κ-deformed scalar field: National Centre for Nuclear Research

National Centre for Nuclear Research
Doctoral Thesis
κ-deformed scalar field

arXiv:2311.00014v1 [hep-th] 31 Oct 2023
Author: Supervisors:
Andrea Bevilacqua Jerzy Kowalski-Glikman
Wojciech Wiślicki
A thesis submitted in fulfillment of the requirements

for the degree of Doctor of Philosophy
in the
Department BP2
2 listopada 2023
iii
Declaration of Authorship
I, Andrea Bevilacqua, declare that this thesis titled, „κ-deformed scalar field” and
the work presented in it are my own. I confirm that:
• This work was done wholly or mainly while in candidature for a research degree
at the National Centre for Nuclear Research.
• Where any part of this thesis has previously been submitted for a degree or any
other qualification at the National Centre for Nuclear Research or any other
institution, this has been clearly stated.
• Where I have consulted the published work of others, this is always clearly
attributed.
• Where I have quoted from the work of others, the source is always given. With
the exception of such quotations, this thesis is entirely my own work.
• I have acknowledged all main sources of help.
• Where the thesis is based on work done by myself jointly with others, I have
made clear exactly what was done by others and what I have contributed myself.
Signed:
Date:
v
NATIONAL CENTRE FOR NUCLEAR RESEARCH
Streszczenie
Andrea Bevilacqua
In the following work we will introduce and discuss in detail a particular model of
complex κ-deformed scalar field, whose behaviour under C, P , T transformation is
particularly transparent from both a formal and phenomenological point of view. We
will begin by introducing the key mathematical structure at the basis of our investiga-
tion, namely the κ-Poincaré (Hopf) algebra and the κ-Minkowski spacetime. We will
then investigate the behaviour of general two-particle states under deformed boost.
After this we will introduce the action of our κ-deformed complex scalar field. From
it, we will derive the equations of motion, as well as the Noether charges due to the
continuous symmetries. The peculiar features of κ-deformation in general, and of our
model in particular, allow for very non-trivial interaction between discrete and con-
tinuous symmetries, of which we will investigate the phenomenological consequences
(particularly in terms of difference of lifetime of decaying particles). To conclude, we
will obtain the κ-deformed propagator of the κ-deformed complex scalar field, and the
imaginary part of the 1-loop contribution to it, ending with additional phenomenolo-
gical consequences. The third chapter is new, unpublished work.
NATIONAL CENTRE FOR NUCLEAR RESEARCH
Streszczenie
Andrea Bevilacqua
W niniejszej pracy wprowadzimy i przedyskutujemy w szczegółach model zespolonego,

κ-zdeformowanego pola skalarnego, którego zachowanie pod wpływem transformacji
C, P i T można jasno przedstawić, zarówno z formalnego, jak z fenomenologicznego
punktu widzenia. Zaczniemy od wprowadzenia struktur matematycznych, leżących u
podstaw naszych badań, a mianowicie algebry κ-Poincarego (Hopfa) oraz czasoprze-
strzeni κ-Minkowskiego. Następnie zbadamy zachowanie ogólnych stanów dwucząst-
kowych pod wpływem zdeformowanego pchnięcia Lorentza. Potem wprowadzimy dzia-
łanie κ-zdeformowanego, zespolonego pola skalarnego, po czym wyprowadzimy rów-
nania ruchu i ładunki Noether odpowiadające symetriom ciągłym. Pewne szczególne
własności κ-deformacji, zarówno te ogólne, jak i dotyczące tylko naszego modelu,
umożliwiają na istnienie nietrywialnch związków między symetriami dyskretnymi i
ciągłymi, z których wyprowadzimy konsekwencje fenomenologiczne i poddamy je ba-
daniom (szczególnie konsekwencje dotyczące różnic czasów życia niestabilnych cząstek
i antycząstek). Obliczymy także κ-zdeformowany propagator κ-zdeformowanego, ze-
spolonego pola skalarnego, jak również część urojoną jednopętlowej poprawki do pro-
pagatora, a następnie zakończymy dyskusją konsekwencji fenomenologicznych. Trzeci
rozdział to nowa, nieopublikowana praca.
ix
Acknowledgements
I would like to thank my supervisors Jurek and Wojtek. Their deep knowledge
and understanding of physics has been the source of many interesting discussions,
and their guidance has made my Ph.D. studies a wonderful experience. For the same
reasons, I would also like to thank Giulia, Michele, Giacomo, and Josua. I would like
to thank Prof. Stanisław Mrówczyński for the interesting discussions, and for giving
me the opportunity to assist him as a tutor in his Statistical Mechanics lectures. I am
also deeply grateful to Przemek for his help in making these four years possible in the
first place. I would also like to thank Monika for all the incredible help, and for the
very interesting discussions during my frequent trips to her office, as well as for the
enjoyable Polish lectures. Additionally, I would like to thank Ubaldo and Anatolii for
the wonderful time spent together. A special thanks goes also to my family for their
help, I wouldn’t have been able to study for so many years without their support.
Finally, my partner Alice has accompanied me throughout all my studies, in a
voyage which has been made even more beautiful (and much more chaotic) by the
birth of our son Levi. To both of them goes all my gratitude.
Part of this work has been supported by funds provided by the Polish National
Science Center, the project number 2019/33/B/ST2/00050, and by the NAWA scho-
larship financed by the STER program. My participation to COST Action CA18108
activities has been financed by COST Action. Part of this work’s bibliography has
been collected using (1).
xi
Spis treści
Declaration of Authorship iii
Streszczenie v
Streszczenie vii
Acknowledgements ix
1 κ-Minkowski and κ-Poincaré 1

1.1 Introduction to κ-deformation . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Formal definition of Hopf algebra . . . . . . . . . . . . . . . . . . . . . 4
1.3 From κ-Poincaré to κ-Minkowski . . . . . . . . . . . . . . . . . . . . . 5
1.4 From κ-Minkowski to κ-Poincaré . . . . . . . . . . . . . . . . . . . . . 9
1.5 Finite boosts for two-particle states . . . . . . . . . . . . . . . . . . . . 16
1.5.1 Fifth column: . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.2 Fourth column: . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5.3 Third column: . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.5.4 Second column: . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.5 First column: . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.5.6 Expansion of the final matrix up to 1/κ . . . . . . . . . . . . . 21
1.5.7 Check of the consistency of the first order expansion of L′g . . . 23
1.6 Two-particle kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2 Complex scalar field on κ-Minkowski spacetime 31

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.2 Weyl maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.3 Action for the complex scalar field . . . . . . . . . . . . . . . . . . . . 34
2.4 Action of derivatives on ⋆ products . . . . . . . . . . . . . . . . . . . . 35
2.4.1 A = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.4.2 A = i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.4.3 A = 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.4.4 A = + . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4.5 ∆†+ = κ2 ∆−1 + . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.5 Integration-by-parts relation for ∂i going ← . . . . . . . . . . . . . . . 38
2.6 Integration-by-parts relation for ∂0 going ← . . . . . . . . . . . . . . . 39
2.7 Integration-by-parts relation for ∂i going → . . . . . . . . . . . . . . . 39
2.8 Integration-by-parts relation for ∂0 going → . . . . . . . . . . . . . . . 40
2.9 Mass terms integration by parts . . . . . . . . . . . . . . . . . . . . . . 41
2.10 EoM and surface terms Π0 , Πi , and Π4 for both orderings of the action 42
2.11 Complex scalar field and its properties . . . . . . . . . . . . . . . . . . 43
2.11.1 C, P, T in the non-deformed context: a short review . . . . . . . 45
2.11.2 C, P, T in the κ-deformed context . . . . . . . . . . . . . . . . . 46
2.12 Off-shell action in momentum space . . . . . . . . . . . . . . . . . . . 47
xii
2.12.1 First ordering of the action . . . . . . . . . . . . . . . . . . . . 47

2.12.2 Second ordering of the action . . . . . . . . . . . . . . . . . . . 48
2.12.3 Sum of the two actions . . . . . . . . . . . . . . . . . . . . . . . 49
2.13 Symplectic form for the two actions . . . . . . . . . . . . . . . . . . . . 49
2.13.1 Example in the non-deformed case . . . . . . . . . . . . . . . . 49
2.13.2 The κ-deformed case . . . . . . . . . . . . . . . . . . . . . . . . 50
2.14 Conserved charges: geometric approach . . . . . . . . . . . . . . . . . . 54
2.14.1 Non-deformed case . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.14.1.1 Translation charges . . . . . . . . . . . . . . . . . . . 54
2.14.1.2 Boost and rotation charges . . . . . . . . . . . . . . . 55
2.14.1.2.1 Crucial step . . . . . . . . . . . . . . . . . . . 58
2.14.2 κ-deformed case . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.14.2.1 Translation charges . . . . . . . . . . . . . . . . . . . 58
2.14.2.1.1 Non-deformed contraction: . . . . . . . . . . 59
2.14.2.1.2 Deformed contraction: . . . . . . . . . . . . . 60
2.14.2.2 Notes on the above (overall) procedure . . . . . . . . 61
2.14.2.3 Boosts . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.14.2.3.1 Crucial step . . . . . . . . . . . . . . . . . . . 64
2.14.2.3.2 The (very) peculiar features of boosts . . . . 66
2.14.2.3.3 The role of ζ(p) . . . . . . . . . . . . . . . . 67
2.14.2.4 Rotations . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.14.3 Is it δA ∧ δB = −δB ∧ δA, or δA ∧ δB = S(δB ∧ δA)? . . . . . 69
2.14.4 And what if we add contractions with vector fields? . . . . . . . 70
2.15 The algebra of the κ-deformed charges . . . . . . . . . . . . . . . . . . 74
2.15.0.1 [Ni , Pj ] = −iηij P0 . . . . . . . . . . . . . . . . . . . 75
2.15.0.2 [Ni , Nj ] = −iϵijk Mk . . . . . . . . . . . . . . . . . . . 76
2.16 C and its commutator with the boosts . . . . . . . . . . . . . . . . . . 80
2.17 CP
T ⇔ Lorentz?
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
2.17.1 Greenberg’s argument . . . . . . . . . . . . . . . . . . . . . . . 84
2.17.2 Jost theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
2.17.3 Comments on Jost and Greenberg theorems . . . . . . . . . . . 88
2.17.3.1 CP T properties of ⟨0|ϕ†1 (x1 ) ⋆ ϕ2 (x2 )|0⟩ . . . . . . . . 89
2.17.3.2 Lorentz properties of ⟨0|ϕ†1 (x1 ) ⋆ ϕ2 (x2 )|0⟩ . . . . . . . 91
2.18 Phenomenological consequences of deformed CP T transformations . . 93
2.18.1 Estimation of deformed decay width . . . . . . . . . . . . . . . 96
3 κ-deformed propagator, and 1-loop correction to it 101

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.2 Feynman propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.2.1 From the generating functional . . . . . . . . . . . . . . . . . . 101
3.2.2 As inverse of the field operator . . . . . . . . . . . . . . . . . . 104
3.3 Example of loop correction to the propagator: non-deformed case . . . 105
3.3.1 Using renormalization . . . . . . . . . . . . . . . . . . . . . . . 105
3.4 Example of loop correction to the propagator: κ- deformed case . . . . 108
3.4.1 Dimensional regularization and cut-off in the κ-deformed context108
3.4.1.1 Summary of the individual diagrams and total 1-loop
amplitude . . . . . . . . . . . . . . . . . . . . . . . . . 125
3.5 Mass distribution width . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
xiii
To A.B. and L.B-B.

1
1
κ-Minkowski and κ-Poincaré
1.1 Introduction to κ-deformation

Immediately after both general relativity and quantum mechanics were formalized,
physicists were already trying to understand what a quantum theory of gravity wo-
uld look like (2). Several approaches were tried, each with their pros and cons (an
extensive discussion of each approach would be out of place here, but one can con-
sult (3) for a review). From a theoretical point of view, the wealth of approaches was
due to the vastly different nature of general relativity and quantum mechanics, both
conceptually and formally. Moreover, from the more pragmatic point of view, effects
of quantum gravitational origin were thought to appear when particles with Plank
energy Ep scatter with impact parameter given by the Plank length lp . This means
that very early on, there was significant doubt that any quantum gravitational effect
could be phenomenologically relevant, even in the distant future (4). Moreover, the
measurement of phenomena happening at distances comparable or smaller than lp
would require apparatuses so large that they inevitably collapse to a black hole (5),
which means there may be some limitation from first principles in the conventional
definition of spacetime as a continuous structure (this was already noted in (4), (6),
(7)). Despite these advances, however, the lack of phenomenological feedback seemed
to be destined to be a long lasting issue in quantum gravity (QG) research. It was
therefore unexpected when it was suggested that, switching the focus to cumulative
effects of QG, a phenomenology was not only possible, but within reach (8). Indeed,
instead of focussing on high energy scattering processes with very small impact pa-
rameter, one could try to study the cumulative effects of QG, which can for example
manifest themselves with a slightly different propagation speed for photons of different
energies (9), (10), (53), (12), (13), (35), and may also affect neutrino properties (15),
(16), (17), (18), (19), (20). One of the most effective ways to go about studying po-
tential QG effects is therefore not to study any model of QG directly, but a potential
effective theory of QG.
One of the most studied model in this sense is based on a deformation of the
Poincaré algebra, called κ-Poincaré algebra, where κ is a deformation parameter with
dimensions of energy. Its origin can be traced back to the 1990’s, particularly to (21),
(22), (23), (24), (25), (26). The idea is to deform the canonical Poincaré algebra to
investigate how symmetries might change in some well defined physical limit of QG.
At the same time, new models were being proposed where in addition to a constant
speed c, there is also a length scale l = κℏ (or l = κ1 if ℏ = 1) which is independent of
the observer (27), (28), (29), (30). This length scale also implies a momentum scale in
2 Rozdział 1. κ-Minkowski and κ-Poincaré
momentum space; the idea is therefore that such observer-independent scale can be
used to define some non-linear dispersion relation while keeping the theory Lorentz
invariant (where ‘Lorentz invariance’ in this context is defined in the appropriate way).
Intuitively speaking, such an additional scale would allow for a non-trivial geometry
of momentum space, which is usually flat in non-deformed theories which lack such a
scale. To give just a small example, one can start from the action of a classical particle
with non-deformed dispersion relation moving in a curved spacetime (75), (76)
Z
S = dτ ẋµ eµa (x)pa − N (η ab pa pb + m2 ). (1.1)
From this, one can by analogy write down the action of a particle moving in flat
spacetime, but curved momentum space
Z
Sκ = − dτ xa Ea µ (k, κ)k̇µ − N (Cκ (k) + m2 ). (1.2)
where Ea µ is the tetrad which is related to the momentum-space metric by
Gµν (k, κ) = η ab Ea µ (k, κ)Eb ν (k, κ), (1.3)
and Cκ (k) is the deformed dispersion relation (once again, here we are using a defor-
mation parameter with dimensions of energy). For example, one could choose

k0 k0
2
Cκ (k) = −4κ sinh 2
+ k2 e κ . (1.4)
2κ
One can show that

Z
k0
Sκ = dτ ẋ0 k0 − e κ x · k̇ + N (Cκ (k) + m2 ) (1.5)
which is invariant under a deformed set of transformations which satisfy the so called
κ-Poincaré algebra (which will be discussed in detail in later sections). The symplectic
form is then given by
xi k0
Z
k0
Ωκ = dτ δk0 ∧ δx0 + e κ δki ∧ δxi − e κ δk0 ∧ δki (1.6)
κ
which implies
k0 1
{x0 , p0 } = 1 {xi , kj } = e− κ δji {x0 , xi } = − xi . (1.7)
κ
We see that, upon quantization, spacetime coordinates do not commute. These type of
non-commutativity defines the so called κ-Minkowski spacetime. This kind of models
showcases interesting phenomena, such as relative locality (33), (34), (35), and are
related to κ-Poincaré algebras (36), (65), (66), (67).
Finally, in 2 + 1 dimensions, one can show that κ-deformation naturally emerges
when considering point-like particles coupled with gravity (see for example (38),(37),
(75) and references therein). This is due to the fact that the Newton constant GN is
dimensionful in 2 + 1 dimensions, and it has dimensions of energy. The procedure is
fairly straightforward and can be summarized in the following steps (76),(75):
1) From the total action of gravity and point-like particles coupled to it, one can
get the explicit solutions of the (finitely many) gravitational degrees of freedom;
1.1. Introduction to κ-deformation 3
2) Substitute these solutions back into the action, obtaining a new action which
describes the motion of point-like particles ‘dressed’ in their own gravitational
field;
3) This new action can then be shown to correspond to a κ-deformed model, much
like the one described by eq. (1.5) (although in the case of 2 + 1 particles with
gravity we have a different kind of deformations).
Furthermore, one can argue that the deformed symmetry group P 3 acting on the
Hilbert space H 3 of 2 + 1 gravity with particles is expected to be a subgroup of the
full group P 4 acting on the Hilbert space H 4 of particles coupled with gravity in 3+1
dimensions (38). All this motivates the study of fields in the context of κ-Poincaré
and κ-Minkowski.
κ-deformation models have also proven to be interesting from the phenomenolo-
gical point of view, ranging from symmetry violation, to string theory, astronomy,
particle physics, and black holes (51), (52), (53), (54), (55), (56), (57), (58).
In this work, we will investigate the kinematical properties of particles in κ-
Minkowski spacetime (chapter 1), and we will then introduce the κ-deformed complex
scalar field, studying its properties and computing the Noether charges (chapter 2).
Finally, we will compute the propagator and the imaginary part of the 1-loop correc-
tion to it (chapter 3). Each of the chapters will contain a final section investigating
possible experimental signature of the theoretical results obtained earlier in the chap-
ter. There have been already several approaches building fields on top of κ-Minkowski
spacetime, or using κ-Minkowski algebra in general, see for example (39), (86), (40),
(41), (42), (43), (44), (45), (46), (47), (48), (80), (49), (50). In contrast to these mo-
dels, the construction which we will present in this work allows for a much simpler,
canonical definition of deformed symmetries in the κ-deformed context. The particu-
larly simple form of the discrete symmetries in turn allows for an interesting (and easy
to interpret) phenomenology.
Of course, key ingredients in all our constructions are the κ-Minkowski spacetime,
and the κ-Poincaré algebra, so we will dedicate the first two section of the current
chapter to a brief exposition of their properties. It turns out, however, that it is not
just the deformed κ-Poincaré algebra to be important, but the Hopf algebra struc-
ture of which the κ-Poincaré algebra is but a part (61). The mathematical aspects
of κ-deformations and the importance of the relevant Hopf algebra have been in-
vestigated in detail in (62), (63), (64), (65), (68), (69), and the important physical
ultra-relativistic and non-relativistic limits in (71), (70), (72). The constructions of
κ-Minkowski spacetime and the κ-Poincaré Hopf algebra, as well as the relations be-
tween them, have been extensively studied for decades. Here we will highlight how
one can obtain κ-Minkowski spacetime starting from the κ-Poincaré Hopf algebra fol-
lowing (59), (37) and references therein as a guide. We will then show the opposite
direction, namely how one can obtain the κ-Poincaré Hopf algebra structure starting
from κ-Minkowski spacetime, following the approach presented in (60). Both the ap-
proaches can be useful when working in the context of κ-deformation, and each of
them sheds light on some feature of κ-deformation in general. Since these notions
are well known in the literature, we will not delve into too much details, referring
the reader to the provided references for a deeper treatment. Nevertheless, we will be
explicit in out computations, and show each important step.
1.2 Formal definition of Hopf algebra

Before beginning our discussion, we give a mathematical definition of a Hopf algebra.
A Hopf algebra is technically defined as a bi-algebra with an antipode, satisfying
several compatibility conditions. To understand what is a bi-algebra with antipode,
we will need to introduce the notion of algebra, co-algebra, and antipode separately.
Using the same notation as (73), we can define a unital algebra as a vector field A
over a field K together with two maps, the product map µ : A ⊗ A → A , and the
unit map η : K → A , which need to satisfy the associativity and unity axiom defined
respectively by the following commutative diagrams (73)
id⊗µ µ
A ⊗A ⊗A A ⊗A A A ⊗A
µ⊗id µ µ 1⊗η
A ⊗A µ A A ⊗A η⊗1
A
A co-algebra can basically be defined by reversing all the arrows. In particular, a

co-algebra is a vector field C over a field K together with two maps, the co-product
map ∆ : C → C ⊗ C and the co-unit map ϵ : C → K, which need to satisfy the
co-associativity and co-unit axiom, which are defined respectively by the following
commutative diagrams (73)
id⊗∆ ∆
C ⊗C ⊗C C ⊗C C C ⊗C
∆⊗id ∆ ∆ 1⊗ϵ
C ⊗C ∆
C C ⊗C ϵ⊗1
C
A bi-algebra B is therefore both an algebra and a co-algebra, with additional

compatibility conditions. In fact, we now have a structure with product µ, unity η,
co-product ∆, and co-unity ϵ, and they must satisfy (73)
∆(µ(a, b)) = µ(∆(a), ∆(b)), ϵ(µ(a, b)) = ϵ(a)ϵ(b) (1.8)
for all a, b ∈ B, and where
∆(id) = id ⊗ id ϵ(id) = 1 (1.9)
where 1 ∈ K. Finally, a Hopf algebra H is a bi-algebra with an additional map

