . . . . . . . . 190
10.1 Problems in the Massless Case . . . . . . . . . . . . . . . . . . . . 190
10.2 Electromagnetic Field . . . . . . . . . . . . . . . . . . . . . . . . 195
11 On the Choice of Lagrangian . . . . . . . . . . . . . . . . . . . . . . . . . 198
12 On the role of the Lorenz condition in the massless case . . . . . . . . . . . 204
12.1 Description of free massless vector fields without the Lorenz con-
dition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
12.2 Analysis of the Euler-Lagrange Equations . . . . . . . . . . . . . . 205
12.3 Dynamical Variables . . . . . . . . . . . . . . . . . . . . . . . . . 208
12.4 The Field Equations . . . . . . . . . . . . . . . . . . . . . . . . . 209
12.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
13 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
V Commutation Relations for Free Fields 213

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
2 Lagrangians, Euler-Lagrange equations and dynamical variables . . . . . . 216
3 On the Uniqueness of the Dynamical Variables . . . . . . . . . . . . . . . 222
Contents vii
4 Heisenberg Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

5 Types of Possible Commutation Relations . . . . . . . . . . . . . . . . . . 234
5.1 Restrictions Related to the Momentum Operator . . . . . . . . . . 235
5.2 Restrictions Related to the Charge Operator . . . . . . . . . . . . . 240
5.3 Restrictions Related to the Angular Momentum Operator(s) . . . . 243
6 Inferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
7 State Vectors, Vacuum and Mean Values . . . . . . . . . . . . . . . . . . . 255
8 Commutation relations for several coexisting different free fields . . . . . . 262
8.1 Commutation relations connected with the momentum operator.
Problems and their possible solutions . . . . . . . . . . . . . . . . 264
8.2 Commutation relations connected with the charge and angular mo-
mentum operators . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
8.3 Commutation Relations between the Dynamical Variables . . . . . 270
8.4 Commutation Relations under the Uniqueness Conditions . . . . . 273
9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
References 277
Index 283
Subject Classes:
Quantum field theory
2000 MSC numbers: 2003 PACS numbers:

81Q99, 81S05, 81T99 03.70.+k, 11.10.Ef, 11.10.-z,
11.90.+t, 12.90.+b
Key-Words:
Quantum field theory, Lagrangian (canonical) quantum field theory
Pictures of motion, Pictures of motion in quantum field theory
Heisenberg, Schrodinger and momentum pictures
Lagrangian, Lagrangians for free fields
Momentum operator, Angular momentum operator
Spin and orbital angular momentum operators
Free quantum fields, Freee scalar, spinor (Dirac), vector and electromagnetic fields
Equations of motion, Euler-Lagrange equations
Equations of motion for free scalar, spinor and vector fields
Heisenberg equations/relations
Klein-Gordon equation, Dirac equation, Proca equation
Maxwell equations, Maxwell-Lorentz equations
Lorenz condition, Lorenz gauge, Coulomb gauge
State vectors and vacuum of free fields, Normal ordering
Commutation relations, Anticommutation relations
Commutation and anticommutation relations for free fields
Paracommutation relations, Parafermi and parabose commutation relations
List of Conventions
References. The book is divided into chapters which have a sequential Roman enumera-
tion. The chapters are divided into sections with a sequential Arabic enumeration, which is
independent in each chapter. Some section are divided into subsection,
In each chapter the subsections, equations, propositions, and so on have a double in-
dependent enumeration of the form m.n or (m.n) for the equations, m,n=1,2,. . . where m
is the number of the section in which the designated item appears and n is its sequential
number in it. So, section 4 and (3.12) (or equation (3.12)) mean respectively section 4 and
equation 12 in section 3 of the current chapter. A suitable item from a chapter different from
the current one is referred as R.m, or R.m.n or (R.m.n) for equations, where R=I,II,. . . is
the Roman number of the chapter in which the item appears; e.g. remark II.5.3 and IV.4
(or section IV.4) mean respectively remark 3 in section 5 of chapter II and section 4 in
chapter IV.
The footnotes are indicated as superscripts in the main text and have independent Arabic
enumeration in each section. When we refer to a footnote, it is on the current page if the
page on which it appears is not explicitly indicated.
Citations. An Arabic number in square brackets, e.g. [16], directs the reader to the list of
references, i.e. in this example [16] means the 16th item from the Bibliography list begin-
ning on page 277.
The ends of the proofs are marked by empty square sign, viz. with .
Einsteins summation convention: in a product of quantities or in a single expression,

a summation over indices repeated on different levels is assumed over the whole range in
which they change. Any exception of this rule is explicitly stated.
Minkowski spacetime and indices. The Minkowski spacetime is denoted by M. The

Greek indices run from 0 to dimM 1 = 3. All Greek indices will be raised and lowered
by means of the standard 4-dimensional Lorentz metric tensor and its inverse with
signature (+ ). The Latin indices a, b, . . . run from 1 to dim M 1 = 3 and, usually,
label the spacial components of some object. The local coordinates of a point x M are
denoted by x , i.e. they ar x0 , x1 , x2 , and x3 . If k are the components of a 4-vector k, the
defined by it 3-vector is described via the spacial components of k by k = (k1 , k2 , k3 ) =
(k1 , k2 , k3 ). Besides, we put k2 := k k = (k0 )2 k2 = k02 k 2 , with k2 := (k1 )2 + (k2 )2 +
xii List of Conventions
(k3 )2 , if there is not a risk to misunderstand this symbol by the second component of the
vector k.
Physical constants. Throughout this paper denotes the Plancks constant (divided by
2) and c is the velocity of light in vacuum.
The imaginary unit is denoted by i.
The Hermitian conjugation, transposition and complex conjugation of operators or

matrices are denoted by the superscripts , and , respectively.
The composition of mappings/operators is marked by the symbol .
Kronecker and Dirac deltas. By f g , or f or f g (:= 1 for f = g, := 0 for f = g)

g
is denoted the Kronecker -symbol, depending on arguments f and g, and n (y), y Rn ,

stands for the n-dimensional Dirac -function; (y) := 1 (y) for y R.
The derivatives with respect to operator-valued (non-commuting) arguments will be

calculated according to the rules of the classical analysis of commuting variables, which
is an everywhere silently accepted practice [1, 2]. As it is demonstrated in the paper [3],
whose concise summary is presented in section I.3, this is not quite correct but does not lead
to incorrect results (except the non-uniqueness of the conserved quantities) when free fields
are concerned. We shall pay attention on that item at the corresponding places in the text.
With an exception of the consequences of a charge symmetric Lagrangian, considered and
explained in Sect. III.10, this method for calculation of derivatives with respect to operators
does not lead to incorrect results in the other cases considered in this book.
The commutator or anticommutator of operators (in their common domain) or matrices

A and B is [A, B] := A B B A or [A, B]+ := A B + B A, respectively. We also put
[A, B] := A B + B A for C, so that [A, B] = [A, B]1.
Preface
The purpose of this monograph is an investigation of free scalar, spinor and vector fields
in the Lagrangian approach to quantum field theory. All initial hypotheses, calculations,
intermediate derivations and results are presented in details. Most of the considerations
are done in momentum picture of motion (or in terms of objects inherent to it) but the
Heisenberg picture of motion is not neglected if it is suitable.
The book can serve as a supplement to the (text)books in which the grounds of La-
grangian quantum field theory are presented or as a continuation of the corresponding ma-
terial presented in them. Its possible audience ranges from graduate and post-graduate
students to scientists interested in quantum field theory or mathematical physics.
In this book is considered only the Lagrangian (canonical) quantum field theory in
which the quantum fields are represented as operators, called field operators, acting on
some Hilbert space, which in general is unknown if interacting fields are studied. These
operators are supposed to satisfy some equations of motion, from them are constructed
conserved quantities satisfying conservation laws, etc. From the view-point of present-day
quantum field theory, this approach is only a preliminary stage for more or less rigorous
formulation of the theory in which the fields are represented via operator-valued distri-
butions, a fact required even for description of free fields. Moreover, in non-perturbative
directions, like constructive and conformal field theories, the main objects are the vacuum
mean (expectation) values of the fields and from these are reconstructed the Hilbert space
of states and the acting on it fields. Regardless of these facts, the Lagrangian (canonical)
quantum field theory is an inherent component of the most of the ways of presentation of
quantum field theory adopted explicitly or implicitly in books like [1, 410]. Besides, the
Lagrangian approach is a source of many ideas for other directions of research, like the
axiomatic quantum field theory [5, 9, 10].
The basic moments of the method, we will follow in describing free fields, are the
following ones:
i. In Heisenberg picture is fixed a (second) non-quantized and non-normally ordered op-

erator-valued Lagrangian, which is supposed to be polynomial (or convergent power
series) in the field operators and their first partial derivatives;
ii. As conditions additional to the Lagrangian formalism are postulated the commuta-
tivity between the components of the momentum operator (see (I.4.4) below) and the
Heisenberg relations between the field operators and momentum operator (see (I.4.1)
below);
xiv Preface
iii. Following the Lagrangian formalism in momentum picture, the creation and annihi-
lation operators are introduced and the dynamical variables and field equations are
written in their terms;
iv. From the last equations, by imposing some additional restrictions on the creation
and annihilation operators, the (anti)commutation relations for these operators are
derived;
v. At last, the vacuum and normal ordering procedure are defined, by means of which
the theory can be developed to a more or less complete form.
The main difference of the above scheme from the standard one is that we postulate the
below-written relations (I.4.4) and (I.4.1) and, then, we look for compatible with them and
the field equations (anti)commutation relations. (Recall, ordinary the (anti)commutation
relations are postulated at first and the validity of the equations (I.4.4) and (I.4.1) is explored
after that [2].)
The books [1, 4, 5] will be used as standard reference works on quantum field theory.
Of course, this is more or less a random selection between the great number of (text)books
and papers on the theme to which the reader is referred for more details or other points of
view. For this end, e.g., [6, 11] or the literature cited in [1, 46, 11] may be helpful. Some
historical remarks, bibliography list of more than 900 items and extended presentation of
quantum field theory can be found in [12]. More recent (text)books, like [1318], are also
worth mentioning in this context.

Any suggestions and comments are welcome. The authors postal address is
Bozhidar Zakhariev Iliev, Laboratory of Mathematical Modeling in Physics,

Institute for Nuclear Research and Nuclear Energy, Bulgarian Academy of
Sciences, Boul. Tzarigradsko chaussee 72, 1784 Sofia, Bulgaria,
his e-mail address is bozho@inrne.bas.bg and http://theo.inrne.bas.bg/bozho/ is the

URL address of his personal World Wide Web site.
Sofia, Bulgaria
11 February, 2007
Acknowledgments
The entire manuscript was typeset by the author by means of the main LATEX 2 document
preparation computer system together with a number of additional to it packages of pro-
grams, including first of all AMS-LATEX, AMS-fonts, B IBTEX, MakeIndex, Index, etc. My
gratitude to all of the numerous persons who created and developed, maintain and distribute
(free!) this valuable and high-quality typesetting system.
I would like to express my deep gratitude to my mother Dr. Lilyana Stefanova Shtereva
for her understanding, support, love and faith in me.
This research was partially supported by the National Science Fund of Bulgaria under
Grant No. F 1515/2005.
Bozhidar Z. Iliev
13 February, 2007
Sofia, Bulgaria
Chapter I
Lagrangian Formalism and

the Momentum Picture
Some features of the Lagrangian formal-

ism in quantum field theory are recalled.
The basic aspects of the momentum picture
of motion in Lagrangian quantum field
theory are given. Under some assumptions,
this picture is a 4-dimensional analogue
of the Schrodinger picture: in it the field
operators are constant, spacetime-indepen-
dent, while the state vectors have a simple,
exponential, spacetime-dependence. The
role of these assumptions is analyzed.
The Euler-Lagrange equations in momen-
tum picture are derived and attention is
paid on the conserved operators in it.
2 Bozhidar Z. Iliev
1. Introduction
The main item of the present chapter is a presentation of the basic aspects of the momentum
picture of motion in Lagrangian quantum field theory, suggested in [19]. In a sense, under
some assumptions, this picture is a 4-dimensional analogue of the Schrodinger picture: in
it the field operators are constant, spacetime-independent, while the state vectors have a
simple, exponential, spacetime-dependence. This state of affairs offers the known merits of
the Schrodinger picture (with respect to Heisenberg one in quantum mechanics [20]) in the
region of quantum field theory. Particular applications of the momentum picture to scalar,
spinor and vector free quantum fields can be found in the next chapters. The material in this
chapter is based on [21]
In Sect. 2 are reviewed some basic moments of the Lagrangian formalism in quantum
field theory. Sect. 3 is devoted to the mathematical problem of calculating derivatives rel-
ative to operator-valued arguments. Their rigorous treatment results in uniqueness of the
dynamical variables, rigorous derivation of all results involving such derivatives, consider-
ation of Lagrangians all of whose classical derivatives vanish, etc. . In Sect 4 are recalled
part of the relations arising from the assumption that the conserved operators are genera-
tors of the corresponding invariance transformations of the action integral; in particular the
Heisenberg relations between the field operators and momentum operator are written.
The momentum picture of motion is defined in Sect 5. Two basic restrictions on the
considered quantum field theories is shown to play a crucial role for the convenience of
that picture: the mutual commutativity between the components of the momentum operator
and the Heisenberg commutation relation between them and the field operators. If these
conditions hold, the field operators in momentum picture become spacetime-independent
and the state vectors turn to have exponential spacetime-dependence. In Sect 6, the attention
is called to the Euler-Lagrange equations and dynamical variables in momentum picture. In
Sect. 7 is given an idea of the momentum representation in momentum picture and the
similarity with that representation in Heisenberg picture is pointed. In Sect. 8 is made a
comparison between the momentum picture in quantum field theory and the Schrodinger
pucture in quantum mechanics.
Some closing remarks are given in Sect. 9. It is pointed that the above-mentioned
restrictions are fundamental enough to be put in the basic postulates of quantum field theory,
which may result in a new way of its (Lagrangian) construction.
2. Lagrangian Formalism
Let us consider a system of quantum fields, represented in Heisenberg picture of motion by
field operators i (x) : F F , with i = 1, . . ., n N, in systems Hilbert space F of states
and depending on a point x in Minkowski spacetime M. Here and henceforth, all quantities
in Heisenberg picture, in which the state vectors are spacetime-independent contrary to the
field operators and observables, will be marked by a tilde (wave) over their kernel
symbols. Let
L = L ( i (x), j (x)) (2.1)
Lagrangian Formalism and the Momentum Picture 3
be the systems Lagrangian, which is supposed to depend on the field operators and their
first partial derivatives.1 We expect that this dependence is polynomial or in a form of
convergent power series, which can be treated term by term. The Euler-Lagrange equations
for the Lagrangian (2.1), i.e.
L ( j (x), l (x)) L ( j (x), l (x))

= 0, (2.2)
i (x) x ( i (x))
are identified with the field equations (of motion) for the quantum fields i (x). In (2.2)
and similar expressions appearing further, the derivatives of functions of operators with re-
spect to operator arguments are calculated in the same way as if the operators were ordinary
(classical) fields/functions, only the order of the arguments should not be changed. This is
a silently accepted practice in the literature [1, 4, 5]. In the most cases such a procedure is
harmless, but it leads to the problem of non-unique definitions of the quantum analogues
of the classical conserved quantities, like the energy-momentum and charge operators. For
some details on this range of problems in quantum field theory, see [3] which is partially
summarized in section 3. In loc. cit. is demonstrated that these problems can be eliminated
by changing the rules of differentiation with respect to not-commuting variables. The pa-
per [3] contains an example of a Lagrangian (describing spin 12 field) whose field equations
are not the Euler-Lagrange equations (2.2) obtained as just described; see also sections 3
and III.10, in particular the Lagrangian (III.10.15).
For the moment, above and below, we consider a quantum field theory before normal
ordering and, possibly, without (anti)commutation relations (see Sect. 5). They will appear
in later stages of the theory development.
Following the standard procedure [1, 4, 5, 11] (see also [3]), from the Lagrangian
(2.1) can be constructed the densities of the conserved quantities of the system, viz. the
energy-momentum tensor T(x), charge current J (x), the (total) angular momentum den-
sity operator
M = L

(x) + S (x), (2.3)
where

L (x) = x T (x) x T (x) (2.4)

and S (x) are respectively the orbital and spin angular momentum density operators, and
others, if such ones exist. The corresponding to these quantities integral ones, viz. the mo-
mentum, charge, (total) angular momentum, orbital and spin angular momentum operators,
are respectively defined by:

1
P := T0(x) d3x. (2.5)
c
x0 =const

1
Q := J0 (x) d3x (2.6)
c
x0 =const
M = L(x) + S (x), (2.7)
1 One can easily generalize the below presented material for Lagrangians depending on higher order deriva-
tives.
4 Bozhidar Z. Iliev

1 0 0
L (x) := {x T (x) x T (x)} d3x (2.8a)
c
x0 =const

1
S(x) := S
0
(x) d3 x (2.8b)
c
x0 =const
and satisfy the conservation laws
d P d Q d M
=0 =0 =0 (2.9)
dx0 dx0 dx0
which, in view of (2.5)(2.8), are equivalent to
P = 0 Q = 0 M = 0 (2.10)
and also to
T = 0 J = 0
M = 0. (2.11)
The Lagrangian, as well as the conserved quantities and their densities, are Hermitian
operators; in particular, such is the momentum operator,
P = P. (2.12)
The reader can find further details on the Lagrangian formalism in, e.g., [1, 35, 11].
3. Derivatives Relative to Operators

and Lagrangian Formalism
This section contains a concise summary of the basic outcomes from the work [3] to which
the reader is referred for more details and results. It will not be used explicitly in this book,
but it justifies some decisions/choices and points how certain mathematical manipulations
can be made rigorous.
As pointed in [3, sections 13], the calculation of derivatives of function of operators
with respect to an operators according to the the known rules for functions of commuting ar-
guments is not mathematically correct. A consequence of applying these rules for functions
of operators is the known non-uniqueness of the expressions for the conserved quantities
in the quantum field theory; in fact, one has to guess or defined these expressions and,
generally, they cannot be derived from the proceeding theory.
To improve the things, one needs a suitable definition of a derivative with respect to
operators. For the goals of the Lagrangian quantum field theory is sufficient the next one.
Definition 3.1. Let F be a C-vector space, {F F } be a subset of the

space of operators acting on F , n N, u1 , . . ., un , u {u1 , . . ., un }, and
f : (u1 , . . ., un )
f (u1 , . . ., un ) : F F be operator-valued function of u1 , . . ., un
which is polynomial (or convergent power series) in its operator arguments. The derivative
of f with respect to u is an n-argument mapping with domain (n-times),
denoted by uf , such that:
f (u ,...,u )
(i) Its value at (u1 , . . ., un ), denoted by uf (u1 , . . ., un ) := 1
u
n
:=

f
u (u1 ,...,un ) , is a mapping {F F } from the subset on the space of opera-
tors on F .
(ii) The mapping u
: f
uf is linear relative to complex-valued functions on M. In
particular, it is C-linear.
(iii) Let v : F F be such that u + v , a N, i1 , . . ., ia {1, . . ., n} and
I := {i {i1 , . . ., ia } : ui = u} be the set of indices which label all operators among
ui1 , . . ., uia equal to u. Then

ui1 uia (v) := ui1 uia u =v . (3.1)
u iI
i
In particular, if I is empty, I = , the r.h.s. of (3.1) is set equal to the zero operator of
F.
Remark 3.1. The restriction u + v is essential one in quantum field theory, in which
the field operators, usually, satisfy some (anti)commutation relations and, hence, in it =
{F F }; for some general remarks on that item, see [22, sec. 21.1]. For instance, let
us find the derivative of A B with respect to A, where A, B : F F are anticommuting

operators, A B = B A. In this particular case A (A B) is defined only on those v : F
F for which (A + v) B = B (A + v), i.e. such that v B = B v, and hence =
(AB)
{z : F F : z B = B z}. In accord with (3.1), we have A (v) = v B = B
v = (BA)
A (v); the evaluation of the derivative on element w {F F }\ leads to a
contradiction, (AB)
A (w) =
(BA)
A (w).
Remark 3.2. From a view-point of functional analysis, the definition 3.1 defines the notion
of partial Frechet derivative of particular kind of functionals employed in quantum field
theory.
In short, definition 3.1 means that a derivative of operator-valued function, polynomial
or convergent power series, of operator arguments with respect to some of its operator
arguments is calculated by differentiating each its term according to (3.1). In particular,
this is valid for Lagrangians of the type we consider in this work.
It is trivial to be verified, the derivatives introduce via definition 3.1 possess all stan-
dard derivative properties; in particular, they satisfy the Leibnitz rule for differentiation of
compositions (products) of functions and the rule for differentiation of composite functions.
If c : M C and u + c(x) idF , x M, it is a trivial corollary of definition 3.1 that
f (u1 , . . ., un ) cl f (u1 , . . ., un )
(c(x) idF ) = c(x) , (3.2)
u u
where all operators are supposed to be linear and u means the classical derivative with
cl
respect to u, i.e. the derivative in the r.h.s. of (3.2) should be calculated as if u1 , . . ., un

were classical fields over M with a preservation of the relative order of all operators. This
6 Bozhidar Z. Iliev
is exactly the definition of a derivative of a function of non-commuting arguments ac-

cepted, e.g., in [23, 2]. For example, we have
3
(v) = v + v + v and
2 2
3
(c(x) idF ) = 3c(x)2 where a := . . . (a-times) for a N and : F F .
cl f f
Notice, u can be identically zero, while u may not vanish; for instance, if f =
g g for some operator g, than 0, while f (v) = v g g v is the zero operator
cl f
only for those operators v which commute with g.

Using definition 3.1, one gets the Euler-Lagrange relation
L L
i (x) x ( i (x)) vi = 0
(3.3)
i
instead of the Euler-Lagrange equations (2.2). Here the operators vi denote the variation of
the fields i . We should emphasize, now this is an equation both for i (x) and vi , contrary
to the standard procedure where one gets, due to the arbitrariness of vi , equations only for
i (x).1 So, the new moment with respect to the old variational principle is that (3.3) puts,
in general, restrictions both on the field operators and on their variations, i.e. the variations
cannot be considered as completely arbitrary (if one does not wont to deal with trivial fields
in some cases). Since the identity operator is always an admissible variation of the fields,
one gets from (3.3) with vi = i idF the classical Euler-Lagrange equations for the field
j
operators as
cl L cl L
= 0. (3.4)
i (x) x ( i (x))
But it should clearly be understood, if one requires (3.3) to be valid for completely
arbitrary variations vi , other restrictions on the field operators may arise.2 The alternative
point of view is to look on (3.4) as on field equations for the field operators and on the
remaining consequences of (3.3), if any, as on restrictions on the admissible variations vi .
The above remarks imply that there may exist Lagrangians for which the formalism gives
completely reasonable field equations which are not the Euler-Lagrange equations for it
(the latter being simply identities with respect to the fields and their variations); in fact, a
charge symmetric Lagrangian for spin 1/2 field exemplifies that case.
The utilization of definition 3.1 implies unique expressions for the conserved operator
quantities in the Lagrangian formalism. In particular, the energy-momentum tensor, charge
current and spin angular momentum density are respectively:

T (x) = i (x) i (x) L (x) (3.5)
i
1
J (x) =
ic i
qi ( i )i (x) i (x) (3.6)

S (x) = i (x)( j (x))Ii
j
, (3.7)
i, j
1 This does not exclude the coincidence of the final equations for ; in particular, such is the case with the
i
free fields described via standard Lagrangians.
2 In some cases, these new restrictions imply the field operators to be proportional to id , i.e., in a sense,
F
leading to classical, not quantum, fields.
where the quantities ( i ), qi and Ii are explained below after equation (4.3) and
j
L
i (x) := : {F F } {F F } (3.8)
( i (x))
is the derivative of L relative to i (x) according to definition 3.1.
4. Heisenberg Relations
The conserved quantities (2.5)(2.7) are often identified with the generators of the corre-
sponding transformations, under which the action operator is invariant [1, 2, 5, 11]. This
leads to a number of commutation relations between the conserved operators and between
them and the field operators. The relations of the latter set are often referred as the Heisen-
berg relations or equations. Part of them are briefly reviewed below; for details, see loc.
cit.
The consideration of P , Q and M as generators of translations, constant phase
transformations and 4-rotations, respectively, leads to the following relations:
i (x)
[ i (x), P ] = i (4.1)
x
[ i (x), Q ] = ( i )qi i (x) (4.2)
[ i (x), M ] = i{x i (x) x i (x) + Ii j (x)}.
j
(4.3)
Here: qi = const is the charge of the ith field, ( i ) = 0 if i = i , ( i ) = 1 if i =

i with ( i ) + ( i ) = 0, and the constants Ii = Ii characterize the transformation
j j
properties of the field operators under 4-rotations. 1 (It is a convention whether to put
( i ) = +1 or ( i ) = 1 for a fixed i.) Besides, the operators P , Q and M satisfy
certain commutation relation between themselves, from which we shall write the following
two:
[ P , P ] = 0 (4.4)
[ Q , P ] = 0. (4.5)
It should be clearly understood, the equations (4.1)(4.5) are from pure geometrical
origin and are completely external to the Lagrangian formalism. However, there are strong
evidences that they should hold in a realistic Lagrangian quantum field theory (see [2, 68]
and [1, 5.3 and 9.4]). Moreover, (most of) the above relations happen to be valid for
Lagrangians that are frequently used, e.g. for the ones describing free fields [2].
However, there is a difference between (4.4)(4.5) and (4.1)(4.3) from a physical view-
point: equation (4.4) (resp. (4.5)) expresses the simultaneous measurability between the
components of the momentum operator (resp. and the charge operator) and therefore are
physically motivated. On the other hand, equations (4.1)(4.3) are only geometrically mo-
tivated at this stage and only a posteriori they can obtain physical meaning.
More details on the Heisenberg relations will be presented in section V.4.
1 The matrices I := [I ij ] are responsible to the spin mixing in states obtained from states which are
eigenstates of the angular momentum operator M see sections III.4 and IV.4.
8 Bozhidar Z. Iliev
5. The Momentum Picture of Motion

Let P be the systems momentum operator, given by equation (2.5). Since P is Hermitian
(see (2.12)), the operator
1
i

U (x, x0) = exp (x x0 ) P , (5.1)

where x0 M is arbitrarily fixed and x M, is unitary, i.e.
U (x0, x) := ( U (x, x0)) = ( U (x, x0))1 =: U 1(x, x0 ). (5.2)
Let X F be a state vector in the systems Hilbert space F and A (x) : F F be an

operator on it. The transformations
X
X (x) = U (x, x0)( X ) (5.3)
A (x)
A (x) = U (x, x0) ( A (x)) U 1(x, x0), (5.4)
evidently, preserve the Hermitian scalar product | : F F C of F and the mean

values of the operators, i.e.
X | A (x)( Y ) = X (x)| A (x)( Y (x)) (5.5)
for any X , Y F and A (x) : F F . Since the physically predictable/measurable results

of the theory are expressible via scalar products in F [1, 4, 11], the last equality implies
that the theorys description via vectors and operators like X and A (x) above is completely
equivalent to the one via the vectors X (x) and operators A (x), respectively. The description
of quantum field theory via X and A (x) will be called the momentum picture (of motion
(of quantum field theory)) [19].
However, without further assumptions, this picture turns to be rather complicated. The
mathematical cause for this is that derivatives of different operators are often met in the
theory and that, as a consequence of (5.4), they transform as
A (x)
U (x, x0 ) ( A (x)) U 1(x, x0 ) = A (x) + [ A (x), H (x, x0 )] (5.6)

H(x, x0) := U (x, x0) U 1(x, x0) (5.7)
from Heisenberg to momentum picture. Here [ A , B ] := A B B A is the commutator

of A , B : F F . The entering in (5.6), via (5.7), derivatives of the operator (5.1) can be
represented as the convergent power series
1 1 1 1 n m nm
U (x, x0 ) = P + (x x0 ) P P (x x0 ) P ,
i i n=1 (i) (n+1)! m=0
n
where ( )n := ( ) ( ) (n-times) and ( )0 := idF is the identity mapping of F ,

which cannot be written in a closed form unless the commutator [ P , P ] has sufficiently
simple form.
In particular, the relation (4.4) entails
1
U (x, x0 ) = P U (x, x0), (5.8)
i
so that (5.7) and (5.6) take respectively the form

1
H(x, x0 ) =
P (5.9)
i
1
A (x)
A (x) + [ A (x), P ] . (5.10)
i
Notice, the equality (5.9) is possible if and only if U (x, x0 ) is a solution of the ini-
tial-value problem (see (5.7) and (5.8))
U (x, x0 )
i = P U (x, x0 ) (5.11a)
x
U (x0, x0 ) = idF , (5.11b)
the integrability conditions for which are exactly (4.4).1 Since (4.4) and (5.1) imply
[ U (x, x0 ), P ] = 0, (5.13)
by virtue of (5.4), we have

P = P, (5.14)
i.e. the momentum operators in Heisenberg and momentum pictures coincide, pro-
vided (4.1) holds.
It is worth to be mentioned, equation (5.14) is a special case of
A (x) = A (x) + [ U (x, x0), A (x)] U 1(x, x0), (5.15)
which is a consequence of (5.4) and is quite useful if one knows explicitly the commutator
[ U (x, x0 ), A (x)] . In particular, if

[ A (x), P ] , P = 0 (5.16)
and (4.4) holds, then, by expanding (5.1) into a power series, one can prove that
1
[ A (x), U (x, x0)] = (x x0 )[ A (x), P ] U (x, x0 ). (5.17)
i
So, in this case, (5.15) reduces to
1
A (x) = A (x) (x x0 )[ A (x), P ] . (5.18)
i
This formula allows to be found an operator in momentum picture if its commutator(s)
with (the components of) the momentum operator is (are) explicitly known, provided (4.4)
and (5.16) hold. The choice A (x) = P reduces (5.18) to (5.14).
1 For a system with a non-conserved momentum operator P (x) the operator U (x,x0 ) should be defined as
the solution of (5.11), with P (x) for P , instead of by (5.1); in this case, equation (4.4) should be replace with
[ P (x), P (x)] + P (x) P(x) = 0. (5.12)
Most of the material in the present section remains valid in that more general situation.
10 Bozhidar Z. Iliev
Of course, a transition from one picture of motion to other one is justified if there
are some merits from this step; for instance, if some (mathematical) simplification, new
physical interpretation etc. occur in the new picture. A classical example of this kind is
the transition between Schrodinger and Heisenberg pictures in quantum mechanics [20] or,
in a smaller extend, in quantum field theory [1]. Until now we have not present evidences
that the momentum picture can bring some merits with respect to, e.g., Heisenberg picture.
On the opposite, there was an argument that, without further restrictions, mathematical
complications may arise in it.
In this connection, let us consider, as a second possible restriction, a theory in which
the Heisenberg equation (4.1) is valid. In momentum picture, it reads
[i (x), P i H ] = i i (x), (5.19)
where i (x) are the field operators in momentum picture, i.e.
i (x)
i (x) = U (x, x0 ) i (x) U 1 (x, x0 ). (5.20)
is the transition from Heisenberg to momentum picture field operators, and the rela-
tions (5.4) and (5.6) were applied. The equation (5.19) shows that, if (5.9) holds, which
is equivalent to the validity of (4.4), then
i (x) = 0, (5.21)
i.e. in this case the field operators in momentum picture turn to be constant,
i (x) = U (x, x0 ) i (x) U 1 (x, x0 ) = i (x0 ) = i (x0 ) =: i . (5.22)
As a result of the last fact, all functions of the field operators and their derivatives,
polynomial or convergent power series in them, become constant operators in momentum
picture, which are algebraic functions of the field operators in momentum picture. This is an
essentially new moment in the theory that reminds to a similar situation in the Schrodinger
picture in quantum mechanics (see [20] and Sect. 8 below).
If P is considered, as usual [1,4], as a generator of 4-translations, then the constancy of
the field operators in momentum picture is quite natural. In fact, in this case, the transition
i (x)
i (x), given by (5.20), means that the argument of i (x) is shifted by (x0 x), i.e.
that i (x)
i (x) = i (x + (x0 x)) = i (x0 ).
Let us turn our attention now to systems state vectors. By definition [1,4], such a vector
X is a spacetime-constant in Heisenberg picture,
X = 0. (5.23)
In momentum picture, the situation is opposite, as, by virtue of (5.3), the operator (5.1)
plays a role of spacetime evolution operator, i.e.

X (x) = U (x, x0)( X (x0)) = e i (x x0 ) P ( X (x0)),
1
(5.24)
with
X (x0) = X (x)|x=x0 = X (5.25)
being considered as initial value of X (x) at x = x0 .

Thus, if X (x0 ) = X is an eigenvector of the momentum operators P = P (x)|x=x0
(= { U (x, x0 ) P U 1(x, x0 )}|x=x0 ) with eigenvalues p , i.e.

P ( X ) = p X = p X (x0 ) = P (x0 )( X (x0 )) , (5.26)
we have the following explicit form of a state vector X :

X (x) = e i (x x0 )p ( X (x0)).
1
(5.27)
It should be understood, this is the general form of all state vectors in momentum pic-
ture, as they are eigenvectors of all (commuting) observables [5, p. 59], in particular, of the
momentum operator.
So, in momentum picture, the state vectors have a relatively simple global descrip-
tion. However, their differential (local) behavior is described via a differential equation that
may turn to be rather complicated unless some additional conditions are imposed. Indeed,
form (5.24), we get
X (x) = H (x, x0 )( X (x)) (5.28)
in which equation the operator H (x, x0 ) is given by (5.7) and may have a complicated ex-
plicit form (vide supra). The equality (5.28) has a form similar to the one of the Schrodinger
equation, but in 4-dimensions, with 4-dimensional Hamiltonian i H (x, x0 ). It is intu-
itively clear, in this context, the operators i H (x, x0 ) should be identified with the com-
ponents of the momentum operator P , i.e. the equality (5.9) is a natural one on this back-
ground.
Thus, if we accept (5.9), or equivalently (4.4), a state vector X (x) in momentum picture
will be a solution of the initial-value problem
X (x)
i = P ( X (x)) X (x)|x=x0 = X (x0 ) = X (5.29)
x
and, respectively, the evolution operator U (x, x0 ) of the state vectors will be a solution
of (5.11). Consequently, the equation (4.4) entails not only a simplified description of the
operators in momentum picture, but also a natural one of the state vectors in it.
The above discussion reveals that the momentum picture is worth to be employed in
quantum field theories in which the conditions
[ P , P] = 0 (5.30a)
[ i (x), P ] = i i (x) (5.30b)
are valid. In that case, the momentum picture can be considered as a 4-dimensional
analogue of the Schrodinger picture [19]: the field operators are spacetime-constant and
the state vectors are spacetime-dependent and evolve according to the 4-dimensional
Schrodinger equation (5.29) with evolution operator (5.1). More details on that item will
be given in Sect. 8 below.
In connection with the conditions (5.30), it should be said that their validity is more a
rule than an exception. For instance, in the axiomatic quantum field theory, they hold identi-
cally as in this approach, by definition, the momentum operator is identified with the gener-
ator of translations [9,10]. In the Lagrangian formalism, to which (5.30) are external restric-
tions, the conditions (5.30) seem to hold at least for the investigated free fields and most
(all?) interacting ones [2]. For example, the commutativity between the components of
the momentum operator, expressed via (5.30a), is a consequence of the (anti)commutation
relations and, possibly, the field equations. Besides, it expresses the simultaneous mea-
surability of the components of systems momentum. The Heisenberg relation (5.30b) is
verified in [2] for a number of Lagrangians. Moreover, in loc. cit. it is regarded as one
of the conditions for relativistic covariance in a translation-invariant Lagrangian quantum
field theory. All these facts point that the conditions (5.30) are fundamental enough to be
incorporated in the basic postulates of quantum field theory, as it is done (more implicitly
than explicitly), e.g., in [1, 5, 11]. Some comments on that problem will be presented in
Sect. 9 (see also [22, chapter 1]).
However, as we mentioned at the end of section 4, there is a difference between (5.30a)
and (5.30b) from physical point of view: the former one is a natural, while the latter one
can be motivated only a posteriori. As we shall see in the next chapters, the Heisenberg
relations (5.30b) is incompatible with the consideration of the field variables i as operators
on a Hilbert space and requires they to be regarded as operator-valued distributions for the
restoration of its compatibility with the Lagrangian formalism.
6. General Aspects of Lagrangian Formalism

in Momentum Picture
In this section, some basic moments of the Lagrangian formalism in momentum picture will
be considered, provided the equations (5.30) hold.
To begin with, let us recall, in the momentum picture, under the conditions (5.30), the
field operators i are constant, i.e. spacetime-independent (which is equivalent to (5.30a)),
and the state vectors are spacetime-dependent, their dependence being of exponential type
(see (5.24) and (5.27)). As a result of this, one can expect a simplification of the formalism,
as it happens to be the case.
Combining (5.6), with A = i , (5.9), (5.21) and (5.14), we see that the first partial
derivatives of the field operators transform from Heisenberg to momentum picture accord-
ing to the rule
1
i (x)
yi := [i , P ] . (6.1)
i
Therefore the operator , when applied to field operators, transforms into i 1
[ , P ]
= i ( P P ), which is a differentiation of the operator space over F . An important
1
corollary of (6.1) is that any (finite order) differential expression of i (x) transforms in
momentum picture into an algebraic one of i . In particular, this concerns the Lagrangian
(which is supposed to be polynomial or convergent power series in the field operators and
their partial derivatives):
L
L : = L (i (x)) := U (x, x0) L ( i (x), j (x)) U 1(x, x0 )

= L U (x, x0 ) i (x) U 1 (x, x0 ), U (x, x0 ) j (x) U 1 (x, x0 )
1
= L i , [ j , P ] .
i
Thus, the Lagrangian (2.1) in momentum picture reads

1
L = L (i) = L (i , y j ) y j = [ j , P] , (6.2)
i
i.e. one has to make simply the replacements i (x)
i and i (x)
yi in (2.1).
Applying the general rule (5.4) to the Euler-Lagrange equations (2.2) and using (5.22)
and (6.1), we find, after some simple calculations,1 the Euler-Lagrange equations in mo-
mentum picture as
L ( , y ) 1 L ( , y )

, P
j l j l
= 0. (6.3)
i i yi y j = i
1
[ j , P ]
A feature of these equations is that they are algebraic, not differential, ones with respect
to the field operators i (in momentum picture), provided P is regarded as a given known
operator. This is a natural fact in view of (5.21).
We shall illustrate the above general considerations on the almost trivial example
of a free Hermitian scalar field , described in Heisenberg picture by the Lagrangian
L = 12 m2 c4 +c2 2 ( )( ) = L ( , y ), with m = const and y = , and sat-
isfying the Klein-Gordon equation ( + m22c2 idF ) = 0, := . In momentum picture

transforms into the constant operator
(x) = U (x, x0 ) U 1 (x, x0 ) = (x0 ) = (x0 ) =: (6.4)
which, in view of (6.3), L = m2 c4 , and yL = c2 2 y is a solution of
m2 c2 [[, P ] , P ] = 0. (6.5)
This is the Klein-Gordon equation in momentum picture, which is considered in detail

in chapter II. As a consequence of (5.14), this equation is valid in Heisenberg picture too,
when it is also a corollary of the Klein-Gordon equation and the Heisenberg relation (5.30b).
The Euler-Lagrange equations (6.3) are not enough for determination of the field oper-
ators i . This is due to the simple reason that in them enter also the components P of the
(canonical) momentum operator (2.5), which are functions (functionals) of the field opera-
tors. Hence, a complete system of equations for the field operators should consists of (6.3)
and an explicit connection between them and the momentum operator. This situation is
considered on particular cases in chapters IIIV.
Since the densities of the conserved operators of a system are polynomial functions
of the field operators and their partial derivatives in Heisenberg picture (for a polynomial
Lagrangian of type (2.1)), in momentum picture they became polynomial functions of i
and y j = i
1
[ j , P ] . When working in momentum picture, in view of (5.4) and (2.5)(2.8),
the following representations turn to be useful:

1
P = P = U 1(x, x0) T0 U (x, x0) d3x (6.6)
c
x0 =const
1 For details, see [19].

1
Q = U 1(x, x0 ) J0 U (x, x0) d3 x (6.7)
c
x0 =const

1
L (x) = U 1(x, x0 ) {x T 0 x T 0} U (x, x0) d3x (6.8)
c
x0 =const

1
S(x) = U 1(x, x0) S
0
U (x, x0 ) d3 x. (6.9)
c
x0 =const
These expressions will be employed essentially in what follows. In particular, the com-
bination of (6.6) and (6.3) (together with an explicit expression for the energy-momentum
tensor T) provide a closed algebraic-functional system of equations for determination of
the field operators i in momentum picture. In fact, this is the system of field equations in
momentum picture. Concrete types of such systems of field equations and their links with
the (anti)commutation (and paracommutation) relations will be investigated further in this
monograph.
d M
The conservation laws ddxQ0 = 0 and dx0 = 0 (or, equivalently, J = 0 and M
=
0), can be rewritten as

Q = 0 M = 0 (6.10)
since (2.6)(2.8) imply a Q = 0 and a M = 0 for a = 1, 2, 3.
As a result of the skewsymmetry of the operators (2.7) and (2.8) in the subscripts
and , their spacial components form a (pseudo-)vectorial operators. If eabc, a, b, c =
1, 2, 3, denotes the 3-dimensional Levi-Civita (totally) antisymmetric symbol, we put
1 2 3 a
M := ( M , M , M ) with M := eabc Mbc and similarly for the orbital and spin angular
momentum operators. Then (2.7) and the below written equation (6.23) imply
M = L(x) + S(x) (6.11)

M(x, x0 ) = L(x) + (x x0 ) P + S(x)
(6.12)
= M(x) + (x x0 ) P,
where x := (x1 , x2 , x3 ) = (x1 , x2 , x3 ), denotes the Euclidean cross product, and P :=

( P 1 , P 2 , P 3 ) = ( P1 , P2 , P3 ).
Obviously, the correction in (6.12) to M can be interpreted as a one due to an additional
orbital angular momentum when the origin, with respect to which it is determined, is change
from x to x0 .
The consideration of Q and M as generators of constant phase transformations and
4-rotations, respectively, leads to the following relations [1, 4, 11] (see (4.2) and (4.3))
[ i (x), Q ] = ( i )qi i (x) (6.13)

[ i (x), M ] = i{x i (x) x i (x) + Ii j (x), }.
j
(6.14)
where the notation was explained after equation (4.3) above. Besides, the operators (2.6)
(2.8) are Hermitian,
Q = Q , M = M , L

= L, S

= S, (6.15)
and satisfy the relations2
[ Q , P ] = 0 (6.16)
[ M , P ] = i{ P P }. (6.17)
Combining the last two equalities with (5.1) and (4.4), we, after a simple algebraic
calculations, obtain3
[ Q , U (x, x0 )] = 0 (6.20)
[ M , U (x, x0 )] = {(x x0 ) P (x x0 ) P } U (x, x0 ). (6.21)
Consequently, in accord with (5.4) or (5.15), in momentum picture the charge and an-
gular momentum operators respectively are
Q (x) = Q := Q (6.22)
M = U (x, x0) M U 1(x, x0 ) = M + [ U (x, x0), M ] U 1(x, x0)
= M + (x x0 ) P (x x0 ) P
= L + (x x0 ) P (x x0 ) P + S = L + S , (6.23)
where
L (x) := U (x, x0) L (x) U 1(x, x0)

(6.24)
S (x) := U (x, x0) S(x) U 1(x, x0)
and (5.14) was taken into account. Notice, the correction to M on the r.h.s. of (6.23) is
typical for the one of classical orbital angular momentum when the origin, with respect to
2 The author is completely aware of the fact that in the literature, for instance in [5, p. 77, eq. (2-87)] or
in [6, eq. (2.187)], the relation (6.17) is written with an opposite sign, i.e. with +i instead of i on its r h.s.
(In this case (6.17) is part of the commutation relations characterizing the Lie algebra of the Poincare group
see, e.g., [9, pp. 143147] or [10, sect. 7.1].) However, such a choice of the sign in (6.17) contradicts to
the explicit form of P and L in terms of creation and annihilation operators (see sections III.5, III.6, IV.6
and IV.7) in the framework of Lagrangian formalism. For this reason and since the relation (6.17) is external
to the Lagrangian formalism, we accept (6.17) as it is written below. In connection with this fact, see below
equations (III.6.28) and (IV.6.13) and the paragraphs containing them.
3 To derive equation (6.21), notice that (6.17) implies [ M , P P ] = n

1 n i=1 i P
i P P1 Pi1 Pi+1 Pn , due to [A,B C] = [A,B] C + B [A,C] , and expand the exponent
in (5.1) into a power series. More generally, if
[A(x), P ] = B (x) with [B (x), P ] = 0, (6.18)
then
1
[A(x), U (x,x0 )] = (x x0 )B (x) U (x,x0 ); (6.19)
i

in particular, [A(x), P ] = 0 implies [A(x), U (x,x0 )] = 0. Notice, we consider (x x0 ) as a real parameter
by which the corresponding operators are multiplied and which operators are supposed to be linear in it.
which it is determined, is changed from x to x0 .4

Let us consider the Heisenberg relations (4.1)(4.3) in momentum picture. As we said
above, the first of them reduces to (5.21) in momentum picture and simply expresses the
constantcy of the field operators i . In momentum picture, by virtue of (5.4), the rela-
tions (4.2) and (4.3) respectively read
[ i , Q ] = ( i )qi i (i ) := ( ) (6.29)
[ i , M (x, x0 )] = x [i , P ] x [i , P ] + iIi j .
j
(6.30)
The first of these equation is evident. To derive the second one, we notice that, by virtue
of the Heisenberg relations/equations (4.1), the equality (4.3) is equivalent to
[ i (x), M ] = x [ i (x), P ] x [ i (x), P ] + iIi j (x)

j
(6.31)
from where (6.30) follows.

It should be emphasized, the Heisenberg relations (4.2) and (4.3), as well as the commu-
tation relations (4.5) and (6.17), are external to the Lagrangian formalism. For this reason,
one should be quite careful when applying them unless they are explicitly proved in the
framework of Lagrangian scheme. So, in a pure Lagrangian approach, to which (6.29)
and (6.30) are external restrictions, one is not allowed to apply (6.29) and (6.30) unless
these equations are explicitly proved for the operators M and P given via (2.5)(2.8)
and (5.4).
In principle, from (6.6)(6.9) and the field equations (i.e. from (6.6) and (6.3)) can be
found the commutation relations between the conserved quantities and the momentum op-
erator, i.e. [ D , P ] with D = P , Q , M. If one succeeds in computing [ D , P ] , one can
calculate [ D , U (x, x0)] and, via (5.15), the operator D = P , Q , M in momentum pic-
ture. If it happens that (5.16) holds for A = D , then one can use simply the formula (5.18).
In particular, this is the case if the commutators [ D , P ] coincide with relations like (4.4)
and (4.5) (see also [2, 5]).5 For instance, if (4.5) holds, then (5.18) yields Q = Q , i.e.
4
In sections III.5 (in particular, see (III.5.19) and the text after it) and III.6, for a free spinor field, it will be
0 1 0 1 0 0 0 0
proved that S = S + S , L = L + L , where S and L are such that S = L = 0,
0 0
M = S + L and
0
[ S , P] = 0, (6.25)
which implies
0
[ S , U (x,s0 )] = 0. (6.26)
Amongst other things, from here follow the equations
0
0
S = S (6.27)
0
0
L = L + (x x0 ) P (x x0 ) P . (6.28)
5 In the next chapters, it will be proved that, in fact, the so-calculated commutators [ D , P ] reproduce

similar relations, obtained from pure geometrical reasons in Heisenberg picture, at least for the most widely
the charge operator remains one and the same in momentum and in Heisenberg pictures.
Obviously, the last result holds for any operator commuting with the momentum operator.
A constant operator C in Heisenberg picture,
C = 0, (6.32)
transforms in momentum picture into an operator C (x) such that

1
C (x) + [ C (x), P ] = 0, (6.33)
i
due to (5.10) and (5.14). In particular, the conserved quantities (e.g., the momentum, charge
and angular momentum operators) are solutions of the equation (6.33), i.e. a conserve oper-
ator need not to be a constant one in momentum picture, but it necessarily satisfies (6.33).
Obviously, a constant operator C in Heisenberg picture is such in momentum picture if and
only if it commutes with the momentum operator,
C (x) = 0 [ C (x), P ] = 0. (6.34)
Such an operator, by virtue of (5.16) and (5.18), is one and the same in Heisenberg and
momentum pictures,
C (x) = C . (6.35)
In particular, the dynamical variables which are simultaneously measurable with the
momentum, i.e. commuting with P , remain constant in momentum picture and, hence,
coincide with their values in Heisenberg one. Of course, such an operator is P = P , as
we suppose the validity of (5.30a), and the charge operator Q = Q , if (4.5) holds.
A (t)
Evidently, equation (6.33) is a 4-dimensional analogue of i t + [ A (t), H (t)] = 0,
which is a necessary and sufficient condition (in Schrodinger picture) for an observable
A (t) to be an integral of motion of a quantum system with Hamiltonian H (t) in non-rela-
tivistic quantum mechanics [20, 24].
7. On the Momentum Representation and

Particle Interpretation
An important role in quantum field theory plays the so-called momentum representation (in
Heisenberg picture) [1, 2, 11]. Its essence is in the replacement of the field operators i (x)
with their Fourier images i (k), both connected by the Fourier transform (kx := k x )1

e i kx i (k) d4k,
1
i (x) = (7.1)
and then the representation of the field equations, dynamical variables, etc. in terms of
i (k).
used Lagrangians. However, for the above purpose, one cannot use directly the last relations, except (4.4) in
this case, because they are external to the Lagrangian formalism, so that they represent additional restriction to
its consequences.
1 For brevity, we omit the inessential for us factor, equal to a power of 2, in the r h.s. of (7.1).
Applying the general rule (5.4) to (7.1), we see that the analogue of i (k) in momentum
picture is the operator
i (k) := e i kx U (x, x0 ) i (k) U 1(x, x0 ),

1
(7.2)
which is independent of x, depends generally on x0 and is such that

i = i (k) d4 k. (7.3)
A field theory in terms of the operators i (k) will be said to be in the momentum repre-
sentation in momentum picture.
The Heisenberg relation (5.30b) in momentum representation, evidently, reads
[ i (k), P ] = k i (k)
(7.4)
[ i (k), P ] = k i (k)
in Heisenberg and momentum picture, respectively.2 Consider a state vector X p with fixed
4-momentum p = (p0 , . . ., p3 ), i.e. for which
P( X p ) = p X p P ( X p ) = p X p . (7.5)
Combining these equations with (7.4), we get

P i (k)( X p ) = (p + k ) i (k)( X p) P i (k)( X p ) = (p + k ) i (k)( X p). (7.6)
So, the operators i (k) and i (k) increase the states 4-momentum p by k . If it hap-
that k0 0, we can say that these operators create a particle with 4-momentum
pens
( k2 + k 2 , k). (Notice k2 = k02 k 2 , k := (k1 , k2 , k3 ), need not to be a constant in the general
1
case, so the mass m := c k2 is, generally, momentum-dependent.) One can introduce the
creation/annihilation operators by

(k) for k 0 i (k) for k0 0

i 0
i (k) := i (k) := . (7.7)
0 for k0 < 0 0 for k0 < 0
(For k = 0 are also possible other definitions for

i (k) and i (k).) Equations (7.6) imply
3

P
i (k)( X p ) = (p k ) i (k)( X p )
(7.8)
P i (k)( X p ) = (p k ) i (k)( X p ).
Thus, if k0 0, we can interpret + +

i (k) and i (k) (resp. i (k) and i (k) ) as operators
creating (resp. annihilating) a particle with 4--momentum k .
2The equations (7.4) are a particular realization of a general rule, according to which any linear combina-
tion, possibly with operator coefficients, of i (x) and their partial derivatives (up to a finite order) transforms
into a polynomial in k , the coefficients of which are proportional to i (k). By virtue of (7.2), the same result
holds in terms of i (k) instead of i (k), i.e. in momentum picture.
3 Notice, generally (7.8) is not equivalent to (7.6) due to (k) = 0 for k < 0.
i 0
If the relations (4.2) (resp. (4.3)) hold, similar considerations are (resp. partially) valid
with respect to state vectors with fixed charge (resp. total angular momentum).
As we see, the description of a quantum field theory in momentum representation is
quite similar in Heisenberg picture, via the operators i (k), and in momentum picture, via
the operators i (k). This similarity will be investigated deeper in the next chapters. The
particular form of the operators i (k) and i (k) can be found by solving the field equations,
respectively (2.2) and (6.3), in momentum representation.
8. The Momentum Picture as 4-Dimensional Analogue of

the Schrodinger One
We have introduced the momentum picture and studied some its aspects on the base of the
Heisenberg one, i.e. the latter picture was taken as a ground on which the former one was
defined and investigated; in particular, the conditions (5.30) turn to be important from this
view-point. At that point, a question arises: can the momentum picture be defined indepen-
dently and to be taken as a base from which the Heisenberg one to be deduced? Below is
presented a partial solutions of that problem for theories in which the equations (5.30) hold.
First of all, it should be decided which properties of the momentum picture, consid-
ered until now, characterize it in a more or less unique way and then they or part of them
to be incorporated in a suitable (axiomatic) definition of momentum picture. As a guid-
ing idea, we shall follow the understanding that the momentum picture is (or should be)
a 4-dimensional analogue of the Schrodinger picture in non-relativistic quantum mechan-
ics. Recall, [20, 24, 25], the latter is defined as a representation of quantum mechanics in
which: (i) the operators, corresponding to the dynamical variables, are time-independent;
(ii) these operators are taken as predefined (granted) in an appropriate way; and (iii) the
wavefunctions are, generally, time-dependent and satisfy the Schrodinger equation
1
= H (), (8.1)
t i
with H being the systems Hamiltonian acting on the systems Hilbert space of states. A
4-dimensional generalization of (i)(iii), adapted for the needs of quantum field theory,
will result in an independent definition of the momentum picture. Since in that theory
the operators of the dynamical variables are constructed form the field operators i , the
latter should be used for the former ones when the generalization mentioned is carried out.
Besides, the field operators satisfy some equations, which have no analogues in quantum
mechanics, which indicates to a nontrivial generalization of item (ii) above.
Following these ideas, we define the momentum picture of quantum field theory as its
representation in which:
(a) The field operators i are spacetime-independent,
(i ) = 0. (8.2)
(b) The state vectors are generally spacetime-dependent and satisfy the following first
order system of partial differential equations
1
() = P (), (8.3)
i
where P are the components of the systems momentum operator (constructed according
to point (c) below see (8.8)). If 0 F and x0 M are fixed, the system (8.3) is supposed
to have a unique solution satisfying the initial condition
|x=x0 = 0 . (8.4)
(c) If D ( i , j ) is the density current of a dynamical variable in (ordinary) Heisen-

berg picture, which is supposed to be polynomial or convergent power series in i and
j , then this quantity in momentum picture is defined to be
1
D = D (i ) := D i , [ j , P ] . (8.5)
i
The corresponding spacetime conserved operator is defined as

1
D := U 1(x, x0) D (i ) U (x, x0) d3x, (8.6)
c x0 =const
where U (x, x0 ) is the evolution operator for (8.3)(8.4), i.e. the unique solution of the
initial-value problem
U (x, x0 ) 1
= P U (x, x0 ) (8.7a)
x i
U (x0, x0 ) = idF (8.7b)
with P corresponding to (8.6) with the energy-momentum tensor T for D ,

1
P := U 1(x,x0) T0(i ) U (x, x0) d3x. (8.8)
c x0 =const
(d) The field operators i are solutions of the (algebraic) field equations, which (in the
most cases) are identified with the Euler-Lagrange equations
L ( , y ) 1 L ( , y )
j l j l
, P = 0, (8.9)
i i yi y j = i
1
[ j , P ]
with L ( j , i
1
[ j , P ] ) being the systems Lagrangian (in momentum picture, defined ac-
cording to (8.5)).
A number or comments on the conditions (a)(d) are in order.
The transition from momentum to Heisenberg picture is provided by the inversion
of (5.3) and (5.4) with U (x, x0 ) given via (8.7), i.e.
X
X = U 1(x, x0)( X (x)) (8.10)
A (x)
A (x) = U 1(x, x0) ( A (x)) U (x, x0). (8.11)
Since (8.7) implies

1
H(x, x0 ) = P (8.12)
i
for the quantities (5.7), the replacement (5.10) is valid. In particular, we have
1
i
y j = [ j , P ] , (8.13)
i
by virtue of (8.2), which justifies the definition (8.5) and the equation (8.9). The Heisenberg
relations (5.30b) follow from this replacement:
[ i (x), P ] = U 1 (x, x0 ) [ i , P ] U (x, x0 ) = i ( i ).
Since the integrability conditions for (8.3) are

1
0 = () () = ( P ()) ( P ())
i
1
= ( ( P ) ( P))() + P ( ()) P ( ()) ,
i
where (8.3) was applied, the existence of a unique solution of (8.3)(8.4) implies (see (8.2);
cf. footnote 1 on page 9)
1
( P ) ( P ) + [ P , P ] = 0. (8.14)
i
As P = 0, due to the conservation of P , the replacement (5.6), with P for A (x),
together with (8.12) entails ( P ) + i
1
[ P, P ] = 0, which, when inserted into (8.14),
gives
( P ) = 0, (8.15)
which expresses the momentum conservation in momentum picture. The substitution of
equation (8.15) into (8.14) results in
[ P , P ] = 0, (8.16)
which immediately implies (5.30a).

As a result of (8.16) and (8.7), we obtain

U (x, x0) = e i (x x0 ) P ,
1
(8.17)
so that
[ U (x, x0 ), P ] = [ U 1 (x, x0 ), P ] = 0 (8.18)
and, consequently
P = U 1(x, x0) P U (x, x0) = P, (8.19)
which implies the coincidence of the evolution operators given by (5.1) and (8.7). The last
conclusion leads to the identification of the momentum picture defined via the conditions
(a)(d) above and by (5.3), (5.4) and (5.30) in Sect. 5.
What regards the conditions (c) and (d) in the definition of the momentum picture, they
have no analogues in quantum mechanics. Indeed, equations (8.5)(8.9) form a closed
system for determination of the field operators (via the so-called creation and annihilation
operators) and, correspondingly, they provide a method for obtaining explicit forms of the
dynamical variables (via the same operators). On the contrary, in quantum mechanics there
is no procedure for determination of the operators of the dynamical variables and they are
defined by reasons external to this theory.
Thus, we see that a straightforward generalization of the Schrodinger picture in quantum
mechanics to the momentum picture in quantum field theory (expressed first of all by (8.2)
and (8.3)) is possible if and only if the equations (5.30) are valid for the system considered.
9. Conclusion
In the present chapter, we have developed, analyzed and summarized the momentum pic-
ture of motion in (Lagrangian) quantum field theory, introduced in [19]. As it was shown,
this picture is (expected to be) useful when the conditions (5.30) are valid in (or compatible
with) the theory one investigates. If this is the case, the momentum picture has proper-
ties that allow one to call it a 4-dimensional Schrodinger picture as the field operators
(and functions which are polynomial in them and their derivatives) in it became space-
time-constant operators and the state vectors have a simple, exponential, dependence on the
spacetime coordinates/points. This situation is similar to the one in quantum mechanics
in Schrodinger picture, when time-independent Hamiltonians are employed [20], the time
replacing the spacetime coordinates in our case.
As we said in Sect 5, there are evidences that the conditions (5.30) should be a part of
the basic postulates of quantum field theory (see also [2, 68]). In the ordinary field the-
ory, based on the Lagrangian formalism to which (anti)commutation relations are added as
additional conditions [1, 4, 11], the validity of (5.30) is questionable and should be checked
for any particular Lagrangian [2]. The cause for this situation lies in the fact that (5.30)
and the (anti)commutation relations are additional to the Lagrangian formalism and their
compatibility is a problem whose solution is not obvious. The solution of that problem
is known to be positive for a lot of particular Lagrangians [2], but, in the general case, it
seems not to be explored. For these reasons, one may try to invert the situation, i.e. to
consider a Lagrangian formalism, to which the conditions (5.30) are imposed as subsidiary
restrictions, and then to try to find (anti)commutation relations that are consistent with the
so-arising scheme.
This program is realized in chapters IIIV for respectively free scalar, spinor and vector
fields. In them, it is demonstrated that the proposed method reproduces most of the known
results, reveals ways for their generalizations at different stages of the theory, and also
gives new results, such as a (second) quantization of electromagnetic field in Lorenz gauge,
imposed directly on the fields operator-valued potentials, and a natural derivation of the
paracommutation relations.
Chapter II
Free Scalar Fields
The chapter contains a

detailed investigation of free neutral
(Hermitian) or charged (non-Hermitian)
scalar fields and the describing them (system
of) Klein-Gordon equation(s) in momentum
picture of motion. A form of the field equa-
tion(s) in terms of creation and annihilation
operators is derived. An analysis of the
(anti-)commutation relations on its
base is presented. The concept
of the vacuum and the evo-
lution of state vectors
are discussed.

1. Introduction
This chapter pursuits a twofold goal. On one hand, it gives a detailed illustration of the
methods of Lagrangian quantum field theory in momentum picture, introduced in [19, 21]
and considered in chapter I, on the simplest examples of free Hermitian (neutral, real) or
non-Hermitian (charged, complex) scalar field. On another hand, it contains an in-depth
analysis of the (system of) Klein-Gordon equation(s) in momentum picture describing such
fields. Most of the known fundamental results are derived in a new way (and in a slightly
modified form), but new consequences are found too.
This chapter is based on the work [26] and is divided into two parts.
Part II.A, involving sections 28, deals with the case of neutral (Hermitian, real) free
scalar field. The contents of its sections is as follows:
In Sect. 2, the material of chapter. I concerning the momentum picture is specialized
in the case of free Hermitian (neutral, real) scalar field; in particular, the Klein-Gordon
equation in momentum picture is derived.
Sect. 3 is devoted to analysis of the Klein-Gordon equation (in momentum picture) in
terms of operators similar to (and, in fact, up to a phase factor and normalization constant,
identical with) the Fourier images of field operators in Heisenberg picture. From them, in
Sect. 4, are constructed the creation and annihilation operators which turn to be identical, up
to a phase factor and, possibly, normalization constant, with the ones known from Heisen-
berg picture. Their physical meaning is discussed (or recalled). In Sect. 5, the 3-dimen-
sional creation and annihilation operators (depending on the 3-momentum) are introduced
and the field equation is written in their terms. It happens to be a tri-linear equation relative
to them. This new form of the field equation is utilized in Sect. 6 for a detailed analysis
of the (additional) conditions leading to the known commutation relations. In particular,
it is proved that, excluding the vanishing field case, the quantization of a free Hermitian
scalar field by anti-commutators is rejected by the field equation without an appeal to the
spin-statistics theorem (or other equivalent to it additional assertion).
Sect. 7 is devoted to the introduction of the concept of vacuum (state) which requires
a modification of the developed theory by a normal ordering of products of creation and/or
annihilation operators. The vacua of Heisenberg and momentum pictures happen to coin-
cide. The problem of state vectors, representing in momentum picture a free Hermitian
scalar field, is considered in Sect. 8. It turns to be rather trivial due to the absence of any
interaction. However, the construction of Fock base is recalled and the basic ideas of scat-
tering theory are illustrated in this almost trivial case.
The second part II.B of the chapter is devoted to the general case of charged or neural
free scalar field. Regardless of some overlap with part II.A, it concentrates mainly on the
case of non-Hermitian field. Most of the proofs in it refer to or are based, at least partially,
on similar ones in the Hermitian case, investigated in part II.A. The problems of the choice
of the initial Lagrangian and the right definitions of the energy-momentum and charge
operators are partially discussed. The layout of part II.B, involving sections 916, is similar
to the one of part II.A.
The description of arbitrary free scalar field is presented in Sect. 9.
An analysis of the system of Klein-Gordon equations describing free scalar fields is
presented in Sect. 10. A feature of the momentum picture is that, in a case of non-Her-
Free Scalar Fields 25
mitian field, the two equations of this system are not separate equations for the field and
its Hermitian conjugate; they are mixed via the momentum operator. The creation and an-
nihilation operators are introduced in Sect. 11 and the field equations are written in their
terms in Sect. 12. They turn to be trilinear equations similar to the ones appearing the
parafield theory. The commutation relations are extracted from them in Sect. 14. These
reveal the non-equivalence of the theories build from different initial Lagrangians. To agree
the results, one needs different additional hypotheses/conditions, depending on the concrete
Lagrangian utilized. The best Lagrangian (of three considered) is pointed out. The charge
and orbital angular momentum operators are considered in Sect. 13. The vacuum is defined
in Sect. 15, where the normal ordering of the dynamical variables is described too. We point
that the last operation is the final step which leads to identical theories build from different
initial Lagrangians. Some problems concerning the state vectors of free scalar fields are
discussed in Sect. 16.
Sect. 17 ends the chapter by pointing to its basic results.
II.A. Neutral Scalar Fields
The purpose of the first part of the present chapter is a detailed exploration of a free neutral
(Hermitian, real, uncharged) scalar field in momentum picture.1
After fixing the notation and terminology, we write the Klein-Gordon equation in mo-
mentum picture and derive its version in terms of creation and annihilation operators. The
famous commutation relations are extracted from it. After defining the vacuum for a free
Hermitian scalar field, some problems concerning the state vectors of such a field are inves-
tigated.
2. Description of Free Neutral Scalar Field

in Momentum Picture
Consider a free neutral (Hermitian) scalar field with mass m. The corresponding to such a
field operator will be denoted by . It is Hermitian, i.e.
(x) = (x), (2.1)

1 A classical real field, after quantization, becomes a Hermitian operator acting on the systems (fields)
Hilbert space of states. That is why the quantum analogue of a classical real scalar field is called Hermitian
scalar field. However, it is an accepted common practice such a field to be called (also) a real scalar field. In
this sense, the terms real and Hermitian scalar field are equivalent and, hence, interchangeable. Besides, since
a real (classical or quantum) scalar does not carry any charge, it is called neutral or uncharged scalar field too.
where the dagger denotes Hermitian conjugation relative to the scalar product | of
fields Hilbert space F , and it is described in the Heisenberg picture by the Lagrangian
1 1
L = L ( , ) = m2 c4 + c2 2( ) ( ), (2.2)
2 2
which in momentum picture transforms into (see (I.6.2))
1 1
L = L (, y )y = 1 [0 , P ] = m2c4 0 0 c2 [0 , P] [0 , P ] . (2.3)
i 2 2
Here 0 is the constant value of the field operator in momentum picture i.e. (see
(I.5.22))1
(x) = U (x, x0) (x) U 1 (x, x0 ) = (x0 ) = (x0 ) =: 0 . (2.4)
Since the operator U (x, x0) is unitary (see (I.5.2) and use the Hermiticity of the momentum
operator), the (momentum) field operator 0 is also Hermitian, i.e.
0 = 0 . (2.5)
So, we have 2
L (0 , y )
:= 1 = ic2 [0 , P ] (2.6)
y y = i [0 , P ]
L L
= = m2 c4 0 . (2.7)
0 0
Consequently, equation (I.6.3), in this particular situation, yields
m2 c2 0 [[0 , P ] , P ] = 0. (2.8)
This is the Klein-Gordon equation in momentum picture. It replaces the usual

Klein-Gordon equation
2 2
+ m c idF (x) = 0, (2.9)
2
:= is the DAlembert operator in Heisenberg picture and idF is the identity
where
mapping of F .
Exercise 2.1. Prove that the DAlembert operator on the space of operator-valued functions
in momentum picture is
2 1
() = () + [ ( ), P ] 2 [[ , P ] , P ] . (2.10)
i
The index 0 in 0 indicates the dependence on x0 , according to (I.6.4).
1
2As pointed on page xii, the calculation of the derivatives in (2.6) and (2.7) below is not quite correct
mathematically. However, the field equation (2.8) is correct; for its rigorous derivation, see [3, section 5.1] and
section I.3.
(Hint: apply (I.5.10) twice.) In particular, we have

1
( A ) = [[ A , P ] , P ] if A = 0
2
and hence (i ) = 12 [[i , P ] , P ] for any field operator i in momentum picture. The
equation (2.8) follows immediately from here and the usual Klein-Gordon equations (2.9).
Equation (2.8) is equivalent to m2 c2 0 [[ 0 , P ] , P ] = 0 which is equivalent to (2.9)
due to the (Heisenberg) relation [ 0 , P ] = i .
The energy-momentum tensorial operator T has two (non-equivalent) forms for free
Hermitian scalar field, viz.
T = L = c2 2( ) ( ) L (2.11a)
1
T = ( + ) L
2 (2.11b)
1
= c2 2 ( ) ( ) + ( ) ( ) L ,
2
which in momentum picture read respectively:
T = c2 [0, P] [0 , P ]
1 (2.12a)
+ c2 m2 c2 0 0 + [0 , P ] [0 , P ]
2
1 2
T = c [0 , P ] [0, P] + [0 , P ] [0 , P]
2 (2.12b)
1
+ c2 m2 c2 0 0 + [0 , P] [0 , P ] .
2
Let us say a few words on the two versions, (2.11a) and (2.11b), of the energy-mo-
mentum operator. The former one is a direct analogue of the classical expression for the
energy-momentum tensor, while the latter variant is obtained from (2.11a) by a Hermitian
symmetrization. The second expression is symmetric and Hermitian, i.e. T = T and
T = T, while the first one is such if and commute for all subscripts and .
However, as we shall see, both forms of T lead to one and the same (Hermitian) momen-
tum operator. In this sense, the both forms of T are equivalent in quantum field theory.
More details on this problem will be given in Sect. 9. 3
Before going on, let us make a technical remark. The derivatives in (2.6) and (2.7) are
calculated according to the rules of classical analysis of commuting variables, which is not
correct, as explained in [3] and partially in section I.3. However, the field equation (2.8)
or (2.9) is completely correct for the reasons given in loc. cit. Besides, the two forms (2.11)
of the energy-momentum tensor operator are also due to an incorrect applications of the
rules mentioned to the region of analysis of non-commuting variables; the correct rigorous
expression turns to be (2.11b). The reader is referred for more detail on that item to [3],
3 The discussed non-uniqueness of the energy-momentum operator, as well as of other conserved currents,
if any, is due to the accepted rules for differentiation with respect to non-commuting variables. A rigorous
treatment of this kind of differentiation restores the uniqueness of the conserved currents see [3] and the next
paragraph.
in particular to section 5.1 in it. The only reason why we use a non-rigorous formalism is
our intention to stay closer to the standard books on Lagrangian quantum field theory. This
approach will turn to be harmless for the most of the problems considered in the present
monograph.
Exercise 2.2. Applying the facts presented in section I.3, rigorously derive the Klein-
Gordon equation (2.8) and prove that the correct expression for the energy-momentum
tensor is (2.12b), not (2.12a).
A free neutral or charged scalar field has a vanishing spin angular momentum and pos-
sesses an orbital angular momentum, which coincides with its total angular momentum.
The orbital angular momentum density operator is

L (x) = L (x) = x T (x) x T (x) (2.13)

L
= L = x T x T (2.14)
in Heisenberg and momentum pictures, respectively. The angular momentum operator in

Heisenberg and momentum picture, respectively, is

1 1 0 0
L := L
0
(x) d3 x = {x T (x) x T (x)} d3x
c c
x0 =const x0 =const

1
= U 1(x, x0 ) {x T 0 x T 0} U (x, x0) d3x (2.15)
c
x0 =const
L(x, x0 ) = U (x, x0) L U 1(x, x0 ). (2.16)
Since the spin angular momentum of a free scalar field is zero, the equations
d L
L = 0 =0 L =0 (2.17)
dx0
express equivalent forms of the conservation law of angular momentum. As the treatment
of the orbital angular momentum for a neutral and charged scalar fields is quite similar, we
shall present a unified consideration of the both cases in Sect. 13.
A simple, but important, conclusion from (2.8) is that the operator
1
M 2 : 0
[[0 , P ] , P ] (2.18)
c2
should be interpreted as square-of-mass operator of the field in momentum (and hence in
any) picture. This does not contradict to the accepted opinion that the square-of-mass oper-
ator is equal to the Lorentz square of the (divided by c) momentum operator. The problem
here is in what is called a momentum operator and in what picture the considerations are
done. Indeed, in Heisenberg picture, we can write
1 1 (i)2
M 2 ( ) = [[ , P ] , P ] = [i , P ] = ( )( (x))
c2 c2 c2
()2 1 1
= 2 (x) = 2 PQM P QM ( (x)) = = 2 [[ , Pt ] , P t ] , (2.19)
c c c
where [27] PQM = i and Pt = PQM + p idF , with p = const and idF being the iden-
tity mapping of F , are respectively the quantum mechanical and translational momentum
operators. So, we see that the conventional identification of M 2 with the square of (divided
by c) momentum operator corresponds to the identification of the last operator with PQM
(or of its square with 2 ). For details on the last item the reader is referred to [27]. It
should be noted, the eigenvalues of the operator (2.18) on the solutions of (2.8) characterize
the field (or its particles), while the eigenvalues of c12 P P on the same solutions are
characteristics of the particular states these solutions represent.
3. Analysis of the Klein-Gordon Equation

Our next aim is to find, if possible the (general) explicit form of the (constant) field oper-
ator 0 . The Klein-Gordon equation (2.8) is not enough for the purpose due to the simple
fact that the (canonical) momentum operator P depends on 0 . To show this, recall the
definition (I.2.5) of P and the expression (I.6.6) for it through the energy-momentum op-
erators T and T, given in Heisenberg and momentum pictures, respectively, by (2.11)
and (2.12). Therefore (2.8), (I.6.6), (2.12), and the explicit relation (I.5.1) form a closed
algebraic-functional system of equations for determination of 0 (and P ).
At the beginning of our analysis of the equations defining 0 , we notice the evident
solution
[0 , P ] = 0 for m = 0 (3.1)
of (2.8) which, by virtue of the Heisenberg equations/relations
(x)
[ (x), P ] = i (3.2)
x
leads to (x) = (x0 ) = const for m = 0. 1, 2 Consequently (in the zero mass case) our
system of equations admits a, generally non-vanishing, solution (see also (I.6.4))
(x) = (x0 ) = const = 0 P = P = 0 for m = 0. (3.3)
Evidently, (3.1) implies

T = 0 L = 0, (3.4)
due to (2.12) and (2.14), and, consequently, the dynamical variables for the solutions (3.1)
vanish, i.e.
P = 0 L = 0. (3.5)
For the general structure of the solutions of (2.8) is valid the following result.
Proposition 3.1. Every common solution 0 of the Klein-Gordon equation (2.8) and the
equations (I.5.30) (with i = ) is of the form

0 = d3 k f+(k)0 (k) + f (k)0 (k) (3.6)
k0 =+ m2 c2 +k2 k0 = m2 c2 +k2
1 For instance, such solutions are 0 = (x) = idF for C.

2 Notice, the scalar field Lagrangian identically vanishes in this case, L = 0, due to (2.3) and (3.1).
which can, equivalently, be rewritten as

0 = d4 k(k2 m2 c2 ) f (k)0 (k). (3.7)
Here: k = (k0 , k1 , k2 , k3 ) is a 4-vector with dimension of 4-momentum, k2 = k k =

k02 k12 k22 k32 = k02 k2 with k being the components of k and k := (k1 , k2 , k3 ) =
(k1 , k2 , k3 ) being the 3-dimensional part of k, () is the (1-dimensional) Dirac delta
function, 0 (k) : F F is a solution of the equation
[0 (k), P ] = k 0 (k), (3.8)
f are complex-valued functions (resp. distributions (generalized functions)) of k for solu-

tions different from (3.3) (resp. for the solutions (3.3)), and f is a complex-valued function
(resp. distribution) of k for solutions
different from (3.3)(resp. for the solutions (3.3)). Be-
sides f (k)|
2 =2 m c + k2 f (k) for solutions different from (3.3).
2 2
k0= 2 2 m c +k
Remark 3.1. In fact, in (3.6) enter only those solutions of (3.8) for which
k2 := k k = k02 k 2 = m2 c2 . (3.9)
Besides, a non-vanishing solution of (3.8) is a solution of (2.8) iff the condition (3.9)
holds. (Proof: write (2.8) with 0 (k) for 0 and use (3.8) twice.)
Remark 3.2. Evidently, the l.h.s. of (3.8) vanishes for the solutions (3.3). Therefore, we
have
0 (x, 0) = 0 (x0 , 0) = const = 0 (0) P = P = 0 (3.10)
where
0 (x, k) := U 1 (x, x0 ) 0 (k) U (x, x0 ). (3.11)
In terms of (3.6), this solution is described by m = 0 and, for example, f (k) = ( 12

a)3 (k), a C or f (k) such that f (k)|k0=|k| = (1 2a)|k|3 (k), a C.
Remark 3.3. Since 0 is Hermitian (see (2.5)), we have

f (k)0 (k) = f(k)0 (k) . (3.12)
k0 = m2 c2 +k2 k0 = m2 c2 +k2
due to (3.6). Also the similar relation
( f (k)0 (k)) = f (k)0 (k), (3.13)
holds as a corollary of (3.7).3

3The same result follows also from the below-written equations (3.19) and (3.21). Indeed, (3.19) and
= imply (k) = (k), which, in view of (3.21), entails (3.13).
0 0
Proof. Since the proposition was proved for the degenerate solutions (3.3) in remark 3.2,
below we shall suppose that (m, k) = (0, 0) (and, hence P = 0).
The equality (3.7) is equivalent to (3.6) (for solutions different from (3.3)) due to (y2
a ) = 2a
2 1
((y a) + (y + a)) for a > 0 [28, eq. 21.9-13].
Let 0 be given by (3.7). Using (3.8), we get

[0 , P ] = d4 kk (k2 m2 c2 ) f (k)0 (k). (3.14)
Inserting this into (2.8) and, again, applying (2.8), we see that (3.7) is a solution of (2.8).
So, it remains to be proved that any solution of (2.8) is of the form (3.7).
The equation (3.8) in Heisenberg picture reads (see (I.5.20))
[ (x, k), P ] = k (x, k). (3.15)
Combining this with (3.2) (with (x, k) for (x)), we find

x )k x )k
(x, k) = e i (x (x0 , k) = e i (x
1 1
0 0 0 (k) (3.16)
as
0 (k) = (x0 , k). (3.17)
Hence, due to (I.5.20), the equation (3.7) in Heisenberg picture reads

d4 k(k2 m2 c2 )e+i k x e i k x0 f (k)0 (k).
1 1
(x) = (3.18)
At the end, recalling that any solution (different from (3.3)) of the Klein-Gordon equa-
tion (2.9), satisfying the Heisenberg relations (I.5.30), admits a Fourier expansion of the
form [1, 4, 5]

1
(x) = d4 k(k2 m2 c2 )ei k x (k) (3.19)
for some operator-valued function (k) satisfying (3.8) under the condition (3.9), viz.4
0
[ (k), P ] = k (k) k2 = m2 c2 , (3.20)
the proof is completed by the identification

1 1
(k) = e i k x0 f (k)0 (k) = e i k x0 f (k) (x0 , k) (3.21)
and subsequent return from Heisenberg to momentum picture.

Meanwhile, we have proved two quite important results. On one hand, by virtue
of (3.21) and the homogeneous character of (3.8), any solution of (2.8) can be written as

0 = (k2 m2 c2 )0 (k) d4 k (3.22)
4 The first equality in (3.20) is the Fourier image of (I.5.30b).

where 0 (k) are appropriately normalized (scaled) solutions of (3.8) which solutions can
1
be identified, up to the phase factor e i x0 k , with the Fourier coefficients of (x), viz.
(see (3.19) and (3.21))
1
(k) = e i x0 k 0 (k). (3.23)
Therefore, by virtue of (3.13) and (3.21), the operators 0 (k) appearing in (3.22) satisfy
the relation (cf. (3.13))
(0 (k)) = 0 (k). (3.24)
On the other hand, this implies that the operator 0 (k)), entering in (3.22), is (up to a
constant) identical with the usual momentum representation of a scaler field (x) in Heisen-
berg picture [1, 4, 5]. Consequently, the operators (k) with k2 = m2 c2 , representing a
free Hermitian scalar field in momentum representation of Heisenberg picture of motion in
quantum field theory, are nothing else but (suitably normalized) solutions of (3.8) under the
condition (3.9), i.e. of the Klein-Gordon equation (2.8), which form a basis in the operator
space of all solutions of (2.8). In short, the momentum representation of a scalar field in
Heisenberg picture is a suitably chosen base for the solutions of the Klein-Gordon equation
in momentum picture. This result is quite important from two aspects. On one hand, it
reveals the real meaning of the momentum representation in Heisenberg picture of quan-
tum field theory. On another hand, it allows us to apply freely in momentum picture all of
the already established results concerning the Fourier images of the field operators, observ-
ables and other operators in Heisenberg picture. In particular, this concerns the frequency
decompositions and creation and annihilation operators theory.
4. Frequency Decompositions and Their Physical

Meaning
Since the decomposition (3.6) (or (3.7) if the solutions (3.3) are excluded) is similar to
the one leading to the frequency decompositions (in Heisenberg picture for free fields) in
quantum field theory, we shall introduce similar (in fact identical) notation. Defining1

f (k)0 (k) for k0 0
0 (k) = , (4.1)
0 for k0 < 0
we see that
0 = +
0 + 0 (4.2)

d3 k
0 = 0 (k)| k0 = m2 c2 +k2
(4.3)

[
0 (k), P ] = k 0 (k) k0 = m2 c2 + k2 (4.4)
due to (3.6) and (3.8). Notice, since 0 and 0 (k) satisfy (2.5), (3.13) and (3.12), we have
(
0 (k)) = 0 (k).

(4.5)
1 Cf. (I.7.7) and notice, we do not exclude the case k0 = 0 as it is done in the literature.
The equation (4.4) implies the interpretation of a free scalar field in terms of particles.
Indeed, if X p is a state vector of a state with 4-momentum p, i.e.
P ( X p ) = p X p , (4.6)
then, applying (4.4), we obtain

P (
0 (k)( X p )) = (p k )0 (k)( X p ) k0 = m2 c2 + k 2 (4.7)

So, + 0 (k) (resp. 0 (k)) can be interpreted as an operator creating (resp.
annihilating)
a particle (quant of the field) with mass m and 4-momentum k with k0 = m2 c2 + k 2 , i.e.
a particle with energy m2 c2 + k 2 and 3-momentum k. In this situation, the vacuum for a
scalar field should be a state X0 with vanishing 4-momentum and such that

0 (k)( X 0 ) = 0 k0 = m2 c2 + k2 . (4.8)
The action of the operators +0 (k) on X0 produces one- or multiple-particle states. More
details concerning the definition of a vacuum and construction of state vectors from it will
be given in sections 7 and 8.
For a comparison with existing literature (see, e.g., [1, 4, 5, 11]), we notice that the
creation/annihilation operators , introduced in momentum representation of Heisenberg
picture of motion, are defined by

(k) for k0 0
(k) = (4.9)
0 for k0 < 0
where (k) is the Fourier image of (x) (see (3.19)). Therefore, by virtue of (3.21) and
proposition 3.1, we have the following connection between the creation/annihilation opera-
tors in Heisenberg and momentum picture:
1

(k) = e i x0 k
0 (k) k0 = m2 c2 + k 2 . (4.10)
Exercise 4.1. Prove that the general formula (I.5.20) implies

k

(k) = e i x U (x, x0) 1
1
0 (k) U (x, x0 ) k0 = m2 c2 + k 2 ,

(4.11)
which, in view of (I.5.30b) and (4.4), is equivalent to (4.10). (Hint: apply the Fourier
decomposition from Sect. 3, then show that x (k) = 0 and, at last, set x = x0 in (4.11).)
5. The Klein-Gordon Equations in Terms of

Creation and Annihilation Operators
Let us return now to our main problem: to be found, if possible, the explicit form of the
field operator 0 (and momentum operator P ).
At first, we shall express P in terms of the creation and annihilation operators 0.

Regardless of the fact that the result is known (see, e.g., [1, eq. (3.26)] or [4, eq. (12.11)]),
we shall reestablish it in a completely new way.
Since (4.2)(4.4) entail

[0 , P ] = d3 k k (+
0 (k) + 0 (k)) | k0 = m2 c2 +k2
,
from (2.12), after some algebra (involving (4.2) and (4.3)), we get the energy-momentum
operator as
1
T = c2 d3 k d3 k (k k + k k )(+
0 (k) + 0 (k))
2
1
(+ 2 2 + +
0 (k ) + 0 (k )) + m c (0 (k) + 0 (k))(0 (k ) + 0 (k )) (5.1a)
2

1 2
T = c d3k d3 k (k k k k + k k )(+
0 (k) + 0 (k))
2

(+ 2 2 + +
0 (k ) + 0 (k )) + m c (0 (k) + 0 (k))(0 (k ) + 0 (k )) , (5.1b)

where k0 = m2 c2 + k 2 , k0 = m2 c2 + k and the two expressions for T correspond to
2
the two its versions in (2.12).