S : H → H such that
S(µ(a, b)) = µ(S(a), S(b)) S(1) = 1 (S ⊗ S) ◦ ∆ = τ ◦ ∆ ◦ S (1.10)
where a, b ∈ H and τ : H ⊗ H → H ⊗ H is the flip map defined by
τ (a ⊗ b) = b ⊗ a (1.11)
1.3. From κ-Poincaré to κ-Minkowski 5
for all a ∈ H .
In what follows we will concentrate on the κ-Poincaré Hopf algebra. All the above
equalities and commutative diagrams are satisfied by the product, co-product, unity,
co-unity, and antipode, but we will not check them explicitly here. Our aim with the
discussion in this section and the next is to highlight how the κ-Poincaré Hopf algebra,
and more in particular its co-product sector, is related to the κ-Minkowski spacetime.
1.3 From κ-Poincaré to κ-Minkowski

The algebra sector of the κ-Poincaré Hopf algebra in the so called bicrossproduct basis
is given by
[Mi , kj ] = iϵijl kl , [Mi , k0 ] = 0, (1.12)
k2

κ i
[Ni , kj ] = iδij (1 − e−2k0 /κ ) + − ki kj , [Ni , k0 ] = iki , (1.13)
2 2κ κ
[Mi , Mj ] = iϵijk Mk , [Mi , Nj ] = iϵijk Nk [Ni , Nj ] = −iϵijk Mk . (1.14)
The co-algebra sector is given by
∆ki = ki ⊗ 1 + e−k0 /κ ⊗ ki , ∆k0 = k0 ⊗ 1 + 1 ⊗ k0 , (1.15)
1
∆Ni = Ni ⊗ 1 + e−k0 /κ ⊗ Ni + ϵijk kj ⊗ Mk , (1.16)
κ
∆Mi = Mi ⊗ 1 + 1 ⊗ Mi (1.17)
and finally the antipode sector is given by
S(k0 ) = −k0 , S(ki ) = −ki ek0 /κ , (1.18)

1
S(Mi ) = −Mi , S(Ni ) = −e k0 /κ
Ni − ϵijk kj Mk . (1.19)
κ
The co-units ϵ(k), ϵ(Ni ), ϵ(Mj ) are all zero. Starting from this Hopf algebra, and
in particular from the co-algebra sector, we can build the spacetime associated to
the algebra. The construction is quite intuitive, because one first introduces the ob-
jects xµ (interpreted as ‘coordinates’, the dual objects of the momenta kµ ), and then
one defines the action of the algebra generators kµ , Ni , Mj on them. Having done
this, the co-algebra sector is used to define the action of kµ , Ni , Mj on products xµ xν ,
and one immediately gets the coordinates commutation relation defining κ-Minkowski
spacetime from the action of kα on xµ xν (or more in general on polynomials of xµ ).
Furthermore, one can also check that such commutation relations are invariant under
the action of Ni , Mj , showing that the Poincaré algebra indeed describes the sym-
metries of κ-Minkowski spacetime. More in detail, using the symbol ▷ to indicate the
action of one of the generators of the κ-Poincaré algebra on spacetime variables, we
have the definitions (37)
kµ ▷ xν = −iδµν , Ni ▷ x̂j = ix̂0 δij , Ni ▷ x̂0 = ixi , kµ ▷ 1 = ϵ(kµ ) = 0

(1.20)
We then use the co-products to define the actions on polynomials of xµ using the so
called Sweedler notation. Writing everything explicitly, we have
X
∆f ▷ g1 g2 = (∆f )(g1 , g2 ) = (fα(1) ▷ g1 )(fα(2) ▷ g2 ) (1.21)
α
where g1 , g2 are some coordinates. The objects with (1) and (2) represent respectively
the ‘first’ and ‘second’ terms in each of the tensor products in the definition of co-
product in eq. (1.15), (1.16), (1.17). To make things more concrete, take for example
ki whose co-product rule is given in eq. (1.15). We have
X
ki ▷ x̂0 x̂j = ((ki )α(1) ▷ x̂0 )((ki )α(2) ▷ x̂j ) (1.22)
α
= ((ki )1(1) ▷ x̂0 )((ki )1(2) ▷ x̂j ) + ((ki )2(1) ▷ x̂0 )((ki )2(2) ▷ x̂j ) (1.23)
= (ki ▷ x̂0 )(1 ▷ x̂j ) + (e−k0 /κ ▷ x̂0 )(ki ▷ x̂j ) (1.24)
k2

k0
= −iδi0 x̂j + 1 − + 02 − . . . ▷ x̂0 (−iδij ) (1.25)
κ 2κ
(k0 ▷ x̂ ) j k0 ▷ (k0 ▷ x̂0 ) j
0
= 0 − ix0 δij + iδi − iδi + . . . (1.26)
κ 2κ2
1
= −ix̂0 δij + δij + 0 − . . . (1.27)
κ
1
= −ix0 δi + δij
j
(1.28)
κ
Notice that we also used the relations k 2 ▷ x = k ▷ (k ▷ x), and recall that ϵ(k) = 0 in
the κ-Poincaré Hopf algebra. In the same way, we also have
X
ki ▷ x̂j x̂0 = ((ki )α(1) ▷ x̂j )((ki )α(2) ▷ x̂0 ) (1.29)
α
= ((ki )1(1) ▷ x̂j )((ki )1(2) ▷ x̂0 ) + ((ki )2(1) ▷ x̂j )((ki )2(2) ▷ x̂0 ) (1.30)
= (ki ▷ x̂j )(1 ▷ x̂0 ) + (e−k0 /κ ▷ x̂j )(ki ▷ x̂0 ) (1.31)
= −ix̂0 δij . (1.32)
X
k0 ▷ x̂0 x̂j = ((k0 )α(1) ▷ x̂0 )((k0 )α(2) ▷ x̂j ) (1.33)
α
= ((k0 )1(1) ▷ x̂0 )((k0 )1(2) ▷ x̂j ) + ((k0 )2(1) ▷ x̂0 )((k0 )2(2) ▷ x̂j ) (1.34)
= (k0 ▷ x̂0 )(1 ▷ x̂j ) + (1 ▷ x̂0 )(k0 ▷ x̂j ) (1.35)
= −ix̂ j
(1.36)
X
k0 ▷ x̂j x̂0 = ((k0 )α(1) ▷ x̂j )((k0 )α(2) ▷ x̂0 ) (1.37)
α
= ((k0 )1(1) ▷ x̂j )((k0 )1(2) ▷ x̂0 ) + ((k0 )2(1) ▷ x̂j )((k0 )2(2) ▷ x̂0 ) (1.38)
j 0 j
= (k0 ▷ x̂ )(1 ▷ x̂ ) + (1 ▷ x̂ )(k0 ▷ x̂ ) 0
(1.39)
1.3. From κ-Poincaré to κ-Minkowski 7
= −ix̂j (1.40)
Furthermore, one can also obtain the relation

X
ki ▷ x̂j x̂k = ((ki )α(1) ▷ x̂j )((ki )α(2) ▷ x̂k ) (1.41)
α
= ((ki )1(1) ▷ x̂j )((ki )1(2) ▷ x̂k ) + ((ki )2(1) ▷ x̂j )((ki )2(2) ▷ x̂k ) (1.42)
= (ki ▷ x̂j )(1 ▷ x̂k ) + (e−k0 /κ ▷ x̂j )(ki ▷ x̂k ) (1.43)
k02

j k k0
= −iδi x̂ + 1 − + 2 − . . . ▷ x̂ (−iδik )
j
(1.44)
κ 2κ
= −iδij x̂k − iδik x̂j (1.45)
One can therefore deduce
δij
ki ▷ x̂0 x̂j − ki ▷ x̂j x̂0 = k0 ▷ x̂0 x̂j − k0 ▷ x̂j x̂0 = 0 (1.46)
κ
ki ▷ x̂j x̂k − ki ▷ x̂k x̂j = 0 (1.47)
and therefore
i j
[x̂0 , x̂j ] = x̂ [x̂i , x̂j ] = 0 (1.48)
κ
which are the commutation relations for the coordinates defining κ-Minkowski spa-
cetime. For the second commutator, notice that we used k0 ▷ x̂i = 0. This algebra is
called an(3) algebra.
One can then show that the commutators (1.48) is invariant under boosts and
rotations. The invariance under rotations is trivial since, as can be seen from eq.
(1.12), (1.14), (1.17), (1.19) the rotation sector is completely not deformed in the κ-
Poincaré Hopf algebra, and so it acts as usual. For the boosts, using eq. (1.20), as well
as the previous computations one gets (61)
1
Ni ▷ x̂0 x̂j = (Ni ▷ x̂0 )(1 ▷ x̂j ) + (e−k0 /κ ▷ x̂0 )(Ni ▷ x̂j ) + ϵijk (kj ▷ x̂0 )(Mk ▷ x̂j )
κ
k2

k0
= ixi x̂j + 1 − + 02 − . . . ▷ x̂0 (ix̂0 δij ) + 0 (1.49)
κ 2κ
x̂0 δij
j
= ixi x̂ + i(x̂0 )2 δij − (1.50)
κ
1
Ni ▷ x̂j x̂0 = (Ni ▷ x̂j )(1 ▷ x̂0 ) + (e−k0 /κ ▷ x̂j )(Ni ▷ x̂0 ) + ϵijk (kj ▷ x̂j )(Mk ▷ x̂0 )
κ
= ix̂0 δij x̂0 + ix̂j xi + 0 (1.51)
= i(x̂0 )2 δij + ix̂j xi (1.52)
and therefore
ix̂j ix̂j
Ni ▷ [x̂0 , x̂j ] = Ni ▷ + i[xi , x̂j ] = Ni ▷ (1.53)
κ κ
which shows that indeed the commutation relation (1.48) are invariant under κ-
deformed boosts. Notice that we used the fact that Mk ▷ x̂0 = 0 (spatial rotations
do not act on the time direction) and eq. (1.48).
Although we used the bicrossproduct basis for the computations up to now, this
is by no means the only choice. In particular, another choice of coordinates for mo-
mentum space is the so called classical basis, which is related to the bicrossproduct
one by the following coordinate change
k0 k2 k0 k0
P0 (k0 , k) = κ sinh + eκ, Pi (k0 , k) = ki e κ (1.54)
κ 2κ
k0 k2 k0
P4 (k0 , k) = κ cosh − eκ. (1.55)
κ 2κ
One can check by direct computations (74), (75) that
−P02 + P2 + P42 = κ2 P+ := P0 + P4 > 0 P4 > 0. (1.56)
In this basis, the algebra sector of the κ-Poincaré Hopf algebra reads
[Mi , Pj ] = iϵijk Pk , [Mi , P0 ] = 0, [Ni , Pj ] = iδij P0 , [Ni , P0 ] = iPi , (1.57)
[Mi , Mj ] = iϵijk Mk , [Mi , Nj ] = iϵijk Nk , [Ni , Nj ] = iϵijk Mk . (1.58)
The co-algebra sector is given by

1
∆Pi = Pi ⊗ P+ + 1 ⊗ Pi , (1.59)
κ
1 X Pk κ
∆P0 = P0 ⊗ P+ + ⊗ Pk + ⊗ P0 , (1.60)
κ P+ P+
k
1 X Pk κ
∆P4 = P4 ⊗ P+ − ⊗ Pk − ⊗ P0 , (1.61)
κ P+ P+
k
κ Pj
∆Ni = Ni ⊗ 1 + ⊗ Ni + ϵijk ⊗ Mk , (1.62)
P+ P+
∆Mi = Mi ⊗ 1 + 1 ⊗ Mi , (1.63)
and the antipodes by
κPi P2
S(Pi ) = − , S(P0 ) = −P0 + , S(P4 ) = P4 , (1.64)
P+ P+
P+ 1
S(Mi ) = −Mi , S(Ni ) = −Ni + ϵijk Pj Mk . (1.65)
κ κ
Once again, the co-units ϵ(k), ϵ(Ni ), ϵ(Mj ) are all zero. Furthermore, notice also
1.4. From κ-Minkowski to κ-Poincaré 9
that rotations are non-deformed in both basis. A very interesting property of the
classical basis is the fact that the algebra coincides with the non-deformed Poincaré
algebra, so the totality of the effect of κ-deformation is found in the co-product and
antipode sector. This also shows that κ-deformation is important at the Hopf algebra
level, and not merely at the algebra level. Because of the construction of κ-Minkowski
showed above, the co-products cover a particularly important role since they allow
the definition of κ-Minkowski spacetime. More details can be found in (75), (59).
1.4 From κ-Minkowski to κ-Poincaré

The first step in going from κ-Minkowski to κ-Poincaré is to find an explicit represen-
tation of the an(3) algebra in eq. (1.48). We choose (with a slight abuse of notation,
we call the explicit matrices representing the algebra with the same names as the
elements of the algebra itself)
0 (ϵi ) T
   
0 0 1 0
i i
x̂0 = −  0T 0̃ 0T  x̂i =  ϵi 0̃ ϵi  . (1.66)
κ κ i T
1 0 0 0 −(ϵ ) 0
The central 0̃ is a 3 × 3 null matrix, while 0 = (0, 0, 0) and ϵi is a unit column vector,
such that (ϵ1 )T = (1, 0, 0), (ϵ2 )T = (0, 1, 0), and (ϵ1 )T = (0, 0, 1). We now compute
the group elements of AN (3)
i 0
êk = eiki x̂ eik0 x̂ . (1.67)
Notice that because of the non-trivial algebra we need to carefully define the expo-
nentiation, i.e. on how we decide to define the group element. Here we choose the so
j 0
called time-to-the-right convention, in which eikj x̂ is to the left of eik0 x̂ . Because of
this choice, the objects k0 , k in the exponent (which are dimensionally momenta) coin-
cide with the bicrossproduct basis introduced in the previous section. With a different
choice of group element, for example
µ
êq = eiqµ x̂ (1.68)
one would get different coordinates for momentum space (in this particular case, we
would be introducing the so called normal basis (75), but we will not address this
further). Furthermore, notice that the representation matrices are 5 × 5.
We have
 2n    2n+1  
0 0 1 1 0 0 0 0 1 0 0 1
0 0 0 = 0 0 0 0 0 0  = 0 0 0 (1.69)
1 0 0 0 0 1 1 0 0 1 0 0
and therefore
 n
X (ik0 x0 )n X 1 k0 n 0 0 1
x̂0
eik0 = = 0 0 0 (1.70)
n! n! κ
n n 1 0 0
cosh kκ0 k0 

0 sinh κ
 
(1.71)
 
=
 0 1 0 

 
sinh kκ0 0 cosh kκ0
Similarly one can check that

2  3
0 ϵT 0 ϵT
  
0 1 0 1 0
ϵ 0 ϵ =  0 0 0  ϵ 0 ϵ = 0 (1.72)
0 −ϵT 0 −1 0 −1 0 −ϵT 0
This behaviour is the reason for the ‘N ’ in the name ’AN (3)’, which stands for ‘nil-
potent’. Therefore
 
k2 kT k2
1 + 2κ 2 κ 2κ 2
i 1
eiki x̂ = 1 + iki x̂i + (iki x̂i )2 =  kκ 1 k
(1.73)
 
2 κ 
k 2 k T k 2
− 2κ2 − κ 1 − 2κ2 .
Multiplying these two matrices we get

k k0
k2 κ0 kT k2
 
cosh kκ0 + 2κ 2e κ sinh kκ0 + 2κ2
e κ
k k0
êk =  k κ0 k (1.74)
 
κe 1 κe
κ 
k0 k0
k2 kT k2
sinh kκ0 − 2κ2
e κ − κ cosh kκ0 − 2κ2
e κ
where we used the relation cosh x + sinh x = ex . Notice now that we can make the
following substitutions
k0 k2 k0 /κ
P0 (k0 , k) = κ sinh + e , (1.75)
κ 2κ
Pi (k0 , k) = ki ek0 /κ , (1.76)
k0 k2 k0 /κ
P4 (k0 , k) = κ cosh − e (1.77)
κ 2κ
to bring the above group element êk into the following form
 
P̃4 κP/P+ P0
1
êP (k) =  P κ × 13×3 P  (1.78)
κ
P̃0 −κP/P+ P4
where
P2 P2
P̃0 = P0 − = −S(P0 ) P̃4 = P4 + (1.79)
P+ P+
Incidentally, one can use the group property êP (k) êQ(l) = êP (k)⊕Q(l) and ê−1
P = êS(P )
to obtain the deformed antipode and sum in the classical basis.
P2 κ2
S(P0 ) = −P0 + = − P4 , (1.80)
P0 + P4 P0 + P4
κP
S(P) = − , S(P4 ) = P4 . (1.81)
P0 + P4
1 PQ κ
(P ⊕ Q)0 = P0 (Q0 + Q4 ) + + Q0 (1.82)
κ P0 + P4 P0 + P4
1
(P ⊕ Q)i = Pi (Q0 + Q4 ) + Qi (1.83)
κ
1 PQ κ
(P ⊕ Q)4 = P4 (Q0 + Q4 ) − − Q0 (1.84)
κ P0 + P4 P0 + P4
One can of course do the same thing in the bicrossproduct basis. Notice the relation-
ship between these equations and the co-products in eq. (1.59), (1.60), (1.61).
Notice that the momenta defined in eq. (1.75) are indeed again the classical basis
defined in eq. (1.54), (1.55). By construction, therefore, they satisfy the constraints in
eq. (1.56). Furthermore, notice that the group elements êP (k) are uniquely identified
once one knows the quantities P0 , Pi , P4 , which are the components on the last column
of the matrix (1.78). Indeed, calling O = (0, 0, 0, 0, κ)T , we have
êP (k) O = (P0 , P, P4 )T (1.85)
and the action of the group AN (3) on O is transitive, meaning that acting with all
g ∈ AN (3) on O reaches all the points of the manifold associated to the group. We
already know that relations (1.56) hold, which means that the 4-dimesional manifold
associated to AN (3) is half of de Sitter space, namely the half where P+ > 0. Notice
that the curvature of such a space is constant and equal to 1/κ2 . Furthermore, since
each point is uniquely described by the coordinates (P0 , P) (after fixing both P0 and
P, P4 is not a free parameter, and can be obtained through the constraint), the group
manifold AN (3) is therefore our momentum space manifold. This gives a clear picture
of how the non-commutative nature of spacetime coordinates is linked to a constant
curvature of momentum space. Notice that to obtain coordinates spanning the other
half of de Sitter space defined by the first constraint in eq. (1.56), one applies the
same group elements to the new origin O ∗ = (0, 0, 0, 0, −κ)T . Alternatively, one can
define a matrix
0
z = eπκx̂ (1.86)
which allows us to define
ê∗k = êk z (1.87)
and then apply these new objects ê∗ to the same origin O = (0, 0, 0, 0, κ). In both
cases, we obtain the coordinates
k0 k2 k0 /κ
P0∗ (k0 , k) = −κ sinh − e , (1.88)
κ 2κ
Pi∗ (k0 , k) = −ki ek0 /κ , (1.89)
k0 k2 k0 /κ
P4∗ (k0 , k) = −κ cosh + e . (1.90)
κ 2κ
These starred coordinates satisfy P+∗ = P0∗ + P4∗ < 0, and one can show that
S(P ∗ ) = −S(P ) (1.91)
and that the deformed sum is not modified if one or both of the momenta is starred.
In other words, we have
1 (∗) (∗) (∗) P(∗) Q(∗) κ (∗)

(P (∗) ⊕ Q(∗) )0 = P0 (Q0 + Q4 ) + (∗) (∗)
+ (∗) (∗)
Q0 (1.92)
κ P0 + P4 P0 + P4
1 (∗) (∗) (∗) (∗)
(P (∗) ⊕ Q(∗) )i = Pi (Q0 + Q4 ) + Qi (1.93)
κ
1 (∗) (∗) (∗) P(∗) Q(∗) κ (∗)