The idea is now the last result to be inserted into (I.6.6) and to commute 0 (k) and

0 (k ) with U (x, x0 ) in order to move 0 (k) and 0 (k ) to the right of U (x, x0 ). Rewrit-
ing (3.8) as 0 (k) P = ( P k idF ) 0 (k), by induction, we derive
0 (k) ( P1 Pn ) = {( P1 k1 idF ) ( Pn kn idF )} 0 (k) (5.2)
for any n N, which, in view of the expansion of the r.h.s. of (I.5.1) into a power series,
implies
x )k
0 (k) U (x, x0) = e i (x
1
0 U (x, x0) 0 (k). (5.3)
So, we have (see (4.1))

x )(k )
i (x
U (x, x0)
1
0 (k) U (x, x0 ) = e 0 (k)

0 (5.4a)

(x x0 )(k k )
+ i
U (x, x0)+
1
0 (k) 0 (k ) U (x, x0 ) = e 0 (k) 0 (k ) (5.4b)

(x x0 )(k k )
i
U (x, x0)
1
0 (k) 0 (k ) U (x, x0 ) = e 0 (k) 0 (k ). (5.4c)
At the end, substituting (5.1) into (I.6.6), applying the derived commutation rules
(5.4), performing the integration over x, which yields -functions like (k k ), and the
integration over k , we finally get after, a simple, but lengthy and tedious, calculation, the
following result

1
P = k | {+ +
0 (k) 0 (k) + 0 (k) 0 (k)} d k,
3
(5.5)
2 k0 = m2 c2 +k2
where we have introduced the 3-dimensional renormalized creation and annihilation oper-
ators

0 (k) := (2c(2) k0 )
3
0 (k) 2 2 2 .
1/2
(5.6)
k0 = m c +k
Notice, the result (5.5) is independent of from what form of T , (2.12a) or (2.12b),
we have started. The only difference of our derivation of (5.5) from similar one in the
literature (see, e.g., [1, sec. 3.2]) is that we have not exclude the massless case and the
degenerate solution (3.3) from our considerations. The operator +
0 (k) (resp. 0 (k)) is
called the creation (resp. annihilation) operator (of the neutral scalar field).
We would like to emphasize on the relations
(
0 (k)) = 0 (k)

(5.7)
which follow from (5.6) and (4.5). Actually these equalities are equivalent to the supposi-
tion that 0 is Hermitian field operator.
As a result of (4.10), the operators (k), corresponding to (5.5) in (the momentum
representation of) Heisenberg picture of motion, are [1, Sec. 3.2]
1/2
(k) : = c(2)3 (2k0 )1/2 (k) 2 2 2
k0 = m c +k
1
(5.8)
=e i x0 k
(k).
2 0
k0 = m c +k
2 2
Similarly, we have also the connection (see (4.11))

1
(k) = e i x k 2 2 2 U 1 (x, x0 )
0 (k) U (x, x0 ). (5.9)
k0 = m c +k
Consequently, the integrand in (5.5) and similar ones which will be met further in this
work, look identically in terms of
0 (k) and (k).
Summarizing the above results, we are ready to state and analyze the following problem.
Problem 5.1. Let

0 = +
(+
0 + 0 = 0 (k) + 0 (k))|k0= m2 c2 +k2 d k
3
1/2 (5.10)
= 2c(2)3 m2 c2 + k2 (+
0 (k) + 0 (k)) d k
3
be a solution of the Klein-Gordon equation (2.8). Find the general explicit form of the
operators
0 (k) which are solutions of the equations (see (4.4))

[
0 (k), P ] = k
0

(k) k0 = m2 c2 + k2 (5.11)
where P is given by (5.5).
Inserting the

representation (5.5) into (5.11)
and writing the expression

k 0 (k) as q 0 (k) (k q) d q with q0 = m2 c2 + q2 , we, after a simple
3 3
algebraic manipulation, obtain

q |q =m2 c2 +q2 [ + +
0 (k), 0 (q) 0 (q) + 0 (q) 0 (q)]
0
3
2
0 (k) (k q) d q = 0 (5.12)
3
Consequently
0 (k) must be solutions of
[ + 3
0 (k), [0 (q), 0 (q)]+ ] 20 (k) (k q) = f (k, q) (5.13)
where [A, B]+ := A B + B A is the anticommutator of operators (or matrices) A and B and
f (k, q) are operator-valued (generalized) functions such that

q |q =m2 c2 +q2 f (k, q) d3 q = 0. (5.14)
0
Looking over the derivation of (5.13), we see its equivalence with the initial system
of Klein-Gordon equation (2.8) and Heisenberg relations (I.5.30b) in momentum picture.
Consequently, (5.13) is the system of field equations in terms of creation and annihilation
operators in momentum picture.
Exercise 5.1. As a simple test of our calculations, prove that the commutativity of the
components P of the momentum operator, expressed by (I.4.4) (see also (I.5.14)), is a
consequence of (5.5), (5.13), and (5.14).
6. Commutation Relations
The equations (5.13) are the corner stone of the famous (anti)commutation relations in
quantum field theory for the considered here free Hermitian scalar field.1 The equa-
tions (5.13) form a system of equations relative to the functions 0 (k). Generally, it has
infinitely many solutions, but, at present, only a selected class of them has a suitable physi-
cal meaning and interpretation. This class will be described a little below.
Since for the physics is essential only the restriction of 0 (k) on the physically real-
izable states, not on the whole systems Hilbert space F , we shall analyze (5.13) in this
case, following the leading idea in similar cases in, e.g., [1, subsec. 10.1], or [4, 70],
or [5, p. 65]. It consists in admitting that the commutator or anticommutator of the creation
and/or annihilation operators (for free fields) is a c-number, i.e. it is proportional to the
identity mapping idF of the systems Hilbert space F of states. In particular, in our case,
this additional hypothesis takes the form of
Assumption 6.1. The operators

0 satisfy the relations
[
0 (k), 0 (q)] = (k, q) idF [
0 (k), 0 (q)] = (k, q) idF (6.1)
where = , [A, B] := A B B A, and

and are complex-valued (generalized)
functions.
1
The idea for arbitrary (Hermitian/real or non-Hermitian/complex, free or interacting) fields remains the
same: one should derive field equations in momentum picture and to solve them relative to the field operators
by using the explicit expression of the (canonical) momentum operator through the field operators.
It should clearly be understood, the conditions (6.1) are additional to the Lagrangian
formalism and do not follow from it. In fact, they or the below written commutation rela-
tions (6.28) (which under some conditions follow from (6.1)) have to be postulated.
Defining
[A, B] = A B + B A C (6.2)
and applying the identity
[A, B C] = [A, B] C + B [A,C] = 1 (6.3)
for = 1, viz.
[A, B C] = [A, B] C B [A,C] , (6.4)
due to = 1 and [A, B]1 = [A, B], we rewrite (5.13) as
[ + +
0 (k), 0 (q)] 0 (q) 0 (q)[0 (k), 0 (q)]
+[ + +
0 (k), 0 (q)] 0 (q) 0 (q)[0 (k), 0 (q)] (6.5)
2
0 (k) (k q)
3
= f (k, q).
Now, substituting the additional conditions (6.1) into (6.5), we get
(1 ) +
0 (q) [0 (k), 0 (q)]
+ (1 )+
0 (q) [0 (k), 0 (q)] 20 (k) (k q) = f (k, q). (6.6)
3
For = +1, this equality reduces to 2

0 (k) (k q) = f (k, q) which, when inserted
3
into (5.14), entails
k | (k) =0 (6.7)
k0 = m2 c2 +k2 0
for any = 0, 1, 2, 3 and k. Consequently, the choice = +1, i.e.
[
0 (k), 0 (q)]+ = + (k, q) idF [
0 (k), 0 (q)]+ = + (k, q) idF , (6.8)
leads, in view of (5.5) and (6.7), to
P = 0 (6.9)
which, by virtue of Klein-Gordon equation (2.8), implies
m2 c2 0 = 0. (6.10)
So, if m = 0, we get the solution (3.3) and, if m = 0, we derive
0 = 0 for m = 0 (6.11)
and, consequently,

0 (k) 0 for m = 0 (6.12)
We interpret the obtained solution (6.11) (or (6.12)) of the Klein-Gordon equation (2.8)
as a complete absence of the physical scalar field.2
Consider now (6.6) with = 1. Writing it explicitly for the upper, +, and lower,
, signs, we see that (6.6) is equivalent to

0 (q) [0 (k), 0 (q)] + [0 (k), 0 (q)] 0 (q)
1

0 (q) (k q) =
3
f (k, q) (6.13)
2
where, in accord with (6.1), the commutators are
[
0 (k), 0 (q)] = (k, q) idF [
0 (k), 0 (q)] = (k, q) idF . (6.14)
Here, for brevity, we have omit the subscript , i.e. we write and for
and
respectively.
The properties of and can be specified as follows.3
Let X p be a state vector with fixed 4-momentum p (see (4.6)). Defining

Xk,q := (
0 (k) 0 (q))( X p )

Yk,q := (
0 (k) 0 (q))( X p ), (6.15)
we, applying (4.7), get

P( Xk,q ) = (p k q ) Xk,q P( Yk,q ) = (p k q ) Yk,q . (6.16)
Noticing that

Xk,q + Xq,k = [
0 (k), 0 (q)] ( X p )

Yk,q
+ Yq,k = [
0 (k), 0 (q)] ( X p )
for = 1, we see that

P [
0 (k), 0 (q)] ( X p ) = (p k q )[0 (k), 0 (q)] ( X p )
(6.17)
P [
0 (k), 0 (q)] ( X p ) = (p k q )[0 (k), 0 (q)] ( X p ).
It is worth to mention, in the derivation of (6.17) no additional hypothesis, like (6.1),

have been used.
Applying (6.17) for = 1 and imposing the additional conditions (6.14), we, due
to (4.6), find
(k + q) (k, q) = 0 (k q)(k, q) = 0 (6.18)

m2 c2 + k2 + m2 c2 + q2 (k, q) = 0
(6.19)
m2 c2 + k2 m2 c2 + q2 (k, q) = 0.
2
Since, for a scalar field, the solutions (3.3) do not lead to any physically predictable results, they, in the
massless case, should also be interpreted as absence of the field.
3 For the initial idea, see [1, subsec. 10.1], where the authors (premeditated or not?) make a number of
implicit assumptions which reduce the generality of the possible (anti)commutation relations.
Here, with necessity, and must be regarded as distributions (generalized func-

tions) as otherwise the equations (6.18) have only the trivial solutions
(k, q) = 0 for k + q = 0 (k, q) = 0 for k q = 0 (6.20)
which, as it can easily be seen, reproduce the trivial solution (6.11) of the Klein-Gordon
equation. Rewriting (6.18) as q (k, q) = k (k, q) and q(k, q) = k (k, q), we see
that
f (q)(k, q) = f (k)(k, q) f (q) (k, q) = f (k)(k, q) (6.21)
for a function f (k) which is supposed to be polynomial or convergent power series.4 As a
result of (6.21), the second equation in (6.19) is equivalent to the identity 0 = 0, while the
first one reduces to
m2 c2 + k 2 (k, q) = 0. (6.22)
Inserting (6.14) into (6.13), we find
1
f (k, q) =
0 (q) (k, q) + 0 (q)( (k, q) (k q)).
3
(6.23)
2
At the end, besides (6.22), the condition (5.14) is the
only one remaining to be satisfied.
Substituting (6.23) into it and using (6.21) for f (q) = m2 c2 + q2 in the case = 0 and for
f (q) = qa in the one with = a = 1, 2, 3, we obtain
3
0 = ka (k, q)
0 (q) + ( (k, q) (k q))0 (q) d q
3
(6.24a)
3
0 = m c +k
2 2 2
+ (k, q) 3
0 (q) + ( (k, q) (k q))0 (q) d q, (6.24b)
where, in the second equation, the term containing vanishes due to (6.22). Since these
equations must be valid for arbitrary k, the integrals in them should vanish if k = 0. Forming
the sum and difference of these integrals in that case, we get

(k, q)
0 (q) d q = 0
3
(6.25a)

((k, q) 3 (k q))
0 (q) d q = 0.
3
(6.25b)
If (m, k) = (0, 0), the standard (Bose-Einstein) commutation relations are extracted from
equations (6.24) (or (6.25) if k = 0) by imposing the following second, after (6.1), addi-
tional condition.
Assumption 6.2a. The equation (6.24) must be valid for arbitrary

0 (q) whose commuta-
tors are c-numbers.
4 We cannot write, e.g., (k,q) = const 3 (k + q) as the equation yg(y) = 0, y R, has a solution
d(y)
g(y) = const (y), but this is not its general solution; e.g., its solutions are g(y) = 0 (y) + 1 y dy +
d2 (y)
2 y2 dy2
+ with 0 ,1 ,... being constant numbers.
Then (6.24) entail
(k, q) = 0 (k, q) = 3 (k q) (6.26)
for k = 0, which convert (6.22) into identity.

Let us say also a few words on the special case when m = 0 and k = 0. In it one
cannot extract any information from (6.24) and (6.22). Recall, if m = 0, the operators
0 (0) describe creation/annihilation of massless particles with zero 4-momentum. Since,
in the Hermitian scalar case, the 4-momentum and mass are the only characteristics of the
fields quanta, the particles corresponding to

0 (k) with m = 0 and k = 0 (6.27)
are unphysical as they cannot lead to some observable consequences. The operators (6.27)
are special kind of the solutions (3.3) of the Klein-Gordon equation and, consequently, can
be interpreted as absence of the field under consideration. Besides, as we proved above
under the hypothesis (6.14), the only restrictions to which 0 (0) with m = 0 should be
subjected are the conditions (6.14) with k = 0, arbitrary q, and any (generalized) functions
and . Thus, to ensure a continuous limit when (m, k) (0, 0), we, by convention,
accept
Assumption 6.2b. The functions (0, q) and (0, q) are given by (6.26) with k = 0.5
So, we have obtained the next solution of the problem 5.1. In momentum picture, the
Klein-Gordon equation (2.8) admits a solution (5.10), i.e.
1/2
0 = 2c(2)3 m2 c2 + k2 (+
0 (k) + 0 (k)) d k,
3
in which the creation/annihilation operators

0 (k) satisfy only the commutation relations
[
0 (k), 0 (q)] = 0 [
0 (k), 0 (q)] = (k q) idF .
3
(6.28)
Exercise 6.1. Verify that, in view of (5.5) and (6.28), the equality (5.11) is identically valid.
We would like to remark, in the above considerations the massless case, i.e. m = 0, is
not neglected. However, obviously, the commutation relations (6.28) exclude the degenerate
solution (6.11) or, equivalently, (6.20) or (6.12).
More generally, one can look for solutions of the relation (5.13), under the condi-
tion (5.14) which do not satisfy the additional conditions (6.1). But this is out of the aims
of this work.6
5 In this way we exclude from the theory a special kind of absent (unphysical) field described by m = 0
and 0 = 0 or 0 (k) = 0.
6 Special types of tri-linear relations, like (5.13), are known as paracommutation relations and were discov-
ered in [29] (See also [30, 31]). However, it seems that at present are not indications that solutions of (5.13),
which do not satisfy (6.28), may describe actually existing physical objects or phenomena [10, 32, 33]. This
is one of the reasons that the quantum field theory deals with (6.28) instead of (5.13) (or the equivalent to it
Klein-Gordon equation (2.8) (in momentum picture) or (2.9) (in Heisenberg picture). In chapter V we shall
demonstrate how the parabose-commutation relations for a free scalar field can be derived from (5.13).
As we have noted several times above, the concepts of a distribution (generalized func-
tion) and operator-valued distribution appear during the derivation of the commutation re-
lations (6.28). We first met them in the relations (5.13). In particular, the canonical com-
mutation relations (6.28) have a sense iff [
0 (k), 0 (k)] is an operator-valued distribution
(proportional to idF ), which is not the case if the field 0 (or 0 ) is an ordinary opera-
tor acting on F . These facts point to inherent contradiction of quantum field theory if the
field variables are considered as operators acting on a Hilbert space. If one traces back this
contradiction, the result is unexpected: it is due to the fact that we have added the Heisen-
berg equation (I.5.30b) to the Lagrangian formalism: 7 all types of commutation relations
originate from (5.11), which is a version of (I.5.30b), and the representation (5.5) in the
derivation of which equation (I.5.30b) was essentially used. The rigorous mathematical set-
ting requires the fields variables to be regarded as operator-valued distributions. However,
such a setting is out of the scope of the present work and the reader is referred to books
like [9, 10, 34, 35] for more details and realization of that program. In what follows, the
distribution character of the quantum fields will be encoded in the Diracs delta function,
which will appear in relations like (5.13) and (6.28).
Ending the discussion of the commutation relations for a free Hermitian scalar field,
we would like to note that the commutation, not anticommutation, relations for it were
derived directly from the Klein-Gordon equation (2.8) without involving the spin-statistics
theorem, as it is done everywhere in the literature [1, 5]. In fact, this theorem is practically
derived here in the special case under consideration. Besides, we saw that the commutation
relations can be regarded as additional restrictions, postulated for the field operators as
stated, e.g., in [5, pp. 5960], which must be compatible with the equations of motion. In
fact, as we proved, these relations are, under some assumptions, equivalent to the equations
of motion, i.e. to the Klein-Gordon equation in our case. Said differently, the commutation
relations convert the field equation(s) into identity (identities) and, in this sense are their
solutions. An alternative viewpoint is the commutation relations (6.28) to be considered
as field equations (under the conditions specified above) with respect to the creation and
annihilation operators as field operators (variables).
To close this section, we have to make the general remark that the tri-linear rela-
tions (5.13) (together with (5.14)) are equivalent to the initial Klein-Gordon equation (in
terms of creation/annihilation operators) and all efforts for the establishment of the com-
mutation relations (6.28) reflect, first of all, the fact of extraction of physically essential so-
lutions of these equations.8 In this sense, we can say that the commutation relations (6.28)
are a reduction of the initial Klein-Gordon equation (2.8) (in momentum picture) or (2.9)
(in Heisenberg picture), under the the assumptions 6.1, 6.2a, and 6.2b.
7 The relation (I.5.30a) expresses the simultaneous measurability of the components of the momentum
operator and hence it is physically motivated. At the same time, the Heisenberg equation (I.5.30b) has a pure
geometrical origin.
8 One may recognize in (5.13) a kind of paracommutation relations which are typical for the so-called
parastatistics [2931, 36].

7. Vacuum and Normal Ordering

The vacuum of a free Hermitian scalar field 0 is a particular its state, described by a state
vector X0 which carries no energy-momentum and, correspondingly, it is characterized by
a constant (in spacetime) state vector, i.e.
P( X0(x)) = 0 (7.1)
X0(x) = X0(x0 ) = X0 (7.2)
where X0 and X0 are the vacua in Heisenberg and momentum pictures respectively. Equa-
tion (7.1) can be taken as a macroscopic definition of the vacuum state vector X0 . Since,
microscopically, the field is considered as a collection of particles (see Sect. 4), the vac-
uum should be considered as a vector characterizing a state with no particles in it. Recalling
the interpretation of creation, +
0 (k), and annihilation, 0 (k), operators from Sect. 4, we
can make the definition (7.1) more precise by demanding

0 (k)( X0 ) = 0 X0 = 0. (7.3)
This is the everywhere accepted definition of a vacuum for a free Hermitian scalar field.
However, it does not agree with the expression (5.5) for the momentum operator and the
commutation relations (6.28). In fact, commuting +
0 (k) and 0 (k), according to (6.28), in
the second term in the integrand in (5.5) and using (7.3), we get

1
P ( X0 ) = k 3 (0) d3 k X0
2
4 for = 0
= X0 3 (0) . (7.4)
( ) for = 1, 2, 3
2 2 2
Of course, this is a nonsense which must be corrected if one wants to exist a vector
X0 with the properties (7.2) and (7.3). The problem can be solved by repairing the r.h.s.
of (5.5), by replacing the commutation relations (6.28) with other relations (compatible
with (5.13) and (5.14)), or by some combination of these possibilities. At this point, we
agree with the established procedure for removing (7.4) from the theory. If one accepts
not to change the logical structure of the theory, the only possibility is a change in the
Lagrangian from which all follows. Since in (7.4) the infinities come from the term 0 (k)
+0 (k) in (5.5), it should be eliminated somehow. The known and, it seems, well working
procedure for doing this, which we accept, is the following one. At first the Lagrangian
and the dynamical variables, obtained from it and containing the field 0 , should be written,
by means of (4.1)(4.3) in terms of creation, +
0 (k), and annihilation, 0 (k), coperators.
Then, any composition (product) of these operators, possibly appearing under some integral
sign(s), must be change so that all creation operators to stand to the left relative to all
annihilation operators.1 The described procedure for transforming compositions of creation
and annihilation operators is known as normal ordering and the result of it is called normal
1 The so-formulated rule is valid only for integer spin particles/fields. By virtue of (6.28), the order of the
different creation/annihilation operators relative to each other is insignificant, i.e. it does not influence the result
of the described procedure.
product of the corresponding operators. This item is discussed at length in the literature to
which the reader is referred for details [1, 4, 5, 37].
The mapping assigning to a composition of creation/annihilation operators their normal
product will be denoted by N and called normal ordering operator. The action of N
on polynomials or convergent power series of creation/annihilation operators is defined by
extending it by linearity;the resulting mapping
being denoted by N too.
Since, evidently, N 0 (k)
0 (k) = +
0 (k)
0 (k), the representation (5.5) of the
momentum operator changes, after normal ordering, into

P = ds kk | + (k)
k0 = m2 c2 +k2 0 0 (k). (7.5)
(Notice, after normal ordering, we retain the notation P for the object resulting from (5.5).
This is an everywhere accepted system of notation in the literature and it is applied to all
similar situations, e.g. for the Lagrangian L or energy-momentum operator T . After
some experience with such a system of doubling the meaning of the symbols, one finds it
useful and harmless.) Now the equality (7.1) is a trivial corollary of (7.5) and (7.3).
As X0 = 0, we shall assume that the vacuum X0 can be normalized to unity, viz.
X0 | X0 = 1, (7.6)
where | : F F C is the Hermitian scalar product of the Hilbert space F of states.

In fact, the value X0 | X0 is insignificant and its choice as the number 1 is of technical
character. In this way, in many calculations, disappears the coefficient X0 | X0 . Prima
facie one can loosen (7.6) by demanding X0 to have a finite norm, but this only adds to the
theory the insignificant constant X0 | X0 which can be eliminated by a rescaling of X0 .2
Let us summarize the above discussion.
Definition 7.1. The vacuum of a free scalar field is its physical state which does not con-
tains any particles and has zero 4-momentum and (total) angular momentum operator. It
is described by a state vector, denoted by X0 (in momentum picture) and also called the
vacuum of the field, such that:
X0 = 0 (7.7a)
X0 = X0 (7.7b)

0 (k)( X0 ) = 0 (7.7c)
X0 | X0 = 1 (7.7d)
for any 3-momentum k.
Since the existence of the vacuum X0 in the Heisenberg picture is a known theorem,
the condition (7.7b), expressing the coincidence of Heisenberg vacuum and momentum
vacuum, ensures the existence of the vacuum X0 in momentum picture. Besides, to make
the theory sensible, we have assumed normal ordering of the creation/annihilation operators
in the Lagrangian and all observables derived from it.
2 If X has an infinite norm, so is the situation with any other state vector obtained from X via action with
0 0
creation operators, which makes the theory almost useless.
The normal ordering of products changes the field equations (5.13) into
1
[ +
0 (k), 0 (q) 0 (q)] 0 (k) (k q) =
3
f (k, q) (7.8)
2
as the anticommutator in (5.13) originates from the expression (5.5) for the momentum
operator (before normal ordering). The conditions (5.14) remains unchanged. However, by
means of (6.4) with = 1 and the commutation relations (6.28), one can verify that (7.8)
and (5.14) are identically valid. This means that, in fact, the commutation relations (6.28)
play a role of field equations under the hypotheses made.
8. State Vectors and Transitions between Them

According to the general theory of chapter I, the general form of a state vector X (x) of a
free Hermitian scalar field in momentum picture is

X (x) = U (x, x0)( X (x0)) = e i (x x0 ) P ( X (x0)),
1
(8.1)
where x0 M is an arbitrarily fixed point in the Minkowski spacetime M, P is given

via (7.5), and the initial value X (x0 ) of X at x0 is identical with the (constant) state vector
X representing the same state as X (x) but in Heisenberg picture, i.e.
X (x0) = X . (8.2)
In particular, if X p is a state vector at x0 with fixed 4-momentum p , i.e.
P ( X p ) = p X p , (8.3)
then its general form at a point x is

X p(x) = e i (x x0 )p X p
1
(8.4)
and P ( X p (x)) = p X p (x).

From the vacuum X0 can be constructed a base for the systems Hilbert space F , called
the Fock base. A general s-particle state containing s particles with momenta k1 . . ., ks (some
of these vectors may be identical) of a scalar field 0 has the form

X (k1, . . ., ks ) = fs(k1 , . . ., ks ) + +
0 (k1 ) 0 (ks ) ( X0 ) (8.5)
for some function f s characterizing the distribution of the particles.1 Notice, by virtue (3.9),
the 4-momenta k1 , . . ., ks are subjected to the conditions ka2 = m2 c2 , a = 1, . . ., s, with m be-
ing the mass of the quanta of the field 0 . An arbitrary state, described via the Klein-Gordon
equation (2.8), can be presented as a superposition of all possible states like (8.5), viz.

X= dk1 . . . dks fs(k1 , . . ., ks ) + +
0 (k1 ) 0 (ks ) ( X0 ). (8.6)
s0
1 In (8.5) we have omit a spacetime dependent factor which the reader may recover, using (8.4); the vec-
tor (8.5) corresponds to the vector X p in (8.4) in a case of s-particle state.

The above results, concerning free Hermitian scalar field, are identical with similar ones
in the momentum representation in ordinary quantum field theory (in Heisenberg picture);
the difference being that now 0 (k) is the field operator in the momentum picture, which, as
we proved, is identical with the Fourier image of the field operator in Heisenberg picture.
So, the Fock base goes without changes from Heisenberg into momentum picture.
One of the main problems in quantum field theory is to find the amplitude for a transition
from some initial state Xi (xi ) into a final state X f (x f ), i.e. the quantities
S f i (x f , xi ) := X f (x f )| Xi (xi ) (8.7)
called elements of the so-called S-matrix (scattering matrix). Ordinary one considers the

limits of (8.7) with x0f + and x0i or, more generally, x f + and xi .
These cases are important in the scattering theory but not for the general theory, described
here, and, respectively, will not be discussed in our work. If we know X f and Xi at some
(0) (0)
points x f and xi , respectively, then, combining (8.1) and (8.7), we get

(0) (0)
S f i (x f , xi ) = X f (x f )|e i (xi x f ) P ( Xi (xi ))
1
(0) (0)
= X f (x f )| U (xi , x f )( Xi(xi )). (8.8)
Consequently (cf. [2, 107]), the operator ( U (x f , xi )) = U 1 (x f , xi ) = U (xi, x f ),

where means (Hermitian) conjugation, has to be identified with the S-operator (often
called also S-matrix). To continue the analogy with the S-matrix theory, we can expand the
exponent in (8.8) into a power series. This yields (see (7.5))

U (xi, x f ) = idF + U (n)(xi , x f ) (8.9)
n=1
1 1
U (n)(xi , x f ) := (xi 1 x f 1 ) . . .(xi n x f n )
n! (i) n

(1) (n)
d3 k(1). . . d3 k(n) k1 kn + (1) (1) + (n) (n)
0 (k )0 (k ) 0 (k ) 0 (k ) (8.10)

(a)
where k0 = m2 c2 + (k(a) )2 , a = 1, . . ., n. This expression is extremely useful in scatter-
ing theory when one deals with states having a fixed number of particles.2
The first thing one notices, is that the vacuum, defined via (7.7), cannot be changed, viz.
U (xi , x f )( X0) X0, (8.11)
and if X1 (x) is a state vector of a state containing at least one particle, then
X1 (x)| X0 0 (8.12)
2 Since we are dealing with a free field, there is no interaction between its quanta (particles) and, hence,
there is no real scattering. However, the method, we present below, is of quite general nature and can be applied
in real scattering problems.
This simple result means that the only non-forbidden transition from the vacuum is into
itself, i.e.
X0 | X0 = 1 = 0. (8.13)
The results just obtained are known as the stability of the vacuum.
In accord with (8.4), (7.5) and (6.28), the vector

X (x, p) = e i (x x0 )p +
1
0 (p)( X0 ) (8.14)
describes a particle at a point x M with 4-momentum p. Similarly, an n-particle state, n

1, in which the ith particle is at a point xi with 4-momentum pi , i = 1, . . ., n, is represented
by3
X (x1, p1 ; . . .; xn , pn)
1 1 n
= exp
n!

i i=1
(xi x0 )(pi ) (+ +
0 (p1 ) 0 (pn ))( X0 ). (8.15)
Exercise 8.1. By applying (6.28), (5.7) and (7.7c), show that the transition amplitude from
m-particle state into n-particle state, m, n N is
1 1 n
X (y1 , q1 ; . . .; yn , qn )| X (x1 , p1 ; . . .; xm , pm ) =
n!
mn exp
i i=1

(xi yi )(pi )
3 (pn qi1 )3 (pn1 qi2 ) . . .3 (p1 qin ), (8.16)

(i1 ,...,in )
where mn is the Kronecker -symbol, i.e. mn = 1 for m = n and mn = 0 for m = n, and

the summation is over all permutations (i1 , . . ., in ) of (1, . . ., n).
The presence of mn in (8.16) means that an n-particle state can be transformed only into
an n-particle state; all other transitions are forbidden. Besides, the -functions in (8.16) say
that if the 4-momentum of a particle changes, the transition is also forbidden. So, the only
change an n-particle state can experience is the change in the coordinates of the particles
it contains. All these results are quite understandable (and trivial too) since we are dealing
with a free field whose quanta move completely independent of each other, without any
interactions between them.
we set m = n and integrate
Notice, if in (8.16) over all momenta, we get a pure phase

factor, equal to exp i i=1 (xi yi )(pi) . Since, in this case, the module of the square
1 n
of (8.16) is interpreted as a probability for the transition between the corresponding states,
this means that the transition between two n-particle states is completely sure, i.e. with
100% probability.
At the end of this section, we note that the states (8.15) are normalized to unity,

d3 q1 . . . d3 qn X (x1 , q1 ; . . .; xn , qn )| X (x1 , p1 ; . . .; xm , pm ) = 1, (8.17)
3 Since the vacuum and creation/annihilation operators in Heisenberg and momentum pictures coincide, we
use the usual Fock base to expand the state vectors. The spacetime depending factor comes from (8.4).
due to (7.7d). However, the norm X (x1 , p1 ; . . .; xn , pn )| X (x1 , p1 ; . . .; xm , pm ) is infinity as

it is proportional to (3 (0))n, due to (8.16). If one works with a vacuum not normalized to
unity, in (8.15) the factor X0 | X0 1/2 will appear. This will change (8.16) with the factor
X0 | X0 1 .
II.B. Arbitrary Scalar Fields
Until now the case of free Hermitian (neutral) scalar field was explored. In the present,
second, part of this chapter, the results obtained for such a field will be transferred to the
general case of free arbitrary, Hermitian (real, neutral, uncharged) or non-Hermitian (com-
plex, charged), scalar field.4
As we shall see, there are two essential peculiarities in the non-Hermitian case. On
one hand, the field operator and its Hermitian conjugate are so mixed in the momentum
operator that one cannot write (in momentum picture) separate field equations for them. On
the other hand, a non-Hermitian field carries a charge. These facts will later be reflected in
the corresponding commutation relations.
9. Description of Free Scalar Field

In Heisenberg picture a free arbitrary, Hermitian or non-Hermitian, scalar field is described
via a field operator (x), which may be Hermitian,
(x) = (x) (9.1a)
or non-Hermitian,
(x) = (x). (9.1b)
The properties of a free non-Hermitian scalar field are, usually [1, 4, 5, 38], encoded in
the Lagrangian
L := m2 c4 (x) (x) + c2 2( (x)) ( (x)), (9.2)
which in momentum picture, according to the general rules of chapter. I (see equa-
tion (I.6.2)), reads
L = m2c4 0 0 c2 [0, P] [0, P ] , (9.3)

4 A classical complex field, after quantization, becomes a non-Hermitian operator acting on the systems
(fields) Hilbert space of states. That is why the quantum analogue of a classical complex scalar field is called
non-Hermitian scalar field. However, it is an accepted common practice such a field to be called (also) a
complex scalar field. In this sense, the terms complex scalar field and non-Hermitian scalar field are equivalent
and, hence, interchangeable. Besides, since a complex (classical or quantum) scalar field carries a charge, it is
called also charged scalar field. Cf. footnote 1 on page 25.
where (see (I.5.22) and cf. (I.6.4))
(x) = U (x, x0 ) (x) U 1 (x, x0 ) = (x0 ) = (x0 ) =: 0 (9.4a)

1
(x) = U (x, x0 ) (x) U

(x, x0 ) = (x0 ) = (x0 ) =:

0 (9.4b)
are the corresponding to (x) and (x), respectively, constant field operators in momen-
tum picture. In general, the operator 0 can be Hermitian,
0 (x) = 0 (9.5a)
or non-Hermitian,
0 = 0 . (9.5b)
as a result of (9.1) and the unitarity of U (x, x0 ).

Defining

1 for = (Hermitian (neutral) field)
( ) := , (9.6)
0 for = (non-Hermitian (charged) field)
we can unify the Lagrangians (9.2) and (2.3) by writing
1 1
L := m2 c4 (x) (x) + c2 2 ( (x)) ( (x)). (9.7)
1 + ( ) 1 + ( )
As an alternative to (9.7), one may also put
1 1
L := m2 c4 (x) (x) + c2 2 ( (x)) ( (x)) (9.8)
1 + ( ) 1 + ( )
There is also one more candidate for a Lagrangian for a free arbitrary scalar field. Since
such a field is equivalent to two independent free Hermitian scalar fields 1 = 1 and
2 = 2 with masses equal to the one of and such that = 1 + i 2 , i being the

imaginary unit, we can set L = L0 ( 1 ) + L0 ( 2 ) with L0 defined by the r.h.s. of (2.2).
Taking into account that = 1 i 2 , one can transform the last Lagrangian into the
form
1
L := m2 c4 (x) (x) + (x) (x)
2(1 + ( ))
(9.9)
1
+ c2 2 ( (x)) ( (x)) + ( (x)) ( (x)) .
2(1 + ( ))
This Lagrangian, which is the half of the sum of (9.7) and (9.8), also reduces to (2.2)
in the Hermitian case = (or 2 = 0 in terms of the Hermitian fields 1 and 2 ).
Evidently, the Lagrangian (9.9) is a symmetrization of the r.h.s. of (9.7) or (9.8) relative
to and with coefficient 12 . The advantage of (9.9) is that in it the field and its
Hermitian conjugate enter in a symmetric way, which cannot be said relative to (9.2),
(9.7) and (9.8).

Going some steps ahead, the consequences of the Lagrangians L , L , L , and L can
be summarized as follows: (i) All of these Lagrangians lead to identical (Klein-Gordon)
field equations for and ; (ii) The energy-momentum, momentum and charge operators
generated by these Lagrangians are, generally, different; (iii) After the establishment of the
commutation relations and a normal ordering of products (compositions), the momentum,

(orbital) angular momentum and charge operators generated by L , L , and L become
identical and equal to one half of the ones induced by L . Therefore, in view of these
assertions, all of the Lagrangians given above can be considered a posteriori as equivalent.1
In momentum picture, by virtue of (I.6.2), the Lagrangians (9.7), (9.8) and (9.9) are:
1 1
L = m2 c4 0 0 c2 [0 , P ] [0 , P ] (9.10)
1 + ( ) 1 + ( )
1 1
L= m2 c4 0 0 c2 [0 , P ] [0 , P ] (9.11)
1 + ( ) 1 + ( )
1
L = m2 c4 0 0 + 0 0
2(1 + ( ))
(9.12)
1
c2 [0 , P ] [0 , P ] + [0 , P ] [0 , P ] ,
2(1 + ( ))
where (cf. (9.6))

1 for 0 = 0 (Hermitian (neutral) field)
(0 ) := = ( ). (9.13)
0 for 0 = 0 (non-Hermitian (charged) field)
From here, we derive: 2

aL L aL L
= = m2 c4 0 = = m2 c4 0 (9.14)
0 0 0 0
aL L a
:= = ic2 [0 , P ] := = ic2 [0 , P ] (9.15)
y y
where a = , , , y := i

1
[0 , P ] and y := i
1
[0 , P ] = (y ) . At this point, all remarks,
made in Sect. 2 in a similar situation, are completely valid too. For a rigorous derivation of
the field equations (9.16) and energy-momentum tensors (9.21) below, the reader is referred
to [3] (see also section I.3).
By means of the above equalities, from (I.6.3), we get the field equations for 0 and 0 :
m2 c2 0 [[0 , P ] , P ] = 0 (9.16a)
m2 c2 0 [[0 , P ] , P ] = 0. (9.16b)
So, regardless of the Lagrangians (9.10)(9.12) one starts from, the fields 0 and 0
satisfy one and the same Klein-Gordon equation. However, in contrast to the Heisenberg
1 However, in Sect. 14, we shall see that the Lagrangian (9.9) carries more information than (9.7), (9.8)
and (9.2). In this sense, it is the best one.

2 Recall, we follow the differentiation rules described on page xii.
picture, in momentum picture these equations are not independent as the momentum op-
erator P , appearing in (9.16) and given via (I.2.5) or (I.6.6), also depends on 0 and 0
through the energy-momentum operator T. Hence, to determine 0 , 0 and P , we need
an explicit expression for T as a function of 0 and 0 .
If was a classical complex/real field, we would have
1
T = { ( ) + ( )} L
1 + ( )
where the means complex conjugation. The straightforward transferring of this expression
in the quantum case results in
1
T(1) = { ( ) + ( )} L . (9.17)
1 + ( )
However, if and and and do not commute, this T is non-Hermitian,

T
= T . This situation can be corrected by a Hermitian symmetrization of the first
two terms in (9.17), which gives
1
T(2) = ( ) + ( )
2(1 + ( ))

+ ( ) + ( ) L . (9.18)
Evidently, if = , and the Lagrangians (9.7) and (9.8) are employed, the equa-
tions (9.17) and (9.18) reduce to (2.11a) and (2.11b), respectively. But these are not the
only possibilities for the energy-momentum operator. Often (see, e.g., [1, eq. (3.34)], or [5,
eq. (2-151)], or [38, eq. (6)]), one writes it in the form
1
T(3) = { ( ) + ( )} L . (9.19)
1 + ( )
As an alternative, we may also choose
1
T(4) = { ( ) + ( )} L . (9.20)
1 + ( )
The last two expressions are suitable as and are proportional to and
respectively.
Partial discussion of the problem how should be defined the quantum energy-momen-
tum tensorial operator T and a similar one for the current (see below), the reader can find
in [38, 2], where a reference to an early work of B. W. Gordon in Z. Phys. (vol. 40, p. 117.
1926) is given.
From the view-point of symmetry, T = T , and Hermiticity T = T , the en-
ergy-momentum operators (9.18)(9.20) are indistinguishable (if one and the same La-
grangian is used in them).
So, if we want to explore all possibilities, we have to look for the consequences of the
four energy-momentum operators (9.17)(9.20) for any one of the three Lagrangians (9.7)
(9.9). However, instead of investigating these 12 cases, we shall study only 3 of them. The
reason is that in [3] we have proved that for a given Lagrangian, from the Schwingers action
principle [5, 39], follow, between other things, also unique expressions for all conserved
quantities in terms of the field operators. In particular, it implies unique energy-momentum
and current operators. These operators, for the Lagrangians (9.7)(9.9) will be pointed
below. Since the discussion of the way of their selection is out of the range of the present
work, the reader may: (i) take this by faith; (ii) consider it as a lucky choice; (iii) look on it
as an additional hypothesis/postulate; (iv) explore the consequences of the other 9 cases to
see that they lead to contradictions in the theory;3 (v) apply the results obtained in [3] and
quoted in section I.3.
Gordon equations (9.16) and prove that the correct expressions for the energy-momentum
tensor are the ones given below by (9.21) in Heisenberg picture and (9.22) in momentum
picture for the Lagrangians we consider. Similarly, obtain the equations (9.39) below for
the corresponding current operators following the recipes given in section I.3.
The correct energy-momentum operators for the Lagrangians (9.7)(9.9) are respec-
tively (9.19), (9.20) and (9.18).4 Explicitly, in view of (9.15), we have
(3) 1
T = c2 2 {( ) ( ) + ( ) ( )} L (9.21a)
1 + ( )
(4) 1
T = c2 2 {( ) ( ) + ( ) ( )} L (9.21b)
1 + ( )
(2) 1
T = c2 2 {( ) ( ) + ( ) ( )
2(1 + ( )) (9.21c)

+ ( ) ( ) + ( ) ( )} L .

In momentum picture, these operators read (see (9.10)(9.15)):

(3) 1
T = c2 {[0 , P ] [0 , P ] + [0 , P ] [0 , P ] }
1 + (0 )
(9.22a)
1
+ c2 {m2 c2 0 0 + [0 , P ] [0 , P ] }
1 + (0 )
(4) 1
T = c2 {[0 , P ] [0 , P ] + [0 , P ] [0 , P ] }
1 + (0 )
(9.22b)
1
+ c2 {m2 c2 0 0 + [0 , P] [0 , P ] }
1 + (0 )
(2) 1
T = c2 {[0 , P ] [0 , P ] + [0 , P ] [0 , P ]
2(1 + (0 ))
+ [0 , P ] [0 , P ] + [0 , P ] [0 , P ] }
(9.22c)
1
+ c2 {m2 c2 0 0 + m2 c2 0 0
2(1 + (0 ))
+ [0 , P ] [0 , P ] + [0 , P ] [0 , P ] }.
3 For instance, for them, generally, there are not one particle states with fixed energy, i.e. there are not one
particle eigenvectors of the zeroth comp onent of the momentum operator.

4 Since (9.2) and (9.7) are proportional, all results for (9.7) can trivially be formulated for (9.2).
Any one of the equations (9.22), together with (9.16), (I.6.6) and (I.5.1) form a complete
system of equations for explicit determination of , and P . It will be analyzed in the
subsequent sections.
Since the Lagrangians of a free general scalar field are invariant under (constant) phase
transformations, such a field carries a, possibly vanishing, charge (see, e.g., [1, 5]. The
(total) charge operator Q is defined by

1
Q := J0 (x) d3x (9.23)
c
x0 =const
where J (x) is a Hermitian operator,
J (x) = J (x), (9.24)
describing the fields current and considered a little below. The charge operator Q is Her-
mitian, i.e.
Q = Q (9.25)
or
Q = Q (9.26)
in momentum picture. Since Q and J are conserved quantities, viz. they satisfy the equiv-
alent conservation laws
d Q
=0 J = 0, (9.27)
dx0
and a Q 0 for a = 1, 2, 3, due to (9.23), we can write
Q = 0. (9.28)
The consideration of Q as a generator of (constant) phase transformations [1, 5] leads

to the following (Heisenberg) equations/relations5
[ , Q ] = q [ , Q ] = q (9.29)
where q is a constant, equal to the opposite charge of the particles (quanta) of (see below
Sect. 11), such that
q=0 for = . (9.30)
The charge operator Q commutes with the momentum operator P ,
[ Q , P ] = 0. (9.31)
5 The equations (9.29) follow from the transformation properties of and too.
In momentum picture, the equations (9.29), evidently, read:
[ , Q ] = q [ , Q ] = q . (9.32)
As a consequence of (9.31) and (I.5.1), the charge operator commutes with the (evolu-
tion) operator U (x, x0 ) responsible to the transition from Heisenberg picture to momentum
one,
[ Q , U (x, x0)] = 0. (9.33)
Combining this with (I.5.4), we get
Q (x) = Q =: Q , (9.34)
that is the charge operators in Heisenberg and momentum pictures coincide.6

Let us note once again (see section I.4), the equations (9.29)(9.34) are from pure geo-
metrical origin and, hence, are external for the Lagrangian formalism in which they need an
independent proof; so, they have to be used with some caution before this proof is explicitly
given.
In momentum picture the equality (9.23) is convenient to be rewritten as (see
(I.5.4))

1
Q= U 1(x, x0) J0 (x) U (x, x0) d3x. (9.35)
c
x0 =x00
The only thing, we need for a complete determination of Q , is the explicit definition of the
(quantum) current J . If was a free classical arbitrary, real or complex, scalar field, we
would have
q
J (x) = ( (x) (x) (x) (x)). (9.36)
i
The straightforward transferring of this result into the quantum case gives
q
J(1)(x) = { (x) (x) (x) (x)}. (9.37a)
i
But, since the current operator must satisfy (9.24), the quantities (9.37a) are not suit-
able for components of a current operator if [ , ] = 0 and/or [ , ] = 0. Evidently,
here the situation is quite similar to the one with the definition of the energy-momentum
operator considered above. So, without going into details, we shall write here a list of three
admissible candidates for a current operator:7
q
J(2) = + (9.37b)
2i
(3) q
J = (9.37c)
i
(4) q
J = . (9.37d)
i
6 Obviously, (9.33) and (9.34) are valid for any operator Q commuting with the momentum operator.
7 For a partial discussion of the problem, see [38, 2]. The expression (9.37c) is the one most often used in
the literature [1, 2, 38]; however, the definition (9.37a) is utilized too, for instance, in [5, p. 99].
Similarly to the case of energy-momentum operator, to any one of the La-

grangians (9.7)(9.9), there corresponds a unique current operator. These operators are
as follows (see [3] or section I.3):
(3) q
J = (9.38a)
i
(4) q
J = (9.38b)
i
(2) q
J = + . (9.38c)
2i
As a consequence of (9.15), these current operators in momentum picture read respec-
tively:
(3) 1
J = qc2 0 [0, P] [0 , P ] 0 (9.39a)
2
(4) 1
J = qc2 [0 , P ] 0 0 [0 , P] (9.39b)
2
(2) 1
J = qc2 0 [0, P] + [0 , P ] 0
4

[0 , P ] 0 0 [0 , P ] . (9.39c)
A free scalar field has no spin angular momentum and possesses a, generally, non-van-
ishing orbital angular momentum, as described in Sect. 2 (in particular, see equa-
tions (2.13)(2.17)). It will be explored in Sect. 13 directly in terms of creation and an-
nihilation operators.
According to the Klein-Gordon equations (9.16), the field operators 0 and 0 are
eigen-operators for the mapping (2.18) with eigenvalues equal to the square m2 of the mass
(parameter) m of the field (more precisely, of its quanta). Therefore the interpretation of
the operator (2.18) as a square-of-mass operator of the field is preserved also in the case of
free arbitrary, Hermitian of non-Hermitian, scalar field. At the same time, the square of the
momentum operator, c12 P P , has an interpretation of a square of mass operator for the
solutions of the field equations, i.e. for the fields states.
10. Analysis of the Field Equations

As we know, the field operator 0 and its Hermitian conjugate 0 satisfy the Klein-Gordon
equations (9.16) which are mixed through the momentum operator P , due to the simulta-
neous presentation of 0 and 0 in the energy-momentum operator(s) (9.22). However, for
0 and 0 are completely valid all of the results of Sect. 3 as in it is used only the Klein-Gor-
don equation (in momentum picture for 0 ) and it does not utilize any hypotheses about the
concrete form of the momentum operator P . Let us formulate the main of them.
Proposition 10.1. The common solutions 0 and 0 of the Klein-Gordon equations (9.16)
and (I.5.30b) can be written as

0 = d3 k f+(k)0 (k)
2 + f (k)0 (k) (10.1a)
k0 =+
2 2 m c +k 2 2k0 =
2
m c +k

0 = d3 k f+ (k)0 (k) + f (k)0 (k) , (10.1b)
k0 =+ m2 c2 +k2 k0 = m2 c2 +k2
where 0 (k), 0 (k) : F F are solutions of
[0 (k), P ] = k 0 (k) [0 (k), P ] = k 0 (k) (10.2)
and, for solutions different from the degenerate solutions
[0 , P ] = 0 [0 , P ] = 0 for m = 0 (10.3)
or, in Heisenberg picture,
(x) = (x0 ) = 0 (x) = (x0 ) = 0 P = P = 0 for m = 0, (10.3 )
the symbols f and f denote complex-valued functions of k and for the solutions (10.3)
they stand for some distributions of k.

Notice, as a result of the restriction k0 = m2 c2 + k 2 in (10.1), only the solutions
of (10.2) for which
k2 = k02 k 2 = m2 c2 (10.4)
are significant.
Let us note that the operator 0 (k) in (10.1) is not the Hermitian conjugate of 0 (k). In
fact, the reader can verify that (10.1) imply the equalities (cf. (3.12))

f (k)0 (k) = f (k)0 (k)
k0 = m c +k
2 2 2
k0 = m2 c2 +k2
(10.5)
f (k)0 (k) = f(k)0 (k)
2.
k0 =2 2 m c +k
2 2 2 k0 = m c +k
However, in the Hermitian case, i.e. for 0 = 0 , the equations (10.1a) and (10.1b) must be
identical and, consequently, we have
f (k) = f (k) 0 (k) = 0 (k) for 0 = 0 . (10.6)
In this case, the equations (10.5) reduce to (3.12).

Proposition 10.2. The solutions of (9.16) have the representations

0 = (k2 m2 c2 )0 (k) d4 k 0 = (k2 m2 c2 )0 (k) d4 k (10.7)
where 0 (k) and 0 (k) are suitably normalized solutions of (10.2) which, up to a phase fac-
1
tor equal to e i x0 k , coincide with the Fourier coefficients of 0 (x) and 0 (x) (in Heisenberg
picture for solutions different from (10.3)).
It should be emphasized, the solutions (10.3) are completely unphysical as they have
zero (energy-)momentum operator (see (9.22) and (I.6.6)), zero total charge (see (9.39))
and zero orbital angular momentum (see (2.15)) and, consequently, they cannot lead to
some physically predictable consequences.
11. Frequency Decompositions and Their Physical

Meaning
The presented in Sect. 4 frequency decompositions of a free Hermitian scalar field are
based on the Klein-Gordon equation, or, more precisely, on (3.6) and (3.8), and do not rely
on a concrete representation of the energy-momentum operator. Hence they can mutatis
mutandis be transferred in the general case of Hermitian or non-Hermitian scalar field. The
basic moments of that procedure are as follows.
Let us put (cf. (4.1) and (I.7.7))

f (k)0 (k) for k0 0
0 (k) :=
0 for k0 < 0
(11.1)

f ((k) (k) for k 0
0 (k) :=
0
0
.
0 for k0 < 0
As an evident consequence of (10.5), we have the equalities

0 (k) = 0 (k) 0 (k) =
0 (k) (11.2)
which mean that the operators 0 (k) are not the Hermitian conjugate of
0 (k). In the
Hermitian case, 0 = 0 , (11.1) reduce to (4.1) due to (10.6).
In view of (10.1) and (10.2), we have:
0 = +
0 + 0 0 = 0 + + 0 (11.3)

d3 k 0 = d3 k0 (k)|
0 = 0 (k)| k0 = m2 c2 +k2 k0 = m2 c2 +k2
(11.4)
[
0 (k), P ] = k 0 (k)
(11.5)
[0 (k), P ] = k 0 (k) k0 = m2 c2 + k2 .
If the equations (9.32), which are external for the Lagrangian formalism, hold, then (11.1)
and (11.2) also imply:

[
0 (k), Q ] = q0 (k) [0 (k), Q ] = q0 (k)
(11.6)
[
0 , Q ] = q0 [0 , Q ] = q0 .
If X p is a state vector characterizing a state with 4-momentum p (see equation (4.6)),

then (11.5) entail

P(
0 (k)( X p )) = (p k )0 (k)( X p ) k0 = m2 c2 + k2
(11.7)
P(0 (k)( X p)) = (p k )0 (k)( X p ) k0 = m2 c2 + k 2 .
+
So, +
0 and 0 create particles with 4-momentum k, while 0 and 0 annihilate

such particles. If = , the operators
0 and 0 coincide, while for = they are

different. In the last case the difference comes from the existence of non-zero charge opera-
tor (9.23) for which the (Heisenberg equations/)relations (9.29) hold. If Xe is a state vector
corresponding to a state with total charge e, i.e.
Q ( Xe ) = e Xe , (11.8)
then, from (9.32) and (11.6), we get:1
Q (0( Xe)) = (e q)0 ( Xe ) Q (0 ( Xe)) = (e + q)0 ( Xe )

Q (
0 ( Xe )) = (e q)0 ( Xe ) Q (0 ( Xe)) = (e + q)0 ( Xe). (11.9)
Q (
0 (k)( Xe )) = (e q)
0 (k)( Xe ) Q (0 (k)( Xe)) = (e + q)0 (k)( Xe ).

Therefore 0 ,
0 and 0 (k) decrease the fields charge by q, while 0 , 0

and
0 (k)increase it by the same quantity.

+ +
So, in a summary, 0 (k) and 0 (k) cre-

ate particles with 4-momentum ( m c + k 2 , k) and charges (q) and (+q),

2 2
respec-

tively, while 0 (k) and 0 (k) annihilate particles with 4-momentum ( m2 c2 + k 2 ,
k) and charges (+q) and (q), respectively.2, 3
12. The Field Equations in Terms of Creation and

Annihilation Operators
If one wants to obtain from the field equations (9.16) a system of equations for the mo-
mentum and field operators, an explicit expression for P , as a function (functional) of the
field operators 0 and 0 , is required. To find it, we shall proceed as in Sect. 5, when the
Hermitian case, 0 = 0 , was investigated.
Since (11.3)(11.5) imply

[0 , P ] = {k (+
0 (k) + 0 (k))}| k0= m2 c2 +k2
d3 k
(12.1)
[0 , P ] = {k (0 + (k) + 0 (k))}| 3
2 d k,
k0 = m2 c2 +k
the energy-momentum operators (9.22), in view of (11.4), can be written as:

(3) 1
T = c2 d3 k d3 k {(k k k k + k k )
1 + (0 )
(12.2a)
(0 + (k) + 0 (k)) (+
0 (k ) + 0 (k ))
+ m2 c2 (0 + (k) + 0 (k)) (+
0 (k ) + 0 (k ))}
1 These considerations do not use concrete forms, like (9.39), of the current operator J .
2 By convention, the particles created by + 0 or annihilated by 0 (k) are called particles, while the ones

created by 0 (k) or annihilated by 0 (k) are called antiparticles.
+
3 Since (11.9) originates from the equation (9.29), which is external to the Lagrangian formalism, one should

accept the given interpretation of 0 (k) and 0 (k) by some reserve. However, this interpretation is confirmed
in the later development of the theory on the base of a notion of n-particle, n N, states (see Sect. 16).

(4) 1
T = c2 d3 k d3 k {(k k k k + k k )
1 + (0 )
+ (12.2b)
(+
0 (k) + 0 (k)) (0 (k ) + 0 (k ))
+
+ m2 c2 (+
0 (k) + 0 (k)) (0 (k ) + 0 (k ))}

(2) 1
T = c2 d3 k d3 k {(k k k k + k k )
2(1 + (0 ))
(0 + (k) + 0 (k)) (+
0 (k ) + 0 (k ))
+ (k k k k + k k ) (12.2c)
+
(+
0 (k) + 0 (k)) (0 (k ) + 0 (k ))
+ m2 c2 (0 + (k) + 0 (k)) (+
0 (k ) + 0 (k ))
+
+ m2 c2 (+
0 (k) + 0 (k)) (0 (k ) + 0 (k ))}

where k0 = m2 c2 + k 2 and k0 = m2 c2 + k .
2
Performing with these expressions the same manipulations as the ones leading
from (5.1) to (5.5), we derive the following expressions for the momentum operator:1, 2

(3) 1
P = k | {0 +(k)
0 (k)
1 + (0 ) k0 = m2 c2 +k2
(12.3a)
+ 0 (k) +
0 (k)} d k
3

(4) 1
P = k | {+
0 (k) 0 (k)
1 + (0 ) k0 = m2 c2 +k2
(12.3b)
+
+
0 (k) 0 (k)} d k
3

(2) 1
P = k | {0 + (k) +
0 (k)+0 (k)0 (k)
2(1 + (0 )) k0 = m2 c2 +k2
(12.3c)
+
+ +
0 (k) 0 (k) + 0 (k) 0 (k)} d k.
3
Here (0 ) is defined by (9.13) and the following shortcuts are introduced:

1/2

0 (k) := 2c(2)3k0 0 (k) 2 2 2
k0 = m c +k
1/2 (12.4)

0 (k) := 2c(2) k0
3
0 (k) 2 2 2 .
k0 = m c +k
+
The operators +
0 (k) and 0 (k) (resp. 0 (k) and 0 (k)) are called the creation (resp.
annihilation) operators (of the field (fields particles)).
1 Notice, the equalities (5.4) remain valid in the general case. Besides, these equations hold if in them some

or all of the operators
0 (k) and 0 (k ) are replaced with 0 (k) and 0 (k ), respectively.
2 For the other 9 combinations of the Lagrangians (9.7)(9.9) and energy-momentum operators (9.17)
(9.20), in the integrands in the r h.s. of (12.3) terms proportional to

0 (k) 0 (k0 ,k) and 0 (k)

0 (k0 ,k) will appear. They are responsible for the contradictions mentioned in Sect. 9 (see, in particular,
footnote 3 on page 51).
Obviously, for a Hermitian field, 0 = 0 , all of the three expressions in (12.3) reduce
to the right Hermitian result (5.5); besides, the 3-dimensional creation/annihilation op-
erators (12.4) reduce to (5.6), as one should expect. In the non-Hermitian case, 0 = 0 ,
the operator (12.3b), as a function of the creation/annihilation operators (12.4), formally
coincides with the momentum operator obtained from the Lagrangian (9.2) in the litera-
ture [1, eq. (3.39)]. However, it should be remarked, our creation and annihilation operators
for the Lagrangian (9.2) with 0 = 0 differ by a phase factor from the ones in Heisenberg
picture used in the literature (see below equations (12.5)). Generally, the three momentum
operators (12.3) are different, but, after normal ordering, they result into one and the same
momentum operator (see Sect. 15, equation (15.2)).
For a comparison with expressions in (the momentum representation of) Heisenberg
picture, it is worth to be noticed that, due to proposition 10.2, the creation/annihilation
operators (k) and (k) in (the momentum representation of) Heisenberg picture are
(cf. (5.8) and (4.10))
1
(k) = e i x0 k 2 2 2 (k)
k0 = m c +k 0
1
(12.5)
(k) = e i x0 k 2 2 2 0 (k).
k0 = m c +k
(Relations like (5.9) are, of course, also valid; (12.5) is a consequence from them for x =

x0 .) Therefore, quadratic expressions, like the ones in the integrands in (12.3), look in one
and the same way in momentum and Heisenberg pictures.
Now we are ready to obtain the field equations in terms of creation and annihilation
operators. Since (11.3)(11.5) are equivalent to (9.16), the equations, we want to derive,
are (11.5) with P given via (12.3). At this stage of the development of the theory, we
will get three, generally different, systems of equations, corresponding to the three La-
grangians, (9.7), (9.8) and (9.9), we started off. But, as after normal ordering the three
momentum operators (12.3) become identical, these systems of equations will turn to be
identical after normal ordering.
In terms of the operators (12.4), the equations (11.3)(11.5) can equivalently be rewrit-
ten as:
1/2
0 = 2c(2)3k0 (+
0 (k) + 0 (k)) d k
3
1/2 (12.6)
0 = 2c(2)3k0 (0 + (k) + 0 (k)) d3 k

[
0 (k), P ] = k 0 (k) [0 (k), P ] = k 0 (k)
(12.7)
k0 = m2 c2 + k2 .

Inserting the equalities (with k0 = m2 c2 + k2 and q0 = m2 c2 + q2 )

k
0 (k) = q
0 (k) (q k) d q
3 3

k 0 (k) = q 0 (k)3 (q k) d3 q
and (12.3) into (12.7), we, after some algebra, obtain the next variants of the systems of
field equations for an arbitrary scalar field in terms of creation and annihilation operators:

q |q =m2 c2 +q2 [ + +
0 (k), 0 (q) 0 (q) + 0 (q) 0 (q)]
0
3 (12.8a)
(1 + (0 )) 0 (k) (q k) d q = 0
3

q |q =m2 c2 +q2 [0 (k), + +
0 (q) 0 (q) + 0 (q) 0 (q)]
0
(12.8b)
(1 + (0 ))0 (k)3 (q k) d3 q = 0

q |q =m2 c2 +q2 [ + +
0 (k), 0 (q) 0 (q) + 0 (q) 0 (q)]
0
3 (12.9a)
(1 + (0 )) 0 (k) (q k) d q = 0
3

q |q =m2 c2 +q2 [0 (k), + +
0 (q) 0 (q) + 0 (q) 0 (q)]
0
(12.9b)
(1 + (0 ))0 (k)3 (q k) d3 q = 0

q |q =m2 c2 +q2 [ +
0 (k), [0 (q), 0 (q)]+ + [0 (q), 0 (q)]+ ]
+
0
3 (12.10a)
2(1 + (0 )) 0 (k) (q k) d q = 0
3

q |q =m2 c2 +q2 [0 (k), [+ +
0 (q), 0 (q)]+ + [0 (q), 0 (q)]+ ]
0
(12.10b)
2(1 + (0 ))0 (k)3 (q k) d3 q = 0.
Consequently, to the Lagrangians (9.7)(9.9) correspond respectively the following

three systems of equations for
0 and 0 :
+
[ +
0 (k), 0 (q) 0 (q) + 0 (q) 0 (q)]

(12.11a)
(1 + (0 ))
0 (k) (q k) = f (k, q)
3
[0 (k), 0 + (q) +
0 (q) + 0 (q) 0 (q)]

(12.11b)
(1 + (0 ))0 (k)3 (q k) = f (k, q)
+
[ +
0 (k), 0 (q) 0 (q) + 0 (q) 0 (q)]
(12.12a)
(1 + (0 ))
0 (k) (q k) = f (k, q)
3
[0 (k), + +
0 (q) 0 (q) + 0 (q) 0 (q)]

(12.12b)
(1 + (0 ))0 (k)3 (q k) = f (k, q)

[ +
0 (k), [0 (q), 0 (q)]+ + [0 (q), 0 (q)]+ ]
+

(12.13a)
2(1 + (0 ))
0 (k) (q k) = f (k, q)
3
[0 (k), [+ +
0 (q), 0 (q)]+ + [0 (q), 0 (q)]+ ]

(12.13b)
2(1 + (0 ))0 (k)3 (q k) = f (k, q),
where the operator-valued (generalized) functions af (k, q) and af (k, q) with a = , ,
must be such that

q | q a
f (k, q) d q =
3
q |q =m2 c2 +q2 af (k, q) d3 q = 0. (12.14)
0= m2 c2 +q2 0
Generally the systems of equations (12.11), (12.12) and (12.13) are different. They
will become equivalent after normal ordering, when the equations in them will turn to be
identical. From the derivation of (12.11)(12.14), it is clear that these systems of equations
are equivalent to the initial system of Klein-Gordon equations (9.16). Thus, we can say that
these systems represent the field equations in terms of creation and annihilation operators.
Exercise 12.1. As a verification of the self-consistence of the theory, prove that the com-
mutativity between the components of any one of the momentum operators (12.3), i.e.
[ P , P ] = 0, (12.15)
(3) (4) (2)
where P = P , P , P , is a consequence of (12.3), (12.11)(12.13) and the iden-
tity (6.4) with = 1.
13. The Charge and Orbital Angular Momentum

Operators
In Sect. 9, we introduced the charge operator (9.23) (see also (9.35)) and pointed to dif-
ferent possible definitions of the defining it current operator. In particular, to the La-
grangians (9.7)(9.9) correspond respectively the current operators (9.39a)(9.39c) in mo-
mentum picture. Below we shall express the charge operator Q through the creation and
annihilation operators (11.1).
Substituting (11.3), (11.4) and (12.1) into (9.39), we get:1

(3) 1
J = qc2 d3 k d3 k k (0 + (k) + 0 (k)) (+
0 (k ) + 0 (k ))
2

k (0 + (k) + 0 (k)) (+
0 (k ) + 0 (k )) (13.1a)

(4) 1 +
J = qc2 d3 k d3 k k (+
0 (k) + 0 (k)) (0 (k ) + 0 (k ))
2
+
k (+
0 (k) + 0 (k)) (0 (k ) + 0 (k )) (13.1b)

(2) 1
J = qc2 d3 k d3 k k (0 + (k) + 0 (k)) (+
0 (k ) + 0 (k ))
4
+
+ k (+
0 (k) + 0 (k)) (0 (k ) + 0 (k ))
k (0 + (k) + 0 (k)) (+
0 (k ) + 0 (k ))
1Since we apply only results based on the general properties of the momentum operator, the below obtained
equalities for the current and charge operators are independent of the concrete choice of energy-momentum
operator, like (9.17)(9.20), and, consequently, of the particular form of the momentum operator, like (12.3).
However, they depend on the Lagrangians (9.7)(9.9) from which the theory is derived.
+
k (+
0 (k) + 0 (k)) (0 (k ) + 0 (k )) . (13.1c)
Now we have to perform the following three steps: (i) insert these expressions with
= 0 into (9.23); (ii) apply (5.4), possibly with 0 (k) and/or 0 (k ) for
0 (k) and/or

0 (k ); (iii) integrate over x, which gives (2) (k + k ) for products of equal-frequency
3 3
operators and (2)33 (k k ) for products of different-frequency operators. As a result of

these steps, we obtain:2

(3)
Q =q d3 k 0 + (k) +
0 (k) 0 (k) 0 (k) (13.2a)

(4) +
Q = q d3 k +
0 (k) 0 (k) 0 (k) 0 (k) (13.2b)

(2) 1
Q = q d3 k 0 + (k) +
0 (k) 0 (k) 0 (k)
2
+
+ +
0 (k) 0 (k) 0 (k) 0 (k) . (13.2c)
Recall, here the 3-dimensional creation/annihilation operators are defined via (12.4).
Notice, if 0 = 0 , the charges (13.2a) and (13.2b), as well as the defining them respective
current operators (9.39a) and (9.39b), vanish as q = 0 in this case, while the charge (13.2c)
and the defining it current operator (9.39c) vanish for two reasons, if 0 = 0 : due to q = 0
and due to the vanishment of the integrand in (13.2c) or the expression in braces in (9.39c).
Generally, the three charge operators (13.2a)(13.2c), originating respectively from the
Lagrangians (9.7)(9.9), are different, but they become identical after normal ordering (see
Sect. 15).
Let us turn our attention now to the orbital angular momentum operator of free scalar
field. In Heisenberg picture, it is given via (2.15) To obtain an explicit expression for
the orbital angular momentum operator L (in Heisenberg picture) through the creation

and annihilation operators 0 (k) and 0 (k) (see (12.4)), we have to do the following:
(i) substitute each of the equalities in (12.2) into the last equality in (2.15); (ii) apply (5.4)
to all terms in the obtained equation; (iii) then integrate over x, which results in -function
terms; (iv) and, at last, perform the integration over k by means of the arising -functions.3
2 When deriving (13.2), one gets the equalities in it with q
1+(0)
for q with (0 ) defined by (9.13). Since
q = 0 for 0
= 0 and (0 ) = 0 for = 0 , we have0 q
1+(0) q in all cases.
3 The integrals, one has to calculate, are of the type

J=

d3 x d3 k d3 k xa U 1 (x,x0 ) {
0 (k)A (k,k ) 0 (k )} U (x,x0 )
, =+,

= x )(k + k )
d3 k d3 k { i (x
1
d3 x 0 (k)A (k,k ) s (k )}xae ,
0 (13.3)
, =+,

where a = 1,2,3, A (k,k ), = 1,2,... , are some functions, k0 = m2 c2 + k2 , and k0 = m2 c2 + k . The
2

integration over x results in (2)3 i (ka + k a ) (k + k ). Simple manipulations with the remaining
3
(y) f (y)
terms, by invoking the equality f (y) y
= y
(y) in the form

3 (y z) 1 3 (y z)
d3 yd3 z f (y,z) = d3 y d3 z f (y,z) f (z,y)
(ya za ) 2 (ya za )
This procedure results in:

(3) i
L = x0 P (3) (3)
x0 P + d3 k 0 + (k) k k 0 (k)
2(1+(0 )) k k

0 (k) k k + (k) (13.5a)
k k 0
k0 = m2 c2 +k2

(4) i
L = x0 P (4) (4)
x0 P + d3 k + (k) k k 0 (k)
2(1+(0 )) 0
k k

+
(k) k k 0 (k) 2 2 2 (13.5b)
0
k k k0 = m c +k
(2)
L = x0 P (2) (2)
x0 P

i
+
+ d k 0 (k) k k
3
0 (k)
4(1 + (0 )) k k

+
+
+ 0 (k) k k 0 (k) 0 (k) k k 0 (k)
k k k k

+
0 (k) k k 0 (k) 2 2 2 . (13.5c)
k k k0 = m c +k

Here the derivatives with respect to k0 , like (k),
k 0 0
must be set equal to zero, and4

A(k) B(k)
A(k)k B(k) := k B(k) + A(k) k
k k k

= k A(k) B(k) (13.6)
k

1
= d3 y d3 z3 (y z) f (y,z),
2 ya za
gives the following result:

J = (2)3
0 x0 )k (1+1)
d3 k d3 k 3 (k + k )e i (x
1
0 0
, =+,
1 ka ka 1
(x0 x00 ) + + x0a + i a + a 0 (k)A (k,k ) 0 (k ) . (13.4)
2 k0 k0 2 k k
The particular form of A (k,k ) is clear from (2.15) and (9.22). So, applying several times (13.4), calculating
the appearing derivatives, and, at last, performing the trivial integration over k or k by means of 3 (k + k ),
one can derive (13.5) after simple, but lengthy and tedious algebraic manipulations.
4 More generally, if : {F F } {F F } is a mapping on the operator space over the systems

Hilbert space, we put A B := (A) B + A (B) for any A,B : F F . Usually [4, 11], this notation is
used for = .
for operators A(k) and B(k) having C1 dependence on k. If the operators A(k) and B(k) tend
to zero sufficiently fast at spacial infinity, then, by integration by parts, one can prove the
equality

d k A(k) k k B(k) 2 2 2
3
k k k0 = m c +k

= 2 d k A(k) k k B(k) 2 2 2
3
k k k0 = m c +k

= 2 d3 k k k A(k) B(k) 2 2 2 . (13.7)
k k k0 = m c +k
By means of these equations, one can reduce (two times) the number of terms in (13.5),
but we prefer to retain the more (anti)symmetric form of the results by invoking the oper-
ation introduced via (13.6).
Since for a neutral scalar field 0 = 0 and (0 ) = 1, in this case the three opera-
tors (13.5) reduce to

i
d3 k +
0
L = x0 P x0 P + (k) k k 0 (k)
4 0
k k

+

0 (k) k k 0 (k) 2 2 2 (13.8)
k k k0 = m c +k
with P given by (5.5).
Exercise 13.1. Verify that the equality (13.8) holds for any one of the energy-momentum
operators (2.12) (or (2.11)) for the Lagrangian (2.3) (or (2.2)).
Exercise 13.2. Show that in terms of the operators (12.5), representing the creati-
on/annihilation operators in (momentum representation of) Heisenberg picture, in the equa-
tions (13.5) and (13.8) the first two terms, proportional to the momentum operator, should
be deleted and tildes over all creation and annihilation operators must be added; for in-
stance, we have:

i
0 +
L = d k 0 (k) k k 0 (k)
3
4 k k

0 (k) k k +

(k) . (13.9)
k k 0
k0 = m2 c2 +k2
Exercise 13.3. Using the explicit formulae (12.3), (13.2) and (13.5), and by means of the
identity (6.4), with = 1, and the field equations (12.11)(12.13), verify the equations:
[ Q , P ] = 0 (13.10)
[ L, P] = i( P P ), (13.11)
where L (resp. Q ) denotes any one of the orbital momentum operators in (13.5) (resp.
charge operators in (14.21)). The first of these equalities confirms the validity of the (ex-
ternal to the Lagrangian formalism) equation (9.31). We emphasize on the sign before i
in the r.h.s. of (13.11), which is opposite to the one usually assumed in the literature, for
instance in [5, p. 77, eq. (2-87)] or in [6, eq. (2.187)].5
From (13.10) and (I.5.1), we get
[ Q , U (x, x0 )] = 0 (13.12)
and, hence,
Q = Q , (13.13)
due to (I.5.4). Therefore the charge operator is one and the same in momentum and Heisen-
berg pictures. In this way, equations (I.4.5), (9.31)(9.34), (I.6.16), (I.6.20) and (I.6.22) are
rigorously proved for arbitrary scalar fields in the framework of the Lagrangian formalism.
Applying (13.11), we get the orbital angular momentum of a scalar field in momentum
picture as (cf. (I.6.23))
L = U (x, x0) L U 1(x, x0 ) = L + [ U (x, x0), L ] U 1(x, x0)

= L + (x x0 ) P (x x0 ) P , (13.14)
due to the equality
[ L, U (x, x0 )] = {(x x0 ) P (x x0 ) P } U (x, x0 ). (13.15)
which is a consequence of (13.11) and (I.5.1).6 Thus, in momentum picture, the equa-
tions (13.5) read:

i
(3)
L = x P (3) (3)
x P +
+
d k 0 (k) k k
3
0 (k)
2(1+(0 )) k k

0 (k) k k + (k) (13.16a)
k k 0
k0 = m2 c2 +k2
5 Equation (13.11) with +i for i in its r h.s. is part of the commutation relations for the Lie algebra of
the Poincare group see, e.g., [9, pp.143147] or [10, sec. 7.1]. However, such a change of the sign in the r h.s.
of (13.11) contradicts to the results and the physical interpretation of the creation and annihilation operators
vide infra.
6 To derive equation (13.15), notice that (13.11) implies [ L , P P ] = n
1 n i=1 i P
i P P1 Pi1 Pi+1 Pn , due to [A,B C] = [A,B] C + B [A,C] , and expand the ex-
ponent in (I.5.1) into a power series. More generally, if [A(x), P ] = B (x) with [B (x), P ] = 0, then
[A(x), U (x,x0 )] = i 1 (x x )B (x) U (x,x ); in particular, [A(x), P ] = 0 implies [A(x), U (x,x )] = 0. No-
0 0 0
tice, we consider (x x0 ) as a real parameter by which the corresponding operators are multiplied and which
operators are supposed to be linear in it.

i
(4)
L = x P (4) (4)
x P + d k 0 (k) k k 0 (k)
3 +
2(1+(0 )) k k

0 (k) k k 0 + (k) 2 2 2 (13.16b)

k k k0 = m c +k

i
L = x P (2)
(2) (2)
x P +
+
d k 0 (k) k k 0 (k)
3
4(1+(0 )) k k

+
+
+ 0 (k) k k 0 (k) 0 (k) k k 0 (k)
k k k k

+
(k) k k 0 (k) 2 2 2 . (13.16c)
0
k k k0 = m c +k
These three angular momentum operators are different but, after normal ordering, they will
be mapped into one and the same operator (see Sect. 15).

The trilinear systems of equations (12.11)(12.13) are similar to the (system of) Klein-Gor-
don equation(s) (5.13) and, correspondingly, will be treated in an analogous way.
Since the equations (12.11b), (12.12b), and (12.13b) can be obtained from
a
(12.11a), (12.12a), and (12.13a) by replacing 0 (k) with 0 (k) and f (k, q) with
a
f (k, q), respectively, all of the next intermediate considerations will be done only for
the former set of equations; only the more essential and final results will be doubled, i.e.

written for 0 (k) and 0 (k).
First of all, applying the identity (6.4) several times, we rewrite (12.11a), (12.12a),
and (12.13a) respectively as (recall, = 1)
+ +
[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)]

+[ + +
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)] (14.1)

(1 + (0 ))
0 (k) (k q) = f (k, q)
3

[ + +
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)]
+ +
+[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)] (14.2)

(1 + (0 ))
0 (k) (k q) = f (k, q)
3
+ +
[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)]
+ +
+[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)]

+[ + +
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)] (14.3)

+[ + +
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)]

2(1 + (0 ))
0 (k) (k q) = f (k, q).
3
Now, following the know argumentation [1, 4, 5], we shall impose the following addi-
tional condition (cf. assumption 6.1).
Assumption 14.1. The commutators, = 1, or anticommutators, = +1, of all combina-
tions of creation and/or annihilation operators are proportional to the identity operator idF
of the Hilbert space F of the considered free arbitrary scalar field.
It is easily seen, the equations (14.1) and (14.2) (and similar ones obtained from them
with 0 (k) for 0 (k)) do not make any difference between the choices = 1 and =
+1. But, for equation (14.3), the situation is completely different. Indeed, for = +1,
which corresponds to quantization of a scalar field by anticommutators, equation (14.3)

reduces to 2(1 + (0 )) 0 (k q) = f (k, q) which, when inserted in (12.14), entails
3
k | 2 2 2 (k) = 0 for any k; a similar result follows from (12.13b), i.e. we have

k0 = m c +k
k | (k) = 0 k | (k) =0 for = +1 (14.4)

k0 = m2 c2 +k2 0 k0 = m2 c2 +k 2 0
which, by (12.3c), implies

(2)
P = 0. (14.5)
Consequently, since (14.5) and the Klein-Gordon equations (9.16) imply
m2 c2 0 = 0 m2 c2 0 = 0, (14.6)
the choice = +1 for (14.3) is possible only for the degenerate (unphysical) solutions (10.3)
(or (10.3 ) in Heisenberg picture) and for the solution
0 = 0 0 = 0 for m = 0, (14.7)
the last of which is equivalent to

0 (k) = 0 0 (k) = 0 for m = 0. (14.8)
According to equations (13.2), (13.16) and (14.5), the degenerate solutions (10.3)
and (14.7) carry no 4-momentum, (orbital) angular momentum and charge and, hence, can-
not be detected. So, these solutions should be interpreted as an absence of the scalar field
and, if one starts from the Lagrangian (9.9), they are the only ones that can be quantized by
anticommutators.1
Let us return to the consideration of equations (14.1)(14.3) and similar ones with
0 (k) for
0 (k) having in mind that = 1 for (14.3). Writing explicitly them for the
upper, +, and lower, , signs, we see that they can equivalently be represented respec-
tively in the forms:

[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)]

+[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)] (14.9)

(1 + (0 ))
0 (k) (k q)
3
= f (k, q)
1 In fact, the last assertion completes the proof of spin-statistics theorem for free arbitrary scalar field.
Notice, in this proof we have not used any additional hypotheses,like charge conjugation/symmetry or positivity
of the Hilbert space metric (cf. [1, sec. 10.2]).

[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)]

+[
0 (k), 0 (q)] 0 (q) 0 (q) [0 (k), 0 (q)] (14.10)

(1 + (0 ))
0 (k) (k q) = f (k, q)
3
0 (q) [
0 (k), 0 (q)] + 0 (q) [0 (k), 0 (q)]
0 (q) [
0 (k), 0 (q)] + 0 (q) [0 (k), 0 (q)] (14.11)
1
(1 + (0 )) 0 (k) (k q) =
3
f (k, q).
2
Let us write explicitly the above-stated assumption 14.1 concerning the (anti)commu-
tators of creation and annihilation operators. We have (cf. equations (6.1)):
[
0 (k), 0 (q)] = a (k, q) idF [0 (k), 0 (q)] = a (k, q) idF
[
0 (k), 0 (q)] = b (k, q) idF [0 (k), 0 (q)] = b (k, q) idF

(14.12)
[
0 (k), 0 (q)] = d (k, q) idF [0 (k),
0 (q)] = d (q, k) idF

[
0 (k), 0 (q)] = e (k, q) idF [0 (k),
0 (q)] = e (q, k) idF

where = 1 for (14.1) and (14.2), = 1 for (14.3), and a
, a , . . . , e are some
complex-valued (generalized) functions, which we have to determine. These last functions
are subjected to a number of restrictions which can be derived in the same way as (6.18)
and (6.19) in the Hermitian case.2 One can easily verify that these restrictions are:
(k + q)a
(k, q) = 0 (k + q)a (k, q) = 0 (k + q)d (k, q) = 0 (14.13a)
(k q)b (k, q) = 0 (k q)b (k, q) = 0

(k q)e
(k, q) = 0 (14.13b)

m2 c2 + k 2 + m2 c2 + q2 (k, q) = 0 for = a
, a , d (14.14a)

m2 c2 + k2 m2 c2 + q2 (k, q) = 0 for = b
, b , e . (14.14b)

Regarding a
, a , . . . ,e as distributions, from (14.13), we derive (cf. (6.21)):
f (q)(k, q) = f (k)(k, q) for = a

, a , d (14.15a)
f (q)(k, q) = f (+k)(k, q) for = b
, b , e (14.15b)
for any function f which is polynomial or convergent power series. In view of (14.15), the
equalities (14.14b) are identically satisfied, while (14.14a) are equivalent to the equations

m2 c2 + k2 (k, q) = 0 for = a
, a , d . (14.16)
Substituting the equalities (14.12) into equations (14.9)(14.11) and similar ones with
0 (k)
for 0 (k), we see that the restrictions (12.14), in view of (14.15), give the following

systems of equations for the unknown (generalized) functions a
, a , . . . , e :

ka d3 q 0 (q)a
(k, q) 0 (q)d (k, q) 0 (q)b (k, q)
(14.17a)
+ 0 (q) e (k, q) (1 + (0 ))3
(k q) = 0

Notice, (6.15)(6.17) remain valid if we replace in them
2
0 (k) and/or 0 (q) with 0 (k) and/or 0 (q)
respectively; see (11.7).

m2 c2 + k2 d3 q 0 (q)a
(k, q) + 0 (q)d (k, q)
(14.17b)

0 (q)b (k, q) + 0 (q) e (k, q) (1 + (0 )) (k q)
3
=0

ka d3 q 0 (q)d (q, k)
0 (q)a (k, q) + 0 (q)b (k, q)

(14.17c)
+0 (q) e (q, k) (1 + (0 )) (k q)
3
=0

m2 c2 + k2 d3 q 0 (q)d(q, k) + 0 (q)a (k, q)

(14.17d)

+
0 (q)b (k, q) + 0 (q) e (q, k) (1 + (0 )) (k q)
3
= 0.
Here: a = 1, 2, 3, = for (14.10) and = 1 for (14.9) and (14.11), = 1 for (14.9)
and (14.10), and = 1 for (14.11) (vide supra). Notice, (14.17a) and (14.17b) corre-
spond to (14.9)(14.11), while (14.17c) and (14.17d) correspond to the same equations
with 0 (k) for
0 (k).
Now we shall assume the following assumption (cf. assumption 6.2a).

Assumption 14.2a. The equations (14.17) are valid for arbitrary
0 (q) and 0 (q).