(P (∗) ⊕ Q(∗) )4 = P4 (Q0 + Q4 ) − (∗) (∗)
− (∗) (∗)
Q0 , (1.94)
κ P0 + P4 P0 + P4
where the (∗) indicates that either P or Q or both or neither can be starred momenta,
and the deformed sum formula works in the same way regardless (78).
We now address the issue of how one can get the κ-Poincaré structure out of the
AN (3) group, and in particular the co-product and antipode sectors. We will not treat
in detail all the coproducts and antipodes for all generators, and we will use either
the classical or the bicrossproduct basis depending on which is easier to work with.
The interested reader is referred back to (60) for more details.
We start with the coproduct and antipode of momenta. As was already discussed
in the previous section, coproducts describe the way the κ-Poincaré algebra act on
products, be it products of coordinates like before, or of states like we will consider
now. In particular, we have
∆kµ (|p⟩ ⊗ |q⟩) = (p ⊕ q)µ |p⟩ ⊗ |q⟩ (1.95)
where p ⊕ q is define through the group multiplication
êp êq = êp⊕q . (1.96)
In the same way, the antipode S(k)µ is related to inverse group elements by the
definition
ê−1
k = êS(k) . (1.97)
One can therefore multiply two copies of eq. (1.74) and deduce the sum of momenta,
and then take the inverse of eq. (1.74) directly to obtain the antipode. Alternatively,
in a more illuminating and less computation intensive way, one can work directly with
the definition in eq. (1.67). Indeed, using the Baker-Campbell-Hausdorff formula one
can show that

s
[X, Y ] = sY =⇒ X Y
e e = exp X + Y (1.98)
1 − e−s
which means that since because of eq. (1.48) we have
i k0
[ik0 x0 , ikx] = −k0 k[x0 , x] = −k0 k x = − (ikx), (1.99)
κ κ
then
!
x0 − kκ0
eik0 eikx = exp ik0 x0 + k0 ikx (1.100)
1−eκ
! !
k0
x0 − kκ0 k0
eiqx eiq0 = exp iq0 x0 + κ
k iqx = exp iq0 x0 + k0 ie κ qx (1.101)
− κ0
1−e 1−eκ
k0
and therefore sending q → ke− κ and q0 → k0 we prove
k
0 − κ0
kx ik0 x0
eik0 x eikx = eie e . (1.102)
Using this relation, we get

k0
i k+e− κ q x
x0 x0 0
êk êq = eikx eik0 eiqx eiq0 =e ei(k0 +q0 )x (1.103)
Furthermore, we can also recover the inverse group element ê−1

k using a similar pro-
cedure. In fact we can use the same steps as before, but at the end we must impose
that both the exponent go to zero, so that we get
k0
q = −e κ k q0 = −k0 (1.104)
and therefore
k0
0
ê−1
k =e
−ie κ k −ik0 x
e (1.105)
We now consider the bicrossproduct coordinate functions on AN (3) defined as (75)

0

fµB (g ∈ AN (3)) = fµB eikx eik0 x := kµ = (k0 , k) (1.106)
These are important because functions on a group have a natural structure of Hopf
algebra (75). In fact, using the same notation as (75), the algebra F un(G) of functions
on G with multiplication and identity defined as usual
(f1 , f2 )(g) = f1 (g)f2 (g) I(g) = 1 (1.107)
This algebra F un(G) is of course commutative and associative (because pointwise

multiplication of functions with codomain in C is of course commutative and associa-
tive). From this we can naturally define all of the additional structure needed for a
Hopf algebra, namely coproduct, co-unit, and antipode
∆ : F un(G) → F un(G × G) ϵ : F un(G) → C S : F un(G) → F un(G)

(1.108)
with the definitions

X
(∆f )(g1 , g2 ) = fα(1) ⊗ fα(2) (g1 , g2 ) = f (g1 g2 ) (1.109)
α
ϵ(g) = f (1) (S(f ))(g) = f (g −1 ) (1.110)
Therefore, using these definitions and the bicrossproduct coordinate functions, we

immediately find
∆ f0B (êk , êq ) = f0B (êk êq ) (1.111)

k0 !
i k+e− κ q x 0
= f0B e ei(k0 +q0 )x (1.112)
= k0 + q0 (1.113)
= f0B ⊗ 1 + 1 ⊗ f0 B
(1.114)

(êk , êq )
∆ fiB (êk , êq ) = fiB (êk êq ) (1.115)


k0 !
i k+e− κ q x
)x0
= fiB e ei(k0 +q0 (1.116)
k0
= k + e− κ q (1.117)
k0

= fiB ⊗ 1 + e− κ ⊗ fiB (êk , êq ) (1.118)
and similarly for the co-unit

0

ϵ(f0B ) = f0B (1) = f0B ei0x ei0x = 0 (1.119)
0

ϵ(fiB ) = fiB (1) = fiB ei0x ei0x = 0 (1.120)
and the antipode

k0

−ie κ k −ik0 x0
S(f0B )(êk ) = f0B (ê−1
k ) = f0B e e = −k0 (1.121)
k0

k0
−ie κ k −ik0 x0
S(fiB )(êk ) = fi (ê−1
k ) = fi e e = −e κ k (1.122)
With a slight abuse of notation, these relations are traditionally written as

k0
∆k0 = k0 ⊗ 1 + 1 ⊗ k0 ∆ki = ki ⊗ 1 + e− κ ⊗ ki (1.123)
k0
S(k)0 = −k0 S(k)i = −e κ ki (1.124)
so that we recover eq. (1.15) and (1.18). In the same way, using the classical basis, we
recover eq. (1.59), (1.60), (1.64).
To get the coproduct of boosts and rotations, we follow a slightly different route.
Indeed, knowing that AN (3) is our momentum manifold, we want to find some action
of SO(1, 3) on it. The trick to find it is to notice that both AN (3) and SO(1, 3) are
subgroups of SO(1, 4). We can then use something called Iwasawa decomposition
so(1, 4) = so(1, 3) ⊕ a ⊕ n
where a is generated by x̂0 , n is generated by x̂i . In other words, given g ∈ SO(1, 4),
then given Kg ∈ SO(1, 3) and g ∈ AN (3), then there exist two unique elements
Kg′ ∈ SO(1, 3) and g ′ ∈ AN (3) such that
Kg g = g ′ Kg′ ⇔ Kg g(Kg′ )−1 = g ′ .
In this way, we obtain the action of SO(1, 3) on a single group element. To get the
co-product, we then need to define the action on the product gh, where g, h ∈ AN (3).
We do this by the following relation.
−1
(gh)′ = Kg ghKgh
′ ′
= (Kg gK ′ g )(Kg′ hKgh ) (1.125)
We will see shortly that Kgh′ is irrelevant for the definition of the coproducts of
SO(1, 3) generators, so that we only need Kg and Kg′ . We are only interested in
infinitesimal transformations, so we assume
Kg ≈ 1 + iξ a ka Kg′ ≈ 1 + iξ a hba (g)kb . (1.126)
Notice that hba (g) is in general momentum dependent, a fact which is responsible for
the non-trivial boost co-product. Concentrating on boosts, and in particular choosing
a boost in the direction 1, and using the classical basis for simplicity, explicitly we
have
 
1 ξ 0 0 0  
ξ 1 0 0 0 P̄4 κP P+
i
P 0
(1.127)
 
0 0 1 0 0 
  Pi κ Pi 


0 0 0 1 0 P̄0 − κPi P4
P+
0 0 0 0 1
ξ¯1 ξ¯2 ξ¯3 0
 
 1
κP′i

P̄ ′ ′
P0 ξ¯1
 4′
′
P+ 1 ρ̄3 −ρ̄2 0
′ ξ¯ −ρ̄ (1.128)
 2 3

=Pi κ Pi  1 ρ̄1 0
κP′i ¯
  
′ ′ ξ 3 ρ̄2 −ρ̄ 1 1 0
P̄0 − P ′ P4  
+
0 0 0 0 1
where the parameters ξ¯i and ρ̄j can be uniquely determined by dong the product
and comparing both sides of the equivalence (75). After doing the computations, one
obtains
 κ 
1 P+ ξ 0 0 0
 κ ξ P2 P3
 P+ 1 P+ ξ P+ ξ 0

κ Pj
Kg′ =  0 − PP+2 ξ 1 0 0 = 1 + ξN1 + ϵijk Mk . (1.129)
 
 P
 P+ P+
 0 − ξ 3
P+ 0 1 0
0 0 0 0 1
In other words, we have up to first order

κ Pj
(gh)′ = (1 + ξN1 )g 1 + ξN1 + ϵijk ′
Mk hKgh (1.130)
P+ P+
Now, notice that a multiplication from the right by Kgh′ cannot influence the momenta
in a group element h. This is due to the fact that Kgh will have all zero entries in the
′
last column and row, except (Kgh′ )

44 = 1, and a multiplication from the right of such
a matrix cannot influence the last row of h, which contains the momenta. Therefore,
we have up to first order

′ κ Pj
(gh) = gh + ξ (N1 g)(h) + g (N1 h) + ϵ1jk g (Mk h). (1.131)
P+ P+
and generalizing this relation to a boost in a generic direction we get immediately the
co-product in eq. (1.62). Finally, for the antipode, one can just switch g 7→ g −1 and
h 7→ g into eq. (1.131), and impose that the result gives 1, which means that we get
the antipode in eq. (1.65). For the rotation, the whole process is repeated, and one
gets that the co-product and the antipode are not deformed, in accordance with eq.
(1.63) and (1.65).
1.5 Finite boosts for two-particle states

What we computed above is the co-product of generators of the algebra on tensor
product states. To get a finite boost for two particles, one needs to exponentiate this
infinitesimal generators to get a finite transformation. However, due to the non-trivial
nature of co-products both in the classical and bicrossproduct basis, this is often not
practical. Luckily for us, however, the Iwasawa decomposition that was used in section
1.4 holds in general, and we only used a first order expansion for the involved matrices
because we only needed the generators co-products (77). Here, we employ once again
the same decomposition, but this time we will use the full transformation matrices to
obtain a finite boost for a two-particle state. Contrary to the previous sections, the
computations and discussions for the rest of the chapter are original results. Because
of the importance of the final result for the kinematics of a two-particle state in the
κ-deformed context, we will explicitly perform all the computations in full.
We start therefore from
Lg = g ′ L′g (1.132)
where the boost on the RHS can in general be momentum dependent.

We need to recover what is L′g . Written explicitly in terms of matrices, and con-
sidering without loss of generality L as a boost in the x1 direction, we have
−βγ 0 0 0
 
γ
−βγ γ 0 0 0
(1.133)
 
L= 0  0 1 0 0 
 0 0 0 1 0
0 0 0 0 1
 
P̃4 κP1 /P+ κP2 /P+ κP3 /P+ P0
P1 κ 0 0 P1 
(1.134)
 
g = P2
 0 κ 0 P2 

P3 0 0 κ P3 
P̃0 −κP1 /P+ −κP2 /P+ −κP3 /P+ P4
P̃4′ κP′1 /P+′ κP′2 /P+′ κP′3 /P+′ P0′

 
P′1 κ 0 0 P′1 
g′ =  ′ ′ (1.135)
 
 P 2 0 κ 0 P 2 
 P′ 0 0 κ P′3 
3
P̃0′ −κP′1 /P+′ −κP′2 /P+′ −κP′3 /P+′ P4′
where P̃4 and P̃0 are defined in eq. (1.79). For what concerns the matrix L′g , we will
use the generic matrix
 
L00 L01 L02 L03 0
L10 L11 L12 L13 0
′
(1.136)
 
L20 L21 L22 L23 0
Lg =  
L30 L31 L32 L33 0
0 0 0 0 1
1.5. Finite boosts for two-particle states 17
1.5.1 Fifth column:

  ′
γP0 − βγP1

P0
γP1 − βγP0  P′1 
 !  ′
(1.137)


 P2  = P 
  2
 P3  P′ 
3
P4 P4′
which translates into the equalities
P0′ = γP0 − βγP1 (1.138)

P′1 = γP1 − βγP0 (1.139)
P′2 = P2 (1.140)
P′3 = P3 (1.141)
P4′ = P4 (1.142)
1.5.2 Fourth column:
L03 P̃4′ + L13 κP′1 /P+ + L23 κP′2 /P+ + L33 κP′3 /P+
   
γκP3 /P+
−βγκP3 /P+  
 !  L03 P′1 + L13 κ 
′  (1.143)
 

 0 =
  L03 P2 + L23 κ 
 κ   ′
L03 P3 + L33 κ 
−κP3 /P+ ′ ′ ′ ′
L03 P̃0 − L13 κP1 /P+ − L23 κP2 /P+ − L33 κP3 /P+
Summing the equations coming from the first and fifth row we get
L03 (P̃4′ + P̃0′ ) = (γ − 1)κP3 /P+ (1.144)
and using eq. (1.79) and (1.138) we get
P3
L03 = (γ − 1)κ (1.145)
P+ P+′
and therefore
P3
L03 = (γ − 1)κ (1.146)
P+ (γP0 − βγP1 + P4 )
From the third row we get

P2
L23 = − L03 (1.147)
κ
and therefore
P2 P3
L23 = (1 − γ) (1.148)
P+ (γP0 − βγP1 + P4 )
From the fourth row we get

P3
L33 = − L03 + 1 (1.149)
κ
and therefore
P23
L33 = (1 − γ) +1 (1.150)
P+ (γP0 − βγP1 + P4 )
Finally, from the second row we get
P3 P′
L13 = −βγ − 1 L03 (1.151)
P+ κ
and therefore
P3 (γP1 − βγP0 )P3

L13 = −βγ + (1 − γ) (1.152)
P+ P+ (γP0 − βγP1 + P4 )
1.5.3 Third column:
L02 P̃4′ + L12 κP′1 /P+ + L22 κP′2 /P+ + L32 κP′3 /P+
   
γκP2 /P+
−βγκP2 /P+  
 !  L02 P′1 + L12 κ 
′  (1.153)
 

 κ =
  L02 P2 + L22 κ 
 0   ′
L02 P3 + L32 κ 
−κP2 /P+ ′ ′ ′
L02 P̃0 − L12 κP1 /P+ − L22 κP2 /P+ − L32 κP3 /P+′
L02 (P̃4′ + P̃0′ ) = (γ − 1)κP2 /P+ (1.154)
P2
L02 = (γ − 1)κ (1.155)
P+ P+′
and therefore
P2
L02 = (γ − 1)κ (1.156)
P+ (γP0 − βγP1 + P4 )

P3
L32 = − L02 (1.157)
κ
and therefore
P2 P3
L32 = (1 − γ) (1.158)
P+ (γP0 − βγP1 + P4 )

P2
L22 = − L02 + 1 (1.159)
κ
and therefore
P22
L22 = (1 − γ) +1 (1.160)
P+ (γP0 − βγP1 + P4 )
P2 P′
L12 = −βγ − 1 L02 (1.161)
P+ κ
and therefore
P2 (γP1 − βγP0 )P2

L12 = −βγ + (1 − γ) (1.162)
P+ P+ (γP0 − βγP1 + P4 )
1.5.4 Second column:
γκP1 /P+ − βγκ L01 P̃4′ + L11 κP′1 /P+ + L21 κP′2 /P+ + L31 κP′3 /P+
   
−βγκP1 /P+ + γκ 
 !  L01 P′1 + L11 κ 
′
 

 0 =
  L01 P2 + L21 κ 

 0   ′
L01 P3 + L31 κ 
−κP1 /P+ ′ ′ ′ ′
L01 P̃0 − L11 κP1 /P+ − L21 κP2 /P+ − L31 κP3 /P+
(1.163)
L01 (P̃4′ + P̃0′ ) = (γ − 1)κP1 /P+ − βγκ (1.164)
P1 βγκ
L01 = (γ − 1)κ ′ − ′ (1.165)
P+ P+ P+
and therefore
P1 βγκ
L01 = (γ − 1)κ − (1.166)
P+ (γP0 − βγP1 + P4 ) γP0 − βγP1 + P4

P3
L31 = − L01 (1.167)
κ
and therefore
P1 P3 βγP3
L31 = (1 − γ) + (1.168)
P+ (γP0 − βγP1 + P4 ) γP0 − βγP1 + P4

P2
L21 = − L01 (1.169)
κ
and therefore
P1 P2 βγP2
L21 = (1 − γ) + (1.170)
P+ (γP0 − βγP1 + P4 ) γP0 − βγP1 + P4
P1 P′
L11 = −βγ − 1 L01 + γ (1.171)
P+ κ
and therefore
P1 (γP1 − βγP0 )P1 βγ(γP1 − βγP0 )

L11 = −βγ + (1 − γ) + +γ (1.172)
P+ P+ (γP0 − βγP1 + P4 ) γP0 − βγP1 + P4
1.5.5 First column:
γ P̃4 − βγP1 L00 P̃4′ + L10 κP′1 /P+ + L20 κP′2 /P+ + L30 κP′3 /P+
   
γP1 − βγ P̃4  
 !  L00 P′1 + L10 κ 
′ (1.173)
 

 P2 =
  L00 P2 + L20 κ 

′
L00 P3 + L30 κ
 P3   
′ ′ ′ ′
L00 P̃0 − L10 κP1 /P+ − L20 κP2 /P+ − L30 κP3 /P+
P̃0
P4 P0 P2 P1
L00 = γ ′ + ′ + (γ − 1) ′ − βγ ′ (1.174)
P+ P+ P+ P+ P+
and therefore
γP4 + P0 P2
L00 = + (γ − 1)
γP0 − βγP1 + P4 P+ (γP0 − βγP1 + P4 )
P1
− βγ (1.175a)
γP0 − βγP1 + P4
From the third row we have

P2
L20 = (1 − L00 ) (1.176)
κ
and therefore
P2 h γP4 + P0
L20 = 1−
κ γP0 − βγP1 + P4
P2 P1 i
+ (γ − 1) − βγ (1.177a)
P+ (γP0 − βγP1 + P4 ) γP0 − βγP1 + P4
From the fourth row we have

P3
L30 = (1 − L00 ) (1.178)
κ
and therefore
P3 h γP4 + P0
L30 = 1−
κ γP0 − βγP1 + P4
P2 P1 i
+ (γ − 1) − βγ (1.179a)
P+ (γP0 − βγP1 + P4 ) γP0 − βγP1 + P4
Finally, from the second row we have
P2

1
L10 = γP1 − βγP4 − βγ − (γP1 − βγP0 )L00 (1.180)
κ P+
and therefore
1h P2 γP4 + P0
L10 = γP1 − βγP4 − βγ − (γP1 − βγP0 )
κ P+ γP0 − βγP1 + P4
P 2 P1 i
+ (γ − 1) − βγ
P+ (γP0 − βγP1 + P4 ) γP0 − βγP1 + P4
(1.181a)
1.5.6 Expansion of the final matrix up to 1/κ

In most phenomenological applications, an expansion of the full results up to first
order in κ1 is in general useful for numerical estimations of the effect of κ-deformation
on canonical, non-deformed quantities. Because of this, we here expand up to first
order the matrix elements computed in the previous section. We make use of the
following relations.
a a
≈ . (1.182)
b + cP4 cκ
More in general, we also have

a + bP4 b b ad c
≈ + − (1.183)
c + dP4 d dκ b d
We also have
aκ a
≈ (1.184)
P+ (b + P4 ) κ
aκ ab
≈a− (1.185)
b + P4 κ
where a, b, c, d do not depend on κ. We have
(1 − γ)
(1.175) 7→ L00 ≈ γ + [(1 + γ)P0 − βγP1 ] (1.186)
κ
γ γ
(1.181) 7→ L10 ≈ −βγ + P1 − (γP1 − βγP0 ) (1.187)
κ κ
(1 − γ)
(1.177) 7→ L20 ≈ P2 (1.188)
κ
(1 − γ)
(1.179) 7→ L30 ≈ P3 (1.189)
κ
γ γ
(1.166) 7→ L01 ≈ −βγ + P1 − (γP1 − βγP0 ) = L10 (1.190)
κ κ
(γ − 1) β2γ2
(1.172) 7→ L11 ≈ γ + βγ P1 − P0 (1.191)
κ κ
βγP2
(1.170) 7→ L21 ≈ (1.192)
κ
βγP3
(1.168) 7→ L31 ≈ (1.193)
κ
P2
(1.156) 7→ L02 ≈ (γ − 1) = −L20 (1.194)
κ
P2
(1.162) 7→ L12 ≈ −βγ (1.195)
κ
(1.160) 7→ L22 ≈ 1 (1.196)
(1.158) 7→ L32 ≈ 0 (1.197)
P3
(1.146) 7→ L03 ≈ (γ − 1) (1.198)
κ
P3
(1.152) 7→ L13 ≈ −βγ (1.199)
κ
(1.148) 7→ L23 ≈ 0 (1.200)

(1.150) 7→ L33 ≈ 1 (1.201)
We can now finally write the matrix in extended form. We have

P2 P
 (1−γ) γ γ

γ+ κ
[(1 + γ)P0 − βγP1 ] −βγ + κ P1 − κ (γP1 − βγP0 ) (γ − 1)κ κ (γ − 1) κ3 0
γ γ (γ−1) 2 2 P P
−βγ + κ P1 − κ (γP1 − βγP0 ) γ + βγ κ P1 − β κγ P0 −βγ κ2 −βγ κ3 0
L′g =  (1−γ) βγP2

κ
P2 κ
1 0 0 
(1−γ) βγP3
κ
P3 κ
0 1 0
0 0 0 0 1
(1.202)
This can be rewritten more simply as a contribution with κ → ∞ plus the first order
in κ1 obtaining
1 ′
L′g = L′g (κ = ∞) + L̃ (1.203)
κ g
where
−βγ 0 0 0
 
γ
−βγ γ 0 0 0
′
(1.204)
 
Lg (κ = ∞) = 
 0 0 1 0 0
 0 0 0 1 0
0 0 0 0 1
(1 − γ) [(1 + γ)P0 − βγP1 ] γP1 − γ(γP1 − βγP0 ) (γ − 1)P2 (γ − 1)P3 0 !