Therefore we see that, if (m, k) = (0, 0), the only solution of (14.17) relative to a
, a ,
. . . , e
is:

a
(0 ) (k, q) = a(0 ) (k, q) = d(0) (k, q) = 0 (14.18a)

b
(0 ) (k, q) = b(0) (k, q) = (0 ) (k q)
3
(14.18b)
e
(0 ) (k, q) = (0 ) (k q),
3
(14.18c)
where (0 ) is defined via (9.13) and:

1 for equation (14.11) with any 0

(0 ) = and for equations (14.9) and (14.10) with 0 = 0

1 for equations (14.9) and (14.10) with 0 = 0 (14.19)

(0 ) for equation (14.10)
(0 ) = .
+1 for equations (14.9) and (14.11)
Evidently, (14.18a) converts (14.16) into identity and, consequently, under the hypotheses
made, (14.18) is the general solution of our problem.
It should be emphasized on the fact that the function (0 ) in (14.18) takes care of what
is the field 0 , Hermitian or non-Hermitian, while the functions (0 ) and (0 ) take care
of from what Lagrangian, (9.7)(9.9), we have started off.
Before commenting on the solutions (14.18), we want to say some words on the case
m = 0 and k = 0 for which the equations (14.17) and (14.16) take the form of the identity
0 = 0 and, consequently, no information can be extracted from them. The above analysis
reveals that, under the additional conditions (14.12), the field equations do not impose some
restrictions on the operators

0 (k) and 0 (k) for m = 0 and k = 0, (14.20)
i.e. these operators must satisfy (14.12) with m = 0 and k = 0 and arbitrary a (0, q),
a (0, q), . . . , e
(0, q) (with = 1 for (14.1) and (14.2) and = 1 for (14.3)). To ensure
a continuous limit (m, k) (0, 0), we shall assume by convention (cf. assumption 6.2b)

Assumption 14.2b. The functions a
(0, q), a (0, q), . . . , e (0, q) are given via (14.18)
with k = 0 (and m = 0).
From physical point of view (see section 11), the operators (14.20) describe creation
or annihilation of massless particles with vanishing 4-momentum and charge q, which is
zero for a Hermitian (neutral) filed and non-zero for a non-Hermitian (charged) one. Con-
sequently, in the non-Hermitian case, the theory admits existence of free, charged, massless
scalar particles with vanishing 4-momentum, which are quanta of free, charged, massless
scalar field. As far as the author of these lines knows, such particles/fields have not been
observed until now. In the Hermitian case, as we pointed in Sect. 6, the operators (14.20)
reduce to (6.27) and describe unphysical particles/fields which are experimentally unob-
servable.
Since the (anti)commutation relations (14.12) are extremely important for quantum field
theory, we shall write them explicitly for the obtained solutions (14.18). As the three ver-
sions of some equations, like (14.17), (14.1)(14.3), (12.11)(12.13), (12.3), etc., origi-
nate from the Lagrangians (9.7)(9.9) we have started off, we shall associate the found
(anti)commutation relations with the initial Lagrangians rather than with the particular
equations utilized in their derivation.
Since equations (14.11), for which (0 ) = 1 and (0 ) = +1, originate from the La-
grangian (9.9), we can assert that the Lagrangian (9.9) implies the following commutation
relations:
[
0 (k), 0 (q)] = 0 [0 (k), 0 (q)] = 0
[
0 (k), 0 (q)] = (0 ) (k q) idF
3
[0 (k), 0 (q)] = (0 )3 (k q) idF

[
0 (k), 0 (q)] = 0 [0 (k),
0 (q)] = 0

[
0 (k), 0 (q)] = (k q) idF
3
[0 (k),
0 (q)] = (kq) idF
3
(14.21)
where 0 denotes the zero operator on F and (0 ) takes care of is the field neutral (0 = 0 ,
(0 ) = 1) or charged (0 = 0 , (0 ) = 0) and ensures a correct commutation relations in
the Hermitian case (see (6.28)).
Since the equations (14.9) are consequences of the Lagrangian (9.7), we can assert that
the following (anti)commutation relations follow from the Lagrangian (9.7):
[
0 (k), 0 (q)] = 0 [0 (k), 0 (q)] = 0
[
0 (k), 0 (q)] = (0 ) (k q) idF
3
[0 (k), 0 (q)] = (0 )3 (k q) idF

[
0 (k), 0 (q)] = 0 [0 (k),
0 (q)] = 0

[
0 (k), 0 (q)] = (k q) idF
3
[0 (k),
0 (q)] = (kq) idF
3
(14.22)
where = 1 (commutation relations) for a Hermitian filed, 0 = 0 , and = 1 (com-

mutation or anticommutation relations) for a non-Hermitian filed, 0 = 0 .
At last, since the Lagrangian (9.8) entails (14.10), the Lagrangian (9.8) implies the
following (anti)commutation relations:
[
0 (k), 0 (q)] = 0 [0 (k), 0 (q)] = 0
[
0 (k), 0 (q)] = (0 ) (k q) idF
3
[0 (k), 0 (q)] = (0 )3 (k q) idF

[
0 (k), 0 (q)] = 0 [0 (k),
0 (q)] = 0

[
0 (k), 0 (q)] = (k q) idF
3
[0 (k),
0 (q)] = (kq) idF
3
(14.23)
where = 1 (commutation relations) for a Hermitian filed, 0 = 0 , and = 1 (com-

mutation or anticommutation relations) for a non-Hermitian filed, 0 = 0 .
It should be emphasized, for a Hermitian (neutral, real) field, when = 1 in (14.22)
and (14.23), the commutation relations (14.21), (14.22), and (14.23) coincide and, due to
(0 ) = 1 in this case, are identical with (6.28); thus, they correctly reproduce the already
established results in Sect. 6. However, for a non-Hermitian (charged) field, for which
(0 ) = 0, we have three independent sets of (anti)commutation relations:
(i) the commutation relations (14.21) correspond to the Lagrangians (9.7) and (9.8), with
the choice = 1 for the both ones, and the Lagrangian (9.9);
(ii) the anticommutation relations (14.22) with = +1 correspond to the Lagrangian (9.7)
with the choice = +1;
(iii) the anticommutation relations (14.23) with = +1 correspond to the Lagrangian (9.8)
with the choice = +1.
The relations (14.23) with = +1 differ from (14.22) with = +1 only in the sign
before the -function in the last row. This is quite understandable as the Lagrangian (9.8)
can be obtained from (9.7) by replacing with and with . If we make the same

change in (14.22), i.e.
0 0 , we see that (14.22) transforms into (14.23). Since (14.22)
and (14.23) are identical (resp. different) for = 1 (resp. = +1), we conclude that
the theory is invariant (resp. non-invariant) under the change 0 0 or, equivalently,

0 0 , called charge conjugation [1,2,5], if and only if it is quantized via commutators
(resp. anticommutators) if one starts from any one of the Lagrangians (9.7) and (9.8). The
theory is always charge symmetric, i.e. invariant under charge conjugation, if one starts
from the Lagrangian (9.9).
Thus, for a free non-Hermitian scalar field, we see a principal difference between the
Lagrangian (9.9), on one hand, and the Lagrangians (9.7) and (9.8), on the other hand: the
first Lagrangian entails quantization with commutators, while the other two imply quan-
tization either with commutators (identical with the one of the previous case) or with an-
ticommutators and one needs a new additional condition/hypothesis to make a distinction
between these two cases. As it is well known, the correct quantization of a free scalar field
is via commutators, not by anticommutators [1, 2, 5]. Thus we accept
Assumption 14.3. If the theory of free non-Hermitian scalar field is derived from the La-
grangian (9.7) or (9.8), then the commutation relations (14.21) hold.
Usually (see loc. cit.), the relations (14.21) are derived, for a charged field, from the
Lagrangian (9.2) by invoking a new additional condition, like charge symmetry, or positiv-
ity of the Hilbert space metric, or spin-statistics theorem.3 The above considerations show
that these additional conditions are not required if one starts from the Lagrangian (9.9);
in fact, these conditions are corollaries from it, as we saw with the charge symmetry and
spin-statistics theorem (saying that a scalar field, which is a spin zero field, must be quan-
tized via commutators). This is not a surprising result, if we recall that the Lagrangian (9.9)
is a sum of the Lagrangians of two (independent) Hermitian scalar fields (see Sect. 9) for
which a similar result was established in Sect. 6. In conclusion, the Lagrangian (9.9) is
richer in consequences than the Lagrangians (9.7) and (9.8);4 the cause for this is that
and enter in (9.9) on equal footing, i.e. (9.9) is invariant under the change ,
which cannot be said relative to (9.7) and (9.8).
Relying on the above discussion, the commutation relations (14.21) will be accepted
from now on in this chapter. As we proved, under the hypotheses made, they are equivalent
to the initial system (9.16) of Klein-Gordon equations. If, by some reason one rejects these
hypotheses, the system (9.16) of Klein-Gordon equations will be equivalent to the trilinear
relations (12.11)(12.13) corresponding to the Lagrangians (9.7)(9.9). But, at present, it
seems that correct description of the real physical world is given by (14.21), not by the more
general trilinear equations mentioned; in particular, there are indications that the so-called
parafields, satisfying trilinear equations similar to (12.13), do not exist in the Nature [33].
Similarly to the said at the end of Sect. 6, the considerations in the present section
naturally lead to the operator-valued distribution character of the field variables and hence
of the creation/annihilation operators. However, such a rigorous treatment is out of the
range of this book, in which it will be incorporated in the appearance of Dirac delta function
in some formulae.
Ending this section, we want to say that, due to the above considerations, the La-
grangian (9.9) is the best one for the correct description of arbitrary, neutral or charged,
free scalar field.

The arguments, leading to a correct definition of a vacuum and the need of normal order-
ing of compositions (products) of creation and/or annihilation operators, are practically the
same as in the Hermitian case, studied in Sect. 7. Without repeating them mutatis mutan-
dis, we shall point only to the difference when the field 0 is non-Hermitian (charged).
There are two of them: (i) since in this case we have two types of annihilation operators,

viz. 0 (k) and 0 (k), the condition (7.3) should be doubled, i.e. to it one must add the
equality 0 (k)( X0 ) = 0, X0 being the vacuum (state vector); (ii) as now the field pos-
sesses a non-vanishing charge operator, the combinations of (13.2) with the commutation
relations (14.21) leads to infinities, like (7.4), for the momentum operator of a Hermitian
3
The particular additional conditions mentioned above are, in fact, equivalent to postulating quantization
via commutators in the case of free scalar field if one starts from someone of the Lagrangians (9.2), (9.7),
and (9.8).
4 And also the Lagrangian (9.2) which is two times (9.7) and is, usually, used in the literature.
field.1
Thus, arguments, similar to the ones in Sect. 7, lead to the following definition of a
vacuum for a free arbitrary scalar field.
Definition 15.1. The vacuum of a free arbitrary scalar field 0 is its physical state that
contains no particles and possesses vanishing 4-momentum, (total) angular momentum and
(total) charge. It is described by a state vector, denoted by X0 (in momentum picture) and
called also the vacuum (of the field), such that:
X0 = 0 (15.1a)
X0 = X0 (15.1b)

0 (k)( X0 ) = 0 (k)( X0 ) = 0 (15.1c)
X0 | X0 = 1. (15.1d)
As we said above, the formulae (12.3) and (13.2), together with the commutation re-
lations (14.21), imply senseless (infinity) values for the 4-momentum and charge of the
vacuum. They are removed by redefining the dynamical variables, like the Lagrangian, mo-
mentum operator and charge operator, by writing the compositions (products) of the field,
and/or creation and/or annihilation operators in normal order, exactly in the same way as
described in Sect. 7. Besides, the definition of the normal ordering operator (mapping)
N is also retained the same as in Sect. 7, with the only remark that now it concerns all
creation/annihilation operators, i.e. (k) and 0 (k).
0 +
Since the evident equalities N 0 (k) 0 (k) = N 0 (k) + (k) = +0 (k)
+ +
+
0
0 (k) and N 0 (k) 0 (k) = N 0 (k) 0 (k) = 0 (k) 0 (k) hold, the three
momentum operators (12.3) transform, after normal ordering, into a single momentum op-
erator, viz. into the operator

1
P = k | {0 + (k) +
0 (k) + 0 (k) 0 (k)} d k .
3
(15.2)
1 + (0 ) k0 = m2 c2 +k2
Similarly, the three charge operators (13.2) transform, after normal ordering, into a single
charge operator, viz. the operator
3
+
Q =q 0 (k) +
0 (k) 0 (k) 0 (k) d k . (15.3)
One can verify the equations

+ + +

N
0 (k) A 0 (k) = N 0 (k) A 0 (k) = 0 (k) A as (k)

N 0 (k) A + + +
0 (k) = N 0 (k) A 0 (k) = as (k) A as (k),
with A = k k . As a result of them the three angular momentum operators (13.16) trans-
form, after normal ordering, into a single orbital angular momentum operator given by
1 Applying (12.3) and (14.21), the reader can easily obtain the versions of (7.4) for a non-Hermitian field.
The results will be senseless infinities, like the ones in (7.4), which are removed via normal ordering (vide
infra).

i
L = x P x P + d k 3
0 + (k)
k k 0 (k)
2(1 + (0 )) k k

+ 0 (k) k k 0 (k) 2 2 2
+
(15.4)
k k k0 = m c +k
where P is given by (15.2). This equation, in Heisenberg picture and expressed via the
Heisenberg creation and annihilation operators (12.5), reads

i
L = 3
d k 0 + (k)
k k
0 (k)
2(1 + ( 0 )) k k

+ 0 (k) k k 0 (k) 2 2 2 . (15.5)
+
k k k0 = m c +k
In a case of neutral (Hermitian) scalar field, when 0 = 0 and (0 ) = 1, the last expression
for the orbital angular momentum operator reproduces the one presented in [11, eq. (3.54)],
due to the first equality in (13.7).
Exercise 15.1. Applying (15.4), (13.7), (12.6) and (14.21), verify the equations
[0 , L ] = x [0 , P ] x [0 , P ] [0 , L] = x [0 , P ] x [0 , P ] , (15.6)
which in Heisenberg picture respectively read
[ (x), L] = ih(x x ) (x) [ (x), L ] = i(x x ) (x). (15.7)
These equations, together with (I.5.30b), express the relativistic covariance of the theory
considered [2]. In this way a proof of the Heisenberg relations (I.4.3), (I.6.14), (I.6.30)
and (I.6.31) for scalar fields is given in the framework of the Lagrangian formalism.
Exercise 15.2. Applying (15.3), (12.6) and (14.21), obtain the equations
[0 , Q ] = q0 [0 , Q ] = q0 , (15.8)
which in Heisenberg picture take the form (9.29). Analogously prove that

[
0 (k), Q ] = q0 (k) [0 (k), Q ] = q0 (k), (15.9)
which, evidently, entail
[ P , Q ] = 0 [ Q , Q ] = 0 [ L , Q ] = 0. (15.10)
These results confirm equations like (I.4.5), and (11.6).
At last, we shall derive the commutation relations between the components of the or-
bital angular momentum operator (15.4), which coincides with the total angular momentum
operator. To simplify the proof and to safe some space, we shall work in Heisenberg picture
and employ the Heisenberg creation and annihilation operators (12.5), which satisfy the
same commutation relations as their momentum picture
counterparts. At first, we notice
that (15.5) and (14.21) imply the equations (p0 := m c + p2 )
2 2

[ (p), L ] = i p p 0 (p)
0
p p

(15.11)
[ 0 (p), L ] = i p p 0 (p).
p p
If ( ) denotes antisymmetrization with respect to the indices and , i.e. a subtraction

of the previous terms combined with the change and +( 0 (p) 0 (p)) means

that one has to add the previous terms by making the change 0 (p) 0 (p), then we
get by applying (15.5), (15.11) and (15.4):
+ (k) (k)
i
[ L , L] = 2i d3 k k 0 k 0
2(1 + ( ) k k
0 + (k)
0 (k)
k k + ( (k) (k)) ( ) ( )
k k 0 0

i (k) (k)
= i d3 k 0 + (k)k 0 0 + (k)k 0
(1 + ( ) k k

+ ( 0 (k) 0 (k)) ( ) ( )

= i L ( ) ( ) = i L ( ) ( ) ,

where k0 := m2 c2 + k 2 in the integrands, the terms containing derivatives with respect to
k were integrated by parts and the antisymmetries relative to and and and were
taken into account. The explicit form of the result obtained is:

[ L , L ] = i L L L + L , (15.12)
which in momentum picture reads

[ L , L ] = i L L L + L . (15.13)
It should be noted the minus sign in the multiplier i in the r.h.s. of (15.12) relative to a
similar one in the last equation in [11, eqs. (3.51)] for a neutral scalar field.
The equations (15.6)(15.13) are valid also before the normal ordering is performed,
i.e. if the orbital angular momentum, momentum and charge operators are replaced with
any one of the corresponding operators in (13.16), (12.3) and (13.2), respectively.
We would like to emphasize, equation (15.12) (or (15.13)) is a consequence of (15.5)
and (15.11), which is equivalent to (15.7), and this conclusion is independent of the validity
of the commutation relations (14.21) and/or the normal ordering (before normal ordering
equation (15.12) follows from (15.11) and (13.5)). Similar result concerns equations (15.8)
and (15.10).
So, we see that, at the very end of building of the theory of free scalar fields, all of the
three Lagrangians (9.7)(9.9) lead to one and the same final theory.2 This is a remarkable
fact which is far from evident at the beginning and all intermediate stages of the theory.
Acting with the operators (15.2), (15.3) and (15.4) on the vacuum X0 , we get
P ( X0 ) = 0 Q ( X0 ) = 0 L ( X0) = 0 (15.14)
which agrees with definition 15.1 and takes off the problem with the senseless expressions
for the 4-momentum, charge and orbital angular momentum of the vacuum before redefin-
ing the dynamical variables via normal ordering.
As a result of the above uniqueness of the momentum operator after normal ordering,
the three systems of field equations (12.11)(12.13), together with the conditions (12.14),
transform after normal ordering into the following unique system of equations:3
+
[ +
0 (k), 0 (q) 0 (q) + 0 (q) 0 (q)]
(15.15a)
(1 + (0 ))
0 (k) (q k) = f (k, q)
3
[0 (k), 0 +(q) +
0 (q) + 0 (q) 0 (q)]
(15.15b)
(1 + (0 ))0 (k)3 (q k) = f (k, q)

q | q
f (k, q) d q =
3
q |q =m2 c2 +q2 f (k, q) d3 q = 0. (15.15c)
0= m2 c2 +q2 0
Applying (6.4) with = 1, one can verify that (15.15) are identically valid due to the com-
mutation relations (14.21). In this sense, we can say that the commutation relations (14.21)
play a role of field equations with respect to the creation and annihilation operators (under
the hypotheses made in their derivation).
16. State Vectors

A state vector of a free arbitrary scalar field is, of course, given via the general formula (8.1)
in which, now, the momentum operator P is given by (15.2). This means that the evolution
operator U (x, x0 ) is
1 x x
U (x, x0) = exp 0
k | 2 2 2
i 1 + (0 ) k0 = m c +k

{0 + (k)
0 (k) + +
0 (k)
0 (k)} d 3
k . (16.1)
A state vector of a state with fixed 4-momentum is, of course, described by (8.4).
Similarly to the neutral field case, the amplitude, describing a transition from an initial
state Xi (xi ) to final state X f (x f ), is (8.7) and admits the representation (8.8) through the
2 Recall (see Sect. 14), the Lagrangians (9.7) and (9.8) require an additional hypothesis, like assump-
tion 14.3, for the establishment of the commutation relations (14.21) and, in this sense the arising from them
theory is not equivalent to the one build from the Lagrangian (9.9).
3 The normal ordering must be applied only to the anticommutators in (12.11)(12.13) as these terms orig-
inate from the corresponding momentum operators (12.3) before normal ordering.
S-matrix U (xi, x f ). The expansion of the exponent in (16.1) into a power series results in
the following series for U (xi, x f ) (cf. (8.9) and (8.10))

U (xi, x f ) = idF + U (n)(xi , x f ) (16.2)
n=1
1 1
U (n)(xi , x f ) := (x1 xf 1 ) . . .(xi n xf n )
n! i(1 + (0 )) n i

(1) (n) (1)
d3 k(1) . . . d3 k(n)k1 kn 0 + (k(1)) (1) + (1)
0 (k ) + 0 (k ) 0 (k )
(n)
0 + (k(n) ) (n) + (n)
0 (k ) + 0 (k ) 0 (k ) (16.3)

(a)
where k0 = m2 c2 + (k (a))2 , a = 1, . . ., n.
According to (8.4) and the considerations in Sect. 11, a state vector of a state containing
n particles and n antiparticles, n , n 0, such that the i th particle has 4-momentum pi
and the i th antiparticle has 4-momentum pi , where i = 0, 1, . . ., n and i = 0, 1, . . ., n , is
given by the equality
X (x; p1; . . .; pn ; p1 ; . . .; pn )

1 1 n
1 n
(x x0 ) (pi ) + (x x0 ) (pi )

= exp
n !n ! i i =1 i i =1
+
0 (p1 ) 0 (pn ) 0 (p1 ) 0 + (pn ) ( X0 ), (16.4)
+ +
where, in view of the commutation relations (14.21), the order of the creation operators is
inessential. If n = 0 (resp. n = 0), the particle (resp. antiparticle) creation operators and
the first (resp. second) sum in the exponent should be absent. In particular, the vacuum
corresponds to (16.4) with n = n = 0. The state vector (16.4) is an eigenvector of the

momentum operator (15.2) with eigenvalue (4-momentum) ni =1 pi + ni =1 pi and is also
an eigenvector of the charge operator (15.3) with eigenvalue (q)(n n ).1
Exercise 16.1. Using (14.21) and (11.2), prove that the transition amplitude between two
states of a charged field, like (16.4), is:
X (y; q1; . . .; qn ; q1 ; . . .; qn )| X (x; p1 ; . . .; pm ; p1 ; . . .; pm )
1 1 n
1 n
= m n m n exp (x y ) (pi ) + (x y ) (pi )

n !n ! i i =1
i i =1
3 (pn qi )3 (pn 1 qi ) . . .3 (p1 qi )
1 2 n
(i1 ,...,in )
3 (pn qi )3 (pn 1 qi ) . . .3 (p1 qi ) (16.5)

1 2 n
(i1 ,...,in )
where the summations are over all permutations (i1 , . . ., in ) of (1, . . ., n) and (i1 ,
. . ., in ) of (1, . . ., n ).
1 Recall (see Sect. 11), the operator +
0 (k) creates a particle with 4-momentum k and charge q, while

+
0 (k) creates a particle with 4-momentum k and charge +q, where, in the both cases, k0 = m2 c2 + k 2 .
The conclusions from this formula are similar to the ones from (8.16) in Sect. 8. For
instance, the only non-forbidden transition from an (n -particle + n -antiparticle)-state is
into (n -particle + n -antiparticle)-state; the both states may differ only in the spacetime
positions of the (anti)particles in them. This result is quite natural as we are dealing with
free particles/fields.
In particular, if Xn denotes any state containing n particles and/or antiparticles, n =
0, 1, . . ., then (16.5) says that
Xn | X0 = n0 , (16.6)
which expresses the stability of the vacuum.
Example 16.1. Consider the one (anti)particle states + (p)( X0 ) and 0 + (p)( X0 ). Apply-
0
ing (15.2), (15.3), (15.4) and (14.21), prove that (p0 := m2 c2 + p2 ):2

P + +
0 (p)( X0 ) = p 0 (p)( X0 ) Q + +
0 (p)( X0 ) = q0 (p)( X0 )
(16.7)
P 0 + (p)( X0) = p 0 + (p)( X0) Q 0 + (p)( X0) = +q0 + (p)( X0 )

L (x) +0 (p)( X0 )
+
= (x p x p )i p p 0 (p)( X0 )
p p
+ (16.8)
L (x) 0 (p)( X0 )
+
= (x p x p )i p p 0 (p)( X0 ) .
p p
+
These results agree completely with the interpretation of +
0 (k) and 0 (k) as operators
creating one (anti)particle states.
17. Conclusion
The main results of this chapter, dealing with a study of free Hermitian or non-Hermitian
scalar fields, may be formulated as follows:
The creation and annihilation operators in momentum representation in momentum pic-

ture are (up to constant phase factor and, possibly, normalization) identical with the
known ones introduced in momentum representation in Heisenberg picture.
The quantization with commutators, not by anticommutators, is derived from the field
equations (in momentum picture) without involving the spin-statistics theorem (or
other equivalent to it additional condition), if one stars from a suitable Lagrangian.
The (system of) field equation(s) in terms of creation and annihilation operators is derived.
It happens to be similar to a kind of paracommutation relations.1
2In Heisenberg picture and in terms of the Heisenberg creation/annihilation operators, in equations (16.8)
the terms proportional to (x p x p ) are absent and tildes over all operators should be added. Equa-
tions (16.7) remain unchanged in Heisenberg picture (in terms of the corresponding Heisenberg operators).
1 More precisely, in chapter V we shall show how the parabose-commutation relations for free arbitrary
scalar field can be derived from the Klein-Gordon equations in terms of creation and annihilation operators.
An analysis of the derivation of the standard commutation relations is given. It is shown

that, under some explicitly presented conditions, they are equivalent to the (system
of) field equation(s) and are not additional to it conditions in the theory.
The description of free scalar fields via operators in Hilbert space contradicts to the as-
sumption that the equations (I.5.30) hold as additional restrictions on the Lagrangian
formalism. This problem is resolved by assuming that the field variables are operator-
valued distributions.
Explicit expressions for the conserved operators, 4-momentum, charge and (orbital or
spin) angular momentum, are obtained.
Chapter III
Free Spinor Fields
Free spinor fields, with

spin 1/2, are explored in details
in the momentum picture of motion in La-
grangian quantum field theory. The field equa-
tions are equivalently written in terms of creation
and annihilation operators and on their base the
anticommutation relations are derived. Some
problems concerning the vacuum and state
vectors of free spinor field are discussed.
Several Lagrangians, describing free
spinor fields, are considered and
the basic consequences
of them are in-
vestigated.

1. Introduction
The chapter is devoted to a detailed investigation of Lagrangian quantum field theory of
a free spinor (spin 12 ) field in momentum picture, introduced in chapter I. It is a direct
continuation of chapter II where similar exploration of free scalar fields was carried out,
and, respectively, most of the methods of chapter II will be mutatis mutandis applied to
problems of free spinor fields. Most of the known fundamental results are derived in a new
way (and in a slightly modified form), but the work contains and new ones; e.g. the field
equations in terms of creation and annihilation operators and a uniform consideration of the
massive and massless cases.
The layout of this chapter, which is based on [40], is as follows.
The description of free spinor field in momentum picture is given in section 2. In Sect. 3,
the (system of) Dirac equation(s) describing a free spinor field is analyzed in terms of op-
erators which, possibly, up to normalization and pure phase factor, are identical with the
Fourier coefficients of the field in Heisenberg picture. From these operators, in Sect. 4, is
constructed a set of operators which admit interpretation as creation and annihilation opera-
tors. Then, the last operators are expressed via a new set of operators with the same physical
interpretation, which operators, up to a phase factor, coincide with the creation/annihilation
operators known from the investigation of free spinor field in Heisenberg picture. The fields
dynamical variables, i.e. the momentum, charge and angular momentum operators, are ex-
pressed via the creation and annihilation operators in Sect. 5, which results in expressions
similar to ones in the momentum representation in Heisenberg picture.
In Sect. 6, the equations of motion for a free spinor field are equivalently written as a
system of operator equations in terms of creation and annihilation operators. These equa-
tions are trilinear ones and their form is similar to the one of the parafermi relations. The
obtained system of equations is analyzed in Sect. 7, where from it, under some explicitly
presented additional conditions, the anticommutation relations for the creation and annihi-
lation operators are derived.
The concept of a vacuum for a free spinor field is introduced in Sect. 8. Some problems
in the theory are pointed and their solution is described via the introduction of normal
ordering of products (compositions) of creation and annihilation operators. In Sect. 9 are
discussed some general aspects regarding state vectors of free spinor field.
In Sect. 10 are investigated different Lagrangians, which do not differ by a full 4-diver-
gence, from which the quantum theory of free spinor fields can be derived. They and their
consequences are compared from different positions and the best one of them is pointed
out. It is the one which is charge symmetric; so that in it is encoded the spin-statistics
theorem for free spinor fields.
Section 11 closes the chapter.
2. Description of Free Spinor Field in Momentum Picture

A spinor field of spin 12 is described by four operators which are collected in a
matrix operator considered as a 4-component column, i.e. (x) := 0 (x), 1 (x),

2 (x), 3 (x) , where is the sign of matrix transposition. In Heisenberg picture, the
Free Spinor Fields 83
theory of free spinor field with mass parameter m is derived from the Lagrangian [1]
1
1
L = ic{ (x) ( (x)) ( (x)) (x)} mc2 (x) (x). (2.1)
2
Here: are the Diracs matrices (see equations (2.2), (2.17), and (2.27)
below), is the
composition of mappings/operators sign, and := 0 , with := 0 , 1 , 2 , 3 ,
is the Dirac conjugate spinor to . Besides, in expressions like and 0 a matrix
multiplication is understood, i.e., in these examples, we have := and

0 := 00 , . . ., 03 .
As a particular realization of the Diracs gamma matrices, we shall use the following
one [1, 4]:

1 0 0 0 0 0 0 1
0 1 0 0 0 1 0
0 = 1 = 0
1 0 1 0 0 1 0 0
1 0 0 1 1 0 0 0
(2.2)
0 0 0 i 0 0 1 0
0 0 i 0 0 0 0 1
2 =
0 i 0 0 = 1 0 0 0 .
3
i 0 0 0 0 1 0 0
These matrices satisfy the basic equation (2.17) below as well as all of subsequent
relations for the -matrices that will appear further in this book.
In momentum picture, in view of (I.5.20), the spinors components and their Dirac
conjugate transform into
(x) = U (x, x0 ) (x) U 1 (x, x0 ) = (x0 ) = (x0 ) =:

(2.3)
(x) = U (x, x0 ) (x) U 1 (x, x0 ) = (x0 ) = (x0 ) =: .
If we identify U (x, x0 ) with U (x, x0 )114 , where 114 := diag(1, 1, 1, 1, ) is the unit (iden-
tity) 4 4 matrix, the last equalities can be rewritten as
(x) = U (x, x0 ) (x) U 1 (x, x0 ) = (x0 ) = (x0 ) =:

1
(2.3 )
(x) = U (x, x0 ) (x) U (x, x0 ) = (x0 ) = (x0 ) =: .
Regardless of the explicit dependence of and on the point x0 M, further it will

not be indicated as insignificant for our present chapter. By virtue of (I.6.2), the La-
grangian (2.1) in momentum picture reads
1
L = L (, , y, y ) = c{ [, P ] [, P ] } mc2 . (2.4)
2
1 The theory of free spinor field can be derived also from the Lagrangian L = ic mc2 , as
it is done in [2,5]. The problem of a selection of Lagrangian for free spinor field will be considered in Sect. 10.
where
1 1
y := [, P ] := [0 , P ] , . . ., [3 , P ]
i i
1 1
y := [, P ] := [0 , P ] , . . ., [3 , P ] .
i i
(Notice, to simplify the notation, here, as we did above with U (x, x0 ), we identify P with
the matrix operator P 114 . For the same reason, we identify with the matrix operator
idF , F being the systems Hilbert space of states.) Therefore, we have: 2
L L 1 L 1
= = c[, P ] mc2 = = + ic
2 y 2
(2.5)
L L 1 L 1
= = + c [, P ] mc2 = = ic.
2 y 2
Consequently, the equations of motion (I.6.3) now read:
[, P ] mc = 0 [, P ] + mc = 0 (2.6)
where we have applied the equality
[ , P ] = 0 (2.7)

as [ , P ] := [ idF , P 114 ] = [[ idF , P ] ]3,=0 0 since are constant complex
numbers. The first of the equations (2.6) is the Dirac equation in momentum picture and
the second one is its Dirac conjugate. They correspond to the famous equations
i mc = 0 i + mc = 0 (2.8)
in Heisenberg picture. It is worth noting, the equations (2.6) are valid in any picture of mo-
tion which cannot be said with respect to (2.8). Indeed, since the transitions to an arbitrary
picture of motion is achieved via a unitary operator U : F F [19] and
[ , U ] = 0 (2.9)
(see the proof of (2.7) above), this statement is a consequence of the algebraic structure
of (2.6).
We shall comment on the choice of the Lagrangians (2.1) and (2.4) in Sect. 10.
We emphasize, the fields and will be treated as independent field variables in this
chapter. However, because of the connection = 0 , most of the relations regarding
can be obtained from the ones concerning by Dirac conjugation, i.e. via the change
( )
( ) := ( ) 0 , where the dots stand for any spinor-matrix operator expression.
2 As pointed on page xii, the calculation of the derivatives in (2.5) is not quite correct mathematically. How-
ever, the field equations (2.6) and the formulae (2.10)(2.13) below are correct. For their rigorous derivation,
see [3] and section I.3.
It is well known, a spinor filed possesses energy-momentum, electric charge and angular
momentum. In Heisenberg picture, the densities of these conserved operator quantities
respectively are:3
1
T = ( ) + ( ) = ic{ ( ) ( ) } (2.10)
2
q
J = { } = qc (2.11)
i

M = x T x T +

S (2.12)

where the spin angular momentum operator density S is4
1

S =+ { (I ) + ( I) } = c { + } ,
4
(2.13)
j=1,2
with
i
:= ( ), (2.14)
2
and the coefficients I and I , characterizing the transformation properties of and
under 4-rotations, are [1, 5].
1 1
I = i I = + i . (2.15)
2 2
Often the spacial components of S
0 , defining the spin angular momentum via
(I.2.8b), are combined into the (pseudo-)vector 12 eabc Sbc

0
with a, b, c = 1, 2, 3. Equa-
tion (2.13) implies
1 1 1
S a := eabc Sbc
0
(x) = + c (x)a (x) = c (x)a (x) (2.16)
2 2 2
where a := eabcbc and the relation
+ = 2 114 (2.17)
was used. 5 So, the vectorial spin operator S = ( S 1 , S 2 , S 3 ), appearing in equations (I.6.11)
and (I.6.12), is

1
S(x) = (x) (x) d3 x := (1 , 2 , 3 ). (2.18)
2
x0 =const
3 See, for instance, [1] or [11]. However, the order of the operators in the compositions below is not
proved in these books, there are only some arguments justifying the made choice. For a rigorous proof
see [3, section 5.5],.
4 We adopt the definition of spin angular momentum operator density from [4, 78, eq. (13.47)] or [11,
sect. 3.3.1, eq. (3.115)]. It agrees with the commutation relation (I.6.17). In [1, eqs. (2.16) and (7.31)] is used
a definition with an opposite sign.
5 For some general properties of the -matrices, see, e.g., [5, 12].
In accord with (I.5.4) and (I.5.22), the dynamical characteristics (2.10)(2.13) in mo-
mentum picture are
1
T = c{ [, P ] [, P ] } (2.19)
2
J = qc (2.20)

M = x T T
x +

S (2.21)
1
S = c { + } . (2.22)
4
Exercise 2.1. Applying the facts presented in section I.3, rigorously derive the Dirac equa-
tions (2.6) and prove that the correct expressions for the energy-momentum tensor, charge
current and spin angular momentum density are (2.19), (2.20) and (2.22), respectively.
In particular, the vector (2.16) takes the form

1 1
S a := eabc Sbc
0
= c a = c a (2.23)
2 2
in momentum picture.
To specify the relations (I.6.29) and (I.6.30), which are external for the Lagrangian
formalism, for a spinor field, we, by convention, put
() = +1 () = 1 (2.24)
and get
[, Q ] = q [, Q ] = q (2.25)
1
[, M (x, x0 )] = x [, P ] x [, P ] + (2.26a)
2
1
[, M (x, x0 )] = x [, P ] x [, P ] (2.26b)
2
due to (2.15). The last terms in (2.26) have their origin in the spin angular momentum and
the remaining ones are due to the orbital angular momentum.
Besides (2.17), the gamma matrices are suppose to satisfy the conditions (0 ) = 0 and
( ) = a for a = 1, 2, 3, i.e. (do not sum over !)
a
:= ( ) = = 0 0 . (2.27)
As a consequence of (2.14) and (2.27), we have (do not sum over and !)
( ) = . (2.28)
Important corollaries from (2.14) and (2.17) are the equalities (do not sum over and
!)
[ , ] = 2i( ) (2.29)
0 0
= , (2.30)
the last of which implies6
0 0a + 0a 0 = 0 0 ab = ab 0 . (2.31)
Combining (2.28) and (2.30), we get
( ) = 0 0 . (2.32)
As result of (2.22) and (2.31), we have

1
S0a
0
=0 Sab
0
= c 0 ab . (2.33)
2
From here, we see that the vector S a and the components S 0 carry one and the same
information. Notice, the equations (2.33) are specific for the Lagrangian (2.1); for instance,
the first of them does not hold for the Lagrangian (10.3) on page 131 see [2, eq. (13.47)].
According to the Dirac equation (2.6), the operators
1 1
M :
+ [, P ] M :
[, P] , (2.34)
c c
when acting on the solutions of (2.6), have the mass parameter m of the spinor field as
eigenvalue. Applying the relations (2.17), the Jacobi identity and (I.4.4), we see that
2 1
M 2 := M M = M := M M :
[[, P ] , P ] = , (2.35)
c2
has a meaning of a square-of-mass operator as its eigenvalues on the solutions of (2.6) are
equal to m2 . Therefore the operator c12 [[(), P ] , P ] , rather than P P , determines the
square of the mass (parameter) of a free spinor field. Obviously, this conclusion is valid in
any picture of motion. This situation is similar to the one with free scalar fields (see the
comments after equation (II.2.18)). It is well known, the operator P P has a sense of a
square-of-mass operator for the fields states (on which it acts).
3. Analysis of the Dirac Equation(s)

The Dirac equations (2.6), together with the equality (I.6.6) and the explicit expres-
sion (2.10) for the energy-momentum operator, form a complete algebraic-functional sys-
tem of equations for determination of the spinor field and its Dirac conjugate = 0
(and its momentum operator P too). This situation is similar to the one concerning a free
charged scalar field, investigated in chapter II; respectively similar results hold for the struc-
ture of spinor field operators. Moreover, there is now a simplification due to the linearity
of the Dirac equations (2.6) relative to the momentum operator, contrary to the scalar field
case, when the corresponding Klein-Gordon equations (II.9.16) are quadratic in momentum
operator.1
6The equations (2.30) and (2.31) are also corollaries from (2.29) with = 0 and 0a = i0 a = ia 0 .
1In Heisenberg picture, the last statement is equivalent to the one that Dirac and Klein-Gordon equations
are respectively first and second order partial differential equations.
To begin with, we single out the degenerate solutions
[, P ] = 0 [, P ] = 0 for m = 0 (3.1)
of the Dirac equations (2.6), which solutions, in view of (I.5.14), (I.5.16) and (I.5.18), in
Heisenberg picture read
(x) = (x0 ) = (= const) (x) = (x0 ) = (= const) for m = 0. (3.2)
According to equations (2.19)(2.22) (see also (2.31)), the energy-momentum, charge

and (spin) angular momentum operators for the solutions (3.1) respectively are:
T = 0 (3.3)
J = qc (3.4)
1
M = S

= c( + ) . (3.5)
4
Since (3.3) and (I.6.6) imply
P = 0, (3.6)
the equalities (3.4) and (3.5) hold in momentum and Heisenberg pictures for any constant
spinors and . Thus, the solutions (3.1) (or (3.2) in Heisenberg picture) describe a mass-
less spinor field with vanishing energy-momentum characteristics which, however, carries,
generally, a non-zero (if = 0) charge and spin angular momentum. It seems that such
spinor fields have not been observed until now.2
Now, we are ready to describe the general structure of the solutions of the Dirac equa-
tions (2.6).
Proposition 3.1. The common solutions of the Dirac equations (2.6) and equa-
tions (I.5.30), with i = , , can be written as

= d3 k f+ (k)(k) + f (k)(k) (3.7a)
k0 =+ m2 c2 +k2 k0 = m2 c2 +k2

= d3 k f + (k)(k) + f (k)(k) (3.7b)
k0 =+ m2 c2 +k 2
k0 = m2 c2 +k 2
or, equivalently as

= d k(k m c ) f (k)(k) =
4 2 2 2
d4 k(k2 m2 c2 ) f (k)(k). (3.8)
Here: k = (k0 , k1 , k2 , k3 ) is a 4-vector with dimension of 4-momentum, k2 = k k = k02

k12 k2 k32 = k02 k2 with k being the components of k and k := (k1 , k2 , k3 ) = (k1 , k2 , k3 )
2 This situation is completely different from a similar one, when free scalar fields are concerned; in the last
case, solutions like (3.1) are unobservable in principle as all their dynamical characteristics vanish see [26]
and chapter II.
being the 3-dimensional part of k, () is the (1-dimensional) Dirac delta function, the
spinor operators (k), (k) : F F are solutions of the equations
[(k), P ] = k (k) [(k), P ] = k (k) (3.9a)

{( k + mc114 )(k)} 2 k =m2 c2
=0 {(k)( k + mc114 )}k2 =m2 c2 = 0,

(3.9b)
f (k) and f (k) are complex-valued functions (resp. distributions (generalized functions))
of k for solutions different from (3.1) (resp. for the solutions (3.1)), and f and f are com-
plex-valued functions (resp. distributions) of k for solutions different from(3.1) (resp. for
the solutions (3.1)). Besides, we have the relations f (k)| =2 m2 c2+k 2 f (k)
k 0 = m 2 c2 +k2
and f (k)| =2 m2 c2+k2 f (k) for solutions different from (3.1).

k0= m c +k
2 2 2
Remark 3.1. Evidently, in (3.7) and (3.8) enter only the solutions of (3.9) for which
k2 := k k = k02 k 2 = m2 c2 . (3.10)
This circumstance is a consequence of the fact that the solutions of (2.6) are also solutions
of the Klein-Gordon equation, viz.
m2 c2 [[, P ] , P ] = 0 m2 c2 [[, P ] , P ] = 0. (3.11)
To prove this, one should apply to the first (resp. second) equation in (2.6) the operator
[(), P ] + mc114 idF () (resp. [(), P ] mc114 idF ()) and, then, to apply the Jacobi
identity, (I.4.4) and (2.17).
Remark 3.2. Obviously, the solutions (3.1) correspond to (3.9a) with P = 0. Hence
(x, 0) = (0) = const (x, 0) = (0) = const P = P = 0 (3.12)
with (see (I.5.4))
(x, k) := U 1 (x, x0 )(k) U (x, x0 ) (x, k) := U 1 (x, x0 )(k) U (x, x0 ). (3.13)
These solutions, in terms of (3.7) or (3.8), are described by m = 0 and, for example,
f (k) = f (k) = ( 12 a)3 (k) for some a C or f (k) = f (k) such that f (k)|k0=|k| =
(1 2a)|k|3 (k), respectively. (Here 3 (k) := (k1 )(k2 )(k3 ) is the 3-dimensional Dirac
delta-function.) To prove that, use the equality (y2 b2 ) = 1b ((y + b) + (y b)) for
b > 0 [28, eq. 21.9-13]. In that case, the equations (3.9a) and (3.9b) reduce respectively to
k (k) = 0 k (k) = 0 (3.14a)

k (k) = 0 (k) k = 0 (3.14b)
as in it m = 0. Since the rank of the matrix k is equal to two for k2 = m2 c2 = 0 but

k = 0 (see below the beginning of the paragraph containing (3.19)), the equations (3.14)
have non-vanishing solutions in this case only if distributions are taken into account.
Remark 3.3. Since = 0 , from (3.7) (resp. (3.8)) is clear that

there should exist some
connection between f (k)(k) and f (k)(k) with k0 = + m2 c2 + k 2 (resp. between
f (k)(k) and f (k)(k)). A simple examination of (3.7) (resp. (3.8)) reveals that the Dirac
conjugation can either transforms these expressions into each other or changes the signs
plus and minus in them according to:

f(k)(k)
2 = f (k)(k) . (3.15a)
k0 = m c +k
2 2 k0 = m2 c2 +k2

f (k)(k)
2 = f (k)(k) (3.15b)
k0 = m c +k
2 2 2 2 2
k0 = m c +k

f (k)(k) = f (k)(k) (3.16a)

f (k)(k) = f (k)(k). (3.16b)
From the below presented proof of proposition 3.1 and the comments after it, it will be clear
that (3.15) and (3.16) should be accepted. Notice, the above equations mean that (k) is
not the Dirac conjugate of (k).
Proof of proposition 3.1 The proposition was proved for the solutions (3.1) in remark 3.2.
So, below we suppose that (k, m) = (0, 0).
The equivalence of (3.7) and (3.8) follows from (y2 b2 ) = 1b ((y + b) + (y b)) for
b > 0.
Since the solutions and of the Dirac equations (2.6) are also solutions of
the Klein-Gordon equations (3.11), the representations (3.7) and the equalities (3.15)
and (3.16), with (k) and (k) satisfying (3.9a), follow from proposition II.3.1 (or propo-
sition II.10.1) describing the structure of the solutions of the Klein-Gordon equation in
momentum picture.3
At the end, inserting (3.7) or (3.8) into (2.6), we obtain the equations (3.9b) due
to (3.9a).
From the proof of proposition 3.1, as well as from the one of proposition II.3.1 (or
proposition II.10.1) the following two conclusions can be made. On one hand, the con-
ditions (3.9a) ensure that (3.7) and (3.8) are solutions of (I.4.1) and the Klein-Gordon
equations (3.11), while (3.9b) single out between them the ones satisfying the Dirac equa-
tions (2.6). On the other hand, since up to a phase factor and, possibly, normalization
constant, the expressions f (k)(k) and f (k)(k) coincide with the Fourier images of re-
spectively (x) and (x) in Heisenberg picture, we can write

x )k
(k2 m2 c2 )(k)ei (x
1
= (k2 m2 c2 )(k) d4 k (x) = 0 d4 k (3.17)
and similarly for (with (k) = (k)), where (k) are suitably normalized solutions
of (3.9). Therefore, up to normalization factor, the Fourier images of (x) and (x) are
1 1
(k) = e i x0 k (k) (k) = e i x0 k (k) (3.18)
3
One can prove the representations (3.7), under the conditions (3.9), by repeating mutatis mutandis the
proof of proposition II.3.1. From it the equalities (3.15) and (3.16) rigorously follow too.
where x0 is a fixed point (see section I.5). So, the momentum representation of free spinor
field in Heisenberg picture is an appropriately chosen operator base for the solutions of
the Dirac equation in momentum picture. This conclusion allows us freely to apply in
momentum picture the existing results concerning that basis in Heisenberg picture.
The equations (3.9b) are well known and explicitly solved in the textbooks where the
Dirac equation, in momentum representation of Heisenberg picture, is explored [1, 4] (for a
summary, see, e.g., [5, asppendix 1] or [1, appendix 3A]). Here are some facts about them,
which we shall need further in the present chapter; for references see loc. cit.
Working in a representation in which 0 is diagonal (see (2.2)), bya direct calculation,
one can prove that the rank of the matrices k + mc114 , with k0 = m2 c2 + k2 , is equal
to 0, if (k, m) = (0, 0) and to 2, if (k, m) = (0, 0).4 Since the -matrices are defined up to
a change
O O1 , O being a non-degenerate matrix (usually taken to be unitary),
from here follows that any one of the equations (3.9b) has r linearly independent solutions,
where r = for (k, m) = (0, 0) and r = 2 for (k, m) = (0, 0). So, in these cases (3.9b) has
respectively infinitely many linearly independent solutions and two linearly independent
solutions. Since the case (k, m) = (0, 0), corresponds to the degenerate solutions (3.1),
which require different treatment, we shall exclude it from our further considerations in this
section.5 In the case m = 0, we shall label these r = 2 linearly independent solutions with
an index s taking the values 1 and 2, s = 1, 2.6 Define the operator spinors s,() (k) and

s,() (k), where the index () indicates the sign of k0 = m2 c2 + k 2 (= 0) in (3.9b), as
linearly independent solutions of respectively the equations

k + mc 1
1 4 s,()(k) = 0
k0 = m2 c2 +k 2
(3.19)
s,() (k) k
2 + mc1 14 = 0.
k0 =
2 2 m c +k
If m = 0, we set by definition s = 0 and label the linearly independent solutions of (3.9b) by

the signs + and . Respectively, for m = 0, we define the operator spinors 0,() (k)
and 0,()(k) as linearly independent solutions of (3.19) with m = 0. As a result of (3.9a),
the solutions of (3.19) satisfy also the equations

[s,() (k), P ] = k (k)
k0 = m2 c2 +k 2 s,()
(3.20)
[s,() (k), P ] = k
2 s,() (k).
k0 =
2 2 m c +k
Since any solution of the first (resp. second) equation in (3.9b) can be represented as a
linear combination of s,() (k) (resp. s,() (k)), we can rewrite (3.7) as

= d3 k fs,+(k)s,(+)(k) + fs, (k)s,()(k) (3.21a)
k0 = m2 c2 +k2
s
4 For (k,m) = (0,0) the statement is evident. Usually (see, e.g., [1]), when m = 0 (but k = 0), it is proved in
a frame in which k = 0 and a subsequent return to a general one; when k = 0 and m = 0, the same method can
be used with the only modification that a frame in which k 1 = k 2 = 0 has to be employed. In the last case, the
Dirac equation is replaced with a system of the so-called Weyl equations [1, 4].
5 About the exploration of the consequences of (3.1), see the paragraph containing equation (4.17) in Sect. 4
below.
6 Usually [1, 4], the index s is referred as the polarization or spin index (parameter, variable).

= d3 k f s,+ (k)s,(+) (k) + f s, (k)s,() (k) (3.21b)
k0 = m2 c2 +k2
s
where f s, (k) and f s, (k) are some complex-valued functions of k for solutions different
from (3.1). Regarding the solutions (3.1), in view of remark 3.2, for them (3.21) holds
too for some distributions fs, (k) and f s, (k) and some operators s,() (k) and s,()(k),
which can be chosen in different ways7
Below we shall need a system of classical, not operator-valued, suitably normalized
solutions of the equations (3.19). The idea of their introduction is to be separated the in-
variant operator properties of the spinor field from its particular matrix representation
as a collection of (operator) components which depend on some concrete reference frame
with respect to which it is studied. This will be done in Sect. 4 by representing the opera-
tor-valued quantum spinors as linear combinations of the mentioned system, in fact a basis,
of classical spinors, the coefficients of which expansion are frame-independent invariant
operators characterizing the spinor field.
Let (k, m) = (0, 0). Consider classical 4-spinors vs, (k) and their Dirac conjugate
v (k) := (vs, (k)) 0 = (vs, (k)) =: v s, (k)0 , where s = 1, 2 for m = 0, s = 0 for m = 0,
s,
and means Hermitian conjugation (i.e. matrix transposition combined with complex
conjugation in the classical case), which are linearly independent solutions of the equations
s,
k + mc11 4 v (k) = 0
k0 =+ m2 c2 +k 2
(3.22)
vs, (k) k
2 + mc1 14 = 0
k0 =+
2 2 m c +k
and satisfy the conditions

s,
v (k) =: v s, (k) vs, (k) =: vs, (k) = v s, (k)0 (3.23)
mc
v s, (k)vs , (k) = ss vs, (k)vs , (k) = ss , (3.24)
m c +k
2 2 2
where s, s = 0 for m = 0 (and k = 0) and s, s = 1, 2 for m = 0. From here the following

relations can be derived [1, subsect. 7.2]:

v s, (k)vs , (k) = 0 (3.25)

v s,
(k){k k mc }v
a b b a a b s,
(k) = 0 a, b = 1, 2, 3
ka{v s, (k)(ab b a )v s , (3.26)
(k)} = 0 a, b = 1, 2, 3
a
( k mc114 )
vs, s,
(k)v (k) =
2k0
22 2
k0 = m c +k
for m = 0. (3.27)
s
These formulae will be applied in different calculations in the next sections.

The explicit form of the spinors vs, (k) for m = 0 can be found in [5, pp. 617618] or
in [11, sect. 2.2.1], where the notation u(k; s) := vs, (k) and v(k; s) := vs,+ (k) is used. Since
( k + mc114 )( k mc114 ) = (k2 m2 c2 )114 and k2 = m2 c2 , we can write

vs, (k) = v s, (k) = A (k)( k 2 2 2 + mc114 )v0 ,
s,
(3.28)
k0 = m c +k
7 For details, see below (4.17) and the paragraph containing it. Recall, the solutions (3.1) can be describe
by (k,m, P ) = (0,0,0) in which case (3.9) (and, hence, (3.19) and (3.20)) reduce to the identity 0 = 0.
where A (k) are some normalization constants, that can be found by using (3.24),8 and vs,
0
are constant spinors given by:

v0 = (1, 0, 0, 0) v0 = (0, 1, 0, 0)
1, 2,
for m = 0 (3.29a)

v1,+
0 = (0, 0, 1, 0) v2,+
0 = (0, 0, 0, 1)

v0,
0 = (0, 0, 1, 0) v0,+
0 = (0, 0, 0, 1) for m = 0. (3.29b)
4. Frequency Decompositions
As a consequence of the results of Sect. 3, one can expect the existence of decompositions
of the constant Dirac spinors and similar to the one in Heisenberg picture [1, 4]. Such
expansions, in fact, exist and can be introduced in almost the same way as it was done for
a free charged scalar field in chapter II; the only differences being the spinor (polarization)
index s in (3.21) and the multi-component character of the spinors.
Let us set

fs, (k)s,() (k) for k0 0
s (k) :=
0 for k0 < 0
(4.1)
f s, (k)s,() (k) for k0 0
s (k) :=
0 for k0 < 0
(k) :=
s (k) (k) := s (k) (4.2)
s s

(k) := (k) 2 2 2
s (k) := s (k) k0 =
2 2 2
m c +k s s k0 = m c +k

(k) := (k) = s (k)
k0 = m2 c2 +k2 (4.3)
s

(k) := (k) 2 2 2 = s (k),
k0 = m c +k
s
where k2 = m2 c2 and (the summation is over) s = 0 for m = 0 (and k = 0) or s = 1, 2 for

m = 0.1
The equalities (3.16) imply

(k) = (k) (k) = (k) (4.4)
which mean that (k) is not the Dirac conjugate of (k).

Combining (4.1), (3.20), (3.21) and proposition 3.1, we get:
= + + = + + (4.5)

Explicitly, we have A (k) = {2mc(mc + m2 c2 + k2 )}1/2 for m = 0; see [5, Appendix A, eq. (A1-29)].
8
For m = 0, one may set A (k) = {2k2 }1/2.

1 On the case (k,m) = (0,0), or, more generally, the case regarding the solutions (3.1) vide infra the
paragraph containing equation (4.17).


:=

d 3
k
s (k) = d k (k) :=
3
d 3
k
s (k) = d3 k (k) (4.6)
s s

[ (k), P ] = k (k) k0 = m2 c2 + k 2 (4.7a)

[ (k), P ] = k (k) k0 = m2 c2 + k 2 . (4.7b)
Notice, now equations (3.9b) are incorporated in the definitions (4.1) via the equa-
tions (3.19).
To reveal the physical meaning of the operators introduced, we shall rewrite (2.25)
and (2.26) in their terms:2
[ (k), Q ] = q (k) [(k), Q ] = q (k) (4.8)
1
[ (k), M (x)] = {(x k x k ) 114 + } (k) (4.9a)
k0 = m2 c2 +k2 2
1
[ (k), M (x)] = (k){(x k x k ) 114 } (4.9b)
k0 = m2 c2 +k2 2
where (4.7) was taken into account. Recall, here Q and M stand for the charge and (total)
angular momentum operators, respectively, and the spin matrices are defined via (2.14).
Besides, the last terms in (4.9) are due to the spin angular momentum while the remaining
ones originate from the orbital angular momentum. We should remind, the equations (4.8)
and (4.9) originate from (I.4.2) and (I.4.3), which are external to the Lagrangian formalism.
Therefore the below-presented results, in particular the physical interpretation of the cre-
ation and annihilation operators, should be accepted with some reserve. However, after the
establishment of the particle interpretation of the theory (see Sections 8 and 9), the results
of this section will be confirmed (see also Sect. 10).
Let X p , Xe and Xm denote state vectors of a spinor field with fixed respectively 4-mo-
mentum p , (total) charge e and (total) angular momentum m (x), i.e.
P ( X p ) = p X p (4.10a)
Q ( Xe ) = e Xe (4.10b)
M (x)( Xm ) = m (x) Xm . (4.10c)
Combining these equations with (4.7)(4.9), we obtain3

P (k)( X p) = (p k )(k)( X p ) k0 = m2 c2 + k2 (4.11a)

P (k)( X p) = (p k )(k)( X p ) k0 = m2 c2 + k2 (4.11b)

Q ( Xe ) = (e q)( Xe ) Q ( Xe ) = (e + q)( Xe ) (4.12a)
2
To derive rigorously (4.8) and (4.9) from (2.25) and (2.26), respectively, one has to take into account that
(k) and (k) are, up to a phase factor and, possibly, normalization factor, the Fourier images of (x) and
(x), respectively (see (3.17)).
3 Expressions like (k)( X ) should be understand as a vector-columns of vectors of the form
s p

s,0 (k)( X p ),... , (k)( X p ) .
s,3

Similarly, the quantity s (k)( X p ) := s,0 (k)( X p ),... ,s,3 (k)( X p ) is a vector-row of vectors.

Q ( Xe) = (e q) ( Xe ) Q ( Xe ) = (e + q) ( Xe) (4.12b)

Q (k)( Xe ) = (e q) (k)( Xe ) Q (k)( Xe ) = (e + q) (k)( Xe) (4.12c)

M (x) (k)( Xm )
1
= {m (x)114 (x k x k ) 2 2 2 114 }(k)( Xm) (4.13a)
k0 = m c +k 2

M (x) (k)( Xm )
1
= (k)( Xm ){m(x)114 (x k x k ) 2 2 2 114 + }. (4.13b)
k0 = m c +k 2
If the field configuration happens to be such that
[, P ] = A [, P ] = A A := 0 A 0 (4.14)
for some 4 4 matrices A , then to (4.13) can be added the equations (cf. (4.12a)
and (4.12b)):

M (x) ( Xm)
1
= {m (x)114 (x A x A ) 2 2 2 114 }( Xm ) (4.15a)
k0 = m c +k 2

M (x) ( Xm)
1
= ( Xm){m(x)114 (x A x A ) 2 2 2 114 + }. (4.15b)
k0 = m c +k 2
In particular, for the exotic solutions (3.1), we have (cf. (4.12)):
1
M (x) ( Xm ) = {m(x)114 }( Xm )
2 (4.16a)
1
M (x) ( Xm ) = ( Xm ){m(x)114 + }
2
1
M (x) ( Xm ) = {m(x)114 } ( Xm)
2 (4.16b)
1
M (x) ( Xm ) = ( Xm ){m(x)114 + }
2
1
M (x) (k)( Xm ) = {m(x)114 } (k)( Xm )
2 (4.16c)
1
M (x) (k)( Xm ) = (k)( Xm ){m(x)114 + }.
2
Besides, the equations (4.12) remain the same for the solutions (3.1) and (4.11) reduce
to the identity 0 = 0 for them.
The equations (4.11) (resp. (4.12)) show that the eigenvectors of the momentum (resp.
charge) operator are mapped into such vectors by the operators (k) and (k) (resp.
, , (k), , , and (k)). However, by virtue of the equalities (4.13)(4.16),
no one of the operators , , (k), , , and (k) maps an eigenvector of the
angular momentum operator into such a vector. The cause for this fact are the matri-
ces (2.15) 4 i.e. I = 12 i and I = + 12 i appearing in (4.13)(4.16), which gen-
erally are non-diagonal [1, 41] and, consequently mix the components of the matrix vectors
( Xm ), ( Xm ), (k)( Xm), ( Xm), ( Xm ), and (k)( Xm ) in (4.13)(4.16). Since
the matrices 12 have a dimension of angular momentum and, obviously, originate
from the pure spinor properties of spin 12 fields, we shall refer to them as spin-mixing an-
gular momentum matrices or simply as spin-mixing matrices; by definition, the spin-mix-
ing matrix of the field (resp. its Dirac conjugate
) is 2 (resp. + 2
1 1
). More

generally,
if X is a state
vector and M (x) (k)( X ) = {l (x) 1
1 4 + s } (k)( X ) or
M (x) (k)( X ) = (k)( X ){l (x)114 + s }, where l and l are some operators
and s and s are matrices, not proportional to the unit matrix 114 , with operator entries,
then we shall say that the operators (k) or (k) have respectively spin-mixing (angular
momentum) matrices s and s relative to the state vector X ; we shall abbreviate this
by saying that the states (k)( X ) and (k)( X ) have spin-mixing matrices s and s ,
respectively.
The other additional terms in, e.g., equation (4.13) are

(x k x k ) 2 2 2 114 .
k0 = m c +k
They do not mix the components of (k)( Xm ) and (k)( Xm ). These terms may be
associated with the orbital angular momentum of the (matrix) state vectors (k)( Xm ) and
(k)( Xm) .
Thus, from (4.11)(4.13), the following conclusions can be made:
i. The operators + (k) and + (k) (respectively (k) and (k)) increase (respec-
tively decrease) the states 4-momentum by the quantity ( m2 c2 + k 2 , k).
ii. The operators , and (k) (respectively , and (k)) decrease (respec-
tively increase) the states charge by q.
iii. The operators + (k) and + (k) (respectively (k) and (k)) increase (respec-
tively decrease) the states orbital angular momentum by

(x k x k ) 2 2 2 .
k0 = m c +k
iv. The operators (k) (respectively (k) ) possess spin-mixing angular momentum
matrices 12 (resp. + 12 ) relative to states with fixed total angular momen-
tum.
In this way, the operators (k) and (k) obtain an interpretation of creation and
annihilation operators of particles (quanta) of a spinor field, namely:
+
(k) (respectively (k)) creates (respectively annihilates) a particle
(a) the operator
with 4-momentum ( m2 c2 + k2 , k), charge (q) (resp. (+q)), orbital angular
1 momentum

(x k x k )
2 , and spin-mixing angular momentum matrix 2 and
k0 =
2 2 m c +k
4 In the general case these are the matrices I = [I ij ] entering in (I.6.14)

+
(b) the operator
(k) (respectively (k)) creates (respectively annihilates) a particle
with 4-momentum ( m2 c2 + k2 , k), charge (+q) (resp. (q)), orbital angular
1 momentum

(x k x k ) , and spin-mixing angular momentum matrix + .
k0 = m2 c2 +k
2 2
Let us say a few words on the solutions (3.1) of the Dirac equations (2.6) in momen-
tum picture. If we try to describe them in terms of ordinary operators, not operator-valued
distributions, in the scheme developed we should put (k, m) = (0, 0). But the operators
(0) and (0) do not change a states 4-momentum (see (4.11) with p = k = 0, m = 0
and P = 0) and orbital angular momentum and produce states with spin-mixing angu-
lar momentum matrix 12 (resp. + 12 ) (see (4.16c)). Since the interpretation of the

(k) and (k) for (k, m) = (0, 0) is connected with a non-vanishing
operators

4-momen-
tum ( m2 c2 + k 2 , k) and/or non-vanishing orbital momentum (x k x k ) 2 2 2 114 ,
k0 = m c +k
the interpretation of (k) and (k) in the limit (k, m) (0, 0) as creation/annihilation
operators is lost. This is exactly the case if we make the limit (k, m) (0, 0) in (3.19),
(3.20), and (4.5)(4.7) (see remark 3.2). But, as we said already, (k) and (k) should
be operator-valued spinor distributions (generalized functions) for the solutions (3.1). For
them (4.11) transform into

k (k)( X0 ) = 0 k (k)( X0 ) = 0 k0 = k2 , (4.17)
as a result of m = 0, P = 0 and (4.7). So, if we suppose these equalities to hold for any
X0 = 0, we get
k (k) = 0 k (k) = 0 k0 = k2 (4.18)
which convert (3.19) and (3.20) into identities, due to m = 0 for the solutions (3.1). So, for
example, we can set
(k) = (2)3/2 a
s (k)v
s,
(k) (k) = (2)3/2 as (k)vs, (k) (4.19)
s s
a
s (k) = (k)s (k)
3
as (k) = 3 (k)
s (k) (4.20)
where
s (k) and s (k) are some operators, v
s,
(k) and vs, (k) are the classical spinors
defined via the equations (3.22)(3.24), and the constant factor (2)3/2 is introduced
for future convenience. (The fact that the solutions (4.19)(4.20) of (4.17) are not the only
ones is inessential for the following.) Consequently, when the solutions (3.1) are concerned,
the operators (operator-valued distributions) (k) (resp. (k)) decrease (resp. increase)
states charge by q and spin-mixing angular momentum matrix by 12 ; they preserve the
vanishing values of the 4-momentum and orbital angular momentum of the states. How-
ever, if we consider the solutions (3.1) as a limiting case when ([, P ] , m) (0, 0) and
([, P ] , m) (0, 0), we, by convention, can say that: (a) the distributions + (k) (resp.
(k)) create (resp. annihilate) particles with vanishing 4-momentum and angular momen-
tum, charge (q), and spin-mixing angular momentum matrix 12 , and (b) the distri-
butions + (k) (resp. (k)) create (resp. annihilate) particles with vanishing 4-momentum
and angular momentum, charge (+q), and spin-mixing angular momentum matrix + 12 .
Until now the operators s,()(k) and s,() (k), s = 0 for m = 0 and s = 1, 2 for
m = 0, entering in definition (4.1), were completely arbitrary linearly independent solutions
of (3.19) and (3.20). As a result, the operators (or operator-valued spinor distributions)

s (k) and s (k) are arbitrary linearly independent solutions of the operator equations

k 2 2 2 + mc114 s (k) = 0
k0 = m c +k
(4.21)

s (k) k k0 =
2 2 2 + mc 11
m c +k
4 =0
which, in form, coincide with the classical equations (3.22). This fact makes it possible to
be separated the pure operator part form the pure matrix-spinor part in the spinor opera-

tors s (k) and s (k). The existence of such decompositions is intuitively clear as one can
expect the dynamical variables to be expressible through some invariant characteristics of
the field (the operator part) while the transformation properties of the field should be ex-
pressible via some frame-dependent objects of the field (the matrix part). This separation

is most conveniently done by expressing s (k) and s (k) as follows (do not sum over s!)

s (k) = (2)
3/2
as (k)vs, (k)
s (k) = (2)
3/2
as (k)vs, (k) (4.22)
or, equivalently,
(k) = (2)3/2 a
s (k)v
s,
(k) (k) = (2)3/2 as (k)vs, (k). (4.23)
s s
Here vs, (k) and vs, (k) are defined via (3.22)(3.24), a
s (k) and as (k) are some (1
1 matrix) operators, for solutions different from (3.1), or (1 1 matrix) operator-valued
distributions, for the solutions (3.1), acting on the systems (spinor fields) Hilbert space F
of states, a
s (k), as (k) : F F , and as (k) and as (k) are such that

as (k) = as (k) as (k) = a s (k), (4.24)
due to (4.4) and (3.23). We have met the settings (4.22) above (see (4.19)) when the special
solutions (3.1) were considered, in which case the distributions a
s (k) and as (k) can be
represented in the form (4.20). The operators a+ +
s (k) and as (k) (resp. as (k) and as (k))
will be referred as the creation (resp. annihilation) operators (of the field).
From (4.11)(4.13)
and (4.23), we, applying (3.22)(3.24), derive the following rela-
tions (with k0 := m c + k 2 ):5
2 2

P a
s (k)( X p ) =(p k )as (k)( X p )
(4.25a)
P as (k)( X p ) =(p k )as (k)( X p )

Q a
s (k)( Xe ) = (e q)as (k)( Xe ) Q as (k)( Xe) = (e + q)as (k)( Xe ) (4.25b)

M (x) a
s (k)( Xm ) = {m (x) (x k x k )}as (k)( Xm )

1

(k)at (k)( Xm)
st,

2 t
(4.25c)
M (x) as (k)( Xm) = {m (x) (x k x k )}a s (k)( Xm )

+ (k)at (k)( Xm ),
1
ts,

2 t
5 The relations involving a (k) are obtained from the similar ones involving (k) by multiplying the
s
latter from the left by vs, (k). Analogously, the relations involving a
s (k) are obtained from the similar ones
involving (k) by multiplying the latter from the right by 0 vs, (k).
where
(k) := v s, (k) vt, (k) = (k) = vs, (k)0 vt, (k)
st, st,
(4.26)
and the indices s and t take the value 0 for m = 0 and the values 1 and 2 for m = 0. The
numbers st,
(k), appearing in (4.25c), generally depend on all of the arguments indicated.
We shall comment on them in Sect. 5.
For the solutions (3.1), the equations (4.25) reduce to:
k a k as (k)( X0 ) = 0
s (k)( X0 ) = 0 (4.27a)

Q a
s (k)( Xe ) = (e q)as (k)( Xe ) Q as (k)( Xe) = (e + q)as (k)( Xe) (4.27b)
1
M (x) a s (k)( Xm ) = m (x)as (k)( Xm ) (k)at (k)( Xm )
st,
2 t
(4.27c)
1
s (k)( Xm )+ (k)at (k)( Xm ),

M (x) as (k)( Xm ) = m (x)a ts,
2 t
where the quantities m (x) in (4.27c) contain only a spin angular momentum.
As a consequence of (4.25), the interpretation of a
s (k) and as (k) is almost the same as

the one of (k) and (k), respectively, with an only change concerning the spin angular
momentum.
If m = 0, then s,t = 0 and the equations (4.27c) say that the particles
created/annihilated by a 0 (k) possess spin angular momentum 2 (k) and
1 00,
the ones created/annihilated by a0 (k) have a spin angular momentum equal to

(k). However, for m = 0, the eigenstates of M are, generally, not mapped
+ 12 00,

into such states by a
0 (k) and a0 (k) as they are mixed through the quantities (k) via
st,
the polarization indices they carry. Now the matrices 12

(k) and + 2 (k), with
1

st, 2

(k) := (k) s,t=1 (4.28)
play a role of polarization-mixing matrices since we can rewrite (4.27c) for m = 0 as

1
M (x) a (k)( Xm ) = {m(x)112
(k)}a (k)( Xm )
2 (4.29)
1

M(x) a (k)( Xm ) = at (k)( Xm ){m(x)112 + (k)},
2
1 0
a
1 (k) a (k)
where 112 := is the unit 2 2 matrix, a(k) := a
, and a (k) := 1 .
2 (k)
01 a2 (k)
So, we can say that a (k) (resp. a (k)) creates/annihilates particles (states) with po-
larization-mixing matrices 12
(k) (resp. + 2 ). This interpretation holds also in the
1

massless case, m = 0, if we set a(k) := a
0 (k), a (k) := a0 (k), and (k) := (k) in
00,
that case. It is worth mentioning, as a consequence of (4.27), the particles and antiparticles
of a free spinor field are always different regardless of their mass m and charge q.
Ending this section, we note that the interpretation of (k), (k), a
s (k), and as (k)
as operators creating/annihilating particles with fixed charge and spin (polarization) mixing
matrices is entirely based on the equations (2.25) and (2.26), or, more generally, on (I.6.29)
and (I.6.30), which are external to the Lagrangian formalism and whose validity depends
on the particular Lagrangian employed. In particular, for the Lagrangian (2.4) (or (2.1)
in Heisenberg picture), the equalities (2.26) do not hold for = 0 and = 1, 2, 3 see,
e.g., (2.33) and (I.6.9). The below-written Lagrangian (10.4) (or (10.3) in Heisenberg pic-
ture), used in [4], is an example of a one for which all of the equalities (2.25) and (2.26)
hold. With an exception of the present sections, the equations (2.25) and (2.26) are not used
in this chapter.
5. The Dynamical Variables in Terms of

The main purpose of this section is a technical one: to be derived expressions for the mo-
mentum, charge and angular momentum operators in terms of the creation and annihilation
operators introduced in Sect. 4. The results will promote the physical interpretation of these
operators and will be applied significantly in the subsequent sections.
As (4.5)(4.7) imply

[, P ] = {k (+
s (k) + s (k))} d k
3
k0 = m2 c2 +k 2
s
(5.1)
[, P ] = {k +
2 (s (k) + s (k))} d k,
3
k0 = m2 c2 +k
s
the energy-momentum operator (2.19), by virtue of (4.5) and (4.6), can be written as

1
T = c d3 k dk k k =m2 c2 +(k )2 + +
s (k) + s (k) s (k )+s (k )
2 s,s
0
+ +
k s (k) +
k0 = m2 c2 +k2 s (k) s (k ) + s (k ) . (5.2)
Since the spinors and satisfy the Klein-Gordon equations (3.11), this is true also
for their components and, consequently, according to (II.5.4), the equality
x )(k + k )
(k) (k ) U (x, x0 ) = e i (x U (x, x0) (k) (k )
1

0 (5.3)

holds for , = +, , k0 = m2 c2 + k2 , k0 = m2 c2 + (k )2 , (k) = ,s (k),
,s (k), with being an index denoting spinors components, and U (x, x0 ) being the op-
erator (I.5.1) by means of which the transition from Heisenberg to momentum picture is
performed.
Substituting (5.2) into (I.6.6), commuting U (x, x0) according to (5.3) until it meets
U 1(x, x0), performing the integration over x (resulting in (2)33 (k k ) and a phase
factor), and, at last, integrating over k , we get1
1 The integrals, appearing in the transition from (I.6.6) to (5.6) and similar ones required for the derivation
of (5.9) and (5.12) below, are of the type

J =

d3 x d3 k d3 k U 1 (x,x0 ) {s (k)A (k,k ) s (k )} U (x,x0 )
, =+, s,s

P = (2)3 d3 kk
k0 =

m2 c2 +k2
+ 0 +
s (k) s (k) s (k) s (k)
0
s,s

i (x0 x00 ) m2 c2 +k2
+ +
2
+ (1 + 2 + 3 )e s (k) s (k)
0

(1 + 2 + 3 )e+ i (x x0 ) m c +k
2 0 0 2
s (k) s (k) . (5.6)
2 2 0
Inserting the expansions (4.22) in the last result and applying (3.24) (respectively
(3.25)) to the different (resp. equal) frequency terms in the obtained expression, we derive
the familiar result (cf. [1, 2])

P = k |
k0 =

m2 c2 +k2
{as + (k) a +
s (k) as (k) as (k)} d k.
3
(5.7)
s
We turn now our attention to the charge operator in Heisenberg picture.2 In view of the
decompositions (4.5)(4.6), the current operator (2.20) reads
+
J = qc d3 k dk +
s (k) + s (k) s (k ) + s (k ) . (5.8)
s,s
Substituting the last equation into (I.6.7), commuting the operator U (x, x0 ) with + s (k)
and +s (k) according to (5.3), integrating over x (which gives (2)
3 3
(k k
) and a phase
factor), and integrating over k , we obtain

Q = q(2)3 d3k + 0 0 +
s (k) s (k) + s (k) s (k)
s,s

i (x0 x00 ) m2 c2 +k 2
+ +
2
+e s (k) s (k)
0

+ e+ i (x x0 ) m c +k
2 0 0 2
s (k) s (k) . (5.9)
2 2 0
At the end, the insertion of (4.22) here entails (cf. [1, 2])

Q = q {as + (k) a +
s (k) + as (k) as (k)} d k
3
(5.10)
s

=
x )(k + k )
d3 k d3 k {s (k)A (k,k ) s (k )}e i (x
1
d3 x 0 , (5.4)
, =+, s,s

where A (k,k ), = 1,2,... , are some matrices, k0 = m2 c2 + k2 , k0 = m2 c2 + k , and (5.3) was applied.
2
Representing the exponent in the integrand as

x00 )(k0 + k0 ) + i 3a=1 xa0 (ka + ka ) i 3a=1 xa (ka + ka )
e i (x
1 0 1 1
e e
and taking into account that the integral over x results in (2)33 (k + k ), we see that, after trivial integration
over k , the above integral takes the form

J = (2)3
0 x0 )k (1+ 1)
d3 ke i (x
1
0 0
, =+, s,s

s (k)A (k,(1)(1)k) s ((1)(1)k). (5.5)
The reader can easily write the concrete form of the matrices A (k,k ) in the particular cases we consider
in this chapter.
2 In momentum picture, it will be found in Sect. 6; see (6.30).
where (3.24) and (3.25) were used.

Now comes the order of the angular momentum to be expressed via the creation and
annihilation operators. Combining (2.21), (2.22) and (4.5)(4.6), we see that
M = L

+ S = x T (x) x T (x) + S

(5.11)
where T is given by (5.2) and the spin angular momentum density is

+
1
S = c d3 k d3 k +
s (k)+s (k) ( + ) s (k )+s (k ) . (5.12)
4 s,s
Taking into account (2.31), substituting (5.12) into (I.6.9) and performing with the r.h.s.
of the obtained equality manipulations similar to the ones leading from (5.8) to (5.9), we
derive the following representation for the spin angular momentum in Heisenberg picture:
S0a = 0 (5.13a)
+
1
Sab = (2)3 d3 k s (k)0 ab 0 +
s (k) + s (k) ab s (k)
2 s,s

+ e i (x
0 x0 ) m2 c2 +k 2
+ +
2
s (k) ab s (k)
0 0

+ e+ i (x x0 ) m c +k
2 0 0 2
s (k) ab s (k)
2 2 0
(5.13b)
where a, b = 1, 2, 3.
Similar transformations of the orbital angular momentum (I.6.8) give its value in
Heisenberg picture as3
3 To prove (5.16), one has to substitute (5.2) into (I.6.8), then to apply (5.3), to calculate the space integral
over x, and, at the end, to integrate over k . The integrals, one has to calculate, are of the type

J=

d3 x d3 k d3 k xa U 1 (x,x0 ) {s (k)A (k,k ) s (k )} U (x,x0 )
, =+, s,s

x )(k + k )
=

d3 k d3 k {s (k)A (k,k ) s (k )}xae i (x
1
d3 x 0 , (5.14)
, =+, s,s

where a = 1,2,3, A (k,k ), = 1,2,... , are some matrices, k0 = m2 c2 + k2 , and k0 = m2 c2 + k . The
2

integration over x results in (2)3 i (ka + k a ) (k + k ).
3
(y) f (y)
Simple manipulation with the remaining terms, by invoking the equality f (y) y = y (y) in the form

3 (y z) 1 3 (y z)
d3 yd3 z f (y,z) = d3 y d3 z f (y,z) f (z,y)
(ya za ) 2 (ya za )

1
= d3 y d3 z3 (y z) a f (y,z),
2 y a z
gives the following result:

J = (2)3
0 x0 )k (1+ 1)
d3 k d3 k 3 (k + k )e i (x
1
0 0
, =+, s,s

1
L(x0 0) = x0 P x0 P + i(2) d k s (k) k k
3 3 + 0
s (k)
2 s,s
k k

+
+
s (k)
0
k k s (k)
k k
k k

e i (x x0 )k0 0 + 0
2 0 0
+ (1 + 2 + 3 )(1 + 2 + 3 )
k0 k0
+
k
+
k + s (k)0 +
s (k )
k k k k
k k

e+ i (x x0 )k0 0 + 0
2 0 0
(1 + 2 + 3 )(1 + 2 + 3 )
k0 k0

k + k + (k) 0
(k ) k =k
, (5.16)
k k k k s s
k0 = m c +k
2 2 2

where (5.6) was taken into account, the derivatives with respect to k0 , like (k),
k 0 s
must
be set equal to zero, and4

A(k) B(k)
A(k)k B(k) := k B(k) + A(k) k
k k k

= k A(k) B(k) (5.17)
k
for (matrix) operators A(k) and B(k) having C1 dependence on k. If the operators A(k) and
B(k) tend to zero sufficiently fast at spacial infinity, then, by integration by parts, one can
prove that

d k A(k) k k B(k) 2 2 2
3
k k k0 = m c +k

= 2 d3 k A(k) k k B(k) 2 2 2
k k k0 = m c +k

= 2 d3 k k k A(k) B(k) 2 2 2 . (5.18)
k k k0 = m c +k
By means of these equations, one can reduce (two times) the number of terms in (5.16), but
we prefer to retain the more (anti)symmetric form of the results by invoking the operation
introduced via (5.17).
1 ka ka 1
(x0 x00 ) + + x0a + i a + a s (k)A (k,k ) s (k ) . (5.15)
2 k0 k0 2 k k
The particular form of A (k,k ) is clear from (I.6.8) and (5.2). So, applying several times (5.15), calculating
the appearing derivatives, and, at last, performing the trivial integration over k or k by means of 3 (k + k ),
one can derive (5.16) after simple, but quite lengthy and tedious algebraic manipulations.
4 Recall, more generally, if : {F F } {F F } is a mapping on the operator space over the systems

used for = .
Let us represent the spin and orbital angular momentum operators (5.13) and
(5.16) as sums of time-dependent and time-independent terms:
0 1 0 1
S = S + S (x0) L = L + L(x0 ), (5.19)
where
1 1
S (x0) = (2)3(1 + 2 + 3 )(1 + 2 + 3 )
2 2 (x0 x0 )m2 c2 +k2 +
3
d k e i 0 s (k)0 +
s (k)
s,s

+ i (x0 x00 ) m2 c2 +k2

2
+e s (k) s (k)
0
(5.20a)
2 0 0
1 k k
L (x ) = i(2)3 d3 k e i (x x0 )k0 0 + 0
1 0
2 s,s
k0 k0
+ (1 + 2 + 3 )(1 + 2 + 3 )
+
+
k + k + (k) 0
(k )
k k k k s s
k k

+ e+ i (x x0 )k0 0 + 0
2 0 0
(1 + 2 + 3 )(1 + 2 + 3 )
k0 k0

k + k + (k) 0
s (k ) k=k . (5.20b)
k k k k s
k0= m2 c2 +k2
Proposition 5.1. The time-dependent components (5.20a) and (5.20b) coincide up to a

sign, i.e.
1 1
S(x0 ) + L (x0) = 0. (5.21)
Proof. Define the quantities

ss , ss ,
(k) := v s, (k) vs , (k) = (k) (5.22)

ss , 0 0 for = 0 or = 0
(x , k) := , (5.23)
2 ss ,
e i (x x0 ) m c +k (k) for , = 0
2 0 0 2 2
in terms of which (5.20a) reads

ss ,+ 0
1
S (x0) = (x , k)as + (k) a+
1
d3 k s (k)
2 s,s
ss ,
+ (x0 , k)as (k) a
s (k) (5.24)
as a result of the substitution of (4.22) into (5.20a). Inserting (3.28) into equation (5.22)
and using (2.29) and (2.17), we get

ss , s ,

(k) = 2iA(k)A(+k)(vs,
0 ) ( k + mc114 )(k k )v0
s ,

= 2iA(k)A (+k)(vs,
0 ) (k k )( k + mc11 4 )v0 . (5.25)
1
Substituting (4.22) into (5.20b), we see, on one hand, that L0a (x0 ) = 0 for a = 1, 2, 3
as a consequence of (3.25). On the other hand, performing the differentiations in the ob-
tained equation, with , = 0, by means of (3.28) (see equation (5.48) below) and ap-
1 1
plying (5.25), we get Lab (x0 ) = Sab (x0 ), a, b = 1, 2, 3. These results, together with
1
S0a(x0 ) = 0 (see (5.13a)), complete the proof of (5.21).
Inserting (4.22) into (5.13b), introducing the quantities

ss , 0 for = 0 or = 0
(k) := ss , (5.26)
(k) for , = 0
ss ,
with (k) defined by (4.26), and using (5.13a) and (5.19), we obtain the time-indepen-
dent part of the spin angular momentum in Heisenberg picture as5

1
S = d3k
ss ,
(k)as + (k) a
ss ,+ +
0
s (k) + (k)as (k) as (k) . (5.27)
2 s,s
We should note the that the quantities (4.26) and hence (5.26), generally, depend on all
of the arguments indicated and, for m = 0, are not diagonal in s and s, i.e. they are not

proportional to ss (= 1 for s = s , = 0 for s = s ). However, if m = 0, the indices s and s
take the single value 0, s, s = 0, so that (5.27) reduces to

1
Sm=0 = + +
0
d3 k 00,
(k)a0 (k) a0 (k) + (k)a0 (k) a0 (k) .
00,+
(5.28)
2
The diagonal part of the quantities (4.26) is real due to (3.23), viz.
ss,
(k) = (k),
ss,
(5.29)
where the asterisk means complex conjugation, but, because of (see (2.28) and do not sum
over and )

ss ,
(k) = v s , (k) vs, (k) = ss, (k), (5.30)
the non-diagonal part of the quantities (4.26) is, generally, complex for some and and
if s = s (for, of course, m = 0). However, using (4.24) and (5.30), one can prove that the
sums over s and s of the first/second terms in the integrand in (5.27) are Hermitian operators
(with, as it is well known, real eigenvalues).
Example 5.1. To get a concrete understanding of the quantities (4.26) and (5.22), we shall
ss
(k) = (k), not all
present below their particular values in some special frames. As ss

of the quantities ss
(k) are independent. Below we shall present only the independent ones
corresponding to (, ) = (0, 1), (0, 2),(0,3),(1,2),(2,3), (3, 1). Similar is the situation

with ss
(k). We shall also omit the argument k to save some space.
5 In momentum picture, the spin angular momentum will be found in Sect. 6 see equation (6.9) below.
If m = 0, there is a frame such thatk = 0. Following [1, 4], in this frame, we choose
see (3.28) with k = 0 and A (0) = 2mc1

v 1, = v1,+ = (0, 0, 0, 1) v 2, = v2,+ = (0, 0, 1, 0)
(5.31)
v 1,+ = v1, = (1, 0, 0, 0) v 2,+ = v2, = (0, 1, 0, 0).
Then, working with the representation (2.2) of the -matrices [1, 4], from equa-
tions (4.26) and (5.22), we, via a straightforward calculation, get:
, ss , s ss

ss = 0 = (1) a = 1, 2, 3 s, s = 1, 2

0a 12

ss , 1 for s =
s

23 = 1 = ss

0 for s = s (5.32)

,
ss (1)s i for s = s

ss = (1) s
(1 )i =

31
0 for s = s

ss , ss ,
ss , ss ,
01 = ss i 02 = (1)sss 03 = (1)s (1 ss )i ab = 0. (5.33)
Similarly, for m = 0, in a frame in which k1 = k2 = 0, we put [1, sec. 7.3]:

v 1, = v1,+ = N 1 (, 0, 1, 0) v 2, = v2,+ = N 1 (0, , 0, 1)
(5.34)
v 1,+ = v1, = N 1 (1, 0, , 0) v 2,+ = v2, = N 1 (0, 1, 0, )

k
3
where := and N = 1 + 2 . Then, in the representation (2.2) of the -ma-
mc+ m2 c2 +k2
trices, from (4.26) and (5.22), we obtain for m = 0 and k0 = m2 c2 + k 2 :
ss , ss ,
0a = 0 12 = (1)sss a = 1, 2, 3 s, s = 1, 2

ss , mc
mc
for s =
s

23 = (1 ) =
ss k 0

k0 0 for s = s (5.35)

ss , mc (1)s mc
k0 i for s = s

31 = (1)s (1 ss )i =

k0 0 for s = s
ss , ss , mc
ss ,
01 = (1 ss )i 02 = (1)s(1 ss ) 03 = (1)s ss i
k0
(5.36)
ss , ss , ss ,
12 = 0 23 = (1)s 2(1 ss ) 2 31 = 2(1 ss )i 2 .
N N
Notice the appearance of the imaginary unit i in the last formulae in (5.32) and (5.35),
which is in conformity with (5.30); the rest of the results agree with
(5.29).
For m = 0, in a frame in which k1 = k2 = 0, we set see (3.28) with A(k) =
12 2 2
2 m c +k
1 1
v 0, = v0,+ = (, 0, 1, 0) v 0,+ = v0, = (0, , 0, 1), (5.37)
2 2
k = k 3 2 {1, +1}, and, correspondingly, we get:

3 3
where :=
mc+ m2 c2 +k2 (k )
0a = 23 = 31 = 0 12 = 1
00, 00, 00, 00,
a = 1, 2, 3 (5.38)

01 = i 02 = 03 = i
00, 00, 00,+
(5.39)
03 = 12 = 0 23 = 1 31
00, 00, 00, 00,
= i.
To calculate the quantities (4.26) and (5.22) in an arbitrary frame, one should use (3.28)
or an equivalent to it explicit form of the spinors vs, (k) with s = 1, 2 for m = 0 or s = 0 for
m = 0.
If we write the spin vector (2.18) as

S = S (k) d3k, (5.40)
the results (5.28), (5.35), (5.38), and (5.39) imply

1
S (k)m=0 = S 3(k)m=0 = a1 + (k) a +
0 3
1 (k) + a2 (k) a2 (k)
2 (5.41)

a1 (k) a+ +
1 (k) + a2 (k) a2 (k)
1
S (k)m=0 = S 3(k)m=0 = a0 + (k) a +
0 3
0 (k) + a0 (k) a0 (k) (5.42)
2
in a frame in which k1 = k2 = 0. From here, for m = 0, follows that the spin projection,
on the third axis (of the chosen frame of reference) is 12 for the particles corresponding
to a1 + (k), a +
1 (k), a1 (k), and a1 (k), while it equals + 2 for the ones corresponding to
1
+
a2 (k), a +
2 (k), a2 (k), and a2 (k). Similarly (see (5.42)), for m = 0, the spin projection, on
the third axis (of the chosen frame of reference) is 12 for the particles corresponding to
any creation/annihilation operator.6 Further details regarding the spin angular momentum
and its interpretation can be found in [2, 5, 11].
Let us express now the orbital angular momentum (5.16) in terms of the operators a s (k)

and as (k). Substituting (4.22) and (5.7) into (5.16) and applying (5.19), (3.23) and (3.24),

we obtain the time-independent part of the orbital angular momentum in Heisenberg picture
as7

L = d3 k(x0 k x0 k )|k =m2 c2 +k2 {as + (k) a +
0
s (k) as (k) as (k)}
0
s
ss ,
1 ss ,+
+ d3 k l (k)as + (k) a +
s (k) + l (k)as (k) as (k)

2 s,s
6 After the normal ordering, this interpretation of a

s (k) and as (k) will be partially changed. In that
connection, the reader may notice some contradiction with the interpretation of the creation and annihilation
operators, given in Sect. 4, and the expressions (5.7), (5.10) and (5.27). It will disappear after normal ordering
of products. For details, see Sect. 8.
7 In momentum picture, the orbital angular momentum will be found in Sect. 6 see equations (6.11)
and (6.14) below.


1
+ i d k as (k) k k a
3 +
s (k)
2 s k k

+ as (k) k k a+s (k) , (5.43)
k k k0 = m2 c2 +k2
where

ss , ss ,
l: = iv s, (k) k k vs , (k) = l
(k) (k)
k k
v s, (k) v s, (k) s , (5.44)
= 2i k
k v (k)
k k
vs , (k) vs , (k)

= +2iv s, (k) k k .
k k

with the restriction k0 = m2 c2 + k 2 done after the differentiation (so that the derivatives
with respect to k0 vanish). The last two equalities in (5.44) are consequences of (see (3.24))

v s, (k) s , vs , (k)
v (k) + v s,
(k) = 0, (5.45)
k k
so that

v s, (k) s , vs , (k)
v s,
(k)k vs , (k) = 2k v (k) = 2k v s,
(k) . (5.46)
k k k
Notice, the equation (5.45) implies (see (3.23))
ss, s s,
l (k) = l (k). (5.47)

ss ,
So, l (k) are real and, by virtue of (4.24), the sums of the first/second terms in the
last integrand in (5.43) are Hermitian.
Exercise 5.1. Using (3.28), prove that the derivatives in the last equality in (5.44) are
vs, (k)
=0
k0
vs, (k) ka A (k) s,
(5.48)
= A (k) a + v0 +A1
0 s,
(k) v (k),
k a
m2 c2 + k 2 ka
s,
where a = 1, 2, 3, A(k) are some normalization constants and v0 are given by (3.29).
The charge, spin and orbital angular momentum operators in momentum picture will be
found in Sect. 6.
Since the right-hand-sides of (5.27) and (5.43) are constant (in spacetime) operators,
we have
0 0
S = 0 L = 0. (5.49)
Besides, by virtue of (5.19), (5.21) and (I.2.7), the total angular momentum operator of
the spinor field under consideration is
0 0
M = S + L (5.50)
and, as a result of (5.49), satisfies the evident conservation equation
M = 0, (5.51)
0 0
which agrees with (I.6.10). We shall call the conserved operators S and L the spin
operator and the angular operator, respectively, of the spinor field. In fact, these invariant
characteristics, not the non-conserved spin and orbital angular momentum, of the field are
the ones which are used practically for the description of a free spinor field; for instance,
the vector components (5.41) and (5.42) are, actually, the only spin characteristics of a free
spinor field examined in the (text)books [1, 4, 11].
Ending this section, we would like to make a comparison with the expressions for the
dynamical variables in terms of the creation/annihilation operators a
s (k) and as (k) in (the
momentum representation of) Heisenberg picture of motion [1, 4, 5, 11]. As a consequence
of (3.18), the analogues of the creation/annihilation operators, defined in terms of spinors
via (4.1) and (4.2), are
1 1

s (k) = e i x0 k
s (k) s (k) = e i x0 k
s (k) k0 = m2 c2 + k2
1 1
(5.52)
(k) = e i x0 k (k) (k) = e i x0 k (k) (k0 = m2 c2 + k 2 )
in Heisenberg picture. Therefore, defining (cf. (4.22))

s (k) =: (2)
3/2
as (k)vs, (k) s (k) =: (2)3/2as (k)vs, (k), (5.53)
we get the creation/annihilation operators in Heisenberg picture as

1 1
a
s (k) = e
i x0 k
as (k) as (k) = e i x0 k as (k) k0 = m2 c2 + k 2 . (5.54)
Evidently, these operators satisfy the equations

a
s (k) = as (k) as (k) = a
s (k), (5.55)
due to (4.24), and have all other properties of their momentum picture counterparts de-
scribed in Sect. 4.
The connection (I.5.4) is not applicable to the creation/annihilation operators, as well
as to operators in momentum representation (of momentum picture), i.e. to ones depending
on the momentum variable k.
Exercise 5.2. Verify, by using the results of Sections 3 and 4, the formulae

a
1
i
U (x, x0) a 1
s (k) = e s (k) U (x, x0 )
x k
1
k0 = m2 c2 + k2 , (5.56)
i 1
as (k) = e x k
U (x, x0) as (k) U (x, x0 )
from which equations (5.54) follow for x = x0 . (Notice, the right hand sides of the equa-
tions (5.56) are independent of x, due to the Heisenberg relations (I.4.1).)
From (5.6)(5.13), (5.27), and (5.52)(5.54), it is clear that all of the obtained expres-
sions for the momentum, charge and spin angular momentum operators in terms of the
(invariant) creation/annihilation operators remain unchanged in Heisenberg picture; to ob-
tain a Heisenberg version of these equations, one has formally to add a tilde over the cre-
ation/annihilation operators in momentum picture. However, this is not the case with the
orbital operator (5.43) because of the existence of derivatives in the integrands in (5.16)
and (5.43).
Exercise 5.3. Prove that in Heisenberg picture and in terms of the operators (5.54), the
term x0 P x0 P in (5.16) should be deleted and tildes over the creation/annihilation
operators must be added. Correspondingly, equation (5.43) in Heisenberg picture reads
ss ,
1 ss ,+
L = +
0
d3 k l (k)a+
s (k) as (k) + l (k)as (k) as (k)
2 s,s

1
3 +
s (k)
2 s k k

+ as (k) k k a+
s (k) . (5.57)
k k k0 = m2 c2 +k2

As we said at the beginning of Sect. 3, the equalities (2.6), (I.6.6) and (2.19) form a closed
algebraic-functional system of equations for determination of the spinor field operators and,
consequently, of the dynamical quantities characterizing a free spinor field. Since, from sec-
tions 4 and 5, we know that the field operators and dynamical variables of a free spinor field
can be expressed uniquely via the creation and annihilation operators a
s (k) and as (k),
which are invariant (frame-independent), in the present section we shall derive a system of
equations for these operators, which system is equivalent to the one just described. In fact,
this procedure will be equivalent to write the field equations (in Heisenberg or momentum
picture) in terms of creation and annihilation operators.
The problem, we want to analyze, is as follows. Given field operators and with
decompositions (see (4.5), (4.6), and (4.22))

= dk +
s (k) + s (k)
s

= (2) 3/2
dk a+
s (k)v
s,+
(k) + a
s (k)v
s,
(k) (6.1a)
s

= dk +
s (k) + s (k)
s

= (2)3/2 dk a+
s (k)v
s,+
(k) + as (k)vs, (k) , (6.1b)
s
find (the explicit equations describing) a

s (k) and as (k) such that the equalities (4.7) hold.
The results of sections 3 and 4 show that (4.7) are the implicit equations of motion for
and as, under the definitions of the quantities in them, they are equivalent to the ini-
tial Dirac equations (2.6). The equations (4.7) and (5.6) form a closed system of equations

with respect to s (k) and s (k). The substitution of (5.6) into (4.7) results into an explicit

system of equations relative to s (k) and s (k); one can easily write it out in full. How-
ever, it is rather complicated, which is due to the non-invariant, frame-dependent, character

of the operators s (k) and s (k). This dependence can be removed by specifying these
operators as in (4.22). In this way one gets a closed system of equations for the invariant,
frame-independent, operators a
s (k) and as (k). It can be derived in the following way.
Inserting (4.22) into (4.7) and taking into account that vs, = 0 and vs, = 0, due to the
linear independence of these spinors (see also the normalization conditions (3.24)), we see
that (4.7) is tantamount to:

[a
s (k), P ] = k as (k) = q a
s (k) (k q) d q
3 3
(6.2a)

[as (k), P ] = k as (k) = q as (k)3 (k q) d3 q (6.2b)

k0 = m2 c2 + k2 q0 = m2 c2 + q2 . (6.2c)
Substituting here (5.7), with an integration variable q for k and summation index t for
s, we get (do not sum over s!)

q q =m2 c2 +q2 a
0
+ +
s (k), at (q) at (q) at (q) at (q)
t (6.3a)

a
s (k)st (k q)
3
d q=0 3

q
q0

= m2 c2 +q2
as (k), at+ (q) at (q) at (q) at+ (q)
t (6.3b)

as (k)st 3 (k q) d3 q = 0.
Consequently, the operators a

s (k) and as (k) must be solutions of

as (k), at+ (q)at (q)at (q)at+(q) a
s (k)st (k q) = f st (k, q)
3
(6.4a)

as (k), at+ (q)at (q)at (q)at+ (q) as (k)st 3 (k q) = fst (k, q), (6.4b)
where fst (k, q) and fst (k, q) are (generalized) functions such that

q q =m2 c2 +q2 fst (k, q) d3 q = 0
0
t
(6.4c)
q
q0 =
f (k, q) d3 q = 0.
m2 c2 +q2 st
t
Since any solution of the Dirac equations (2.6) can be written in the form (6.1) with
a
s (k) and as (k) being solutions of (6.4), we can assert that the system (6.4) is equivalent
to the initial system (2.6) under the subsidiary condition (I.4.1). In this sense, (6.4) is
the system of field equations in terms of creation and annihilation operators in momentum
picture. Comparing it with the system(s) (II.12.13) of field equations in terms of creation
and annihilation operators for a free charged scalar field, we see that the structure of these
two types of algebraic-functional equations is quite similar; the only essential difference
being in the signs before the second terms in the commutators in (6.4a) and (6.4b) as in a
case of free charged scalar field in this place stands +as (q) a+
s (q) instead of the term
+
as (q) as (q).
1
As a first application of the field equations (6.4), we shall calculate the commutator
[ S , P ] between the spin angular momentum operator in Heisenberg picture, S , and
the momentum operator P = P . Applying the equalities (5.27), with q for the integration
variable k, and (5.7), with summation variable t for s, we get
0 1
[ S, P] =
2
d3 qd3 kk
k0 =

m2 c2 +k2
ss ,
{ (q)B ss ,+ +
sst (q, k) + (q)Bsst (q, k)}
s,s ,t
where
+
B +
sst (q, k) := [as (q) as (q), at (k) at (k) at (k) at (k)] .

Using the identity [A B,C] = [A,C] B + A [B,C] , (6.4a) and (6.4b), we see that (do
not sum over s and s !)

B
sst (q, k) = (st s t ) (q k)as (q) as (q) + as (q) f st (q, k) + as (q) f st (q, k).

3
0
Substituting this result into the above expression for [ S, P] , using (6.4c) (with k for
q and vice versa) and summing over t, we obtain
0
[ S, P ] = 0. (6.5)
So, for free spinor fields the spin operator in Heisenberg picture commutes with the
momentum operator in Heisenberg or momentum picture (see (I.5.14)).
Exercise 6.1. Making similar manipulations with the time-dependent part (5.24) of the spin
angular momentum, prove the relation

[ S(x ), P ] = 0
1 0
d3 k m2 c2 + k 2
s,s
,+
+ ss , 0

(x , k)as (k) as (k) + (x , k)as (k) as (k) , (6.6)
ss 0 +
which, by virtue of (6.5) and (5.19), implies

[ S, P ] = 0 d3 k m2 c2 + k2
s,s
,+
+ ss , 0

(x , k)as (k) as (k) + (x , k)as (k) as (k) . (6.7)
ss 0 +
1 This comparison is based on the first choice of a Lagrangian and the third choice of an energy-momen-
tum operators in section II.9. Regarding some problems of a choice of Lagrangian(s) and structure of conserved
quantities see Sect. 10.
By virtue of (I.5.1) (or footnote I. 3 on page 15) and (6.5), we have

0
[ S, U (x, x0 )] = 0 (6.8)
where U (x, x0 ) is the operator responsible for the transition from Heisenberg to momentum
picture. So, (I.5.4) implies
0
0
S = S (6.9)
i.e. the spin operators are identical in Heisenberg and momentum pictures.2
Exercise 6.2. Combining the above results with (5.50), (5.21), (I.6.17), (I.6.21)
and (I.6.23), derive the following equalities:
0 0
M = L + S M = 0L + 0S (6.10)
0
L = L + (x x0 ) P (x x0 ) P
0
(6.11)
0
[ L, P ] = [ 0L , P ] = i{ P P }. (6.12)
0
[ L, U (x, x0 )] = [ 0L, U (x, x0)] (6.13)
= {(x x0 ) P (x x0 ) P } U (x, x0 ).
As a result of (5.43) and (6.11), the explicit form of the orbital operator in momentum
picture in terms of creation and annihilation operators is

L =
0
d3 k(x k x k )|
k0 =

m2 c2 +k2
{as + (k) a +
s
ss ,
1 ss ,+
+ d3 k l (k)as + (k) a +
s (k) + l (k)as (k) as (k)

2 s,s

1
3 +
s (k)
2 s k k

+ as (k) k k a+

(k) . (6.14)
k k s
k0 = m2 c2 +k2
For the purposes of the present work, the equalities (6.9) and (6.14) are very important
as they, in view of (5.27), give explicit representations of the spin and orbital operators (and,
hence, of the total angular momentum operator) in momentum picture via the creation and
annihilation operators (in momentum picture).
If we split 0L into a sum of two conserved operators as
L = 2L + 3L
0
(6.15)

1
L := + d3k l
2 ss ,
(k)a+ ss ,+ +
s (k) as (k) + l (k)as (k) as (k) (6.16)
2 s,s
2 From (6.9) does not follow the conservation of the spin operator; it is a consequence from the considera-
tions in Sect. 5.

3
L := d3 k(x k x k )|
k0 =

m2 c2 +k2
{a+ +
s (k) as (k) as (k) as (k)}
s

1
+ i d3 k as + (k) k k a s (k)
2 s k k

+ as (k) k k a+

(k) , (6.17)
k k s
k0 = m2 c2 +k2
then, from the proof of (6.5), it immediately follows

2
[ L, P ] = 0, (6.18)
which implies
2
[ L, U (x, x0 )] = 0 (6.19)
2
2
L = L . (6.20)
Combining the above results with (6.11)(6.13), we obtain

3
3
L = L + (x x0 ) P (x x0 ) P (6.21)
3
[ L, P ] = [ 3L , P ] = i{ P P }. (6.22)
3
[ L, U (x, x0 )] = [ 3L, U (x, x0)] (6.23)
= {(x x0 ) P (x x0 ) P } U (x, x0 ).
As a result of (6.21), in Heisenberg picture and in terms of the Heisenberg creation and
annihilation operators (5.54), the terms proportional to the momentum operator in (6.17)
disappear,

1
L = i as + (k) k k a
3
s (k)
3
d k
2 s k k

+ as (k) k k a+

s (k) , (6.24)
k k k0 = m2 c2 +k2
So, the conclusion can be made that the non-commutativity between the total angular
momentum or orbital operator with the momentum operator is entirely due to the opera-
tor (6.17) or, more precisely, to the terms containing derivatives in (6.17), i.e. in the expres-
0
sion for the orbital operator L .
3
It is important to be noted, the operator L is a carrier of the pure orbital angular mo-
2 0
mentum of a system while the sum L + S represents its pure spin angular momentum.
Exercise 6.3. Using the explicit results (5.7), (5.10) and (5.43), in a way similar to the
derivation of (6.5) (by essentially employing the equation of motion (6.4)), prove via a
direct calculation the relations:
[ P , P ] = 0 (6.25)
[ Q , P ] = 0 (6.26)
[ L , P ] = i( P P ).
0
(6.27)
We emphasize ones again, the relations (6.25)(6.27) are external to the Lagrangian
formalism and their validity depends on the Lagrangian one employs. The equation (6.27)
agrees with (6.12) and, by virtue of (6.5), (6.9), (I.6.24) and (I.2.7) implies
[ M , P ] = i{ P P }. (6.28)
which justifies the choice of the sign on the r.h.s. of (I.6.17). Notice, (6.14) agrees
with (4.13) and (4.25c) (see also Sect. 9) which will not be the case if on the r.h.s. of (I.6.17)
stands +i instead of i.
Since (6.26) implies (see footnote I. 3 on page 15 or (I.5.1))
[ Q , U (x, x0)] = 0, (6.29)
from the general equation (I.5.4), with A = Q , we get
Q = Q . (6.30)
So, the charge operator is one and the same in momentum and Heisenberg pictures.
7. Anticommutation Relations
As we said in Sect. 6, the main difference between (6.4) and a similar system of equations
for a free charged scalar field is in the sign in the second terms in the commutators in them.
There is also and a second difference, which is and the last one. We have in mind the pres-
ence of the polarization (spin) indices s and t (taking the values 1 and 2 for m = 0 and the
value 0 for m = 0) in (6.4). However, this second difference is (formally) insignificant as
we, for instance, can introduce new variables,
say and , and put = (s, k) and = (t, q).
So, if we set ( ) := st 3 (k q) and d := t d3 q, the mentioned second differ-
ence will disappear. Hence we can treat the polarization and momentum variables on equal
footing. These reasonings allow us to transfer the investigation of the (anti)commutation
relations in section II.14 for scalar fields to spinor ones mutatis mutandis, almost automati-
cally.
Applying the identity
[A, B C] = [A, B] C B [A,C] , (7.1)
where = 1 and [A, B]1 := [A, B] := A B B A for any operators A and B (with
common domain), we rewrite (6.4) as1
+ +
[a
s (k), at (q)] at (q) at (q) [as (k), at (q)]

[a + +
s (k), at (q)] at (q) + at (q) [as (k), at (q)] (7.2a)
a
3
1 Here and below, do not sum over s in expressions like a

s (k)st .
+
+
[a
+ +
s (k), at (q)] at (q) + at (q) [as (k), at (q)]

[a (7.2b)
as (k)st 3 (k q) = fst (k, q).
Writing these equalities explicitly for the upper, +, and lower, , signs, we see that
they can equivalently be represented respectively as:

[a

[a
s (k), at (q)] at (q) at (q) [as (k), at (q)] (7.3a)
a
3


[a


[a (7.3b)
as (k)st 3 (k q) = fst (k, q).
Following the book [1, subsect. 10.1], we shall assume the following additional condi-
tion.
Assumption 7.1. The (anti)commutators of creation and annihilation operator are propor-
tional to the identity operator of systems Hilbert space F of states.2 Namely, we assume
the following conditions:
[a
s (k), at (q)] = g,st (k, q) idF [as (k), at (q)] = g,st

(k, q) idF
[a
s (k), at (q)] = b,st (k, q) idF [as (k), at (q)] = b,st

(k, q) idF

(7.4)
[a
s (k), at (q)] = d,st (k, q) idF [as (k), at(q)] = d,st

(q, k) idF

[a
s (k), at (q)] = e,st (k, q) idF [as (k), at(q)] = e
,st (q, k) idF

where = 1 and g
,st , g,st , . . . , e,st are some complex-valued (generalized) functions.
These last (generalized) functions in (7.4), which have to be determined, are subjected
to a number of restrictions which were derived in sections II.6 and II.14. Their explicit form
is:

(k + q)g
,st (k, q) = 0 (k + q)g,st (k, q) = 0 (k + q)d,st (k, q) = 0 (7.5a)
(k q)b
,st (k, q) =0 (k q)b
,st (k, q) =0 (k q)e
,st (k, q) =0 (7.5b)

m2 c2 + k2 + m2 c2 + q2 (k, q) = 0 for = g
,st , g,st , d,st (7.6a)

m2 c2 + k 2 m2 c2 + q2 (k, q) = 0 for = b
,st , b,st , e,st . (7.6b)
2 The authors of [1] refer to [24, 21], where the introduction of commutation relations in quantum me-
chanics is discussed for systems which have a classical analogue. The introduction of anticommutation rela-
tions, which are adequate for description of spinor fields, cannot be confirmed in this way as half-integer spin
(Fermi-Dirac) systems have not classical analogues.

Regarding g
,st , g,st , . . . ,e,st as distributions, from (7.5), we derive:

f (q)(k, q) = f (k)(k, q) for = g
,st , g,st , d,st (7.7a)
f (q)(k, q) = f (+k)(k, q) for = b
,st , b,st , e,st (7.7b)
for any function f which is polynomial or convergent power series. In view of (7.7), the
equalities (7.6b) are identically satisfied, while (7.6a) are equivalent to the equations

m2 c2 + k2 (k, q) = 0 for = g,st , g,st , d,st . (7.8)
Substituting (7.4) into (7.3) and taking into account the equalities (7.7), we obtain,
from (6.4c), the following system of equations for the unknown (generalized) functions

g
,st , g,st , . . . , e,st :

ka d3 q at (q)d,st

(k, q) at (q)g
,st (k, q)
t (7.9a)

at (q)b
,st (k, q) at (q) e,st (k, q) st (k q)
3
=0

m2 c2 + k2 d3 q at (q)d,st

(k, q) at (q)g
,st (k, q)
t (7.9b)

at (q)b,st (k, q) at (q) e

,st (k, q) st (k q)
3
= 0

ka
d3 q at (q)g,st (k, q) at (q)d,st

(q, k)
t (7.9c)

,st (q, k) st (k q)
3
= 0

m2 c2 + k2 d3 q at (q)g
,st (k, q) at (q)d,st (q, k)
t (7.9d)

,st (q, k) st (k q)
3
= 0,
where = 1 is a free parameter and a = 1, 2, 3.

Now we shall impose a second, after (7.4), additional condition.
Assumption 7.2a. The equations (7.9) are valid for arbitrary at(q) and at (q).
In this case, for (k, m) = (0, 0), the equations (7.9) have the following unique solution

with respect to the g
,st , g,st , . . . , e,st :

g
,st (k, q) = g,st (k, q) = b,st (k, q) = b,st (k, q) = d,st (k, q) = 0 (7.10a)
e
,st (k, q) = st (k q).
3
(7.10b)
Evidently, (7.10a) convert (7.8) into identities, and, consequently, under the hypotheses
made, (7.10) is the general solution of our problem (for (k, m) = (0, 0)).
Let us consider now the case (k, m) = (0, 0), when (7.9) and (7.6) reduce to the identity
0 = 0. In it, the operators
a
s (k) and as (k) for m = 0 and k = 0, (7.11)

must satisfy (7.4) in which the (generalized) functions g
,st (0, q), g,st (0, q), . . . , e,st (0, q)
remain as free parameters of the theory. To insure a continuous limit (k, m) (0, 0), we
shall admit the following hypothesis.

Assumption 7.2b. The (generalized) functions g
,st (0, q), g,st (0, q), . . . , e,st (0, q) are
given via equations (7.10) with k = 0 (and m = 0).
Physically (see Sect. 4), the operators (7.11) describe creation/annihilation of massless
particles (of massless spinor field) with vanishing 4-momentum and angular momentum,
but carrying charge q and having different spin operators. Thus, the theory admits ex-
istence of free, charged, spin 12 , massless particles with vanishing spacetime dynamical
variables (4-momentum and angular momentum).3 This is the quantized version of the
exotic solutions (3.1) (see the paragraph containing equation (4.17) in Sect. 4). Mean-
while, we notice that the solutions (3.1) are rejected by (7.10) and (7.4) (see (4.17)(4.20)).
The developed until now theory cannot make a distinction between the cases = 1
and = +1 in (7.4), corresponding to quantization with commutators and anticommuta-
tors, respectively. For the purpose, a new additional conditions (hypotheses, postulate) is
required. As such additional condition can serve the spin-statistics theorem (or an equiva-
lent to it assertion, like, e.g., the charge symmetry) [1,4,5]. According to it, the half-integer
spin fields should be quantized via anticommutators. For this reason, we accept
Assumption 7.3. The number = 1 should be set equal to one,
= +1 (7.12)
Now the equalities (7.4), (7.10) and (7.12) imply the following anticommutation rules:
[a
s (k), at (q)]+ = 0 [a
s (k), at (q)]+ = 0
[a
s (k), at (q)]+ = 0 [a
s (k), at (q)]+ = 0
[a
s (k), at (q)]+ = 0 [a
s (k), at (q)]+ = 0
[a
s (k), at (q)]+ = st (k q) idF
3
[a
s (k), at (q)]+ = st (k q) idF .
3
(7.13)
These quantization rules will be accepted hereafter in the present chapter. In Sect. 10, we
shall show how they can be derived without invoking the spin-statistics theorem.
As we have noted several times above, the concepts of a distribution (generalized func-
tion) and operator-valued distribution appear during the derivation of the commutation re-
lations (7.13). We first met them in the tri-linear relations (6.4). In particular, the canonical
commutation relations (7.13) have a sense iff the anticommutators in them are operator-
valued distributions (proportional to idF ), which is not the case if the field is described via
ordinary operators acting on F . These facts point to inherent contradiction in quantum field
theory if the field variables are considered as operators acting on a Hilbert space. 4 The rig-
orous mathematical setting requires the fields variables to be regarded as operator-valued
distributions. However, such a setting is out of the scope of the present work and the reader
3It seems, until now such particles/fields have not been observed.
4 Tracking back this contradiction, one can come to the conclusion that its origin is in the addition of the
equations (I.5.30) to the classical scheme of the Lagrangian formalism.
is referred to books like [9, 10, 34, 35] for more details and realization of that program. In
what follows, the distribution character of the quantum fields will be encoded in the Diracs
delta function, which will appear in relations like (6.4) and (7.13).
As a first application of the anticommutation relations (7.13), we shall calculate the
commutator between the components (5.27) of the spin operator.
Exercise 7.1. Prove the following commutation relations between quadratic combinations
of creation and annihilation operators:
[as (k) a
s (k), at (p) at (p)]
= {st at (p) a
s (k) + s t as (k) at (p)} (k p)

3

[as (k) a
s (k), at (p) at (p)]

= {st a
s (k) at (p) + s t at (p) as (k)} (k p)

3

[a
s (k) as (k), at (p) at (p)] (7.14)
= {st at(p) as (k) + s t a
s (k) at (p)} (k p)
3
[as (k) a
s (k), at (p) at (p)] = 0

[as (k) a
s (k), at (p) at (p)] = 0

[a
s (k) as (k), at (p) at (p)] = 0.
These equalities are simple corollaries of the identities [A, B C] = [A, B] C + B

[A,C] and [B C, A] = [A, B]+ C + B [A,C]+, applied in this order to the left-hand-
sides of (7.14), and (7.13).
Now, via a direct calculation by means of (5.27) and (7.14), we get

ss ,
1 s t, , s t,
[ 0S , 0S ] = 2 d3 k (k) (k)ss (k) (k) as + (k) at(k)
4 s,s ,t
ss ,+ s t,+ ,+ st,+ +

+ (k) (k) ss (k) (k) a s (k) at (k) . (7.15)
In particular, the equation
[ 0S , 0S ] = 0 for 0 + 0 + 0 + 0 1 (7.16)
is an evident consequence of (5.26) and (7.15). Besides, in the massless case, when the
polarization indices take the single value 0, we obtain

[ 0S, 0S] m=0 = 0. (7.17)
For m = 0, the r.h.s. of (7.15) is, generally, a non-zero operator, which means that
in the massive case not all of the spin operator components are simultaneously measur-
able, contrary to the massless one. More details on this problem will be given in Sect. 9
(see (9.14)(9.19)).
Exercise 7.2. For m = 0, the summation over s in (7.15) can be performed explicitly
by means of (5.26), (4.26), (3.27), (3.22), and [k , 0 ]+ = 2k0 2i0 (k k )
(see (2.29)). Verify that the result reads:

{
ss ,
(k)
s t,
(k) ss ,
(k) (k)}
s t,
s
0
= v s, (k) [ , ] i (k k ) (k k ) v s, (k)
k0

= 2i (k) (k) (k) + st,
st, st, st,
(k)
i
v s, (k) 0 (k k )
k0

(k k ) v s, (k) for m = 0, (7.18)

where [ , ] = 2i + (see (2.29)) should used
and one should set , , , = 1, 2, 3 in the context of (7.15), due to (5.26).
The commutativity between the spin and charge operators, i.e.
[ 0S, Q ] = 0, (7.19)
is an almost trivial corollary from (5.10), (5.27), and (7.14). This implies that the spin and
charge are simultaneously measurable quantities.
Important corollaries from (7.13) are the commutation relations between the field oper-
ators and the total angular momentum operator M , viz.
1
[, M (x, x0 )] = x [, P ] x [, P ] + (7.20a)
2
1
[, M (x, x0 )] = x [, P ] x [, P ] , (7.20b)
2
which in Heisenberg picture read
1
[ (x), M] = x x (x) + (x) (7.21a)
x x 2
1
[ (x), M] = x x (x) (x) , (7.21b)
x x 2
and, together with (I.4.1) for i (x) = (x), (x), represent the relativistic covariance of
the theory considered [2, 80].
Since (7.20b) is a consequence of (7.20a), = 0 , and (2.32), we shall prove only
the equation (7.20a); besides, for brevity, only the massive case, m = 0, will be considered.5
We shall prove the following lemma from which equation (7.20a) follows, due to (6.10).
Lemma 7.1. The following two relations are valid:
[, 0S]
5 One can prove independently (7.20b) in a similar way. For an alternative proof of (7.21), see [7, 19.1].

2
+ 12 i0
1

d3 p 1 (p p )(+ (p) + (p))
p0
= p0 = m2 c2 +p2 (7.22a)

for , = 1, 2, 3

0 for = 0 or = 0
[, 0L] = x [, P ] x [, P ]

2 i
1 0

d3 p 1 (p p )(+(p) + (p))
p0
p0 = m2 c2 +p2 (7.22b)

for , = 1, 2, 3

1
2 for = 0 or = 0.
Proof. For = 0 or = 0, (7.22a) is an evident corollary of (5.26) and (5.27). Let , =

1, 2, 3. Substituting the equations (6.1a) and (5.27) (see also (6.9)) in [, 0S] and applying
the identity [A, B C] = [A, B]+ C B [A,C]+, the anticommutation relations (7.13), and
then (5.26), (3.28) and (3.27), we get (, = 1, 2, 3)
p mc
1 p +mc 0
[, S] = d3 p
0
0 + (p) + (p) 2 2 2 .
2 2p0 2p0 p0= m c +p
Since [p , 0 ]+ = 2p0 2i0 (p p ) (use [A, B C]+ = [A, B]+ C B

[A,C] , and (2.29)), the last equality implies (7.22a) for , = 1, 2, 3, as a result of (4.3),
(4.1), and (3.19) (or (4.22) and (3.22)).
Performing similar manipulations with the l.h.s. of (7.22b), by applying (5.43) for (5.27)
and (5.44) for (5.26), we see that the last displayed equation should be replace with
p mc
1
[, L] = x [, P ] x [, P ] + i(2)
0
2
3/2
d3 p p
p0
0 2
s
v (p) +
s,+
p + mc 0 v (p)
s,

a (p) + p 2
a (p) ( ) 22 2,
p s
p0 p s
p0 = m c +p
where ( ) means that one should subtract the previous terms in the braces with the
change , (5.18) has been applied, and integration by parts of terms proportional to
as (p) vs, (p)
p has been performed. Since [p , ]+ = 2p0 114 (see (2.17)) and 0, the
0
p0
equation (7.22b) follows from the last equality, due to (3.22), (4.22), 0a = i0 a for a =
1, 2, 3, (see (2.14) and (2.17)), and the fact that for = 0 (resp. = 0) the derivatives with
respect to (resp. ) identically vanish, so that, e.g., for = 0 and = 0 the terms denoted
by ( ) in the last equation identically vanish.
It is easy to be seen, the terms containing commutators with the momentum operator
in (7.20) are due to the orbital operator (see (6.14)) and, more precisely, the operator (6.17)
is entirely responsible for their appearance.
We should also mentioned the equations
[, Q ] = q [, Q ] = q, (7.23)
which are trivial corollaries from (6.1), (6.10), the identity [A, B C] = [A, B]+ C B
[A,C]+, and (7.13). Similarly one can prove that
[a
s (k), Q ] = qas (k) [as (k), Q ] = qas (k). (7.24)
Equations (7.23) and (7.24) entail that, if is any 4 4 matrix and G (k, k ) (resp. G (x))
is any matrix-valued operator build from k, k (resp. x) and the derivatives with respect to
them, then (, = +.)
[, Q ] = 0 [ (x) G (x) (x), Q ] = 0

(7.25)
[as (k) G (k, k ) as (k ), Q ] = 0 [as (k) G (k, k ) as (k ), Q ] = 0.
In particular, we have
[ P , Q ] = 0 [ Q , Q ] = 0
(7.26)
[ 0S, Q ] = [ 0L, Q ] = [ M , Q ] = 0,
as a result of (5.7), (5.10), (5.27), (5.43), (5.50), and (6.9)(6.11).

Ending this section, we would like to mention the relation

[ M , M ] = i M M M + M , (7.27)
which can be proved via a quite long and tedious direct calculation based on (6.10), (5.27),
(6.11), (6.14) and (7.13).6 It should be noted the opposite sign of the r.h.s. of (7.27) relative
to similar relations in the literature; see, e.g., [7, eq. (9.1.15)] or [5, eq. (2.84)]. This sign,
as well as the whole structure of (7.27), is in agreement with the first 4 terms, proportional
to 2i, in (7.18) (see also (7.15) or similar result obtained in II.13).
An alternative prove of equation (7.27) is based on (7.21), which is equivalent to (k0 =
m2 c2 + k 2 )
1
[ (k), M ] = i k k (k) + i (k)
k k 2
1
(7.28)

[ (k), M ] = i k k (k) i (k) ,
k k 2

where (k) = e i x0 k (k) and (k) = e i x0 k (k) are the Heisenberg analogues
1 1
of (k) and (k), respectively, (see (5.52)). Combining (7.28) with (5.11)(5.21) and
applying the identity [A, B C] = [A, B] C + B [A,C] , one can easily verify (7.27) with-
out invoking (7.13). This quite more simple derivation of (7.27) is remarkable with the fact
that if one imposes (7.21) (or (7.20) in momentum picture) as a subsidiary restriction on the
Lagrangian formalism, it immediately implies (7.27) for a spin 12 field regardless of the va-
lidity of the anticommutation relations (7.13). Similarly, (7.23) entails (7.26) irrespectively
of the validity of (7.13).
6 The proof in Heisenberg picture is more simple, but also too long.

The introduction of the vacuum (state) of a free scalar fields was discussed in momentum
picture in sections II.7 and II.15.1 Here is a brief mutatis mutandis summary of the argu-
ments leading to a correct definition of the vacuum of free spinor field.
The vacuum of a field is a particular its state which describes, in a sense, the absence
of the field itself. Since the field is considered as a collection of particles, the vacuum
should contain no particles of the field. Therefore the conserved dynamical characteristics
of the vacuum should vanish as an absent particle has zero 4-momentum, no charge, etc.
Besides, since the vacuum does not contains any particles, the action of an annihilation
operator on it should produce the zero state vector as one cannot destroy something that
does not exist. On the contrary, the action of a creation operator on the vacuum should
produce a non-zero vector describing a state with one particle in it and, consequently, the
vacuum cannot be represented by the zero vector. Applying these heuristic ideas to a free
spinor field, we may say that its vacuum should be represented by a non-vanishing state
vector which has zero 4-momentum, charge and total angular momentum and the action of
an annihilation operator on it results in the zero vector. So, denoting by X0 the state vector
representing the vacuum, which is also called the vacuum, we should have
P( X0 ) = 0 Q ( X0 ) = 0 M( X0) = 0
(8.1)
X0 = 0, a
s (k)( X0 ) = as (k)( X0 ) = 0.
Besides, by virtue of P ( X0 ) = 0, (I.5.1) and (I.5.3), one can expect the vacua in
Heisenberg and momentum picture to coincide, i.e.
X 0 = X0 . (8.2)
However, one can easily see that the conditions (8.1) do not agree with the ex-
pressions (5.7), (5.10), (5.27), and (5.43) for the conserved quantities of a spinor field.
Indeed, substituting in them (see the commutation relations (7.13)) as (k) a+ s (k) =
a+ s (k) a
s (k) + ss 3
(k k) and applying the so-obtained operators on X 0 , we, in view
of (8.1), get senseless combinations of infinities; e.g., for the charge operator the result
is Q ( X0 ) q3 (0)3 X0 instead of the expected Q ( X0 ) = 0. The problem originates
from the terms as (k) a+ s (k) in the obtained expressions for the dynamical variables. The
accepted and well working procedure for its removal is known as normal ordering of prod-
ucts (compositions) of creation and/or annihilation operators. It is described at length in the
literature [1, 4, 5, 37] and consists in the following, when applied to a free spin 12 field. The
Lagrangian and dynamical variables should be written in terms of creation and annihilation
operators and, then, any composition (product) of such operator must be replaced by its
normally ordered form. By definition, the normal form of a composition of creation and/or
annihilation operators is called a composition of the same operators, in which all creation
operators stand to the left relative to all annihilation operators, multiplied by minus one or
plus one, depending on is the permutation, bringing the operators from the initial compo-
1 For similar problems in Heisenberg picture, see [1, 4, 5].
sition to the final one, odd or even, respectively.2 The normal form of a composition of
creation and/or annihilation operators is known as their normal product (composition) and
the mapping assigning to a product of such operators their normal product will be denoted
by N and it is called normal ordering (operator). The action of N on polynomials or
convergent power series in creation and/or annihilation operators is extended by linearity.
The dynamical variables after normal ordering are denoted by the same symbols as before
this operation.
Since, obviously,
+ + +
N a
s (k) at (k) = N at (k) as (k) = at (k) as (k)

(8.3)
N as (k) at+(k) = N at+ (k) as (k) = at+ (k) as (k),
equations (5.7), (5.10), (5.27), (5.41), (5.42) and (6.14), after normal ordering, take respec-
tively the forms:

P = k |
k0 =

m2 c2 +k2
{as + (k) a +
s (k) + as (k) as (k)} d k
3
(8.4)
s

Q = q {as + (k) a +
s (k) as (k) as (k)} d k
3
(8.5)
s

1
0
S = d3k
ss ,
(k)a+ ss ,+ +
s (k) as (k) (k)as (k) as (k) (8.6)
2 s,s
0 3 1
= S3 (k)m=0 = a1 + (k) a
S m=0
+
1 (k) + a2 (k) a2 (k)
2 (8.7)

+ a+ +
1 (k) a1 (k) a2 (k) a2 (k)
0 3 1
S m=0 = S3 (k)m=0 = a0 + (k) a +
0 (k) + a0 (k) a0 (k) (8.8)
2

L =
0
d3 k(x k x k )|
k0 =

m2 c2 +k2
{as + (k) a +
s (k) + as (k) as (k)}
s
ss ,
1 ss ,+
+ d3 k l (k)as + (k) a +
s (k) l (k)as (k) as (k)
2 s,s

1
+ i d3 k as + (k) k k a
s (k)
2 s k k

+ as (k) k k as (k) 2 2 2 . (8.9)
+
k k k0 = m c +k
Accepting the above-described normal ordering procedure, we can formalize the defi-
nition of the vacuum of a free spinor field as follows.
Definition 8.1. The vacuum of a free spinor field is its physical state that contains no
particles and has vanishing 4-momentum, (total) charge and (total) angular momentum. It
2 The so-formulated definition holds only for half integer spin fields/particles. The relative order of the
creation/annihilation operators with respect to each other is insignificant due to the anticommutation rela-
tions (7.13).
is described by a state vector, denoted by X0 (in momentum picture) and called also the
vacuum (of the field), such that:
X0 = 0 (8.10a)
X0 = X0 (8.10b)
a
s (k)( X0 ) = as (k)( X0 ) = 0 (8.10c)
X0 | X0 = 1 (8.10d)
where | : F F C is the (Hermitian) scalar product of systems (fields) Hilbert space

of states.
As we noted in section II.7, the condition (8.10d) is of technical character and, usually,
is imposed for pure computational convenience.
Taking into account the expressions of the dynamical variables after normal ordering,
we see that (cf. (8.1))
P ( X0 ) = 0 Q ( X0 ) = 0 M ( X0) = 0L ( X0) = 0S( X0) = 0 (8.11)
which equalities solve the problem with the senseless eigenvalues of the conserved quanti-
ties, corresponding to the vacuum before normal ordering.
The normal form of the dynamical variables solves also two other problems we pre-
meditated did not mentioned earlier.
The first problem is connected with the positivity of the energy operator which is iden-
tified, up to a constant, with the zeroth component of the momentum operator, viz.
E := c P0 . (8.12)
Before normal ordering, according to (5.7), it is

E = m2 c2 + k2 {as + (k) a +
s (k) as (k) as (k)} d k
3
(8.13)
s
which is not positive defined. After normal ordering, in view of (8.4), it takes the form

E = m2 c2 + k2 {as + (k) a +
s (k) + as (k) as (k)} d k
3
(8.14)
s
which is a Hermitian operator with positive eigenvalues.

The second problem concerns the interpretation of the operators a
s (k) and as (k)
as creation/annihilation operators (see Sect. 4). Since the field can be thought as a col-
lection of particles, one can expect that the dynamical variables should be expressible as
sums/integrals of the corresponding individual characteristics of these particles; this is rig-
orously expressed via results like (5.6)(5.16) and (5.41)(5.43). However, if we want to
retain the interpretation of a
s (k) and as (k), introduced in Sect. 4, the signs before the
second terms in the braces in (5.7) and (5.10) should be opposite. The reason being quite
simple: the 4-momentum of a system of two particles, one created by a+ s (k) and another
one by as +(k), should be a sum of the 4-momenta of these particles, while the charge of
this system should be the difference of the charges of the particles, measured in the units q.3
As we see from (8.4) and (8.5), this problem automatically vanishes after normal ordering.
The equality (8.7) partially changes the interpretation of the creation and annihilation
operators with respect to the polarization (spin) index s (see (5.41) and the conclusions after
it). So, in the frame in which (8.7) is derived, the projection of the spin vector on the third
+
axis is + 12 (resp. 12 ) for the particles corresponding to a+
1 (k), a1 (k), a2 (k), and a2 (k)
+
(resp. a +
1 (k), a1 (k), a2 (k), and a2 (k)).
The general formulae (8.6), with = 0 or = 0, and (8.9) for the spin and orbital
operators agree with (4.25c) and the interpretation of a s and as

given in Sect. 4. This
problem will be discussed and solved in Sect. 10; in particular, see equation (9.10) below.
The normal ordering changes not only the dynamical variables, but also the field equa-
tions (6.4). Since the quadratic combinations of creation and annihilation operators in the
commutators in (6.4) originate from the momentum operator (see (6.2)), the field equa-
tions (6.4), after normal ordering, will read

as (k), at +(q)at(q)+at+ (q) at (q) a
s (k)ts (k q) = f st (k, q)
3
(8.15a)

as (k), at+ (q)at (q)+at+ (q)at (q) as (k)ts 3 (k q) = fst (k, q) (8.15b)

q q0=m2c2+q2 fst (k, q) d3 q = 0
t
(8.15c)
q q =m2c2+q2 fst (k, q) d3q = 0.
0
t
However, applying (7.1) with = +1, one can verify that (8.15) is identically valid as
a result of the anticommutation relations (7.13). This means that, under the hypotheses
made when deriving (7.13), the anticommutation relations (7.13) play a role of field equa-
tions with respect to the creation and annihilation operators, considered as field operators
(variables).
The normal ordering influences the r.h.s. of (7.15) too. In fact, applying (8.6) and (7.14),
we, via a direct calculation, see that
, +
1 s t, ss , s t,
[ 0S , 0S ] = 2 d3 k ss (k) (k) (k) (k) as (k) at (k)

4 s,s ,t
ss ,+ s t,+ ss ,+ st,+
(k) (k) (k) (k) at+ (k) as (k) , (8.16)
which agrees with (8.3). Therefore, in the massless case, we have

[ 0S , 0S ] m=0 = 0 (8.17)
as in it s = s = t = 0, i.e. the relation (7.17) is preserved after normal ordering of products.

Notice, the equality
[ 0S , 0S ] = 0 for 0 + 0 + 0 + 0 1 (8.18)
is an evident corollary from (8.6) and (5.26).

3Recall, a+
s (k) produces particles with charge
q, while as + (k) produces ones with charge +q. The
4-momentum of the both kind of particles is ( m2 c2 + k 2 ,k).
However, the commutativity between the charge and spin operators, expressed
by (7.19), is not influenced by the normal ordering procedure (see (8.5), (8.6) and (7.14)).
Similarly, the relations (7.20)(7.23) remain valid after normal ordering.
9. State Vectors
The description of the state vectors of a free spinor field is almost identical with the one of
state vectors of free charged (with non-zero charge) scalar field considered in section II.16.
Formally, the only essential difference is in the polarization index s carried by the creation
and annihilation operators of a spinor field.
In momentum picture, in accord with the general theory of chapter I, the state vectors
of a spinor field are spacetime-dependent, contrary to the field operators and dynamical
variables constructed from them. In view of (I.5.24), the spacetime-dependence of a state
vector X (x) is
X (x) = U (x, x0)( X (x0)) (9.1)
where x0 is an arbitrarily fixed spacetime point and the evolution operator U (x, x0) :
F F is
1

U (x, x0) = exp (x x0 )
i
k | {as + (k) a
s (k) + a +
s (k) a
s (k)} d 3
k . (9.2)
k0 = m2 c2 +k 2
s
due to (I.5.1) and (8.4) (see also (I.5.11)(I.5.24)). The operator (9.2) plays also a role of
an S-matrix determining the transition amplitudes between any initial and final states, say
Xi (xi ) and X f (x f ) respectively. In fact, we have
(0) (0)
S f i (x f , xi ) := X f (x f )| Xi (xi ) = X f (x f )| U (xi , x f )( Xi (xi )). (9.3)
For some purposes, the following expansion of U (xi , x f ) into a power series may turn
to be useful:

U (xi, x f ) = idF + U (n)(xi , x f ) (9.4)
n=1

1
U (xi , x f ) := (xi 1 xf 1 ) . . .(xi n xf n )
(n) (1) (n)
d3 k(1) . . . d3 k(n) k1 kn
n! s1 ,...,sn
+ (1)
as1 (k ) as1 (k ) + a+
(1) (1) (1)
s1 (k ) as1 (k )

asn+ (k(n)) a (n) + (n)
sn (k ) + asn (k ) asn (k )
(n)
(9.5)

(a)
where k0 = m2 c2 + (k (a))2 , a = 1, . . ., n.
According to (I.5.27) and the considerations in Sect. 4, a state vector of a state contain-
ing n particles and n antiparticles, n , n 0, such that the i th particle has 4-momentum
pi and polarization si and the i th antiparticle has 4-momentum pi and polarization si ,
where i = 0, 1, . . ., n and i = 0, 1, . . ., n, is given by the equality
X (x; p1, s1; . . .; pn , sn ; p1 , s1 ; . . .; pn , sn )

1 1 n
1 n
= exp (x x0 ) (pi ) + (x x0 ) (pi )

n !n ! i i =1 i i =1
+
as (p1 ) as (pn ) as (p1 ) as+ (pn ) ( X0 ), (9.6)
+ +
1 n 1 n
where, in view of the anticommutation relations (7.13), the order of the creation opera-
tors is essential. Besides, the vector (9.6) vanishes if two of the particles of the state it
describes are identical, i.e. , e.g. for a charged field, if either n 2 and two of the pairs
(p1 , s1 ), . . ., (pn , sn ) coincide or n 2 and two of the pairs (p1 , s1 ), . . ., (pn , sn ) coincide;
to prove this, apply (7.13).1 If n = 0 (resp. n = 0), the particle (resp. antiparticle) creation
operators and the first (resp. second) sum in the exponent should be absent. In particular,
the vacuum corresponds to (9.6) with n = n = 0. The state vector (9.6) is an eigenvector

of the momentum operator (8.4) with eigenvalue (4-momentum) ni =1 pi + ni =1 pi and is
also an eigenvector of the charge operator (8.5) with eigenvalue (q)(n n ).2
Exercise 9.1. Using (7.13) and (4.24), prove that the transition amplitude between two
states, like (9.6), is:
X (y; q1,t1 ; . . .; qn ,tn ; q1 ,t1 ; . . .; qn ,tn )

| X (x; p1 , s1 ; . . .; pm , sm ; p1 , s1 ; . . .; pm , sm )
1 1 n
1 n

n !n ! i i =1
i i =1
i1 ,...,i s t 3 (pn qi )s
n n i 1
t
n 1 i
3 (pn 1 qi ) . . .s1 t 3 (p1 qi )
2 i n
(i1 ,...,in ) 1 2 n
i1 ,...,i s t 3 (pn qi )s

n n i1 1
t
n 1 i
(i1 ,...,in ) 2
3 (pn1 qi ) . . .s1 t 3 (p1 qi ) (9.7)

2 i n
n
. . ., in ) of (1, . . ., n ) and i1 ,...,i (resp. i1 ,...,i ) equals to +1 or 1 depending on is the
n n
permutation (n , . . ., 1)
(i1 , . . ., in ) (resp. (n , . . ., 1)
(i1 , . . ., in )) even or odd, respec-
tively.
The conclusions from the formula (9.7) are similar to the ones concerning free scalar
fields in section II.16. For instance, the only non-forbidden transition from an (n -parti-
cle + n -antiparticle) state is into (n -particle + n -antiparticle) state; the both states may
differ only in the spacetime positions of the (anti)particles in them. This result is quite
natural as we are dealing with free particles/fields.
1 This is a demonstration of the so-called Pauli principle: no more than one particle can be in a given state
of a system consisting of fermions.
2 Recall (see Sect. 4), the operator a+ (k) creates a particle with 4-momentum k and charge q, while
s
+
as (k) creates a particle with 4-momentum k and charge +q, where, in the both cases, k0 = m2 c2 + k 2 . See
also equations (9.9)(9.11) below.

0, 1, . . ., then (9.7) says that
Xn | X0 = n0 , (9.8)
+ +
Example 9.1. Consider the one (anti)particle states at (p)( X0 ) and at (p)( X0 ). Apply-
ing (8.4)(8.9) and (7.13), we find (p0 := m2 c2 + p2 ):3

P at+(p)( X0) = p at+ (p)( X0 ) Q at+ (p)( X0) = qat+ (p)( X0)
(9.9)
P at + (p)( X0) = p at + (p)( X0) Q at + (p)( X0 ) = +qat + (p)( X0 )
1
0
S at+(p)( X0) = ts,+ +
(p)as (p)( X0 )
2 s
(9.10)
1
0
S at + (p)( X0) = + st, +
(p)as (p)( X0 )
2 s
1
S 3m=0 at+ (p)( X0) = (1)t {t1 a+ +
1 (p) + t2 a2 (p)}( X0 )
2 (9.11)
1
S 3 m=0 at + (p)( X0) = +(1)t {t1 a1 + (p) + t2 a2 + (p)}( X0 )
2
3
+ 1
S m=0 a0 (p)( X0 ) = + a+ (p)( X0 )
2 0 (9.12)
1
S 3m=0 a0 + (p)( X0 ) = a+ (p)( X0 )
2 0
1
0
L at+ (p)( X0) = (x p x p) l ts,+
(p)
2 s
+
i p p at (p)( X0 )
p p
1
(9.13)
0
L at + (p)( X0) = (x p x p ) + l st,
(p)
2 s
+
i p p at (p)( X0 ) ,
p p
where (9.11) and (9.12) are valid in a frame such that p1 = p2 = 0 and, consequently, in
which (5.35) and (5.38) are valid. Notice, the one-(anti)particle massive states are not, gen-
erally, eigenstates of the spin (angular momentum) operator. However, in a special frame in
which p1 = p2 = 0 they are eigenstates of the third spin vector component. Conversely, the
massless one-(anti)particle states are always eigenstates of the spin (angular momentum)
operator. It should be remarked the agreement of (9.9)(9.13) with (4.25).4
3 The easiest way to derive (9.13) is by applying (I.6.8), (5.2), (5.3) and (7.13). Notice, in Heisenberg picture

and in terms of the Heisenberg creation/annihilation operators (5.54), equations (9.13) read L at+ (p)( X0 ) =
+
0 and L at (p)( X0 ) = 0 which is quite understandable in view of the fact that L is, in a sense, the
average orbital momentum with respect to all spacetime points, while L (x,x0 ) is the one relative to x and x0 ;
the dependence on x0 being hidden in L , at+ (p) and at+ (p).
4 If the r h.s. of (I.6.17) is with an opposite sign, this agreement will be lost. Besides, due to (5.26), the r h.s.
of (9.10) vanishes for = 0 or = 0, which is not generally the case with the spinor terms in (4.25c).
The equations (9.9)(9.13) confirm the interpretation, given in Sect. 4, of the oper-
ators a
s (k) and as (k) as ones describing creation/annihilation of fields (anti)particles.
+
as (k)( X0 ) can be interpreted as one representing a par-
For instance, the state vector
ticle with 4-momentum ( m2 c2 + k2 , k), charge (q) and polarization-mixing matrices
st,+
12 [ (k)]2 0m
s,t=1 .
Acting with (8.16) on the state vectors at+(p)( X0 ) and at +(p)( X0 ) and using (7.13),
we obtain

[ 0S, 0S] at+ (p)( X0 )
1 ,+ s t ,+ ts ,+ s t ,+ +
= 2 ts (p) (p) (p) (p) at (p)( X0 ) (9.14a)
4 s ,t

[ 0S, 0S] at + (p)( X0 )
1 s t, s t,
= + 2 ts , (p) (p) ts , (p) (p) at + (p)( X0 ). (9.14b)
4 s ,t
In particular, since for m = 0 from (5.32) (or (7.18) with k = 0) follows (do not sum
over b!)
s t,
ssab , (0)bcs t, (0) ssbc , (0)ab (0) = 2ist,
ca (0), (9.15)
s
where (abc) is an even permutation of (123), we see that (9.10) and (9.14) imply

[ 0Sab , 0Sbc] X = i 0Sca X (9.16)
where m = 0, (abc) = (123), (231),(312), and X = at+ (0)( X0 ), at+ (0)( X0 ). The other
non-vanishing components of (9.14) with m = 0 and p = 0 can be obtained from (9.16)
by using the skewsymmetry of 0Sab in the indices a and b (see (4.26), (5.26), and (9.10)).
Introducing the conserved spin 3-vector operator
S := abc 0Sbc,
0 a
(9.17)
we can rewrite (9.16) as

3
[ 0S a , S b ] ( X ) = i abc 0S c( X )
0
for m = 0 (9.18)
c=1
with a, b = 1, 2, 3. Notice, in the massless case, we have

[ 0S a , S b ] m=0 = 0,
0
(9.19)
due to (8.17). Therefore, in the massless case all of the spin 3-vector components are
simultaneously measurable, contrary to the massive one, when neither pair of them is si-
multaneously measurable (if the third one does not vanish).
10. On the Choice of Lagrangian and Its Consequences

The developed until now theory of a free spinor field (in momentum picture) is based on
the Lagrangian
1
L = ic{ (x) ( (x)) ( (x)) (x)} mc2 (x) (x). (10.1)
2
in Heisenberg picture, or
1
L = L (, , y , y ) = c{ [, P] [, P] } mc2 (10.2)
2
in momentum picture. In this Lagrangian, the field and its Dirac conjugate are con-
sidered as independent variables. However, that choice of a Lagrangian for description of a
free spinor field is not the only possible one. For instance, in [2, 78] it is chosen as1
L0 = ic (x) ( (x)) mc2 (x) (x) (10.3)

L0 = c [, P ] mc2 (10.4)
in respectively Heisenberg and momentum pictures. Since
L0 L0 L0
= = mc2 = +ic
y
(10.5)
L0 L0 L0
= = c [, P ] mc2 = 0,
y
the equations of motion (I.6.3) for the Lagrangian (10.4) coincide with the system of Dirac
equations (2.6) for the Lagrangian (10.2). As L and L0 differ by a full 4-divergence,
L0 L = { 12 ic (x) (x)}, these two Lagrangians give rise (under some conditions
at infinity) to identical action integrals and, consequently, to identical theories.2 However,
the Lagrangian (10.1) has two advantages: on one hand, it is Hermitian contrary to (10.3)
and, on the other hand, it vanishes when the field variables satisfy the field equations, which
simplifies some calculations.
Remark 10.1. Before going on, we emphasize that the calculation of the derivatives in (10.5)
and in (10.8) below according to the rules of the classical analysis of commuting variables
is not quite correct. This method is harmless when the Euler-Lagrange equations are con-
sidered but it requires some additional operator ordering rules for the expressions of the
conserved quantities; for details, see [3] and section I.3. However, this approach breaks
down for the Lagrangian (10.18) (or (10.15)) which we shall investigate below; the cause
being that all its derivatives vanish if they are calculated according to the rules mentioned.
For the correct treatment of that Lagrangian, the methods developed in [3] are required (see,
in particular, [3, section 5.6]).
1 In [2] as an independent variable is chosen (x), not the pair ((x),(x)).
2 The expressions for the conserved currents (energy-momentum, charge current and angular momentum
density) arising from the Lagrangians (10.1) and (10.3) differ significantly. But the conserved, time-indepen-
dent, quantities (momentum, charge and angular momentum operators) for the both Lagrangians are identical.
It is almost evident, the operator and its Dirac conjugate do not enter in (10.1)
(10.4) on an equal footing, i.e., in some sense, is first and is second in order (count-
ing from the left to the right) operator in these Lagrangians. One can revert the situation by
considering, for example, the Hermitian Lagrangian
1
L1 = ic{ (x) ( (x))( (x)) (x)}mc2 (x) (x) (10.6)
2
1
L1 = c{ [ , P] [ , P ] } mc2 (10.7)
2
in Heisenberg and momentum picture, respectively.3 A straightforward calculation gives:
L1 L1 1 L 1
= = + c [ , P ] mc2 = ic
2 y
2
(10.8)
L1 L1 1 L1 1

=
= c[, P ] mc2
= + ic .
2 y 2
Therefore, in view of (I.6.3), the field equations for the Lagrangian (10.7) are
[ , P ] + mc = 0 [ , P ] mc = 0. (10.9)
So, as one can expect, we get the system of Dirac equations (2.6) with interchanged
places of and .4
In the Lagrangian (10.6), the fields and differ by their positions too; now is first
and is second. However, since we are dealing with a free field, we, due to the charge
symmetry [1, 4], can try to start from a Lagrangian which describes in a symmetric way the
fields particles and antiparticles. Since is connected with fields particles and with its
antiparticles, a hypothesis can be made that this may be achieved via a Lagrangian in which
and enter symmetrically. For example, as a possible candidate, the half sum of (10.2)
and (10.6) can be taken, viz.
1
L2 = c{ [, P] [, P] + (x) ( (x)) ( (x)) (x)}
4
1
mc2 { + (x) (x)}. (10.10)
2
Obviously, this Lagrangian has all required properties, including the symmetry
, which in matrix notation reads and . However, it turns that the field
described by the Lagrangian (10.10) has nothing to do with a spinor field. Indeed, one may
verify that the field equations for it are
[ , P] mc = 0 [ , P] mc = 0

3 Here and below the matrix transposition sign is required to be ensured a proper matrix multiplication
see the conventions at the beginning of Sect. 2.
4 Recall, the matrix transposition sign in (10.9) serves to provide a proper matrix multiplication as we
consider as vector-column and as vector-row. If we write (10.9) and (2.6) in components, it will be
obvious that these systems of 8 equations are identical up to notation.
or, in a matrix form,
[, P ] mc = 0 [, P ] + mc = 0,
where := 12 ( ( )) is the antisymmetric part of . Besides, now and do not sat-
isfy the Klein-Gordon equations (3.11) and the Lagrangian (10.10) happens to be singular.5
The last fact means that (10.10) describes a system with constraints that requires a different
treatment (see, e.g., [42]).
Similar will be the consequences of any Lagrangian symmetric in and : it will not
describe the field we are investigating here. One of the formal reasons for such a conclusion
is that the field equations for and , implied by a Lagrangian symmetric in them, will
be identical, contrary to our expectation that they should be (equivalent to) (2.6). The
physical reason for this situation is that the change particleantiparticle is describe by
a
s (k) as (k), not by , which is due to the half integer spin of Dirac spinor field
(and its transformation properties under Lorentz transformations).6
It is known (see, for instance, [1, 13.4], [4, 99], and [5, p. 114]), in terms of and
, the change particleantiparticle is describe by the replacement

:= C = (C )
:= C1 = (C1 ), (10.11)
called charge conjugation, where the matrix C satisfies the conditions
C1 C = := ( ) C = C (10.12)
and, in a representation, like (2.2), in which = (1) , = 0, 1, 2, 3, can be chosen as

C = b0 2 with b C\{0}. The Lagrangian (10.1), under the conditions (7.13) and after
(resp. before) normal ordering, is invariant (resp. changes sign) under the change (10.11).
So, if we want, from the very beginning, to have a suitable description of particle-an-
tiparticle properties of a spinor field, we have to describe it via the variables and , not
via and . In these new variables, due to = C1 (see (10.11) and (10.12)), the
Lagrangian (10.1) reads
1
L = ic{ (x)C1 ( (x)) ( (x))C1 (x)}
2
(x)C1 (x). (10.13)
+ mc2
We would like to emphasize on the change of the signs and the appearance of the matrix
C in (10.13) with respect to (10.1). An alternative to this Lagrangian is a one with changed
positions of and , viz.
1
L = ic{ (x)C1 ( (x))
( (x))C1 (x)}

2
5 For instance, in the representation (2.2) of the -matrices, in which 0 is diagonal, we have in Heisenberg
L L
2
picture (0 ) = (0 )
2
= 0.
6 One should compare this situation with the one for free scalar field where the change is equiv-
alent to the transformation particleantiparticle; see chapter II for considerations in momentum picture and,
e.g., [1] for treatment in Heisenberg picture.
+ mc2 (x)C1 (x).

(10.14)
Notice, the last Lagrangian is completely different from (10.6) and cannot be obtained

from it by putting (x) = C1 (x).
Evidently, the variables and do not enter
in (10.13) and (10.14) on equal footing. We shall try to symmetrize the situation by
considering a Lagrangian which is the half sum of the last two ones, i.e.
1
L = ic{ (x)C1 ( (x)) + ( (x))C1 (x)
4
(x)C1 (
(x)) + ( (x))C1
(x)}
1 (x)C1 (x) + (x)C1 (x)}.
+ mc2 { (10.15)
2
According to (I.6.2), the Lagrangians (10.13)(10.15) in momentum picture respec-
tively are:
1
L = c{C1 [, P] [ , P] C1 } + mc2 C1 (10.16)
2
1
L = c{C1 [, P] [ , P] C1 } + mc2 C1 (10.17)
2
1
L = c{C1 [, P] + [ , P] C1 C1 [, P]
4
1
+ [ , P ] C1 } + mc2 {C1 + C1 }. (10.18)
2
The consequences of these Lagrangians will be described and compared below. But,
before proceeding with this task, we must make an important remark regarding the La-
grangian (10.15) (or (10.18) in momentum picture). If one computes its derivatives with
respect to and and their partial derivatives, according to the silently accepted rules valid
for classical fields, one will see that all of them vanish.7 In a sense, this will mean that the
Lagrangian (10.15) is constant as a functional of the field variables, which, obviously, is
not the case. This paradox is explained in the work [3] in which it is proved that it is due
to an incorrect transferring of the differentiation rules from the classical analysis of com-
muting variables and functions of them to the quantum one of non-commuting variables,
such as the operators on Hilbert spaces, and functions of them. In this work, it is shown
that the derivatives of Lagrangians like (10.15) should be computed in a way different from
the one of the classical analysis (and, in fact, are operators on the operator space over F );
in particular (see [3, section 5.6]), the Euler-Lagrange equations for the Lagrangian (10.15)
are identities, like 0 = 0, but it implies field equations according to a rigorously modified
action principle. Without going into details, the new procedure, when applied to the La-
grangian (10.15), is equivalent to the standard one with new definitions of the derivatives
of (10.15) with respect to its operator arguments. Here is the list of the derivatives of the
7 It is quite possible, the first reaction to such a result will be the rejection of the Lagrangian (10.15) as
it leads to identities, like 0 = 0, instead to field equations and to identically vanishing dynamical variables.
However, this will be a hasty conclusion vide infra.
Lagrangians (10.16)(10.18):
L L L 1
= = := + cC1 [, P ] mc2C1
2
L L L 1
= = := cC1 [, P ] + mc2C1
2
(10.19)
L L L 1
= = = := icC1
y y y 2
L L L 1
= = = := + icC1 .
y
y
y
2
Notice, in this list the expressions for the derivatives of L are definitions, while the
derivatives of L and L are derived, using the equality
1
C = C1 (10.20)
implied by (10.12), according to the rules of the analysis of classical fields.
Exercise 10.1. Convert (10.19) into the more usual Heisenberg picture of motion.
Exercise 10.2. Show that, as a result of (10.19), the field equations (I.6.3) for all of the
Lagrangians (10.16)(10.18) read:
[, P ] mc = 0 [, P ] mc = 0. (10.21)
Consequently, as one may expect, the spinor field and its charge conjugate := C =
C0 are solutions of one and the same Dirac equation.8
Regardless of the identical equations of motion obtained from the Lagrangians
(10.16)(10.18), these Lagrangians entail completely different conserved quantities (dy-
namical variables). In [3] is shown that to any operator Lagrangian in quantum field the-
ory there corresponds a unique set of conserved operators.9 The corresponding to the La-
grangians (10.13)(10.15) energy-momentum, current and spin angular momentum density
operators are:

T =
( ) + ( )

1 (10.22a)
= ic{ C1 ( ) + ( )C1 }
2
q } = qc
C1 ope
J = { (10.22b)
i
}
S = { (I ) + (I )

1 1 (10.22c)
= c C { + }
4
8 It is a simple algebra to be proved, the substitution := C transforms the second equation in (2.6) into
the second equation in (10.21).

9 The meaning of the last theorem is that there is a criterion for selecting the order of all operators in
compositions (products) of operators appearing in the conserved quantities obtained from the Schwingers
variational principle [5] and/or from the first Noether theorem [1]. For rigorous derivation of equations (10.22)
(10.24) below, apply the results of [3, sec. 4] (see also section I.3) to the Lagrangians under consideration here.

T = ( )
( )
1 (10.23a)
= ic{ C1 ( ) + ( )C1 }

2
q
J = { } = +qc C1 (10.23b)
i
+ (I ) }
S = +{ (I )

1 (10.23c)
= c C1 { + }

4
1
T = {
( ) ( ) ( ) + ( ) }

2
1 C1 ( ) + ( )C1 (10.24a)
= ic{
4
C1 ( ) + ( )C1 }
1q }
J = { + +
+
2 i (10.24b)
1 C1 C1 }
= qc{
2
1

S = { (I ) (I )
2
}
(I )
+ (I ) (10.24c)
1 1
= c C ( + ) + C1 ( + )
,
8
where (10.19), () = () = +1, I = I = 12 i (see (2.15)), C

C
1
=
(see (2.14) and (10.12)), and (10.20) were used. The corresponding expressions for the
orbital angular momentum are obtained according to the general formula (I.2.8a).
Exercise 10.3. Applying the facts presented in section I.3, rigorously derive the expres-
sions (10.22)(10.24).
Exercise 10.4. Write (10.22)(10.24) in momentum picture; formally, this can be done
by respectively y := 1 [, P ] and y := 1 [, P ] and by
by replacing and i i
removing the tildes (waves) from the remaining symbols.
Looking over (10.15) (or (10.18)) and (10.22)(10.24), one can notice that the La-
grangian, energy-momentum and spin angular momentum density operators are symmetric
while the charge current operator is antisymmetric under the charge conjugation, i.e. under
the change , as it should be [1, 4, 5]. In this way, we see that the symmetry par-
ticle-antiparticle for a free spinor field is encoded into a Lagrangian which is symmetric
relative to the spinor operators describing the corresponding fields/particles.
Since the field and its charge conjugate := C satisfy identical Dirac equa-
tions for all of the Lagrangians (10.16)(10.18) (see (10.21)), the field and its Dirac
conjugate := 0 = C1 satisfy the system of Dirac equations (2.6) for these
Lagrangians. Thus for and are valid all considerations and results in sections 3
and 4.10 As the Dirac conjugate field enters linearly everywhere in sections 3 and 4,
all of the material in these sections is valid mutatis mutandis for the charge conjugate field
= C = (C ) = (C) whose components are linear combinations of those of ;
the only thing one should do, to rewrite this material in terms of , is to replace with
or, more formally, to write the breve accent sign for the line over a symbols sign.
For example, the decomposition (3.21b) implies

= d3 k fs,+ (k)s,(+)(k) + fs, (k)s,()(k) (10.25)
s
where fs,(k) are some (, possibly, generalized) functions and s,() (k) are spinor opera-
tors. It is essential to be mentioned the connection

s,() (k) = C
s,() (k) = s,() (k)C (10.26)
which is a consequence of (3.21b) and (10.11). Further, the creation/annihilation operators

for should be introduced via the second formula in (4.1) with for and f for f and, by
virtue of (10.26), are

s (k) = C s (k) = s (k)C . (10.27)
All of the remaining material of Sect. 4 concerning s (k) is then automatically

valid for s (k), provided the replacement
is made. In particular, the inter-
pretation
of
s (k) as creation/annihilation operators of (anti)particles with 4-momentum

( m c +k , k), charge q, orbital angular momentum (x k x k )
2 2 2 14 , and
21
2 2 k0 = m c +k
spin mixing angular momentum matrices 12 is valid. The invariant, frame-indepen-

dent, creation/annihilation operators as (k) for
s (k) are the same as for s (k) and, as a
consequence of (10.27) and (4.22), are introduce via the equation

s (k) = (2)
3/2
as (k)C(vs, (k)) = (2)3/2as (k) vs, (k)C (10.28)
where vs, (k) are defined by (3.22)(3.24). So, the equalities (4.22) and the following from
them interpretation of a
s (k) and as (k) hold without changes.
All of the above said, concerning and its (frequency) decompositions, is valid for any
one of the Lagrangians (10.16)(10.18) as they entail identical field equations. However,
form (10.22)(10.24), it is evident that the energy-momentum, charge and angular momenta
densities for these Lagrangians are completely different. The corresponding to them 4-mo-
mentum, charge and angular momenta can be calculated on the base of the results of Sect. 5
and the decomposition (10.28) in the following way.
At first, notice that the Lagrangians (10.16)(10.18) vanish if and satisfy the field
equations (10.21), i.e.
L = L = L = 0 if [, P ] mc = 0, = , , (10.29)
and L = L as = C (see (10.2)). Moreover, L (resp. L ) and the results follow-
ing from it can be obtained by interchanging the positions of and , possibly with some
10 The only exception being the concrete forms of the expression in the r.h.s. of (3.4) and (3.5) for the
degenerate solutions (3.1).

indices and arguments, (resp. by forming the half sum of L and L and the results cor-
responding to them). Consequently, the results (5.2), (5.6), (5.8), (5.9), (5.11)(5.16), all
with 1
s = s ) C , are valid for L = L , but for L and L the just describe changes
must be made.
Applying (10.28), (4.22) and (10.12), we derive:

(s (k)) s (k ) = (s (k))C1 s (k )

= +(2)3 as (k) vs, (k)vs , (k ) as (k ) (10.30a)

(s (k))C1 (k ) s

= (2)3 as (k) vs , (k ) C1 C vs, (k) as (k ) (10.30b)

(s (k))C1 (k ) s

= +(2)3 as (k) as (k ) vs , (k ) vs, (k) (10.30c)

(s (k))C1 0 ab s (k )

= (2)3 as (k) as (k ) vs , (k )0 ab vs, (k) (10.30d)
where , = +, , is a constant 44 matrix (e.g., = , 0 ab ), or an operator (e.g., =
k k ) or matrix operator (e.g., = 0 k k ),11 s, s = 1, 2 for m = 0, and s, s = 0 for m = 0.
The first of these equalities, together with the derivation of (5.7), (5.10), (5.27), and (5.43),
implies the validity of (5.7), (5.10) and (5.27) for L . The equations (10.30c) and (10.30d),
in view of the derivation of (5.7), (5.10), (5.27), and (5.43), mean that the change
is equivalent to any one of the replacements
s (k) s (k) or as (k) as (k), both
ss ,
combined with the change of the sign of the charge operator and the changes (k)
s s, ss , s s,
(k) and l (k)
l (k). The equations (10.30), combined with the above-made
conclusions, (10.22)(10.24), (5.7), (5.10), (5.27) and (6.14), imply that the momentum,
charge, spin and orbital (angular momentum) operators for the Lagrangians (10.16)(10.18)
respectively are:

P = k | {a+ +
s (k) as (k) as (k) as (k)} d k
3
(10.31a)
k0 = m2 c2 +k2
s

Q = +q {as + (k) a +
s (k) + as (k) as (k)} d k
3
(10.31b)
s
ss ,
1 ss ,+
= + d3 k
0
S (k)as +(k)a +
s (k)+ (k)as (k)as (k) (10.31c)
2 s,s

=
0
L d3 k(x k x k )| {as + (k) a +
k0 = m2 c2 +k2
s
ss ,
1 ss ,+
+ d3 k l (k)as + (k) a +
s (k) + l (k)as (k) as (k)

2 s,s

1
3+
s (k)
2 s k k
11 If is an operator, one should take care of the direction to which it acts, as well as on what arguments it
acts; e.g., if = k , then the r h.s. of (10.30b) is equal to a sum of two terms, due to the left action of on

vs , (k ) and the right action of on a
s (k ).

+ as (k) k k a+ (k) , (10.31d)
k k s
k0 = m2 c2 +k2

P = k |
k0 =

m2 c2 +k2
{a+ +
3
(10.32a)
s

Q = q {a+ +
s (k) as (k) + as (k) as (k)} d k
3
(10.32b)
s
ss ,+
1 ss ,
= d3 k (k)a+
0 +
S s (k)as (k)+ (k)as (k)as (k) (10.32c)
2 s,s

=
0
L d3 k(x k x k )| {a+ +
s (k) as (k) as (k) as (k)}
k0 = m2 c2 +k2
s
ss ,+
1 ss ,
d3 k l (k)a+ +
s (k) as (k) + l (k)as (k) as (k)
2 s,s

1
+ i d k as (k) k k as (k)
3 +
2 s k k

+
+ a
s (k) k k as (k) 2 2 2 , (10.32d)
k k k0 = m c +k

1
P = k | {a+ +
s (k) as (k) as (k) as (k)
2 s k0 = m2 c2 +k2
(10.33a)
+ a+ +
3

1
Q
= + q {a+ +
s (k) as (k) + as (k) as (k)
2 s (10.33b)
a+ +
s (k) as (k) as (k) as (k)} d k.
3
ss ,
1
= +
0
S d3 k (k)[as + (k), a
s (k)]
4 s,s (10.33c)
ss ,+ +

+ (k)[a
s (k), as (k)]
0 1 0

L = L + 0L . (10.33d)
2
(To save some space, we do not write the evident, but rather long, explicit formula for
0
L; the reader can easily write it down with some patience. For the same reason we
do not write at all the total angular momentum operators; they are simply sums of the
corresponding spin and orbital (angular momentum) operators. Besides, we have omit the
evident expressions for the time-dependent part of the spin angular momentum (see (5.24),
which is insignificant here and below.) Notice, in the above formulae the operators a s
and as do not depend on the Lagrangians (10.16)(10.18) we started off. As one can
expect, the r.h.s. of (10.33a), (10.33c) and (10.33d) (resp. (10.33b)) is symmetric (resp.
antisymmetric) with respect to the change particleantiparticle.
Since the obtained expressions for the momentum operators corresponding to the La-
grangians (10.16)(10.18) are different, the field equations for them in terms of the creation
and annihilation operators a
s (k) and as (k) will be different, regardless of their coinci-
dence in terms of and (see (10.21)). These equations can be derive in the same way as
we did in Sect. 6 for the Lagrangian (10.2), but the results at our disposal give us the possi-
bility to write them without any calculations. As the expressions (10.31a) and (5.7) for the
momentum operators of (10.16) and (10.2), respectively, coincide and the Lagrangians L
and L are equal up to a change of the independent variables in them, from the decomposi-
tions (6.1), a similar one for (see (10.25) and (10.28)), and the derivation of (6.4) follows
that the field equations for the Lagrangian L , given by (10.16), in terms of a s and as

coincide with the ones for L , given by (10.2). Therefore, the equations of motion for L
in terms of creation and annihilation operators are (6.4). Since (10.31a) and (10.32a) differ
by the change a
s as , the field equations for L , given via (10.17), can be obtained

form (6.4) by making the replacement as (q) as (q), i.e. they are

as (k), at+(q) at (q)at (q) at + (q) a
3
(10.34a)

as (k), at+(q)at (q)at (q)at + (q) as (k)ts 3 (k q) = f st (k, q),

(10.34b)

where f st (k, q) and f st (k, q) are (generalized) functions such that

q q =m2 c2 +q2 f st (k, q) d3 q = 0
0
t
(10.34c)

q q =m2 c2 +q2 f st (k, q) d3 q = 0.
0
t
At last, as (10.33a) is the half sum of (10.31a) and (10.32a), the field equations for L ,
given via (10.18), in terms of creation and annihilation operators can be obtained from (6.4)
by replacing the first terms in it by the half sum of the first terms in left-hand-sides of (6.4)
and (10.34), i.e. these equations are:

as (k), at+ (q) at(q) at (q) at+ (q)

+
+ a +
s (k), at (q) at (q) at (q) at (q) (10.35a)

2a
3

as (k), at+ (q) at(q) at (q) at+ (q)

+ as (k), at+(q) at (q) at (q) at + (q) (10.35b)

2as (k)ts 3 (k q) = f st (k, q),

where f st (k, q) and f st (k, q) are (generalized) functions such that

q q =m2 c2 +q2 f st (k, q) d3 q = 0
0
t
(10.35c)
q
q0 =

m2 c2 +q2

f st (k, q) d3 q = 0.
t
Consider now the problem regarding the possible (anti)commutation relations for the
Lagrangians (10.16)(10.18). As we saw in Sect. 7, the field equations (6.4a)(6.4b)
are equivalent to (7.2) (with = 1). Similarly, applying (7.1), we can rewrite equiva-
lently (10.34a)(10.34b) and (10.35a)(10.35b) respectively as

[a + +

+
[a
+
(10.36a)

a
3
[as (k), at+(q)] at (q) at+ (q) [a

s (k), at (q)]
[as (k), at(q)] at + (q) + at (q) [a +
s (k), at (q)] (10.36b)

as (k)st 3 (k q) = f st (k, q)
+
[a
+

[a + +

+[a + +

(10.37a)
+
[a
+

2a
3
[as (k), at+ (q)] at (q) at + (q) [a

s (k), at (q)]
[as (k), at (q)] at+ (q) + at (q) [a +
s (k), at (q)]
[as (k), at+(q)] at (q) at+ (q) [as (k), at (q)] (10.37b)
[as (k), at(q)] at + (q) + at (q) [as (k), at+ (q)]

2as (k)st 3 (k q) = f st (k, q).
The analysis of (10.36) and (10.37) is practically identical with the one of (7.2) in
Sect. 7 with one very important new consequence from (10.37). Imposing the additional
conditions (7.4), we see that equations (10.37) reduce to
+ +
(1 + ) [a

[a + +
s (k), at (q)] at (q) + at (q) [as (k), at (q)] (10.38a)

2a
3
+
(1 + ) [a +

[as (k), at (q)] at+ (q) + at (q) [as (k), at+(q)] (10.38b)

2as (k)st 3 (k q) = f st (k, q).
Let us see what entails (10.38) for = 1, which corresponds to quantization of a

spinor field via commutators, not via anticommutators. Inserting (10.38) with = 1
into (10.35c), we get
k | a (k) =0 k | a (k) =0 for = 1, (10.39)

k0 = m2 c2 +k2 s k0 = m2 c2 +k2 s
which, by virtue of (10.33a), imply
P = 0 for = 1 (10.40)
which, in its turn, reduce the Dirac equations (10.21) to
mc = 0 mc = 0 for = 1. (10.41)
Thus, we see that, when one starts from the Lagrangian (10.18), the only free spinor
fields that are possible to be quantized via commutators are (3.1), in the massless case m = 0,
and the missing field = = 0 in the massive case m = 0. The former solutions of the
field equations were investigated at length in the previous sections and the latter solution,
= = 0, of the field equations is completely unphysical as it cannot lead to any physically
observable consequences. But, moreover, these solutions are rejected by the properties (7.5)
and (7.6) of the (generalized) functions in the right-hand-sides of the equations in (7.4).
Therefore, the Lagrangian (10.18) does not admit quantization by commutators, contrary to
the Lagrangians (10.16) and (10.17), and, consequently, for it we must put = +1 in (7.4).
The further analysis of (10.36) with = 1 and (10.37) with = +1 is practically
identical to the one of (7.2) and leads to the anticommutation relations (7.13).12
Consequently, the Lagrangians (10.16)(10.18) lead to identical anticommutation rela-
tions, viz. (7.13). The only difference between them is that (10.18) does not require as an
additional condition quantization via anticommutators, or an equivalent to it hypothesis, like
the spin-statistics theorem or charge symmetry, as this condition is encoded in it from the
very beginning. In short, assumption 7.3 is not needed in a case of the Lagrangian (10.18)
as it is a consequence from it.
Since the anticommutation relations for the Lagrangians (10.16)(10.18) are identical,
the vacuum and normal ordering for them should be introduced in an identical way, viz. as it
was described in Sect. 8. As a result of (10.31)(10.33), after normal ordering, the operators
of the dynamical variables for them become identical and are given by (8.4)(8.9).
Let us summarize. The Lagrangians (10.16)(10.18) (or (10.13)(10.15) in Heisenberg
picture) are equivalent in a sense that they entail identical final quantum field theories.
The principle difference between them is that in (10.18) is encoded additionally the charge
symmetry of (or, equivalently, the spin-statistics theorem for) a free spinor field and there
is not a need to impose it as additional condition on a later stage of theorys development.
11. Conclusion
In this chapter, we made a more or less comprehensive investigation of free spin 12 quantum
fields in momentum picture of motion in Lagrangian quantum field theory. The methods
we used and the results obtained are near (similar) to the ones in chapter II, concerning free
charged scalar fields. The main difference from the scalar case comes from the multi-com-
ponent character of a spinor field, which is due to its non-zero spin.
12 One can derive (7.13) without any calculations by taking into account that the l h.s. of (10.36) is equal to
the l h.s. of (7.2) combined with the change a

s (q) as (q) and that the l h.s. of (10.37) is equal to the sum
of the left-hand-sides of (7.2) and (10.36). This implies similar changes in (7.13) under which it is invariant.
We have written the field equations in terms of creation and annihilation operators. In
this form they turn to be trilinear algebraic(-functional) equations relative to the creation
and annihilation operators. On this base, the standard anticommutation relations were de-
rived under some explicitly presented additional conditions. Under these conditions, the
anticommutation relations are tantamount to the initial field equations for free spinor fields.
We have also briefly studied the state vectors, vacuum and normal ordering procedure for
these fields.
An analysis of the initial Lagrangian, from which the quantum theory of free spinor
field is derived, is presented. The consequences of several Lagrangians are described and
the best one of them is single out. It is the one which is charge symmetric, i.e. it is the one
which is invariant under the change particleantiparticle described in suitable variables.
This is the Lagrangian (10.15) in Heisenberg picture or (10.18) in momentum picture. It
entails, before quantization and normal ordering, the field equations (10.35) in terms of
creation and annihilation operators. Evidently, these equations can be rewritten as

+
+
[at (q), at(q)] , a
s (k) + [at (q), at (q)] , as (k)

(11.1a)
= 2a s (k)st (k q) f st (k, q)
3

+
+
[at (q), at(q)] , a
s (k) + [at (q), at (q)] , as (k)

(11.1b)
= 2as (k)st 3 (k q) f st (k, q).
Trilinear relations, like equations (11.1), are typical for the so-called parastatistics and
parafield theory [2933], where they play a role of (para)commutation relations. In chap-
ter V we shall show how from (11.1) the parafermi commutation relations for a free spinor
field can be derived.
We shall end with the evident remark that the particles and antiparticles of a free spinor
field (, described via some of the Lagrangians considered in this paper) are always different,
due to the fact that the charge of fields particles and antiparticles is q and +q, respectively,
and their spin and orbital (angular momentum) operators are always non-zero and different
(see, e.g, (9.9), (9.10), (5.27) and (5.38)).
Chapter IV
Free Vector Fields

Free vec-
tor fields, satis-
fying the Lorenz con-
dition, are investigated in
details in the momentum picture
of motion in Lagrangian quantum field
theory. The field equations are equivalently
written in terms of creation and annihilation op-
erators and on their base the commutation relations are
derived. Some problems concerning the vacuum and state vec-
tors of free vector field are discussed. Special attention
is paid to peculiarities of the massless case; in par-
ticular, the electromagnetic field is explored.
Several Lagrangians, describing free
vector fields, are considered and
the basic consequences of
them are compared
and pointed
out.