1 ′ 1 γP1 − γ(γP1 − βγP0 ) βγ(γ − 1)P1 − β 2 γ 2 P0 −βγP2 −βγP3 0
L̃g = (1 − γ)P2 βγP2 0 0 0 (1.205)
κ κ (1 − γ)P3 βγP3 0 0 0
0 0 0 0 0
1.5.7 Check of the consistency of the first order expansion of L′g

Notice that if L′g is a legitimate Lorentz transformation, then by definition one must
have
T
L′g ηL′ g = η (1.206)
where η = diag(+−−−const). We know that the zeroth order in perturbation satisfies

the relation
T
L′g (κ = ∞)ηL′ g (κ = ∞) = η (1.207)
by definition, which means that we must have (ignoring terms of order 1/κ2 )
L′g (κ = ∞)η(L̃′ )Tg + L̃′g η[L′g (κ = ∞)]T = 0 (1.208)
We have
(1 − γ)P2 (1 − γ)P3
 
A B 0
 C D βγP2 βγP3 0
′ ′ T
(1.209)
 
Lg (κ = ∞)η(L̃ )g = (1 − γ)P2 βγP2
 0 0 0
(1 − γ)P3 βγP3 0 0 0
0 0 0 0 0
where
A = γ(1 − γ) [(1 + γ)P0 − βγP1 ] + βγ[γP1 − γ(γP1 − βγP0 )] (1.210)

C = −βγ(1 − γ) [(1 + γ)P0 − βγP1 ] − γ[γP1 − γ(γP1 − βγP0 )] (1.211)

B = γ[γP1 − γ(γP1 − βγP0 )] + βγ[βγ(γ − 1)P1 − β γ P0 ] 2 2
(1.212)
D = −βγ[γP1 − γ(γP1 − βγP0 )] − γ[βγ(γ − 1)P1 − β γ P0 ] 2 2
(1.213)
Furthermore we have
(γ − 1)P2 (γ − 1)P3
 
Ã B̃ 0
 C̃ D̃ −βγP2 −βγP3 0
L̃′g η(L′g (κ = ∞))T =  (1.214)
 
(γ − 1)P 2 −βγP 2 0 0 0
(γ − 1)P3 −βγP3 0 0 0
0 0 0 0 0
where
Ã = γ(1 − γ) [(1 + γ)P0 − βγP1 ] + βγ[γP1 − γ(γP1 − βγP0 )] (1.215)

B̃ = −βγ(1 − γ) [(1 + γ)P0 − βγP1 ] − γ[γP1 − γ(γP1 − βγP0 )] (1.216)
C̃ = γ[γP1 − γ(γP1 − βγP0 )] + βγ[βγ(γ − 1)P1 − β 2 γ 2 P0 ] (1.217)
D̃ = −βγ[γP1 − γ(γP1 − βγP0 )] − γ[βγ(γ − 1)P1 − β γ P0 ] 2 2
(1.218)
The sum of the two matrices is therefore

 
A + Ã B + B̃ 0 0 0
C + C̃ D + D̃ 0 0 0
(1.219)
 
 0 0 0 0 0
 
 0 0 0 0 0
0 0 0 0 0
so that we only need to compute the remaining quantities. We do it one by one. First
notice that A = Ã, so we need to show that A = 0. Indeed, we have
A = γ(1 − γ) [(1 + γ)P0 − βγP1 ] + βγ[γP1 − γ(γP1 − βγP0 )] (1.220)

2
= P1 [−βγ (1 − γ) + βγ − βγ ] 2 3
(1.221)
+ P0 [γ(1 − γ 2 ) + β 2 γ 3 ] (1.222)
=0 (1.223)
where we used the fact that 1 − γ 2 = (1 − β 2 − 1)/(1 − β 2 ) = −β 2 γ 2 . Secondly, notice

that also D = D̃, and therefore we must have D = 0. Indeed we have
D = −βγ[γP1 − γ(γP1 − βγP0 )] − γ[βγ(γ − 1)P1 − β 2 γ 2 P0 ] (1.224)

2 3
= P1 [−βγ + βγ − βγ + βγ ] 3 2
(1.225)
2 3
+ P0 [−β γ + β γ ] 2 3
(1.226)
=0 (1.227)
Then we need to check B + B̃ and C + C̃. Starting with the first one we have
B + B̃ = γ[γP1 − γ(γP1 − βγP0 )] + βγ[βγ(γ − 1)P1 − β 2 γ 2 P0 ] (1.228)

+ {−βγ(1 − γ) [(1 + γ)P0 − βγP1 ] − γ[γP1 − γ(γP1 − βγP0 )]} (1.229)
= βγ[βγ(γ − 1)P1 − β 2 γ 2 P0 ] − βγ(1 − γ) [(1 + γ)P0 − βγP1 ] (1.230)

1.6. Two-particle kinematics 25
= P1 [β 2 γ 3 − β 2 γ 2 + β 2 γ 2 − β 2 γ 3 ] + P0 [−β 3 γ 3 − βγ(1 − γ 2 )] (1.231)

=0 (1.232)
where we used the fact that 1 − γ 2 = (1 − β 2 − 1)/(1 − β 2 ) = −β 2 γ 2 . Finally, we have
C + C̃ = −βγ(1 − γ) [(1 + γ)P0 − βγP1 ] − γ[γP1 − γ(γP1 − βγP0 )]

+ γ[γP1 − γ(γP1 − βγP0 )] + βγ[βγ(γ − 1)P1 − β 2 γ 2 P0 ] (1.233)
= B + B̃ = 0 (1.234)
which concludes the check. Therefore, the matrix

1 ′
L′g = L′g (κ = ∞) + L̃ (1.235)
κ g
satisfies
L′g η(L′g )T = η (1.236)
and it is therefore a valid Lorentz transformation.
1.6 Two-particle kinematics

We now have the finite boost for a two-particle state, and we can therefore turn to
understanding the phenomenological consequences of our computations.
As we already discussed in the context of the boost co-product in section 1.4,
multiplication of the matrices Lg, L′g h from the right by an arbitrary Lorentz matrix
does not change the entries of their last columns so that |(Lg(L′g )−1 )(P )⟩ = |(Lg)(P )⟩
and in writing the states we can neglect the Lorentz group element on the right of
g. Therefore we define the action of the finite Lorentz transformation L on a two
particle state as (77)
∆L ▷ |g(P )⟩ ⊗ |h(Q)⟩ = |(Lg)(P )⟩|(L′g h)(Q)⟩ (1.237)
where (Lg)(P ) denotes (in the case of the boost along the first axis) the four top
components of the last column of the product of the matrices L (which is a canonical
boost) and g (eq. (1.134)), in components explicitly
Lg(P ) = (γP0 − βγP1 , −βγP0 + γP1 , P2 , P3 ) (1.238)
The components of L′g h can be similarly computed using eq. (1.203), (1.204), and
(1.205), as we will discuss in more details shortly.
We now consider the decay of a particle of mass M , originally at rest, into two par-
ticles of mass m. Starting from the first order expansion in 1/κ of the four-momentum
deformed composition rule (P ⊕ Q)0 and (P ⊕ Q)i we get
PQ Pi Q0
(P ⊕ Q)0 ≈ P0 + Q0 + (P ⊕ Q)i ≈ Pi + Qi + (1.239)
κ κ
and we need to impose
PQ Pi Q0
P0 + Q0 + =M Pi + Qi + = 0. (1.240)
κ κ
We need to find the spatial momenta such that these two relations are satisfied. The
conservation of spatial momenta tells us that the momenta P and Q are parallel, which
means that we can align them with one of our axis, and we can write Pµ = (P0 , P, 0, 0)
and Qµ = (Q0 , Q, 0, 0) without loss of generality. We have therefore
PQ P Q0
P0 + Q0 + =M P +Q+ = 0. (1.241)
κ κ
From the conservation of momenta, one can get the relation
QQ0
P ≈ −Q + (1.242)
κ
which substituted back into the sum of energies gives (at first order in 1/κ)

p 1 3M
2
Q ≈ M − 4m 2 + (1.243)
2 8κ
Substituting back into (1.242), we also get

p 1 M
P ≈ − M − 4m2 2 + . (1.244)
2 8κ
√
Notice that the modulus P differs from that of Q by M 2 − 4m2 M/4κ. Notice also
that another solution would have been to switch the RHS of eq. (1.243) and (1.244),
which however amounts to the same solution. Notice that in the formal limit κ → ∞
the spatial momenta are equal and opposite. The deformation effects is here considered
only up to first order in 1/κ, but can be understood without needing to approximate.
Indeed, since we are imposing P ⊕ Q = 0, it is clear that one must either have P = P ′
and Q = S(P ′ ) or vice versa. What we obtain in eq. (1.243) and (1.244) is just a first
order expansion of this all-orders result. We will see in chapter 2 that this is not only
a general kinematical property due to the presence of κ-deformation, but it is also
comes out naturally when dealing with particles and antiparticles at the same time.
We can now boost the two particles. Since the boost is in the x1 direction but
the two particles can go in arbitrary direction in the center of mass frame, we con-
sider spherical coordinates with the polar axis oriented along the x axis, so that the
coordinates of the momenta Pµ and Qµ are given by
Pµ = (P0 , P cos θ, P sin θ sin ϕ, P sin θ cos ϕ) =: (P0 , P1 , P2 , P3 ) (1.245)

Qµ = (Q0 , Q cos θ, Q sin θ sin ϕ, Q sin θ cos ϕ) =: (Q0 , Q1 , Q2 , Q3 ). (1.246)
When computing the boost, we have two possibilities, depending on which particle
is the first and which is second in (1.237), namely L ▷ Pµ and L′g ▷ Qµ , or L ▷ Qµ
and L′g ▷ Pµ . Since the non-deformed boost acts on momenta as usual, we will only
concentrate in the action of L′g . Furthermore, since the second case can be obtained
from the first one by simply switching P ←→ Q, we will only consider the first one.
Using equations (1.133), (1.203), (1.204), (1.205), (1.243), (1.244), we can write down
the first order expansion in powers of 1/κ of the momenta boosted using L′g . We have
1h
(L′g ▷ Q)0 = γ(Q0 − βQ1 ) + (γ − 1)(P2 Q2 + P3 Q3 − γP1 Q1 + βγQ0 P1 )
κ i
+ P0 (Q0 − Q0 γ 2 + βγ 2 Q1 ) (1.247)
1h
(L′g ▷ Q)1 = γ(Q1 − βQ0 ) + − βγ(P2 Q2 ) + P3 Q3 − P1 (Q0 − βQ1 )(γ − 1)γ
κ i
+ P0 (Q1 − Q1 γ 2 + Q0 βγ 2 ) (1.248)
1
(L′g ▷ Q)2 = Q2 + [P2 (Q0 − Q0 γ + Q1 βγ)] (1.249)
κ
1
(L′g ▷ Q)3 = Q3 + [P3 (Q0 − Q0 γ + Q1 βγ)] (1.250)
κ
One can also explicitly write the expressions of L ▷ P and L′g ▷ Q obtained by substi-
tuting eq. (1.243), (1.244) in the formulae, which are the following.
1 p p
(L′g ▷ Q)0 = γ β cos(θ) M 2 − 4m2 + 5M 2 − 4m2
2
1 h p p
+ 5βγ cos(θ) M 2 − 4m2 3M − 2 5M 2 − 4m2
40κ
12γM p
+ 4m2 5γ 2 − √ − 5 + 3M γ 5M 2 − 4m2 − 15 γ 2 − 1 M
5M 2 − 4m2
i 1 2
− 5(γ − 1)2 cos(2θ) 4m2 − M 2 + O (1.251)
κ
1 p p
(L′g ▷ Q)1 = − γ β 5M 2 − 4m2 + cos(θ) M 2 − 4m2
2
1 h p
+ √ βγ (γ − 2) cos(2θ) 4m2 − M 2 5M 2 − 4m2
8κ 5M 2 − 4m2
p p
+ 4m2 3M − γ 5M 2 − 4m2 − 3M 2 M − 3γ 5M 2 − 4m2
p p
+ cos(θ) M 2 − 4m2 M 10(γ − 1)M − 3γ 5M 2 − 4m2
i 1 2
− 8(γ − 1)m2 + O (1.252)
κ
1 p
(L′g ▷ Q)2 = sin(θ) M 2 − 4m2 cos(ϕ)
2
1h p p
− sin(θ) M 2 − 4m2 cos(ϕ) 2βγ cos(θ) M 2 − 4m2
8κ
p i 1 2
+ 2(γ − 1) 5M 2 − 4m2 − 3M + O (1.253)
κ
1 p
(L′g ▷ Q)3 = sin(θ) M 2 − 4m2 sin(ϕ)
2
1h p p
− sin(θ) M 2 − 4m2 sin(ϕ) 2βγ cos(θ) M 2 − 4m2
8κ
p i 1 2
+ 2(γ − 1) 5M 2 − 4m2 − 3M + O (1.254)
κ
1 p p
(L ▷ P )0 = γ β cos(θ) M 2 − 4m2 + 5M 2 − 4m2
2
√ 2 −4m2

γM β cos(θ) M 2 − 4m2 + √M
5M 2 −4m2
1 2
+ +O (1.255)
8κ κ
1 p p
(L ▷ P )1 = − γ β 5M 2 − 4m2 + cos(θ) M 2 − 4m2
2
√

β (4m2 −M 2 ) 2 2
γM √5M 2 −4m2 − cos(θ) M − 4m 1 2
+ +O (1.256)
8κ κ
1 p
(L ▷ P )2 = − sin(θ) M 2 − 4m2 cos(ϕ)
2 √
M sin(θ) M 2 − 4m2 cos(ϕ) 1 2
− +O (1.257)
8κ κ
1 p
(L ▷ P )3 = − sin(θ) M 2 − 4m2 sin(ϕ)
2 √
M sin(θ) M 2 − 4m2 sin(ϕ) 1 2
− +O (1.258)
8κ κ
Notice that, in absence of deformation, the components of momenta along x2 and x3
(i.e. those components perpendicular to the boost direction) of L′g ▷ Q and L ▷ P are
equal and opposite. However, deformation introduces additional terms which make
even these components not equal and opposite after boost, which in turn gives a
deformed angular distribution. To show this, we plot the angular dependence of the
modulus of the deformed boosted momenta to compare it to the non-deformed case.
In figures below all the quantities on the axes are expressed in GeV.
In the center of mass (COM) frame, the two momenta are distributed on a sphere.
Because of deformation, however, the spatial momenta P and Q are different in mo-
dulus (see eq. (1.243), (1.244)). In Figure 1.1 we show the distribution of momenta in
the COM frame, together with the distribution in the non-deformed case. To highlight
the qualitative difference, for the next plots we chose hypothetically M = 10 GeV,
m = 0.1 GeV, κ = 102 GeV. Furthermore, we chose the domains ϕ ∈ − 10 π 3π π
, 2 + 10
and ϕ ∈ 0, 2 for the distributions of the deformed P and the deformed Q respec-
3π
tively, in order to clearly show how they are related to the non-deformed one. Notice
that the non-deformed case contains only a single surface because both spatial mo-
menta have the same modulus. We then boost our momenta. In particular, Pµ will
be canonically boosted while Qµ will be boosted in a deformed way (recall however
that both Pµ and Qµ still have some effects of deformation in their modulus). In this
case we chose γ = 5 while the other parameters remain the same. We also restricted
the domain in ϕ of all distributions to better show the features of the surfaces, and
we obtain Figure 1.2. Notice that the deformed boost not only modifies the ampli-
tude of the boosted momentum, but also its angular distribution. Finally, notice that
switching P ↔ Q in the above formula makes manifest that the κ-deformed boost
contribution is much greater than the center of mass deformation of the moduli.
From a phenomenological point of view, eq. (1.247) and (1.250) show that the
difference between the two boosted momenta mostly depend on the κ1 contribution
coming from the deformed boost (large values of boost parameter γ highlight these
contributions). This in turn is important for the study of two-particle correlations and
interference patterns, since these are often expressed as a function of the difference of
momenta (or lifetime) of the involved particles. Of particular relevance, the study of
favoured mesons from the decays Φ0 (1020) → K 0 K̄ 0 , Ψ(3770) → D0 D̄0 , Υ(10580) →
B 0 B̄ 0 and Υ(10860) → Bs0 B̄s0 .
Non-Deformed P,Q
Deformed Q
Deformed P
Rysunek 1.1: Angular distribution of deformed and non-deformed

momenta in the COM frame, with parameters M = 10 GeV, m =
0.1 GeV, κ = 102 GeV. Notice the onion-like structure, with the non-
deformed momentum being innermost, followed by the deformed P ,
and finally the deformed q. Here, P and Q are antipodal also in the κ-
deformed context (see eq. (1.240)). To give a more concrete numerical
estimate, the modulus of the non-deformed momentum is 5.00 GeV,
the modulus of P is 5.12 GeV, while the modulus of Q is 5.37 GeV.
Non-deformed P,Q
Deformed boosted Q
Boosted P
Rysunek 1.2: Angular distribution of deformed and non-deformed

momenta in the boosted frame, with parameters γ = 5, M = 10 GeV,
m = 0.1 GeV, κ = 102 GeV. Recall that both P and Q feel the effects
of deformation already in the modulus (see eq. (1.244) and (1.243)).
When boosting, P gets boosted with a canonical non-deformed boost,
while Q is instead boosted in a deformed way.
31
2
Complex scalar field on
κ-Minkowski spacetime
2.1 Introduction
Given κ-Minkowski spacetime, one can begin building physical quantities on top of it.
In particular, one can introduce fields over κ-Minkowski spacetime. The whole process
is however not trivial, and because of the many computations involved, it will be easy
to lose track of the progress or the direction of our investigations. As such, in the
remainder of this section we will provide the reader with a roadmap of this chapter.
This chapter will be divided in the following main sections:
i) In order to introduce a field in κ-Minkowski spacetime, one could directly work in
the bicrossproduct basis using the coordinate representation given in eq. (1.66).
It is however much simpler to work with canonical, commuting coordinates, with
the caveat that one needs to switch to the classical basis. One can do this through
the so called Weyl map (the canonical product of functions is also deformed to
a non-commutative ⋆ product under Weyl map). Therefore, the first step will
be to introduce the Weyl map and its properties, after which the rest of the
discussion will be more transparent. This will be done in section 2.2;
ii) We are now in the position to introduce the action of a complex scalar field,
and we will do so immediately after having introduced the Weyl map. Because
of the properties of the ⋆ product, our total action S will be defined in terms of
two possible partial actions S1 and S2 . This will be done in section 2.3;
iii) The first thing to understand about the action is how to compute physically
meaningful quantities starting from it, like for example the equations of motion
or the Noether charges. However, because of the non-trivial co-product, the
derivatives do not satisfy the Leibniz rule. We will therefore first introduce
the action of derivatives on ⋆ products from in section 2.4. After this we can
immediately introduce the integration-by-parts relations, from section 2.5 to 2.9;
iv) We now have both an action, and the integration-by-parts relation which means
that we can vary the action introduced in section 2.3 to obtain both the equations
of motion (EoM) and the surface terms. We will do this in section 2.10;
v) We can now build the on-shell field since we know the EoM. Once we have it,
we can also discuss its properties under the discrete transformations C, P , and
T . We will do this in section 2.11;
32 Rozdział 2. Complex scalar field on κ-Minkowski spacetime
vi) Going a bit off-topic, it is useful to introduce at this point the off-shell momentum-
space action, in section 2.12. This is not needed for the rest of this chapter, but
will be needed in chapter 3 when we will obtain the propagator;
vii) Since we now have the EoM and the on-shell field, the other important quantities
to obtain from the action defined in section 2.3 are the charges. We will proceed
along the following steps:
a) To compute the charges, we introduce the symplectic form in section 2.13.