1. Introduction
This chapter is devoted to an exploration of two types of free vector fields in the momen-
tum picture of Lagrangian quantum field theory: massive vector fields and massless vector
fields, the latter satisfying the Lorenz1 condition as addition to the Lagrangian formalism.
Since the massive free vector fields satisfy the Lorenz condition as a consequence of the
Euler-Lagrange equations, both kinds of fields are treated on almost equal footing in the
present chapter. However, the massless case has its own peculiarities to which special at-
tention is paid. Most of the known results, concerning the mentioned fields in Heisenberg
picture, are reproduced in momentum picture of motion. New results are obtained too. For
example, the field equations in terms of creation and annihilation operators and the (second)
quantization of electromagnetic field in Lorenz gauge (imposed on the fields potentials as
operators), a special case of which is the quantization in Coulomb gauge.
Here is the organization of the material in this chapter.
The description of free vector fields in momentum picture is presented in Sect. 2.
The structure of the solutions of the field equations is analyzed in Sect. 3. Decompo-
sitions of these solutions, equivalent to the Fourier decompositions in Heisenberg picture,
are established. A suitably normalized system of classical solutions of the field equations
is constructed. The creation and annihilation operators for the fields considered are intro-
duced in Sect. 4 on a base of the decompositions just mentioned and a system of classical
solutions. A physical interpretation of these operators is derived from the Heisenberg rela-
tions, which are external to the Lagrangian formalism. At this point, the first problem with
the massless case, concerning the angular momentum operator, appears. In Sect. 5, the op-
erators of the dynamical variables of free vector fields (satisfying the Lorenz condition) are
calculated in Heisenberg picture of motion in terms of creation and annihilation operators
in momentum picture. Special attention is paid to the spin angular momentum operator and
the above mentioned problem is analyzed further.
In Sect. 6, the field equations are equivalently rewritten in terms of creation and annihi-
lation operators. As a consequence of them, the dynamical variables in momentum picture
are found. It should be mentioned, in the massless case, the creation and annihilation op-
erators corresponding to the degrees of freedom, parallel to the 4-momentum variable,
do not enter in the field equations. In Sect. 7, the commutation relations for free vector
fields satisfying the Lorenz conditions are derived. They also do not include the opera-
tors just-mentioned. The commutators between the components of spin angular momentum
operator and between them and the charge operator are calculated on the base of the estab-
lished commutation relations. It is pointed that these relations play a role of field equations
under the hypotheses they are derived. To the normal ordering procedure and definition of
vacuum is devoted Sect. 8. Problems, regarding state vectors and physical interpretation of
creation and annihilation operators in the Lagrangian formalism, are considered in Sect. 9.
Some peculiarities of the massless case are explored in Sect. 10. It is pointed that,
1 The Lorenz condition and gauge (see below equation (2.6b)) are named in honor of the Danish theoretical
physicist Ludwig Valentin Lorenz (18291891), who has first published it in 1867 [43] (see also [44, pp. 268-
269, 291]); however this condition was first introduced in lectures by Bernhard G. W. Riemann in 1861 as
pointed in [44, p. 291]. It should be noted that the Lorenz condition/gauge is quite often erroneously referred
to as the Lorentz condition/gauge after the name of the Dutch theoretical physicist Hendrik Antoon Lorentz
(18531928) as, e.g., in [5, p. 18] and in [45, p. 45].
Free Vector Fields 147
generally, new suppositions are required for the treatment of creation and annihilation op-
erators, connected with the degrees of freedom parallel to the 4-momentum, which are
the cause for the problems arising in the massless case. Two such hypotheses are analyzed.
The obtained formalism is applied to a description of the electromagnetic field. In fact,
it provides a new quantization of this field in which the Lorenz conditions are imposed
directly on the field operators, which is completely different with respect to the one used
in Gupta-Bleuler quantization. It is shown that for an electromagnetic field no problems
arise, due to a suitable definition of the normal ordering procedure. The basic relations of
quantum field theory of free electromagnetic field are written explicitly.
Sect. 11 contains a discussion of some Lagrangians suitable for description of free vec-
tor fields satisfying the Lorenz condition. The basic consequences of these Lagrangians are
pointed and compared. As a best Lagrangian is pointed the one which is charge-symmet-
ric and, hence, in which the spin-statistics theorem is encoded. It is proved that the quantum
field theories, arising from the considered Lagrangians, became identical after the normal
ordering procedure is applied.
In Sect. 12 is analyzed the role of the Lorenz condition, when studying massless free
vector fields. This is done by investigating a massless vector field with a Lagrangian equal
to the one of a massive vector field with vanishing mass and without imposing the Lorenz
condition as a subsidiary condition on the field operators.
Sect. 13 closes the chapter.
2. Description of Free Vector Field in Momentum Picture

A vector field U is described by four operators U := U (x), called its components, which
transform as components of a 4-vector under Poincare transformations. The operators U

are Hermitian, U = U , for a neutral field and non-Hermitian, U = U , for a charged
one. Since the consideration of U 0 , . . ., U 3 as independent scalar fields meets as an obsta-
cle the non-positivity of the energy (see, e.g., [1, 4.1] or [46, 2a]), the Lagrangian of a
free vector field is represented as a sum of the Lagrangians, corresponding to U 0 , . . ., U 3
considered as independent scalar fields, and a correction term(s) ensuring the energy pos-
itivity (and, in fact, defining U as spin 1 quantum field). As pointed in the discussion
in [1, 4.1 and 5.3], the Lagrangian of a free vector field (and, possibly, conditions addi-
tional to the Lagrangian formalism) can be chosen in different ways, which lead to identical
theories, i.e. to coinciding field equations and dynamical variables.1
Between a number of possibilities for describing a massive vector field of mass m = 0,
we choose the Lagrangian as [2]
m2 c4 c2 2 c2 2
L = U U ( U )( U )+ ( U )( U ), (2.1)
1 + ( U ) 1 + ( U ) 1 + ( U )
where the function takes care of is the field neutral (Hermitian) or charged (non-Hermi-
1 This may not be the case when interacting fields are considered.
tian) according to

1 for U = U (Hermitian (neutral) field)
( U ) := . (2.2)
0 for U = U (non-Hermitian (charged) field)
Since

( U ) ( U ) ( U ) ( U ) = { U ( U ) U ( U )},

the theory arising from the Lagrangian (I.4.1) is equivalent to the one build from [6, 11]
m2 c4 c2 2
( U )( U )

L = U U
1 + ( U ) 1 + ( U )
c
2 2

( U ) ( U )

+ (2.3)
1 + ( U )
m2 c4 c2 2
= U U F F ,
1 + ( U ) 2(1 + ( U ))
where
F := U U . (2.4)
The first two terms in (2.1) (or in the first row in (2.3)) correspond to a sum of four in-
dependent Lagrangians for each of the components U 0 , . . ., U 3 , considered as free scalar
fields [1, 2, 26]. The remaining terms in (2.1) or (2.3) represent the afore-mentioned cor-
rection which reduces the independent components (degrees of freedom) of a vector field
from 4 to 3 and ensures the positivity of the fields energy [1, 2, 11].
Before proceeding with the description in momentum picture, we notice that the Eu-
ler-Lagrange equations for the Lagrangians (2.1) and (2.3) coincide and are
U ) 2 ( U ) = m2 c2 U + 2 F
0 = m2 c2 U + 2 ( (2.5a)

0=
U ) 2 ( U )
m2 c2 U + 2 ( = m2 c2 U + 2

F , (2.5b)

where := is the DAlembert operator (in Heisenberg picture). For m = 0, these

equations, known as the Proca equations for U and U , can be written equivalently as
U = 0
(m2 c2 + 2 ) U = 0
(m2 c2 + 2 ) (2.6a)

U = 0 U = 0, (2.6b)
due to F 0 and m = 0, and show that there is a bijective correspondence between

U and F (in the case m = 0) [46, 2]. Therefore the field operators U and U are
solutions of the Klein-Gordon equations (2.6a) with mass m (= 0) and satisfy the condi-
tions (2.6b), known as the Lorenz conditions. We shall say that a vector field satisfies the
Lorenz condition, if the equations (2.6b) hold for it.
According to (I.6.2), the Lagrangians (2.1) and (2.3) in momentum picture are
m2 c4
L= U U
1 + ( U )
(2.7)
c2
+ [ U , P ] [ U , P ] [ U , P ] [ U , P]
1 + ( U )
m2 c4
L= U U
1 + ( U )
c2
+ [ U , P ] [ U , P ] [ U , P ] [ U , P ] (2.8)
1 + ( U )
m2 c4 c2 2
= U U F F ,
1 + ( U ) 2(1 + ( U ))
respectively, where
U (x) := U (x, x0) U (x) U 1(x, x0)

(2.9)
U (x) := U (x, x0) U (x) U 1(x, x0)

1 for U = U (Hermitian (neutral) field)
( U ) := = ( U ) (2.10)
0 for U = U (non-Hermitian (charged) field)
1
F(x) = U (x, x0) ( F (x)) U 1(x, x0) = {[ U , P ] [ U , P] }. (2.11)
i
F (x) = U (x, x0) ( F (x)) U 1(x, x0 )
1 (2.12)
= {[ U , P ] [ U , P ] } = ( F(x)) .
i
Regarding U and U as independent variables, from (2.7), we get2
L L
= = m2 c4 U
U U
L
: = = ic2 [ U , P ] ic2 [ U , P ]
y
(2.13)
L L

= = m2 c4 U
U U
L
: = = ic2 [ U , P] ic2 [ U , P ]
y
with
1 1
y := [U , P ] y := [ U, P ] .
i i
2 As noted on page xii, the derivatives in (2.13) are calculating according to the classical rules of commuting
variables, which requires additional rules for ordering the operators in the expressions for dynamical variables;
for details, see [3] and section I.3. The Lagrangian (2.8) has different derivatives, but leads to the same field
equations and dynamical variables and, for this reason, will not be considered further in this chapter.
Therefore the field equations (I.6.3) now read
m2 c2 U [[ U , P ] , P ] + [[ U , P ] , P ] = 0 (2.14a)
m2 c2 U [[ U , P ] , P ] + [[ U , P ] , P ] = 0 (2.14b)
or, using the notation (2.4),
m2 c2 U i[ F , P ] = 0 m2 c2 U i[ F

, P ] = 0. (2.15)
These are the systems of the Proca equations in momentum picture for a massive free
spin 1 (vector) fields.
Since the equality = (valid when applied on C2 functions or operators) in
momentum picture takes the form (see (I.5.10))
[[ , P ] , P ] = [[ , P ] , P ] , (2.16)
from (2.15), (2.16) and F = F follow the equalities
m2 [ U , P ] = 0 m2 [ U , P ] = 0. (2.17)
Consequently, in the massive case, i.e. m = 0, the system of Proca equations (2.14)
splits into the system of Klein-Gordon equations in momentum picture (see (II.2.8))
m2 c2 U [[ U , P ] , P ] = 0 m2 c2 U [[ U , P ] , P ] = 0 (2.18)
and the system of Lorenz conditions
[ U , P ] = 0 [ U , P ] = 0 (2.19)
for the field operators U and U . This result is a momentum picture version of (2.6).
From technical point of view, it is quite important as it allows a partial application of most
of the results obtained for free scalar fields, satisfying (systems of) Klein-Gordon equa-
tion(s), to the case of massive vector fields. Evidently, for the solutions of (2.18)(2.19),
the Lagrangian (2.7) reduces to3
m2 c4 c2
L= U U + [ U , P ] [ U , P ] , (2.20)
1 + ( U ) 1 + ( U )
which equals to a sum of four Lagrangians corresponding to U 0 , . . ., U 3 considered as free

scalar fields.
The above consideration show that the Lagrangian theory of massive free vector field
can be constructed equivalently from the Lagrangian (2.20) under the additional condi-
tions (2.19). This procedure is realized in Heisenberg picture in [1].
Consider now the above theory in the massless case, i.e. for m = 0. It is easily seen,
all of the above conclusions remain valid in the massless case too with one very important
exception. Namely, in it the equations (2.17) are identically valid and, consequently, in
3 The same result holds, up to a full divergence, for the Lagrangian (2.8) too.
this case the massless Proca equations, i.e. (2.14) with m = 0, do not imply the Klein-Gor-
don equations (2.18) and the Lorenz conditions (2.19).4 However, one can verify, e.g. in
momentum representation in Heisenberg picture, that the Lorenz conditions (2.19) are com-
patible with the massless Proca equations (2.14) with m = 0; said differently, the system of
Klein-Gordon equations (2.18) with m = 0 and Lorenz conditions (2.19), i.e.
[[ U , P ] , P ] = 0 [[ U , P ] , P ] = 0 (2.21)
[ U , P ] = 0 [ U , P ] = 0, (2.22)
does not contain contradictions and possesses non-trivial solutions. Moreover, one can con-
sider this system of equations as the one describing a free electromagnetic field in Lorenz
gauge before second quantization, i.e. before imposing a suitable commutation relations
between the fields components.
For these reasons, in the present chapter, with an exception of Sect. 12, we shall con-
sider a quantum field theory build according to the Lagrangian formalism arising from the
Lagrangian (2.7) to which, in the massless case, are added the Lorenz conditions (2.19)
as additional requirements. In other words, with an exception of Sect. 12, vector fields
satisfying the Lorenz conditions will be explored in this chapter.
A free vector field possesses energy-momentum, (possibly vanishing) charge, and angu-
lar momentum. The corresponding to them density operators, the energy-momentum tensor
T, current density J and (total) angular momentum density M , in Heisenberg picture
for the Lagrangian (2.7) are as follows:5
1
T = ( U ) + ( U ) L
1 + ( U )
(2.23)
c2 2
= ( U ) ( U ) + ( U ) ( U ) L = T
1 + ( U )
q
J = U U = iqc2 ( U ) U U ( U ) (2.24)
i
M = L

+ S , (2.25)
where the Lorenz conditions (2.19) were taken into account, q is the charge of the field (of
fields particles), and

L := x T x T (2.26)
1

U )

S := (I U ) + (I
1 + ( U )
2 c2
( U ) U ( U ) U (2.27)

=
1 + ( U )

U ( U ) + U ( U )

4 As it is well know [1, 4, 11], in this important case a gauge symmetry arises, i.e. an invariance of the
theory under the gauge transformations U

U + K , K being a C2 operator, in Heisenberg picture or
U
U + K + i1 [ K , P ] in momentum picture (see (I.5.10) and (I.5.14)).
5 As a consequence of (2.19), the expressions (2.23)(2.26) are sums of the ones corresponding to
U 0 ,... , U 3 considered as free scalar fields [1,2,11]. For a rigorous derivation of (2.23)(2.27), see the general
rules described in [3] or section I.3.
with the numbers

I = I = (2.28)
being characteristics of a vector field under 4-rotations [5, eq. (0-43)]. It should be noticed,
since the energy-momentum operator (2.23) is symmetric, T = T , the spin and orbital
angular momentum density operators satisfy the continuity equations

S = 0 L =0 (2.29)
and, consequently, the spin and orbital angular momentum operators of a free vector field
are conserved ones, i.e.
d d
S = 0 L = 0. (2.30)
dx0 dx0
According to (I.5.4), (2.9), (2.13), and (2.19), the densities of the dynamical character-
istics of a free vector field in momentum picture are:
c2
T = [ U , P ] [ U , P ] + [ U , P ] [ U , P ]
1 + ( U )
(2.31)
c2 2 2
m c U U + [ U , P ] [ U , P ]
1 + ( U )

J = qc2 [ U , P] U U [ U , P ] (2.32)

L = x T x T (2.33)
ic2
S = [ U , P ] U [ U , P ] U
1 + ( U ) (2.34)

U [ U , P ] + U [ U , P ] .
Gordon equations (2.21) and the expressions (2.31), (2.32) and (2.34).
Comparing (2.31) and (2.32) with the corresponding expressions for U 0 , . . ., U 3 , con-
sidered as free scalar fields (see chapter II), we see that the terms originating from U 1 ,
U 2 and U 3 enter in (2.31) and (2.32) with right signs if U 1 , U 2 and U 3 were free scalar
fields. But the terms, in which U 0 enters, are with signs opposite to the ones if U 0 was a
free scalar field. In particular, this means that the contribution of U 0 in the fields energy
is negative. All this points to the known fact that U 0 is a carrier of an unphysical degree
of freedom, which must be eliminated (via the Lorenz conditions (2.19)). The second new
moment, with respect to the scalar field case, is the existence of a, generally, non-vanishing
spin angular momentum density operator (2.34) to which a special attention will be paid in
Sect. 5.
Since U are solutions of the Klein-Gordon equations (2.18), the operator c12 [[ ,
P] , P ] has a meaning of a square-of-mass operator of the vector field under considera-
tion. At the same time, the operator c12 P P has a meaning of square-of-mass operator
of the fields states (state vectors).
We shall specify the relations (I.6.29) and (I.6.30) for a vector field by putting ( U ) =
+1 and ( U ) = 1. Therefore the relations (I.6.29) and (I.6.30), which are external for
the Lagrangian formalism, take the form
[ U , Q ] = q U [ U , Q ] = q U (2.35)
[ U , M (x, x0 )] = x [ U , P ] x [ U , P ] + i( U U ) (2.36a)

[ U , M (x, x0 )] = x [ U , P ] x [ U , P ] + i( U U ), (2.36b)
where (2.28) was used. It is clear, the last terms in (2.36) are due to the spin angular
momentum, while the other ones originate from the orbital angular momentum.
3. Analysis of the Field Equations

The analysis of the Dirac equations in section III.3 can mutatis mutandis be applied to the
case of vector fields satisfying the Lorenz condition. This can be done as follows
At first, we distinguish the degenerate solutions
[ U , P ] = 0 [ U , P ] = 0 for m = 0 (3.1)
of the Klein-Gordon equations (2.18), which solutions, in view of (I.5.16), (I.5.14), (I.5.18)
and (I.5.22), in Heisenberg picture read

U (x) = U (x0) = U (= const ) U (x) = U (x0 ) = U (= const) for m = 0. (3.2)
According to equations (2.31)(2.34), the energy-momentum, charge and angular mo-

mentum density operators for the solutions (3.1) respectively are:

T = 0 J = 0 L = S = M =0 (3.3)
Since (3.3) and (I.6.6)(I.6.9) imply
P = 0 Q = 0 L = S = M = 0, (3.4)
the solutions (3.1) (or (3.2) in Heisenberg picture) describe a massless vector field with
vanishing dynamical characteristics. Such a field cannot lead to any predictable observable
results and, in this sense is unphysical.1
The further analysis of the field equations will be done similarly to the one of free
spinor fields in section III.3. For the purpose, one should replace the Dirac equations with
the Lorenz conditions (2.19) and take into account that now the field equations are (2.18)
(2.19), not only (2.19).2
Taking into account the above facts, we can describe the structure of the solutions of
the field equations (2.18)(2.19) as follows.
1This case is similar to the one of free scalar fields describe in section II.3 see equation (II.3.1). Note, the
so-arising situation is completely different from a similar one, when free spinor Dirac fields are concerned as
in it solutions, like (III.3.1), are in principle observable see section III.3.
2 It is interesting to be noted, the Dirac equation i mc = 0, being the -matrices and a

4-spinor, in the massless case, m = 0, takes the form of a Lorenz condition, viz. U = 0 with U = .
Proposition 3.1. The solutions of the equations (2.18)(2.19) and (I.5.30) can be written
as (do not sum over !)

U = d3 k f,+(k) U (k)
2 + f , (k) U (k) (3.5a)
2 2 k0 =+ m c +k 2 2 2
k0 = m c +k

U = d3 k
f,+ (k) U (k) +
f, (k) U (k) (3.5b)
k0 =+ m2 c2 +k2 k0 = m2 c2 +k2
or, equivalently as

U = d4 k(k2 m2 c2 ) f (k) U (k) U = d4 k(k2 m2 c2 ) f (k) U (k). (3.6)
Here: k = (k0 , k1 , k2 , k3 ) is a 4-vector with dimension of 4-momentum, k2 = k k = k02

k12 k2 k32 = k02 k2 with k being the components of k and k := (k1 , k2 , k3 ) = (k1 , k2 , k3 )
being the 3-dimensional part of k, () is the (1-dimensional) Dirac delta function, the
operators U (k), U (k) : F F are solutions of the equations
[ U (k), P ] = k U (k) [ U (k), P ] = k U (k) (3.7a)

{k U (k)} 2 2 2 = 0
k =m c
{k U (k)} 2 2 2 = 0,
k =m c
(3.7b)

f ,(k) and f, (k) are complex-valued functions (resp. distributions (generalized func-
tions)) of k for solutions different from (3.1) (resp. for the solutions (3.1)), and f and f
are complex-valued functions (resp. distribution) of k for solutions different from (3.1) (resp.
for the solutions (3.1)). Besides, we have the relations

f (k)| = 2 m2 c2 + k 2 f,(k)
2 2 k0=
2
m c +k
and
f (k)| =2
m2 c2 + k 2 f, (k)
k0= m2 c2 +k2
for solutions different from (3.1).
Remark 3.1. Evidently, in (3.5) and (3.6) enter only the solutions of (3.7) for which
k2 := k k = k02 k 2 = m2 c2 . (3.8)
This circumstance is a consequence of the fact that U the solutions of the Klein-Gordon
equations (2.18).
Remark 3.2. Obviously, to the solutions (3.1) corresponds (3.7a) with P = 0. Hence
U (x, 0) = U (0) = const U (x, 0) = U (0) = const P = P = 0 (3.9)
with (see (I.5.4))
U (x, k) := U 1(x, x0) U (k) U (x, x0)

(3.10)
U (x, k) := U 1(x, x0 ) U (k) U (x, x0).
These solutions, in terms of (3.5) or (3.6), are described by m = 0 and, for example,

f ,(k) = f, (k) = ( 12 a)3 (k) for some a C or f (k) = f (k) such that f (k)|k0=|k| =
(1 2a)|k|3 (k), respectively. (Here 3 (k) := (k1 )(k2 )(k3 ) is the 3-dimensional Dirac
delta-function. Note the equality (y2 b2 ) = 1b ((y + b) + (y b)) for b > 0.)
Remark 3.3. Since U := ( U ) , from (3.5) (resp. (3.6)) is clear that there should ex-

ist some connection between f , (k) U (k) and f,
(k) U (k) with k0 = + m2 c2 +k2
(resp. between f (k) U (k) and f (k) U (k)). A simple examination of (3.5) (resp.
(3.6)) reveals that the Hermitian conjugation can either transform these expressions into
each other or change the signs plus and minus in them according to:

f (k) U (k)
= f, (k) U (k) (3.11a)
k0 = m c +k
2 2 2
k0 = m2 c2 +k2

f,(k) U (k) = f,(k) U (k) (3.11b)
k0 = 2 2 m c +k
2 2 2 2
k0 = m c +k

f (k) U (k) = f (k) U (k) (3.12a)

f (k) U (k) = f (k) U (k). (3.12b)
From the below presented proof of proposition 3.1 and the comments after it, it will be clear
that (3.11) and (3.12) should be accepted. Notice, the above equations mean that U (k) is
not the Hermitian conjugate of U (k).
Proof of proposition 3.1 The proposition was proved for the solutions (3.1) in remark 3.2.
So, below we suppose that (k, m) = (0, 0).
The equivalence of (3.5) and (3.6) follows from (y2 b2 ) = 1b ((y + b) + (y b)) for
b > 0.
Since U and U are solutions of the Klein-Gordon equations (2.18), the representa-
tions (3.5) and the equalities (3.11) and (3.12), with U (k) and U (k) satisfying (3.7a),
follow from proposition II.3.1 describing the structure of the solutions of the Klein-Gordon
equation in momentum picture.3
At the end, inserting (3.5) or (3.6) into (2.19), we obtain the equations (3.7b), due
to (3.7a).
From the proof of proposition 3.1, as well as from the one of proposition II.3.1, the fol-
lowing two conclusions can be made. On one hand, the conditions (3.7a) ensure that (3.5)
and (3.6) are solutions of (I.5.30) and the Klein-Gordon equations (2.18), while (3.7b) sin-
gle out between them the ones satisfying the Lorenz conditions (2.19). On other hand, since
up to a phase factor and, possibly, normalization constant, the expressions f (k) U (k) and

f (k) U (k) coincide with the Fourier images of respectively U (x) and U (x) in Heisen-
berg picture, we can write

U (x) = (k2 m2 c2 ) U (k)ei (x x0 )k d4 k
1
U = (k m c ) U (k) d k
2 2 2 4
(3.13)

and similarly for U (with U (k) = U (k)), where U (k) are suitably normalized
solutions of (3.7). Therefore, up to normalization factor, the Fourier images of U (x) and
U (x) are
1 1
U (k) = e i x0 k U (k) U (k) = e i x0 k U (k) (3.14)
3 One can prove the representations (3.5), under the conditions (3.7), by repeating mutatis mutandis the
proof of proposition II.3.1. From it the equalities (3.11) and (3.12) rigorously follow too.
where x0 is a fixed point (see chapter I). So, the momentum representation of free vector
field (satisfying the Lorenz condition) in Heisenberg picture is an appropriately chosen
operator base for the solutions of the equations (2.18)(2.19) and (I.5.30) in momentum
picture. This conclusion allows us freely to apply in momentum picture the existing results
concerning that basis in Heisenberg picture.
As anyone of the equations (3.7b) is a linear homogeneous equation with respect to U
and U , each of these equations has exactly three linearly independent solutions, which
will be labeled by indices s, s,t, . . . taking the values 1, 2 and 3, s, s ,t = 1, 2, 3.4 Define

the operator-valued vectors U s,() (k) and U s,() (k) , where s = 1, 2, 3 and the index ()

indicates the sign of k0 = m2 c2 + k 2 in (3.7b), as linearly independent solutions of the
equations

k
k
2 U s,() (k) = 0 2 U s,() (k) = 0. (3.15)
k0 = m c +k
2 2 2 2k0 = m c +k
As a consequence of (3.7a), they also satisfy the relations

[ U s,()(k), P ] = k U s,() (k) [ U s,() (k), P ] = k U s,()(k). (3.16)
Since any solution of the first (resp. second) equation in (3.7b) can be represented as a linear
combination of U s,()(k) and U s,()
(k), s = 1, 2, 3, we can rewrite (3.5) as (do not sum
over !)

U = d3 k f,s,+(k) U ,s,+ (k) + f,s, (k) U ,s, (k)
k0 =+

m2 c2 +k2
(3.17a)
s

U = d3 k f,s,+ (k) U ,s,+ (k)+ f,s,

(k) U ,s, (k)
k0 =+

m2 c2 +k2
, (3.17b)
s

where f ,s,(k) and f,s, (k) are some complex-valued (generalized) functions of k such
that

f,s, (k) U ,s, (k) 2 2 2 = f,(k) U (k) 2 2 2
k0 =+ m c +k k0 = m c +k
s
(3.18)
f,s,

(k) U ,s, (k)
k0 =+

m2 c2 +k2

= f, (k) U (k)
k0 =

m2 c2 +k2
.
s
In what follows, we shall need a system of classical, not operator-valued, suitably

normalized solutions of the equations (3.7b), which equations reflect the Lorenz condi-
tions (2.19). (Cf. section III.3, where similar solutions were introduced for a spinor field.)
The idea of their introduction lies in the separation of the frame-independent properties of
a free vector field from the particular representation of that field in a particular frame of
reference. It will be realized below in Sect. 4.
Consider the equation
k
2 v (k) = 0 (3.19)
k0 =+
2 2 m c +k
4 These indices, which will be referred as the polarization or spin indices, have nothing common with the
spacial indices a,b, = 1,2,3 labeling the spacial components of 4-vectors or tensors.
where v (k) is a classical 4-vector field (over the R3 k-space). This is a single linear and ho-
mogeneous equation with respect to the four functions v (k), = 0, 1, 2, 3. Therefore (3.19)
admits three linearly independent solutions. Define vs (k), with s = 1, 2, 3 and = 0, 1, 2, 3,
as linearly independent solutions of

k
2 v (k) = 0
s
(3.20)
k0 =+
2 2 m c +k
satisfying the conditions

ss ss 1 for m = 0
vs (k)v,s (k) = (1 0m s3 ) = (3.21)
1s + 2s for m = 0
where v,s (k) := vs (k). In more details, the relations (3.21) read

vs (k)v,s (k) = ss
for m = 0 (3.22a)

s ss if (s, s ) = (3, 3) ss

if s, s = 1, 2

,s
v (k)v (k) m=0 = = (3.22b)
0 if (s, s) = (3, 3) 0 otherwise
Exercise 3.1. Prove via a direct calculation that an explicit solution of (3.20)(3.21), for
k = 0, is provided by
v1a (k) = e1a (k) v2a (k) = e2a (k)

ka m2 c2 + k 2 1/2
v3a (k) =
k2 m2 c2 + k 0m
2
ka 2 2 1 for m = 0 (3.23)
mc
= m c + k 1
2
k 2 2 for m = 0
k
1 3 k2 1/2
vs0 (k) = ka vsa (k) =
m c + k 0m
2 2 2
3s
m2 c2 + k a=1
2
where the vectors e1a (k) and e2a (k) are such that
es (k) es (k) = esa (k)esa (k) = ss

a
(3.24)
e (k) v (k) =
s 3
ea (k)v3a (k) =
s
0 for s, s = 1, 2,
a
i.e. the 3-vectors v1 (k), v2 (k) and v3 (k) form an orthogonal (orthonormal for m = 0) basis
in the R3 k-space with v3 (k) being proportional to (having the direction of) k. Here 0m := 0
for m = 0 and 0m := 1 for m = 0. If k = 0, one can put
vs0 (0) = 0 v1a (0) = e1a (0) v2a (0) = e2a (0)

abc e1b (0)e2c (0) for m = 0 (3.25)
va (0) = (1 0m ) eb (0)ec (0) =
3 abc 1 2
bc 0 for m = 0
with e1a (0) and e2a (0) satisfying (3.24) with k = 0; in particular, one can put e1a (0) = 1a
and e2a (0) = 2a .5
Exercise 3.2. Applying the equations (3.14)(3.25) in a 3-frame specified by the unit
vectors defined by vsa (k), s = 1, 2, 3, prove that the solutions (3.23) and (3.25) of (3.20)
(3.21)satisfy the following relations for summation with respect to the polarization index
(k0 = m2 c2 + k2 ):
3 k k
vs (k)vs (k) k=0 = +
m c + k 2 0m
2 2
s=1

1 + 0m
for = = 0
1 0m for = = 1, 2, 3 (3.26a)

1 otherwise
3
vs (0)vs(0)
s=1

if m = 0 and , = 1, 2, 3 or if m = 0 and , = 1, 2
= (3.26b)
0 otherwise

for , = 1, 2
v (k)v(k) = 0 otherwise
s s
. (3.27)
s=1,2
k (k)
If we introduce the vectors := (1, 0, 0, 0) and k := 0, k 2 =
(k)k 2

k
with (k) := k and k k=0 := (0, 0, 0, 1), then the covariant form of (3.27) is
vs (k)vs (k) = + k k , (3.28)

s=1,2
which, for m = 0, coincides with [2, eq. (14.53)], where vs (k)|m=0 with s = 1, 2 is denoted
by (k, s), but it is supposed that 0 (k, s) := 0 and the value s = 3 is excluded by definition.
Notice, the last multiplier in (3.26a) can be written in a covariant form as (1 + 0m ). The
easiest way for proving these equalities for , = 1, 2, 3 is in a frame in which k1 = k2 = 0.
The rest of the equations are consequences of the ones with , = 1, 2, 3, (3.23) and (3.25).
4. Frequency Decompositions and

The frequency decompositions of a free vector field, satisfying the Lorenz condition, can
be introduced similarly to the ones of a free scalar field in section II.11, if (3.5) is used, or

The expression for v3a (k) in (3.23) is not defined for k = 0. Indeed,
5
the limit of k a / k 2 , when k 0,
depends on how k approaches to the zero vector 0; one can force ka / k to tends to any real number by an
2
appropriate choice of the limiting process k 0. For instance lim0 ka 2 = 1a +2a

2
if k = (1,,0) for
k 1+
some R.
of a free spinor field in section III.5, if (3.17) is used, i.e. if the spin of the field is taken into
account. Respectively, we put:

f (k) U (k) for k0 0
U (k) := ,

0 for k0 < 0
(4.1)
f
, (k) U
(k) for k 0
U (k) := 0
0 for k0 < 0

f,s,(k) U ,s,() (k) for k0 0
U ,s (k) :=
0 for k0 < 0
(4.2)
f
(k) U
(k) for k 0
U ,s (k) := ,s, ,s,() 0
,
0 for k0 < 0
where k2 = m2 c2 and s = 1, 2, 3. As a consequence of (3.18), we have

3 3
(k) = U ,s (k)
U U (k) = U ,s (k).

(4.3)
s=1 s=1
These operators satisfy the equations

U (k) = U (k) U (k) = U
(k) (4.4)
due to (3.11).
It will be convenient for the following the definitions (4.1) and (4.2) to be specified
when k0 = + m2 c2 + k2 0:
3
(k) := U (k) k =+ m2 c2 +k2 = U ,s (k)
U
0
s=1

U
,s (k) := U ,s (k) k0 =+ m2 c2 +k2
(4.5)
3
U (k) := U (k)k =+m2 c2 +k2
0
= U ,s (k)
s=1

U ,s (k) := U ,s(k)k =+m2 c2 +k2 .
0
Combining (4.1)(4.3), (3.5), (3.7) and (3.17), we get
U = U +
+ U U = U + + U (4.6)

:=
U d3 k U
,s (k) = d3 k U
(k)
s
(4.7)
U := d3 k U ,s (k) = d3 k U
(k)
s

[ U
(k), P ] = k k0 = m2 c2 +k2 U (k) (4.8a)

[ U (k), P ] = k 2 2 2 U (k). (4.8b)
k0 = m c +k
The equations (3.7b) are incorporated in the above equalities via (4.2), (4.3) and (3.17).
The physical meaning of the above-introduced operators is a consequence of (4.8) and
the equations
[ U
(k), Q ] = q U (k) [ U (k), Q ] = q U (k) (4.9)
[ U
(k), M (x)]
= {(x k x k ) + i( )} U
k0 = m2 c2 +k2 (k) (4.10a)
[ U (k), M (x)]

= {(x k x k ) + i( )} U (k), (4.10b)
k0 = m2 c2 +k2
which follow from (2.35) and (2.36). Recall, these equations are external to the Lagrangian
formalism, but, in general, they agree with the particle interpretation of the theory. There-
fore the below-presented results, in particular the physical interpretation of the creation and
annihilation operators, should be accepted with some reserve. However, after the establish-
ment of the particle interpretation of the theory (see Sect 9), the results of this section will
be confirmed.
Let X p , Xe and Xm denote state vectors of a vector field with fixed respectively 4-mo-
mentum p , (total) charge e and (total) angular momentum m (x), i.e.
P ( X p ) = p X p (4.11a)
Q ( Xe ) = e Xe (4.11b)
M (x)( Xm ) = m (x) Xm . (4.11c)
Combining these equations with (4.6)(4.10), we obtain

P U (k)( X p ) = (p k ) U
(k)( X p ) k0 = m2 c2 + k 2 (4.12a)

P U (k)( X p) = (p k ) U (k)( X p ) k0 = m2 c2 + k2 (4.12b)

Q U ( Xe ) = (e q) U ( Xe) Q U ( Xe ) = (e + q) U ( Xe) (4.13a)

Q U
( Xe ) = (e q) U ( Xe ) Q U ( Xe) = (e + q) U ( Xe) (4.13b)

Q U
(k)( Xe ) = (e q) U (k)( Xe )
(4.13c)
Q U (k)( Xe) = (e + q) U (k)( Xe)
M (x) U (k)( X m ) = m (x) (x k x k )
k0 = m2 c2 +k2

(4.14a)
i( ) U (k)( Xm )

M (x) U (k)( Xm ) = m (x) (x k x k )k =m2 c2 +k2

0
(4.14b)

i( ) U (k)( Xm ).
The equations (4.12) (resp. (4.13)) show that the eigenvectors of the momentum (resp.
charge) operator are mapped into such vectors by the operators U
(k) and U (k) (resp.
U, U
, U (k), U , U , and U (k)). However, by virtue of the equalities (4.14),

no one of the operators U , U

, U (k), U , U , and U (k) maps an eigenvec-

tor of the angular momentum operator into such a vector. The cause for this fact are
the matrices (2.28), 1 i.e. I := [ ]3,=0 , appearing in (4.14), which gen-
1 In the general case at this place the matrices I = [I ij ], entering into (I.6.14), should appear.
erally are non-diagonal and, consequently mix the components of the vectors U ( Xm ),
U
( Xm ), U (k)( Xm ), U ( Xm ), U ( Xm ), and U (k)( Xm ) in (4.14). Since the matri-

ces iI have a dimension of angular momentum and, obviously, originate from the pure
spin properties of vector fields, we shall refer to them as spin-mixing angular momentum
matrices or simply as spin-mixing matrices; by definition, the spin-mixing matrix of the
field U and its Hermitian conjugate U is iI. More generally, if X is a state vec-

tor and M (x) U
(k)( X ) = {l (x) + s } U (k)( X ) or M (x) U (k)( X ) =

{l (x) + s } U (k)( X ) where l and l

are some operators and s := [s]3,=0
and s := [s
],=0 are matrices, not proportional to the unit matrix 114 , with operator en-
3
tries, then we shall say that the operators U

(k) or U (k) have, respectively, spin-mixing
(angular momentum) matrices s and s relative to the state vector X ; we shall abbreviate
this by saying that the states U
(k)( X ) and U (k)( X ) have spin-mixing matrices s
and s , respectively.
The other additional terms in equations (4.14) are (x k x k ) 2 2 2 114 .
k0 = m c +k
They do not mix the components of U
(k)( Xm ) and U (k)( Xm ). These terms
may be associated with the orbital angular momentum of the state vectors U (k)

( Xm ) and U (k)( Xm).
Thus, from (4.12)(4.14), the following conclusions can be made:
i. The operators U + +
(k) and U (k) (respectively U (k) and U (k)) increase

(respectively decrease) the states 4-momentum by the quantity ( m c2 + k 2 ,
2
k).
ii. The operators U , U

and U (k) (respectively U , U and U (k)) decrease

(respectively increase) the states charge by q.
iii. The operators U + +

(k) and U (k) (respectively U (k) and U (k)) in-
crease (respectively decrease) the states orbital angular momentum by (x k
x k ) 2 2 2 .
k0 = m c +k
iv. The operators U

(k) and U (k) possess spin-mixing angular momentum matrices
iI relative to states with fixed total angular momentum.
In this way, the operators U
(k) and U (k) obtain an interpretation of creation and
annihilation operators of particles (quanta) of a vector field, viz.
(a) the operator U + (k) (respectively U
(k)) creates (respectively annihilates) a parti-
cle with 4-momentum ( m2 c2 + k , k), charge (q) (resp. (+q)), orbital angular
2
momen-
tum (x k x k )
2 , and spin-mixing angular momentum matrices iI and
k0 = 2 2m c +k
(b) the operator U + (k) (respectively U (k)) creates (respectively annihilates) a par-

ticle with 4-momentum ( m2 c2 + k 2 , k), charge (+q) (resp. (q)), orbital angular
momen-
tum (x k x k ) , and spin-mixing angular momentum matrices iI .
k0 = m c +k
2 2 2
,
Since U s,() (k) and U s,()(k) are, by definition, arbitrary linearly independent solu-
tions of (3.15), the operators U
,s (k) and U ,s (k) are linearly independent solutions of the
operator equations (see (4.1) and (4.6))

k 2 2 2 U ,s (k) = 0 k U ,s(k) = 0, (4.15)
k0 = m c +k k0 = m2 c2 +k 2
which, by virtue of (4.3), imply

k U k U (k) = 0.
k0 = m2 c2 +k2 (k) = 0 k0 = m2 c2 +k 2
(4.16)
The fact that the classical vector fields vs (k) are linearly independent solutions of the
same equations (see (3.20)) gives us the possibility to separate the invariant, frame-inde-
pendent, operator part of the field operators and the frame-dependent their properties by
writing (do not sum over s!)
1/2
U,s (k) := 2c(2) 3
m2 c2 + k 2 as (k)vs (k)
1/2 (4.17)
U ,s (k) := 2c(2)3 m2 c2 + k 2 as (k)vs (k),
which is equivalent to expend U

and U as
1/2 3
U
(k) := 2c(2)
3
m2 c2 + k 2 as (k)vs (k)
s=1
(4.18)
1/2 3
U (k) := 2c(2)3 m2 c2 + k 2 as (k)vs (k),
s=1
where a
s (k), as (k) : F F are some operators such that

(a
s (k)) = as (k) (as (k)) = as (k),

(4.19)
1/2
due to (4.4). The normalization constant 2c(2)3 m2 c2 + k 2 is introduced
in (4.17) and (4.18) for future convenience (see Sect. 5). The operators a+ s (k) and as + (k)
(resp. a
s (k) and as (k)) will be referred as the creation (resp. annihilation) operators (of
the field).
The physical meaning of the creation and annihilation operators is similar to the one of
U
and U . To demonstrate this, we insert (4.18) into (4.12)(4.14) and, using (3.22),
we get:

P a
s (k)( X p ) = (p k )as (k)( X p ) 1, 2, 3 for m = 0
s= (4.20a)
P as (k)( X p ) = (p k )as (k)( X p )

1, 2 for m = 0

Q a
s (k)( Xe ) = (e q)as (k)( Xe ) 1, 2, 3 for m = 0
s= (4.20b)
Q as (k)( Xe) = (e + q)as (k)( Xe) 1, 2 for m = 0

M(x) a
s (k)( Xm ) = {m(x) (x k x k )}as (k)( Xm )

3

+ i (k)at (k)( Xm)
st

t=1

M (x) as (k)( Xm ) = {m(x) (x k x k )}as (k)( Xm)

3

+ i st (k)at (k)( Xm )

(4.20c)

t=1

2 2

at (k) = 0 at (k) = 0 if m = 0
3t 3t

t=1 t=1

1, 2, 3 for m = 0
s=
1, 2 for m = 0
where
st (k) : = v,s (k)I

t
v (k) = v,s (k)(
)v (k)
t
(4.21)
= vs (k)vt (k) vs (k)vt (k) = st (k) = ts
(k)
with s,t = 1, 2, 3.2

As a consequence of (4.20), the interpretation of a
s (k) and as (k) is almost the same

as the one of U (k) and U (k), respectively, with an only change regarding the angular
momentum in the massless case. Equations (4.20) do not say anything about the dynamical
+
characteristics of the states a+
3 ( Xm ) and a3 ( Xm ) for a vanishing mass. All this indicates
possible problems with the degree of freedom arising from the value s = 3 of the polar-
ization variable in the massless case. Indeed, as we shall see below in Sect. 10, this is an
unphysical variable; this agrees with the known fact that a massless vector field possesses
only two, not three, independent components.
5. The Dynamical Variables in

Terms of Creation and Annihilation Operators
The Lagrangian (2.7) (under the Lorenz conditions (2.19)), the energy-momentum opera-
tor (2.31), the current operator (2.32), and the orbital angular momentum operator (2.33)
are sums of similar ones corresponding to the components U 0 , U 1 , U 2 and U 3 of a vector
field, considered as independent free scalar fields (see chapter II). Besides, the operators
U
(k) andU (k), defined via (4.5), (4.1) and (4.2), are up to the normalization constant
1/2
2c(2)3 m2 c2 + k2 equal to the creation/annihilation operators for U 0 , U 1 , U 2
and U 3 (considered as independent scalar fields described in chapter II). Consequently, we
can automatically write the expressions for the momentum, charge and angular momentum
2
Notice, the last equations in (4.20c), valid only in the massless case, impose, generally, 6 conditions on

any one of the pairs of operators a+ +
1 (k) and a2 (k) and a1 (k) and a2 (k). However, one should not be worried
about that as these conditions originate from the external to the Lagrangian formalism equations (I.6.30) and,
consequently, they may not hold in particular theory based on the Lagrangian formalism; see the paragraph
containing equations (5.18) below.
operators in terms of U
(k) and U (k) by applying the results obtained in chapter II
for arbitrary free scalar fields. In this way, we get the following representations for the
momentum operator P , charge operator Q and orbital angular momentum operator L :1

1
P = k 2c(2)3k0
1 + ( U ) k0 = m2 c2 +k2 (5.1)
{ U + (k) U , (k) + U (k) U ,+ (k)} d3 k

Q = q d3k 2c(2)3 m2c2 + k 2
(5.2)
U + (k) U , (k) U (k) U ,+ (k)

i
L = x0 P x0 P d3 k 2c(2)3 m2 c2 + k2
2(1 + ( U ))

U + (k)
k k U , (k)
k k

U (k) k k U ,+ (k) 2 2 2 , (5.3)
k k k0 = m c +k
where

A(k) B(k)
A(k)k B(k) := k B(k) + A(k) k
k k k

= k A(k) B(k) (5.4)
k
for operators A(k) and B(k) having C1 dependence on k (and common domains). 2
Now we shall express these operators in terms of the creation and annihilation operators
as (k) and as (k), introduced in Sect. 4. For the purpose, one should substitute (4.18)
into (5.1)(5.3) and to take into account the normalization conditions (3.22) for the vectors
vs (k), s = 1, 2, 3. The result of this procedure reads:
1
P =
1 + ( U )
30m
k |
k0 =

m2 c2 +k2
{a+ +
s (k) as (k) + as (k) as (k)} d k
3
(5.5)
s=1

1 The choice of the Lagrangian (2.7) corresponds to the Lagrangian L and energy-momentum operator
T(3) in chapter II. So, the (3)
below-presented operators are consequences of the expressions for P , Q ,
(3)
(3)
and L , in chapter II.
{F F } {F F } is a mapping on the operator space over the systems
2 More generally, if :

used for = .
30m
Q = q {a+ +
3
(5.6)
s=1
30m
1
L = d3 k(x0 k x0 k )|k0=m2 c2+k2
1 + ( U ) s=1
{as + (k) a +
s (k) + as (k) as (k)}
3
i
+
1 + ( U ) s,s =1
ss
d3 k l (k) as + (k) a +
s (k) as (k) as (k)

i 30m
+
2(1 + ( U )) s=1
d k 3
k
as + (k)
k k a
k s (k)

as (k) k k a+s (k) , (5.7)
k k k0 = m2 c2 +k2
where

ss 1 s ss s s
l : = v (k) k k v,s (k) = l
(k) (k) = l (k)
2 k k
vs (k) vs (k)
= k k v,s (k) (5.8)
k k
v,s (k) v,s (k)

= +vs (k) k k .
k k

with the restriction k0 = m2 c2 + k 2 done after the differentiation (so that the derivatives
with respect to k0 vanish). The last two equalities in (5.8) are consequences of (see (3.22))

vs (k) ,s v,s (k)
v (k) + v (k)
s
= 0, (5.9)
k k
so that

,s vs (k) ,s v,s (k)
vs (k)k v (k) = 2k v (k) = 2k vs
(k) . (5.10)
k k k
Since, vs (k) are real (see (3.19)(3.25)), the definition (5.8) implies

ss
l (k) = l
ss
(k) = l
ss
(k), (5.11)

where the asterisk denotes complex conjugation. So, l ss (k) are real and, by virtue
of (4.19), the sums of the first/second terms in the last integrand in (5.7) are Hermitian.
A peculiarity of (5.5)(5.7) is the presence in them of the Kronecker symbol 0m , which
equals to zero in the massive case, m = 0, and to one in the massless case, m = 0. Thus,
in the massless case, the modes with polarization s = 3 do not contribute to the momentum
and charge operators, but they do contribute to the orbital angular momentum operator only
via the numbers (5.8) in the second term in (5.7). Notice, in this way the arbitrariness in
the definition (3.20)(3.21) of the vectors vs (k) enters in the orbital angular momentum
operator. This is more or less an expected conclusion as the last operator is generally a
frame-dependent object.
Let us turn now our attention to the spin angular momentum operator (I.6.9) with density
operator S given by (2.34). Substituting (4.6)(4.7) into (2.34), we get the following
representation of the spin angular momentum density operator:

ic2

S = d3k d3 k k 2 2 2 U + (k)+ U (k) U + (k )+ U
(k )
1 + ( U ) k0 = m c +k
+
+
k 2 U (k) + U (k) U (k ) + U (k )
k0 = m c +k
2 2

k 2 2 2 U + (k) + U (k) U +

(k ) + U (k )
k0 = m c +k

+ k 2 2 2 U + (k) + U (k) U +
(k ) + U (k ) . (5.12)
k0 = m c +k
We shall calculate the spin angular momentum operator in Heisenberg picture by insert-
ing (5.12), with = 0, into (I.6.9). Then one should move the operators U
and U to
the right of U (x, x0 ) according to the relation

(k) (k ) U (x, x0 ) = e i (x x0 )(k+ k ) U (x, x0 ) (k) (k )
1
(5.13)

where , = +, , k0 = m2 c2 + k2 , k0 = m2 c2 + (k )2 , (k) = U (k), U (k), and
U (x, x0) being the operator (I.5.1) by means of which the transition from Heisenberg to
momentum picture is performed. This relation is valid for any (k) such that [ (k), P ] =
k (k) see (4.8) and equation (II.5.4). At last, performing the integration over x, which
results in the terms (2)33 (k k ), and the trivial integration over k by means of the
-functions 3 (k k ), we find:

i
S = d3 k 2c(2)3 m2 c2 + k2 U + (k) U
(k)
1 + ( U )

+ U (k) U + + +
(k) + U (k) U (k) U (k) U (k) . (5.14)
To express S via the creation and annihilation operators, we substitute (4.18) into (5.14)
and get
3 +
i
1 + ( U ) s,s
+
S = (k) as (k) as (k) as (k) as (k) ,
d3 kss (5.15)
=1
where the functions
st (k) := vs (k)vt (k) vs (k)vt (k) = st (k) = ts

(k) = + (k)
ts
(5.16)
were introduced earlier by (4.21).

From (5.15), we observe that S, generally, depends on the mode with polarization
index s = 3 even in the massless case. Evidently, a necessary and sufficient condition for
the independence of the spin angular momentum form this mode is
s3 s3 +
d3 k (k)as + (k) a
3 (k) (k)as (k) a3 (k)
s=1,2
3s
+ a3 + (k) 3s +
(k)as (k) a3 (k) (k)as (k) = 0. (5.17)
In particular, (5.17) is fulfilled in the massless case if

2 2
3sas (k) = 0 3sas (k) = 0 for m = 0, (5.18)
s=1 s=1
which are exactly the conditions appearing in (4.20c).3 Thus, a strange situation arises: the
massless modes with s = 3 do not contribute to the momentum and charge operators, but
they do contribute to the spin and orbital angular momentum operators unless additional
conditions, like (5.18), are valid. However, one can prove that (5.18) is either equiva-
lent to a
s (k) = as (k) = 0, if s = 1, 2 and (k, m) = (0, 0), or it is identically valid, if
(k, m) = (0, 0). In view of (5.5)(5.8) and (5.15), this means that (5.18) may be valid only
4
for free vector fields with vanishing momentum, charge and spin and angular momentum
operators, which fields are completely unphysical as they cannot lead to any physically ob-
servable/predictable results. Thus, only for such unphysical massless free vector fields the
Heisenberg relations (2.36) (for the Lagrangian (2.7)) may be valid. For these reasons, the
relations (2.36) and (5.18) will not be considered further in the present chapter.
However, the above conclusions do not exclude the validity of the more general
than (5.18) equation (5.17). We shall return on this problem in Sect. 10.
Example 5.1. To get a concrete idea of the spin angular momentum operator (5.15), we
shall calculate the quantities (5.16) for particular choices of the vectors e1 (k) and e2 (k)
in (3.23)(3.24).
To begin with, we notice that, as a result of the antisymmetry of st (k) in the super-
3(31) 4(41)
scripts and subscripts, only 2 2 = 18 of all of the 32 42 = 144 of these quanti-
ties are independent. As such we shall choose the ones with (s,t) = (1, 2), (2, 3),(3,1) and
(, ) = (0, 1), (0, 2), (0, 3), (1, 2), (2, 3), (3, 1).
For k = 0, we choose a frame such that

k = (0, 0, 0, k3) 3 2 e1a (k) = 1a e2a (k) = 2a . (5.19)
k = k 0
3 Recall, the equation (5.18) was derived in Sect. 4 on the base of the relations (2.36), which are external to
the Lagrangian formalism.

4 Let k = 0. Substituting (5.16), with t = 3, into the first equation in (5.18) and, then, using (3.23), we find
(ka e1b (k) kb e1a (k))a 2 2

1 (k) + (kaeb (k) kb ea (k))a2 (k) = 0 a,b = 1,2,3. (*)
Multiplying this equality with e1a (k) or e2a (k) and summing over a = 1,2,3, we, in view of (3.24), get kb a
1 (k) =
kb a
2 (k) = 0 for any b = 1,2,3. Therefore a1 (k) = a2 (k) = 0, as we supposed k = 0. If k = 0, repeating the
above method by using (3.25) for (3.23), we get
(1 0m){(3a e1b (0) 3b e1a (0))a 3 2 3 2

1 (0) + (a eb (0) b ea (0))a2 (0)} = 0 (**)
instead of (*). This equation is identically valid for m = 0, but for m = 0 it, by virtue of (3.24), implies

a
1 (k) = a2 (k) = 0. The assertion for the operators a1 (k) and a2 (k) can be proved similarly; alternatively,
it follow from the just proved results and (4.19).
1/2
m2 c2 +k2
Let := k
3
m2 c2 +k2 0m
= v3,3 (k) = v33 (k). The results of a straightforward calcu-
k2
lation, by means of (5.16), of the chosen independent quantities st (k) are presented in
table 5.1. Similar results for k = 0 in a frame in which
vs0 (0) = 0 e1a (0) = 1a e2a (0) = 2a v3a (0) = 3a (1 0m ) (5.20)
are given in table 5.2.
Table 5.1. The quantities (5.16) for k = 0 in the basis (5.19).
(,)
(s,t) (0,1) (0,2) (0,3) (1,2) (2,3) (3,1)
(1,2) 0 0 0 1 0 0
(2,3) 0 k3
0 0 0
m2 c2 +k2
(3,1) k3
0 0 0 0
m2 c2 +k2
Table 5.2. The quantities (5.16) for k = 0 in the basis (5.20).
(,)
(s,t) (0,1) (0,2) (0,3) (1,2) (2,3) (3,1)
(1,2) 0 0 0 1 0 0
(2,3) 0 0 0 0 1 0m 0
(3,1) 0 0 0 0 0 1 0m
Consider now the so-called spin vector(s). Since S is antisymmetric in and , S =

S , the spin angular momentum operator has 6 independent components which define
two 3-dimensional vectors, viz.
3
1 1 abc bc
Ra := S0a S a := abc S bc =
2 2
S (5.21)
b,c=1
where abc is the 3-dimensional Levi-Civitas symbol (which equals to +1 (resp. -1) if
(a, b, c) is an even (resp. odd) permutation of (1, 2, 3) and to zero otherwise). Defining
the cross (vector) product of 3-vectors A and B in Cartesian coordinates by (A B)a :=
abcAb Bc , where Ab = Ab are the covariant Cartesian components of A = (A1 , A2 , A3 ),
from (5.14) we find

i
R = d3 k 2c(2 m2 c2 + k 2 )3 U 0 + (k) U (k)
1 + ( U ) (5.22)
+ + +

+ U 0 (k) U (k) + U (k) U 0 (k) U (k) U 0 (k)

i
S = d3 k 2c(2 m2 c2 + k2 )3
1 + ( U ) (5.23)
+

+

U (k) U (k) U (k) U (k) ,
where U (k) := ( U 1,(k), U 2, (k), U 3, (k)) := ( U

1 (k), U 2 (k), U 3 (k)), U

(k)

:= ( U 1 (k), U 2 (k), U 3 (k)), and means a cross product combined with operator

composition, e.g. (A B)1 = A2 B3 A3 B2 for operator-valued vectors A and B.
To write the spin vectors R and S in terms of creation and annihilation operators, we
notice that the vectors es (k) := (es1 (k), es2 (k), es3 (k)), s = 1, 2, 3, with e3 (k) := e1 (k) e2 (k)
form an orthonormal basis of the R3 k-space, such that (see (3.23)(3.25) and do not sum
over s)
vs (k) := (vs1 (k), vs2 (k), vs3 (k)) = s (k)es (k)

m2 c2 +k2 1 for m = 0 (5.24)
(k) = (k) := 1 (k)k=0 :=
1 2 3
(0) :=
3
.
m2 c2 +k 2 0m 0 for m = 0
Then, we have
1 abc st
bc (k) = abcvsb (k)vtc (k) = s (k)t (k)abcesb (k)etc(k)
2 a
= s (k)t (k) es (k) et (k)
kb
st0a (k)k=0 = vs0 (k)vta (k) vsa (k)vt0 (k) = stba (k)
m2 c2 + k 2
kb 1
= baccd f std f (k)
22
m c +k
2 2
kb
= + abc(vs (k) vt (k))c
m2 c2 + k 2
st0a (0) = 0,
where (3.23), (3.25), and the equality abce f c = ea bf eb af were applied. Therefore,
from (5.21) and (5.15), we get
c
i
Ra = abc d3 krb (k) a +(k) a (k) a (k) a+ (k) (5.25)
1 + ( U )

i
S = d3 k a + (k) a (k) a (k) a+(k) , (5.26)
1 + ( U )
where the operator-valued vectors
3 3
a (k) := vs (k)as (k) a (k) := vs (k)as (k) (5.27)
s=1 s=1
were introduced, the index c in { }c means the cth component of { }, and the function

kb
for (k, m) = (0, 0)
r (k) :=
b m2 c2 +k 2 (5.28)
0 for (k, m) = (0, 0)
takes care of the above-obtained expressions for st0a (k).

In connection with the particle interpretation of the creation and annihilation operators,
the component S 3 is of particular interest as, for k = 0, its integrand describes the spin
projection on the direction of the 3-momentum k. From (5.26), we obtain

i
S =
3
d3 k a1 + (k) a +
2 (k) a2 (k) a1 (k)
1 + ( U ) (5.29)
+ +

a1 (k) a2 (k) + a2 (k) a1 (k) .
Since in (5.29) enter only mixed products, like a1 (k) a 2 (k), the states/(anti)parti-
cles created/annihilated by as (k) and a (k) for s = 1, 2 do not have definite projection on
s
k (for k = 0). This situation can be improved by introducing operators b
s (k) and bs (k)
such that [1, eq. (4.28)]
b
1 (k) + b2 (k) b
1 (k) b2 (k)
a
1 (k) = a
2 (k) = i a
3 (k) = b3 (k)
2 2
(5.30)
b (k)+b2 (k) b (k)b2 (k)
a1 (k) = 1 a2 (k) = +i 1 a3 (k) = b3 (k).
2 2
In terms of the operators b
s (k) and bs (k), the momentum (5.5), charge (5.6) and third
spin vector projection (5.29) operators take respectively the forms:
30m
1
P = k |k0=m2 c2+k2 {b+
1 + ( U ) s=1
+
s (k)bs (k)+bs (k)bs (k)} d k
3
(5.31)
30m
Q = q {bs +(k) b +
s (k) bs (k) bs (k)} d k
3
(5.32)
s=1
3
2
S 3 =
1 + ( U ) s=1
(1)s+1 bs + (k) b +
s (k) bs (k) bs (k) d k. (5.33)

From these formulae is clear that the states (particles) created/annihilated by bs (k) and
bs (k) have 4-momentum ( m2 c2 + k 2 , k), charge q, and spin projection on the direction
of movement equal to (1 + ( U ))1 for s = 1, 2 or equal to zero if s = 3.5
We would like now to make a comparison with the expressions for the dynamical vari-
ables in terms of the creation/annihilation operators a
s (k) and as (k) in (the momentum
representation of) Heisenberg picture of motion [1, 4, 5, 11]. As a consequence of (3.14),
the analogues of the creation/annihilation operators, defined in terms of the vector field
frequency operators via (4.1) and (4.2), are
1 1
U ,s(k) = e i x0 k U
,s (k) U ,s (k) = e i x0 k
U
,s (k) k0 = m2 c2 + k 2
1 1
(5.34)
U (k) = e i x0 k U
(k) U (k) = e i x0 k
U
(k) (k 0 = m 2 c2 + k 2 )
in Heisenberg picture. Therefore, defining (cf. (4.17))

1/2
U ,s (k) =: 2c(2)3 m2 c2 + k 2 as (k)vs (k)
(5.35)
U ,s (k) =: 2c(2)3 m2 c2 + k 2 1/2as (k)vs(k),

5 One can get rid of the factor (1 + ( U ))1 by rescaling the operators b
s (k) and bs (k) by the factor
(1 + ( U )) .
1/2
we get the creation/annihilation operators in Heisenberg picture as

1 1
a
s (k) = e
i x0 k
as (k) as (k) = e i x0 k as (k) k0 = m2 c2 + k 2 . (5.36)
Evidently, these operators satisfy the equations

a
s (k) = as (k) as (k) = a
s (k), (5.37)
due to (4.19), and have all other properties of their momentum picture counterparts de-
scribed in Sect. 4.
The connection (I.5.4) is not applicable to the creation/annihilation operators, as well
as to operators in momentum representation (of momentum picture), i.e. to ones depending
on the momentum variable k.
Example 5.2. Verify, by using the results of sections 3 and 4, the formulae

a
1
i
U (x, x0) a 1
s (k) = e s (k) U (x, x0 )
x k
1
k0 = m2 c2 + k2 , (5.38)
as (k) = e i x k U (x, x0) as (k) U 1(x, x0 )
from which equations (5.36) follow for x = x0 . (Notice, the right hand sides of the equa-
tions (5.38) are independent of x, due to the Heisenberg relations (I.4.1).)
From (5.5), (5.6), (5.15) and (5.34)(5.36), it is clear that all of the obtained expressions
for the momentum, charge and spin angular momentum operators in terms of the (invariant)
creation/annihilation operators remain unchanged in Heisenberg picture; to obtain a Heisen-
berg version of these equations, one has formally to add a tilde over the creation/annihilation
operators in momentum picture. However, this is not the case with the orbital operator (5.7)
because of the presents of derivatives in the integrands in (5.7).
Exercise 5.1. Prove that, in terms of the operators (5.36), in (5.7) the term x0 P x0 P ,
i.e. the first sum in it, should be deleted and tildes over the creation/annihilation operators
must be added. Correspondingly, equation (5.7) will read
3
i
L =
1 + ( U ) s,s =1
ss
d3 k l (k) as +(k) a +
s (k) as (k) as (k)

i 30m
+
2(1 + ( U )) s=1
3
d k as + (k)
k k a
k k s (k)

as (k) k k a+
s (k) . (5.39)
k k k0 = m2 c2 +k2

For a free vector field (satisfying the Lorenz condition), the equalities (2.18), (2.19), (I.6.6)
and (2.31) form a closed algebraic-functional system of equations for determination of the

field operators U and U . As these operators and the fields dynamical variables are
expressible in terms of the creation and annihilation operators, it is clear that the mentioned
system of equations can equivalently be represented in terms of creation and annihilation
operators. The derivation of the so-arising system of equations and some its consequences
are the main content of this section.
As a result of (4.6), (4.7) and (4.17) (or (4.18)), we have the decompositions
3
U = d3 k U +
,s (k) + U ,s (k)
s=1
3 1/2 + s
= d3 k 2c(2)3 m2 c2 + k 2 as (k) + a
s (k) v (k)
s=1
(6.1)
3
U = d k U ,s+ (k) + U ,s (k)
3
s=1
3 1/2 +
= d3 k 2c(2)3 m2 c2 + k 2 as (k) + as (k) vs (k).
s=1
The definitions of the quantities entering in these equations ensure the fulfillment of the
equations (3.7b). Therefore the conditions (3.7a), which are equivalent to (4.8), are the only
restrictions on the operators a
s (k) and as (k). Inserting (4.18) into (4.8), multiplying the

result by v,s (k), summing over , and applying the normalization conditions (3.22), we
obtain:
30m
[a
s (k), P ] = k as (k) = q a
s (k)st (k q) d q
3 3
(6.2a)
t=1
30m
[as (k), P ] = k as (k) = q as (k)st 3 (k q) d3 q (6.2b)
t=1

1, 2, 3 for m = 0
s= k0 = m2 c2 + k2 q0 = m2 c2 + q2 . (6.2c)
1, 2 for m = 0
Substituting in (6.2) the equation (5.5), with integration variable q for k and summation
index t for s, we get (do not sum over s!)
30m

q q =m2 c2 +q2 a
0
+ +
s (k), at (q) at (q) + at (q) at (q)
(6.3a)
t=1

(1 + ( U ))a
s (k)st (k q)
3
d q=0
3
30m

q q =m2 c2 +q2 as (k), at + (q) at(q) + at (q) at+(q)
0 (6.3b)
t=1

(1 + ( U ))a
s (k)st (k q)
3
d q = 0.
3

1, 2, 3 for m = 0
s= . (6.3c)
1, 2 for m = 0
Consequently, the operators a

s (k) and as (k) must be solutions of

as (k), at + (q)at (q) + at (q) at+ (q)
(6.4a)
(1 + ( U ))a
3

as (k), at + (q)at (q) + at (q) at+ (q)

(6.4b)
(1 + ( U ))a
3

1, 2, 3 for m = 0
s,t = , (6.4c)
1, 2 for m = 0
where fst (k, q) and fst (k, q) are (generalized) functions such that
30m
q q =m2 c2 +q2 fst (k, q) d3 q = 0
0
t=1
(6.4d)
30m
q f (k, q) d3 q =
q0 = m2 c2 +q2 st
0.
t=1
Since any solution of the field equations (2.18)(2.19) can be written in the form (6.1)
with a
s (k) and as (k) being solutions of (6.4) or, equivalently, of (6.3), the system of
equations (6.4) or (6.3) is equivalent to the initial system of field equations, consisting of
the Klein-Gordon equations (2.18) and the Lorenz conditions (2.19).1 In this sense, (6.4)
or (6.3) is the system of field equations (of a vector field satisfying the Lorenz condition)
in terms of creation and annihilation operators in momentum picture. If we neglect the
polarization indices, this system of equations is identical with the one for an arbitrary free
scalar field, obtained in section II.12. The reader may also wish to compare (6.4) with a
similar system of field equations for a free spinor field, found section III.6.
It is important to be mentioned, in the massless case, m = 0, the field equations (6.4)
contain only the polarization modes with s = 1, 2 and, consequently, they do not impose any

restrictions on the operators a3 (k) and a3 (k). We shall comment on this phenomenon in
Sect. 10.
The commutators of the dynamical variables with the momentum operator can easily be
found by means of the field equations (6.4). Indeed, from (5.5), (5.6), and (5.15) is evident
that for the momentum, charge and spin operators these commutators are expressible as
integrals whose integrands are linear combinations of the terms
30m
d3 kk
,
Bss (q) := [as (q) a +
s (q), at (k)
k0 = m2 c2 +k 2
t=1
at (k) + at (k) at+(k)]

, (6.5)
if m=0
where s, s = 1,2,3
1,2 if m=0 for the momentum and charge operators and s, s = 1, 2, 3 for the
spin angular momentum operator. Applying the identity [A B,C] [A,C] B + A [B,C] ,
1 Recall, the Heisenberg relations (I.4.1) are incorporated in the constancy of the field operators and are
reflected in (3.7a) or (6.2a).

for operators A, B and C, and (6.3) (which is equivalent to (6.4)), we get (do not sum over s
and s !)
30m
d3 kk
,
Bss (q) := (1 + ( U )) (st s t )3 (k q)a
s (q) as (q)
k0 = m2 c2 +k2
t=1
30m
= (1 + ( U ))q q =m2 c2 +q2 (st s t )as (q) a
s (q)
0
t=1
1,2,3 if m=0
for s, s = 1,2 if m=0 . Obviously, the summation over t in the last expression results in the
multiplier ss s s = 1 1 0, and hence

, 1, 2, 3 if m = 0
Bss (q) = 0 for s, s = . (6.6a)
1, 2 if m = 0
Similarly, when m = 0 and s or s is equal to 3, but (s, s) = (3, 3), we find

Bss (q)m=0 = (1 + ( U ))qq =m2 c2 +q2
,

0

a (q) as (q) if s = 3 and s = 1, 2
3
as (q) a 3 (q) if s = 1, 2 and s = 3

[a (q), P ] a
s (q) if s = 3 and s = 1, 2
+(1 + ( U )) 3 . (6.6b)
as (q) [a3 (q), P ]

if s = 1, 2 and s = 3
,
At last, the case s = s = 3 is insignificant for us as the quantities B33 (q) have a vanish-
ss
ing contribution in (5.15) and (5.7), due to ss 33 (k) = l33 (k) = 0 (see (5.16) and (5.8)). Now,
it is trivial to be seen that (5.5), (5.6), (5.15), and (5.8), on one hand, and (6.6), on another
hand, imply the commutation relations
[ P , P ] = 0 (6.7)
[ Q , P ] = 0 (6.8)
[ S, P] = 0m SC (6.9)
[ L, P] = i{ P P } 0m LC (6.10)
where
,
i
S
C :=
(1 + ( U ))2 s=1,2 (k) B3s (k)
d3 k3s

B,+ , ,+
3s (k) Bs3 (k) + Bs3 (k) m=0

(6.11)
2i
L
C :=
(1 + ( U )) s=1,2
2
d3 kl
3s
(k) B, ,+
3s (k) B3s (k)

B, ,+
s3 (k) + Bs3 (k) m=0
and we have also used (5.16), (5.8) and the equality

30m 30m
d3 q d3 kk
k0 =

m2 c2 +k2
s=1 t=1

[as (q)q a (q), at +(k) at (k) + at (k) at+ (k)]
q s
30m
= 2(1 + ( U )) d3 kk 2 2 2 as (k) a s (k), (6.12)
k0 = m c +k
s=1
which can be proved analogously to (6.6a) and which is responsible for the term propor-
tional to i in (6.10).
As in Sect 5, the polarization along the 3-momentum, characterized by s = 3, is a cause
for the appearance of abnormal terms in (6.9) and (6.10) in the massless case, m = 0.2
Combining (6.9), (6.10) and (I.2.7), we get the commutator between the total angular mo-
mentum M and the momentum operator P as
[ M , P ] = i{ P P } + 0m { SC LC }, (6.13)
in which the terms mentioned also change the ordinary commutation relation in the mass-
less case. We should also pay attention on the sign before the constant i in (6.13) which
sign agrees with a similar one in (I.6.17) and is opposite to the one, usually, accepted in the
literature [5, 6, 9, 10].
The expressions for the dynamical variables in momentum picture can be found from
equations (6.7)(6.10) and the general rule (I.5.4) with U (x, x0 ) being the operator (I.5.1).
However, the spin and orbital angular momentum operators in momentum picture cannot be
written, generally, in a closed form for m = 0, due to the presents of the terms proportional
to 0m in (6.9) and (6.10). To simplify the situation, below it will be supposed that
[ SC , P] = 0 [ LC , P] = 0. (6.14)
If required for some purpose, the reader may generalize, as an exercise, the following results
on the base of (I.5.1) and (I.5.4) in a case if (6.14) do not hold.
Since (6.8)(6.10), (6.14) and (I.5.1) entail (see footnote I. 3 on page 15)
[ Q , U (x, x0 )] = 0 (6.15)
1
[ S , U (x, x0 )] = 0m(x x0 ) SC U (x, x0 ) (6.16)
i
[ L , U (x, x0 )] = {(x x0 ) P (x x0 ) P } U (x, x0)
1 (6.17)
0m (x x0 ) LC U (x, x0 ),
i
by virtue of A (x) = A (x) [ A (x), U (x, x0 )] U 1 (x, x0 ) (see (I.5.4)), it follows that the
charge, spin and orbital angular momentum operators in momentum picture respectively
2 The reader may wish to compare (6.9) and (6.10) with similar relations for a free Dirac field in section III.6
or for a free scalar field in section II.13 (with S = 0 in the last case). The mentioned abnormal terms destroy
also the ordinary commutation relation between the total angular momentum M = L + S and the
momentum operator P ; see equation (6.13) below and, e.g., [5, 6, 9, 10].
are:
Q = Q (6.18)
1
S = S 0m (x x0 ) SC . (6.19)
i
1
L = L + (x x0 ) P (x x0 ) P + 0m (x x0 ) LC . (6.20)
i
Explicitly, by virtue of (5.7), the orbital angular momentum operator is
30m
1
L = d3k(x k x k )|k0=m2 c2+k2 {as +(k) as (k) + as (k) a+s (k)}
1 + ( U ) s=1
3
i
+
1 + ( U ) s,s =1
ss
d3 k l (k) as + (k) a
s (k) a
s (k) a +
s (k)

i 30m

+
2(1 + ( U )) s=1
d k 3
as + (k)
k k a
k k s (k)

as (k) k k a+ (k) . (6.21)
k k s
k0 = m2 c2 +k2
As we see again, these results differ in the massless case from the ones, expected from

the outcome of chapters II and III, by terms depending on the operators a 3 (k) and a3 (k)
with polarization variable s = 3.
Comparing the field equations (6.4) with the similar ones (II.12.13) for an arbitrary free
scalar field, obtained in section II.12, we see that the only difference between them is that
the creation and annihilation operators depend on the polarization indices in the vector field
case, which indices are missing when scalar fields are concerned. It is a simple observation,
a polarization variable, say s, is coupled always to a momentum variable, say k, and can be
considered as its counterpart. This allows s and k to be treated on equal footing in order
that one takes into account that k R3 is a continuous variable, while s is a discrete one,
taking the values s = 1, 2, 3 for a massive vector field and s = 1, 2 for a massless vector field
satisfying the Lorenz condition. Therefore the transformations
k
(s, k)
0 (k)
as (k) 0 (k)
as (k)
30m (7.1)
d k
3
d3 k 3 (k q)
st 3 (k q),
s=1

where 0 (k) and 0 (k) are the creation/annihilation operators for a free scalar field, allow
us to transfer automatically all results regarding the field equations of a free scalar field to
free vector field (satisfying the Lorenz condition). The same conclusion is, evidently valid
and with respect to results in which the momentum and charge operators are involved.1 As
a particular realization of these assertions, the commutation relations for a free vector field
(satisfying the Lorenz condition) will be considered below; in other words, a second quan-
tization of such a field will be performed by their means. The reader can find a motivation
for an introduction of these relations in books like [1, 4, 5, 11].
Before writing the commutation relations for a free vector field satisfying the Lorenz
condition, we would like to state explicitly the additional to Lagrangian formalism con-
ditions, imposed on the field operators, which reduce the field equations (6.4) to these
relations.
Assumption 7.1. The (anti)commutators between all creation and/or annihilation operators
are proportional to the identity operator idF of the systems (fields) Hilbert space of states
F.
This hypothesis reduces the field equations (6.4) to a certain algebraic-functional sys-
tem of equations which can be obtained from a similar one for a scalar field, derived in
section II.14, by means of the rules (7.1).
Assumption 7.2a. We demanded the last mentions system of equations to be an identity
with respect to the creation and annihilation operators.
A consequence of this restriction is that the (anti)commutators between creation and/or
annihilation operators are uniquely defined as operators proportional to idF . The men-
tioned system of equations does not give any information about the operators a s (0)|m=0
and as (0)|m=0 , which describe massless particles with vanishing 4-momentum and, pos-
sibly, non-vanishing charge and spin.
Assumption 7.2b. We assume the operators a
s (0)|m=0 and as (0)|m=0 to satisfy the same
(anti)commutation relations as the creation/annihilation operators for (k, m) = (0, 0).
At this stage of the theory development it remains undetermined whether a vec-
tor field should be quantize via commutators or anticommutators; the Lagrangian
(2.7), we started off, is insensitive with respect to that choice. To be achieved a confor-
mity with the experimental data, we accept the following additional restriction.
Assumption 7.3. A vector field is quantized via commutators, not via anticommutators.2
This condition can be incorporated in the Lagrangian formalism by a suitable choice of
a Lagrangian. It follows from the Lagrangian (2.7), we started off, if the field considered is
neutral/Hermitian. For details, see Sect. 11.
As a result of the described additional hypotheses, the field equations (6.4) reduce to
the following system of commutation relations, which is obtainable from a similar system
of equations (II.14.21) for a free scalar field, derived in section II.14, via the changes (7.1):
[a
s (k), at (q)] = 0 [as (k), at (q)] = 0
1 However, when the angular momentum operator is concerned, one should be quite careful as the
changes (7.1) will produce (5.7) without the integral depending on lss (k). Moreover, the rules (7.1) cannot
be applied at all to results in which the spin is involved; e.g. they will produce identically vanishing spin
angular momentum operator of free vector fields instead of the expression (5.15).
2 Equivalently, one may demand a charge symmetry of the theory, the validity of the spin-statistics theorem,
etc.
[a
s (k), at (q)] = ( U )st (kq) idF
3
[as (k), at (q)] = ( U )st 3 (kq) idF
[a
s (k), at (q)] = 0 [as (k), at(q)] = 0
[a
s (k), at (q)] = st (kq) idF
3
[as (k), at(q)] = st 3 (kq) idF (7.2)
where, as it was said above, the values of the polarization indices depend on the mass
parameter m according to
1, 2, 3 for m = 0
s,t = , (7.3)
1, 2 for m = 0
the zero operator of F is denoted by 0, and ( U ) takes care of is the field neutral/Hermitian
( U = U , ( U ) = 1) or charged/non-Hermitian ( U = U , ( U ) = 0) and ensures correct
commutation relations in the Hermitian case, when as (k) = a s (k).
Let us emphasize once again, the commutation relations (7.2) are equivalent to the field
equations (6.4) and, consequently, to the initial system of equations (2.18)(2.19) under the
made hypotheses. If by some reason one or more of these additional to the Lagrangian
formalism conditions is rejected, the trilinear system of equations (6.4), which is more
general than (7.2), should be considered.

A feature of (7.2) is that, in the massless case, the operators a
3 (k) and a3 (k), i.e. the
polarization modes with s = 3 (along the vector k), do not enter in it and hence, these opera-
tors remain completely arbitrary. In that sense, these modes remain not second quantized,
i.e. the Lagrangian formalism does not give any information about the (anti)commutation
relations between themselves or between them and other creation and annihilation opera-
tors.
As we have noted in chapters II and III, the concepts of a distribution (generalized
function) and operator-valued distribution appear during the derivation of the commutation
relations (7.2). In particular, the canonical commutation relations (7.2) have a sense iff the
commutators of the creation and/or annihilation operators are operator-valued distributions
(proportional to idF ), which is not the case if the fields considered are described via ordi-
nary operators acting on F . These facts point to inherent contradiction of quantum field
theory if the field variables are considered as operators acting on a Hilbert space. 3
The rigorous mathematical setting requires the fields variables to be regarded as oper-
ator-valued distributions. However, such a setting is out of the scope of the present work
and the reader is referred to books like [9, 10, 34, 35] for more details and realization of that
program. In what follows, the distribution character of the quantum fields will be encoded
in the Diracs delta function, which will appear in relations like (6.4) and (7.2).
As an application of the commutation relations (7.2), we shall calculate the commuta-
tors between the components of the spin operator and between them and the charge operator.
Exercise 7.1. Prove that the following commutation relations between quadratic combina-
3 Again, the reason is that to the Lagrangian formalism are added the external to it conditions (I.5.30).
tions of creation and/or annihilation operators are valid:

[as (k) a
s (k), at (p) at (p)]
= {st at (p) a
s (k) s t as (k) at (p)} (k p)

3

[as (k) a
s (k), at (p) at (p)]

= {st a
s (k) at (p) s t at (p) as (k)} (k p)

3

[a
s (k) as (k), at (p) at (p)]
= {st at(p) as (k) s t a
s (k) at (p)} (k p)
3

(7.4)
[as (k) a
s (k), at (p) at (p)]

=( U ){st a
s (k)at (p) s t at (p)as (k)} (k p)

3

[as (k) a
s (k), at (p) at (p)]
=( U ){st at (p)a

s (k) s t as (k)at (p)} (k p)
3

[a
s (k) as (k), at (p) at (p)]
=( U ){st as (k)at (p) s t at(p)a
s (k)} (k p),
3
where the polarization indices s, s , t, and t take the values 1, 2 and 3 for m = 0 and
1 and 2 for m = 0. These equalities are simple corollaries of the identities [A, B C] =
[A, B] C + B [A,C] and [B C, A] = [B, A] C + B [C, A] , applied in this order to the
left-hand-sides of (7.4), and (7.2).
Applying (5.15), (6.19) and (7.4), we find:
2
[ S, S ] = [ S, S ] m=0 =
(1 + ( U ))2
30m +

s t ss s t +
d3 k ss
(k) (k) (k) (k) as (k) at (k) as (k) at (k)
s,s ,t=1
s t
+ +

+ ( U )ss (k) (k) as (k) at (k) as (k) at (k)
ss
+ +
+ ( U )ts
(k) (k) as (k) at (k) as (k) at (k)
+ 0m f(a
3 , a3 ), (7.5)
where f (a
3 , a3 ) is a term whose integrand is a homogeneous expression in a3 (k) and
a3 (k) and which term is set equal to zero for m = 0. The summation over s in (7.5) can
be performed explicitly by means of (5.16):
30m

s t
(k) (k) = v (k)v (k)v (k) ( ) ( )
ss s t
(7.6)
s =1
30m

(k) (k) (k) (k)
ss st ss st
s =1
= v (k)st (k) ( ) ( ), (7.7)
where
30m
v (k) := vs (k)vs (k) (7.8)
s=1
is given via (3.26) and (3.27) and the symbol ( ) means that we have to subtract all
of the previous terms by making the change , i.e. an antisymmetrization over and
must be performed.
Similar calculations, based on (5.15) and (5.6), show that4

iq( U ) 30m
[ S, Q ] =
1 + ( U ) s,t=1
+ +
(k) as (k) at (k) as (k) at (k)
d3 kts

+ as + (k) at (k) as (k) at + (k) + 0m f(a
3 , a3 ),

where f (a
3 , a3 ) is a term whose integrand is a homogeneous expression in a3 (k) and
a3 (k) and which term is set equal to zero for m = 0. Hence, recalling that q = 0 for a
Hermitian field and ( U ) = 0 for a non-Hermitian one and, consequently, q( U ) 0, we
get

[ S, Q ] = 0m f (a
3 , a3 ) (7.9)
in Heisenberg picture. Therefore, we obtain

[ S, Q ] = 0m f (a
3 , a3 ) (7.10)
in momentum picture (see (I.5.4)). In particular, we have

[ S, Q ] = 0 if m = 0 or if m = 0 and a
3 (k) = a3 (k) = 0; (7.11)
so, for a massive vector field, the spin and charge operators commute.
As other corollary of (7.2), we shall establish the equations

[ U , M (x, x0 )] m=0
= x [ U , P ] x [ U , P ] + i( U U ) (7.12a)

[ U , M (x, x0 )] m=0
= x [ U , P ] x [ U , P ] + i( U U ) (7.12b)
which are part of the conditions ensuring the relativistic covariance of the theory consid-
ered [2]. The condition m = 0 is essential one. For m = 0, additional terms, depending on
a
s (k) and as (k), s = 1, 2, 3, should be added to the right hand sides of (7.12); these terms
are connected with the gauge symmetry of the massless case see [2, 8.4] for some
details about that situation for an electromagnetic field. The equations (7.12) immediately
follow from the following lemma, due to (4.6), (4.7) and (2.25).
Lemma 7.1. The equalities

[ U , S(x, x0 )]
m=0
4 1 (1 + ( U ))qst for st (k).

This result can formally be obtained from (7.5) with i
3
= i d3 p vt(p)vt(p) U+
(p) + U (p) ( ) (7.13a)
t=1

[ U , L(x, x0 )] m=0
3
= x [ U , P] x [ U , P ] i d3 p + vt (p)vt (p)
t=1

U+
(p) + U
(p) ( ) (7.13b)

and similar ones with U for U hold. Here and below p0 := m2 c2 + p2 and the symbol
( ) means that we have to subtract all of the previous terms by making the change
, i.e. the previous expression has to be antisymmetrized relative to and .
Proof. Equation (7.13a) is a simple corollary of (6.1), (7.2) and (5.16). To derive (7.13b),
we substitute (6.1) and (5.7) in its l.h.s. and then, after an integration by parts of the terms
(p)
proportional to ap
s
and using (5.5) and (5.8), we obtain
[ U , L(x, x0 )] = x [ U , P ] x [ U , P ]
3
+ i d3 p{2c(2)3 m2 c2 + p2 }1/2
s=1
3 v,s (p)
+ vt (p)vt (p) p a +
(p) + a
(p) ( ) .
t=1 p s s
Since from (3.26a) with m = 0 5 and (3.19) it follows
3 v,s (p) v,s (p)

1
+ vt (p)vt (p) p = p p p

t=1 p m=0 m2 c2 p
1 3
= 2 2 p p v,s (p) = + vt (p)vt (p) vs (p) for m = 0,
m c t=1
the last equality implies (7.13b), due to (4.17) and (4.5).

As it was said above, the relations (7.12) are not valid for m = 0 in the theory consid-
ered, unless some additional terms are taken into account. This fact is connected with the
quantization method adopted in the present chapter for massless vector fields. Other such
methods may restore the validity of (7.12) for m = 0; for example, such is the Gupta-Bleuler
quantization of electromagnetic field [1,7], as it is proved in [7, 19.1] (for interacting elec-
tromagnetic and spin 12 fields).
An interesting result is that equations (7.12), regardless of the condition m = 0, imply
the relation

[ M , M ] = i M M M + M . (7.14)
5 This is the place where the supposition m = 0 is essentially used and the proof brakes down if m = 0.
To prove this, we notice that (7.12), in momentum representation in Heisenberg picture,

is equivalent to (see (5.34))

[ U(k), M ] = i k k U(k)
k k
+ i U (k) U(k)

(7.15)
[ U (k), M ] = i k k U (k)
k k
+ i U (k) U (k) ,

where k0 = m2 c2 + k 2 . Now, applying (5.3), (5.14) and the identity [A B,C]
= A [B,C] + [A,C] B, one can prove, after a trivial but long calculations, that

[ S , M ] = i S S S + S
(7.16)
[ L , M ] = i L L L + L ,
from where equation (7.14) follows in Heisenberg picture as M = L + S . Notice,

this derivation of (7.14) demonstrates that (7.14) is a consequence of the validity of (7.12)
regardless of the fulfillment of the commutation relations (7.2). Similarly, equations (7.18)
below imply (7.19) regardless of the validity of (7.2).
By virtue of the identity [A, B C] = [A, B] C + B [A,C] , the relations
if m=0
[a
s (k), Q ] = qas (k) [as (k), Q ] = qas (k) s = 1,2,3
1,2 if m=0 (7.17)
are trivial corollaries from (5.6) and the commutation relations (7.2). From here and (6.1),
we get
[ U , Q ] = q U + 0m ( ) [ U , Q ] = q U + 0m ( ) , (7.18)
with ( ) and ( ) denoting expressions which are linear and homogeneous in a 3 (k) and
a3 (k). So, the equations (2.35) are consequences of the Lagrangian formalism under con-

sideration if m = 0 or if m = 0 and a3 (k) = a3 (k) = 0. Moreover, the relations
(7.17) entail
if m=0
the commutativity of bilinear functions/functionals of as (k) and as (k), s = 1,2,3 1,2 if m=0 ,
with the charge operator Q . In particular, we have (see (5.5)(5.7) and (5.15)):
[ P , Q ] = 0 [ Q , Q ] = 0
(7.19)
[ S, Q ] = 0m ( ) [ L , Q ] = 0m ( ) [ M , Q ] = 0m ( ).

So, if m = 0 or if m = 0 and a
3 (k) = a3 (k) = 0, the spin, orbital and total angu-
lar momentum operators commute with the charge operator, as it is stated by (7.11); the
momentum and charge operators always commute.

For a general motivation regarding the introduction of the concepts of vacuum and normal
ordering, the reader is referred, e.g., to [1,4,5] (see also chapters II and III). Below we shall
concentrate on their formal aspects in an extend enough for the purposes of the present
chapter.
Definition 8.1. The vacuum of a free vector field U (satisfying the Lorenz condition) is its
physical state that contains no particles and has vanishing 4-momentum, (total) charge and
(total) angular momentum. It is described by a state vector, denoted by X0 (in momentum
picture) and called also the vacuum (of the field), such that:
X0 = 0 (8.1a)
X0 = X0 (8.1b)
a
s (k)( X0 ) = as (k)( X0 ) = 0 (8.1c)
X0 | X0 = 1 (8.1d)
where | : F F C is the (Hermitian) scalar product of systems (fields) Hilbert space

of states and s = 1, 2, 3.
It is known, this definition is in contradiction with the expressions for the dynamical
variables, obtained in Sect. 5, as the latter imply infinite, instead of vanishing, character-
istics for the vacuum; see (5.5)(5.7), (5.15) and (7.2). To overcome this problem, one
should redefine the dynamical variables of a free vector field via the so-called normal or-
dering of operator products (compositions) of creation and/or annihilation operators. In
short, this procedure, when applied to free vector fields (satisfying the Lorenz condition),
says that [1, 4, 5, 37]:
(i) The Lagrangian and the fields dynamical variables, obtained from it and containing
the field operators U and U , should be written in terms of the creation and annihilation
operators via (4.1)(4.7).
(ii) Any composition (product) of creation and/or annihilation operators, possibly ap-
pearing under some integral sign(s), must be changed so that all creation operators to stand
to the left of all annihilation operators.1
The just described procedure is known as normal ordering (of products) and the result
of its application on some operator is called its normal form; in particular, its application
on a product of creation and/or annihilation operators is called their normal product. The
mapping assigning to an operator its normal form, obtained from it according to the above
procedure, will be denoted by N and it is called normal ordering operator and its action
on a product of creation and/or annihilation operators is defined according to the rule (ii)
given above. The action of N on polynomials or convergent power series of creation and/or
annihilation operators is extended by linearity. Evidently, the order of the creation and/or
annihilation operators in some expression does not influence the result of the action of N
on it. The dynamical variables after normal ordering are denoted by the same symbols as
before it.
It should be noticed, the normal ordering procedure, as introduced above, concerns all
degrees of freedom, i.e. the ones involved in the field equations (6.4) and the operators

a3 (k)|m=0 and a3 (k)|m=0, in the massless case. This simplifies temporary the considera-
tion of massless vector fields, but does not remove the problems it contains see Sect. 10
below. Moreover, in Sect. 10 arguments will be presented that the afore-given definition
of normal ordering agrees with the description of electromagnetic field and that the men-
tioned operators should anticommute with the other ones. In principle, one can consider the
1 The relative order of the creation/annihilation operators is insignificant as they commute according to (7.2).
normal ordering operation only for creation and annihilation operators involved in the field

equations (6.4), i.e. by excluding its action on the operators a3 (k)|m=0 and a3 (k)|m=0 .
This will add new problems with the spin and orbital angular momentum of the vacuum as
they will be, possibly, finite, but completely undetermined.
From the evident equalities
+ + +
N a
s (k) at (k) = N at (k) as (k) = at (k) as (k)

N as (k) at+ (k) = N at+ (k) as (k) = at+ (k) as (k)
(8.2)

N bk b = b+ k b b = a
s , as
k k
and the equations (5.5)(5.7), (5.15), and (6.18)(6.20), we see that the dynamical variables
of a free vector field (satisfying the Lorenz condition) take the following form after normal
ordering:
1
P =
1 + ( U )
30m
k |
k0 =

m2 c2 +k2
{a+ +
s (k) as (k) + as (k) as (k)} d k
3
(8.3)
s=1
30m
Q =q {a+ +
3
(8.4)
s=1
30m
1
L =
1 + ( U ) d3 k(x k x k )|
k0 =

m2 c2 +k2
{as + (k) a
s (k)
s=1
+ a+
s (k) as (k)}
3
i
+
1 + ( U ) s,s =1
ss
d3 k l (k) as + (k) a +
s (k) as (k) as (k)

30m
i
+
2(1 + ( U )) s=1 d k as (k) k k a
3 +
k k s (k)

+ as (k) k k as (k) 2 2 2
+
k k k0 = m c +k
1
+ 0m (x x0 ) N ( LC ) (8.5)
i
3 +
i

+
S = (k) as (k) as (k) as (k) as (k)
d3 kss
1 + ( U ) s,s =1
1
0m (x x0 ) N ( SC ), (8.6)
i
where N ( LC ) and N ( SC ) can easily be found by means of (8.2), (6.11) and (6.6),
but we shall not need the explicit form of these operators. Similarly, the vectors of the
spin (5.25) and (5.26) take the following form after normal ordering (see also (6.19)):
c
i
Ra = abc d3 krb (k) a + (k) a (k) + a+ (k) a (k)
1 + ( U )
1
(x x0 ) N ( SC0a ) (8.7)
0m
i

i
S= d3 k a + (k) a (k) + a+(k) a (k) 0m S (8.8)
1 + ( U )
1 abc
with S a := i (x x0 ) N ( SCbc). In particular, the third component of S is (cf. (5.29))

i
S3 = d3 k a1 + (k) a +
2 (k) a2 (k) a1 (k)
1 + ( U ) (8.9)

a+ +
2 (k) a1 (k) + a1 (k) a2 (k) .
The substitutions (5.30) transform the integrand in the last equality into a diagonal form
and preserves the ones into (8.3) and (8.4), i.e.
1
P =
1 + ( U )
30m
k |
k0 =

m2 c2 +k2
{b+ +
s (k) bs (k) + bs (k) bs (k)} d k
3
(8.10)
s=1
30m
Q =q {b+ +
s (k) bs (k) as (k) bs (k)} d k
3
(8.11)
s=1
3
2
S =
3

1 + ( U ) s=1
(1)s+1 bs + (k) b +
s (k) bs (k) bs (k) d k. (8.12)
From the just written expressions for the dynamical variables after normal ordering is
evident that
P ( X0 ) = 0 Q ( X0 ) = 0 M ( X0) = L ( X0) = S( X0) = 0 (8.13)
and, consequently, the conserved quantities of the vacuum, the 4-momentum, charge, spin
and orbital angular momenta, vanish, as required by definition 8.1.
Besides the dynamical variables, the normal ordering changes the field equations (6.4)
too. As the combinations quadratic in creation and annihilation operators in the commuta-
tors in (6.4) come from the momentum operator (see (6.2)), the field equations (6.4) after
normal ordering, by virtue of (8.3), read:

as (k), at+ (q) at (q) + at+ (q) at (q)
(8.14a)
(1 + ( U ))a
3

as (k), at+ (q) at(q) + at+ (q) at (q)

(8.14b)
(1 + ( U ))as (k)st (k q) = f st (k, q),
3

1, 2, 3 for m = 0
s,t = (8.14c)
1, 2 for m = 0
30m
q q =m2 c2 +q2 fst (k, q) d3 q = 0
0
t=1
(8.14d)
30m
q
q0 =
f (k, q) d3 q
m2 c2 +q2 st
= 0.
t=1
However, one can verify that (8.14) hold identically due to the commutation rela-
tions (7.2). Once again, this demonstrates that (7.2) play a role of field equations under
the suppositions made for their derivation.
As a result of (8.6), (6.19) and (7.4), we see that the commutation relations (7.5) after
normal ordering transform into
2
[ S, S ] = [ S, S ] m=0 =
(1 + ( U ))2
30m +

ss s t
d3k st (k)ss
(k) (k) (k) as (k) at (k) + a+
s (k) at (k)
s,s ,t=1
N
( U )a+ +
s (k) at (k) + ( U )as (k) at (k) + 0m f (a3 , a3 ), (8.15)
N
with f (a
3 , a3 ) being a term whose integrand is a homogeneous expression in a3 (k)
and a3 (k) and which term is set equal to zero for m = 0.
The commutation relations between the spin angular momentum operator and the
charge one, expressed by (7.9)(7.11) before normal ordering, take the following form after
normal ordering
N
[ S, Q ] = 0m f (a
3 , a3 ) (8.16)
N

[ S, Q ] = 0m f (a
3 , a3 ) (8.17)

[ S , Q ] m=0 = 0. (8.18)
These relations can be checked by means of (8.6), (8.4) and (7.4); alternatively, they are
corollaries of (8.15) with st (k) replaced by i
1
(1 + ( U ))qst .
9. State Vectors and Particle Interpretation

The description of the state vectors of a free vector field satisfying the Lorenz condition
is practically identical with the one of a free spinor field, presented in section III.9. To be
obtained the former case from the latter one, the following four major changes should be
made: (i) the polarization indices should run over the range 1, 2 and 3, if m = 0, or 1 and 2, if
m = 0; (ii) the commutation relations (7.2) must replace the corresponding anticommutation
(k)
(k) and l (k)
ones (III.7.13) for a free spinor field; (iii) the replacements st, st st,
st
l (k) of the spin and orbital momentum coefficients should be made; (vi) the additional
terms, depending on the polarization s = 3 should be taken into account, when the spin and
orbital angular momentum operators of a massless field are considered. According to these
changes, we present below a mutatis mutandis version of the corresponding considerations
in section III.9 (see also section II.16).
In momentum picture, in accord with the general theory of chapter. I, the state vectors
of a vector field are spacetime-dependent, contrary to the field operators and dynamical
variables constructed from them. In view of (I.5.24), the spacetime-dependence of a state
vector X (x) is
X (x) = U (x, x0)( X (x0)) (9.1)
where x0 is an arbitrarily fixed spacetime point and the evolution operator U (x, x0 ) : F
F is
1
(x x0 ) k | 2 2 2

U (x, x0) = exp k0 = m c +k
i s
(9.2)
+ +
{as (k) as (k) + as (k) as (k)} d k .
3
due to (I.5.1) and (8.3) (see also (I.5.11)(I.5.24)). The operator (9.2) plays also a role of
an S-matrix determining the transition amplitudes between any initial and final states, say
Xi (xi ) and X f (x f ) respectively. In fact, we have
(0) (0)
S f i (x f , xi ) := X f (x f )| Xi (xi ) = X f (x f )| U (xi , x f )( Xi (xi )). (9.3)
For some purposes, the following expansion of U (xi , x f ) into a power series may turn
to be useful:

U (xi, x f ) = idF + U (n)(xi , x f ) (9.4)
n=1
30m
1 1
(xi x f 1 ) . . .(xi n x f n )
(1) (n)
U (n)(xi , x f ) := d3 k(1) . . . d3 k(n)k1 kn
n! s1 ,...,sn =1
+ (1)
as1 (k ) as1 (k ) + a+
(1) (1) (1)
s1 (k ) as1 (k )

asn+ (k(n)) a (n) + (n)
sn (k ) + asn (k ) asn (k )
(n)
(9.5)

(a)
where k0= m2 c2 + (k (a))2 , a = 1, . . ., n.
According to (I.5.27) and the considerations in Sect. 4, a state vector of a state contain-
ing n particles and n antiparticles, n , n 0, such that the i th particle has 4-momentum
pi and polarization si and the i th antiparticle has 4-momentum pi and polarization si ,
where i = 0, 1, . . ., n and i = 0, 1, . . ., n, is given by the equality
X (x; p1, s1; . . .; pn , sn ; p1 , s1 ; . . .; pn , sn )

1 1 n
1 n
= exp (x x0 ) (pi ) + (x x0 ) (pi )

n !n ! i i =1 i i =1
+ + +
as (p1 ) a+ s (p
n s

) a (p1 ) a (pn ) ( X0 ),

s
(9.6)
1 n 1 n
where, in view of the commutation relations (7.2), the order of the creation operators is
inessential. If n = 0 (resp. n = 0), the particle (resp. antiparticle) creation operators and
the first (resp. second) sum in the exponent should be absent. In particular, the vacuum
corresponds to (9.6) with n = n = 0. The state vector (9.6) is an eigenvector of the mo-

mentum operator (8.3) with eigenvalue (4-momentum) ni =1 pi + ni =1 pi and is also an
eigenvector of the charge operator (8.4) with eigenvalue (q)(n n ).1
Exercise 9.1. Using (7.2) and (4.19), verify that the transition amplitude between two states
of a vector field, like (9.6), is:
X (y; q1,t1 ; . . .; qn ,tn ; q1 ,t1 ; . . .; qn ,tn )

| X (x; p1 , s1 ; . . .; pm , sm ; p1 , s1 ; . . .; pm , sm )
1 1 n
1 n
n !n ! i i =1 i i =1
s t 3 (pn qi )s
n i 1
t
n 1 i
3 (pn 1 qi ) . . .s1 t 3 (p1 qi )
2 i n
(i1 ,...,in ) 1 2 n
s t 3 (pn qi )s

n i1 1
t
n 1 i
3 (pn1 qi ) . . .s1 t 3 (p1 qi ) (9.7)
2 i n
(i1 ,...,in ) 2 n
. . ., in ) of (1, . . ., n ).
The conclusions from (9.7) are similar to the ones concerning free scalar or spinor fields
considered in chapters II and III. For instance, the only non-forbidden transition from an
(n -particle + n -antiparticle) state is into (n -particle + n -antiparticle) state; the both states
may differ only in the spacetime positions of the (anti)particles in them. This result is quite
natural as we are dealing with free particles/fields.
0, 1, . . ., then (9.7) says that
Xn | X0 = n0 , (9.8)
Consider the one (anti)particle states at+(p)( X0 ) and at + (p)( X0 ), with
t = 1, 2, 3, if
m = 0, or t = 1, 2, if m = 0. Applying (8.3)(8.10) and (7.2), we find (p0 := m2 c2 + p2 ):2

P at+(p)( X0) = p at+ (p)( X0 ) Q at+ (p)( X0) = qat+ (p)( X0)
(9.9)
P at + (p)( X0) = p at + (p)( X0) Q at + (p)( X0 ) = +qat + (p)( X0 )
30m
S at+(p)( X0) = i +
(p)as (p)( X0 ) 0m ( )( X0 )
ts
s=1
(9.10)
30m
S at + (p)( X0) = i +
(p)as (p)( X0 ) 0m ( ) ( X0 )
ts
s=1
1 Recall (see Sect. 4), the operator a+
s (k)
creates a particle with 4-momentum k and charge q, while

a+
s (k) creates a particle with 4-momentum k and charge +q, where, in the both cases, k0 = m2 c2 + k 2 . See
also equations (9.9)(9.12) below.
2 The easiest way to derive (9.12) is by applying (I.6.8), (6.2), (6.3) and (7.2). Notice, in Heisenberg picture

and in terms of the Heisenberg creation/annihilation operators (5.36), equations (9.12) read L at+ (p)( X0 ) =
+
0 and L at (p)( X0 ) = 0 which is quite understandable in view of the fact that L is, in a sense, the
average orbital momentum with respect to all spacetime points, while L (x,x0 ) is the one relative to x and x0 ;
the dependence on x0 being hidden in the L , at+ (p) and at+ (p).

S at+ (p)( X0 ) = i{t2a+
3 +
1 (p) t1 a2 (p)}( X0 )
(9.11)
S 3 at + (p)( X0 ) = i{t2a1 + (p) t1 a2 +(p)}( X0)
+
L(x) at+(p)( X0) = (x p x p )i p p at (p)( X0 )
p p
30m
i l ts
(p) a+
s (p)( X0 ) 0m ( )( X0 )
s=1
+
(9.12)

L (x) at + (p)( X0) = (x p x p )i p p at (p)( X0 )
p p
30m
i l ts
(p) as +(p)( X0 ) 0m ( ) ( X0 )
s=1
where q( U ) 0 was used and ( ) and ( ) denote expressions whose integrands are

homogeneous with respect to a 3 (p) and a3 (p). It should be noted the agreement of (9.9)
(9.12) with (4.20).3
If one uses the operators b
s (k) and as (k) instead
of a
s (k) and as (k) (see
(5.30)), the equations (9.9) and (9.11) will read (p0 := m2 c2 + p2 ):

P bt+ (p)( X0) = p bt+(p)( X0) Q bt+(p)( X0) = qbt+ (p)( X0 )
(9.13)
P bt + (p)( X0) = p bt + (p)( X0) Q bt +(p)( X0) = +qbt + (p)( X0)

S 3 bt+ (p)( X0) = {t1 b+ +
1 (p) t2 b2 (p)}( X0 )
(9.14)
S 3 bt + (p)( X0) = {t1 b1 + (p) t2 b2 + (p)}( X0).
We should emphasize, the equations (9.9)(9.14) do not concern themassless case,
m = 0, with polarization t = 3, i.e. the (state?) vectors a+ (p) ( X0 ) and a + (p) ( X0 )
3 m=0 3 m=0
have undetermined 4-momentum, charge, spin and orbital angular
momentum. This is
quite understandable as the operators a (p) and a
(p) do not enter in the field
3 m=0 3 m=0
equations (6.4) and, consequently, remain completely arbitrary. (See Sect. 10 for more
details on this situation.)
On the formulae (9.9)(9.14) is based the particle interpretation of quantum field theory
(in Lagrangian formalism) of free vector field satisfying the Lorenz condition. According
to them, the state vectors produced by the operators a+ +
the vacuum X0
s (p) and bs (p) from
can be interpreted as ones representing particles with 4-momentum ( m2 c2 + p2 , p) and
charge (q); the spin and orbital angular momentum of these vectors is not definite. The
states a+s (p)( X0 ) do not have a definite projection of the vector of spin on the direction
of movement, i.e. along p (for p = 0), but, if m = 0, the one of b+ s (p)( X0 ) is equal to
(1) , if s = 1, 2, or to zero, if s = 3. Similarly, the state vectors as + (p)( X0 ) and
s+1
bs + (p)( X0 ) should be interpreted as ones representing particles, called antiparticles (with

respect to the ones created by a+ +
s (p) and bs (p)), with the same characteristics but the
charge, which for them is equal to (+q). For this reason, the particles and antiparticles of a
neutral (Hermitian), q = 0, field coincide.
Notice, the above interpretation of the creation and annihilation operators does not con-
cern these operators for a massless field and polarization along the 3-momentum, i.e. for
3 If the r h.s. of (I.6.17) is with an opposite sign, this agreement will be lost.
(m, s) = (0, 3). Besides, in the massless case, the just said about the spin projection of the
states b+ +
s (p)( X0 ) and bs (p)( X0 )) is not valid unless the last terms, proportional to 0m ,
in (9.10) vanish.
10. The Massless Case and

Electromagnetic Field in Lorenz Gauge
A simple overview of the preceding sections reveals that the zero-mass case, m = 0, is more
or less an exception of the general considerations. The cause for this is that the Lorenz
condition, expressed by (2.19), is external to the Lagrangian formalism for a massless free
vector field, contrary to the massive case, m = 0. However, this condition does not contra-
dict to the formalism and, as we demonstrated, it can be developed to a reasonable extend.
10.1. Problems in the Massless Case

Practically, the only problem we have met in the massless case is the one with the physical

meaning/interpretation of the operators a 3 (k)|m=0 and a3 (k)|m=0 . The first indications
1
for it were the last two equations in (4.20c) (valid if m = 0), which were derived from
the external to the Lagrangian formalism equation (I.6.30) and, hence, can be neglected if
one follows rigorously the Lagrangian field theory; moreover, in Sect. 5, we proved that
the equations (5.18) imply a
s (k) = as (k) = 0 for s = 1, 2, which, in view of the further
development of the theory, is unacceptable. The really serious problem with the opera-

tors a3 (k)|m=0 and a3 (k)|m=0 is that they have vanishing contributions to the momentum
operator (5.5) and charge operator (5.6), but they have, generally, non-vanishing ones to
the orbital and spin angular momentum operators (5.7) and (5.15), respectively.2 (See also
the vectors of spin (5.25) and (5.26) in which these operators enter via (5.27).) In con-
nection with the (possible) interpretation in terms of particles, this means that a3 (k)|m=0
and a3 (k)|m=0 describe creation/annihilation of neutral massless particles with vanish-
ing 4-momentum and charge, but, generally, non-zero spin and orbital angular momen-
tum. Besides, the last two characteristics of the (hypothetical) particles with state vec-
+
tors a+3 (k)|m=0( X0 ) and a3 (k)|m=0( X0 ) can be completely arbitrary since the field equa-

tion (6.4) do not impose on the operators a 3 (k)|m=0 and a3 (k)|m=0 any restrictions.
3
The above discussion leads to the following conclusion. The Lagrangian formalism,
without further assumptions/hypotheses, cannot give any information about the operators

a3 (k)|m=0 and a3 (k)|m=0 and, consequently, leaves them as free parameters of the quan-
tum field theory of massless free vector field satisfying the Lorenz condition and described
by the Lagrangian (2.7). Thus, these operators are carries of a completely arbitrary de-
grees of freedom, which have a non-vanishing contribution to the spin and orbital angular
momentum operators (5.15) and (5.7), respectively, unless before normal ordering we have
1 From pure mathematical viewpoint, everything is in order and no problems arise.
2 To save some space, here our considerations do not take into account the normal ordering, i.e. they concern
the theory before it; vide infra.
3 However, the third component of the vector of spin S does not depend on a (k)|
3 m=0 and a3 (k)|m=0
see (5.29) and (9.11).
+
+
(k) m=0 as (k) a3 (k) as (k) a3 (k)
d3 ks3
s=1,2

a3 +(k) a +
s (k) + a3 (k) as (k) m=0 = 0 (10.1a)

d3 kl
s3
(k)m=0 as + (k) a +
3 (k) as (k) a3 (k)
s=1,2

a3 + (k) a +
s (k) + a3 (k) as (k) m=0 = 0, (10.1b)
where we have used the skewsymmetry of the quantities (5.8) and (5.16). The just-presented
considerations concern the theory before (second) quantization, i.e. before imposing the
commutation relations (7.2), and normal ordering. However, since the quantization proce-

dure does not concern the operators a 3 (k)|m=0 and a3 (k)|m=0 (see Sect. 7), the above-said
remains completely valid after these procedures, provided one takes into account the ex-
pressions (8.3)(8.8) for the dynamical variables after normal ordering. In particular, after
normal ordering, the spin and orbital angular momentum operators (8.6) and (8.5), respec-

tively, will be independent of a
3 (k)|m=0 and a3 (k)|m=0 iff
+
+
(k) m=0 as (k) a3 (k) a3 (k) as (k)
d3 ks3
s=1,2

a3 +(k) a +
s (k) + as (k) a3 (k) m=0 = 0 (10.2a)

d3 kl
s3
(k)m=0 as + (k) a +
3 (k) a3 (k) as (k)
s=1,2

a3 + (k) a +
s (k) + as (k) a3 (k) m=0 = 0. (10.2b)
Thus, if one wants to construct a sensible physical theory of a massless free vector field
satisfying the Lorenz condition, new assumptions to the Lagrangian formalism should be
added. At this point, there is a room for different kinds of speculations. Here are two such
possibilities.
One can demand, as an additional condition, the fulfillment of
Conjecture 10.1. The field equations (6.4) hold for m = 0 and any polarization indices, i.e.
for s,t = 1, 2, 3, instead only for s,t = 1, 2 obtained from the Lagrangian formalism.
,
This assumption will entail the vanishment of all of the quantities Bss (q), i.e. (6.6) will
be replaced with
,
Bss (q) 0 s, s = 1, 2, 3 (10.3)
which, in its turn, leads to (see (6.11))
S
C = LC = 0 (10.4)
and, consequently, to the commutativity of the spin angular momentum and momentum
operators, etc. (see (6.9)(6.20)). Other consequence of the above assumption will be the
validity of the commutation relations (7.2) for arbitrary polarization indices s,t = 1, 2, 3
in the massless case. As a result of them and the normal ordering procedure, the vec-
+
tors a+ 3 (k)|m=0 ( X0 ) and a3 (k)|m=0( X0 ) will describe states with vanishing 4-momentum
and charge and, generally, non-vanishing spin and orbital angular momentum. It seems,
states/particles with such characteristics have not been observed until now. This state of
affairs can be improved by adding to the integrands in (8.3) and (8.4) terms proportional to
a3 + (k) a +
3 (k) and a3 (k) a3 (k) in the massless case, but such a game with adjustment
of theorys parameters is out of the scope of the present work.
A second possible solution of the problem(s) with the zero-mass case, we would like to
explore, does not require drastical changes of the formalism as the preceding one.
Conjecture 10.2. The equations (10.2) or (10.1) are added to the Lagrangian formalism as
subsidiary conditions depending if respectively the normal ordering is or is not taken into
account.
In this way only the operators of spin and orbital angular momentum are changed, viz.
before normal ordering they read (see (5.15), (5.7), (6.19), (6.20) and notice that the above
assumption entails (10.4))
i
S = S =
1 + ( U )
30m +

+
(k) as (k) as (k) as (k) as (k)
d3 kss (10.5)
s,s =1
1
L (x x0 ) P + (x x0 ) P = L =
1 + ( U )
30m

d3 k(x0 k x0 k )|
k0 = m2 c2 +k2
{a+ +
s (k) as (k) + as (k) as (k)}
s=1
30m
i
1 + ( U ) s,s
ss
+ d3 k l (k) as + (k) a
s (k) a
s (k) a +
s (k)
=1

i 30m

+
2(1 + ( U )) s=1
d k 3
as + (k)
k k a
k k s (k)

as (k) k k a+
s (k) (10.6)
k k k0 = m2 c2 +k2
and, after normal ordering, they take the form (see (8.6) and (8.5))
30m +
i
1 + ( U ) s,s
+
d3 kss (10.7)
=1
30m
1
L = d3 k(xk x k )|k0=m2 c2+k2
1 + ( U ) s=1
{as + (k) a +
s (k) + as (k) as (k)}
30m
i
1 + ( U ) s,s
ss
+ d3 k l (k) as + (k) a
s (k) a +
s (k) a
s (k)
=1

i 30m
+
2(1 + ( U )) s=1
d k 3
as + (k)
k k a
k k s (k)

+ a+ (k) k k as (k) 2 2 2 . (10.8)
s
k k k0 = m c +k
So, formally, the replacement 3s,s =1

3 0m
s,s =1 should be made and the terms propor-
tional to 0m should be deleted. As a result of these changes, all terms proportional to 0m
in all equations, starting from (6.9) onwards, will disappear, i.e., for any m, we have
0m ( ) = 0, (10.9)
where the dots stand for some expressions, which depend on a 3 (k)|m=0 and

a3 (k)|m=0 for m = 0 and are set to zero for m = 0; in particular, the equations (10.4)
hold (see (6.11) and (6.5)), but (10.3) do not.

In this way, the operators a3 (k)|m=0 and a3 (k)|m=0 disappear from all dynamical vari-
ables. So, if we extend the particle interpretation on them,4 these operators will describe
creation/annihilation of massless particles with vanishing 4-momentum, charge, spin and
orbital angular momentum. Naturally, such particles are completely unobservable. Thus,
+
the properties of the states a+3 (k)|m=0 ( X0 ) and a3 (k)|m=0 ( X0 ) are similar to the ones of
the vacuum (see (8.13)), but their identification with the vacuum X0 requires additional and,
in a sense, artificial hypotheses for a self-consistent development of the theory.5
So, assuming the validity of (10.1) or (10.2), we see that these equations are the only

place in the theory, where the operators a 3 (k)|m=0 and a3 (k)|m=0 essentially appear.
6
In fact, these equations should be regarded as equations of motion for the mentioned op-
erators, which operators do not enter in the field equations (6.4) or in the commutation
relations (7.2). The equations (10.1) or (10.2) possess always the trivial solution

a
3 (k)|m=0 = 0 a3 (k)|m=0 = 0, (10.10)
which agrees with the definition 8.1 of the vacuum, but they may have and other solutions.

Since, at the moment, it seems that the operators a 3 (k)|m=0 and a3 (k)|m=0 cannot lead to
some physically measurable results, we shall not investigate the problem for existence of
solutions of (10.1) or (10.2), different from (10.10).
Regardless of the fact that equations (10.1) or (10.2) exclude a contribution of

a
3 (k)| m=0 and a3 (k)|m=0 from the dynamical variables, the initial operators U and U

4 Such an extension requires the fulfillment of the commutation relations (7.2) for s = 3 and/or t = 3 in the
massless case.
5 For instance, in a case of a neutral field, when a (k) = a (k), one can satisfy (10.1) by requiring
s s
+
[a3 (k),a +
s (k)]+ |m=0 = [a3 (k),as (k)]+ |m=0 , with s = 1,2 and [A,B]+ := A B + B A; in particular, this will
be valid if we assume [a
3 (k),as (k)]+ |m=0 = 0 for s = 1,2.
6 The initial operators U and U depend on these operators too vide infra.

depend on them via the operators U

(k) and U (k) see (4.18) and (4.6)(4.7). As a
consequence of this, the commutation relations between different combinations of U (k)

and U (k) also depend on a3 (k)|m=0 and a3 (k)|m=0 in the massless case. Moreover, one
cannot calculate these relations without additional assumptions, like (10.10) or the validity
of (7.2) for any s,t = 1, 2, 3 in the massless case. Besides, the result depends essentially
on the additional condition(s) one assumes in the massless case. For instance, if we as-
sume (7.2) to hold for any s,t = 1, 2, 3, when m = 0, then, for (k, m) = (0, 0), we get

[ U (k), U (k )]
1 3 3
= f (, ) 2c(2)3 m2 c2 + k 2 (k k ) {vs (k)vs (k)}, (10.11)
s=1
where , = , , , and we have applied (4.17) and (7.2) in the form

[as (k), at (k )] = f (, )st 3 (k k ) (10.12a)

0 for , = ,
f (, ) := (U) for (, ) = (, ), (, ) . (10.12b)

1 for (, ) = (, ), (, )
Combining (10.11) and (3.26), we obtain ((k, m) = (0, 0))
1 3
[ U (k), U (k )] m=0 = f (, ) 2c(2)3 m2 c2 + k 2

(k k )
#
+ 2 2
k k
for k=0
m c
for k = 0 and , = 1, 2, 3 (10.13a)
0 otherwise
1 3

[ U (k), U (k )] m=0 = f (, ) 2c(2)3 m2 c2 + k 2 (k k )
$ # %
k k 2 for ==0
+ 2 0 for ,=1,2,3 . (10.13b)
k 1 otherwise
On the other hand, if we assume (10.10), then (10.11) must be replaced with

[ U (k), U (k )]
1 3 30m
= f (, ) 2c(2)3 m2 c2 + k 2 (k k ) {vs (k)vs (k)}, (10.14)
s=1
1,2,3 if m=0
as a consequence of (4.17), (10.12) with s,t = 1,2 if m=0 , and (10.10). So, for m = 0, the
relation (10.14) reduces to (10.13a) (see (3.26)),7 but for m = 0 it reads ((k, m) = (0, 0))

[ U (k), U (k )] m=0
1 3
= f (, ) 2c(2)3 m2 c2 + k 2 (k k ) 0 for ,=1,2
otherwise
, (10.15)
due to (3.27). Evidently, the equations (10.13a) and (10.15) coincide for , = 1, 2 but
otherwise are, generally, different.
7 For m = 0 and k = 0, as one can expect, the commutation relations (7.2) and (10.13a) reproduce, due
to (5.36), the known ones for a massive free vector field in Heisenberg picture [1, 4, 11].
10.2. Electromagnetic Field

It is now time to be paid special attention to the electromagnetic field.8 As it is well
known [1, 4, 6, 11], this field is a massless neutral vector field whose operators, called the
electromagnetic potentials, are usually denoted by A and are such that
A = A. (10.16)
The (second) quantization of electromagnetic field meets some difficulties, described in
loc. cit., the causes for which are well-described in [2, 82] (see also [7]). The closest to our
approach is the so-called Gupta-Bleuler quantization [1,2,7,11,12] in the way it is described
in [1]. However, our method is quite different from it as we quantize only the independent
degrees of freedom, as a result of which there is no need of considering indefinite metric,
time (scalar) photons and similar objects.9 The idea of most such methods is to be started
from some Lagrangian, to be applied the standard canonical quantization procedure [2,
5], and, then, to the electromagnetic potentials to be imposed some subsidiary conditions,
called gauge conditions, by means of which is (partially) fixed the freedom in the field
operators left by the field equations.
In our scheme, the free electromagnetic field is described via 4 Hermitian operators A
(for which (A) = 1 see (2.2)), the Lagrangian (2.7) with m = 0, i.e.
1 1
L = c2[ A , P] [ A , P ] c2 [ A , P] [ A , P ] (10.17)
2 2
and the Lorenz conditions (2.19) with U = A , i.e.
[ A , P ] = 0. (10.18)
It should be emphasized, the Lorenz condition (10.18) is imposed directly on the field
operators, not on the physical states etc. as in the Gupta-Bleuler formalism. So, (10.16)
(10.18) describe an electromagnetic field in Lorenz gauge.
Thus, to specialize the general theory form the preceding sections to the case of elec-
tromagnetic field, one should put in it (see (10.16) and sections 4 and 5)
m=0 U = A ( U ) = 1 A = A
(10.19)
A = A A (k) = A (k) as (k) = a
s (k).
It is important to be emphasized, the equations (10.19) reduce (10.1) to

d3 ks3(k) [a+s (k), a3 (k)]+ [as (k), a+3 (k)]+ = 0 (10.20a)
s=1,2

d3 kl
s3
(k) [a+ +
s (k), a3 (k)]+ [as (k), a3 (k)]+ = 0, (10.20b)
s=1,2
where [A, B]+ := A B + B A is the anticommutator of operators A and B. Since in (10.20)

enter the anticommutators [a s (k), a3 (k)]+ and a3 (k) are actually free parameters, the con-

tributions of a3 (k) in the spin and orbital momentum operators can be eliminated via the
following change in the theory.
8 In fact, the description of electromagnetic field was the primary main reason for the inclusion of the
massless case in the considerations in the preceding sections.

9 The only such problem we meet is connected with the longitudinal photons vide infra.
Conjecture 10.3. Redefine the normal products of creation and/or annihilation operators by
assigning to them an additional (with respect to the definition in Sect. 8) multiplier (sign)
equal to (1) f , where f is equal to the number of transpositions of the operators a 3 (k),
relative to a
1 (k) and a
2 (k), required to be obtained the normal form of a product, i.e. all
creation operators to be to the left of all annihilation ones.
Evidently, such a (redefined) normal ordering procedure transforms (10.20) into identi-
ties and, consequently, after it all dynamical variables become independent of the operators
a3 (k). A similar result will be valid if the normal ordering procedure, as defined in Sect. 8,
holds only for the operators a
1 (k) and a2 (k) and the operators a3 (k) anticommute with
them,
[a
3 (k), as (k)]+ = 0 for s = 1, 2. (10.21)
Notice, if we put (cf. (10.10))
a
3 (k) = 0, (10.22)
the definitions of vacuum, normal ordering procedure, and equations (10.21) (and
hence (10.1) and (10.2)) will be satisfied. Besides, these choices will naturally exclude from
the theory the longitudinal photons, represented in our theory by the vector a+ 3 (k)( X0 ),
which have identically vanishing dynamical characteristics.
The above discussion shows that the operators a 3 (k) can naturally be considered as
fermi or bose operators that anticommute with a s (k), s = 1, 2, i.e. with anomalous com-
mutation relations between both sets of operators [22, appendix F].10 Besides, this agrees
with the above-modified normal ordering procedure, which will be accepted below in the
present section.
As a result of (10.19), the commutation relations (7.2) for an electromagnetic field (in
Lorenz gauge) read11
[a 3
s (k), at (k )] = 0 [as (k), at (k )] = st (k k ) for s,t = 1, 2 (10.23)
and the operators of the dynamical variables, after the (redefined) normal ordering is per-
formed, for this field are (see (10.20) and (8.3)(8.8) with m = 0 and ( U ) = 1)

P = k |
k0 =
a+(k) a (k) d3 k
2 s
k s (10.24)
s=1,2
Q =0 (10.25)

L = d3 k(x k x k )|
k0 =

k2
s=1,2

a+
s (k) as (k) + i d3 k l
ss
(k)a+
s (k) as (k)
s,s =1,2
10 Since this excludes the operators a3 (k) from all physically significant quantities, one is free to choose
bilinear or other commutation relations between a 3 (k).
11 According to our general considerations, the relations (10.23) are equivalent to the Maxwell(-Lorentz)
equations, but written in terms of creation and annihilation operators.


1
3 +
(k) (10.26)
2 s=1,2 k k s
k0 = k2

+
S = i (k)as (k) as (k)
d3 kss (10.27)
s,s =1,2
R =0 (10.28)

S1 = S2 = 0 S 3 = i d3 k a+ +
1 (k) a2 (k) a2 (k) a1 (k) , (10.29)
where, for the derivation of (10.28) and (10.29), we have used (5.28), (3.23), (3.25), and
that, formally, (10.27) corresponds to (8.6) with a
3 (k) = 0. Thus, the operators a3 (k) do
not enter in all of the ;dynamical variables.12
From (10.24)(10.29), it is evident that the particles of an electromagnetic field, called
photons, coincide with their antiparticles, which agrees with the general considerations in
Sect. 9. Besides, the state vectors b+ +
X0 ) and b2 (k)( X0 ), with bs (k) given by (5.30),
1 (k)(
describe photons with 4-momentum ( k2 , k), zero charge, and vectors of spin R = 0 and
S = (0, 0, +) or S = (0, 0, ) (see (9.14)), i.e. their spin vector S is collinear with k with
projection value + or , respectively, on its direction.
It is worth to be mentioned, the commutation relations (6.7)(6.10) for an electromag-
netic field take their ordinary form, i.e.
[ P , P ] = 0 (10.30)
[ Q , P ] = 0 (10.31)
[ S, P ] = 0 (10.32)
[ L, P ] = i{ P P }, (10.33)
as a consequence of which the dynamical variables in momentum picture are (see (6.18)
(6.20))
Q = Q (10.34)
S = S (10.35)
L = L + (x x0 ) P (x x0 ) P. (10.36)
At last, as we said above, the commutators [ A (k), A (k)] and [ A(k), A(k)]

cannot be computed without knowing the explicit form of [a
s (k), a3 (k)] and [as (k),

a3 (k)] for s = 1, 2, 3. For instance, the additional conditions (10.22) lead to (10.15) with
A for U and , = , .
The so-obtained quantization rules for electromagnetic field, i.e. equations (10.23) to-
gether with (10.22), coincide with the ones when it is quantized in Coulomb gauge [2, 6], in
12 This situation should be compared with similar one in the Gupta-Bleuler formalism. In it the contribution
of the time (scalar) and longitudinal photons, the last corresponding to our states a
3 (k)( X0 ), is removed
from the average 4-momentum of the admissible states, but, for example, the longitudinal photons have a
generally non-vanishing part in the vector of spin S see, e.g., [1, eq. (12.19)].
which is assumed
3
A0 = 0 [ P a , Aa ] =0 (10.37)
a=1
in momentum picture. This is not accidental as (10.37) is a special case of (2.19). In fact,
by virtue of (I.5.9), it is equivalent to ka Aa = 0 (with k2 = k02 k 2 = 0) which is tantamount
to

a+3 (k) + a3 (k) for k = 0
0 = (a3 (k) + a3 (k))(k va (k)m=0 ) =
+ a 3
, (10.38)
0 for k = 0
due to (4.6), (4.7), (4.17) and (3.23)(3.24). Therefore any choice of a 3 (k) such that
a+
3 (k) + a
3 (k) = 0 reduces the Lorenz gauge to the Coulomb one. However, the partic-
ular choice (10.22) completely reduces our quantization method to the one in Coulomb
gauge, as a little more derailed comparison of the both methods reveals. We shall end this
discussion with the remark that the choice (10.22) is external to the Lagrangian formalism
and, of course, it is not the only possible one in that scheme.
11. On the Choice of Lagrangian

Our previous exploration of free vector fields was based on the Lagrangian (see (2.1))
m2 c4
L = L =

U U
1 + ( U )
c2 2
+ ( U ) ( U ) + ( U ) ( U ) (11.1)
1 + ( U )

in Heisenberg picture. In it the field operators U and their Hermitian conjugate U do not

enter on equal footing: in a sense, U are first and U are second in order (counting

from left to right) unless the field is neutral/Hermitian. Since U and U are associated
with the operators a s and as

(see Sect. 4), which create/annihilate fields particles and
antiparticles, respectively, the Lagrangian (11.1) describes the particles and antiparticles in
a non-symmetric way, which is non-desirable for a free field as for it what should be called
a particle or antiparticle is more a convention than a natural distinction. This situation is
usually corrected via an additional condition in the theory, such as the charge symmetry,
spin-statistics theorem, etc. Its sense is the inclusion in the theory of the symmetry par-
ticle antiparticle, which in terms of the creation and annihilation operators should be
expressed via theory invariance under the change a
s (k) as (k). As we demonstrated
in chapters II and III for free scalar and spin 12 fields, this symmetry/invariance can be in-
corporated in the initial Lagrangian, from which the theory is constructed. Below we shall
show how this can be achieved for free vector fields satisfying the Lorenz condition.
As an alternative to the Lagrangian (11.1), one can consider
m2 c4
L = U U
1 + ( U )
c2 2
+ ( U ) ( U ) + ( U ) ( U ) = L U U (11.2)
1 + ( U )

in which the places, where the operators U and U are situated in (11.1), are interchanged.
In terms of particles and antiparticles, this means that we call fields particles antiparticles
and vice versa, or, equivalently, that the change a
s (k) as (k) has been made. Obviously,
the Lagrangian (11.2) suffers from the same problems as (11.1). However, judging by our
experience in chapter II and partially in chapter III, we can expect that the half-sum of the
Lagrangians (11.1) and (11.2), i.e.
m2 c4 c2 2
L = ( U ) ( U )

U U + U U +
2(1 + ( U )) 1 + ( U )
1
+( U )( U )( U )( U )+( U )( U ) = ( L + L ), (11.3)
2

is one of the Lagrangians we are looking for, as it is invariant under the change U U .
To any one of the Lagrangians (11.1)(11.3), we add the Lorenz conditions
U = 0 U = 0,

(11.4)

which are symmetric under the transformation U U and for m = 0 are additional
conditions for the Lagrangian formalism, but for m = 0 they are consequences from the
field equations (see Sect. 2 and below).
According to the general rules of chapter I (see (I.5.4) and (I.6.2)), the La-
grangians (11.1)(11.3) and the Lorenz conditions (11.4) in momentum picture respectively
are:
m2 c4
L = U U
1 + ( U )
c2
+ [ U , P ] [ U , P ] [ U , P ] [ U , P] (11.5)
1 + ( U )
m2 c4
L = U U
1 + ( U )
c2
+ [ U , P ] [ U , P ] [ U , P ] [ U , P ] (11.6)
1 + ( U )
m2 c4 c2
L = U U + U U + [ U , P ] [ U , P ]
2(1 + ( U )) 2(1 + ( U ))

[ U , P ] [ U , P ] + [ U , P ] [ U , P ] [ U , P ] [ U , P ] (11.7)
[ U , P ] = 0 [ U , P ] = 0. (11.8)
The (classical) derivatives of the above Lagrangians happen to coincide and are as fol-
lows:1
L L L
= = = m2 c4 U
U U U

L L L
= = = m2 c4 U
U U U
(11.9)
L L L
= = = = ic 2
[ U , P ] ic [ U , P ]
2
y y y
L L L
= = = = ic2 [ U , P ] ic2 [ U , P]
y y y
with y := i1
[ U , P ] and y := i
1
[ U , P ] .
As a consequence of (11.9), the field equations of the Lagrangians (11.5)(11.7) coin-
cide and are given by (2.14) (and (11.8) as an additional conditions/equations for m = 0)
which, for m = 0, split into the Klein-Gordon equations (2.18) and the Lorenz condi-
tions (11.8). From here it follows that the material of sections 3 and 4 remains valid without
any changes for the Lagrangian theories arising from any one of the Lagrangians (11.5)
(11.7) (under the Lorenz conditions in the massless case).
The densities of the operators of the dynamical variables for the Lagrangian (11.1) are
given via (2.23)(2.27). Similarly, the energy-momentum tensor, (charge) current and spin
angular momentum operators for the Lagrangians (11.2) and (11.3) respectively are:2
1
T = ( U ) + ( U ) L
1 + ( U )
(11.10a)
c2 2
= ( U ) ( U ) + ( U ) ( U ) L
1 + ( U )
q
J = U U
i (11.10b)

= iqc2 ( U ) U + U ( U )
1

S := (I U ) + (I

U )
1 + ( U )
2 c2 (11.10c)
= ( U ) U
1 + ( U )

( U ) U U ( U ) + U ( U )

1 This assertion is valid if the derivatives are calculated according to the classical rules of analysis of com-
muting variables, as it is done below and discussed on page xii. Such an approach requires additional rules for
ordering of the operators entering into the expressions for the dynamical variables, as the ones presented below.
Both of these assumptions, in the particular cases we are considering here, have their rigorous explanation in
a different way for computing derivatives of non-commuting variables, as it is demonstrated in [3], to which
paper the reader is referred for further details (see also section I.3).
2 Excluding the spin operators, the other density operators can be obtained, by virtue of (11.4), as sums
of similar ones corresponding to U 0 , U 1 , U 2 and U 3 and considered as free scalar fields see chapter II.
See [3] and section I.3 for a rigorous derivation.
1
T = ( U ) + ( U )
2(1 + ( U ))

+ ( U ) + ( U ) L (11.11a)
1
= T + T
2
q
J = U U + U U
2i
1 (11.11b)
= J + J
2
1

U )

S := (I U ) + (I
2(1 + ( U ))
(11.11c)
1
+ (I U ) + (I

U ) =
S + S .
2
Thus, we see that the dynamical variables derived from L can be obtained from the
ones for L = L by making the change U U and reversing the currents sign. Besides,
the dynamical variables for L are equal to the half-sum of the corresponding ones for
L = L and L . So, symbolically we can write (see also (I.2.5), (I.2.6)(I.2.8) and (I.5.4))
1
D = D U D = ( D + D ), (11.12)
U

2

where D = T, J , S , L , P , Q , S, L and the minus sign in the first equality stand
only for D = J , Q . If we express the dynamical variables in terms of creation and
annihilation operators, which are identical for the Lagrangians we consider (vide infra),
then (11.12) takes the form
1
D = D as (k)as (k) D = ( D + D ), (11.13)
2
with D = P , Q , S , L .3 To save some space, we shall write explicitly only the con-
served operator quantities for the Lagrangian (11.6). Combining (5.5)(5.7) and (5.15)
with the rule (11.13), we get in Heisenberg picture (and before normal ordering):
1
P =
1 + ( U )
30m
k |
k0 =

m2 c2 +k2
{a+ +
s (k) as (k) + as (k) as (k)} d k
3
(11.14a)
s=1
30m

Q = q {a+ +
3
(11.14b)
s=1
3 +
i
1 + ( U ) s,s
+
d3 kss (11.14c)
=1
3For D = P , Q , the first equation in (11.13) is evident (see (5.1) and (5.2)), but for D = S , L some
simple manipulations are required for its proof see (5.8) and (5.16).
30m
1

L = d3 k(x0 k x0 k )|k0=m2 c2+k2
1 + ( U ) s=1
{a+ +
s (k) as (k) + as (k) as (k)}
3
i
1 + ( U ) s,s
ss +
+ d3 k l (k) a+
s (k) as (k) as (k) as (k)
=1

i 30m
+
2(1 + ( U )) s=1
d3 k a+
s (k) k k
k k
as (k)

as (k) k k as + (k) 2 2 2 . (11.14d)

k k k0 = m c +k
Let us turn now our attention to the field equations in terms of creation and annihilation
operators for the Lagrangians L , L and L . For L = L they are given by (6.4). To de-
rive them for L and L , one should repeat the derivation of (6.4) from (6.2) with P and
P = 12 ( P + P ), respectively, for P. In this way, from (6.2) with P = P , P, (11.14a)
and (11.13) with D = P , we obtain the field equations derived from the Lagrangians L
and L respectively as:

as (k),at+(q) at (q) + at (q) at +(q)

(11.15a)
(1 + ( U ))a s (k)st (k q) = f st (k, q)
3

as (k),at+(q) at (q) + at (q) at +(q)
(11.15b)

(1 + ( U ))a s (k)st (k q) = f st (k, q)
3

as (k),at + (q) at (q) + at (q) at+ (q)

+
+ a +
s (k), at (q) at (q) + at (q) at (q) (11.16a)

2(1 + ( U ))a
3

as (k),at + (q) at (q) + at (q) at+ (q)

+ as (k), at+(q) at (q) + at (q) at + (q) (11.16b)

2(1 + ( U ))a
s (k)st (k q) = f st (k, q),
3
where the polarization indices s and t take the values

1, 2, 3 for m = 0
s,t = (11.17)
1, 2 for m = 0
and the operator-valued (generalized) functions af (k, q) and af (k, q), with a = , , ,
are such that

q |q =m2 c2 +q2 af (k, q) d3 q = q |q =m2 c2 +q2 af (k, q) d3 q = 0. (11.18)
0 0
Equations, similar to (6.7)(6.10), can be derived form (11.15) and (11.16) and, conse-
quently, expressions for the dynamical variables in momentum picture, similar to (6.18)
(6.20), can easily be obtained from these equations.
As we see, the dynamical variables and the field equations in terms of creation and anni-
hilation operators for the Lagrangians (11.1)(11.3) are completely different for a non-Her-
mitian field, U = U or as (k) = a
s (k), and, in this sense, the arising from them quantum
field theories of free vector field satisfying the Lorenz condition are different. A step toward
the identification of these theories is achieved via the second quantization procedure, i.e.
by establishing/imposing for/on the creation and annihilation operators commutation re-
lations, like (7.2) for the Lagrangian (11.1). These relations for the Lagrangians (11.2)
and (11.3) can be derived analogously to the ones for (11.1), i.e. by making appropriate
changes in the derivation of the commutation relations for an arbitrary free scalar field,
given in chapter II (see also section III.7, where free spinor fields are investigates). With-
out going into details, we shall say that this procedure results into the commutation rela-
tions (7.2) for any one of the Lagrangians (11.1)(11.3) (under the Lorenz conditions (11.4)
in the massless case). In this way, the systems of field equations (6.4), (11.15) and (11.16)
became identical and equivalent to (7.2).
It should be emphasized, the derivation of (7.2) for the Lagrangians (11.1)(11.3) is
not identical (cf. section II.14): the Lagrangian (11.3) does not admit quantization via
anticommutators, contrary to (11.1) and (11.2). So, the establishment of (7.2) for L and L
requires as an additional hypothesis the quantization via commutators or some equivalent to
it assertion, like the charge symmetry, spin-statistics theorem, etc. [1]. Said differently, this
additional assumption is not needed for the Lagrangian (11.3) as it entails such a hypothesis
in the framework of the Lagrangian formalism. The initial cause for this state of affairs is
that the symmetry particle antiparticle is encoded in the Lagrangian L via its invariance
under the change U U . In particular, since for a neutral field we, evidently, have
L = L = L if U = U , (11.19)
for such a field, e.g. for the electromagnetic one, the spin-statistics theorem and other equiv-
alent to it assertions are consequences form the Lagrangian formalism investigated in the
present paper.
Since the commutation relations for the Lagrangians (11.1)(11.3) are identical, we
assume the normal ordering procedures and the definitions of the vacuum for them to be
identical, respectively, and to coincide with the ones given in Sect. 8.
Applying the normal ordering procedure to the dynamical variables corresponding to
the Lagrangians (11.1)(11.3) (see (5.5)(5.8), (5.15), (11.14) and (11.13)), we see that,
after this operation, they became independent of the Lagrangian we have started, i.e. sym-
bolically we can write
D = D = D = D D = P, Q , L, S , (11.20)
where the operators for L = L are given by (8.3)(8.6). (To prove these equations for
D = L , S, one has to use the antisymmetry of the quantities (5.8) and (5.16).)
Let us summarize at the end. The Lagrangians (11.1)(11.3), which are essentially dif-
ferent for non-Hermitian fields, generally entail quite different Lagrangian field theories
unless some additional conditions are added to the Lagrangian formalism. In particular,
these theories became identical if one assumes the commutation relations (7.2), the normal
ordering procedure and the definition of vacuum, as given in Sect. 8. The Lagrangian (11.3)
has the advantage that the spin-statistics theorem (or charge symmetry, etc.) is encoded in
it, while, for the Lagrangians (11.1) and (11.2) this assertion should be postulated (im-
posed) as an additional condition to the Lagrangian formalism. For a neutral free vector
field satisfying the Lorenz condition, e.g. for electromagnetic field in Lorenz gauge, the
spin-statistics theorem is a consequence from the Lagrangian formalism.4
12. On the Role of the Lorenz Condition in the Massless Case

Until now, in the description of free massless vector fields, we supposed that they satisfy

the Lorenz condition, i.e. the equations U = 0 and U = 0 in Heisenberg picture
or (2.19) in momentum one, as subsidiary restrictions to the Lagrangian formalism. Such
a theory contains some physical, not mathematical, problems which were summarized and
partially analyzed in Sect. 10. The present section is devoted to a brief exploration of a
Lagrangian formalism for free massless vector field without the Lorenz condition as an
additional restriction. As we shall see, in this case the problems inherent to a formalism
with the Lorenz conditions remain and new ones are added to them.
For other point of view on the topic of this section, see, e.g., [6, 7.1].
Most of the considerations in this section will be done in Heisenberg picture which will
prevent the exposition from new details (which are not quite suitable for the purpose).
12.1. Description of Free Massless Vector Fields

without the Lorenz Condition
The description of a free massless vector field coincides with the one of a free massive
vector field, given in Sect. 2, with the only difference that the fields mass parameter m
(which is equal to the mass of fields particles, if m = 0) is set equal to zero,
m = 0. (12.1)
In particular, the Lagrangian formalism can start from the Lagrangian (see (2.1) and (2.7))
c2 2 c2 2
L = ( U ) ( U ) + ( U ) ( U ) (12.2)
1 + ( U ) 1 + ( U )
c2
L= [ U , P ] [ U , P ] [ U , P ] [ U , P ] (12.3)
1 + ( U )
in Heisenberg and momentum picture, respectively. The Euler-Lagrange equations for this
Lagrangian are the following massless Proca equations (see (2.5) and (2.14))
U ) ( U ) = 0
2 ( U ) ( U ) = 0
( (12.4)

4 Recall (see chapter II), the proof of the spin-statistics theorem (charge symmetry, etc.) for the La-
grangian (11.3) requires as a hypothesis, additional to the Lagrangian formalism, the assumption 7.1, i.e. the
assertion that the commutators or anticommutators of all combinations of creation and/or annihilation operators
to be proportional to the identity mapping of systems Hilbert space of states, i.e. to be c-numbers.
[[ U , P ] , P ] [[ U , P ] , P ] = 0
(12.5)
[[ U , P ] , P ] [[ U , P ] , P ] = 0
in Heisenberg and momentum picture, respectively. (Recall, := .) As pointed in

Sect. 2, these equations do not imply that the field operators satisfy the massless Klein-Gor-
don equations and the Lorenz conditions (see (2.6), (2.14), (2.18) and (2.19) with m = 0).
The common solutions of the massless Klein-Gordon equations and the Lorenz conditions
for the field operators are solutions of the massless Proca equation, but the opposite is not
necessary, i.e. the latter system of equations is more general than the former one. This is the
cause why, for solutions of (2.5), the Lagrangian (12.3) cannot be reduced to (2.20) with
m = 0 in the general case (unless the Lorenz conditions (2.19) are imposed on the solutions
of (12.5) as additional conditions).
The general expressions for the densities of the dynamical variables through the gener-
alize momenta (see the first equalities in (2.23)(2.27)) remain, of course, valid in the
massless case too, but their particular dependence on the field operators is different from the
second equalities in (2.23)(2.27), as now (2.13), without the Lorenz conditions on these
operators, should be used. Thus, the dynamical variables of a massless vector field are:
c2 2
T = ( U ) ( U ) + ( U ) ( U ) L
1 + ( U )
(12.6)
c2 2
+ ( U ) ( U ) + ( U ) ( U )
1 + ( U )

J = iqc2 ( U ) U U ( U )

(12.7)
( U ) U + U ( U )
M = L

+ S (12.8)

L : = x T x T (12.9)
2 c2
U ( U ) U

S = ( U )
1 + ( U )
(12.10)
U ( U ) + U ( U ) ( U ) U

+ U ( U ) + ( U ) U + U ( U ) .

Notice, in (12.6) the Lagrangian L must be replaced by its value given by (12.2), not
by (2.20) with m = 0. Evidently, the Lorenz conditions (2.6b) reduce the equations (12.6)
(12.10) to (2.23)(2.27), respectively. If needed, the reader can easily write the above
equations in momentum picture by means of the general rules of chapter I.
It should be remarked, as the energy-momentum tensor (12.6) is non-symmetric, the
spin and orbital angular momentum are no longer conserved quantities.
12.2. Analysis of the Euler-Lagrange Equations

Since the solutions of the Euler-Lagrange equations (12.4) (or (12.5)) generally do not sat-
isfy the Klein-Gordon equation, we cannot apply to free massless vector fields the methods
developed for free scalar fields. To study the equations (12.4), we shall transform them into
algebraic ones in the momentum representation in Heisenberg picture [1, 2, 11].
Define the Fourier images u (k) and u (k), k R4 , of the field operators via the Fourier
transforms

1 1
e i kxu (k) d4 k e i kx u (k) d4 k,
1 1
U (x) = U (x) = (12.11)
(2)2 (2)2

where d4 k := dk0 dk1 dk2 dk3 and kx := k x . Since U (x) is the Hermitian conjugate of

U (x), U (x) := ( U (x)), we have
u (k) = (u (k)) . (12.12)
Substituting (12.11) into (12.4), we find the systems of equations
k2 u (k) k k u (k) = 0 k2 u (k) k k u (k) = 0, (12.13)
which is equivalent to (12.4). Here and below k2 := k k . (From the context, it will be clear
that, in most cases, by k2 we have in mind k k , not the second contravariant component of
k.)
Let us consider the classical analogue of the equations (12.13), i.e.
k2 v (k) k k v (k) (k2 k k )v (k) = 0, (12.14)
where v (k) is a classical, not operator-valued, vector field over the k-space R4 . This is a
linear homogeneous system of 4 equations for the 4 variables v0 (k), v1 (k), v2 (k) and v3 (k).
Since the determinant of the matrix of (12.14) is1
det[k2 k k ]3,=0 = (k2 )3 (k2 + k2 ) 0,
the system of equation (12.14) possesses always a non-zero solution relative to v (k). Be-
sides, the form of this determinant indicates that the value k2 = 0 is crucial for the number
of linearly independent solutions of (12.14). A simple algebraic calculation reveals that the
rank r of the matrix [k2 k k ]3,=0 , as a function of k, is: r = 0 if k = 0, r = 1 if k2 = 0
and k = 0 for some = 0, 1, 2, 3 and r = 3 if k2 = 0. Respectively, the number of linearly
independent solutions of (12.14) is infinity if k = 0, three if k2 = 0 and k = 0 for some
= 0, 1, 2, 3, and one if k2 = 0.
For k2 = 0, the system (12.14) reduces to the equation (3.20) with m = 0, which was
investigated in Sect. 3. For k2 = 0, it has the evident solution
k
w (k) := i ( k2 = 0 ), (12.15)
k2
which is normalized to 1,
w (k)w (k) = 1 ( k2 = 0 ), (12.16)

3
1 One can easily prove that det c z z ,=0 = c0 c1 c2 c3 + z20 c1 c2 c3 + c0 z21 c2 c3 + c0 c1 z22 c3 + c0 c1 c2 z23
for a diagonal matrix [c ] = diag(c0,c2 ,c2 ,c3 ) and any 4-vector z . Putting here c = k 2 and z = k , we
get the cited result.
and any other solution of (12.14) is proportional to w (k), as defined by (12.15).

Therefore, for any k , we can write the general solution of (12.14) as
3
v (k) = 0k2 s (k)vs (k)m=0 + (1 0k2 )4 (k)w (k), (12.17)
s=1
where 1 (k), . . ., 4 (k) are some functions of k = (k0 , . . ., k3 ), vs (k) with s = 1, 2, 3 are
defined by (3.20)(3.22) (see also (3.23)(3.25)), and the Kronecker delta-symbol 0k2 (:= 1
for k2 = 0 and := 0 for k2 = 0) takes care of the number of linearly independent solutions
of (12.14).
Returning to the operator equations (12.13), we can express their solutions as
u (k) = (2)2{ic2 (2)3 }1/2

3
0k2 as (k)vs (k)m=0 + (1 0k2 )a4 (k)w (k)
s=1
(12.18)
u (k) = (2)2{ic2 (2)3 }1/2
3
0k2 as (k)vs (k)m=0 + (1 0k2 )a4 (k)w (k) ,
s=1
where a1 (k), . . ., a4 (k) are some operator-valued functions of k, which, by (12.12), are such
that
a(k) = (a(k)) = 1, 2, 3, 4, (12.19)
and the factor (2)2{ic2 (2)3 }1/2 is introduced for future convenience. (The operators
a1 (k), . . ., a4 (k) are closely related to the creation and annihilation operators, but we shall
not consider this problem here.)
At last, combining (12.19) and (12.11), we can write the solutions of the field equa-
tions (12.4) as
3
U (x) =(2)2{ic2(2)3}1/2 e as (k)vs (k)m=0 d4 k
1
i kx
s=1
k 2 =0

+ (2)2 {ic2 (2)3}1/2 e i kxa4 (k)w (k) d4 k
1
k 2 =0
(12.20)
3
3 1/2
e i kxas (k)vs (k)m=0 d4 k
1
U (x) =(2) {ic (2) }
2 2
s=1
k 2 =0

+ (2)2 {ic2 (2)3}1/2 e i kxa4 (k)w (k) d4 k.
1
k 2 =0
Since the Lorenz conditions (see (2.6))

U (x) = 0 U (x) = 0 (12.21)
in momentum representation in Heisenberg picture read (see (12.11)),
k u (k) = 0 k u (k) = 0, (12.22)
we see that they are equivalent to the selection of solutions of (12.13) with
k2 = 0. (12.23)
Said differently, the Lorenz conditions on the field operators are equivalent to the im-
position of the restrictions
a4 (k) = 0 a4 (k) = 0 ( k2 = 0 ), (12.24)
due to (12.18) (or (12.20)). Here the operators a4 (k) and a4 (k) may be considered as a
measure of the satisfaction of the Lorenz conditions by the field operators. Therefore the
sum of the terms containing an integral over the hyperboloid k2 = 0 in (12.20) corresponds
to field operators satisfying the Lorenz condition and, consequently, to them is valid the
theory developed in the preceding sections. In particular, up to a constant, the operators
as (k), s = 1, 2, 3, are sums of the creation and annihilation operators (in Heisenberg picture
see (5.36)) of a free massless vector field satisfying the Lorenz condition.
12.3. Dynamical Variables

To reveal the meaning of the operators a4 (k) and a4 (k) in (12.20), we shall express the
fields dynamical variables in terms of a (k) and a (k), = 1, 2, 3, 4. For the purpose, the
decompositions (12.20) should be inserted into the expressions (12.6)(12.10) and, then, the
conserved operators (I.2.5)(I.2.8) to be calculated. It is not difficult to be seen, a dynamical
variable D , with D = P , Q , L , S for respectively the momentum, charge, orbital and
spin angular momentum operators, has the following structure:
0 04 4
D = D + D + D . (12.25)
0
Here D := k2 =0 d4 k k 2 =0 d4 k { } is the dynamical variable under the condi-
tions (12.24), i.e. if the field operators were supposed to satisfy the Lorenz condition; the
04
second term is of the form D := k2 =0 d4k k2 =0 d4 k { } with the expression in braces
being a linear combination of terms like as (k) a4(k) and a4 (k ) as(k), where s = 1, 2, 3;
4
and the structure of the last term is D := k2 =0 d4 k k2 =0 d4 k { } with the expression in
braces being proportional to the operator a4 (k) a4 (k ).
Exercise 12.1. By means of the explicit formulae (3.23)(3.25) and (12.15), prove that
4
D = 0 D = P, Q , L , S. (12.26)
Thus, if we regard a4 (k) and a4 (k ) as independent degrees of freedom (possibly con-
nected with some particles), then their pure (free) contribution to the dynamical variables
is vanishing. However, the second term in (12.25) is generally non-zero.
Exercise 12.2. Via algebraic calculations show that
3
1

d4 k 3 (k + k )e i (k0 k0 )x
04 1 0
P = d4 k
1 + ( U ) s=1 k 2=0 k 2 =0
k0 k
k02 3s k0 {as (k) a4 (k) + a4 (k ) as (k)} (12.27a)
k0 2 k02
3
Q = q

d4 k 3 (k + k )e i (k0 k0 )x
04 1 0
d4 k
2
s=1 k =0 k =0
2
k0
k02 3s k0 {as (k) a4 (k ) + a4 (k ) as(k)} (12.27b)
k0 2 k02
3
i

d4 k 3 (k + k )e i (k0 k0 )x
04 1 0
S = d4 k
1 + ( U ) s=1 k 2 =0 k 2 =0
1
{as (k) a4 (k) + a4 (k ) as (k)}
2
k0 k0 2
2 3s

(1) 0
k (k
k )v s
(k)k
k0 k0

0 0 0 a a

for (, ) = (0, a), (a, 0) with a = 1, 2, 3 (12.27c)

(k0 k0 )(k v (k) k v (k))
s s

otherwise.
Notice, the expression k0 2 k02 in (12.27) is different from zero as k0 2 k02 = k0 2 k2 =
k0 2 k = k 2 = 0, due to 0 = k2 = k02 k 2 and the -function 3 (k + k ) in (12.27).
2
12.4. The Field Equations

Recall now that we consider quantum field theories in which the Heisenberg relations (I.4.1)
hold as a subsidiary restriction on the field operators. Consequently, the system of field
equations consists of the Euler-Lagrange equations (12.4), the Heisenberg relations

U (x) U (x)
[ U (x), P] = i [ U (x), P ] = i (12.28)
x x
and the explicit connection between P and the field operators, i.e. (see (12.25)(12.27))
0 04
P = P + P (12.29)
0 04
with P given by the r.h.s. of (5.5) and P defined via (12.27a).
Since the expansions (12.20) take care of the Euler-Lagrange equation (12.4), the
equations (12.28) remain the only restrictions on the field operators. Substituting equa-
tion (12.20) into (12.28), we get
[as (k), P ] = k as(k) s = 1, 2 k2 = 0 (12.30a)

[a4 (k), P ] = k a4 (k) k = 0.
2
(12.30b)
One can verify that (12.30a) is equivalent to (6.2) with m = 0 and P given by (12.29).

Notice, the operators a3 (k) and a3 (k) (or a
3 (k) and a3 (k)), with k = 0, enter in the
2
field equations (12.30) and into the dynamical variables via (12.29) and (12.27) (see
also (5.7) and (5.15)). However, a particle interpretation of the degrees of freedom con-
nected with a3 (k) and a3 (k) fails as they enter in (12.30), (5.7) and (5.15) only in combina-
tions/compositions with a (k) and a (k) with = 1, 2, 4. In that sense, the operators a3 (k)
and a3 (k) serve as coupling constants with respect to the remaining ones. Similar is the
situation with the operators a4 (k) and a4 (k), with k2 = 0, regarding the dynamical variables,
but these operators are more dynamical as they must satisfy the equations (12.30b).
The explicit equations of motion for a(k) and a (k), = 1, 2, 3, 4, can be obtained by
inserting (12.29) (see also (12.27a) and (5.5) with m = 0) into (12.30). The result will be
similar to (6.3) or (6.4), with m = 0, but additional terms, depending on a4 (k) and a4 (k)
with k2 = 0, will be presented. We shall not write these equations as they will not be used
further and it seems that a4 (k) = a4 (k) = 0, when they coincide with (6.4), is the only their
physically meaningful solution (see Subsect. 12.5 below).
12.5. Discussion
Equations (12.26) and (12.27) show that the operators a4 (k ) and a4 (k ) (with k 2 = 0) do
not have their own contributions to the dynamical variables, but they do contribute to them
via the combinations as (k) a4(k ) and a4 (k ) as(k), with k2 = 0, k 2 = 0 and s = 1, 2, 3.
In a sense, the operators a4 (k) and a4 (k ) act as operator-valued coupling constants for
the operators as (k) and as (k) (and hence for as (k) and a s (k)), as via them they bring an
additional contribution to the dynamical variables with respect to vector fields satisfying
the Lorenz condition. In this aspect, the operators a4 (k ) and a4 (k ) are similar to a 3 (k) and

a3 (k) (or a3 (k) and a3 (k)) (for details, see Sect. 10).
Consequently, the absence of the Lorenz conditions brings new problems, in addition

to the similar ones with the operators a 3 (k) and a3 (k) (see Sect. 10). One can say that
a massless free vector field, which does not satisfy the Lorenz condition, is equivalent to
a similar field satisfying that restriction and with self-interaction determined by a4 (k) and
a4 (k ). One can try to get rid of the contributions of the last operators in the dynamical
variables by imposing on them some subsidiary conditions, called gauges in terms of the

initial operators U (x) and U (x). The problem of gauge freedom of massless vector fields
is well-studied in the literature [1,4,6] to which the reader is referred. In particular, one can
set the operators (12.27) equals to zero by demanding
as (k) a4 (k ) + a4 (k ) as (k) = 0 as (k) a4 (k ) a4 (k) as (k) = 0 (12.31)
for k2 = 0, k 2 = 0, and s = 1, 2, 3, or, equivalently,
as (k) a4 (k ) = 0 a4 (k ) as (k) = 0 ( k2 = 0, k = 0, s = 1, 2, 3 ).

2
(12.32)
For instance, these equalities are identically satisfied if the Lorenz condition, e.g. in the
form (12.24), is valid.
Let us summarize. A theory of massless free vector field, based on the La-

grangian (12.2) contains as free parameters the operators a 3 (k) and a3 (k) and, partially,
a4 (k ) and a4 (k ), with k2 = 0 and k = 0. These operators have, generally, non-vanishing
2
contributions to the dynamical variables only via their compositions with the (physical)
operators a
s (k) and as (k) (or as (k) and as (k) ), with s = 1, 2, and via the combinations

a3 (k) a4 (k ) and a4 (k ) a3 (k). As a result, these operators describe degrees of freedom
with vanishing own dynamical characteristics and, consequently, they do not admit particle
interpretation.2 As we saw in Sect. 10, the operators a
s (k) as (k), with s = 1, 2, 3 and
k = 0, describe reasonably well massless free vector fields satisfying the Lorenz condi-
2
tion, in particular the electromagnetic field. From this point of view, we can say that the

operators a
3 (k), a3 (k), a4 (k ) and a4 (k ) describe some selfinteraction of the field, but it
seems such a selfinteracting, massless, free vector field is not known to exist in the Nature
for the time being. The easiest way for exclusion of that selfinteraction from the theory
is the pointed operators to be set equal to zero, i.e. on the field operators to be imposed
the Lorenz conditions and (10.10). However, it is possible that other restrictions on the
Lagrangian formalism may achieve the same goal.
At the end, the above considerations point that the Lorenz condition should be imposed
as an addition (subsidiary) condition to the Lagrangian formalism of massless free vec-
tor fields, in particular to the quantum theory of free electromagnetic field. Besides, the
conditions (10.10) also seems to be necessary for a satisfactory description of these fields.
13. Conclusion
A more or less detailed Lagrangian quantum field theory of free vector fields, massless
in Lorenz gauge and massive ones, in momentum picture was constructed in the present
chapter. Regardless of a common treatment of the both types of fields, the massless case
has some specific features and problems. The Lorenz conditions are external to the La-
grangian formalism of massless vector fields, but they are compatible with it. However,
for an electromagnetic field, which is a neutral massless vector field, in Lorenz gauge, we
have obtained a problem-free description in terms of creation and annihilation operators,
i.e. in terms of particles. This description is similar to the Gupta-Bleuler quantization of
electromagnetic field, but is quite different from the latter one and it is free of the problems
this formalism contains. Our formalism reproduces, under suitable additional conditions,
the quantization of electromagnetic field in Coulomb gauge.
Between the Lagrangians, considered for a suitable description of free vector fields
satisfying the Lorenz conditions, we have singled out the Lagrangian (11.3). It is invari-
ant under the transformation particleantiparticle, described in appropriate variables, so
that in it is encoded the charge symmetry (or spin-statistics theorem). The field equations
in terms of creation and annihilation operators for this Lagrangian are (11.16) (under the
conditions (11.17) and (11.18)). They can equivalently be rewritten as

+
+
[at (q),at (q)]+, a
s (k) + [at (q), at (q)]+ , as (k)

(13.1a)
= 2(1 + ( U ))a s (k)st (k q) f st (k, q)
3

+

[at (q),at (q)]+, as (k) + [at+(q), at (q)]+, as (k)
(13.1b)

= 2(1 + ( U ))a s (k)st (k q) f st (k, q).
3
2 If one assigns particle interpretation of the discussed operators, then they will have vanishing 4-momen-
tum, charge and spin and hence will be unobservable.
Trilinear equations of this kind are typical for the so-called parastatistics and para-
field theory [2933], in which they play a role of (para)commutation relations. In the next
chapter we intend to demonstrate how the parabose commutation relations for free vector
fields (satisfying the Lorenz condition) can be obtained from (13.1).
Chapter V
Commutation Relations
for Free Fields
Possible (algebraic) commutation relations in the
Lagrangian quantum theory of free (scalar, spinor
and vector) fields are considered from mathematical view-
point. As sources of these relations are employed the Heisen-
berg equations/relations for the dynamical variables and a spe-
cific condition for uniqueness of the operators of the dynamical vari-
ables (with respect to some class of Lagrangians). The paracommutation
relations and some their generalizations are pointed
as ones that entail the validity of all Heisenberg
equations. The simultaneous fulfillment of the
Heisenberg equations and the uniqueness require-
ment turn to be impossible. This problem is solved via
a redefinition of the dynamical variables, similar to the
normal ordering procedure and containing it as a
special case. That implies corresponding changes in
the admissible commutation relations. The intro-
duction of the concept of the vacuum makes narrow
the class of the possible commutation relations; in
particular, the mentioned redefinition of the dynamical variables is
reduced to normal ordering. As a last restriction on that class is
imposed the requirement for existing of an effective procedure
for calculating vacuum mean values. The standard bilinear
commutation relations are pointed as the only known
ones that satisfy all of the mentioned conditions
and do not contradict to the existing data.
1. Introduction
The main subject of this chapter is an analysis of possible (algebraic) commutation relations
in the Lagrangian quantum theory of free fields. These relations are considered only from
mathematical view-point and physical consequence of them, like the statistics of many-par-
ticle systems, are not investigated.
The canonical quantization method finds its origin in the classical Hamiltonian mechan-
ics [24,42] and naturally leads to the canonical (anti)commutation relations [2,5,11]. These
relations can be obtained from different assumptions (see, e.g., [1] and chapters IIIV) and
are one of the basic corner stones of the present-day quantum field theory.
Theoretically there are possible also non-canonical commutation relations. The best
known example of them being the so-called paracommutation relations [22, 29, 33]. But,
however, it seems no one of the presently known particles/fields obeys them.
In the present chapter is shown how different classes of commutation relations, un-
derstood in a broad sense as algebraic connections between creation and/or annihila-
tion operators, arise from the Lagrangian formalism, when applied to three types of La-
grangians describing free scalar, spinor and vector fields. 1 Their origin is twofold. On
one hand, a requirement for uniqueness of the dynamical variables (that can be calcu-
lated from Lagrangians leading to identical Euler-Lagrange equation) entails a number of
specific commutation relations. On another hand, any one of the so-called Heisenberg
relations/equations [2, 5], implies corresponding commutation relations; for example, the
paracommutation relations arise from the Heisenberg equations regarding the momentum
operator, when charge symmetric Lagrangian is employed.2 The combination of the both
methods leads to strong, generally incompatible, restrictions on the admissible types of
commutation relations.
The introduction of the concept of vacuum, combined with the mentioned uniqueness of
the operators of the dynamical variables, changes the situation and requires a redefinition
of these operators in a way similar to the one known as the normal ordering [1, 2, 5, 11],
which is its special case. Some natural assumptions reduce the former to the letter one; in
particular, in that way are excluded the paracommutation relations. However, this does not
reduce the possible commutation relations to the canonical ones. Further, the requirement
to be available an effective procedure for calculating vacuum mean (expectation) values, to
which reduce all predictable results in the theory, puts new restriction, whose only realistic
solution at the time being seems to be the standard canonical (anti)commutation relations.
The layout of this chapter, whose prototype is [48], is as follows.
In Sect. 2 are reviewed some basic results from the previous chapters, part of which can
be found also in works like [1,2,5,11]. In particular, the explicit expression of the dynamical
variables via the creation and annihilation operators are presented (without assuming some
commutation relations or normal ordering) and it is pointed to the existence of a family of
such variables for a given system of Euler-Lagrange equations for free fields. The last fact
1 It sees that the problem for determination of the commutation relations from the equations of motion is
first set by Wigner in [47]. As noted in this paper and as we shall see further, this procedure does not lead to
unique commutation relations and requires additional hypotheses to make them more specific.
2 Ordinary [2, 5], the commutation relations are postulated and the validity of the Heisenberg relations is
then verified. We follow the opposite method by postulating the Heisenberg equations and, then, looking for
commutation relations that are compatible with them.
Commutation Relations for Free Fields 215
is analyzed in Sect. 3, where a number of its consequences, having a sense of commutation

relations, are drawn. The Heisenberg relations and the commutation relations between
the dynamical variables are reviewed and analyzed in Sect. 4. It is pointed that the letter
should be consequences from the former ones. Arguments are presented that the Heisenberg
equation concerning the angular momentum operator should be split into two independent
ones, representing its orbital and spin parts, respectively.
Sect. 5 contains a method for assigning commutation relations to the Heisenberg equa-
tions. It is shown that the Heisenberg equation involving the orbital part of the angular
momentum gives rise to a differential, not algebraic, commutation relation and the one con-
cerning the spin part of the angular momentum implies a complicated integro-differential
connections between the creation and annihilation operators. Special attention is paid to
the paracommutation relations, whose particular kind are the ordinary ones, which ensure
the validity of the Heisenberg equations concerning the momentum operator. Partially is
analyzed the problem for compatibility of the different types of commutation relations de-
rived. It is proved that some generalization of the paracommutation relations ensures the
fulfillment of all of the Heisenberg relations.
Sect. 6 is devoted to consequences from the commutation relations derived in Sect. 5
under the conditions for uniqueness of the dynamical variables presented in Sect. 3. Gen-
erally, these requirements are incompatible with the commutation relations. To overcome
the problem, it is proposed a redefinition of the dynamical variables via a method similar
to (and generalizing) the normal ordering. This, of course, entails changes in the commu-
tation relations, the new versions of which happen to be compatible with the uniqueness
conditions and ensure the validity of the Heisenberg relations.
The concept of the vacuum is introduced in Sect. 7. It reduces (practically) the redefini-
tion of the operators of the dynamical variables to the one obtained via the normal ordering
procedure in the ordinary quantum field theory, but, without additional suppositions, does
not reduce the commutation relations to the standard bilinear ones. As a last step in specify-
ing the commutation relations as much as possible, we introduce the requirement the theory
to supply an effective way for calculating vacuum mean values of (anti-normally ordered)
products of creation and annihilation operators to which are reduced all predictable results,
in particular the mean values of the dynamical variables. The standard bilinear commu-
tation relation seem to be the only ones know at present that survive the last conditions,
however their uniqueness in this respect is not investigated.
Sect. 8 deals with the same problems as described above but for systems containing at
least two different quantum fields. The main obstacle is the establishment of commutation
relations between creation/annihilation operators concerning different fields. Argument is
presented that they should contain commutators or anticommutators of these operators. The
major of corresponding commutation relations are explicitly written and the results obtained
turn to be similar to the ones just described, only in multifield version.
Section 9 summarizes some of main results in the chapter.
The fulfillment of the relations (I.5.30) will not be supposed in this chapter until Sect. 5
(see also Sect. 4).
Let a
s (k) and as (k) be the creation/annihilation operators of some free particular field
(see, e.g., Sect. 2 below for a detailed explanation of the notation). We have the connections
1
as (k) = e
i x0 k
as (k)
1
k0 = m2 c2 + k 2 (1.1)
as (k) = e i x0 k as (k)
due to (II.4.10), (II.12.5), (III.5.54) and (IV.5.36).