The role of this object is twofold. On one hand it allows us to compute
the creation/annihilation operators algebra. On the other hand, we can
use it to compute the charges using the geometric approach, which we
describe in section 2.14. This will lead us to introduce deformed rules for the
contraction of vector fields with forms. Using them, we will obtain both the
translation (section 2.14), boost (section 2.14.2.3) and the rotation charges
(section 2.14.2.4);
b) We will conclude the computation of the charges with some mathematical
considerations on the properties of the Poisson brackets of creation/annihilation
operators, and the role of the antipode in the interaction between forms
and vector fields (section 2.14.3 and 2.14.4).
viii) We now finally have all the charges, as well as the creation/annihilation operators
algebra. The natural next step is to check what is the algebra satisfied by the
charges. We check it in section 2.15;
ix) We now have all the charges coming from continuous symmetries, and as we
discussed in point v) we also have discrete symmetries. We are therefore ready
to discuss the interaction between them. The first surprising result is the non-
commutativity of the C operator with the boost charge Ni . We show the detailed
computations of this fact in section 2.16;
x) Another important and non-trivial result is related to the Greenberg’s theorem,

which (in canonical axiomatic QFT) shows how invariance CP T transformation
is equivalent to Lorentz invariance. Because our result of point ix), we already
know that in our model CP T invariance is subtly violated. We investigate the
details of Greenberg’s theorem and its application to the deformed context in
section 2.17;
xi) We will conclude with some phenomenological considerations concerning possi-

ble signatures of the fact that [C, Ni ] ̸= 0 in section 2.18.
The Weyl map introduced in point i) of the above list is a quantity which (at least in
canonical quantum mechanics) has been known for decades. Here we just review some
of the material in the literature which we need to proceed forward, but it does not
constitute original work. Similarly, the deformation of the Lorenz rule for derivatives
and the integration-by-parts obtained in point iii) have been already discussed in (86)
in the context of the bicrossproduct basis, and only their translation to the classical
basis discussed in this work is original. All other points are original results.
2.2 Weyl maps

We begin with point i) in the roadmap. In our context, a Weyl map is a function which
allows for every object built on top of the group manifold AN (3) to be translated into
2.2. Weyl maps 33
commutative spacetime with ⋆ product. This includes functions of x, integrals, deri-

vatives. Since this is a standard construction, we only pinpoint the most important
features and definitions, referring the reader to (86), (80), (85), (78) for a more com-
plete discussion. Conceptually speaking, a Weyl map is not strictly needed for any of
the computations in the present and following chapters. Indeed, since AN (3) is a Lie
group, there are well defined concepts of differentiation and integration, which suffice
for our purpose. Nevertheless, being able to switch to a more conventional spacetime,
where coordinates are not matrices any more, greatly simplifies the computations, and
allows for a more clear interpretation of the results. For clarity, since we will need to
talk about the parity transformation P , from here on we will indicate the classical
basis with p and not P .
By definition, we have
W (êk (x̂)) = ep(k) (x), W −1 (ep(k) (x)) = êk (x̂), (2.1)
where
ep(k)(x) = e−i(ωp t−px) , (2.2)
and where k are momenta in the bicrossproduct basis, and p are the momenta in
the classical basis as a function of the k momenta, see eq. (1.75), (1.88) (recall also
the definition of ê∗ in eq. (1.87)). Of course, we need to take care also of the sum
of momenta and antipodes, and we have the definitions (we will only write down the
transformation in one direction, since the inverse is then obvious, and we also avoid
writing the dependence on both x and x̂, since they are also obvious)
W ((êk )−1 ) = eS(p(k)) , W (êk êl ) = ep(k)⊕q(l) =: ep ⋆ eq (2.3)
where the second equation defines the ⋆ product. Notice that, since the product in
AN (3) is not commutative, then neither is the ⋆ product. Furthermore, notice that
since by definition we have
e†p ⋆ ep = ep ⋆ e†p = 1 (2.4)
then we have
e†p = eS(p) . (2.5)
Furthermore, calling ∂ˆ the derivatives in the context of the AN (3) group defined by
∂ˆµ êk = ipµ (k) êk , ∂ˆ4 êk = i(κ − p4 (k)) êk
∂ˆµ ê∗k = ip∗µ (k) ê∗k , ∂ˆ4 ê∗k = i(κ − p∗4 (k)) ê∗k (2.6)
we define the action of the Weyl map action by
W (∂ˆµ êk ) = ∂µ ep W (∂ˆ4 êk ) = ∂4 ep , ∂4 ep = i(κ − p4 )ep . (2.7)
Finally, we have the definition

Zb Z
êk = d4 x W (êk (x̂)) (2.8)
R4
for plane waves. Notice that the integral on the LHS is not an integral over AN (3).
This in turn allows us to define fields using the Fourier transform (86), (80), (85),
(78), (81), (82)
d4 p
Z
ϕ(x) = dµ(p)ϕ̃(p)ep (x) dµ(p) = (2.9)
AN (3) p4 /κ
restricted by the constraints in eq. (1.56). Of course, one can define the same thing at
the group level and not in canonical spacetime by
Z
ϕ̂(x̂) = dµ(p)ϕ̃(p)êk (x̂) (2.10)
AN (3)
with the definitions
W (ϕ̂(x̂)) = ϕ(x) W (ϕ̂(x̂)ψ̂(x̂)) = ϕ(x) ⋆ ψ(x) (2.11)
and therefore
Zb Z
ϕ̂(x̂)ψ̂(x̂) = d4 x ϕ(x) ⋆ ψ(x). (2.12)
R4
In other words, we can safely translate quantities defined purely in terms of (de-
rivatives and integrals of) group elements in AN (3) into more familiar terms, with
canonical Riemannian integrals defined in R4 and derivatives acting on canonical plane
waves, paying attention to the fact that products of different objects need to be defi-
ned as ⋆ product. Of course, such a set of definitions is not unique, for example here
we chose to relate canonical plane waves with group elements defined according to the
time-to-the-right convention (recall eq. (1.67)). Instead, we could have chosen a Weyl
map with the definition
µ
W −1 (ep ) = eipµ x̂ (2.13)
which, as we saw in eq. (1.68), corresponds to a different choice of basis in momentum

space. Different Weyl maps can offer different advantages and disadvantages, depen-
ding on the problem at hand, for a more in depth discussion see (86), (80), (85),
(78).
In what follows, we will use the Weyl map defined as above, and we will assume
that it has already been applied.
2.3 Action for the complex scalar field

We can now go to point ii) in the roadmap in section 2.1. Because of what has been
said in section 2.2, and recalling that we need to get back the non-deformed action in
the formal κ → ∞ limit, we have two possible choices for the action, namely
Z
S1 = d4 x (∂ µ ϕ)† ⋆ (∂µ ϕ) − m2 ϕ† ⋆ ϕ (2.14)
R 4
Z
S2 = d4 x (∂µ ϕ) ⋆ (∂ µ ϕ)† − m2 ϕ ⋆ ϕ† . (2.15)
R4
2.4. Action of derivatives on ⋆ products 35
Because of this, and for reasons that will become apparent later, we choose the follo-
wing action.
Z
1
S= d4 x [(∂ µ ϕ)† ⋆ (∂µ ϕ) + (∂ µ ϕ) ⋆ (∂µ ϕ)† − m2 (ϕ† ⋆ ϕ + ϕ ⋆ ϕ† )] . (2.16)
2 R4
We now need to extract the equations of motion (EoM) and the conserved Noether
charges associated with continuous symmetries out of this action. Notice that the
quantities ∂ † are non-trivial objects which will be defined (and whose properties will
be investigated) in the next section.
2.4 Action of derivatives on ⋆ products

We now have a star product in a commutative spacetime, and we want to know
how derivatives act on the star product of two functions. We are at point iii) of our
roadmap in section 2.1. We know that the action cannot be Leibnizian, since otherwise
one would get the following contradiction.
i(p ⊕ q)µ ep⊕q = ∂µ (ep ⋆ eq ) = (∂µ ep ) ⋆ eq + ep ⋆ ∂µ eq = i(p + q)µ ep⊕q (2.17)
Indeed, from the first equality in eq. (2.17) we already see that the action must be
defined in terms of the co-product. Since we are using plane waves whose momenta
are in the classical basis, we will need to model the action of derivatives on products
over eq. (1.59), (1.60), (1.61). We define therefore the following κ-deformed Leibniz
rules, in analogy with those defined already1 in (86).
1
∂0 (ϕ ⋆ ψ) = (∂0 ϕ) ⋆ (∆+ ψ) + κ(∆−1 −1
+ ϕ) ⋆ (∂0 ψ) + i(∆+ ∂i ϕ) ⋆ (∂i ψ) (2.18)
κ
1
∂i (ϕ ⋆ ψ) = (∂i ϕ) ⋆ (∆+ ψ) + ϕ ⋆ (∂i ψ) (2.19)
κ
1
∆+ (ϕ ⋆ ψ) = (∆+ ϕ) ⋆ (∆+ ψ) . (2.20)
κ
Furthermore, we introduce the objects ∂ † which are related to antipodes. More in

detail, are defined in such a way that
∂ep ∝ pep ⇔ ∂ † ep ∝ S(p)ep . (2.21)
As such, we expect them to be related to eq. (1.64). We define
∂i† = κ∆−1
+ ∂i , ∂0† = ∂0 − i∆−1 2
+ ∂ , ∂4† = −∂4 , ∆†+ = κ2 ∆−1
+ (2.22)
∆+ = i∂0 + κ + i∂4 . (2.23)

1
In (86) the co-product rules in the bicrossproduct basis are used to define the action of derivatives
on products, instead of the classical basis that we use here. Furthermore, the model considered in
(86) is completely different from the present one, since they consider only one of the two orderings of
the action (the equivalent of our eq. (2.14)) and they have a completely different field decomposition.
The action of derivatives on products, however, is only a manifestation of the co-product, and it is
therefore model-independent.
Notice that for the definition of ∆+ we used the conventions highlighted in eq. (2.6).
If we define the linear combinations
∗ ∗ )x)
Ξ(x) = ap e−i(ωp t−px) + b†p∗ ei(S(ωp )t−S(p ≡ Ξ(+) (x) + Ξ(−) (x) (2.24)
∗ ∗ x)
Ξ† (x) = a†p e−i(S(ωp )t−S(p)x) + bp∗ ei(ωp t−p ≡ Ξ†(+) (x) + Ξ†(−) (x) (2.25)
then we have
† †
(∂A Ξ)† = ∂A Ξ. (2.26)
The reason for this particular choice of linear combination Ξ will become clear after
we introduce the complex fields in subsequent sections (at this point, a, a† , b, b† are
just some complex coefficients). We check the validity of eq. (2.26) for A = 0, i, 4.
2.4.1 A=0
In this case
∗ ∗ )x)
∂0 Ξ = −iωp ap e−i(ωp t−px) + iS(ωp∗ )b†p∗ ei(S(ωp )t−S(p (2.27)
which means that

∗ ∗ x)
(∂0 Ξ)† = iωp a†p e−i(S(ωp )t−S(p)x) − iS(ωp∗ )bp∗ ei(ωp t−p . (2.28)
On the other hand
∂0† Ξ† = (∂0 − i∆−1 2

+ ∂ )Ξ
†
(2.29)

i
= −iS(ωp ) + S(p) a†p e−i(S(ωp )t−S(p)x)
2
S(ωp ) + p4
| {z }
=iS(S(ωp ))

∗ i ∗ 2 ∗ ∗
+ iωp + ∗ ∗ (p ) bp∗ ei(ωp t−p x) (2.30)
−ωp − p4
| {z }
=−iS(ωp∗ )
where we also used S(p4 ) = p4 which is true both off-shell and on-shell, so that we
only need to verify that S(S(ωp )) = ωp . We have
1
S(S(ωp )) = −S(ωp ) + S(p)2 (2.31)
S(ωp ) + p4
p2 ωp + p 4 κ 2 p 2
= ωp − + (2.32)
ωp + p4 κ2 (ωp + p4 )2
= ωp (2.33)
where we used
p2 κ2
S(p0 ) + p4 = −p0 + + p4 = . (2.34)
p0 + p4 p0 + p4
Notice the important detail that the definition ∂ˆ4 ê∗k = i(κ−p∗4 (k)) ê∗k in equation (2.6)
translated through the Weyl map states that
∗ ∗ x) ∗ ∗ x)
∂4 e−i(ωp t−p = i(κ − p∗4 (k)) e−i(ωp t−p (2.35)
2.4. Action of derivatives on ⋆ products 37
while in eq. (2.29), when we use ∆−1 + in the second term in the RHS, we are applying
∗ ∗
∂4 to ei(ωp t−p x) = e−i(ωp t−px) , so that in the second term in the RHS of eq. (2.29)
we get the quantity
1 1
= . (2.36)
ωp + p 4 −ωp − p∗4
∗
2.4.2 A=i
In this case
∗ ∗ )x)
∂i Ξ = ipi ap e−i(ωp t−px) − iS(p∗ )i b†p∗ ei(S(ωp )t−S(p (2.37)
and therefore
∗ ∗ x)
(∂i Ξ)† = −ipi a†p e−i(S(ωp )t−S(p)x) + iS(p∗ )i bp∗ ei(ωp t−p . (2.38)
On the other hand we have
∂i† Ξ† = (κ∆−1
+ ∂i )Ξ
†
(2.39)

i
= κ S(p)i a†p e−i(S(ωp )t−S(p)x)
S(ωp ) + p4
| {z }
=−iS(S(p))i

i ∗ ∗ ∗
+ −κ p bp∗ ei(ωp t−p x) (2.40)
−ωp∗ − p∗4 i
| {z }
=iS(p)i
and also
i ωp + p4 κpi
κ S(p)i = −iκ = −ipi (2.41)
S(ωp ) + p4 κ2 ωp + p4
which concludes the verification.
2.4.3 A=4
We have
∗ ∗ )x)
∂4 Ξ = i(κ − p4 )ap e−i(ωp t−px) + i (κ − S(p4 )) b†p∗ ei(S(ωp )t−S(p (2.42)
| {z }
=κ−p4
where in the first term in the RHS we just used eq. (2.6) and in the second term we
used the fact that
∗ ∗ )x)
ei(S(ωp )t−S(p = e−i(S(ωp )t−S(p)x) (2.43)
and then we used again the definition (2.6). Therefore we have

∗ ∗ x)
(∂4 ϕ)† = −i(κ − p4 )a†p e−i(S(ωp )t−S(p)x) − i(κ − p4 )bp∗ ei(ωp t−p (2.44)
∂4† ϕ† = −∂4 ϕ† (2.45)

∗ ∗ x)
= −i(κ − S(p4 ))a†p e−i(S(ωp )t−S(p)x) − i(κ − p4 )bp∗ ei(ωp t−p (2.46)
which concludes the check.
2.4.4 A=+
From subsection (2.4.1) and (2.4.3) together with the definition (2.23) we have that
(∆+ ϕ)† = ∆†+ ϕ† . In fact, we know that (∂0 ϕ)† = ∂0† ϕ† then we also know that
(α∂0 ϕ)† = α† (∂0 ϕ)† = α† ∂0† ϕ† for any constant α, and the same goes for ∂4 .
2.4.5 ∆†+ = κ2 ∆−1

+
We have
∆†+ = −i∂0† + κ − i∂4† (2.47)

i∆−1 2
(2.48)

= −i ∂0 − + ∂ + κ + i∂4
∂2
= −i∂0 − + κ + i∂4 (2.49)
i∂0 + κ + i∂4
(2.50)
Furthermore, recall that by definition we have
(κ + i∂4 )2 = κ2 − ∂02 + ∂ 2 (2.51)
and therefore we end up with
∂02 − ∂ 2 + κ2 − ∂02 + ∂ 2 κ2
∆†+ = = (2.52)
∆+ ∆+
which proves the formula.
2.5 Integration-by-parts relation for ∂i going ←

We now need explicitly what are the integration by parts relations that one can get
using eq. (2.18), (2.19), (2.20), (2.22), (2.23). This is a preparatory step, and we will
need it after the introduction of the action of the complex scalar field in order to
compute the continuous charges.
For the spatial integration by parts we have
h i ∆ h i
+
(∂i ϕ)† ⋆ (∂i ψ) = ∂i (∂i ϕ)† ⋆ ψ − (∂ 2 ϕ)† ⋆ ψ . (2.53)
κ
which can be obtained using the following steps

(2.19)
h i 1
(∂i ϕ)† ⋆ (∂i ψ) = ∂i (∂i ϕ)† ⋆ ψ − [∂i (∂i ϕ)† ] ⋆ (∆+ ψ) (2.54)
κ
(2.20)
h i h i
= ∂i (∂i ϕ) ⋆ ψ − ∆+ (∆−1
†
+ ∂i (∂i ϕ) ) ⋆ ψ
†
(2.55)
h i ∆ h i
+
= ∂i (∂i ϕ)† ⋆ ψ − (κ∆−1
+ ∂ i (∂ i ϕ) †
) ⋆ ψ (2.56)
κ
(2.22)
h i ∆+ h i
= ∂i (∂i ϕ)† ⋆ ψ − (∂i† (∂i ϕ)† ) ⋆ ψ (2.57)
κ
2.6. Integration-by-parts relation for ∂0 going ← 39
h i ∆ h i
+
= ∂i (∂i ϕ)† ⋆ ψ − (∂ 2 ϕ)† ⋆ ψ . (2.58)
κ
2.6 Integration-by-parts relation for ∂0 going ←

We do the same thing for the temporal part. We have
∂0 h i ∆+ h 2 † i
(∂0 ϕ)† ⋆ (∂0 ψ) = (∆+ (∂0 ϕ)† ) ⋆ ψ − i∂i [(∆−1 ∂
+ i 0∂ ϕ)†
⋆ ψ] − (∂ 0 ϕ) ⋆ ψ
κ κ
(2.59)
which can be proven using the following steps.
κ (∂0 ϕ)† ⋆ (∂0 ψ) (2.60)

(2.18)
h i 1
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − (∆+ ∂0 (∂0 ϕ)† ) ⋆ (∆+ ψ) − i(∂i (∂0 ϕ)† ) ⋆ (∂i ψ) (2.61)
κ
(2.20)
h i
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − ∆+ [(∂0 (∂0 ϕ)† ) ⋆ ψ] − i(∂i (∂0 ϕ)† ) ⋆ (∂i ψ) (2.62)
(2.19)
h i
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − ∆+ [(∂0 (∂0 ϕ)† ) ⋆ ψ] − i∂i [(∂i (∂0 ϕ)† ) ⋆ ψ]
i 2
+ (∂ (∂0 ϕ)† ) ⋆ (∆+ ψ) (2.63)
κ
(2.20)
h i
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − ∆+ [(∂0 (∂0 ϕ)† ) ⋆ ψ] − i∂i [(∂i (∂0 ϕ)† ) ⋆ ψ]
+ i∆+ [(∆−1 2 †
+ ∂ (∂0 ϕ) ) ⋆ ψ] (2.64)
h i
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − i∂i [(∂i (∂0 ϕ)† ) ⋆ ψ]
n o
− ∆+ [(∂0 − i∆−1 + ∂ 2
)(∂ 0 ϕ) †
] ⋆ ψ (2.65)
(2.22)
h i
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − i∂i [∆+ κ−1 (∂i ∂0 ϕ)† ⋆ ψ]
n o
− ∆+ [(∂0 − i∆−1 + ∂ 2
)(∂ 0 ϕ) †
] ⋆ ψ (2.66)
(2.22)
h i
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − iκ∂i [(∆−1 + ∂i ∂0 ϕ) ⋆ ψ]
†
n o
− ∆+ [(∂0 − i∆−1 2 †
+ ∂ )(∂0 ϕ) ] ⋆ ψ (2.67)
(2.22)
h i n o
†
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − iκ∂i [(∆−1 + ∂ i ∂ 0 ϕ) †
⋆ ψ] − ∆ + [∂ 0 (∂ 0 ϕ) †
] ⋆ ψ (2.68)
h i h i
= ∂0 (∆+ (∂0 ϕ)† ) ⋆ ψ − iκ∂i [(∆−1 + ∂ i ∂ 0 ϕ) †
⋆ ψ] − ∆ + (∂ 2 †
0 ϕ) ⋆ ψ . (2.69)
2.7 Integration-by-parts relation for ∂i going →

Working in the same way, we obtain
(∂i ψ) ⋆ (∂i ϕ)† = κ∂i (ψ ⋆ [∆−1 † 2 †

+ (∂i ϕ) ]) − ψ ⋆ (∂ ϕ) . (2.70)
Indeed, we have
1 (2.19)
(∂i ψ) ⋆ (∂i ϕ)† = ∂i (ψ ⋆ [∆−1 † −1 †
+ (∂i ϕ) ]) − ψ ⋆ [∆+ ∂i (∂i ϕ) ] (2.71)
κ
and therefore
(2.19)
(∂i ψ) ⋆ (∂i ϕ)† = κ∂i (ψ ⋆ [∆−1 † −1 †
+ (∂i ϕ) ]) − ψ ⋆ [κ∆+ ∂i (∂i ϕ) ] (2.72)
(2.22)
= κ∂i (ψ ⋆ [∆−1 † 2 †
+ (∂i ϕ) ]) − ψ ⋆ (∂ ϕ) . (2.73)
2.8 Integration-by-parts relation for ∂0 going →