Further it will be assumed a
s (k) and as (k) to be defined in Heisenberg picture, in-

dependently of as (k) and as (k), by means of the standard Lagrangian formalism. What
concerns the operators a
s (k) and as (k), we shall regard them as defined via (1.1); this
makes them independent from the momentum picture of motion. The fact that the so-de-
fined operators a
s (k) and as (k) coincide with the creation/annihilation operators in mo-
mentum picture (under the conditions (I.5.30)) will be inessential in the almost whole text
in this chapter.
At last, we ought to give an explanation why this chapter appears in the present mono-
graph when in it almost all considerations are done, in fact, in Heisenberg picture with
possible, but not necessary, usage of the creation and annihilation operators in momentum
picture. First of all, we essentially employ the obtained in chapters IIIV expressions for
the dynamical variables in momentum picture for three types of Lagrangians. The cor-
responding operators in Heisenberg picture, which in fact is used in this chapter, can be
obtained via a direct calculation, as it is partially done in, e.g., [1] for one of the mentioned
types of Lagrangians. The important point here is that in Heisenberg picture it suffice to
be used only the standard Lagrangian formalism, while in momentum picture one has to
suppose the commutativity between the components of the momentum operator and the
validity of the Heisenberg relations for it (see equations (I.4.4) and (I.4.1)). Since for the
analysis of the commutation relations we intend to do the fulfillment of these relations is
not necessary (they are subsidiary restrictions on the Lagrangian formalism), the Heisen-
berg picture of motion is the natural one that has to be used. For this reason, the expression
for the dynamical variables obtained in chapters IIIV will be used simply as their Heisen-
berg counterparts, but expressed via the creation and annihilation operators in momentum
picture. The only real advantage one gets in this way is the more natural structure of the
orbital angular momentum operator. As the commutation relations considered below are
algebraic ones, it is inessential in what picture of motion they are written or investigated.
2. Lagrangians, Euler-Lagrange Equations

and Dynamical Variables
In chapters IIIV we have investigated the Lagrangian quantum field theory of respectively
scalar, spin 12 and vector free fields. The main Lagrangians from which it was derived are
respectively (see chapters IIIV or, e.g. [1, 2, 5, 11]):
1
Lsc = Lsc ( , ) = m2 c4 (x) (x)
1 + ( )
(2.1a)
1
+ c2 2 ( (x)) ( (x))
1 + ( )
1
(x)C1 ( (x))
Lsp = Lsp = ic{
( , )
2 (2.1b)
(x))C1 (x)} + mc2
( (x)C1 (x)
m2 c4
Lv = Lv ( U , U ) =

U U
1 + ( U )
(2.1c)
c2 2
+ ( U ) ( U ) + ( U ) ( U )
1 + ( U )
Here it is used the following notation: (x) is a scalar field, a tilde (wave) over a symbol
means that it is in Heisenberg picture, the dagger denotes Hermitian conjugation, :=
:= C := C( 0 ) is its charge conjugate with
( 0 , 1 , 2 , 3 ) is a 4-spinor field,
being the Dirac gamma matrices and the matrix C satisfies the equations C1 C = and
C = C, U is a vector field, m is the fields mass (parameter) and the function

1 for A = A (Hermitian operator)
(A) := , (2.2)
0 for A = A (non-Hermitian operator)
with A : F F being an operator on the systems Hilbert space F of states, takes care of
is the field charged (non-Hermitian) or neutral (Hermitian, uncharged). Since a spinor field
is a charged one, we have ( ) = 0; sometimes below the number 0 = ( ) will be written
explicitly for unification of the notation.
We have explored also the consequences from the charge conjugate Lagrangians
Lsc = Lsc ( , ) := Lsc ( , ) (2.3a)

:= L ( ,
Lsp = Lsp ( , ) sp
) (2.3b)
Lv = Lv ( U , U ) := Lv ( U , U ), (2.3c)
as well as from the charge symmetric Lagrangians

1 1
Lsc = Lsc ( , ) := Lsc + Lsc = Lsc( , ) + Lsc ( , ) (2.4a)
2 2
1
1
Lsp = Lsp ( , ) := Lsp + Lsp = + L ( ,
Lsp ( , ) ) (2.4b)
sp
2 2
1 1
Lv = Lv ( U , U ) := Lv + Lv = Lv ( U , U ) + Lv ( U , U ) . (2.4c)
2 2
It is essential to be noted, for a massless, m = 0, vector field to the Lagrangian formalism
are added as subsidiary conditions the Lorenz conditions

U = 0 U = 0 (2.5)
on the solutions of the corresponding Euler-Lagrange equations. Besides, if the opposite is

not stated explicitly, no other restrictions, like the (anti)commutation relations, are supposed
to be imposed on the above Lagrangians. And a technical remark, for convenience, the
fields , and U and their charge conjugate , and U , respectively, are considered
as independent field variables.
Let L denotes any one of the Lagrangians (2.1) and L (resp. L ) the corresponding
to it Lagrangian given via (2.3) (resp. (2.4)). Physically the difference between L and L is
that the particles for L are antiparticles for L and vice versa. Both of the Lagrangians L
and L are not charge symmetric, i.e. the arising from them theories are not invariant under
the change particleantiparticle (or, in mathematical terms, under some of the changes
U U ) unless some additional hypotheses are made. Contrary to
, ,
this, the Lagrangian L is charge symmetric and, consequently, the formalism on its base
is invariant under the change particleantiparticle.1
The Euler-Lagrange equations for the Lagrangians L , L and L happen to coincide
(see chapters IIIV):2
L L L L L L
= 0, (2.6)
x ( ) x ( ) x ( )

U , U for respectively scalar, spinor and vector field.
where = , , , ,
Since the creation and annihilation operators are defined only on the base of Euler-La-
grange equations [1,2,5,11] (see also chapters IIIV), we can assert that these operators are
identical for the Lagrangians L , L and L . We shall denote these operators by a s (k)
+ +
and as (k) with the convention that as (k) (resp. as (k)) creates a particle (resp. antiparti-
cle) with 4-momentum ( m c + k2 , k), polarization s (see below) and charge (q) (resp.
2 2
(+q)) 3 and as (k) (resp. a s (k)) annihilates/destroys such a particle (resp. antiparticle).
Here and henceforth k R3 is interpreted as (anti)particles 3-momentum and the values of
the polarization index s depend on the field considered: s = 1 for a scalar field, s = 1 or
s = 1, 2 for respectively massless (m = 0) or massive (m = 0) spinor field, and s = 1, 2, 3 for
a vector field.4 Since massless vector fields modes with s = 3 may enter only in the spin
and orbital angular momenta operators (see chapter IV), we, for convenience, shall assume
that the polarization indices s,t, . . . take the values from 1 to 2 j + 1 0m (1 0 j ), where
j = 0, 12 , 1 is the spin for scalar, spinor and vector field, respectively, and 0m := 1 for m = 0
and 0m := 0 for m = 0;5 if the value s = 3 is important when j = 1 and m = 0, it will be
commented/considered separately. Of course, the creation and annihilation operators are
different for different fields; one should write, e.g., j a s
(k) for a
s (k), but we shall not use
such a complicated notation and will assume the dependence on j to be an implicit one.
1 Besides, under the same assumptions, the Lagrangian L does not admit quantization via anticommu-
tators (commutators) for integer (half-integer) spin field, while L and L do not make difference between
quantization via commutators or anticommutators.
2 Rigorously speaking, the Euler-Lagrange equations for the Lagrangian (2.4b) are identities like 0 = 0
see [3, section5.6]. However, bellow we shall handle this exceptional case as pointed in section III.10.
3 For a neutral field, we put q = 0.
4 For convenience, in chapter III, we have set s = 0 if m = 0 and s = 1,2 if m = 0 for a spinor field. For a
massless vector field, one may set s = 1,2, thus eliminating the unphysical value s = 3 for m = 0 see [1,2]

and chapter IV. In chapter II, for a scalar field, the notation
0 (k) and 0 (k) is used for a1 (k) and a1 (k),
respectively.
5 In this way the case ( j,s,m) = (1,3,0) is excluded from further considerations; if ( j,m) = (1,0) and
q = 0, the case considered further in this work corresponds to an electromagnetic field in Coulomb gauge, as
the modes with s = 3 are excluded (see section IV.10). However, if the case ( j,s,m) = (1,3,0) is important for
some reasons, the reader can easily obtain the corresponding results by applying the ones from chapter IV.
The following settings will be frequently used throughout this chapter:

0 for scalar field
j := 12 for spinor field

1 for vector field

1 for q = 0 (neutral (Hermitian) field) (2.7)
:=
0 for q = 0 (charged (non-Hermitian) field)

+1 for integer j (bose fields)
:= (1)2 j =
1 for half-integer j (fermi fields)
[A, B] := [A, B]1 := A B B A, (2.8)
where A and B are operators on the systems Hilbert space F of states.
The dynamical variables corresponding to L , L and L are, however, completely
different, unless some additional conditions are imposed on the Lagrangian formalism (see
chapters IIIV). In particular, the momentum operators P , charge operators Q , spin

operators S and orbital operators L , where = , , , for these Lagrangians are (see
chapters IIIV):
2 j+10m (10 j )
1
P =
1+ d3 kk |
k0 =

m2 c2 +k2
{as + (k) a
s (k)
(2.9a)
s=1
+ as (k) a+
s (k)}
2 j+10m (10 j )
1
P =
1+ d3 kk |
k0 =

m2 c2 +k2
{a+
s (k) as (k)
(2.9b)
s=1
+ a +
s (k) as (k)}
2 j+10m (10 j )
1
P =
2(1 + ) d3 kk |
k0 =

m2 c2 +k2
{[as + (k), a
s (k)]
(2.9c)
s=1
+ [a+
s (k), as (k)]}

2 j+10m (10 j )
Q = +q d3 k{as + (k) a +
s (k) as (k) as (k)} (2.10a)
s=1
2 j+10m (10 j )
Q = q d3 k{a+ +
s (k) as (k) as (k) as (k)} (2.10b)
s=1
2 j+1 (1 )
0m 0j
1
Q = q
2 d3 k{[a+ +
s (k), as (k)] [as (k), as (k)] } (2.10c)
s=1
(1) j1/2 j 2 j+10m (11 j ) ss ,

S =
1+
d3 k (k)as + (k) a
s (k)
(2.11a)
s,s =1
,+
+

+ ss
(k)as (k) as (k)

2 j+10m (11 j )
(1) j1/2 j ss ,+

S =
1+ d3 k (k)a+
s (k) as (k)
(2.11b)
s,s =1
ss ,
+ (k)a +
s (k) as (k)
2 j+10m (11 j )
(1) j1/2 j ss ,

S =
2(1 + ) d3 k (k)[as + (k), a
s (k)]
(2.11c)
s,s =1
ss ,+ +

+ (k)[a
s (k), as (k)]

L =x0 P x0 P
2 j+10m (11 j )
(1) j1/2 j ss ,
+
1+ d3 k l (k)as + (k) a
s (k)
s,s =1
ss ,+
+ l (k)as (k) a+
s (k)
2 j+10m (10 j )

i
+
2(1 + ) d3 k as + (k) k k a
k k s (k)
s=1

as (k) k k a+

(k) (2.12a)
k k s
k0 = m2 c2 +k 2

L =x0 P x0 P
2 j+10m (11 j )
(1) j1/2 j ss ,+
+
1+ d3 k l (k)a+
s (k) as (k)
s,s =1
ss ,
+ l (k)a +
s (k) as (k)
2 j+10m (10 j )

i
+
2(1 + ) d3 k a+
s (k) k
k
k
k
as (k)
s=1

as (k) k k as + (k) 2 2 2

(2.12b)
k k k0 = m c +k

L =x0 P x0 P
2 j+10m (11 j )
(1) j1/2 j ss ,
+
2(1 + ) d3 k l (k)[as +(k), a
s (k)]
s,s =1
ss ,+ +

+ l (k)[a s (k), as (k)]
2 j+10m (10 j )

i
+
4(1 + ) 3 +
d k as (k) k k as (k)
k k
s=1

+
as (k) k k as (k) + as (k) k k as (k)
+
k k k k

as (k) k k a+ (k) . (2.12c)
k k s
k0 = m2 c2 +k 2
we have used the following notation: (1)

Here n+1/2 := (1)n i for all n N and i :=
+ 1,

A(k) B(k)
A(k)k B(k) := k B(k) + A(k) k
k k k

= k A(k) B(k) (2.13)
k
ss , ss ,
for operators A(k) and B(k) having C1 dependence on k,6 and (k) and l (k) are
7
some functions of k such that

ss , ss , ss , ss ,
(k) = (k) l (k) = l (k)

ss , ss ,
(k) = l (k) = 0 for j = 0 (scalar field)
ss , ss ,+ ss
(k) = (k) =: (k)
(2.14)
=
ss
(k) = ss
(k) for j = 1 (vector field)

ss , ss ,+
l (k) = l (k) =: l
ss
(k)

= l
ss
(k) = l
ss
(k) for j = 1 (vector field).
A technical remark must be make at this point. The equations (2.9)(2.12) were de-
rived in chapters IIIV under some additional conditions, represented by equations (I.4.4)
and (I.4.1), which are considered bellow in Sect. 4 and ensure the effectiveness of the mo-
mentum picture of motion [21] (see chapter I). However, as it is partially proved, e.g.,
in [1], when the quantities (2.9)(2.12) are expressed via the Heisenberg creation and anni-
hilation operators (see (1.1)), they remain valid, up to a phase factor, and without making
the mentioned assumptions, i.e. these assumptions are needless when one works entirely in
Heisenberg picture. For this reason, we shall consider (2.9)(2.12) as pure consequence of
the Lagrangian formalism.
We should emphasize, in (2.11) and (2.12) with S and L , = , , , are denoted

the spin and orbital, respectively, operators for L , which are the spacetime-independent
parts of the spin and orbital, respectively, angular momentum operators [49] (see also chap-

ter III); if the last operators are denoted by S and L , the total angular momentum
operator of a system with Lagrangian L is [49]

M = L + S = L

+ S , = , , (2.15)

and S = S (and hence L = L ) iff S is a conserved operator or, equivalently, iff
the systems canonical energy-momentum tensor is symmetric.8
6 More generally, if : {F F } {F F } is a mapping on the operator space over the systems

used for = .
7 For the explicit form of these functions, see (III.4.26), (III.5.26), (IV.5.16), (III.5.44) and (IV.5.8); see
also equation (5.57) below and example III. 5.1 on page 105.
8 In [49] and in chapter III the spin and orbital operators are labeled with an additional left superscript
Going ahead (see Sect. 5), we would like to note that the expressions (2.9c) and, conse-
quently, the Lagrangian L are the base from which the paracommutation relations were
first derived [29].
And a last remark. Above we have expressed the dynamical variables in Heisenberg
picture via the creation and annihilation operators in momentum picture. If one works
entirely in Heisenberg picture, the operators (1.1), representing the creation and annihilation
operators in Heisenberg picture, should be used. Besides, by virtue of the equations
(a
s (k)) = as (k) (as (k)) = as (k)

(2.16)

as (k) = as (k) as (k) = a
s (k), (2.17)
some of the relations concerning as (k), e.g. the Euler-Lagrange and Heisenberg equa-
tions, are consequences of the similar ones regarding a s (k). In view of (1.1), we shall con-
sider (2.9)(2.12) as obtained form the corresponding expressions in Heisenberg picture by
making the replacements a
s (k)
as (k) and as (k)
as (k). So, (2.9)(2.12) will have,
up to a phase factor, a sense of dynamical variables in Heisenberg picture expressed via the
creation/annihilation operators in momentum picture.
3. On the Uniqueness of the Dynamical Variables

Let D = P , Q , S, L denotes some dynamical variable, viz. the momentum, charge,
spin, or orbital operator, of a system with Lagrangian L . Since the Euler-Lagrange equa-
tions for the Lagrangians L , L and L coincide (see (2.6)), we can assert that any field

satisfying these

equations

admits at least three classes of conserved operators, viz. D , D
and D = 2 D + D . Moreover, it can be proved that the Euler-Lagrange equations for
1
the Lagrangian
L, := L + L + = 0 (3.1)
do not depend on , C and coincide with (2.6). 1 Therefore there exists a two parameter
family of conserved dynamical variables for these equations given via
D, := D + D + = 0. (3.2)
Evidently L = L 1 , 1 and D = D 1 , 1 . Since the Euler-Lagrange equations (2.6) are
2 2 2 2
linear and homogeneous (in the cases considered), we can, without a lost of generality,
restrict the parameters , C to such that
+ = 1, (3.3)

, which, for brevity, is omitted in the present chapter as in it only these operators, not S and L , will

be considered. Notice, the operators S and L are, generally, time-dependent while the orbital and spin
ones are conserved, as a result of which the total angular momentum is a conserved operator too [49] (see also
chapter III).
1 The Lagrangian (3.1) may have symmetries which are not ones for L and L ; respectively, this is true
for the field theories arising from these Lagrangians. Particular examples illustrating that fact were presented
in sections II.14, III.10 and IV.11.
which can be achieved by an appropriate renormalization (by a factor ( + )1/2) of the

field operators. Thus any field satisfying the Euler-Lagrange equations (2.6) admits the
family D, , + = 1, of conserved operators. Obviously, this conclusion is valid if
in (3.1) we replace the particular Lagrangians L and L (see (2.1) and (2.3)) with any two
Lagrangians (of one and the same field variables) which lead to identical Euler-Lagrange
equations. However, the essential point in our case is that L and L do not differ only
by a full divergence, as a result of which the operators D, are different for different pairs
(, ), + = 1.2
Since one expects a physical system to possess uniquely defined dynamical character-
istics, e.g. energy and total angular momentum, and the Euler-Lagrange equations are con-
sidered (in the framework of Lagrangian formalism) as the ones governing the spacetime
evolution of the system considered, the problem arises when the dynamical operators D, ,
+ = 1, are independent of the particular choice of and , i.e. of the initial Lagrangian
one starts off. Simple calculation show that the operators (3.2), under the condition (3.1),
are independent of the particular values of the parameters and if and only if
D = D . (3.4)
Some consequences of the condition(s) (3.4) will be considered below, as well as pos-
sible ways for satisfying these restrictions in the Lagrangian formalism.
Combining (2.9)(2.12) with (3.4), for respectively D = P , Q , S , L , we see that
a free scalar, spinor or vector field has a uniquely defined dynamical variables if and only if
the following equations are fulfilled:
2 j+10m (10 j ) +
d3 k k
k0 =

m2 c2 +k2
as (k) a +
s (k) as (k) as (k)
s=1

a+ +
s (k) as (k) + as (k) as (k) = 0 (3.5)
2 j+10m (10 j )

q d3 k as + (k) a +
s (k) as (k) as (k)
s=1

+ a+ +
s (k) as (k) as (k) as (k) = 0 (3.6)
2 j+10m (11 j )
ss , ss ,
d3 k +
s (k) as (k) (k)as (k) as (k)
(k)a+
s,s =1
ss ,+ ss ,+
(k)a+ +
s (k) as (k) + (k)as (k) as (k) = 0 (3.7)

2 j+10m (11 j )
ss, ss ,
d3 k l (k)as + (k) a +
s,s =1

ss ,+ ss ,+
l (k)a+ +
s (k) as (k) + l (k)as (k) as (k)
2 Note, no commutativity or some commutation relations between the field operators and their charge (or
Hermitian) conjugate are presupposed, i.e., at the moment, we work in a theory without such relations and
normal ordering.
2 j+10m (10 j)

1
+
2 3 +
d k as (k) k k as (k)
k k
s=1

+
+ a
s (k) k k as (k)
k k

+
as (k) k k as (k)
k k

as (k) k k a+

(k) = 0. (3.8)
k k s
k0 = m2 c2 +k2
In (3.6) is retained the constant factor q as in the neutral case it is equal to zero and,
consequently, the equation (3.6) reduces to identity.
The equations (3.5)(3.8) can be regarded as subsidiary conditions on the Lagrangian
formalism and can serve as equations for (partial) determination of the creation and anni-
hilation operators. The system of integral equations (3.5)(3.8) is quite complicated and
we are not going to investigate it in the general case. Below we shall restrict ourselves to
analysis of only those solutions of (3.5)(3.8), if any, for which the integrands in (3.5)
(3.8) vanish. This means that we shall replace the system of integral equations (3.5)(3.8)
with respect to creation and annihilation operators with the following system of algebraic
equations (do not sum over s and s in (3.12) and (3.13)!):
as +(k) a + + +
s (k) as (k) as (k) as (k) as (k) + as (k) as (k) = 0 (3.9)

as +(k) a + +
s (k) as (k) as (k) + as (k) as (k)
as (k) a+
s (k) = 0 if q = 0 (3.10)

as + (k) k k a
s (k) + as (k) k k as + (k)
k k k k

a+ (k) k k as (k)
s
k k

as (k) k k a+

(k) = 0 (3.11)
k k s
k0 = m2 c2 +k2
ss , ss ,

s (k) as (k) (k)as (k) as (k)
(k)a+ +
s,s
ss ,+ ss ,+
(k)a+ +
s (k) as (k) + (k)as (k) as (k) = 0 (3.12)

ss ,
l (k)as + (k) a ss , +
s,s
ss ,+ ss ,+
l (k)a+ +
s (k) as (k) + l (k)as (k) as (k) = 0 (3.13)
Here: s = 1, . . ., 2 j + 1 0m (1 0 j ) in (3.9)(3.11) and s, s = 1, . . ., 2 j + 1 0m (1

1 j ) in (3.12) and (3.13). (Notice, by virtue of (2.14), the equations (3.12) and (3.13) are
identically valid for j = 0, i.e. for scalar fields.) Since all polarization indices enter in (3.5)
and (3.6) on equal footing, we do not sum over s in (3.9)(3.11). But in (3.12) and (3.13) we
have retain the summation sign as the modes with definite polarization cannot be singled out
in the general case. One may obtain weaker versions of (3.9)(3.13) by summing in them
over the polarization indices, but we shall not consider these conditions below regardless of
the fact that they also ensure uniqueness of the dynamical variables.
At first, consider the equations (3.9)(3.11). Since for a neutral field, q = 0, we have
as (k) = a s (k), which physically means coincidence of fields particles and antiparticles,
the equations (3.9)(3.11) hold identically in this case.
Let consider now the case q = 0, i.e. the investigated field to be charged one. Using the
standard notation (cf. (2.8))
[A, B] := A B + B A, (3.14)
for operators A and B and C, we rewrite (3.9) and (3.10) as
[as +(k), a +
s (k)] [as (k), as (k)] = 0

(3.9 )
[as +(k), a +
s (k)] + [as (k), as (k)] = 0

if q = 0, (3.10 )
which are equivalent to

s (k), as (k)] = 0
[a if q = 0. (3.15)
Exercise 3.1. Differentiating (3.15) and inserting the result into (3.11), verify that (3.11) is
tantamount to

as + (k), k k a (k)
k k s

a+ (k), k
k a
(k) =0 if q = 0, (3.16)
s
k k s
k0 = m2 c2 +k2
Consider now (3.12) and (3.13). By means of the shorthand (3.14), they read
ss , ss ,+
+
s (k), as (k)] + (k)[as (k), as (k)] = 0
(k)[a+
(3.17)
s,s
ss , ss ,+
l (k)[a+ +
s (k), as (k)] + l (k)[as (k), as (k)] = 0.

(3.18)
s,s
For a scalar field, j = 0, these conditions hold identically, due to (2.14). But for j = 0
they impose new restrictions on the formalism. In particular, for vector fields, j = 1 and
= +1 they are satisfied iff (see (2.14))
+
s (k), as (k)] [as (k), as (k)]
[a+
[as+ (k), a +
s (k)] + [as (k), as (k)] = 0. (3.19)
One can satisfy (3.17) and (3.18) if the following generalization of (3.15) holds
[as (k), a
s (k)] = 0. (3.20)
For spin j = 12 (and hence = 1 see (2.7)), the conditions (3.12) and (3.13) cannot
be simplified much, but, if one requires the vanishment of the operator coefficients after
ss , ss ,
(k) and l (k), one gets
1
as (k) a
s (k) = 0 j= = 1. (3.21)
2
Excluding some special cases, e.g. neutral scalar field (q = 0 and j = 0), the equa-
tions (3.15) and (3.21) are unacceptable from many viewpoints. The main of them
is that they are incompatible with the ordinary (anti)commutation relations (see, e.g.,
e.g. [1, 2, 11, 22] or Sect. 5, in particular, equations (5.13) bellow); for example, (3.21)
means that the acts of creation and annihilation of (anti)particles with identical character-
istics should be mutually independent, which contradicts to the existing theory and experi-
mental data.
Now we shall try another way for achieving uniqueness of the dynamical variables for
free fields. Since in (3.9)(3.13) naturally appear (anti)commutators between creation and
annihilation operators and these (anti)commutators vanish under the standard normal order-
ing [1, 2, 11, 22], one may suppose that the normally ordered expressions of the dynamical
variables may coincide. Let us analyze this method.
Recall [1, 2, 5, 11], the normal ordering operator N (for free field theory) is a linear
operator on the operator space of the system considered such that to a product (composition)
c1 cn of n N creation and/or annihilation operators c1 , . . . cn it assigns the operator
(1) f c1 cn . Here (1 , . . ., n ) is a permutation of (1, . . ., n), all creation operators
stand to the left of all annihilation ones, the relative order between the creation/annihilation
operators is preserved, and f is equal to the number of transpositions among the fermion
operators ( j = 12 ) needed to be achieved the just-described order (normal order) of the
operators c1 , . . ., cn in c1 cn .3 In particular this means that
+
N a+ s (k)at (p) = as (k)at (p) N as + (k)at(p) = as + (k)at (p)
+ +
N a s (k)at (p) = at (p)as (k)

N as (k)at+(p) = at+ (p)as (k) (3.22)
and, consequently, we have

N [as (k), at(p)] = 0 N [a
s (k), at (p)] = 0, (3.23)
due to := (1)2 j = 1 (see (2.7)). (In fact, below only the equalities (3.22) and (3.23),
not the general definition of a normal product, will be applied.)
Applying the normal ordering operator to (3.9 ), (3.10 ), (3.17) and (3.18), we, in view
of (3.23), get the identity 0 = 0, which means that the conditions (3.9), (3.10), (3.12)
and (3.13) are identically satisfied after normal ordering. This is confirmed by the ap-
plication of N to (2.9) and (2.10), which results respectively in (see (3.22))
3 We have slightly modified the definition given in [1, 2, 5, 11] because no (anti)commutation relations are
presented in our exposition till this point in this chapter. In this chapter we do not concern the problem for
elimination of the unphysical operators a
3 (k) and a3 (k) from the spin and orbital momentum operators
when j = 1; for details, see chapter IV, where it is proved that, for an electromagnetic field, j = 1 and q = 0,
one way to achieve this is by adding to the number f above the number of transpositions between a s (k),
s = 1,2, and a
3 (k) needed for getting normal order.
1
N ( P ) = N ( P ) =
1+
2 j+10m (10 j )
d3 kk |
k0 =

m2 c2 +k2
{a+ +
s (k) as (k) + as (k) as (k)} (3.24)
s=1
1
N ( Q ) = N ( Q ) =
1+
2 j+10m (10 j )
d3 k{as + (k) a +
s (k) as (k) as (k)}. (3.25)
s=1
Therefore the normal ordering ensures the uniqueness of the momentum and charge oper-
ators, if we redefine them respectively as
P := N ( P ) Q := N ( Q ). (3.26)
Exercise 3.2. Putting := k k k k and using (3.22), prove that

N a+ +
s (k) as (k) = as (k) as (k)

N as +(k) a +
s (k) = as (k) as (k)
(3.27)
+
N a
s (k) as (k)
= as + (k) a
s (k)

N as (k) a+ +
s (k) = as (k) as (k).
As a consequence of these equalities, the action of N on the l.h.s. of (3.11) vanishes.

Combining this result with the mentioned fact that the normal ordering converts (3.12)
and (3.13) into identities, we see that the normal ordering procedure ensures also uniqueness
of the spin and orbital operators if we redefine them respectively as:
(1) j1/2 j
S := N ( S ) := N ( S )=
1+
2 j+10m (11 j )
ss ,
d3 k (k)as + (k) a
s (k)
s,s =1
ss ,+
+ (k)a+
s (k) as (k) (3.28)
(1) j1/2 j
L := N ( L ) := N ( L ) = x0 P x0 P +
1+
2 j+10m (11 j )
ss, ss ,+
d3 k l (k)as + (k) a +
s (k) + l (k)as (k) as (k)
s,s =1
2 j+10m (10 j )

i
+
2(1 + ) d3 k as + (k) k k a
k k s (k)
s=1

+ as (k) k k as (k) 2 2 2 ,
+
(3.29)
k k k0 = m c +k
where (2.14) was applied.

4. Heisenberg Relations
The conserved operators, like momentum and charge operators, are often identified with the
generators of the corresponding transformations under which the action operator is invari-
ant [1, 2, 5, 11]. As pointed in section I.4, this leads to a number of commutation relations
between the components of these operators and between them and the field operators. The
relations of the letter set are known/referred as the Heisenberg relations or equations. Both
kinds of commutation relations are from pure geometric origin and, consequently, are com-
pletely external to the Lagrangian formalism; one of the reasons being that the mentioned
identification is, in general, unacceptable and may be carried out only on some subset of the
systems Hilbert space of states [27, 49]. Therefore their validity in a pure Lagrangian the-
ory is questionable and should be verified [2]. However, the considered relations are weaker
conditions than the identification of the corresponding operators and there are strong evi-
dences that these relations should be valid in a realistic quantum field theory [1, 2]; e.g., the
commutativity between the momentum and charge operators (see below (4.18)) expresses
the experimental fact that the 4-momentum and charge of any system are simultaneously
measurable quantities.
It is known [1, 2], in a pure Lagrangian approach, the field equations, which are usually
identified with the Euler-Lagrange, 1 are the only restrictions on the field operators. Besides,
these equations do not determine uniquely the field operators and the letter can be expressed
through the creation and annihilation operators. Since the last operators are left completely
arbitrary by a pure Lagrangian formalism, one is free to impose on them any system of
compatible restrictions. The best known examples of this kind are the famous canonical
(anti)commutation relations and their generalization, the so-called paracommutation rela-
tions [22,29]. In general, the problem for compatibility of such subsidiary to the Lagrangian
formalism system of restrictions with, for instance, the Heisenberg relations is open and
requires particular investigation [2]. For example, even the canonical (anti)commutation
relations for electromagnetic field in Coulomb gauge are incompatible with the Heisenberg
equation involving the (total) angular momentum operator unless the gauge symmetry of
this field is taken into account [2, 84]. However, the (para)commutation relations are,
by construction, compatible with the Heisenberg relations regarding momentum operator
(see [29] or below Subsect. 5.1). The ordinary approach is to be imposed a system of equa-
tions on the creation and annihilation operators and, then, to be checked its compatibility
with, e.g., the Heisenberg relations. In the next sections we shall investigate the opposite
situation: assuming the validity of (some of) the Heisenberg equations, the possible restric-
tions on the creation and annihilation operators will be explored. For this purpose, below
we briefly review the Heisenberg relations and other ones related to them.
Consider a system of quantum fields i (x), i = 1, . . ., N N, where i (x) denote the
components of all fields (and their Hermitian conjugates), and P , Q and M be its mo-
mentum, charge and (total) angular momentum operators, respectively. The Heisenberg
1
Recall, there are Lagrangians whose classical Euler-Lagrange equations are identities. However, their
correct and rigorous treatment [3] reveals that they entail field equations which are mathematically correct and
physically sensible.
relations/equations for these operators are [1, 2, 5, 11]
i (x)
[ i (x), P ] = i (4.1)
x
[ i (x), Q ] = e( i )qi i (x) (4.2)
[ i (x), M ] = i{x i (x) x i (x)} + i i
Ii i (x). (4.3)
i
Here: q = const is the fields charge, e( i ) = 0 if i = i , e( i ) = 1 if i = i with

i = I i characterize the transformation properties
e( i ) + e( i ) = 0, and the constants Ii i
of the field operators under 4-rotations. (If e( i ) = 0, it is a convention whether to put
e( i ) = +1 or e( i ) = 1 for a fixed i.)
We would like to make some comments on (4.3). Since its r.h.s. is a sum of two opera-
tors, the first (second) characterizing the pure orbital (spin) angular momentum properties
of the system considered, the idea arises to split (4.3) into two independent equations, one
involving the orbital angular momentum operator and another one concerning the spin an-
gular momentum operator. This is supported by the observation that, it seems, no process
is known for transforming orbital angular momentum into spin one and v.v. (without de-
stroying the system). So one may suppose the existence of operators M or and M sp such

that
[ i (x), M
or
] = i{x i (x) x i (x)} (4.4)
[ i (x), Msp ] = i i
Ii i (x) (4.5)
i
M = Mor + M .
sp
(4.6)
However, as particular calculations demonstrate (see chapters III and IV and [7]), nei-
ther the spin (resp. orbital) nor the spin (resp. orbital) angular momentum operator is a suit-
able candidate for M (resp. M
sp or ). If we assume the validity of (4.1), then equations (4.4)
and (4.5) can be satisfied if we choose
Mor(x) = L
ext
:= x P x P (4.7)
Msp(x) = M(0) (x) := M L
ext
= S + L {x P x P } (4.8)
with M satisfying (4.3). These operators are not conserved ones. Such a representation
is in agreement with the equations (2.12), according to which the operator (4.7) enters ad-
ditively in the expressions for the orbital operator.2 The physical sense of the operator (4.7)
is that it represents the orbital angular momentum of the system due to its movement as
a whole. Respectively, the operator (4.8) describes the systems angular momentum as a
result of its internal movement and/or structure.
Since the spin (orbital) angular momentum is associated with the structure (movement)
of a system, in the operator (4.8) are mixed the spin and orbital angular momenta. These
2 This is evident in the momentum picture of motion, in which x stands for x0 in (2.12) see chapters II
IV.
quantities can be separated completely via the following representations of the operators
Mor and Msp in momentum picture (when (4.1) holds)
Mor = x P x P + L
int
(4.9)
Msp = M (x P x P) L
int
, (4.10)
where L int describes the internal orbital angular momentum of the system considered
and depends on the Lagrangian we have started off. Generally said, L int
is the part of the
orbital angular momentum operator containing derivatives of the creation and annihilation
operators. In particular, for the Lagrangians L , L and L (see Sect. 2), the explicit
forms of the operators (4.9) and (4.10) respectively are (see (2.11) and (2.12)):
M or = x P x P
2 j+10m (10 j )

i
+
2(1 + ) d k as (k) k k a
3 +
k k s (k)
(4.11a)
s=1

as (k) k k a+ (k)
k k s
k0 = m2 c2 +k2
or
M = x P x P
2 j+10m (10 j )

i
+
2(1 + ) d 3
k a +
s (k) k
k
k
k
as (k)
(4.11b)
s=1

+
a (k) k k as (k) 2 2 2
s
k k k0 = m c +k
or
M = x P x P
2 j+10m (10 j )

i
+
4(1 + ) d k as (k) k k a
3 +
k k s (k)
s=1

as (k) k k as + (k)

(4.11c)
k k

+ a+ s (k) k k as (k)
k k

as (k) k k a+ s (k) .
k k k0 = m2 c2 +k2
(1) j1/2 j
M sp =
1+
2 j+10m (11 j )
ss , ss ,
d3 k ( (k) + l (k))a+
s (k) as (k)
(4.12a)
s,s =1
ss ,+
ss ,+
+ ( (k) + l (k))as (k) a+
s (k)
(1) j1/2 j
M sp =
1+
2 j+10m (11 j )
ss ,+ ss ,+
d3 k ( (k) + l (k))a+
s (k) as (k)
(4.12b)
s,s =1
ss , ss ,
+ ( (k) + (k))a +
s (k) as (k)
(1) j1/2 j
M sp =
2(1 + )
2 j+10m (11 j )
ss , ss , (4.12c)
d3 k ( (k) + l (k))[a+
s (k), as (k)]
s,s =1
ss ,+
ss ,+ +

+ ( (k) + l (k))[a
s (k), as (k)] .
Obviously (see chapter I), the equations (4.12) have the same form in Heisenberg pic-
ture in terms of the operators (1.1) (only tildes over M and a must be added), but the
equations (4.11) change substantially due to the existence of derivatives of the creation and
annihilation operators in them (chapters IIIV):
i
M or =
2(1 + )
2 j+10m (10 j )

d k 3
a +
s (k) k
k
k
k
as (k) (4.13a)
s=1

as (k) k k a+ s (k)
k k k0 = m2 c2 +k2
i
M or =
2(1 + )
2 j+10m (10 j )

d3 k a+ s (k) k
k
k
k
as (k) (4.13b)
s=1

as (k) k k as + (k) 2 2 2

k k k0 = m c +k
i
M or =
4(1 + )
2 j+10m (10 j )

d3 k as + (k) k k a
k k s (k)
s=1

as (k) k k as + (k)
(4.13c)
k k

+ a+ (k) k k as (k)
s
k k

+

as (k) k k as (k) 2 2 2 .
k k k0 = m c +k
From (4.13) and (4.12) is clear that the operators M and M so defined are con-
or sp
served (contrary to (4.7) and (4.8)) and do not depend on the validity of the Heisenberg
relations (4.1) (contrary to expressions (4.11) in momentum picture).
The problem for whether the operators (4.12) and (4.13) satisfy the equations
(4.4) and (4.5), respectively, will be considered in Sect. 5.
There is an essential difference between (4.4) and (4.5): the equation (4.5) depends
i
on the particular properties of the operators i (x) under 4-rotations via the coefficients Ii
(see (4.25) below), while (4.4) does not depend on them. This is explicitly reflected in (4.11)
and (4.12): the former set of equations is valid independently of the geometrical nature
of the fields considered, while the latter one depends on it via the spin (polarization)
ss , ss ,
functions (k) and l (k). Similar remark concerns (4.3), on one hand, and (4.1)
and (4.2), on another hand: the particular form of (4.3) essentially depends on the geometric
properties of i (x) under 4-rotations, the other equations being independent of them.
It should also be noted, the relation (4.3) does not hold for a canonically quantized
electromagnetic field in Coulomb gauge unless some additional terms in its r.h.s., reflecting
the gauge symmetry of the field, are taken into account [2, 84].
As it was said above, the relations (4.1)(4.3) are from pure geometrical origin. How-
ever, the last discussion, concerning (4.4)(4.8), reveals that the terms in braces in (4.3)
should be connected with the momentum operator in the (pure) Lagrangian approach. More
precisely, on the background of equations (2.11a)(2.12), the Heisenberg relation (4.3)
should be replaced with
[ i (x), M ] = x [ i (x), P ] x [ i (x), P ] + i Ii

i
i (x), (4.14)
j
which is equivalent to (4.3) if (4.1) is true. An advantage of the last equation is that it
is valid in any picture of motion (in the same form) while (4.3) holds only in Heisenberg
picture.3 Obviously, (4.14) is equivalent to (4.5) with M defined by (4.8).
sp
The other kind of geometric relations mentioned at the beginning of this section are con-
nected with the basic relations defining the Lie algebra of the Poincare group [9, pp. 143
3 In other pictures of motion, generally, additional terms in the r h.s. of (4.3) will appear, i.e. the functional
form of the r h.s. of (4.3) is not invariant under changes of the picture of motion, contrary to (4.14).
147], [10, sect. 7.1]. They require the fulfillment of the following equations between the
components P of the momentum and M of the angular momentum operators [5,7,9,10]:
[ P , P ] = 0 (4.15)
[ M , P] = i( P P ). (4.16)

[ M , M ] = i M M M + M . (4.17)
We would like to pay attention to the minus sign in the multiplier (i) in (4.16)
and (4.17) with respect to the above references, where i stands instead of i in these
equations. When (a representation of) the Lie algebra of the Poincare group is considered,
this difference in the sign is insignificant as it can be absorbed into the definition of M .
However, the change of the sign of the angular momentum operator, M
M , will
result in the change i
i in the r.h.s. of (4.3). This means that equations (4.15), (4.16)
and (4.3), when considered together, require a suitable choice of the signs of the multiplier
i in their right hand sides as these signs change simultaneously when M is replaced
with M . Since equations (4.3), (4.16) and (4.17) hold, when M is defined accord-
ing to the Noethers theorem and the ordinary (anti)commutation relations are valid (see
chapters IIIV), we accept these equations in the way they are written above.
To the relations (4.15)(4.17) should be added the equations [5, p. 78]
[ Q , P ] = 0 (4.18)
[ Q , M ] = 0, (4.19)
which complete the algebra of observables and express, respectively, the translational and
rotational invariance of the charge operator Q ; physically they mean that the charge and
momentum or the charge and angular momentum are simultaneously measurable quantities.
Since the spin properties of a system are generally independent of its charge or momen-
tum, one may also expect the validity of the relations4
[ S , P ] = 0 (4.20)
[ S , Q ] = 0. (4.21)
But, as the spin describes, in a sense, some of the rotational properties of the system,
equality like [ S , L ] = 0 is not likely to hold. Indeed, the considerations in chapters II
IV reveal that (4.20) and (4.21), but not the last equation, are true in the framework of
the Lagrangian formalism with added to it standard (anti)commutation relations. Notice,
if (4.20) and (4.21) hold, then, respectively, (4.16) and (4.19) are equivalent to
[ L , P ] = i( P P ) (4.22)
[ Q , L] = 0. (4.23)
It is intuitively clear, not all of the commutation relations (4.1)(4.3) and (4.15)
(4.21) are independent: if D denotes some of the operators P , Q , M , S
4 Recall, S (resp. L ) is the conserved spin (resp. orbital) operator, not the generally non-conserved spin

(resp. orbital) angular momentum operator [49].
or L and the commutators [ i (x), D ] , i = 1, . . ., N, are known, then, in princi-

ple, one can calculate the commutators [( 1 (x), . . ., N (x)), D ] , where ( 1 (x), . . .,
N (x)) is, for example, any function/functional bilinear in 1 (x), . . ., N (x); to prove this
fact, one should apply the identity [A, B C] = [A, B] C + B [A,C] a suitable number
of times. In particular, if D1 and D2 denote any two (distinct) operators of the dynamical
variables, and [ i (x), D1 ] is known, then the commutator [ D1 , D2 ] can be calculated ex-
plicitly. For this reason, we can expect that:
(i) Equation (4.1) implies (4.15), (4.16), (4.18), (4.20) and (4.22).
(ii) Equation (4.2) implies (4.18), (4.19), (4.21), and (4.23).
(iii) Equation (4.3) implies (4.16), (4.17), and (4.19).
Besides, (4.3) may, possibly, entail equations like (4.17) with S or L for M, with an excep-
tion of M in the l.h.s., i.e.

[ S , M ] = i S S S + S
(4.24)
[ L , M ] = i L L L + L .
The validity of assertions (i)(iii) above for free scalar, spinor and vector fields, when
respectively

i (x)
(x), (x) Ii
i

I = 0 e( ) = e( ) = +1 (4.25a)
i i
i (x)
(x), (x)
Ii
I = I =
2 (4.25b)
e( ) = e( )
= +1

i (x)
U (x), U (x) Ii
i

I = I =
(4.25c)
e( U ) = e( U ) = +1,
where := 2i [ , ] with being the Dirac -matrices [1, 41], is proved in chapters II
IV, respectively. Besides, in them is proved that equations (4.24) hold for scalar and vector
fields, but not for a spinor field.5
Thus, we see that the Heisenberg relations (4.1)(4.3) are stronger than the commu-
tation relations (4.15)(4.23), when imposed on the Lagrangian formalism as subsidiary
restrictions.
5. Types of Possible Commutation Relations

In a broad sense, by a commutation relation we shall understand any algebraic relation be-
tween the creation and annihilation operators imposed as subsidiary restriction on the La-
grangian formalism. In a narrow sense, the commutation relations are the equations (5.13),
with = 1, written below and satisfied by the bose creation and annihilation operators. As
anticommutation relations are known the equations (5.13), with = +1, written below and
satisfied by the fermi creation and annihilation operators. The last two types of relations are
often referred as the bilinear commutation relations [22]. Theoretically are possible also
5 The problem for the validity of assertions (i)(iii) or equations (4.24) in the general case of arbitrary fields
(Lagrangians) is not a subject of the present work.

trilinear commutation relations, an example being the paracommutation relations [22, 29]
represented below by equations (5.18) (or (5.20)).
Generally said, the commutation relations should be postulated. Alternatively, they
could be derived from (equivalent to them) different assumptions added to the Lagrangian
formalism. The purpose of this section is to be explored possible classes of commutation
relations, which follow from some natural restrictions on the Lagrangian formalism that are
consequences from the considerations in the previous sections. Special attention will be
paid on some consequences of the charge symmetric Lagrangians as the free fields possess
such a symmetry [1, 2, 5, 11].
As pointed in Sect 2, the Euler-Lagrange equations for the Lagrangians L , L and
L coincide and, in quantum field theory, the role of these equations is to be singled out
the independent degrees of freedom of the fields in the form of creation and annihilation
operators a
s (k) and as (k) (which are identical for L , L and L ). Further specialization
of these operators is provided by the commutation relations (in broad sense) which play a
role of field equations in this situation (with respect to the mentioned operators).
Before proceeding on, we would like to simplify our notation. As a spin variable, s
say, is always coupled with a 3-momentum one, k say, we shall use the letters l, m and n to
denote pairs like l = (s, k), m = (t, p) and n = (r, q). Equipped with this convention, we shall

write, e.g., a
l for as (k) and al for as (k). We set lm := st 3 (k p) and a summation

sign like l should be understood as s d3 k, where the range of the polarization variable
s will be clear from the context (see, e.g., (2.9)(2.12)).
5.1. Restrictions Related to the Momentum Operator

First of all, let us examine the consequences of the Heisenberg relation (4.1) involving the
momentum operator. Since in terms of creation and annihilation operators it reads (see the
previous chapters or, e.g., [1])

[a
s (k), P ] = k a
s (k) [a
s (k), P ] = k a
s (k) k0 = m2 c2 + k 2 , (5.1)
the field equations in terms of creation and annihilation operators for the Lagrangians (2.1),
(2.3) and (2.4) respectively are (see (5.1) and (2.9)):
2 j+10m (10 j )
q q =m2 c2 +q2 a
0
+
s (k), at (q) at (q)
(5.2a)
t=1
3
+ at (q) at+(q) (1 + )a
s (k)st (k q) d q
3
=0
2 j+10m (10 j )
q q =m2 c2 +q2 as (k), at + (q) at (q)

0 (5.2b)
t=1

+ at (q) at+(q) (1 + )as (k)st 3 (k q) d3 q =0
2 j+10m (10 j )
q q =m2 c2 +q2 a
0
+
s (k), at (q) at (q)
(5.3a)
t=1

+ at (q) at + (q)
(1 + )a
s (k)st (k q) d q = 0
3 3
2 j+10m (10 j )
q q =m2 c2 +q2 as (k), at+(q) at (q)

0 (5.3b)
t=1

+ at (q) at + (q)
(1 + )as (k)st 3 (k q) d3 q = 0
2 j+10m (10 j )
q q =m2 c2 +q2 a
0
+
s (k), [at (q), at (q)]
(5.4a)
t=1
3
+ [at+ (q), at (q)] (1 + )a
s (k)st (k q) d q
3
=0
2 j+10m (10 j )
q q =m2 c2 +q2 as (k), [at+ (q), at(q)]

0 (5.4b)
t=1

+ [at+ (q), at (q)] (1 + )as (k)st 3 (k q) d3 q = 0,
where j and are given via (2.7), the generalized commutation function [, ] is defined
by (3.14), and the polarization indices take the values
s,t = 1, . . ., 2 j + 1 0m (1 0 j )

1 for j = 0 or for j = 12 and m = 0
= 1, 2 for j = 12 and m = 0 or for j = 1 and m = 0 . (5.5)

1, 2, 3 for j = 1 and m = 0
The b versions of the equations (5.2)(5.4) are consequences of the a versions and
the equalities

(a
l ) = al

(a
l ) = al

(5.6)

[A, B] = [A , B ]

for [A, B] = [B, A] = 1. (5.7)
Exercise 5.1. Applying (5.2)(5.4) and the identity
[A, B C] = [A, B] C B [A,C] for = 1 (5.8)
for the choice = 1, prove via a direct calculation that
[ P , P ] = 0 [ Q , P ] = 0 [ S, P] = 0
(5.9)
[ L , P ] = i{ P P } [ M , P] = i{ P P },
where the operators P , Q , S , L , and M denote the momentum, charge, spin, or-
bital and total angular momentum operators, respectively, of the system considered and are
calculated from one and the same initial Lagrangian. This result confirms the supposition,
made in Sect. 4, that the assertion (i) before (4.24) holds for the fields investigated here.
Below we shall study only those solutions of (5.2)(5.4) for which the integrands in
them vanish, i.e. we shall replace the systems of integral equations (5.2)(5.4) with the
following systems of algebraic equations (see the above convention on the indices l and m
and do not sum over indices repeated on one and the same level):

+
al , am am + am a+
m (1 + )lm al = 0 (5.10a)

+
al , am am + am a+ m (1 + )lm al = 0 (5.10b)

+
al , am am + a +
m am (1 + )lm al = 0 (5.11a)

+
al , am am + a +
m am (1 + )lm al = 0 (5.11b)

+
al , [am , am ] + [a+
m , am ] 2(1 + )lm al = 0

(5.12a)

+
al , [am , am ] + [a+m , am ] 2(1 + )lm al = 0.

(5.12b)
It seems, these are the most general and sensible trilinear commutation relations one
may impose on the creation and annihilation operators.
First of all, we should mentioned that the standard bilinear commutation relations, viz.
(see [1, 2, 5, 11] and sections II.6, II.14, III.7, and IV.7)
[a
l , am ] = 0 [a
l , am ] = 0
[a
l , am ] = (1)
2 j+1
lm idF [a
l , am ] = (1)
2 j+1
lm idF
[a
l , am ] = 0 [a
l , am ] = 0
[a
l , am ] = (1)
2 j+1
lm idF [a
l , am ] = (1)
2 j+1
lm idF , (5.13)
provide a solution of any one of the equations (5.10)(5.12), in a sense that, due to (2.7)
and (5.8), with = , any set of operators satisfying (5.13) converts (5.10)(5.12) into
identities.
Exercise 5.2. Prove the last assertion.
Besides, this conclusion remains valid also if the normal ordering is taken into account,
i.e. if, in this particular case, the changes am a+ + +
m
am am and am am
am am
+
are made in (5.10)(5.12).

Now we shall demonstrate how the trilinear relations (5.12) lead to the paracommuta-
tion relations. Equations (5.12) can be split into different kinds of trilinear commutation
relations into infinitely many ways.
Exercise 5.3. Verify that the system of equations

+
al , [am , am ] (1 + )lm a
l =0 (5.14a)

+
al , [am , am ] (1 + )lm a
l =0 (5.14b)

+
al , [am , am ] (1 + )lm al = 0 (5.14c)

+
al , [am , am ] (1 + )lm al = 0 (5.14d)
provides a solution of (5.12). Besides, prove that these relations are incompatible with the
standard (anti)commutation relations (5.13) and, in this sense, are not suitable as subsidiary
restrictions on the Lagrangian formalism.
For our purposes is more important the result in

Exercise 5.4. Prove that the equations

+ +
al , [am , am ] + 2lm a+
l =0 (5.15a)

+ +
al , [am , am ] + 2lm a+
l =0 (5.15b)

+
al , [am , am ] 2lm a
l =0 (5.15c)

+
al , [am , am ] 2lm a
l =0 (5.15d)
and their Hermitian conjugate provide a solution of (5.12), which is compatible with (5.13),
i.e. if (5.13) hold, the equations (5.15) are converted into identities.
The idea of the paraquantization is in the following generalization of (5.15)

+ +
al , [am , an ] + 2ln a+
m=0 (5.16a)

+ +
al , [am , an ] + 2ln a+
m=0 (5.16b)

+
al , [am , an ] 2lm an =0 (5.16c)

+
al , [am , an ] 2lm an = 0 (5.16d)
which reduces to (5.15) for n = m.
Exercise 5.5. Verify that (5.16) is a generalization of (5.13) in a sense that any set of
operators satisfying (5.13) converts (5.16) into identities, the opposite being generally not
valid.1
Suppose that the field considered consists of a single sort of particles, e.g. electrons or
photons, created by bl := al and annihilated by bl := al . Then the equation Hermitian
conjugated to (5.15a) reads
[bl , [bm, bm] ] = 2lm bm . (5.17)
This is the main relation from which the paper [29] starts. The basic paracommutation
relations are [22, 29, 33, 36]:
[bl , [bm, bn ] ] = 2lm bn (5.18a)

[bl , [bm, bn ] ] = 0. (5.18b)
The first of them is a generalization (stronger version) of (5.17) by replacing the second
index m with an arbitrary one, say n, and the second one is added (by hands) in the
theory as an additional assumption. Obviously, (5.18) are a solution of (5.15) and therefore
of (5.12) in the considered case of a field consisting of only one sort of particles.
The equations (5.15) contain also the relativistic version of the paracommutation rela-
tions, when the existence of antiparticles must be respected [22, sec. 18.1]. Indeed, noticing
+
that the fields particles (resp. antiparticles) are created by bl := a+
l (resp. cl := al ) and
1 Other generalizations of (5.15) are also possible, but they do not agree with (5.13). Moreover, the reader
may wish to prove that any other (non-trivial) arrangement of the indices in (5.16) is incompatible with (5.13).
annihilated by bl := al (resp. cl := a
l ), from (5.15) and the Hermitian conjugate to them
equations, we get
[bl , [bm, bm] ] = 2lm bm [cl , [cm, cm ] ] = 2lm cm (5.19a)

[bl , [cm , cm] ] = 2lm bm [cl , [bm, bm] ] = 2lm cm . (5.19b)
Generalizing these equations in a way similar to the transition from (5.17) to (5.18), we
obtain the relativistic paracommutation relations as (cf. (5.16))
[bl , [bm, bn ] ] = 2lm bn [bl , [bm, bn ] ] = 0 (5.20a)

[cl , [cm, cn ] ] = 2lm cn [cl , [cm, cn ] ] = 0 (5.20b)
[bl , [cm, cn ] ] = 2ln bm [cl , [bm, bn ] ] = 2ln cm . (5.20c)
The equations (5.20a) (resp. (5.20b)) represent the paracommutation relations for the
fields particles (resp. antiparticles) as independent objects, while (5.20c) describe a pure
relativistic effect of some interaction (or its absents) between fields particles and an-
tiparticles and fixes the paracommutation relations involving the bl s and cl s, as pointed
in [22, p. 207] (where bl is denoted by al and cl by bl ). The relations (5.17) and (5.20)
for = +1 (resp. = 1) are referred as the parabose (resp. parafermi) commutation
relations [22]. This terminology is a natural one also with respect to the commutation
relations (5.16), which will be referred as the paracommutation relations too.
As first noted in [29], the bilinear commutation relations (5.13) provide a solution
of (5.20) (or (5.18) in the nonrelativistic case) but the latter equations admit also an in-
finite number of other solutions. Besides, by taking Hermitian conjugations of (some of)
the equations (5.18) or (5.20) and applying generalized Jacobi identities, like
[[A, B],C] + [[A,C]/ , B]/
2 [[B,C]/, A]1/ = 0 = 0
[A, [B,C], ] + [B, [C, A], ]
(5.21)
+ [C, [A, B], ] = 0 , , = 1
[[A, B],C] + [[B,C], A] + [[C, A], B] = 0 = 1
[[A, B], [C, D]] = [[A, B],C], D] + [[A, B] , D],C]1/ = 0,
one can obtain a number of other (para)commutation relations for which the reader is re-
ferred to [22, 29, 36].
Of course, the paracommutation relations (5.16), in particular (5.18) and (5.20) as their
stronger versions, do not give the general solution of the trilinear relations (5.12). For
instance, one may replace (5.12) with the equations

+ +
al , [am , an ] + [a+ +
m , an ] + 2(1 + )ln am = 0

(5.22a)

+
al , [am , an ] + [a+
m , an ] 2(1 + )lm an = 0.

(5.22b)
and their Hermitian conjugate, which in terms of the operators bl and cl introduced above
read
[bl , [bm, bn ] + [cm , cm ] ] = 2(1 + )lm bn (5.23a)

[cl , [bm, bn ] + [cm , cm ] ] = 2(1 + )lm cn , (5.23b)
and supplement these relations with equations like (5.18b). Obviously, equations
(5.16) convert (5.22) into identities and, consequently, the (standard) paracommutation re-
lations (5.20) provide a solution of (5.23). On the base of (5.23) or other similar equations
that can be obtained by generalizing the ones in (5.10)(5.12), further research on particular
classes of trilinear commutation relations can be done, but, however, this is not a subject of
the present chapter.
Let us now pay attention to the fact that equations (5.10), (5.11) and (5.12) are generally
different (regardless of existence of some connections between their solutions). The cause
for this being that the momentum operators for the Lagrangians L , L and L are gen-
erally different unless some additional restrictions are added to the Lagrangian formalism
(see Sect. 3). A necessary and sufficient condition for (5.10)(5.12) to be identical is
[a + +
l , [am , am ] [am , am ] ] = 0, (5.24)
which certainly is valid if the condition (3.9 ) on page 225, viz.

+
m , am ] [am , am ] = 0,
[a+ (5.25)
ensuring the uniqueness of the momentum operator, holds. If one adopts the standard bilin-
ear commutation relations (5.13), then (5.25), and hence (5.24), is identically valid, but in
the framework of, e.g., the paracommutation relations (5.16) (or (5.20) in other form) the
equations (5.25) should be postulated to ensure uniqueness of the momentum operator and
therefore of the field equations.
On the base of (5.10) or (5.11) one may invent other types of commutation relations,
which will not be investigated in this book because we shall be interested mainly in the
case when (5.10), (5.11) and (5.12) are identical (see (5.24)) or, more generally, when the
dynamical variables are unique in the sense pointed in Sect. 3.
5.2. Restrictions Related to the Charge Operator

The consequences of the Heisenberg relations (4.2), involving the charge operator for a
charged field, q = 0 (and hence = 0 see (2.7)), will be examined in this subsection. In
terms of creation and annihilation operators it is equivalent to (see [1] and/or chapters IIIV)
[a
s (k), Q ] = qas (k) [as (k), Q ] = qas (k), (5.26)
the values of the polarization indices being specified by (5.5). Substituting here (2.10), we
see that, for a charged field, the field equations for the Lagrangians L , L and L (see
Sect. 2) respectively are:
2 j+10m (10 j )
d3 p{[a + +
s (k), at (p) at (p) at (p) at (p)]
(5.27a)
t=1
a
s (k)st (k p)} =
3
0
2 j+10m (10 j )
d3 p{[a + +
(5.27b)
t=1
+as (k)st 3 (k p)} = 0
2 j+10m (10 j )
d3 p{[a + +
(5.28a)
t=1
+a
s (k)st (k p)} = 0
3
2 j+10m (10 j )
d3 p{[a + +
(5.28b)
t=1
as (k)st 3 (k p)} = 0
2 j+10m (10 j )
d3 p{[a + +
s (k), [at (p), at (p)] [at (p), at (p) ]
(5.29a)
t=1
2a
s (k)st (k p)} = 0
3
2 j+10m (10 j )
d3 p{[a + +
s (k), [at (p), at (p)] [at (p), at (p) ]
(5.29b)
t=1
+2as (k)st 3 (k p)} = 0.
Using (5.27)(5.29) and (5.8), with = = 1, or simply (5.26), one can easily verify
the validity of the equations
[ P , Q ] = 0 [ L , Q ] = 0
(5.30)
[ S , Q ] = 0 [ M , Q ] = 0,
where the operators P , Q , S , L and M are calculated from one and the same initial
Lagrangian according to (2.9)(2.12). This result confirms the validity of assertion (ii)
before (4.24) for the fields considered.
Following the above considerations, concerning the momentum operator, we shall now
replace the systems of integral equations (5.27)(5.29) with respectively the following
stronger systems of algebraic equations (by equating to zero the integrands in (5.27)
(5.29)):

+
al , am am am a+
m lm al = 0 (5.31a)

+
al , am am am a+ m + lm al = 0

(5.31b)

+
al , am am a +
m am + lm al = 0 (5.32a)

+
al , am am a +
m am lm al = 0 (5.32b)

+
al , [am , am ] [a+
m , am ] 2lm al = 0

(5.33a)

+
al , [am , am ] [a+m , am ] + 2lm al = 0.

(5.33b)
These trilinear commutation relations are similar to (5.10)(5.12) and, consequently,

can be treated in analogous way.
Exercise 5.6. By invoking the identity (5.8), prove that the standard bilinear commutation
relations (5.13) convert (5.31)(5.33) into identities.
Thus (5.13) are stronger version of (5.31)(5.33) and, in this sense, any type of com-
mutation relations, which provide a solution of (5.31)(5.33) and is compatible with (5.13),
is a suitable candidate for generalizing (5.13). To illustrate that idea, we shall proceed
with (5.33) in a way similar to the derivation of the paracommutation relations from (5.12).
Exercise 5.7. Verify that the equations (cf. (5.14) with = 0, as now q = 0)
[a +
l , [am , am ] ] + lm am = 0 (5.34a)
[a +
l , [am , am ] ] lm a
m =0 (5.34b)
and their Hermitian conjugate provide a solution of (5.33), but they do not agree with the
standard (anti)commutation relations (5.13) in a sense that (5.13) does not provide a solution
of (5.34).
A solution of (5.33) compatible with (5.13) is given by the equations (5.15), with = 0
as the field considered is charged one see (2.7). Therefore equations (5.16), with = 0,
also provide a compatible with (5.13) solution of (5.33), from where immediately follows
that the paracommutation relations (5.20), with = 0, convert (5.33) into identities. To
conclude, we can say that the paracommutation relations (5.20), in particular their special
case (5.13), ensure the simultaneous validity of the Heisenberg relations (4.1) and (4.2) for
free scalar, spinor and vector fields.
Similarly to (5.22), one may generalize (5.33) to

+ +
al , [am , an ] [a+ +
m , an ] 2ln am = 0

(5.35a)

+
al , [am , an ] [a+
m , an ] 2lm an = 0.

(5.35b)
Exercise 5.8. Prove that equations (5.35) agree with (5.13), (5.15), (5.16) and (5.20), but
generally do not agree with (5.22), with = 0, unless the equations (5.16), with = 0, hold.
More generally, we can assert that (5.33) and (5.12), with = 0, hold simultaneously if
and only if (5.15), with = 0, is fulfilled. From here, again, it follows that the paracommu-
tation relations ensure the simultaneous validity of (4.1) and (4.2).
Let us say now some words on the uniqueness problem for the Heisenberg equations
involving the charge operator. The systems of equations (5.31)(5.33) are identical iff

+
al , [am , am ] + [a+
m , am ] = 0, (5.36)
which, in particular, is satisfied if the condition
+
m , am ] + [am , am ] = 0,
[a+ (5.37)
ensuring the uniqueness of the charge operator (see (3.10 )), is valid. Evidently, equa-
tions (5.36) and (5.24) are compatible iff

+

al , [am , am ] = 0 a
l , [am , am ] = 0 (5.38)
which is a weaker form of (3.15) ensuring simultaneous uniqueness of the momentum and
charge operator.
5.3. Restrictions Related to the Angular Momentum Operator(s)

It is now turn to be investigated the restrictions on the creation and annihilation operators
that follow from the Heisenberg relations (4.3) concerning the angular momentum operator.
They can be obtained by inserting the equations (2.11) and (2.12) into (4.3). As pointed in
Sect. 4, the resulting equalities, however, depend not only on the particular Lagrangian
employed, but also on the geometric nature of the field considered; the last dependence
m ss m
being explicitly given via (4.25) and the polarization functions ss
(k) and l (k) (see
also (2.14)).
Consider the terms containing derivatives in (4.3),

Li
or
:= i x x i (x). (5.39)
x x
If i (k) denotes the Fourier image of i (x), i.e.

x
d4 ke i k
1
i (x) =
i (k), (5.40)
with being a normalization constant, then the Fourier image of (5.39) is

or
L i = i k k (k). (5.41)
k k i
Comparing this expression with equations (2.12), we see that the terms containing
derivatives in (2.12) should be responsible for the term (5.39) in (4.3).2 For this reason,
we shall suppose that the momentum operator M admits a representation
M = Mor + Msp (5.42)
such that the operators Mor and M satisfy the relations (4.4) and (4.5), respectively. Thus
sp

we shall replace (4.3) with the stronger system of equations (4.4)(4.5). Besides, we shall
admit that the explicit form of the operators Mor and M sp are given via (4.13) and (4.12)

for the fields investigated in the present work.
Let us consider at first the orbital Heisenberg relations (4.4), which is independent
of the particular geometrical nature of the fields studied. Substituting (4.13) and (5.40)
into (4.4), using that (k), with k2 = m2 c2 , is a linear combination of as (k) with clas-
i
sical, not operator-valued, functions of k as coefficients (see [1] and/or chapters IIIV) and
introducing for brevity the operator

(k) := k
k , (5.43)
k k
we arrive to the following integro-differential systems of equations:
2 j+10m (10 j )

d3 p ( (p) + (q))([a +
s (k), at (p) at (q)
t=1
2 The terms proportional to the momentum operator in (2.12) disappear if the creation and annihilation
operators (1.1) in Heisenberg picture are employed (see also chapters IIIV).

at (p) at+ (q)] ) q=p p =m2 c2 +p2 = 2(1 + ) (k)(a
s (k)) (5.44a)
0
2 j+10m (10 j )

d3 p ( (p) + (q))([a +
t=1

at (p) at+ (q)] ) q=p p =m2 c2 +p2 = 2(1 + )(k)(a
s (k)) (5.44b)
0
2 j+10m (10 j )

d3 p ( (p) + (q))([a +
t=1

at (p) at + (q)] ) q=p p =m2 c2 +p2 = 2(1 + ) (k)(a
s (k)) (5.45a)
0
2 j+10m (10 j )

d3 p ( (p) + (q))([a +
t=1

at (p) at + (q)] ) q=p p =m2 c2 +p2 = 2(1 + )(k)(a
s (k)) (5.45b)
0
2 j+10m (10 j )

d3 p ( (p) + (q))([a +
s (k), [at (p), at (q)]
t=1

+ [at+ (p), at (q)] ] ) q=p p =m2 c2 +p2 = 4(1 + ) (k)(a
s (k)) (5.46a)
0
2 j+10m (10 j )

d3 p ( (p) + (q))([a +
s (k), [at (p), at (q)]
t=1

+ [at+ (p), at (q)]] ) q=p p =m2 c2 +p2 = 4(1 + ) (k)(a
s (k)), (5.46b)
0

where k0 = m2 c2 + k 2 is set after the differentiations are performed (see (5.43)). Fol-
lowing the procedure of the previous considerations, we replace the integro-differential
equations (5.44)(5.46) with the following differential ones:

( (m) + (n))([a
l , am an
+

am a+
n ] ) n=m = 2(1 + )lm (l)(al ) (5.47a)

( (m) + (n))([a +
l , am an

am a+
n ] ) n=m = 2(1 + )lm (l)(al ) (5.47b)

( (m) + (n))([a +
l , am an

a +
m an ] ) n=m = 2(1 + )lm (l)(al ) (5.48a)

( (m) + (n))([a +
l , am an

a +
m an ] ) n=m = 2(1 + )lm (l)(al ) (5.48b)


( (m) + (n))([a +
l , [am , an ]

+ [a+
m , an ] ] ) n=m = 4(1 + )lm (l)(al ) (5.49a)

( (m) + (n))([a +
l , [am , an ]

+ [a+
m , an ] ] ) n=m = 4(1 + )lm (l)(al ), (5.49b)

where we have set (cf. (5.43))

(l) := (k) = k
k if l = (s, k) (5.50)
k k

and k0 = m2 c2 + k2 is set after the differentiations are performed.
Remark 5.1. Instead of (5.47)(5.49) one can write similar equations in which the operator
(m) or + (n) is deleted and the factor + 12 or 12 , respectively, is added on their
right hand sides. These manipulations correspond to an integration by parts of some of the
terms in (5.44)(5.46).
The main difference of the obtained trilinear relations with respect to the previous ones
considered above is that they are partial differential equations of first order.
The relations (5.49) agree with the equations (5.16) in a sense that if (5.16) hold,
then (5.49) become identically valid. Indeed, since

( (m) + (n))(a
m ln ) n=m = 2lm (m)(am )

(5.51)
( (m) + (n))(alm )
n = +2lm (m)(a),
n=m m
d f (x)
due to (5.50), (5.43) and the equality d(x)
dx f (x) = (x) dx for a C function f , the ap-
1

plication of the operator ( (m) + (n)) to (5.16) and subsequent setting n = m en-
tails (5.49). In particular, this means that the paracommutation relations (5.20) and, more-
over, the standard (anti)commutation relations (5.13) convert (5.49) into identities. There-
fore the orbital Heisenberg relations (4.4) hold for scalar, spinor and vector fields satisfy-
ing the bilinear or para commutation relations.
It should be noted, the paracommutation relations are not the only trilinear commutation
relations that are solutions of (5.49).
Exercise 5.9. Prove that the trilinear relations

+ +

al , [am , an ] = a+ + +
l , [am , an ] = (1 + )ln am (5.52a)

+
+
al , [am , an ] = al , [am , an ] = +(1 + )lm a+ n, (5.52b)
which reduce to (5.14) for n = m, do not agree with (5.13), but convert (5.49) into identities
(see (5.51)).
Exercise 5.10. Prove that the equations (5.22) are compatible with the paracommutation
relations and, as a result of (5.51), convert (5.49) into identities.
Prima facie one may suppose that any solution of (5.12) provides a solution of (5.49),
but this is not the general case. A counterexample is provided by
Exercise 5.11. Prove that the commutation relations

+
al , [am , an ] + [a+
m , an ] 2(1 + )ln am = 0, (5.53)
which reduce to (5.12) for n = m, satisfy (5.49) with a+

l for al , and do not satisfy (5.49)
with a
l for al (see (5.51) and cf. (5.22)).
From (4.13) follows that the operator M or

is independent of the Lagrangian L , L or
L one starts off if and only if (see (3.11))

( (m) + (n)) [a+ +
m , an ] [am , an ] n=m
= 0. (5.54)
This condition ensures the coincidence of the systems of equations (5.47), (5.48)
and (5.49) too. However, the following necessary and sufficient condition for the coin-
cidence of these systems is expressed by the weaker equations

( (m) + (n)) a + +
l , [am , an ] [am , an ] n=m
= 0. (5.55)
It is now turn to be considered the spin Heisenberg relations (4.5).

Recall, the field operators i for the fields considered here admit a representation (see
chapters IIIV) t,+
i = d3 p vi (p)at+(p) + vi (p)at(p) ,
t,
(5.56)
t
where is a normalization constant and vi (p) are classical, not operator-valued, com-
t,
plex or real functions which are linearly independent. The particular definition of vt, i (p)
depends on the geometrical nature of i and can be found in the previous chapters (see
t,
also [1]), where the reader can find also a number of relations satisfied by vi (p). Here
t, t,+ t,
we shall mention only that vi (p) = 1 for a scalar field and vi (p) = vi (p) =: vti (p) =
(vti (p)) for a vector field.
ss , ss ,
The explicit form of the polarization functions (k) and l (k) (see Sect. 2, in
t,
particular (2.14)) through vi (k) are (see chapters IIIV):
(1) j s ,
ss ,
(k) =
j + j0 i,i
(vs, i
i (k)) Ii vi (k)

(5.57)
(1) j s ,
ss ,
2 j + j0
s,
l (k) = (v (k)) k k v (k),
i
i
k k i
ss , ss ,
with an exception that 0a (k) = a0 (k) = 0, a = 1, 2, 3, for a spinor field, j = 12
(see chapter III). Evidently, the equations (2.14) follow from the mentioned facts (see
also (4.25)).
Substituting (5.56) and (4.12) into (4.5), we obtain the following systems of integral
equations (corresponding respectively to the Lagrangians L , L and L ):
ss ,
(1) j+1 j
1 + s,s d3 k
t,
d3 pvi (p) ( (k)
,t

ss , ss ,+
ss ,+
+ l (k))[at(p), a+ +
s (k) as (k)] +( (k) + l (k))[at (p), as (k) as (k)]

=

vi (p)at(p) (5.58)
t,
d3 pIi
i
i t
ss,+
(1) j+1 j

1 + s,s d3 k
t,
d3 pvi (p) ( (k)
,t

ss ,+ ss ,
ss ,
+ l (k))[at(p), a+ +
s (k) as (k)] +( (k) + l (k))[at (p), as (k) as (k)]

=

d3 pIi
i
vt,
i (p)at (p) (5.59)
i t
ss ,
(1) j j+1

2(1 + ) s,s
d3 k d3 pvt,
i (p) ( (k)
,t
ss ,

+ l (k)) at(p), [a+
s (k), as (k)]
ss ,+ ss ,+

+ ( (k) + l (k)) at (p), [a +
s (k), as (k)]

=

d3 pIi
i
vt,
i (p)at (p). (5.60)
i t
For the difference of all previously considered systems of integral equations, like (5.2)
(5.4), (5.27)(5.29) and (5.44)(5.46), the systems (5.58)(5.60) cannot be replaced by
ones consisting of algebraic (or differential) equations. The cause for this state of affairs
is that in (5.58)(5.60) enter polarization modes with arbitrary s and s and, generally, one
cannot diagonalize the integrand(s) with respect to s and s ; moreover, for a vector field,
the modes with s = s are not presented at all (see (2.14)). That is why no commutation
relations can be extracted from (5.58)(5.60) unless further assumptions are made. Without
going into details, below we shall sketch the proof of the assertion that the commutation
relations (5.16) convert (5.60) into identities for massive spinor and vector fields.3 In
particular, this entails that the paracommutation and the bilinear commutation relations
provide solutions of (5.60).
Let (5.16) holds. Combining it with (5.60), we see that the latter splits into the equations
st,
(1) j j
1+
d3 pvi (p) ( (p) + l (p)) + ( (p) + l (p)) a+
t,+ st, ts,+ ts,+
s (p),
s,t

= i
Ii d3 pvs,+ +
i (p)as (p) (5.61a)
i s
ts,
(1) j+1 j
1+ i (p) ( (p) + l (p)) + ( (p) + l (p)) as (p),
d3 pvt, ts, st,+ st,+
s,t

= Ii

i
d3 pvs,
i (p)as (p). (5.61b)
i s
t,
3 The equations (5.58)(5.60) are identities for scalar fields as for them I = 0 and vi (k) = 1, which
reflects the absents of spin for these fields.
Inserting here (5.57), we see that one needs the explicit definition of vs,
i (k) and formulae
for sums like ii (k) := s vi (k)(vi (k)) , which are specific for any particular field and
s, s,
can be found in chapters IIIV. In this way, applying (4.25), (2.7) and the mentioned
results from the previous chapters, one can check the validity of (5.61) for massive fields
in a way similar to the proof of (4.3) in chapters IIIV for scalar, spinor and vector fields,
respectively.
We shall end the present subsection with the remark that the equations (3.17) and (3.18),
which together with (3.15) ensure the uniqueness of the spin and orbital operators, are
sufficient conditions for the coincidence of the equations (5.58), (5.59) and (5.60).
6. Inferences
To begin with, let us summarize the major conclusions from Sect. 5. Each of the Heisenberg
equations (4.1)(4.3), the equations (4.3) being split into (4.4) and (4.5), induces in a nat-
ural way some relations that the creation and annihilation operators should satisfy. These
relations can be chosen as algebraic trilinear ones in a case of (4.1) and (4.2) (see (5.10)
(5.12) and (5.31)(5.33), respectively). But for (4.4) and (4.5) they need not to be algebraic
and are differential ones in the case of (4.4) (see (5.47)(5.49)) and integral equations in the
case of (4.5) (see (5.58)(5.60)). It was pointed that the cited relations depend on the initial
Lagrangian from which the theory is derived, unless some explicitly written conditions hold
(see (5.24), (5.37) and (5.55)); in particular, these conditions are true if the equations (3.9)
(3.13), ensuring the uniqueness of the corresponding dynamical operators, are valid. Since
the charge symmetric Lagrangians (2.4) seem to be the ones that best describe free fields,
the arising from them (commutation) relations (5.12), (5.33), (5.49) and (5.60) were studied
in more details. It was proved that the trilinear commutation relations (5.16) convert them
into identities, as a result of which the same property possess the paracommutation rela-
tions (5.20) and, in particular, the bilinear commutation relations (5.13). Examples of trilin-
ear commutation relations, which are neither ordinary nor para ones, were presented; some
of them, like (5.14), (5.34) and (5.52), do not agree with (5.13) and other ones, like (5.16),
(5.22) and (5.35), generalize (5.20) and hence are compatible with (5.13). At last, it was
demonstrated that the commutators between the dynamical variables (see (4.15)(4.23)) are
uniquely defined if a Heisenberg relation for one of the operators entering in it is postulated.
The chief aim of the present section is to be explored the problem whether all of the
reasonable conditions, mentioned in the previous sections and that can be imposed on the
creation and annihilation operators, can hold or not hold simultaneously. This problem is
suggested by the strong evidences that the relations (4.1)(4.3) and (4.15)(4.23), with a
possible exception of (4.3) (more precisely, of (4.5)) in the massless case, should be valid
in a realistic quantum field theory [1, 2, 5, 911]. Besides, to the arguments in loc. cit., we
shall add the requirement for uniqueness of the dynamical variables (see Sect. 3).
As it was shown in Sect. 5, the relations (4.1), (4.2), (4.4) and (4.5) are compatible if
one starts from a charge symmetric Lagrangian (see (2.4)), which best describes a free field
theory; in particular, the commutation relations (5.16) (and hence (5.20) and (5.13)) ensure
their simultaneous validity.1 For that reason, we shall investigate below only commutation
1 The special case(s) when (4.5) may not hold for a massless field will not be considered below.
relations for which (4.1), (4.2), (4.4) and (4.5) hold. It will be assumed that they should
be such that the equations (5.10)(5.12), (5.31)(5.33), (5.47)(5.49) and (5.58)(5.60),
respectively, hold.
Consider now the problem for the uniqueness of the dynamical variables and its consis-
tency with the commutation relations just mentioned for a charged field. It will be assumed
that this uniqueness is ensured via the equations (3.9)(3.11).
The equation (3.15), viz.
[am , a
m ] = 0, (6.1)
is a necessary and sufficient conditions for the uniqueness of the momentum and charge
operators (see Sect. 3 and the notation introduced at the beginning of Sect. 5). Before
commenting on this relation, we would like to derive some consequences of it. Applying
consequently (5.8) for = , (6.1) and the identity
[A, B C]+ = [A, B] C B [A,C] = 1 (6.2)
for = +, , we, in view of (6.1), obtain
[a+ + + + + +
m , [am , am ] ] = [am , [am , am ] ]+ = (1 )[am , am ] am

(6.3)
[a + +
m , [am , am ] ] = [am , [am , am ] ]+ = (1 )[am , am ] am .
+
Forming the sum and difference of (5.12a), for = 0, and (5.33a), we see that the system
of equations they form is equivalent to
[a+ +
l , [am , am ] ] = 0 [a +
l , [am , am ] ] = 0 (6.4a)
[a+ +
l , [am , am ] ] + 2lm a+
l =0 [a +
l , [am , am ] ] 2lm a
l = 0. (6.4b)
Combining (6.4b), for l = m, with (6.3), we get

+ + + +
m , am ] am + 2am = 0 (1 )[am , am ] am 2am = 0.
(1 )[a (6.5)
Obviously, these equations reduce to
a
m=0 (6.6)
for bose fields as for them = +1 (see (2.7)). Since the operators (6.6) describe a com-
pletely unobservable field, or, more precisely, an absence of a field at all, the obtained result
means that the theory considered cannot describe any really existing physical field with
spin j = 0, 1. Such a conclusion should be regarded as a contradiction in the theory. For
fermi fields, j = 12 and = 1, the equations (6.5) have solutions different from (6.6) iff am
are degenerate operators, i.e. with no inverse ones, in which case (6.4a) is a consequence
of (6.5) and (6.1) (see (5.8) and (6.3) too).
The source of the above contradiction is in the equation (6.1), which does not agree with
the bilinear commutation relations (5.13) and contradicts to the existing correlation between
creation and annihilation of particles with identical characteristics (m = (t, p) in our case)
as (6.1) can be interpreted physically as mutual independence of the acts of creation and
annihilation of such particles [1, 10.1].
At this point, there are two ways for repairing of the theory. On one hand, one can
forget about the uniqueness of the dynamical variables (in a sense of Sect. 3), after which
the formalism can be developed by choosing, e.g., the charge symmetric Lagrangians (2.4)
and following the usual Lagrangian formalism; in fact, this is the way the parafield theory is
build [22,29]. On another hand, one may try to change something at the ground of the theory
in such a way that the uniqueness of the dynamical variables to be ensured automatically.
We shall follow the second method. As a guiding idea, we shall have in mind that the
bilinear commutation relations (5.13) and the related to them normal ordering procedure
provide a base for the present-day quantum field theory, which describes sufficiently well
the discovered elementary particles/fields. On this background, an extensive exploration of
commutation relations which are incompatible with (5.13) is justified only if there appear
some evidences for fields/particles that can be described via them. In that connection it
should be recalled [22, 33], it seems that all known particles/fields are described via (5.13)
and no one of them is a para particle/field.
Using the notation introduced at the beginning of Sect. 3, we shall look for a linear
mapping (operator) E on the operator space over the systems Hilbert space F of states
such that
E ( D ) = E ( D ). (6.7)
As it was shown in Sect. 3, an example of an operator E is provided by the normal
ordering operator N . Therefore an operator satisfying (6.7) always exists. To any such
operator E there corresponds a set of dynamical variables defined via
D = E ( D ). (6.8)
Let us examine the properties of the mapping E that it should possess due to the re-
quirement (6.7).
First of all, as the operators of the dynamical variables should be Hermitian, we shall
require

E ( B ) = E ( B ) (6.9)
for any operator B , which entails
D = D, (6.10)
due to (2.9)(2.12) and (6.8).

As in Sect. 3, we shall replace the so-arising integral equations with corresponding
algebraic ones. Thus the equations (3.5)(3.20) remain valid if the operator E is applied to
their left hand sides.
Consider the general case of a charged field, q = 0. So, the analogue of (3.15) reads

E [am , a
m ] = 0, (6.11)
which equation ensures the uniqueness of the momentum and charge operators. Respec-
tively, the condition (3.11) transforms into

( (m) + (n)) E ([a+ +
m , an ] ) E ([am , an ] ) n=m
= 0, (6.12)
which, by means of (6.11) can be rewritten as (cf. (3.16))

(n) E ([a+ +
m , an ] ) E ([am , an ] )
= 0. (6.13)
n=m
At the end, equations (3.17) and (3.18) now should be written as

ss , ss ,+
(k) E [as +(k), a
s (k)] + (k) E [a s (k), a +
s (k)] =0 (6.14)
s,s
ss , ss ,+
l (k) E [as + (k), a +
s (k)] + l (k) E [as (k), as (k)] = 0. (6.15)
s,s
These equations can be satisfied if we generalize (6.11) to (cf. (3.20))

E [as (k), a
s (k)] = 0 (6.16)
for any s and s . At last, the following stronger version of (6.16)

E [am , a
n ] = 0, (6.17)
for any m = (t, p) and n = (r, q), ensures the validity of (6.14) and (6.15) and thus of the
uniqueness of all dynamical variables.
It is time now to call attention to the possible commutation relations. The replace-
ment D , D , D
D := E ( D ) = E ( D ) = E ( D ) results in corresponding changes
in the whole of the material of Sect. 5. In particular, the systems of commutation rela-
tions (5.10)(5.12), (5.31)(5.33), (5.47)(5.49) and (5.58)(5.60) should be replaced re-
spectively with:2

+

al , E (a+
m am ) + E (am am ) (1 + )lm al = 0

(6.18)

+

al , E (a+
m am ) E (am am ) lm al = 0

(6.19)

( (m) + (n))([a, E (a+ +
m an ) E (am an )] )

l n=m
= 2(1 + )lm (l)(a
l ) (6.20)
ss ,
(1) j+1 j
1 + s,s d k 3 t,
d3 pvi (p) ( (k)
,t
ss ,
+ l (k))[at(p), E (a+
s (k) as (k))]
ss ,+ ss ,+
+ ( (k) + l (k))[at(p), E (a +
s (k) as (k))]

=

vi (p)at(p).
t,
d3 pIi
i
(6.21)
i t
Due to the uniqueness conditions (6.11)(6.14), one can rewrite the terms E (a
m am )
in (6.18)(6.21) in a number of equivalent ways; e.g. (see (6.11))
1
E (a
m am ) = E (am am ) = E ([a
m , am ] ). (6.22)
2
2 To save some space, we do not write the Hermitian conjugate of the below-written equations.
Consider the general case of a charged field, q = 0 (and hence = 0). The system of
equations (6.18)(6.19) is then equivalent to

al , E (a
m am ) = 0 (6.23a)

+ +
+
al , E (a
m am ) + lm al = 0 (6.23b)

al , E (a+
m am ) lm al = 0. (6.23c)
These (commutation) relations ensure the simultaneous fulfillment of the Heisenberg

relations (4.1) and (4.2) involving the momentum and charge operators, respectively. To en-
sure also the validity of (6.20), with = 0, and, consequently, of (4.4), we generalize (6.23)
to

al , E (a
m an ) = 0 (6.24a)

+ +
+
al , E (a
m an ) + lm an = 0 (6.24b)

al , E (am an ) lm an = 0,
+
(6.24c)
for any l = (s, k), m = (t, p) and n = (t, q) (see also (5.51)). In the way pointed in Sect. 5,
one can verify that (6.24) for any l = (s, k), m = (t, p) and n = (r, p) entails (6.21) and
hence (4.5). At last, to ensure the validity of all of the mentioned conditions and a suitable
transition to a case of Hermitian field, for which q = 0 and = 1 (see (2.7)), we general-
ize (6.24) to

+
+
al , E (a+
m an ) + ln am = 0 (6.25a)

+

al , E (a
m an ) ln am = 0 (6.25b)

+ +
+
al , E (a
m an ) + lm an = 0, (6.25c)

al , E (am an ) lm an = 0
+
(6.25d)
where l, m and n are arbitrary. As a result of (6.17), which we assume to hold, and
al = a
l (see (2.7)), the equations (6.25a) and (6.25c) (resp. (6.25b) and (6.25d)) be-
come identical when = 1 (and hence al = a l ); for = 0 the system (6.25) reduces
to (6.24). Recalling that = (1)2 j (see (2.7)), we can rewrite (6.25) in a more compact
form as

al , E (a
m an ) + (1)
2 j+1
ln a
m =0 (6.26a)

al , E (am an ) (1)
2 j+1
lm an = 0. (6.26b)
Since the last equation is equivalent to (see (6.17)) and use that = (1)2 j )

al , E (a
m an ) + (1)
2 j+1
ln a
m = 0, (6.26b )
it is evident that the equations (6.26a) and (6.26b) coincide for a neutral field.
Let us draw the main moral from the above considerations: the equations (6.17) are
sufficient conditions for the uniqueness of the dynamical variables, while (6.26) are such
conditions for the validity of the Heisenberg relations (4.1)(4.5), in which the dynamical
variables are redefined according to (6.8). So, any set of operators a

l and E , which are
simultaneous solutions of (6.17) and (6.26), ensure uniqueness of the dynamical variables
and at the same time the validity of the Heisenberg relations.
Consider the uniqueness problem for the solutions of the system of equations consisting
of (6.17) and (6.26).
Exercise 6.1. Writing (6.17) as
1
E (a
m an ) = E (an am ) = E ([a
m , an ] ), (6.27)
2
which reduces to (6.22) for n = m, and using = (1)2 j (see (2.7)), verify that (6.26) is
equivalent to

+
al , E ([a+ +
m , an ] ) + 2ln am = 0 (6.28a)

+
+
al , E ([a+
m , an ] ) + 2ln am = 0 (6.28b)

+

al , E ([am , an ] ) 2lm an = 0 (6.28c)

al , E ([a+
m , an ] ) 2lm an = 0. (6.28d)
The similarity between this system of equations and (5.16) is more than evident: (6.28)
can be obtained from (5.16) by replacing [, ] with E ([, ]).
As it was said earlier, the bilinear commutation relations (5.13) and the identification of
E with the normal ordering operator N ,
E= N, (6.29)
convert (6.27)(6.28) into identities.
Exercise 6.2. By invoking (5.8), for = , check the last assertion via a direct calculation
(see also (3.23)).
However, this is not the only possible solution of (6.27)(6.28). For example, if, in the
particular case, one defines an anti-normal ordering operator A as a linear mapping such
that
A (a+
m an ) := an am
+
A (am+ a +
n ) := an am
(6.30)
A (a + +
m an ) := am an A (a +
m an ) := am an ,
+
then the bilinear commutation relations (5.13) and the setting E = A provide a solution
of (6.27)(6.28); to prove this, apply (5.8) for = . Evidently, a linear combination of
N and A , together with (5.13), also provides a solution of (6.27)(6.28).3 Other solution
of the same system of equations is given by E = idF and operators a l satisfying (5.16), in
3 If we admit a
l to satisfy the anomalous bilinear commutation relations (7.27) (see below), i.e. (5.13)
with for and (1)2 j for (1)2 j+1, then E = N , A also provides a solution of (6.27)(6.28). However,
as it was demonstrated in chapters IIIV, the anomalous commutation relations are rejected if one works with
the charge symmetric Lagrangians (2.4).
particular the paracommutation relations (5.20), and am a, ,

n = an am . The problem

for the general solution of (6.27)(6.28) with respect to E and al is open at present.
Let us introduce the particle and antiparticle number operators respectively by
(see (6.27), (6.9) and (2.16))
1
Nl := E [a+ +
l , al ] = E (al al ) = ( Nl ) =: Nl

2 (6.31)
1

Nl := E [al + , a +
l ] = E (al al ) = ( Nl ) =: Nl .

2
As a result of the commutation relations (6.28), with n = m, they satisfy the equations4
[ Nl , a+ +
m ] = lm al (6.32a)
[ Nl , a+ +

m ] = lm al (6.32b)
[ Nl , a+
m ] = lm al + (6.32c)
[ Nl , a+ = lm al +.

m ] (6.32d)
Combining (2.9)(2.12) and (4.11)(4.13) with (6.8), (6.27) and (6.31), we get the fol-
lowing expressions for the operators of the (redefined) dynamical variables:
1
1+
P = k | 2 2 2 ( Nl + Nl )

l = (s, k) (6.33)
k0 = m c +k
l
Q = q ( Nl + Nl ) (6.34)
l
(1) j1/2 j
S = Nnm + mn, Nmn)}m=(s,k)

{mn,+
(6.35)
1+ m,n n=(s ,k)
(1) j1/2 j
L = x0 P x0 P +
1+ mn,+
{l Nnm + l
mn,
Nmn)}m=(s,k)
m,n n=(s ,k)
i
(l) + (m) ( Nl + Nl ) m=l=(s,k)

+ (6.36)
2(1 + ) l
(1) j1/2
j
Msp =
1+ {(mn,+
+ l ) Nnm
mn,+
m,n

+ ( + l ) Nmn )}m=(s,k)
mn, mn,
(6.37)
n=(s ,k)
i
Mor = (l) + (m) ( Nl + Nl ) m=l=(s,k) .

(6.38)
2(1 + ) l
Here (l) is defined via (5.50), we have set
ss , ss ,
:= (k) l for m = (s, k) and n = (s , k),
mn, mn,
:= l (k) (6.39)
4 The equations (6.32a) and (6.32b) correspond to (6.28a) and (6.28b), respectively, and (6.32c) and (6.32d)
correspond to the Hermitian conjugate to (6.28c) and (6.28d), respectively.
and (see (6.27))
1
Nlm := E [a+ +
l , am ] = E (al am ) = ( Nml ) =: Nml

2 (6.40)
1

Nlm := E [a+
l , am ] = E (al am ) = ( Nml ) =: Nml
+
2
are respectively the particle and antiparticle transition operators (cf. [36, sec. 1] in a case
of parafields). Obviously, we have
Nl = Nll Nl = Nll .

(6.41)
The choice (6.29), evidently, reduces (6.33)(6.36) to (3.24), (3.25), (3.28) and (3.29),
respectively.
In terms of the operators (6.38), the commutation relations (6.28) can equivalently be
rewritten as (see also (6.9))
[ Nlm , a+ +
n ] = mn al (6.42a)
[ Nlm , a+ +

n ] = mn al (6.42b)
[ Nlm , a+
n ] = mn al + (6.42c)
[ Nlm , a+ = mn al + .

n ] (6.42d)
If m = l, these relations reduce to (6.32), due to (6.39).

We shall end this section with the remark that the conditions for the uniqueness of the
dynamical variables and the validity of the Heisenberg relations are quite general and are
not enough for fixing some commutation relations regardless of a number of additional
assumptions made to reduce these conditions to the system of equations (6.27)(6.28).
7. State Vectors, Vacuum and Mean Values

Until now we have looked on the commutation relations only from pure mathematical view-
point. In this way, making a number of assumptions, we arrived to the system (6.27)
(6.28) of commutation relations. Further specialization of this system is, however, almost
impossible without making contact with physics. For the purpose, we have to recall [1,
2, 5, 11] that the physically measurable quantities are the mean (expectation) values of the
dynamical variables (in some state) and the transition amplitudes between different states.
To make some conclusions from these basic assumption of the quantum theory, we must
rigorously said how the states are described as vectors in systems Hilbert space F of states,
on which all operators considered act.
For the purpose, we shall need the notion of the vacuum or, more precisely, the as-
sumption of the existence of unique vacuum state (vector) (known also as the no-particle
condition). Before defining rigorously this state, which will be denoted by X0 , we shall
heuristically analyze the properties it should possess.
First of all, the vacuum state vector X0 should represent a state of the field without any
particles. From here two conclusions may be drawn: (i) as a field is thought as a collection
of particles and a missing particle should have vanishing dynamical variables, those of
the vacuum should vanish too (or, more generally, to be finite constants, which can be set
equal to zero by rescaling some theorys parameters) and (ii) since the operators a l and
al are interpreted as ones that annihilate a particle characterize by l = (s, k) and charge
q or +q, respectively, and one cannot destroy an absent particle, these operators should
transform the vacuum into the zero vector, which may be interpreted as a complete absents
of the field. Thus, we can expect that
D ( X0 ) = 0 (7.1a)
a
l ( X0 ) =0 al ( X0 ) = 0. (7.1b)
+
Further, as the operators a+
l and al are interpreted as ones creating a particle character-
+
ize by l = (s, k) and charge q or +q, respectively, state vectors like a+
l ( X0 ) and al ( X0 )
should correspond to 1-particle states. Of course, a necessary condition for this is
X0 = 0, (7.2)
due to which the vacuum can be normalize to unit,
X0 | X0 = 1, (7.3)
where | : F F C is the Hermitian scalar (inner) product of F . More generally, if

+
M (a+
l1 , al2 , . . .) is a monomial only in i N creation operators, the vector
+
l1 l2 ... := M (a+
l1 , al2 , . . .)( X0 ) (7.4)
may be expected to describe an i-particle state (with i1 particles and i2 antiparticles, i1 +i2 =
+
i, where i1 and i2 are the number of operators a+ + +
l and al , respectively, in M (al1 , al2 , . . .)).
Moreover, as a free field is intuitively thought as a collection of particles and antiparticles,
it is natural to suppose that the vectors (7.4) form a basis in the Hilbert space F . But the
validity of this assumption depends on the accepted commutation relations; for its proof,
when the paracommutation relations are adopted, see the proof of [22, p. 26, theorem I-1].
Accepting the last assumption and recalling that the transition amplitude between two
states is represented via the scalar product of the corresponding to them state vectors, it
is clear that for the calculation of such an amplitude is needed an effective procedure for
calculation of scalar products of the form
l1 l2 ... |m1 m2 ... := X0 |( M (a+ + +

l1 , al2 , . . .)) M (am1 , am2 , . . .) X0 ,
+
(7.5)
with M and M being monomials only in the creation operators. Similarly, for compu-
tation of the mean value of some dynamical operator D in a certain state, one should be
equipped with a method for calculation of scalar products like
+
l1 l2 ... | D m1 m2 ... := X0 |( M (a+ +
l1 , al2 , . . .)) D M (am1 , am2 , . . .) X0 .
+
(7.6)
Supposing, for the moment, the vacuum to be defined via (7.1), let us analyze (7.1)
(7.6). Besides, the validity of (6.27)(6.28) will be assumed.
From the expressions (6.8) and (2.9)(2.12) for the dynamical variables, it is clear that
the condition (7.1a) can be satisfied if

E (a
m an )( X0 ) = 0, (7.7)
which, in view of (6.27), is equivalent to any one of the equations
E (a
m an )( X0 ) = 0 (7.8a)
E ([a
m , an ] )( X0 ) = 0. (7.8b)
Equation (7.7) is quite natural as it expresses the vanishment of all modes of the vac-
uum corresponding to different polarizations, 4-momentum and charge. It will be accepted
hereafter.
By means of (7.8) and the commutation relations (6.28) in the form (6.42), in particu-
lar (6.32), one can explicitly calculate the action of any one of the operators (6.33)(6.38)
on the vectors (7.4): for the purpose one should simply to commute the operators Nlm (or
Nl = Nll ) with the creation operators in (7.4) according to (6.42) (resp. (6.32)) until they
act on the vacuum and, hence, giving zero, as a result of (7.8) and (6.42) (resp. (6.32)). In
particular, we have the equations (k0 = m2 c2 + k 2 ):

P a+ +
l ( X0 ) = k al ( X0 ) P al + ( X0) = k al + ( X0) l = (s, k) (7.9)
+ +
+ +
Q a ( X0) = qal ( X0) Q al ( X0 ) = +qal ( X0) (7.10)
+ l
S a ( X0 )
l l=(s,k)
(1) j1/2 j ml,

=
1+ {lm,+ +
+ } m=(t,k) am |m=(t,k) ( X0 )
t
+ (7.11)
S a ( X0 )
l l=(s,k)
(1) j1/2 j
=
1+ ml, +
+ } m=(t,k) am |m=(t,k) ( X0 )
{lm,+
t

L a+ + +
l l=(s,k) ( X0 ) = (x0 k x0 k )(al )( X0 ) i (l)(al ) ( X0 )
(1) j1/2 j ml,

{l + l }m=(t,k) a+ m |m=(t,k) ( X0 )
lm,+
+
1+ t
(7.12)
L a +
l ( X0 ) = (x0 k x0 k )(a+ )( X0 ) i (l)(a+) ( X0 )
l=(s,k) l l
(1) ml,
j1/2
j
+
1+ {l + l }m=(t,k) am+ |m=(t,k)( X0 )
lm,+

(1) j1/2 j
Msp a+
l l=(s,k) ( X0 ) =
1+ {(lm,+ lm,+
+ l )
t
ml, +
+ ( + l )}
ml,
am |m=(t,k) ( X0 )
m=(t,k)
(7.13)
sp
(1) j j1/2
a +
M l l=(s,k) ( X0) =
1+ {(lm,+ lm,+
+ l )
t
ml, +
ml,
+ ( + l )} a |m=(t,k)( X0 )
m=(t,k) m

Mor a+ +
l ( X0 ) = i (l)(al ) ( X0 )
(7.14)
Mor al + ( X0) = i (l)(a+
l ) ( X0 ).
These equations and similar, but more complicated, ones with an arbitrary monomial
+
in the creation operators for a+ l or al are the base for the particle interpretation of the
quantum theory of free fields. For instance, in view of (7.9) and (7.10), the state vec-
+ +
al ( X0 ) and al ( X0 ) are interpreted as ones representing particles with 4-momentum
tors
( m2 c2 + k2 , k) and charges q and +q, respectively; similar multiparticle interpretation
can be given to the general vectors (7.4) too.
The equations (7.9)(7.12) completely agree with similar ones obtained in chapters II
IV on the base of the bilinear commutation relations (5.13).
By means of (7.7), the expression (7.6) can be represented as a linear combination of

terms like (7.5). Indeed, as D is a linear combinations of terms like E (a m an ), by
means of the relations (6.28) we can commute each of these terms with the creation (resp.
+ +
annihilation) operators in the monomial M (a+ +
m1 , am2 , . . .) (resp. ( M (al1 , al2 , . . .)) =

M (al1 , al2 , . . .)) and thus moving them to the right (resp. left) until they act on the vacuum

X0, giving the zero vector see (7.7). In this way the matrix elements of the dynamical
variables, in particular their mean values, can be expressed as linear combinations of scalar
products of the form (7.5). Therefore the supposition (7.7) reduces the computation of mean
values of dynamical variables to the one of the vacuum mean value of a product (composi-
tion) of creation and annihilation operators in which the former operators stand to the right
of the latter ones. (Such a product of creation and annihilation operators can be called their
antinormal product; cf. the properties (6.30) of the antinormal ordering operator A .)
The calculation of such mean values, like (7.5) for states , = X0 , however, cannot
be done (on the base of (6.27)(6.28), (7.7) and (7.1a)) unless additional assumptions are
made. For the purpose one needs some kind of commutation relations by means of which
the creation (resp. annihilation) operators on the r.h.s. of (7.5) to be moved to the left (resp.
right) until they act on the left (resp. right) vacuum vector X0 ; as a result of this operation,
the expressions between the two vacuum vectors in (7.5) should transform into a linear
combination of constant terms and ones with no contribution in (7.5). (Examples of the last
type of terms are E (a
m a ) and normally ordered products of creation and annihilation
operators.) An alternative procedure may consists in defining axiomatically the values of
all or some of the mean values (7.5) or, more stronger, the explicit action of all or some
of the operators, entering in the r.h.s. of (7.5), on the vacuum.1 It is clear, both proposed
schemes should be consistent with the relations (6.27)(6.28), (7.1b) and (7.7)(7.8).
Let us summarize the problem before us: the operator E in (6.27)(6.28) has to be fixed
and a method for computation of scalar products like (7.5) should be given provided the
vacuum vector X0 satisfies (7.1b), (7.2), (7.3) and (7.7). Two possible ways for exploration
of this problem were indicated above.
Consider the operator E . Supposing E (a
m an ) to be a function only of am and an ,

we, in view of (7.1b), can write E (am an ) = f (am an ) b with b = an (upper sign)

or b = am (lower sign) and some functions f . Applying (6.27), we obtain (do not sum
1 Such an approach resembles the axiomatic description of the scattering matrix [1, 9, 10].
over l)
+ + +
E (a+
m al ) = f (am , al ) al E (a+
m al ) = f (am , al ) al
+
E (a + + +
l am ) = f (am , al ) al E (a +
l am ) = f (am , al ) al .

Since E is a linear operator, the expression E (a
m an ) turns to be a linear and homo-
geneous function of am and an , which immediately implies f (A, B) = A for operators

A and B and some constants C. For future convenience, we assume = 1, which

can be achieved via a suitable renormalization of the creation and annihilation operators.2
Thus, the last equations reduce to
+
E (a+
m al ) = am al E (a+ +
m al ) = am al (7.15a)

E (a + +
l am ) = am al E (a + +
l am ) = am al . (7.15b)
Evidently, these equations convert (6.27), (7.7) and (7.8) into identities. Compar-
ing (7.15) and (3.22), we see that the identification
E=N (7.16)
of the operator E with the normal ordering operator N is quite natural. However, for our
purposes, this identification is not necessary as only the equations (7.15), not the general
definition of N , will be employed.
As a result of (7.15), the commutation relations (6.28) now read:
[a+ + +
l , am an ] + ln am = 0 (7.17a)
[a+ + +
l , am an ] + ln am = 0 (7.17b)
[a +
l , am an ] lm a
n =0 (7.17c)
[a +
l , am an ] lm a
n = 0. (7.17d)
(In a sense, these relations are one half of the (para)commutation relations (5.16): the
latter are a sum of the former and the ones obtained from (7.17) via the changes a+ m
+ + +
an
an am and am an
an am ; the last relations correspond to (6.28) with

E = A , A being the antinormal ordering operator see (6.30). Said differently, up to the
replacement a i

2a
l for all l, the relations (7.17) are identical with (5.16) for = 0;
as noted in [36, the remarks following theorem 2 in sec. 1], this is a quite exceptional
case from the view-point of parastatistics theory.) By means of (5.8) for = , one can
verify that equations (7.17) agree with the bilinear commutation relations (5.13), i.e. (5.13)
convert (7.17) into identities.
The equations (7.15) imply the following explicit forms of the number operators (6.31)
and the transition operators (6.40):

Nl = a+
l al

Nl = al + a
l (7.18)
Nlm = a+
Nlm = al + a

l am m. (7.19)
2 Since + = 0 or/and = 0 implies D = 0, due to (6.8), these values are excluded for evident reasons.
As a result of them, the equations (6.33)(6.36) are simply a different form of writing
of (3.24), (3.25), (3.28) and (3.29), respectively.
Let us return to the problem of calculation of vacuum mean values of antinormal ordered
products like (7.5). In view of (7.1b) and (7.3), the simplest of them are
X0 | idF ( X0 ) = X0 | M ( X0 ) = 0 (7.20)
where C and M + (resp. M ) is any monomial of degree not less than 1 only in

the creation (resp. annihilation) operators; e.g. M = a
l , al , al1 al2 , al1 al2 . These
equations, with = 1, are another form of what is called the stability of the vacuum:
if Xi denotes an i-particle state, i N {0}, then, by virtue of (7.20) and the particle
interpretation of (7.4), we have
Xi | X0 = i0 , (7.21)
i.e. the only non-forbidden transition into (from) the vacuum is from (into) the vacuum.
More generally, if Xi ,0 and X0, j denote respectively i -particle and j -antiparticle states,
with X0,0 := X0 , then
Xi ,0 | X0, j = i 0 0 j , (7.22)
i.e. transitions between two states consisting entirely of particles and antiparticles, respec-
tively, are forbidden unless both states coincide with the vacuum. Since we are dealing with
free fields, one can expect that the amplitude of a transitions from an (i -particle + j -an-
tiparticle) state Xi , j into an (i -particle + j -antiparticle) state Xi , j is
Xi , j | Xi, j = i i j j , (7.23)
but, however, the proof of this hypothesis requires new assumptions (vide infra).
Let us try to employ (7.17) for calculation of expressions like (7.5). Acting with (7.17)
and their Hermitian conjugate on the vacuum, in view of (7.1b), we get
+
a+ +
m (an al +ln idF )( X0 ) = 0 an + (a
m al lm idF )( X0 ) = 0
+
(7.24)
am+ (a + +
n al +ln idF )( X0 ) = 0 an (am al lm idF )( X0 ) = 0.

These equalities, as well as (7.17), cannot help directly to compute vacuum mean val-
ues of antinormally ordered products of creation and annihilation operators. But the equa-
tions (7.24) suggest the restrictions3
al a+
m ( X0 ) = lm X0 a +
l am ( X0 ) = lm X0
(7.25)
a +
l am ( X0 ) = lm X0 al am+ ( X0 ) = lm X0
to be added to the definition of the vacuum. These conditions convert (7.24) into identities
and, in this sense agree with (7.17) and, consequently, with the bilinear commutation rela-
tions (5.13). Recall [22,29], the relations (7.25) are similar to ones accepted in the parafield
theory and coincide with that for parastatistics of order p = 1; however, here we do not sup-
pose the validity of the paracommutation relations (5.20) (or (5.16)). Equipped with (7.25),
3 Since the operators a
l and al are, generally, degenerate (with no inverse ones), we cannot say that (7.24)
implies (7.25).
one is able to calculate the r.h.s. of (7.5) for any monomial M (resp. M ) and monomials
M (resp. M ) of degree 1, deg M = 1 (resp. deg M = 1).4 Indeed, (7.25), (7.1b) and (7.3)
entail:
+ +
X0 |a
l am ( X0 ) = X0 |al am ( X0 ) = lm
X0 |a + +
l am ( X0 ) = X0 |al am ( X0 ) = lm

+
X0 |( M (a+ +
l1 , al2 , )) am ( X0 )

+
(7.26)
= X0 |( M (a+ +
l1 , al2 , )) am ( X0 ) = 0

deg M 2
X0 |a + +
l M (am1 , am2 , )( X0 )
= X0 |al M (a+ +
m1 , am2 , )( X0 ) = 0 deg M 2.
Hereof the equation (7.23) for i + j = 1 (resp. i + j = 1) and arbitrary i and j (resp.
iand j ) follows.
However, it is not difficult to be realized, the calculation of (7.5) in cases more general
than (7.20) and (7.26) is not possible on the base of the assumptions made until now.5 At
this point, one is free so set in an arbitrary way the r.h.s. of (7.5) in the mentioned general
case or to add to (7.17) (and, possibly, (7.25)) other (commutation) relations by means of
which the r.h.s. of (7.5) to be calculated explicitly; other approaches, e.g. some mixture of
the just pointed ones, for finding the explicit form of (7.5) are evidently also possible. Since
expressions like (7.5) are directly connected with observable experimental results, the only
criterion for solving the problem for calculating the r.h.s. of (7.5) in the general case can be
the agreement with the existing experimental data. As it is known [1, 2, 5, 11], at present
(almost?) all of them are satisfactory described within the framework of the bilinear com-
mutation relations (5.13). This means that, from physical point of view, the theory should
be considered as realistic one if the r.h.s. of (7.5) is the same as if (5.13) are valid or is
reducible to it for some particular realization of an accepted method of calculation, e.g. if
one accepts some commutation relations, like the paracommutation ones, which are a gen-
eralization of (5.13) and reduce to them as a special case (see, e.g., (5.20)). It should be
noted, the conditions (7.1b)(7.3) and (7.25) are enough for calculating (7.5) if (5.16), or its
versions (5.17) or (5.20), are accepted (cf. [29]). The causes for that difference are replace-
ments like [a+ +
m , an ]
2am an , when one passes from (5.16) to (7.17); the existence of

+
terms like an a+m al in (5.16) are responsible for the possibility to calculate (7.5).
If evidences appear for events for which (7.5) takes other values, one should look, e.g.,
for other commutation relations leading to desired mean values. As an example of the last
type can be pointed the following anomalous bilinear commutation relations (cf. (5.13))
[a
l , am ] = 0 [al , a
m ] = 0
[a
l , am ] = (1) lm idF
2j
[al , a
m ] = (1) lm idF
2j
[a
l , am ] = 0 [al , a
m ] = 0
4For deg M = 0 (resp. deg M = 0) see (7.20).
5It should be noted, the conditions (7.1b)(7.3) and (7.25) are enough for calculating (7.5) if the rela-
tions (5.16), or their version (5.20), are accepted (cf. [29]). The cause for that difference is in replacements
+
like [a+ + +
m ,an ]
2am an , when one passes from (5.16) to (7.17); the existence of terms like an am al
in (5.16) is responsible for the possibility to calculate (7.5), in case (5.16) hold.
[a
l , am ] = (1) lm idF
2j
[al , a
m ] = (1) lm idF ,
2j
(7.27)
which should be imposed after expressions like E (a

m an ) are explicitly calculated.
These relations convert (7.17) and (7.25) into identities and by their means the r.h.s. of (7.5)
can be calculated explicitly, but, as it is well known [1, 2, 5, 11, 34] they lead to deep contra-
dictions in the theory, due to which should be rejected.6
At the time being, it seems, the bilinear commutation relations (5.13) are the only known
commutation relations which satisfy all of the mentioned conditions and simultaneously
provide an evident procedure for effective calculation of all expressions of the form (7.5).
(Besides, for them and for the paracommutation relations the vectors (7.4) form a base, the
Fock base, for the systems Hilbert space of states [22].) In this connection, we want to
mention that the paracommutation relations (5.16) (or their conventional version (5.20)), if
imposed as additional restrictions to the theory together with (7.17), reduce in this particular
case to (5.13) as the conditions (7.25) show that we are dealing with a parafield of order
p = 1, i.e. with an ordinary field [22, 33].7
Ending this section, let us return to the definition of the vacuum X0 . It, generally,
depends on the adopted commutation relations. For instance, in a case of the bilinear com-
mutation relations (5.13) it consists of the equations (7.1a)(7.3), while in a case of the
paracommutation relations (5.16) (or other ones generalizing (5.13)) it includes (7.1a)(7.3)
and (7.25).
8. Commutation Relations for

Several Coexisting Different Free Fields
Until now we have considered commutation relations for a single free field, which can be
scalar, or spinor or vector one. The present section is devoted to similar treatment of a
system consisting of several, not less than two, different free fields. In our context, the
fields may differ by their masses and/or charges and/or spins; e.g., the system may consist
of charged scalar field, neutral scalar field, massless spinor field, massive spinor field and
massless neural vector field. It is a priori evident, the commutation relations regarding only
one field of the system should be as discussed in the previous sections. The problem is to
be derived/postulated commutation relations concerning different fields. It will be shown,
the developed Lagrangian formalism provides a natural base for such an investigation and
makes superfluous some of the assumptions made, for example, in [33, p. B 1159, left
column] or in [22, sec. 12.1], where systems of different parafields are explored.
To begin with, let us introduce suitable notation. With the indices , , = 1, 2, . . ., N
will be distinguished the different fields of the system, with N N, N 2, being their
number, and the corresponding to them quantities. For brevity, when talking on the field
with label , we shall call it -field. Let q and j be respectively the charge and spin of
6 As it was demonstrated in chapters IIIV, a quantization like (7.27) contradicts to (is rejected by) the
charge symmetric Lagrangians (2.4).

7 Notice, as a result of (7.17), the relations (5.16) correspond to (6.28) for E = A , with A being the
antinormal ordering operator (see (6.30)).

the -th field. Similarly to (2.7), we define

0 for scalar -th field

j := 12 for spinor -th field

1 for vector -th field

1 for q = 0 (neutral (Hermitian) field) (8.1)
:=
0 for q = 0 (charged (non-Hermitian) field)

2 j +1 for integer j (bose fields)
:= (1) = .
1 for half-integer j (fermi fields)
Suppose L is the Lagrangian of the -field. For definiteness, we assume L for all
to be given by one and the same set of equalities, viz. (2.1), or (2.3) or (2.4). To save
some space, below the case (2.4), corresponding to charge symmetric Lagrangians, will be
considered in more details; the reader can explore other cases as exercises.
Since the Lagrangian of our system of free fields is
N
L := L , (8.2)
=1
the dynamical variables are

N
D= D (8.3)
=1
and the corresponding system of Euler-Lagrange equations consists of the independent

equations for each of the fields of the system (see (2.6) with L for L ). This allows an
introduction of independent creation and annihilation operators for each field. The ones for

the -th field will be denoted by a ,s (k) and a,s (k); notice, the values of the polarization
variables generally depend on the field considered and, therefore, they also are labeled with
index for the -field. 1 For brevity, we shall use the collective indices l , m and n , with

l := (, s, k) etc., in terms of which the last operators are a l and al , respectively. The

particular expressions for the dynamical operators D are given via (2.9)(2.12) in which
the following changes should be made (see footnote 1):

j
j
s
s s
s

ss , ,
ss , , (8.4)
(k)

s s
(k) l s s
(k)
l (k).
The content of sections 3 and 4 remains valid mutatis mutandis, viz. provided the just
pointed changes (8.4) are made and the (integral) dynamical variables are understood in
conformity with (8.3).
1 One may also write k for k, but this is not essential as mathematically k is simply a variable in R3 ,
k R3 . However, if one deals with the 4-vector k, then one should write k instead of k as k = (k0,k), where

k0 = m2 + k 2 with m being the mass parameter of the -field.
8.1. Commutation Relations Connected with the Momentum

Operator. Problems and Their Possible Solutions
In sections 57, however, substantial changes occur; for instance, when one passes
from (5.12) or (5.15) to (5.16). We shall consider them briefly in a case when one starts
from the charge symmetric Lagrangians (2.4).
The basic relations (5.12), which arise from the Heisenberg relation (4.1) concerning
the momentum operator, now read (here and below, do not sum over , and/or and/or if
the opposite is not indicated explicitly!)

+
al , [am , am ] + [a+
m , a ] (1 + )l m a
m l = 0 (8.5a)

+ +
al , [am , am ] + [am , am ] (1 + )l m al = 0. (8.5b)
It is trivial to be seen, the following generalizations of respectively (5.14) and (5.15)

+
al , [am , am ] (1 + )l m a
l = 0 (8.6a)

+
al , [am , am ] (1 + )l m a
l = 0 (8.6b)

+
al , [am , am ] (1 + )l m al = 0 (8.6c)

+
al , [am , am ] (1 + )l m al = 0 (8.6d)

a+ +
l , [am , am ]
+ 2l m a+
l = 0 (8.7a)

a+ + +
l , [am , am ] + 2 l m al = 0 (8.7b)

+
al , [am , am ] 2 l m a
l = 0 (8.7c)

+
al , [am , am ] 2l m a
l = 0 (8.7d)
provide a solution of (8.5) in a sense that they convert it into identity. As it was said
in Sect. 5, the equations (8.6) (resp. (8.7)) for a single field, i.e. for = , agree (resp.
disagree) with the bilinear commutation relations (5.13).
The only problem arises when one tries to generalize, e.g., the relations (8.7) in a way
similar to the transition from (5.15) to (5.16). Its essence is in the generalization of expres-
sions like [am , a ] and l m a
m l . When passing from (5.15) to (5.16), the indices l and
m are changed so that the obtained equations to be consistent with (5.13); of course, the
numbers and are preserved because this change does not concern the field regarded. But
the situation with (8.7) is different in two directions:
(i) If we change the pair (m, m ) in [am , a ] with (m, n), then with what the num-
m
ber should be replace? With , or or with something else? Similarly, if the mentioned
change is performed, with what the multiplier in l m a l should be replaced? The
problem is that the numbers and are related to terms like am a

m
and a
m
a
m
, in
the momentum operator, as a whole and we cannot say whether the index in and
originates from the first of second index m in these expressions.

(ii) When writing (m , n) for (m , m ) (see (i) above), then shall we replace l m a l
with l m a
n , or l n a

m , or
m n l a
? For a single field, = = , this problem is solved
by requiring an agreement of the resulting generalization (of (5.16) in the particular case)
with the bilinear commutation relations (5.13). So, how shall (5.13) be generalized for sev-
eral, not less than two, different fields? Obviously, here we meet an obstacle similar to the
one described in (i) above, with the only change that should stand for .
Let bl and cl denote some creation or annihilation operator of the -field. Consider
the problem for generalizing the (anti)commutator [bl , cl ] . This means that we are
looking for a replacement
[bl , cl ]
f (bl , cm ; , ), (8.8)
where the functions f are such that

f (bl , cm ; , )= = [bl , cl ] . (8.9)
Unfortunately, the condition (8.9) is the only restriction on f that the theory of free
fields can provide. Thus the functions f , subjected to equation (8.9), become new free
parameters of the quantum theory of different free fields and it is a matter of convention
how to choose/fix them.
It is generally accepted [22, appendix F], the functions f to have forms maximum
similar to the (anti)commutators they generalize. More precisely, the functions
f (bl , cm ; , ) = [bl , cm ] (8.10)
where C are such that
= , (8.11)
are usually considered as the only candidates for f . Notice, in (8.10), are functions in
and , not in l and/or m . Besides, if we assume to be function only in and ,
then the general form of is
= u + (1 u ) + v (1 ) u , v C, (8.12)
due to (8.1) and (8.11). (In view of (5.13), the value = +1 (resp. = 1) corresponds
to quantization via commutators (resp. anticommutators) of the corresponding fields.)
Call attention now on the numbers which originate and are associated with each term
[bl , cm ] . With every change (8.8) one can associate a replacement

g(bl , cm ; , ), (8.13)
where the function g is such that

g(bl , cm ; , )= = . (8.14)
Of course, the last condition does not define g uniquely and, consequently, the function
g, satisfying (8.14), enters in the theory as a new free parameter. Suppose, as a working
hypothesis similar to (8.10)(8.11), that g is of the form
g(bl , cm ; , ) = , (8.15)
where are complex numbers that may depend only on and and are such that
= . (8.16)
Besides, if we suppose to be functions only in and , then
= x + y + (1 x y ) x , y C, (8.17)
as a result of (8.1) and (8.16).

Let us summarize the above discussion. If we suppose a preservation of the algebraic
structure of the bilinear commutation relations (5.13) for a system of different free fields,
then the replacements
[bl , cl ]
[bl , cm ] = (8.18a)

= (8.18b)
should be made; accordingly, the relations (5.13) transform into:
[a
l , am ] = 0 [al , am ] = 0

[a
l , am ] = l m idF
1

[al , am ] = l m idF 1

[a
l , am ] = 0 [al , a ] = 0
m

[a
l , am ] = l m idF
1

[al , a ]
m
= l m idF 1

, (8.19)

where 1 (resp. ) in 1 corresponds to the choice of the upper (resp. lower) signs. If
we suppose additionally (resp. ) to be a function only in and (resp. in and
), then these numbers are defined up to two sets of complex parameters:
= u + (1 u ) + v (1 ) u , v C (8.20a)
= x + y + (1 x y ) x , y C. (8.20b)
A reasonable further specialization of and may be the assumption their ranges

to coincide with those of and , respectively. As a result of (8.1), this supposition is
equivalent to
v = u , u + 1, u 1, u u C (8.21a)
(x , y ) = (0, 0), (0, 1),(1,0),(1,1). (8.21b)
Other admissible restriction on (8.20) may be the requirement and to be sym-

metric, viz.
( , ) = ( , ) = ( , ) (8.22a)

( , ) = ( , ) = ( , ), (8.22b)
which means that the -th and -th fields are treated on equal footing and there is no a
priori way to number some of them as the first or second one.2 In view of (8.20), the
conditions (8.22) are equivalent to
1
u = v C (8.23a)
2
y = x . (8.23b)
If both of the restrictions (8.21) and (8.23) are imposed on (8.20), then the arbitrariness
of the parameters in (8.20) is reduced to:
1 1 1 1
(u , u) = , , , (8.24a)
2 2 2 2
(x , y ) = (0, 0), (1, 1) (8.24b)
and, for any fixed pair (, ), we are left with the following candidates for respectively
and :
1
+ := (+1 + + ) (8.25a)
2
1
:= (1 + + + ) (8.25b)
2

0 := + (8.25c)

1 := + . (8.25d)
When free fields are considered, as in our case, no further arguments from mathematical
or physical nature can help for choosing a particular combination ( , ) from the four
possible ones according to (8.25) for a fixed pair (, ). To end the above considerations of
and , we have to say that the choice
1

( , ) = (+ , 0 ) = (+1 + + ), + (8.26)
2
is known as the normal case [22, appendix F]; in it the relative behavior of bose (resp.
fermi) fields is as in the case of a single field, i.e. they are quantized via commutators
(resp. anticommutators) as ( , ) = (+1, 0) (resp. ( , ) = (1, 0)), and the one
of bose and fermi field is as in the case of a single fermi field, viz. the quantization is

via commutators as ( , ) = (+1, 0). All combinations between and 0,1 different
from (8.26) are referred as anomalous cases. Above we supposed the pair (, ) to be fixed.
If and are arbitrary, the only essential change this implies is in (8.25), where the choice
of the subscripts +, , 0 and 1 may depend on and . In this general situation, the normal
case is defined as the one when (8.26) holds for all and . All other combinations are
referred as anomalous cases; such are, for instance, the ones when some fermi and bose
operators satisfy anticommutation relations, e.g. (8.19) with = 1 for + = 0,
or some fermi fields are subjected to commutation relations, like (8.19) with = +1
2 However, nothing can prevent us to make other choices, compatible with (8.18), in the theory of free fields;
for instance, one may set = and = 12 ( + ) .

for = = 1. For some details on this topic, see, for instance, [22, appendix F], [9,
chapter 20] and [34, sect 4-4]. Fields/operators for which = +1 (resp. = 1), with
= , are referred as parabose (resp. parafermi) in the parafield theory [22, 33]. One can
transfer this terminology in the general case and call the fields/operators for which = +1
(resp. = 1), with = , relative bose (resp. fermi) fields/operators.
Further the relations (8.19) will be referred as the multifield bilinear commutation rela-
tions and it will be assumed that they represent the generalization of the bilinear commuta-
tion relations (5.13) when we are dealing with several, not less than two, different quantum
free fields. The particular values of and in them are insignificant in the following;
if one likes, one can fix them as in the normal case (8.26). Moreover, even in (8.19) the
definition of is completely inessential at all, as always appears in combinations like
l m (see (8.19) or similar relations, like (8.27), below), which are non-vanishing if
= , but then = ; so one can freely write for in all such cases.
Equipped with (8.19) and (8.18), we can generalize (8.7) in different ways. For exam-
ple, the straightforward generalization of (5.16) is:

+ +
al , [am , an ] + 2l n a+
m
=0 (8.27a)

+ +
al , [am , an ] + 2 l n a+
m
=0 (8.27b)

+
al , [am , an ] 2 l m a n = 0 (8.27c)

+
al , [am , an ] 2l m a
n = 0. (8.27d)
However, generally, the relations (8.19) do not convert (8.27) into identities. The reason
is that an equality/identity like (cf. (5.8))
[bl , cm dn ] = [bl , cm ] dn + cm [bl , dn ] , (8.28)
where bl , cm and dn are some creation/annihilation operators and C, can be valid

only for
= = 1/ ( = 0), (8.29)
which, in particular, is fulfilled if = and = 1. So, the agreement between (8.19)

and (8.27) depends on the concrete choice of the numbers . There exist cases when even
the normal case (8.26) cannot ensure (8.19) to convert (8.27) into identities; e.g. when the
-th field and -th fields are fermion ones and the -th field is a boson one. Moreover,
it can be proved that (8.19) and (8.27) are compatible in the general case if unacceptable
equalities like a
l am = 0 hold.
One may call (8.27) the multifield paracommutation relations as from them a corre-
sponding generalization of (5.18) and/or (5.20) can be derived. For completeness, we shall
record the multifield version of (5.20):
[bl , [bm , bn ] ] = 2l m bn [bl , [bm , bn ] ] = 0 (8.30a)

[cl , [cm , cn ] ] = 2l m cn [cl , [cm , cn ] ] = 0 (8.30b)
[bl , [cm , cn ] ] = 2 l n bm [cl , [bm , bn ] ] = 2 l n cm . (8.30c)
For details regarding these multifield paracommutation relations, the reader is referred
to [22, 33], where the case = = = 0 is considered.
Exercise 8.1. Write the multifield versions of the commutation relations (5.22) or (5.23),
which provide examples of generalizations of (8.7) and hence of (8.19) and (8.27).
8.2. Commutation Relations Connected with the Charge and

Angular Mmomentum Operators
In a case of several, not less than two, different fields, the basic trilinear commutation
relations (5.33), which ensure the validity of the Heisenberg relation (4.2) concerning the
charge operator, read:

+
al , [am , am ] [a+
m , a ] 2l m a
m l = 0 (8.31a)

+ +
al , [am , am ] [am , am ] + 2l m al = 0. (8.31b)
Of course, these relations hold only for those fields which have non-vanishing charges,
i.e. in (8.31) is supposed (see (8.1))
= 0 = 0 ( q q = 0). (8.32)
The problem for generalizing (8.31) for these fields is similar to the one for (8.7) in the
case of non-vanishing charges, = 0. Without repeating the discussion of Subsect. 8.1,
we shall adopt the rule (8.18) for generalizing (anti)commutation relations between cre-
ation/annihilation operators of a single field. By its means one can obtain different general-
izations of (8.31).
Exercise 8.2. Prove that the commutation relations.

+ +
al , [am , an ] [a+
m , a
n ]

2l n a+
m
=0 (8.33a)

+ +
al , [am , an ] [am , an ] 2l m an = 0 (8.33b)
and their Hermitian conjugate contain (8.31) and (5.35) as special cases and agree
with (8.19) if = and = +1.
Besides, the multifield paracommutation relations (8.27) for charged fields, = =

= 0, convert (8.33) into identities and, in this sense, (8.33) agree with (contain as special
case) (8.27) for charged fields.
Exercise 8.3. Show that the commutation relations

+
al , [am , an ] + l n am
=0 (8.34a)

+
al , [am , an ] l n a
m
= 0, (8.34b)
which are a multifield generalization of (5.34), do not agree with (8.27) for charged fields
and, consequently, with (8.33).
The consideration of commutation relations originating from the orbital Heisenberg

equation (4.4) is analogous to the one of the same relations regarding the charge operator.
The multifield version of (5.49) is:

( (m ) + (n))([a +
l , [am , an ]

+ [a+ , a ] ] )
= 4(1 + )l m (l )(a
m n n =m l ) (8.35a)

( (m ) + (n))([a +
l , [am , an ]

+ [a+ ,
] ] )
= 4(1 + ) (l )(a
m a n n =m l m l ) (8.35b)
where

(l ) := (k) = k
k if l = (, s, k). (8.36)
k k
Exercise 8.4. Applying (5.51), with m for m and n for n, check that the multifield para-
commutation relations (8.27) convert (8.35) into identities and hence provide a solution
of (8.35) and ensure the validity of (4.4), when system of different free fields is considered.
Exercise 8.5. Show that a solution of (8.35), which does not agree with (8.27), is provided
by the following multifield generalization of (5.52):

+ +

al , [am , an ] = a+ + +
l , [am , an ] = (1 + )l n am (8.37a)

+
+ +
al , [am , an ] = al , [am , an ] = +(1 + )l m an , (8.37b)
which provides also a solution of (8.5).
Notice, the evident multifield version of (5.53) agrees with (8.5), but disagrees
with (8.35) when the lower signs are used.
At last, the multifield exploration of the spin Heisenberg relations (4.5) is a mutatis
mutandis (see (8.35)) version of the corresponding considerations in the second part of Sub-
sect. 5.3. The main result here is that the multifield bilinear commutation relations (8.19),
as well as their para counterparts (8.27), ensure the validity of (4.5).
8.3. Commutation Relations between the Dynamical Variables

The aim of this subsection is to be discussed/proved the commutation relations
(4.15)(4.24) for a system of at least two different quantum fields from the view-point
of the commutation relations considered in subsections 8.1 and 8.2.
To begin with, we rewrite the Heisenberg relations (4.1), (4.2) and (4.4) in terms of
creation and annihilation operators for a multifield system [1, 2]:

[a
l , P ] = k al [al , P ] = k al (8.38)

[a
l , Q ] = q al [al , Q ] = q al (8.39)

[a
l , M ] = i(l ) al
or
[al , M
or
] = i(l ) al , (8.40)
where l = (, s
, k), (l ) is defined by (8.36), k = (k0 , k), q is the charge of the
-field, and k0 = m2c2 + k2 is set in (8.38) and (8.40) (after the differentiations are per-
formed in the last case). The corresponding version of (4.5) is more complicated and de-
pends on the particular field considered (do not sum over s !):
st ,+
,s (k), M ] = ig

f s [a (k)a+
sp
,t (k)
t
s t ,

+ (k)a ,t (k)
s
, (8.41)
f [a
,s (k), Msp ] = ih
s t
(k)a,t+ (k)
t
s t ,+

+ (k)a ,t (k) ,
where f s = 1, 0, +1 (depending on the particular field), g := h := 1

j + j 0

j +1 s t ,+ s t ,
(1) and (k) and (k) are some functions which strongly depend on

s t ,
the particular field considered, with (k) being related to the spin (polarization)
s t ,
functions (k) (see (2.14) and (2.11)).3
Exercise 8.6. Using (4.6), (8.40) and (8.41), write the Heisenberg relations (4.3) in a form
similar to (8.38)(8.41).
The commutation relations involving the momentum operator are:
[ P , P ] = 0 [ Q , P ] = 0
[ S, P ] = [ M , P ] = 0
sp
(8.45)
[ L, P] = [ M
or
, P ] = [ M , P ] = i{ P P}.
We claim that these equations are consequences from (8.38) and the explicit expres-
sions (2.9)(2.12) and (4.11)(4.13) for the operators of the dynamical variables of the free
fields considered in the present work. In fact, since (8.38) implies
[b
l cm , P ] = 0 l = (, s, k), m = (, s, k) (8.46a)
3 If i (k) are the Fourier images of the -th field and
s ,+ s ,
i (k) = vi (k)a+ ,s (k) + vi (k)a
,s (k) , (8.42)
s
s ,
where vi (k) are linearly independent functions normalize via the condition
s , t ,
vi (k) vi (k) = s t f s , (8.43)
i

with f s = 1 for j = 0, 12 and f s = 0,1 for ( j ,s ) = (1,3) or ( j ,s ) = (1,1),(1,2), respectively, then
+ s t , 1 s ,+ i t ,
(k) := v
g i,i i
(k) Ii vi (k)
(8.44)
s t , 1 s , i t ,
(k) := v
g i,i i
(k) Ii vi (k),
s t ,
s t , (k) s t ,
i given via (4.25). Besides,
with Ii (k) = with an exception that (k) = 0 for j = 1
2
and (,) = (a,0),(0,a) with a = 1,2,3.

[b
l (l ) cm , P ] = 2(k k )bl cm , (8.46b)

where b
l , cl = al , al and (l ) is defined via (8.36) and (2.13), the verification
of (8.45) reduces to almost trivial algebraic calculations. Further, we assert that any sys-
tem of commutation relations considered in Subsect. 8.1 entails (8.45): as these relations
always imply (8.5) (or similar multifield versions of (5.10) and (5.11) in the case of the
Lagrangians (2.1) or (2.3), respectively) and, on its turn, (8.5) implies (4.1), the required
result follows from the last assertion and the remark that (4.1) and (8.38) are equivalent.
Exercise 8.7. As an additional verification of the validity of (8.45), prove these equations
by invoking the identity (5.8) and any system of commutation relations mentioned in Sub-
sect. 8.1, in particular (8.19) and (8.27).
The commutation relations concerning the charge operator read:
[ P , Q ] = 0 [ Q , Q ] = 0
[ L, Q ] = [ S , Q ] = 0 (8.47)
[ M Q] = [ M , Q ] = [ M , Q ] = 0.
sp
or
,
These equations are trivial corollaries from (2.9)(2.12) and (4.11)(4.13) and the ob-
servation that (8.39) implies
[al a
m
, Q ] = [a
l a m , Q ] = 0 if q = q , (8.48)
due to (5.8) for = 1. Since any one of the systems of commutation relations mentioned
in Subsect. 8.2 entails (8.31) (or systems of similar multifield versions of (5.31) and (5.32),
if the Lagrangians (I.4.1) or (2.3) are employed), which is equivalent to (8.39), the equa-
tions (8.47) hold if some of these systems is valid.
Exercise 8.8. Alternatively, one can prove via a direct calculation that the commutation
relations arising from the charge operator entail the validity of (8.47); for the purpose the
identity (5.8) and the explicit expressions for the dynamical variables via the creation and
annihilation operators should be applied.
At last, consider the commutation relations involving the different angular momentum
operators:
[ P , S ] = [ P , M ] = 0
sp
[ P , L ] = [ P , M
or
] = [ P, M ] = +i{ P P }
[ Q , L ] = [ Q , S] = [ Q , M ] = [ Q , M ] = [ Q , M ] = 0
or sp
(8.49)
[ S , M ] = i S S S + S

[ L , M ] = i L L L + L

[ M , M ] = i M M M + M .
(The other commutators, that can be form from the different angular momentum operators,
are complicated and cannot be expressed in a closed form.) The proof of these relations
is based on equations like (see (8.40) and (5.8))

[bl cm , M
or
] = i (l ) bl cm l = (, s, k), m = (, s , k), (8.50)
+
with bl , cl = a+
l , al , al , al , and similar, but more complicated, ones involving the other
angular momentum operators. It, generally, depends on the particular field considered and
will be omitted.
As it was said in Subsect. 5.3, the Heisenberg relations concerning the angular momen-
tum operator(s) do not give rise to some (algebraic) commutation relations for the creation
and annihilation operators. For this reason, the only problem is which of the commutation
relations discussed in subsections 8.1 and 8.2 imply the validity of the equations (8.49) (or
part of them). The general answer of this problem is not known but, however, a direct cal-
culation by means of (8.7), if it holds, and (5.8) shows the validity of (8.49). Since (8.19)
and (8.27) imply (8.7), this means that the multifield bilinear and para commutation rela-
tions are sufficient for the fulfillment of (8.49).
To conclude, let us draw the major moral of the above material: the multifield bilinear
commutation relations (8.19) and the multifield paracommutation relations (8.27) ensure
the validity of all standard commutation relations (8.45), (8.47) and (8.49) between the
operators of the dynamical variables characterizing free scalar, spinor and vector fields.
8.4. Commutation Relations under the Uniqueness Conditions

As it was said at the end of the introduction to this section, the replacements (8.4) ensure the
validity of the material of Sect. 3 in the multifield case. Correspondingly, the considerations
in Sect. 6 remain valid in this case provided the changes
l
l m
m n
n
lm
l m = l m (8.51)
[bm, bm ]
[bm , bm ] [bm , bn ]
[bm , bn ] ,
with bm (or bm ) being any creation/annihilation operator, and, in some cases, (8.4) are
made.4 Without going into details, we shall write the final results.
The multifield version of (6.27)(6.28) is:
1
E (a
m
a
n ) = E (an am ) = E ([a , a ] )
m n
(8.52)
2

a+ + +
l , E ([am , an ] ) + 2l n am = 0

(8.53a)

+
al , E ([a+ , a ] ) + 2 l n a+
m n m
=0 (8.53b)

+

al , E ([am , an ] ) 2 l m an = 0

(8.53c)

al , E ([a+
m , a n ] )

2l m a
n = 0 (8.53d)
=. (8.53e)
As one can expect, the relations (8.53a)(8.53d) can be obtained from the multifield
paracommutation relations (8.27) via the replacement [, ]
E ([, ] ). It should be paid
special attention on the equation (8.53e). It is due to the fact that in the expressions for
4 As a result of (6.11), (6.16) and (6.17), in expressions like (6.18)(6.26) the number should be replace
by , where and are the corresponding field indices of the creation/annihilation operators on which the
operator E acts, i.e. E (bm bn )
E (bm bn ).
the dynamical variables do not enter cross-field-products, like al + a m

for = , and
it corresponds to the condition (ii) in [33, p. B 1159]. The equality (8.53e) is quite im-
portant as it selects only that part of the E -transformed multifield paracommutation rela-
tions (8.27) which is compatible with the bilinear commutation relations (8.19) (see (8.28)
and (8.29)). Besides, (8.53e) makes (8.53a)(8.53d) independent of the particular definition
of (see (8.11)).
The equations (8.52) are the only restrictions on the operator E ; examples of this op-
erator are provided by the normal (resp. antinormal) ordering operator N (resp. A ), which
has the properties (cf. (3.22) (resp. (6.30))
+
N a+ m a n := a +
m a n N am an := am+ a n
+ + (8.54)
N am an := an am N am an := an am
+ +

+
A a+m a n := a n a
+
m
A am an := a +
n a m
(8.55)
A am
an+ := a m
an+ A am a+ +
n := am an .
The material of Sect. 7 has also a multifield variant that can be obtained via the replace-
ments (8.51) and (8.4). Here is a brief summary of the main results found in that way.
The operator E should possess the properties (8.54) and, in this sense, can be identified
with the normal ordering operator,
E= N. (8.56)
As a result of this fact and = (see (8.11)), the commutation relations (8.53) take the
final form:

+ +
al , am an + l n a+
m
=0 (8.57a)

+ +
al , am an + l n a+m
=0 (8.57b)

+
al , am an l m a n
=0 (8.57c)

+
al , am an l m an = 0 (8.57d)
which is the multifield version of (7.17) and corresponds, up to the replacement a

l
2a
l , to (8.27) with = 0.
The vacuum state vector X0 is supposed to be uniquely defined by the following equa-
tions (cf. (7.1b)(7.3)):

a
l X 0 = 0 al X 0 = 0 (8.58a)
X0 = 0 (8.58b)
X0 | X0 = 1 (8.58c)
al a+
m
( X0 ) = l m X0 a +
l am ( X0 ) = l m X0

(8.58d)
a +
l am ( X0 ) = l m X0 a +
l am ( X0 ) = l m X0 .
The Hilbert space F of state vectors is a direct sum of the Hilbert spaces F of the
different fields and it is supposed to be spanned by the vectors
l 1 l 2 ... = M (a+1 , a+2 , . . .)( X0 ) (8.59)

1 2 l1 l2
with M (a+1 , a+2 , . . .) being arbitrary monomial only in the creation operators.
l1 l2
Since (8.58a), (8.56) and (8.54) imply the multifield version of (7.7), the computation
of the mean values of (7.6), with l1
l11 etc., of the dynamical variables is reduced to the
one of scalar products like (cf. (7.5))
l 1 l 2 ... |m1 m2 ...

1 2 1 2

= X0 | M (a+1 , a+2 , . . .) M (a+1 , a+2 , . . .)( X0 ) (8.60)
l1 l2 m1 m2
of basic vectors of the form (8.59). By means of the basic properties (8.58) of the vacuum,
one is able to calculate the simplest forms of the vacuum mean values (8.60), viz. the mul-
tifield versions (see (8.51)) of (7.20) and (7.26). But more general such expression cannot
be calculated by means of (8.57)(8.58). Prima facie one can suppose that the multifield
commutation relations (8.19), which ensure the vectors (8.59) to form a base of the systems
Hilbert space of states, can help for the calculation of (8.60) in more complicated cases. In
fact, this is the case which works perfectly well and covers the available experimental data.
In this connection, we must mention that the applicability of (8.19) for calculation of (8.60)
is ensured by the compatibility/agreement between (8.19) and (8.57): by means of (5.8) for
= , one can check that (8.19) converts (8.57) into identities.5
The commutation relations (8.57) admit as a solution also the multifield version of
the anomalous bilinear commutation relations (7.27) but it, as we said earlier, leads to
contradictions and must be rejected. The existence of solutions of (8.57) different from
it and (8.19) seems not to be investigated. If there appear data which do not fit into the
description by means of (8.19), one should look for other, if any, solutions of (8.57) or
compatible with (8.57) effective procedures for calculating vacuum mean values like (8.60).
9. Conclusion
In this chapter we have investigated two sources of (algebraic) commutation relations in
the Lagrangian quantum theory of free scalar, spinor and vector fields: the uniqueness of
the dynamical variables (momentum, charge and angular momentum) and the Heisenberg
relations/equations for them. If one ignores the origin of the former one, which is the ordi-
nary case, the paracommutation relations or some their generalizations seems to be the most
suitable candidates for the most general commutation relations that ensure the validity of all
Heisenberg equations. The simultaneous consideration of the both sources mentioned re-
veals, however, their incompatibility in the general case. The outlet of this situation is in the
redefinition of the operators of the dynamical variables, similar to the normal ordering pro-
cedure and containing it as a special case. That operation ensures the uniqueness of the new
(redefined) dynamical variables and changes the possible types of commutation relations.
5 Recall, equations (8.19) and (8.27), or (8.53a)(8.53d), for = are generally incompatible. For in-
stance, excluding some special cases, like systems consisting of only fermi (bose) fields or one fermi (bose)
field and arbitrary number of bose (fermi) fields, the only operators satisfying (8.19) and (8.27) for = and
having normal spin-statistics connection are such that bm cn = 0, with = and bm and cn being any cre-
ation/annihilation operators, which, in particular, means that no states with two particles from different fields
can exist.
Again, the commutation relations, connected with the Heisenberg relations concerning the
(redefined) momentum operator, entail the validity of all Heisenberg equations.
Further constraints on the possible commutation relations follow from the definition/in-
troduction of the concept of the vacuum (vacuum state vector). They practically reduce the
redefined dynamical variables to the ones obtained via normal ordering procedure, which
results in the explicit form (7.17) of the admissible commutation relations. In a sense, they
happen to be one half of the paracommutation ones. As a last argument in the way for
finding the unique true commutation relations, we require the existence of procedure for
calculation of vacuum mean values of anti-normally ordered products of creation and anni-
hilation operators, to which the mean values of the dynamical variables and the transition
amplitudes between different states are reduced. We have pointed that the standard bilinear
commutation relations are, at present, the only known ones that satisfy all of the conditions
imposed and do not contradict to the existing experimental data.
The consideration of a system of at least two different quantum free fields meets a new
problem: the general relations between creation/annihilation operators belonging to differ-
ent fields turn to be undefined. The cause for this is that the commutation relations for any
fixed field are well defined only on the corresponding to it Hilbert subspace of the systems
Hilbert space of states and their extension on the whole space, as well as the inclusion in
them of creation/annihilation operators of other fields, is a matter of convention (when free
fields are concerned); formally this is reflected in the structure of the dynamical variables
which are sums of those of the individual fields included in the system under consideration.
We have, however, presented argument by means of which the a priori existing arbitrari-
ness in the commutation relations involving different field operators can be reduced to the
standard one: these relations should contain either commutators or anticommutators of
the creation/annihilation operators belonging to different fields. A free field theory cannot
make difference between these two possibilities. Accepting these possibilities, the admissi-
ble commutation relations (8.57) for system of several different fields are considered. They
turn to be corresponding multifield versions of the ones regarding a single field. Similarly
to the single field case, the standard multifield bilinear commutation relations seem to be
the only known ones that satisfy all of the imposed restrictions and are in agreement with
the existing data.
References
[1] N. N. Bogolyubov and D. V. Shirkov. Introduction to the theory of quantized fields.

Nauka, Moscow, third edition, 1976. In Russian. English translation: Wiley, New
York, 1980.
[2] J. D. Bjorken and S. D. Drell. Relativistic quantum fields, volume 2. McGraw-Hill

Book Company, New York, 1965. Russian translation: Nauka, Moscow, 1978.
[3] Bozhidar Z. Iliev. On operator differentiation in the action principle in quantum field
theory. In Stancho Dimiev and Kouei Sekigava, editors, Proceedings of the 6th Inter-
national Workshop on Complex Structures and Vector Fields, 36 September 2002, St.
Knstantin resort (near Varna), Bulgaria, Trends in Complex Analysis, Differential
Geometry and Mathematical Physics, pages 76107. World Scientific, New Jersey-
London-Singapore-Hong Kong, 2003.
http://arXiv.org e-Print archive, E-print No. hep-th/0204003, April 2002.
[4] J. D. Bjorken and S. D. Drell. Relativistic quantum mechanics, volume 1 and 2.

McGraw-Hill Book Company, New York, 1964, 1965. Russian translation: Nauka,
Moscow, 1978.
[5] Paul Roman. Introduction to quantum field theory. John Wiley&Sons, Inc., New
York-London-Sydney-Toronto, 1969.
[6] Lewis H. Ryder. Quantum field theory. Cambridge Univ. Press, Cambridge, 1985.
Russian translation: Mir, Moscow, 1987.
[7] A. I. Akhiezer and V. B. Berestetskii. Quantum electrodynamics. Nauka, Moscow,

1969. In Russian. English translation: Authorized English ed., rev. and enl. by the
author, Translated from the 2d Russian ed. by G.M. Volkoff, New York, Interscience
Publishers, 1965. Other English translations: New York, Consultants Bureau, 1957;
London, Oldbourne Press, 1964, 1962.
[8] Pierre Ramond. Field theory: a modern primer, volume 51 of Frontiers in physics.
Reading, MA Benjamin-Cummings, London-Amsterdam-Don Mills, Ontario-Sidney-
Tokio, 1 edition, 1981. 2nd rev. print, Frontiers in physics vol. 74, Adison Wesley
Publ. Co., Redwood city, CA, 1989; Russian translation from the first ed.: Moscow,
Mir 1984.
278 References
[9] N. N. Bogolubov, A. A. Logunov, and I. T. Todorov. Introduction to axiomatic quan-

tum field theory. W. A. Benjamin, Inc., London, 1975. Translation from Russian:
Nauka, Moscow, 1969.
[10] N. N. Bogolubov, A. A. Logunov, A. I. Oksak, and I. T. Todorov. General principles of

quantum field theory. Nauka, Moscow, 1987. In Russian. English translation: Kluwer
Academic Publishers, Dordrecht, 1989.
[11] C. Itzykson and J.-B. Zuber. Quantum field theory. McGraw-Hill Book Company,
New York, 1980. Russian translation (in two volumes): Mir, Moscow, 1984.
[12] Silvan S. Schweber. An introduction to relativistic quantum field theory. Row, Peter-
son and Co., Evanston, Ill., Elmsford, N.Y., 1961. Russian translation: IL (Foreign
Literature Pub.), Moscow, 1963.
[13] Michio Kaku. Quantum field theory. A modern Introduction. Oxford University Press,
New York - Oxford, 1993.
[14] V. Parameswaran Nair. Quantum field theory. A modern perspective. Springer, New
York, 2005.
[15] Warren Siegel. Fields. http://arXiv.org e-Print archive, E-print No. hep-th/99122005,
2005. See also
http://insti.physics.sunysb.edu/ siegel/Field3Archive.zip, and
http://insti.physics.sunysb.edu/ siegel/Field3.pdf.
[16] Mark Srednicki. Quantum field theory. Cambridge University Press, Cambridge,
2007. See also http://arXiv.org e-Print archive, E-print No. hep-th/0409035 and
http://arXiv.org e-Print archive, E-print No. hep-th/0409036.
[17] Steven Weinberg. Foundations, volume I of The quantum theory of fields. Cambridge
University Press, Cambridge, 1995.
[18] A. Zee. Quantum field theory in a nutshell. Princeton University Press, Princeton and
Oxford, 2003.
[19] Bozhidar Z. Iliev. Pictures and equations of motion in Lagrangian quantum field
theory. In Charles V. Benton, editor, Studies in Mathematical Physics Research, pages
83125. Nova Science Publishers, Inc., New York, 2004.
http://arXiv.org e-Print archive, E-print No. hep-th/0302002, February 2003.
[20] A. M. L. Messiah. Quantum mechanics, volume I and II. Interscience, New York,
1958. Russian translation: Nauka, Moscow, 1978 (vol. I) and 1979 (vol. II).
[21] Bozhidar Z. Iliev. Momentum picture of motion in Lagrangian quantum field theory.
International Journal of Theoretical Physics, Group Theory, and Nonlinear Optics,
12(2):????, 2007. To appear.
http://arXiv.org e-Print archive, E-print No. hep-th/0311003, November, 2003.
References 279
[22] Y. Ohnuki and S. Kamefuchi. Quantum field theory and parafields. University of
Tokyo Press, Tokyo, 1982.
[23] W. Pauli and W. Heisenberg. Zur Quantendynamik der wellenfelder (On the quantum
dynamics of wavefields). Z. Phys., 56(1):161, 1929. In German. Russian translation
in [50, pp. 3088].
[24] P. A. M. Dirac. The principles of quantum mechanics. Oxford at the Clarendon Press,
Oxford, fourth edition, 1958. Russian translation in: P. Dirac, Principles of quantum
mechanics, Moscow, Nauka, 1979.
[25] V. A. Fock. Fundamentals of quantum mechanics. Mir Publishers, Moscow, 1978.

Russian edition: Nauka, Moscow, 1976.
[26] Bozhidar Z. Iliev. Lagrangian quantum field theory in momentum picture. In

O. Kovras, editor, Quantum Field Theory: New Research, pages 166. Nova Science
Publishers, Inc., New York, 2005.
http://arXiv.org e-Print archive, E-print No. hep-th/0402006, February 1, 2004.
[27] Bozhidar Z. Iliev. On momentum operator in quantum field theory. In Frank Columbus
and Volodymyr Krasnoholovets, editors, Frontiers in quantum physics research, pages
143156. Nova Science Publishers, Inc., New York, 2004.
http://arXiv.org e-Print archive, E-print No. hep-th/0206008, June 2002.
[28] G. Korn and T. Korn. Mathematical Handbook. Mc Graw-Hill Book Company, New
York, second edition, 1968. Russian translation: Nauka, Moscow, 1973,.
[29] H. S. Green. A generalized method of field quantization. Phys. Rev., 90(2):270273,

1953.
[30] D. V. Volkov. On quantization of half-integer spin fields. Zh. Eksperim. i Teor. Fiz.
(Journal of experimental and theoretical physics), 36(5):15601566, 1959. In Rus-
sian. English translation: Soviet Phys.JETP, vol. 9, p. 1107, 1959.
[31] D. V. Volkov. S-matrix in the generalized quantization method. Zh. Eksperim. i Teor.
Fiz. (Journal of experimental and theoretical physics), 38(2):518523, 1960. In Rus-
sian. English translation: Soviet Phys.JETP, vol. 11, p. 375, 1960.
[32] O. W. Greenberg and A. M. I. Messiah. Symmetrization postulate and its experimental

foundation. Phys. Rev., 136B(1B):248267, 1964.
[33] O. W. Greenberg and A. M. I. Messiah. Selection rules for parafields and the absence
of para particles in nature. Phys. Rev., 138B(5B):11551167, 1965.
[34] R. F. Streater and A. S. Wightman. PCT, spin and statistics and all that. W. A.
Benjamin, Inc., New York-Amsterdam, 1964. Russian translation: Nauka, Moscow,
1966.
[35] Res Jost. The general theory of quantized fields. American Mathematical Society,
Rhode Island, 1965. Russian translation: Mir, Moscow, 1967.
280 References
[36] A. B. Govorkov. Parastatistics and internal symmetries. In N. N. Bogolyubov, editor,

Physics of elementary particles and atomic nuclei, volume 14, No. 5, of Particles and
nuclei, pages 12291272. Energoatomizdat, Moscow, 1983. In Russian.
[37] G. C. Wick. Calculation of scattering matrix. Physical Review, 80(2):268, 1950.

Russian translation in [51, pp. 245253].
[38] W. Pauli and V. Weisskopf. Uber die Quantizierung der skalaren relativistischen
Wellengleichung (On the quantization of the scalar relativistic wave equation). Helv.
Phys. Acta, 7:709731, 1934. In German. Russian translation in [50, pp. 194214].
[39] J. Schwinger. The theory of quantized fields. I. Physical Review, 82(6):914, 1951.
Russian translation in [51, pp. 115137].
[40] Bozhidar Z. Iliev. Lagrangian quantum field theory in momentum picture. II. Free
spinor fields.
http://arXiv.org e-Print archive, E-print No. hep-th/0405008, May 1, 2004.
[41] J. D. Bjorken and S. D. Drell. Relativistic quantum mechanics, volume 1. McGraw-

Hill Book Company, New York, 1964. Russian translation: Nauka, Moscow, 1978.
[42] P. A. M. Dirac. Lectures on quantum mechanics. Belfer graduate school of science,

Yeshiva University, New York, 1964. Russian translation in: P. Dirac, Principles of
quantum mechanics, Moscow, Nauka, 1979.
[43] Ludwig Valentin Lorenz. Uber die Intensitat der Schwingungen des Lichts mit den
elektrischen Stromen. Annalen der Physik und Chemie, 131:243263, 1867.
[44] Edmund Whittaker. A history of the theories of aether and electricity, volume 1. The
classical theories. of Harper torchbooks / The science library. Harper & brothers,
New York, 1960. Originally published by Thomas Nelson & Son Ltd, London, 1910;
revised and enlarged 1951. See also the 1989 edition: New York: Dover, 1989.
[45] M. Gockeler and T. Schucker. Differential geometry, gauge theories, and gravity.
Cambridge Univ. Press, Cambridge, 1987.
[46] W. Pauli. Relativistic field theories of elementary particles. Rev. Mod. Phys., 13:203
232, 1941. Russian translation in [50, pp. 372423].
[47] E. Wigner. Do the equations of motion determine the quantum mechanical commuta-
tion relations? Physical Review, 77(5):711712, 1950.
[48] Bozhidar Z. Iliev. Lagrangian quantum field theory in momentum picture. IV. Com-
mutation relations for free fields.
http://arXiv.org e-Print archive, E-print No. arXiv:0704.066 [hep-th], Aprl 1, 2007.
[49] Bozhidar Z. Iliev. On angular momentum operator in quantum field theory. In Frank
Columbus and Volodymyr Krasnoholovets, editors, Frontiers in quantum physics re-
search, pages 129142. Nova Science Publishers, Inc., New York, 2004.
http://arXiv.org e-Print archive, E-print No. hep-th/0211153, November 2002.
References 281
[50] Wolfgang Pauli. Works on quantum theory. Articles 19281958. Nauka Publ.,
Moscow, 1977. In Russian.
[51] Novejshee razvitie kvantovoj electrodinamiki. Sbornik statej (The newest development
of quantum electrodynamics. Collection of papers), editor Ivanenko D. D., Moscow,
1954. IL (Foreign Literature Pub.). In Russian.
Index
angular momentum in Heisenberg picture, 3

density in momentum picture, 15
for spinor field, 85, 86 Coloumb gauge, 197198
for vector field, 151 commutation relations, 213276
in Heisenberg picture, 3 anomalous bilinear, 261262
of spinor field, 109 bilinear, 234, 237, 262
in Heisenberg picture, 113 anomalous, 261262
in momentum picture, 113 multifield, 266, 268
operator connected with
in Heisenberg picture, 3 angular momentum operator, 243
in momentum picture, 15 248
vector in Heisenberg picture, 14 charge operator, 240242
vector in momentum picture, 14 momentum operator, 235240
anomalous bilinear commutation rela- differential, 244245
tions, 261262 for several coexisting different fields,
anomalous cases, 267 262275
anticommutation relations, 234 anomalous cases, 267
anticommutation relations for spinor between dynamical variables,
field, 115122, 140142 270273
comments on, 118119 connected with angular momen-
definition of, 118 tum operators, 270
derivation of, 115118 connected with charge operator,
antiparticle number operator, 254 269270
antiparticle transition operators, 255 connected with momentum opera-
tor, 264269
charge current density normal case, 267
for arbitrary scalar field, 5354 under uniqueness conditions, 273
for spinor field, 85, 101 275
for vector field, 151, 152 general types, 236237, 241, 244
in Heisenberg picture, 3, 6 247
charge operator in broad sense, 234
for scalar arbitrary field, 5262 in narrow sense, 234
after normal ordering, 73 integral, 246247
for spinor field, 101, 115 parabose, 239
after normal ordering, 124 parafermi, 239
for vector field, 165 possible types, 234248
after normal ordering, 184 trilinear, 143, 211212, 235, 237
283
284 Index
240, 242, 245246, 251ff, 259 evolution operator, 113

multifield, 264ff, 268270, 273 momentum, 112, 174, 233
274 spin angular momentum compo-
multifield differential, 270 nents, 119120, 126, 130, 179,
trilinear related to 186
charge operator, 241 commutation relations for
momentum operator, 236237 electromagnetic field, 196
orbital angular momentum opera- scalar arbitrary field, 6672
tor, 244245 comments on, 72
spin angular momentum operator, derivation of, 6670
246247 scalar neutral field, 3641
commutation relations between comments on, 4041
angular momentum and defintion of, 40
charge, 182 derivation of, 3640
evolution operator, 15 vector field, 176182
momentum, 15, 115, 175, 233 commnets on, 178
angular momentum components, 75, definition of, 177
122, 181, 233 derivation of, 176177
charge and conservation laws, 4
angular momentum, 122, 233 for spinor field angular momentum,
evolution operator, 15, 115 104, 108, 109
momentum, 7, 15, 64, 115, 122, constant and conserved operators, 17
174, 233 creation and annihilation operators, 18
orbital angular momentum, 233 interpretation of, 1819
orbital operator, 122 creation and annihilation operators for
spin operator, 122 scalar arbitrary field, 56, 58
dynamical variables, 270273 charge operator in terms of, 62
evolution operator and equations of motion in terms of,
orbital angular momentum, 181 6061
spin angular momentum, 180 in Heisenberg picture, 59
momentum and charge, 182 interpretation of, 5657
momentum operator components, 7, momentum operator in terms of,
114, 174, 233 58
orbital angular momentum and scalar neutral field, 3233, 35
angular momentum, 234 equations of motion in terms of, 36
angular momentum operators, 182 in Heisenberg picture, 33
charge, 182, 233 interpretation of, 33
evolution operator, 113 momentum operator in terms of,
momentum, 64, 113, 115, 174, 34
233 spinor field, 98, 109110
orbital angular momentum compo- charge operator in terms of, 101
nents, 75 dynamical variables in terms of,
spin angular momentum and 100110
angular momentum, 182, 234 in Heisenberg picture, 109
charge, 120, 180, 182, 186, 233 interpretation of, 9597, 125126,
Index 285
129130 for free fields, 218

vector field, 160163 in Heisenberg picture, 3
dynamical variables in terms of, in momentum picture, 13
163171 Euler-Lagrange relation, 6
in Heisenberg picture, 171 evolution operator
interpretation of, 162163, 188 in momentum picture, 8, 10
190 of scalar arbitrary field, 77
of scalar neutral field, 45
DAlembert operator in of spinor field, 127
Heisenberg picture, 26 of vector field, 187
momentum picture, 26
derivatives relative to operators, 47 field operator(s)
definition of, 4 in Heisenberg picture, 2
Dirac equation(s), 140 in momentum picture, 10
analysis of, 8793 of scalar arbitrary field, 47
in Heisenberg picture, 84 of scalar neutral field, 26
in momentum picture, 84, 135 of spinor field, 82
in terms of cration and annihilation of vector field, 147
operators, 111112, 140 free fields
dynamical variables particle interpretation of, 257258
action on 1-particle states, 257258 relative parabose, 268
commutation relations between, relative parafermi, 268
270273 several coexisting, 262ff
of a spinor field, 135136, 138139 state vectors of, 255ff
of free fields, 219222 vacuum of, 255ff
of vector field, 200203 frequency decompositions of
uniqueness, 222227, 249ff scalar arbitrary field, 5657
after normal ordering, 226227 scalar neutral field, 3233
algebraic equations for, 224225 spinor field, 93100
integral equations for, 223224 vector field, 158160
electromagnetic field, 195198 gamma matrices, 83

commutation relations for, 196 properties of, 8587
dynamical variables of, 196197
Lagrangian for, 195 Heisenberg picture of motion, 2
Lorenz condition for, 195 angular momentum operator vector
energy-momentum tensor operator, 3 in, 14
in Heisenberg picture, 6 Heisenberg relations in Heisenberg pic-
of scalar arbitrary field, 5052 ture, 7, 228234
of scalar neutral field, 27 Heisenberg relations in Heisenberg pic-
of spinor field, 85, 100 ture for
of vector field, 151, 152 angular momentum operator, 7, 14,
equations of motion, 3 74, 229232
of arbitrary scalar field, 49 angular momentum operator of
Euler-Lagrange equations spinor field, 120
classical, 6 charge operator, 7, 14, 229
286 Index
charge operator of scalar field, 52 charge symmetric in momentum

momentum operator, 7, 229 picture, 134
orbital angular momentum operator, standard in Heisenberg picture, 83,
74 131
Heisenberg relations in momentum pic- standard in momentum picture, 83,
ture for 131
angular momentum, 120 vector field, 198204
angular momentum operator, 16, 74 charge symmetric in Heisenberg
of spinor field, 86 picture, 199
of vector field, 153 charge symmetric in momentum
angular momentum operator of vec- picture, 199
tor field, 180 standard in Heisenberg picture,
charge operator, 16, 74 147148
of scalar field, 53 standard in momentum picture,
of spinor field, 86, 121 149
of vector field, 153, 182 Lagrangians of free fields, 216218
orbital angular momentum operator, Lorenz condition(s)
74 in Heisenberg picture, 148
orbital operator, 121 in momentum picture, 150
spin operator, 120 role in the massless case, 204211
Klein-Gordon equation momentum operator

analysis of, 2932 in Heisenberg picture, 3
in Heisenberg picture, 26 in momentum picture, 9, 13
in momentum picture, 13, 26 of scalar arbitrary field, 58
in terms of creation and annihilation of scalar neutral field, 34
operators, 36 after normal ordering, 43
Klein-Gordon system of equations for ar- of spinor field, 101
bitrary scalar field, 49 after normal ordering, 124
of vector field, 164
Lagrangian formalism, 27 after normal ordering, 184
in momentum picture, 1217 momentum picture of motion, 822
Lagrangian in angular momentum operator in, 15
Heisenberg picture, 2 angular momentum operator vector
momentum picture, 13 in, 14
Lagrangian of as a 4-analogue of the Shcrodinger
electromagnetic field, 195 one, 1922
scalar arbitrary field charge operator in, 15
in Heisenberg picture, 4748 definition of, 1920
in momentum picture, 47, 49 Dirac equation in, 84
scalar neutral field Euler-Lagrange equations in, 13
in Heisenberg picture, 26 evolution operator in, 8, 10
in momentum picture, 26 field operators in, 10
spinor field, 131142 Klein-Gordon equation in, 13
charge symmetric in Heisenberg Lagrangian formalism in, 1217
picture, 134
Index 287
Lagrangian in, 13 particle interpretation, 258

momentum operator in, 9, 13 particle number operator, 254, 259
momentum representation of, 1719 particle transition operators, 255
operators in, 8 Proca equation(s)
Proca equation in, 150 in Heisenberg picture, 148
state vectors in, 8, 10, 11 in momentum picture, 150
transformation to, 8
momentum representation of momentum relative parabose fields, 268
picture of motion, 1719 relative parafermi fields, 268
multifield bilinear commutation relations, scalar field arbitrary, 4779
266, 268 charge current density, 5354
multifield paracommutation relations, charge operator of, 5262
268 after normal ordering, 73
no-particle condition, 255 creation and annihilation operators
normal case, 267 for, 58
normal ordering, 42 description of, 4754
normal ordering for energy-momentum tensor of, 5052
scalar arbitrary field, 73 equations of motion, 49
scalar neutral field, 4243 analysis of, 5455
spinor field, 123124 in terms of creation and annuula-
vector field, 183184 tion operators, 6061
frequency decompositions, 5657
orbital angular momentum density Lagrangian in
in Heisenberg picture, 3 Heisenberg picture, 4748
of spinor field, 102 momentum picture, 47, 49
of vector field, 151, 152 momentum operator of, 58
orbital angular momentum operator after normal ordering, 73
in Heisenberg picture, 4 orbital angular momentum operator
of scalar arbitrary field of
in Heisenberg picture, 63, 74 in Heisenberg picture, 63, 74
in momentum picture, 6566, 74 in momentum picture, 6566, 74
of scalar neutral field, 28 state vectors of, 7678
of spinor field, 113114 vacuum of, 73
in Heisenberg picture, 102104, scalar field neutral, 2347
107108, 110 creation and annihilation operators
in momentum picture, 113 for, 35
of vector field, 165 description in momentum picture,
after normal ordering, 184 2529
energy-momentum operator, 27
parabose fields, 268 evolution operator of, 45
paracommutation relations, 235 field operator, 26
multifield, 268 frequency decompositions, 3233
nonrelativistic, 238 Lagrangian in
relativistic, 239 Heisenberg picture, 26
parafermi fields, 268 momentum picture, 26
288 Index
momentum operator of, 34 spin angular momentum vector den-

after normal ordering, 43 sity, 85, 86
normal ordering for, 4243 spin operator of
state vectors of, 4447 after normal ordering, 124
vacuum of, 43 vacuum of, 124
spin angular momentum density stability of vacuum, 46, 129, 188, 260
in Heisenberg picture, 3, 6 state vectors
of spinor field, 86 in momentum picture, 10, 11
of vector field, 151, 152 of free fields, 255ff
spin angular momentum operator of scalar arbitrary field, 7678
in Heisenberg picture, 4 of scalar neutral field, 4447
of spinor field, 102107 of spinor field, 127130
of vector field, 166170 of vector field, 186190
after normal ordering, 184185
spinor field, 81143 vacuum
state vectors of, 127130 mean values, 258ff
angular momentum density, 85, 86 properties of, 255ff
charge current density, 85, 86 stability of, 46, 129, 188, 260
charge operator of, 101 vacuum of
after normal ordering, 124 free fields, 255ff
creation and annihilation operators scalar arbitrary field, 73
for, 98, 109110 scalar neutral field, 43
description of, 8287 spinor field, 124
energy-momentum tensor of, 85, 86 vector field, 183
evolution operator of, 127 vector field, 145212
frequency decompositions, 93100 angular momentum density of, 151
Lagrangian of, 131142 charge current density of, 151, 152
charge symmetric in Heisenberg charge operator of, 165
picture, 134 after normal ordering, 184
charge symmetric in momentum creation and annihilation operators
picture, 134 for, 160163
standard in Heisenberg picture, 83 description of, 147153
standard in momentum picture, 83 dynamical variables of, 200203
momentum operator of, 101 energy-momentum tensor of, 151,
after normal ordering, 124 152
orbital angular momentum operator equations of motion
in Heisenberg picture, 102104, analysis of, 153158
107108 in terms of creation and annihila-
orbital operator of tion operators, 173
after normal ordering, 124 frequency decompositions, 158160
spin angular momentum density, 85, Lagrangian of, 198204
86 charge symmetric in Heisenberg
spin angular momentum of picture, 199
in Heisenberg picture, 102107 charge symmetric in momentum
spin angular momentum vector, 85 picture, 199
Index 289
standard in Heisenberg picture,

147148
standard in momentum picture,
149
massless, 172198
analysis of field equations without
Lorenz condition, 205208
description without Lorenz condi-
tion, 204205
dynamical variables without
equations of motion of, 151
equations of motion without
problems with, 190194
momentum operator of, 164
orbital angular momentum density
of, 151, 152
orbital angular momentum operator
of, 165
spin angular momentum density of,
151, 152
spin angular momentum operator of,
166170
vacuum of, 183

Quantum Field Theory PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quantum Field Theory PDF

Uploaded by

Copyright:

Available Formats

L AGRANGIAN Q UANTUM F IELD T HEORY

Nova Science Publishers, Inc.

NOTICE TO THE READER

Library of Congress Cataloging-in-Publication Data

Published by Nova Science Publishers, Inc.  New York

I Lagrangian Formalism and the Momentum Picture 1

II Free Scalar Fields 23

II.A Neutral scalar fields 25

II.B Arbitrary scalar fields 47

III Free Spinor Fields 81

IV Free Vector Fields 145

V Commutation Relations for Free Fields 213

4 Heisenberg Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

2000 MSC numbers: 2003 PACS numbers:

Einsteins summation convention: in a product of quantities or in a single expression,

Minkowski spacetime and indices. The Minkowski spacetime is denoted by M. The

The imaginary unit is denoted by i.

The Hermitian conjugation, transposition and complex conjugation of operators or

The composition of mappings/operators is marked by the symbol .

Kronecker and Dirac deltas. By f g , or f or f g (:= 1 for f = g, := 0 for f = g)

is denoted the Kronecker -symbol, depending on arguments f and g, and n (y), y Rn ,

The derivatives with respect to operator-valued (non-commuting) arguments will be

The commutator or anticommutator of operators (in their common domain) or matrices

i. In Heisenberg picture is fixed a (second) non-quantized and non-normally ordered op-

Bozhidar Zakhariev Iliev, Laboratory of Mathematical Modeling in Physics,

his e-mail address is bozho@inrne.bas.bg and http://theo.inrne.bas.bg/bozho/ is the

Lagrangian Formalism and

Some features of the Lagrangian formal-

L ( j (x), l (x)) L ( j (x), l (x))

and satisfy the conservation laws

which, in view of (2.5)(2.8), are equivalent to

3. Derivatives Relative to Operators

Definition 3.1. Let F be a C-vector space, {F F } be a subset of the

respect to u, i.e. the derivative in the r.h.s. of (3.2) should be calculated as if u1 , . . ., un

is exactly the definition of a derivative of a function of non-commuting arguments ac-

only for those operators v which commute with g.

Here: qi = const is the charge of the ith field, ( i ) = 0 if i = i , ( i ) = 1 if i =

5. The Momentum Picture of Motion

where x0 M is arbitrarily fixed and x M, is unitary, i.e.

U (x0, x) := ( U (x, x0)) = ( U (x, x0))1 =: U 1(x, x0 ). (5.2)

Let X F be a state vector in the systems Hilbert space F and A (x) : F F be an

evidently, preserve the Hermitian scalar product |  : F F C of F and the mean

X | A (x)( Y ) = X (x)| A (x)( Y (x)) (5.5)

for any X , Y F and A (x) : F F . Since the physically predictable/measurable results

from Heisenberg to momentum picture. Here [ A , B ] := A B B A is the commutator

where ( )n := ( ) ( ) (n-times) and ( )0 := idF is the identity mapping of F ,

so that (5.7) and (5.6) take respectively the form

by virtue of (5.4), we have

A (x) = A (x) + [ U (x, x0), A (x)] U 1(x, x0), (5.15)

[ P (x), P (x)] + P (x) P(x) = 0. (5.12)

[i (x), P i H ] = i i (x), (5.19)

where i (x) are the field operators in momentum picture, i.e.

i (x) = U (x, x0 ) i (x) U 1 (x, x0 ) = i (x0 ) = i (x0 ) =: i . (5.22)

being considered as initial value of X (x) at x = x0 .

6. General Aspects of Lagrangian Formalism

Thus, the Lagrangian (2.1) in momentum picture reads

(x) = U (x, x0 ) U 1 (x, x0 ) = (x0 ) = (x0 ) =: (6.4)

which, in view of (6.3), L = m2 c4 , and yL = c2 2 y is a solution of

This is the Klein-Gordon equation in momentum picture, which is considered in detail

0), can be rewritten as

M = L(x) + S(x) (6.11)

where x := (x1 , x2 , x3 ) = (x1 , x2 , x3 ), denotes the Euclidean cross product, and P :=

[ i (x), Q ] = ( i )qi i (x) (6.13)

Published by Nova Science Publishers, Inc. New York

evidently, preserve the Hermitian scalar product | : F F C of F and the mean

X | A (x)( Y ) = X (x)| A (x)( Y (x)) (5.5)