One can show that
(∂0 ψ) ⋆ (∂0 ϕ)† = ∂0 (ψ ⋆ [κ∆−1 † −1 †

+ (∂0 ϕ) ]) − i∂i (ψ ⋆ [∆+ ∂i (∂0 ϕ) ])

i ∂4
− [ψ ⋆ (∂02 ϕ)† ] + ∂0 + i [ψ ⋆ (∂02 ϕ)† ] (2.74a)
κ κ
Indeed we have
1 (2.18)
(∂0 ψ) ⋆ (∂0 ϕ)† = ∂0 (ψ ⋆ [∆−1 † −1 −1 †
+ (∂0 ϕ) ]) − κ(∆+ ψ) ⋆ (∆+ ∂0 (∂0 ϕ) )
κ
− i(∆−1 −1 †
+ ∂i ψ) ⋆ (∆+ ∂i (∂0 ϕ) ) (2.75)
(2.20)
= ∂0 (ψ ⋆ [∆−1 † −1 †
+ (∂0 ϕ) ]) − ∆+ [ψ ⋆ (∂0 (∂0 ϕ) )]
−1
− iκ−1 ∆+ [(∂i ψ) ⋆ (∂i (∂0 ϕ)† )] (2.76)
(2.19)
= ∂0 (ψ ⋆ [∆−1 † −1
+ (∂0 ϕ) ]) − ∆+ [ψ ⋆ (∂0 (∂0 ϕ) )]
†
h i
− i∆−1
+ ∂i (ψ ⋆ [∆ −1
+ ∂ i (∂ 0 ϕ)†
]) − ψ ⋆ [∆−1 2
+ ∂ (∂ 0 ϕ) †
] (2.77)
= ∂0 (ψ ⋆ [∆−1 † −1 −1 †
+ (∂0 ϕ) ]) − i∆+ ∂i (ψ ⋆ [∆+ ∂i (∂0 ϕ) ])
− ∆−1 −1 2 †
+ [ψ ⋆ (∂0 − i∆+ ∂ )(∂0 ϕ) ] (2.78)
(2.22)
= ∂0 (ψ ⋆ [∆−1 † −1 −1 †
+ (∂0 ϕ) ]) − i∆+ ∂i (ψ ⋆ [∆+ ∂i (∂0 ϕ) ])
− ∆−1 2 †
+ [ψ ⋆ (∂0 ϕ) ] (2.79)
We now expand ∆−1 + (recall that ∆+ = i∂0 + κ + i∂4 ). By expanding we mean writing
the differential operator κ∆−1
+ in terms of a sum of derivatives in the numerator. In
other words, after applying κ∆−1
+ to a star product, we get some function of momenta
at the denominator, and we expand such a function around a point in momentum
space which is convenient for us. At this point, we convert this function back to
action of derivatives on the original star product. The net effective result is that we
have expanded out original differential operator as if it was already a function of
momenta. We have
1
κ∆−1
+ = (2.80)
i ∂κ0 + 1 + i ∂κ4
Notice that we cannot expand around (∂0 , ∂i ) = (0, 0) (here we abuse notation a bit
by addressing derivatives as if they are already momenta, but one should always keep
in mind our previous discussion) because in this case m = 0 on-shell, and we are
2.9. Mass terms integration by parts 41
assuming m ̸= 0 instead. Furthermore, notice that if we expand around ∂0 = 0, then

the expansion up to first order would be
δκ∆−1
+
κ∆−1 −1
+ = κ∆+ (∂0 = 0) + ∂0 + O(∂02 ) (2.81)
δ∂0
∂0 =0
where we used eq. (2.6), and in particular the zeroth order would become
1
κ∆−1
+ (∂0 = 0) = q ̸= 1 (2.82)
m2
1− 1+ κ2
which is however problematic, because one would not recover the correct equation
of motion (since the temporal component of the kinetic term will have a different
rescaling than the spatial part). We need to find an expansion point p such that
†
κ∆−1
+ (p) = 1. Therefore, we expand around ∂0 = −∂4 = ∂4 . We have
δκ∆−1
+
κ∆−1 −1
+ = κ∆+ (∂0 = −∂4 ) + (∂0 − (−∂4 )) + O((∂0 + ∂4 )2 ) (2.83)
δ∂0
∂0 =−∂4
i
i ∂4
=1− κ
2 (∂0 + ∂4 ) = 1 − ∂0 − i (2.84)
κ κ
i ∂κ0 + 1 + i ∂κ4 ∂0 =−∂4
Coming back to our integration-by-parts relation, without loss of generality we can

write
(∂0 ψ) ⋆ (∂0 ϕ)† = ∂0 (ψ ⋆ [κ∆−1 † −1 † 2 †

+ (∂0 ϕ) ]) − i∂i (ψ ⋆ [∆+ ∂i (∂0 ϕ) ]) − [ψ ⋆ (∂0 ϕ) ]

i ∂4
+ i∂i ∂0 + i (ψ ⋆ [∆−1 †
+ ∂i (∂0 ϕ) ])
κ κ

i ∂4
+ ∂0 + i [ψ ⋆ (∂02 ϕ)† ] (2.85)
κ κ

However, the term i∂i κi ∂0 + i ∂κ4 (ψ ⋆ [∆−1 + ∂i (∂0 ϕ) ]) can be ignored because it con-
†
tain a double total derivative (we are considering ∂4 as an independent derivative with
respect to the others, because indeed p4 can have a value independent from the value
of p0 or pi ). This proves eq. (2.74).
Notice that, since we do not have any prefactor in front of the kinetic part of the
on-shell operator, we also don’t need to modify the mass term to obtain the correct
on-shell relation, like it was needed in the previous ordering. Furthermore, notice also
that we only expanded the ∆−1 + which acted globally on each star product, while
there would be no meaning in expanding the factors ∆−1 + inside the star product.
Furthermore, this expansion does not mean that our results are approximate. Indeed,
higher order term in the expansion of κ∆−1 + would contribute factors proportional to
∂ 2 or higher, and all these contributions vanish in any case because of the integral.
2.9 Mass terms integration by parts

Recalling that ∆+ = i∂0 + κ + i∂4 for the first ordering we have
i∂0 2 † i∂4 2 † ∆+ 2 †
m2 ϕ† ⋆ ψ = − (m ϕ ⋆ ψ) − (m ϕ ⋆ ψ) + (m ϕ ⋆ ψ) (2.86)
κ κ κ
which can be proven trivially from

2 † ∆+ ∆+ 2 †
m ϕ ⋆ψ =− − 1 (m2 ϕ† ⋆ ψ) + (m ϕ ⋆ ψ). (2.87)
κ κ
For the second ordering we don’t have a contribution coming from this reasoning.
2.10 EoM and surface terms Π0 , Πi , and Π4 for both or-

derings of the action
We now have all the necessary tools, i.e. eq. (2.53), (2.59), (2.70), (2.74), (2.86), to
integrate by parts both orderings of the action, i.e. eq. (2.14) and (2.15), which we
rewrite here for simplicity
Z
(2.14) → S1 = d4 x (∂ µ ϕ)† ⋆ (∂µ ϕ) − m2 ϕ† ⋆ ϕ (2.88)
R4
Z
(2.15) → S2 = d4 x (∂µ ϕ) ⋆ (∂ µ ϕ)† − m2 ϕ ⋆ ϕ† . (2.89)
R4
We are at point iv) of our roadmap in section 2.1. We can now compute the variation
of the action in eq. (2.16)
Z
1
δS1 = d4 x (∂µ δϕ)† ⋆ ∂ µ ϕ + (∂µ ϕ)† ⋆ ∂ µ δϕ − m2 δϕ† ⋆ ϕ − m2 ϕ† ⋆ δϕ (2.90)
2 R4
which can be immediately rewritten using integration by parts (2.53), (2.59), (2.86)
as2
Z (
1 ∆+ h † µ † i
4
(∂µ (∂ ) − m2 )ϕ† ⋆ δϕ + ∂A ΠA

δS1 = d x − 1 ⋆ δϕ
2 R4 κ
)
κ h † i
†
†

δϕ ⋆ (∂µ ∂ µ − m2 )ϕ + ∂A δϕ† ⋆ ΠA (2.91)

− 1
∆+
where
1
Π01 = (Π0 )1 = (∆+ ∂0† + im2 )ϕ† (2.92)
κ
Πi1 = −(Πi )1 = (−∂i (1 + i∆−1
+ ∂0 ))ϕ
†
(2.93)
m 2 ϕ†
Π41 = (Π4 )1 = −i . (2.94)
κ
For the other ordering we proceed in a similar manner, with the only difference that
we use the integration-by-parts relations (2.70), (2.74). We have
Z
1
δS2 = d4 x ∂ µ ϕ ⋆ (∂µ δϕ)† + ∂ µ δϕ ⋆ (∂µ ϕ)† − m2 ϕ ⋆ δϕ† − m2 δϕ ⋆ ϕ† (2.95)
2 R4
2
One can integrate by parts one of the two terms, and then obtain the other by taking the
Hermitian conjugate. Of course, the same relation is obtained if one starts with the co-product rules
for the derivatives ∂ † , and then proceeds from there.
2.11. Complex scalar field and its properties 43
which can be rewritten as

Z (
1 h i
d x − δϕ ⋆ (∂µ† (∂ µ )† − m2 )ϕ† + ∂A δϕ ⋆ ΠA
4

δS2 =
2 R4
)
h i
†
†
− (∂µ ∂ µ − m2 )ϕ ⋆ δϕ† + ∂A ΠA ⋆ δϕ† (2.96)
where

κ † i † 2
Π02= (Π0 )2 = ∂ + (∂ ) ϕ† (2.97)
∆+ 0 κ 0
κ
Πi2 = −(Πi )2 = − (∂ † + i∂i ∂0† )ϕ† (2.98)
∆+ i
(∂0† )2 †
Π42 = (Π4 )2 = +i ϕ. (2.99)
κ
We see that the equations of motion are the canonical Klein-gordon equation
(∂µ ∂ µ − m2 )ϕ = 0 (∂µ† (∂ µ )† − m2 )ϕ† = 0. (2.100)
Notice that in momentum space, ∂µ† (∂ µ )† ↔ S(p)µ S(p)µ = pµ pµ ↔ ∂µ ∂ µ ), where one

can check explicitly that
S(p)µ S(p)µ = S(ωp )2 − S(p)2 (2.101)

2 2
p2

κp
= −ωp + − − (2.102)
ωp + p4 ωp + p4
2 4
ωp p p κ2 p2
= ωp2 − 2 + − (2.103)
ωp + p4 (ωp + p4 )2 (ωp + p4 )2
−(ωp + p4 )2
= ωp2 + p2 (2.104)
(ωp + p4 )2
= pµ pµ (2.105)
and therefore
S(p)µ S(p)µ = pµ pµ (2.106)
The fields satisfying such equations are usually written as linear combinations of plane
waves and their Hermitian conjugate. Here, because of the curved nature of momentum
space, it is not so automatic. Nevertheless, we will see in the next section that fields
satisfying these equations, in the context of κ-deformation, can be described in a
particularly clear way.
2.11 Complex scalar field and its properties

We are now in point v) of our roadmap. As it is clear from eq. (1.56), the AN (3) group
manifold is a 4-dimensional submanifold of a 5-dimensional space. Furthermore, we
have plane waves, i.e. group elements g ∈ AN (3). To define a field on it which satisfies
the EoM in eq. (2.100) we can use both these facts and define (78)
Z
ϕ(x) = d5 p 2κδ(pA pA + κ2 )θ(p+ )θ(p4 )δ(pµ pµ − m2 )ϕ̃(p)e−ipx (2.107)
where ϕ(x) is a function in commutative spacetime. Notice that the quantity
d5 p 2κδ(pA pA + κ2 ) (2.108)
is the metric in R5 restricted to the group manifold with the constraint in eq. (1.56)
(the κ is only present for dimensional reasons). In the same way, the Heaviside θ
functions restrict ourselves to the correct part of the manifold, namely where P+ > 0
and P4 > 0. Finally, the remaining Dirac delta imposes the on-shell condition. In order
to write this field in a more familiar way, one can first split the constraints as follows.
δ(pµ pµ − m2 ) = δ(pµ pµ − m2 )θ(p0 − m) + δ(pµ pµ − m2 )θ(−p0 − m) (2.109)
We use the notation

Z
ϕ+ (x) = d5 p 2κδ(pA pA + κ2 )θ(p+ )θ(p4 )δ(pµ pµ − m2 )θ(p0 − m)ϕ̃(p)e−ipx (2.110)
Z
ϕ− (x) = d5 p 2κδ(pA pA + κ2 )θ(p+ )θ(p4 )δ(pµ pµ − m2 )θ(−p0 − m)ϕ̃(p)e−ipx
(2.111)
for simplicity. We can then change variable in ϕ− by sending p 7→ S(p). Because of

eq. (2.106) and (1.64), we know that the arguments of the Dirac delta are invariant.
One can also show starting from eq. (1.64) that
p2 κ2
S(p+ ) = −p0 + + S(p)4 = (2.112)
p+ p+
so that also the Heaiside θ(p+ ) and θ(p4 ) are left invariant. One can then change va-
riable p 7→ −p = p∗ and pick up some additional factors coming from the determinant
of the Jacobian obtained in the coordinate change p 7→ S(p). Furthermore, one can
apply the Dirac deltas, reducing the integration from d5 p to d3 p, with other factors
added to the integrand (more details can be found in (78)). The result is that the field
can be written as
d3 p d3 p∗
Z Z
∗
ϕ(x) = p ζ(p) ap e −i(ωp t−px)
+ p ζ(p)b†p∗ ei(S(ωp∗ )t−S(p )x) (2.113)
2ωp 2|ωp∗ |
d3 p d3 p∗
Z Z
† ∗
ϕ (x) = p ζ(p) a†p e−i(S(ωp )t−S(p)x) + p ζ(p)bp∗ ei(ωp∗ t−p )x
2ωp 2|ωp∗ |
(2.114)
Notice that if p > 0, then S(p∗ ) > 0, so that both ϕ(x) and ϕ† (x) are a linear
combination of positive and negative energy states. Notice that there is an additional
function ζ(p) which will be important when talking about the action. Furthermore,
notice also that the ∗ can be eliminated by all the expressions by simply change
variable p∗ 7→ −p. Therefore, it may be convenient to use instead the fields
d3 p d3 p
Z Z
ϕ(x) = p ζ(p) ap e−i(ωp t−px) + p ζ(p)b†p e−i(S(ωp )t−S(p)x) (2.115)
2ωp 2ωp
2.11. Complex scalar field and its properties 45
d3 p d3 p
Z Z
†
ϕ (x) = p ζ(p) a†p e−i(S(ωp )t−S(p)x) + p ζ(p)bp e−i(ωp t−p)x (2.116)
2ωp 2ωp
for simplicity. The reason for using the variables p∗ instead of −p is to make clear that
the minus sign in −p needs to also be present in front of p4 , and not only p0 , p. This
is important because we need the relation defined in (1.91) to hold. Having clarified
this point, we can safely switch from p∗ to −p without issues.
2.11.1 C, P, T in the non-deformed context: a short review

We provide a short review of the definition of C, P, T transformations in the non-
deformed case, which will be the basis for our treatment of the deformed case. The
non-deformed fields are the same as eq. (2.115), (2.116) but in the limit κ → ∞, so
we have
d3 p d3 p
Z Z
−i(ωp t−px)
ϕU (x) = p ap e + p b†p ei(ωp t−p)x) (2.117)
2ωp 2ωp |
d3 p d3 p
Z Z
ϕ†U (x) = p a†p ei(ωp t−px) + p bp e−i(ωp t−p)x . (2.118)
2ωp 2ωp
(P ) By definition, a P transformation is defined by (t, x) 7→ (t, −x) and therefore

Pϕ(t, x)P −1 = ϕ(t, −x).
d3 p
Z
−1
PϕU (t, x)P = p Pap P −1 e−i(ωp t−px) + Pb†p P −1 ei(ωp t−px)
2ωp
(2.119)
Z 3
d p
!
= p ap e−i(ωp t+px) + b†p ei(ωp t+px) (2.120)
2ωp
which means that
Pap P −1 = a−p Pbp P −1 = b−p (2.121)
and the same for a† , b† .
(T ) By definition, a T transformation is defined by (t, x) 7→ (−t, x) and therefore

T ϕ(t, x)T −1 = ϕ(−t, x). Recall that this is an anti-unitary operator. One can
show easily this because T xT −1 = x, T pT −1 = −p imply
T !
[x, p] = i −→ −[x, p] = T iT −1 = −i. (2.122)
We have
d3 p
Z
−1
T ϕ(t, x)T = p T ap T −1 e+i(ωp t−px) + T b†p T −1 e−i(ωp t−px) (2.123)
2ωp
d3 p
Z
!
= p ap e−i(−ωp t−px) + b†p ei(−ωp t+px) (2.124)
2ωp
which means that
T ap T −1 = a−p T bp T −1 = b−p (2.125)

(C) Contrary to the previous two transformation, this is not related to spacetime
coordinates. Its action is defined by the relation C ϕC −1 = ϕ† . We have
d3 p
Z
−1
C ϕ(t, x)C = p C ap C −1 e−i(ωp t−px) + C b†p C −1 e+i(ωp t−px) (2.126)
2ωp
d3 p † +i(ωp t−px)
Z
!
= p ap e + bp e−i(ωp t+px) (2.127)
2ωp
which means that
C ap C −1 = bp C bp C −1 = ap (2.128)
2.11.2 C, P, T in the κ-deformed context

The P and T transformations can consistently be defined as acting like in the non-
deformed case (they leave [x̂0 , x̂i ] = κi x̂i invariant). Therefore, they behave as in the
non-deformed case analysed before
T ϕ(t, x)T −1 = ϕ(−t, x) =⇒ T ap T −1 = a−p (2.129)

−1 −1
Pϕ(t, x)P = ϕ(t, −x) =⇒ Pap P = a−p (2.130)
The only non-trivial transformation is therefore the C transformation. However, be-

cause of the presence of the antipode S(·) in the fields in eq. (2.115), (2.116), also
C can be shown to behave like in the non-deformed case (in its action on a, a† , b, b† )
when acting on fields.
Indeed, take the ϕ+ (x) component first
d3 p
Z
−1
C ϕ+ (x)C = p ζ(p) C ap C −1 e−i(ωp t−px) . (2.131)
2ωp

d3 p
Z
ϕ†− (x) = p ζ(p)bp e−i(ωp t−p)x (2.132)
2ωp
which means that
C ap C −1 = bp . (2.133)
At the same time, we also have
d3 p
Z
−1
C ϕ− (x)C = p ζ(p)C b†p C −1 e−i(S(ωp )t−S(p)x) (2.134)
2ωp
and
d3 p
Z
ϕ†+ (x) = p ζ(p) a†p e−i(S(ωp )t−S(p)x) (2.135)
2ωp
2.12. Off-shell action in momentum space 47
which means that
C b†p C −1 = a†p . (2.136)
Notice that the definition of the C transformation is trivial also in the deformed case,
and we stress once again that this is due to the presence of the antipode in the field
definition. This is in contrast with other approaches to charge conjugation, which do
not give rise to such simple transformation rules.
2.12 Off-shell action in momentum space

It is instructive (and we will need it in chapter 3) to compute the off-shell action in
momentum space (we are in point vi) of our roadmap in section 2.1). This gives a
unique insight on the on-shell relation in momentum space, as well as a clear picture of
the invariance of the action under both discrete C, P, T transformation, and continuous
κ-Poincaré transformations.
To do so we will need the off-shell field. We cannot use the same construction as
before because it includes the Dirac delta δ(pµ pµ − m2 ). Therefore, we define
d4 p
Z n o
ϕof f (x) = ζ(p) ap e−i(ωp t−px) + b†p e−i(S(ωp )t−S(p)x) (2.137)
I+ p4 /κ
d4 p
Z n o
ϕ†of f (x) = ζ(p) a†p e−i(S(ωp )t−S(p)x) + bp e−i(ωp t−px) . (2.138)
I+ p4 /κ
Since we have two ordering of the action, namely eq. (2.14) and (2.15), we will compute
them separately.
2.12.1 First ordering of the action

† †
Using the fact that (∂A ϕ)† ≡ ∂A ϕ and substituting eq. (2.137), (2.138) into (2.14)
we get
(
d4 p d4 q
Z Z
S= d x 4
(pµ q µ − m2 )a†p aq e−i(S(ωp )⊕ωq )t ei(S(p)⊕q)x (2.139)
p4 /κ q4 /κ
− (pµ S(q)µ − m2 )a†p b†q e−i(S(ωp )⊕S(ωq ))t ei(S(p)⊕S(q))x (2.140)
− (S(p)µ q µ − m2 )bp aq e−i(ωp ⊕ωq )t ei(p⊕q)x (2.141)
)
+ (S(p)µ S(q)µ − m2 )bp b†q e−i(ωp ⊕S(ωq ))t e−i(p⊕S(q))x ζ(p)ζ(q) (2.142)
For the moment we will ignore the terms (2.140) and (2.141). Their treatment is quite
non-trivial, and we will defer their treatment to future studies. Therefore, integrating
in d4 x,
(
d4 p d4 q
Z
S= (pµ q µ − m2 )a†p aq δ(S(ωp ) ⊕ ωq )δ(S(p) ⊕ q) (2.143)
p4 /κ q4 /κ
)
+ (S(p)µ S(q)µ − m2 )bp b†q δ(ωp ⊕ S(ωq ))δ(p ⊕ S(q)) ζ(p)ζ(q) (2.144)
(
d4 p d4 q
Z
= (pµ q µ − m2 )a†p aq δ(S(p) ⊕ q) (2.145)
p4 /κ q4 /κ
)
+ (S(p)µ S(q)µ − m2 )bp b†q δ(p ⊕ S(q)) ζ(p)ζ(q) (2.146)
where we used the notation
δ(S(ωp ) ⊕ ωq )δ(S(p) ⊕ q) = δ(S(p) ⊕ q) (2.147)

δ(ωp ⊕ S(ωq ))δ(p ⊕ S(q)) = δ(p ⊕ S(q)). (2.148)
We now use a known result in the literature (85) which states that
|p+ |3 p4
δ(p ⊕ S(q)) = δ(S(p) ⊕ q) δ(S(p) ⊕ q) = δ(p − q) (2.149)
κ3 κ
The additional pκ4 factor eliminates one of the factors in one of the two metrics, so
that the final expression for the firs off-shell action in momentum space is
d4 p 2 † |p+ |
3
Z
2 †
S= µ µ 2
ζ (p) (pµ p − m )ap ap + (S(p)µ S(p) − m )bp bp 3 (2.150)
p4 /κ κ
where the red term is the factor coming from the delta.
2.12.2 Second ordering of the action

Using the switched convention, i.e. substituting eq. (2.137) and (2.138) into eq. (2.15),
we obtain
(
d4 p d4 q
Z Z
S= d x 4
(pµ q µ − m2 )aq a†p e−i(ωq t−qx) ⋆ e−i(S(ωp )t−S(p)x) (2.151)
p4 /κ q4 /κ
)
+ (S(p)µ S(q)µ − m2 )b†q bp ei(S(ωq )t−S(q)x) ⋆ ei(ωp t−px) ζ(p)ζ(q) (2.152)
(
d4 p d4 q
Z Z
= 4
d x (pµ q µ − m2 )aq a†p e−i(ωq ⊕S(ωp ))t ei(q⊕S(p))x (2.153)
p4 /κ q4 /κ
)
+ (S(p)µ S(q)µ − m2 )b†q bp e−i(S(ωq )⊕ωp )t e−i(S(q)⊕p)x ζ(p)ζ(q) (2.154)
The two missing terms have been eliminated for the same reason as before, and we
will study them in detail in forthcoming publications.
These are exactly the same terms as in the previous calculations with only two
exceptions:
1) In the previous action we had the products a†p aq and bp b†q , while with this
switched convention we have the other order, i.e. aq a†p and b†q bp ;
2) The Dirac deltas are switched with respect to the computations with the pre-
vious ordering. More explicitly, the exponential multiplying the term aq a†p here
is the same that was multiplying the term bp b†q , and vice versa.
2.13. Symplectic form for the two actions 49
Therefore, all the computations proceed in the same way but the extra coefficient is
now found with the a, a† operators, i.e. we have
d4 p 2 |p+ |3
Z
2 † †
S= ζ (p) µ µ 2
(pµ p − m )ap ap + (S(p)µ S(p) − m )bp bp (2.155)
p4 /κ κ3
Notice that the on-shell conditions are not swapped, because their formulation only
† †
depends on the property (∂A ϕ)† = ∂A ϕ , and not on the ordering of the star products.
2.12.3 Sum of the two actions

Using the definition of the total action given by eq. (2.16), and substituting eq. (2.150)
and (2.155), we obtain (recall that we are still in the classical case, but upon quanti-
zation one should remember that normal-ordering should be applied)
d4 p |p+ |3
Z n o
S= 1 + 3 ζ 2 (p) (pµ pµ − m2 )a†p ap + (S(p)µ S(p)µ − m2 )b†p bp .
p4 /κ κ
(2.156)
If we decided to go on-shell, the computations would proceed as usual, but with an

additional Dirac delta imposing the shell condition, which means that the action would
be exactly zero.
2.13 Symplectic form for the two actions

We are now at point vii)a) of our roadmap in section 2.1.
2.13.1 Example in the non-deformed case

Let us consider a complex scalar field with the Lagrangian

L = − ∂ µ ϕ† ∂µ ϕ + m2 ϕ† ϕ (2.157)
The variation of this Lagrangian is

δL = − ∂ µ δϕ† ∂µ ϕ + ∂ µ ϕ† ∂µ δϕ + m2 δϕ† ϕ + m2 ϕ† δϕ (2.158)
This variation contains terms proportional to the field equations and the total deri-
vative term
δL = EoM − ∂ µ δϕ† ∂µ ϕ − ∂µ ∂ µ ϕ† δϕ (2.159)
Therefore the variation of the action is

Z Z
δS = EoM − dσµ δϕ† ∂ µ ϕ + ∂ µ ϕ† δϕ (2.160)
M ∂M
The second, boundary term in this expression is the so-called presymplectic structure
or Liouville form θ, if we choose the boundary ∂M to be a Cauchy surface, for example
the surface of constant time t = 0. Then we can find the symplectic form
Z
1 † ∂ ∂ †
ω = δθ = 3
d x δϕ ∧ δϕ − δϕ ∧ δϕ (2.161)
2 t=0 ∂t ∂t t=0
Now we take the field decomposition
d3 p
Z
ϕ(t, x) = √ ap e−i(Et−px) + b†p ei(Et−px) (2.162)
2E
so that
d3 p
Z
δϕ(t, x) = √ δap e−i(Et−px) + δb†p ei(Et−px) (2.163)
2E
and plug it to (2.161) to obtain the symplectic form, whose inverse defines for us the
Poisson brackets, and (upon quantization) the commutators.
Z Z
i
ω= d x d3 p d3 q δa†p e−ipx + δbp eipx ∧ −δaq eiqx + δb†q e−iqx
3
2
Z Z
i
− d x d3 q d3 p δa†p e−ipx − δbp eipx ∧ δaq eiqx + δb†q e−iqx
3
2
Defining Z
Ω= d3 pd3 q ω(p, q) (2.164)
we find
Ω(p, q) ∼ −δa†p ∧ δaq − δb†p ∧ δbq (2.165)
from which the commutational relations
[ap , a†q ] ∼ [bp , b†q ] ∼ δ(p − q) (2.166)
2.13.2 The κ-deformed case

The extension of this reasoning to the case of fields in the context of κ-Poincaré
algebra is straightforward. In particular we have already computed the surface term
in previous sections, and we are only interested in those terms whose boundary is
spacelike, i.e. we only need to consider the following presymplectic forms.
Z †
θ1 = − d3 x Π01 ⋆ δϕ + δϕ† ⋆ Π01
R3
( )
Z
1
1 †
† † † † †
=− 3
d x 2
∆+ ∂0 + im ϕ ⋆ δϕ + δϕ ⋆ 2
∆+ ∂0 + im ϕ . (2.167)
R3 κ κ
Z
†
θ2 = − d3 x δϕ ⋆ Π02 + Π02 ⋆ δϕ†
R3
Z ( † )
κ † i † 2 κ † i †
=− d3 x δϕ ⋆ ∂ + (∂ ) ϕ† + ∂ + (∂ )2 ϕ† ⋆ δϕ† .
R3 ∆+ 0 κ 0 ∆+ 0 κ 0
(2.168)
Notice that the sign in the RHS of these two presymplectic forms is such that we get
the correct symplectic form in the canonical limit κ → ∞ (and therefore the correct
canonical commutator between a and a† ,a nd same for b and b† ). Furthermore, we
are not writing down explicitly the wedge product for the moment because at the
moment it is not present, it will be when we will consider δθ1 and δθ2 . To simplify
the computations, notice that
∆+ † im2

Π01 = ∂ + ϕ† (2.169)
κ 0 κ
†
2 ϕ
= ∆+ (∂0 − i∆−1 + ∂ 2
) + im (2.170)
κ
∆+ i 2
= ∂0 + (m − ∂ 2 ) ϕ† (2.171)
κ κ
i∂ 2

∆+
= ∂0 − 0 ϕ† (2.172)
κ κ
i∂02

i∂0 + κ + i∂4
= ∂0 − ϕ† (2.173)
κ κ

κ + i∂4
= ∂0 ϕ† . (2.174)
κ
and in the same way one gets

!†
∆+ † im2 κ † i(∂0† )2

κ + i∂4
∂0 = ∂ + = ∂ + (2.175)
κ κ 0 κ ∆+ 0 κ
and therefore
!
κ † i(∂0† )2
†
κ + i∂4
∂ + = ∂0 . (2.176)
∆+ 0 κ κ
Notice that the quantity ∂4 acting on any plane wave brings down iκ − ip4 , and since
p4 = S(p4 ), then κ + i∂4 → κ + i(iκ − ip4 ) = p4 (recall that S(p4 ) = p4 ). The
presymplectic forms can therefore be simplified to
Z
p4
θ1 = − d3 x ∂0 ϕ† ⋆ δϕ + δϕ† ⋆ ∂0† ϕ (2.177)
R3 κ
Z
p4
θ2 = − d3 x δϕ ⋆ ∂0† ϕ† + ∂0 ϕ ⋆ δϕ† . (2.178)
R3 κ
Taking the delta, we get

Z
p4 ∧ ∧

δθ1 = − d3 x δ∂0 ϕ† ⋆ δϕ − δϕ† ⋆ δ∂0† ϕ (2.179)
R3 κ
Z
p4 ∧ ∧

δθ2 = − d3 x −δϕ ⋆ ∂0† δϕ† + ∂0 δϕ ⋆ δϕ† . (2.180)
R3 κ
∧
The notation ⋆ means that we are still taking the ⋆ product among plane waves, and
also the wedge product between δa, δa† , δb, δb† . We can now explicitly compute both
symplectic forms using the fields in eq. (2.115), (2.116). Starting from the first, we
have
Z
p4 n ∧ ∧
o
δθ1 = − δ∂0 ϕ† ⋆ δϕ − δϕ† ⋆ δ∂0† ϕ (2.181)
R3 κ
d3 p d3 q
Z Z
p4
= −i p p ζ(q) ζ(p) ×
R3 2ωp 2ωq κ
n
(−S(ωq ) − S(ωp ))a†q ∧ ap e−i(S(ωq )⊕ωp )t ei(S(q)⊕p)x (2.182)
+ (−ωq − ωp )bq ∧ b†p e−i(ωq ⊕S(ωp ))t ei(q⊕S(p))x (2.183)
− (S(ωq ) + ωp )a†q ∧ b†p e−i(S(ωq )⊕S(ωp ))t ei(S(q)⊕S(p))x (2.184)
o
− (ωq + S(ωp ))aq ∧ bp e−i(ωq ⊕ωp )t ei(q⊕p)x (2.185)
We now use
Z
e−i(S(ωq )t−S(q)x) ⋆ e−i(ωp t−px) = δ (p − q) (2.186)
R3
|q+ |3
Z
e−i(ωp t−px) ⋆ e−i(S(ωq )t−S(q)x) = δ (p − q) (2.187)
R3 κ3
so that the first two terms give
d3 p 2 p4 p3+
Z
† †
δθ1 = i ζ (p) 2S(ωp )ap ∧ ap + 3 2ωp bp ∧ bp (2.188)
2ωp κ κ
and the other two go away because the Dirac delta impose on-shell that ωq = −S(ωp ).
Indeed, one can check for example that
Z q
+
e−iωp t+ipx ⋆ e−iωq t+iqx = e−i(ωp ⊕ωq )t δ(p ⊕ q) = e−i(ωp ⊕ωq )t δ p + q (2.189)
R3 κ
If we first integrate in d3 p, the quantity |q+ |3 /κ3 can be brought outside the delta
3
obtaining |qκ+ |3 δ(p + q qκ+ ). Therefore we have
κ3 −i(ωp ⊕ωq )t
q
−i(ωp ⊕ωq )t +
κ
e δ p+q = e δ p+ q (2.190)
κ |q+ |3 q+
The Dirac deltas imply also a modification of the energy ωp . In fact, we have
κ2 2
ωq2 − q2 = m2 = ωp2 − p2 = ωp2 − 2 q (2.191)
q+
and therefore
κ2 2
ωp2 = m2 + 2
2 q = S(ωq ) . (2.192)
q+
In other words, the Dirac delta imposes
ωp = −S(ωq ). (2.193)
Notice that the LHS contains a minus sign because on-shell we have ωp > 0 and
S(ωq ) < 0, so taking the square root we need to take this into account. The other
exponential is treated in the same way, since one can change variables p 7→ S(p),
obtaining the same result, so that the time-dependent parts go away.
For the second one, we proceed in a similar manner, and we have

Z
p4 n ∧ † † ∧
o
δθ2 = δϕ ⋆ ∂0 δϕ − δ∂0 ϕ ⋆ δϕ† (2.194)
R3 κ
d3 p d3 q
Z Z
p4
=i p p ζ(q) ζ(p) ×
R3 2ωp 2ωq κ
n
(ωq + ωp )δap ∧ δa†q e−i(ωp ⊕S(ωq ))t ei(p⊕S(q))x (2.195)
+ (S(ωp ) + S(ωq ))δb†p δδbq e−i(S(ωp )⊕ωq )t ei(S(p)⊕q)x (2.196)
+ (S(ωp ) + ωq )δb†p δδa†q e−i(S(ωp )⊕S(ωq ))t ei(S(p)⊕S(q))x (2.197)
o
+ (ωp + S(ωq ))δbp δδaq e−i(ωp ⊕ωq )t ei(p⊕q)x (2.198)
so that, proceeding in the same way as before, we have
d3 p 2 p4 p3+
Z
δθ2 = ζ (p) 2ωp δap ∧ δa†p + 2S(ωp )δb†p ∧ δbp (2.199)
2ωp κ κ3
Because of the definition of the action, the total symplectic form will be
δθ1 + δθ2
δθT = (2.200)
2
and therefore we have
d3 p 2 |p+ |
3
Z n
p4 † †
o
δθT = i ζ(p) ωp − S(ω p ) ap ∧ a p − bp ∧ bp . (2.201)
2ωp κ3 κ
This result can be simplified using the definition
ζ(p)2 |p+ |3

p4
α(p) = ωp − S(ωp ) (2.202)
2ωp κ3 κ
and we get
Z n o
δθT = i d3 p α(p) ap ∧ a†p − b†p ∧ bp . (2.203)
This result would imply the Poisson brackets

n o 1 κ 2ωp
ap , a†q = −i |3
δ(p − q) (2.204)
ξ(p)2 p4 |p+
ωp − S(ωp )
κ3
n o 1 κ 2ωp
bp , b†q = −i |p+ |3
δ(p − q). (2.205)
ξ(p)2 p4 ωp − S(ωp )
κ3
which can also be simplified using eq. (2.202) as follows

n o i
ap , a†q = − δ(p − q) (2.206)
α(p)
n o i
bp , b†q = − δ(p − q). (2.207)
α(p)
Notice that, if one uses the canonical quantization map [A, B] = iℏ{A, B} and assume
ℏ = 1, then we would get the commutators
h i 1
ap , a†q = δ(p − q) (2.208)
α(p)
h i 1
bp , b†q = δ(p − q). (2.209)
α(p)
where the LHS is positive and the coefficient goes to 1 for κ → ∞, as it should. We
will comment on how to technically get these commutations relations in section 2.14.4
below.
2.14 Conserved charges: geometric approach

The starting point is the symplectic form in eq. (2.203). For simplicity, we will call
Ω = δθT = δθ1 +δθ
2
2
. In order to get the charges from the symplectic form we need to
do the following.
Assuming that the charges come from a symmetry described by some continuous
vector field ξ in spacetime, then we compute (87) (this approach is also used quite
often in GR, for example see (89), (90))
!
−δξ ⌟ Ω = δQξ (2.210)
where δ is the exterior derivative in phase space (i.e. the set of solutions of the EoM)
and Qξ is the charge associated to the vector ξ. The object δξ is a vector field in
phase space generated by the killing vector field ξ. In other words, δξ A measures the
infinitesimal variation of the object A in phase space due to the symmetry due to the
action along ξ in spacetime. The symbol ⌟ is used to indicete contraction of the vector
field with forms. Notice that each charge obtained in this way is a symmetry of the
symplectic form. Indeed one can easily show that
£δξ Ω = δξ ⌟ δΩ + δ(δξ ⌟ Ω) = 0 + δ 2 Qξ = 0 (2.211)
where we used the fact that δ 2 = 0 since δ is the exterior derivative in phase space,
and also δΩ = 0 because Ω is an exact 2-form. The charges built in this way are
independent of the dynamics because the symplectic form Ω and the vector field δξ
are both time independent and non-dynamical by construction, and they are defined
on-shell.
To see how the procedure of getting the charges using this formalism, we first
tackle the problem in the canonical non-deformed QFT context.
2.14.1 Non-deformed case

In this case, the symplectic form is given by the κ → ∞ limit of eq. (2.203) above,
and it reads
Z
Ωκ→∞ = i d3 p (δap ∧ δa†p − δb†p ∧ δbp ). (2.212)
2.14.1.1 Translation charges

In this case, the translation can be described by the vectors ∂0 or ∂i , depending on
whether we are talking about time translation or spatial translations.
2.14. Conserved charges: geometric approach 55
Starting with the time translations, we first need to understand what δ∂0 ap is (it
will be analogous for b). We know that after a translation we have
ap 7→ eiϵωp ap = ap + iϵωp ap (2.213)
and therefore
δ∂0 ap = iϵωp ap ⇔ δ∂0 a†p = −iϵωp a†p (2.214)
where the second expression can be obtained from the first one by simply taking
the hermitian conjugate. Therefore, we have (the wedge product goes away after the
contraction because the creation/annihilation operators are not themselves forms)
Z
−δ∂0 ⌟ Ω κ→∞
= −i d3 p (δ∂0 ap δa†p − δap δ∂0 a†p − δ∂0 b†p δbp + δb†p δ∂0 bp ) (2.215)
Z
= −i d3 p [iϵωp ap δa†p − δap (−iϵωp a†p )
− (−iϵωp b†p )δbp + δb†p iϵωp bp ] (2.216)

Z
= ϵ d3 p [ωp ap δa†p + δap ωp a†p + ωp b†p δbp + δb†p ωp bp ] (2.217)
Z
= ϵδ d3 p ωp (ap a†p + b†p bp ) (2.218)
and therefore we get the canonical charge apart from the irrelevant prefactor ϵ.
Z n o
P0κ→∞
= d3 p ωp a†p ap + bp b†p (2.219)
We can also check that eq. (2.213) gives the correct field transformation.
d3 p h µ
Z i
δ ∂µ ϕ = p iϵ pµ ap e−i(ωp t−px) − iϵµ pµ b†p ei(ωp t−px)
2ωp
= ϵµ ∂µ ϕ(x). (2.220)
Of course, once we quantize these charges, the operators will need to be put in normal
order. The same exact reasoning will work for the spatial translation charges, it is
sufficient to substitute ωp with pi everywhere.
2.14.1.2 Boost and rotation charges

In this case, the procedure is again the same, but this time we need to use as a
vector field ξ the generator of boost transformation, and then we need to understand
how creation/annihilation operator change under this transformation, i.e. we need to
understand what is the object δΛ ap .
In the case of boosts the field Fourier components transform as follows
∂ap pi
δ B ap = iωp λi + i ap λi ,
∂pi 2ωp
†
i ∂ap pi
δ B a†p = iωp λ + i a†p λi (2.221)
∂pi 2ωp
and analogous expressions for bp , b†p . We have therefore

Z
B
−δ ⌟ Ω κ→∞
= i d3 p (δ B a†p δap − δa†p δ B ap − δ B bp δb†p + δbp δ B b†p )
!
Z † †
∂a p ∂a p ∂b p ∂bp
= λi d3 p ωp δa† − δap i − δbp + δb†p i
∂pi p ∂p ∂pi ∂p
Z
pi
+ λi d3 p ap δa†p − δap a†p + δb†p bp − b†p δbp
2ωp
!
Z † †
1 i ∂a p ∂a p ∂b p ∂bp
= λ δ d3 p ωp a† − ap i − bp + b†p i (2.222)
2 ∂pi p ∂p ∂pi ∂p
which gives the non-deformed boost charge

!
∂a†p ∂b†p
Z
1 ∂ap † ∂bp
Ni κ→∞
= 3
d p ωp i
a p − a p i
− i
bp + b†p i . (2.223)
2 ∂p ∂p ∂p ∂p
The case of rotations R is analogous, and starting from the transformations

∂ ∂ †
δ R ap = iϵijk ρi pj ap δ R a†p = iϵijk ρi pj a (2.224)
∂pk ∂pk p
and analogously for bp , b†p , one gets the charge

!
∂a†p ∂b†p
Z
1 ∂ap † ∂bp
Mi = 3
d p ϵijk p j
ap − ap − bp + b†p . (2.225)
8 ∂pk ∂pk ∂pk ∂pk
Once again, we can obtain the field transformations from eq. (2.221) and (2.224), and
one respectively gets
" #
d3 p ∂b†p
Z
B ∂ap pi −i(ωp t−px) † pi
δ ϕ(x) = iλi p ωp i
+ ap e + ωp i
+ bp ei(ωp t−px)
2ωp ∂p 2ωp ∂p 2ωp
3
Z
d p ∂ −i(ωp t−px) ∂ i(ωp t−px)
= −iλi p a p ωp e + b†p ωp e
2ωp ∂pi ∂pi
∂
= iλi xi ϕ(x) (2.226)
∂t
and
d3 p
Z
R i j ∂ −i(ωp t−px) † ∂ i(ωp t−px)
δ ϕ(x) = iρ ϵik p ap pk e + bp pk e
2ωp ∂pj ∂pj
∂
= iρi ϵik j xj k ϕ(x) (2.227)
∂x
which are the standard spacetime boost and rotation transformations.
Notice that for the boosts we actually have antisymmetrization when writing them
in spacetime. In fact, we only wrote down the contribution from ωp ∂p ∂
i
, but we should
also include the contribution of pi ∂ωp . However, since we are only integrating in d3 p,
∂
we first need to transform the derivative in ωp into a derivative in pi (which will

allow us to integrate by parts). Moreover, since we are on-shell (which means that
p2 + m2 ), one can show by chain rule that

p
ωp =
∂ ∂ωp ∂ pi ∂ pi ∂
= = =− (2.228)
∂pi ∂pi ∂ωp ωp ∂ωp ωp ∂ωp
and we get the same contribution as before.

Given the importance of the boost for the following discussion, as a double check
of the correctness of eq. (2.223), we also sketch the computation of the boost charge
from scratch using Noether theorem in non-deformed QFT. The action is
L = (∂µ ϕ)† (∂ µ ϕ) − m2 ϕ† ϕ (2.229)
and the on-shell fields for such an action are given by
d3 p
Z
ϕ= p ap e−i(ωp t−px) + b†p ei(ωp t−px) (2.230)
2ωp
d3 p † i(ωp t−px)
Z
†
ϕ = p ap e + bp e−i(ωp t−px) . (2.231)
2ωp
The energy-momentum tensor is given by

∂L ∂L
Tµν = ∂µ ϕ + ∂µ ϕ† − δµν L (2.232)
∂(∂ν ϕ) ∂(∂ν ϕ† )
1 1 1 1
= (∂µ ϕ)† (∂ ν ϕ) + (∂µ ϕ)(∂ ν ϕ)† − (∂α ϕ)† (∂ α ϕ) + m2 ϕ† ϕ. (2.233)
2 2 2 2
By Noether theorem, the charge is given by
Z
L0i = i (x0 Ti0 − xi T00 ) (2.234)
R3
(the additional i factor is added by hand, and it does not modify the discussion) and
using eq. (2.232) and eq. (2.230), (2.231) one can do the explicit computations, which
we will not do explicitly here. Notice however the following subtlety. In the term xi T00 ,
the xi factor can be described as a derivative in momentum space acting on the spatial
exponent, but since this exponent will contain p − q (the two momenta are necessary
because Tµν contains product of fields, each of which is an integral in momentum
space, so we need two integration variables) we have two ways of going so. This will
result in two possible ways to compute the coefficient xi T00 , namely
(
Z Z
d3p d 3q
− xi T00 = − p p (ωp ωq + pq + m2 )×
R3 2ωp 2ωq
)
∂
† −i(ωp −ωq )t † i(ωp −ωq )t
× −iap aq e + ibp bq e δ(p − q) (2.235)
∂pi
(
d3 p d3 q
Z Z
− xi T00 = − p p (ωp ωq + pq + m2 )×
R3 2ωp 2ωq
)
∂
× iap a†q e−i(ωp −ωq )t − ib†p bq ei(ωp −ωq )t δ(p − q) (2.236)
∂qi
After integration by parts, in both cases one term will eliminate the coefficient x0 Ti0 ,
and the final charges will be
( !)
∂b†p
Z
(p) ∂ap †
3
Ni = − d p ωp − i ap + bp . (2.237)
∂p ∂pi
( !)
∂a†p
Z
(q) ∂bp
Ni = d3 p ωp −ap i + b†p i . (2.238)
∂p ∂p
However, this means that we can write without loss of generality that the boost charge
Ni is given by
1 (p) (q)
Ni = (Ni + Ni ) (2.239)
2
which is exactly eq. (2.223).
2.14.1.2.1 Crucial step Notice that in the last passage of Eq. (2.222) we used
the fact that (we consider only the a, a† case since the same reasoning applies to b, b† )
!
∂ap † ∂a†p 1 ∂ap † ∂a†p 1 ∂ † †

δa − δap = δ a − ap + ap δa − δa p a p . (2.240)
∂pi p ∂pi 2 ∂pi p ∂pi 2 ∂pi p
The second term on the RHS of the last equation is still not a total derivative be-
cause of the presence of ωp inside the integrand. After integration by parts, this term
contributes a factor
pi
− ap δa†p − δap a†p (2.241)
2ωp
to the integrand, which cancels the second term in the second passage of (2.222).
2.14.2 κ-deformed case

Here we work with the full symplectic form in eq. (2.203). In obtaining this equation
we used the following property of the wedge product of two 1-forms e and f .
e ∧ f = −f ∧ e. (2.242)
In particular, this property was used when summing the contribution to the symplec-
tic form coming from the two different orderings of the action. This is a non-trivial
assumption in the presence of deformation, and we will discuss it more in section
2.14.3.
2.14.2.1 Translation charges

The idea is basically the same, with the only exception that now we need to consider
the presence of antipodes. Because of this, the first thing to do is to understand how
the creation/annihilation operator transform under translation.
Judging by the definition of the fields in eq. (2.115) and (2.116), we assume the
following behaviour
ap 7→ eiϵωp ap a†p 7→ eiϵS(ωp ) a†p b†p 7→ eiϵS(ωp ) b†p bp 7→ eiϵωp bp (2.243)

and the same for spatial momenta, or in other words
δ∂0 ap = iϵωp ap δ∂0 a†p = iϵS(ωp )a†p

δ∂0 b†p = iϵS(ωp )b†p δ∂0 bp = iϵωp bp (2.244)
These transformations are correspond to the field transformation
d3 p
Z h i
δ∂µ ϕ = p ζ(p) iϵµ pµ ap e−i(ωp t−px) + iϵµ S(p)µ b†p e−i(S(ωp )t−S(p)x)
2ωp
= −ϵµ ∂µ ϕ(x) (2.245)
for the index µ = 0. However, for these transformations to be well defined, then in
order to ensure the invariance of the products aa† and bb† we need to interpret them
as a star products. This star product is defined in the phase space and not in canonical
spacetime, otherwise the steps leading to the computation of the charges are faulted
by the fact that we cannot write ae(·) ⋆ be[·] = abe(·) ⋆ e[·] . However, things are easily
solved by defining
ae(·) ⋆ be[·] = (a ⋆P S b) e(·) ⋆ e[·] . (2.246)
In fact, we have
def.

ap a†p := ap ⋆P S a†p 7→ (ap ⋆P S a†p ) · eiϵωp ⋆ eiϵS(ωp ) = ap ⋆P S a†p (2.247)
This definition of aa† as a⋆P S a† does not influence the creation/annihilation operators
algebra or the definition of normal ordering, so everything remains well defined.
However, there is now the issue of how to treat the contraction of a vector field δξ
with the form δap ∧ δa†p . In fact, the objects inside the wedge product are now to be
treated in the deformed case.
2.14.2.1.1 Non-deformed contraction: While we are in the context of forms,

the product between the forms δa and δa† is treated using the canonical wedge product
with its properties. However, when we contract with δξ , one has canonically that
δξ ⌟ (δap ∧ δa†p ) = δξ ap δa†p − δap δξ a†p (2.248)
and the wedge product is not present. Notice that the change of sign in the second
term can be understood in terms of the following steps.
• The vector field always acts on the first objects it encounters in the wedge
product. Because of this, to compute its action on the second term of the wedge
product, we first need to use the formula
v ∧ w = (−)p·q w ∧ v (2.249)
where v and w are respectively a p-form and a q-form. In our case, we have the
wedge product of two 1-forms, and therefore we have
δap ∧ δa†p = −δa†p ∧ δap . (2.250)
• Now the vector field acts on the first component, giving
−(δξ a†p )δap . (2.251)

• Now the object δξ a†p is just some function, i.e. a 0-form, and therefore it com-
mutes with δa, so its placement is arbitrary.
2.14.2.1.2 Deformed contraction: In this case, we define:3
δξ ⌟ (δap ∧ δa†p ) = (δξ ap )δa†p + δap [S(δξ )a†p ]. (2.252)
The main difference between this axiom and the one in eq. (2.248) is the presence of
the antipode instead of the normal minus sign in the second term. Notice that, since
κ→∞
S(AB) = S(A)S(B), and since S(A)S(B) −−−→ (−A)(−B) = AB, the object S(δξ )
in eq. (2.252) must contain an additional minus sign when necessary, for consistency
with the non-deformed axiom. For example, if δξ = ϵpi ∂p ∂
i
where ϵ is just some
constant, then S(δξ ) = −ϵS(p)i ∂S(p)i , which has the correct κ → ∞ limit.
∂
Coming back to the charges, we first notice that we can rewrite

Z
Ω = i d3 p α (δap ∧ δa†p − δb†p ∧ δbp ) (2.253)
Z
= −i d3 p α (δa†p ∧ δap − δbp ∧ δb†p ). (2.254)
Then, contracting with the vector field δ∂µ , using the assumption in eq. (2.252), and
using eq. (2.244) we finally get
Z
−δ∂0 ⌟ Ω = i d3 p α (δ∂0 a†p δap + δa†p S(δ∂0 )ap − δ∂0 bp δb†p − δbp S(δ∂0 )b†p ) (2.255)
Z
= i d3 p α [iϵS(ωp )a†p δap + δa†p (iϵS(ωp )ap )
− (iϵωp b†p )δbp − δb†p iϵωp bp ] (2.256)

Z
= −ϵ d3 p α [S(ωp )a†p δap + δa†p S(ωp )ap − ωp b†p δbp − δb†p ωp bp ] (2.257)
Z
= ϵδ d3 p α [−S(ωp )ap a†p + ωp b†p bp ] (2.258)
which, apart from the irrelevant ϵ factor, gives the correct time translation charge (of
course, in order to compare the final expression for the charges, normal ordering is
always assumed. As stated before, the normal ordering is not affected by whether we
consider the product of creation/annihilation operators as aa† or a ⋆ a† .). Notice also
that we have used the fact that δ(AB) = (δA)B +AδB. Of course, the same reasoning
gives the correct space translation charges, since the procedure works in exactly the
same way. We get therefore the following translation charges
Z n o
P0 = d3 p α(p) −S(ωp )a†p ap + ωp bp b†p (2.259)
Z n o
Pi = d3 p α(p) −S(p)i a†p ap + pi bp b†p (2.260)
3
We write them down for a, but the same goes for b
Notice that by construction we have the formal limit (ζ goes to 1 for κ → ∞ by

definition)
lim α(p) = 1. (2.261)

κ→∞
Furthermore, since ωp > 0 and −S(ωp ) > 0, the energy has the correct sign, and in
the limit κ → ∞ we get back the correct translation charges of non-deformed complex
scalar field. As a final comment, in (78) the same charges were calculated using a direct
approach based on the Noether theorem. However, it became clear while writing this
thesis that the direct approach presented in (78) only reproduces some of the terms
of the full charges computed here (the charges in (78) are however still correct, in
light of the geometric approach presented here). The source of this incongruence is an
interesting topic, deserving of further investigation, and it will be tackled in future
publications.
2.14.2.2 Notes on the above (overall) procedure

There are several things to notice about the above procedure.
1) The most obvious thing to address is that there seems to be no reason as to why
we chose to write
Z
Ω = −i d3 p α (δa†p ∧ δap − δbp ∧ δb†p ) (2.262)
and not for example

Z
Ω = −i d3 p α (δa†p ∧ δap + δb†p ∧ δbp ) (2.263)
which seems at least as good. The only sensible answer to this is to look at the
single symplectic forms δθ1 and δθ2 , we have
δθ1 ∝ δϕ† ∧ δϕ δθ2 ∝ δϕ ∧ δϕ† (2.264)
In both cases, the off diagonal terms go away and we are only left with the
wedge products between a, a† and between b, b† . However, both in δθ1 and δθ2 ,
since ϕ contains a and b† (and of course ϕ† contains a† and b), each time that
we have δa ∧ δa† we also have δb† ∧ δb. Therefore, the correct final symplectic
charge must be of the kind in eq. (2.262) and not the one in eq. (2.263).
Of course, in non-deformed QFT, either (2.262) or (2.263) are equally fine, but in
that case there is no antipode in the formula for the contraction of the symplectic
form and an external vector field δξ , while there is one in the deformed case in
(2.252).
Therefore, because of the presence of the antipode in (2.252), it seems that the
structure of the symplectic form in the deformed case is more rigid than in the
non-deformed case, because although we are free to chose the overall sign, we
cannot arbitrarily change the order of only a part of the symplectic form before
contracting with δξ .
2) Having tackled the first important point, then there is the issue of why we chose
the overall sign as in eq. (2.254) and not the one in (2.253). The answer is that
one can go from (2.253) to (2.254) by means of the following exchange
a↔b a† ↔ b† (2.265)
so that, performing the same exact computations, instead of the charge

Z
Q1 = d3 p α [−S(ωp )ap a†p + ωp b†p bp ] (2.266)
we would end up with the charge

Z
Q2 = d3 p α [−S(ωp )b†p bp + ωp ap a†p ]. (2.267)
However, for all practical purposes, the charge Q2 is the same as Q1 because
both a, a† and b, b† live on the same manifold. The important thing is that one
of them has energy ωp and the other −S(ωp ) while living on the same manifold,
the name of the operators are not important in this context.
3) One last thing to notice is that the assumption (2.252) is not in contradiction
with the properties of the differential operator dF = ϵA ∂A + ωµν Lµν that we
used in the computation of the charges (at leat at first sight). In fact, while dF
is the exterior derivative in spacetime, we are dealing with (the properties of)
the exterior derivative δ in phase space, which is however the space of solutions
to the EoM. Therefore, the behaviour of δ can be different than the one of dF .
2.14.2.3 Boosts
Analogously to what has been done before, we can now concentrate on the boost
charges. Once again, we need to understand what is the action of δΛ on a, a† , b, b† .
One approach would be to start from the assumed Lorentz transformations
∂
δ B ϕ(x) = iλi xi ϕ(x)
∂t
d3 p
Z
∂ −i(ωp t−px)
= −iλi p ζ(p) ap ωp e
2ωp ∂pi
d3 p
Z
∂
− iλi p ζ(p) b†p S(ωp ) e−i(S(ωp )t−S(p)x) . (2.268)
2ωp ∂S(p)i
and integrating by parts obtaining

" p !#
d3 p 2ωp ∂
Z
∂ap ωp
B
δ ϕ(x) = iλi p ωp ζ(p) + ap p ζ(p) e−i(ωp t−px)
2ωp ∂pi ζ(p)ωp ∂pi 2ωp
" !#
∂b†p
p
d3 p 2ωp
Z
† ∂ S(ωp )
+ iλi p S(ωp ) ζ(p) + bp p ζ(p) e−i(S(ωp )t−S(p)x) .
2ωp ∂S(p)i ζ(p)S(ωp ) ∂S(p)i 2ωp
( p !)
d3 p 2ωp ∂
Z
∂ap ωp
B
δ ϕ(x) = iλi p ζ(p) ωp + ap p ζ(p) e−i(ωp t−px)
2ωp ∂pi ζ(p) ∂pi 2ωp
( !)
∂b†p
p
d3 p 2ωp ∂
Z
† S(ωp )
+ iλi p ζ(p) S(ωp ) + bp p ζ(p) e−i(S(ωp )t−S(p)x) .
2ωp ∂S(p)i ζ(p) ∂S(p)i 2ωp
Therefore, limiting ourselves to ϕ since ϕ† is analogous, we would have the following

relations

∂ 1 ∂ √
B
δ ap = −iλ ωpi
+ √ (ζ(p) ωp ) ap (2.269)
∂pi ζ(p) ωp ∂pi
√
ωp

B † ∂ ∂ S(ωp )
i
δ ap = −iλ S(ωp ) + √ ζ(p) a†p (2.270)
∂S(p)i ζ(p)S(ωp ) ∂S(p)i ωp

∂ 1 ∂ √
B
δ bp = −iλ ωpi
+ √ (ζ(p) ωp ) bp (2.271)
∂pi ζ(p) ωp ∂pi
√
ωp

∂ ∂ S(ωp )
δ B b†p = −iλi S(ωp ) + √ ζ(p) b†p (2.272)
∂S(p)i ζ(p)S(ωp ) ∂S(p)i ωp
We will however now see that the above transformations need to be modified into

∂ 1 1 ∂[ωp S(α(p))]
B i
δ ap = −iλ ωp + ap (2.273)
∂pi 2 ωp ∂pi

B † ∂ 1 1 ∂[S(ωp )α(p)] †
i
δ ap = −iλ S(ωp ) + ap (2.274)
∂S(p)i 2 S(ωp ) ∂S(p)i

∂ 1 1 ∂[ωp α(p)]
δ B bp = −iλi ωp + bp (2.275)
∂pi 2 ωp ∂pi

B † ∂ 1 1 ∂[S(ωp )S(α(p))] †
i
δ bp = −iλ S(ωp ) + bp (2.276)
∂S(p)i 2 S(ωp ) ∂S(p)i
in order to obtain a conserved charge. The reasoning here is as follows. We know

that the action is invariant under κ-Poincaré transformations, which means that by
Noether theorem there are conserved charges. More in detail, there is a one-to-one
correspondence between symmetries and conserved charges. One usually goes in one
direction, obtaining the conserved charges as a result of exploiting the symmetries of
the action. Here we go in the other direction, obtaining the charges, and then defining
the symmetry of the system by showing that charges are conserved.
Notice that in this whole process, there is no way to know a priori the transforma-
tion laws of the fields under a symmetry of the action. A symmetry of the action does
not necessarily translate to a symmetry of the field inside the action; quite the oppo-
site, the fields change under some transformation, and the action is such combination
of fields and their derivatives which is invariant. Eq. (2.268) represents a canonical
field transformation under non-deformed boost; it is therefore not surprising that it
does not give rise to a charge since it is not a symmetry of our system.
We now first show that eq. (2.273), (2.274), (2.275), (2.276) give rise to a conserved
charge, and then we will show to what field transformation they correspond to. We
have
Z
−δ B ⌟ Ω = −i d3 p α(p) δ B a†p δap + δa†p S(δ B )ap − δ B bp δb†p − δbp S(δ B )b†p
!
∂a†p ∂b†p
Z
i 3 † ∂ap ∂bp †
= −λ d p α(p) S(ωp ) δap − δap S(ωp ) − ωp i δbp + δbp ωp i
∂S(p)i ∂S(p)i ∂p ∂p
Z
1 ∂[S(ωp )α(p)] † †
1 ∂[ω α(p)]
p † †

+ λi d3 p a p δap − δap ap − bp δbp − δbp bp
2 ∂S(p)i 2 ∂pi
(2.277)
( " # " #)
∂a†p ∂b†p ∂bp †
Z
1 † ∂ap
= − λi δ 3
d p α(p) S(ωp ) ap − ap + ωp bp i − b
2 ∂S(p)i ∂S(p)i ∂p ∂pi p
(2.278)
which gives the boost charge

( " # " #)
∂a†p ∂b†p ∂bp †
Z
1 3 † ∂ap
Ni = − d p α(p) S(ωp ) ap − ap + ωp bp i − b .
2 ∂S(p)i ∂S(p)i ∂p ∂pi p
(2.279)
2.14.2.3.1 Crucial step Once again, in order to be able to take the exterior
derivative outside, we used the fact that
!
† †
∂ap ∂a p ∂ap ∂ap
δa† − δap =δ a† − ap
∂S(p)i p ∂S(p)i ∂S(p)i p ∂S(p)i
∂δap † ∂δa†p
− a + ap (2.280)
∂S(p)i p ∂S(p)i
and
∂ap † ∂a†p ∂
† †

δa − δap = ap δa − δap a
∂S(p)i p ∂S(p)i ∂S(p)i p p
∂δap † ∂δa†p
+ a − a p (2.281)
∂S(p)i p ∂S(p)i
and therefore, adding (2.280) and (2.281) we finally get

!
∂ap † ∂a†p 1 ∂ap † ∂a†p
δap − δap = δ a − ap
∂S(p)i ∂S(p) i 2 ∂S(p)i p ∂S(p)i
1 ∂
† †

+ a δa
p p − δap p .
a (2.282)
2 ∂S(p)i
Of course, the same exact reasoning works for the b, b† operators. Here, notice that
after integration by parts the second term produces the following contribution to
−δ∂Λ ⌟ Ω
1 ∂[S(ωp )α(p)] † †
1 ∂[ω α(p)]
p † †

− ap δap − δap a p + bp δbp − δbp bp , (2.283)
2 ∂S(p)i 2 ∂pi
and once again it eliminates the relevant term in the computation of the boost charge.
We can now turn to the transformation of the field due to eq. (2.273), (2.274),
(2.275), (2.276). We have
∂
δ B ϕ(x) = iλi xi ϕ(x)
∂t (
3 √
ωp ∂
Z
d p 1 1 ∂[ωp S(α)] ωp
− iλi ζ ωp − √ ζ ap e−i(ωp t−px)
∂pi ζωp ∂pi
p
2ωp 2 ωp ωp
√ )
ωp

1 1 ∂[S(ωp )S(α)] ∂ S(ωp ) † −i(S(ωp )t−S(p)x)
− S(ωp ) − √ ζ bp e
2 S(ωp ) ∂S(p)i ζS(ωp ) ∂S(p)i ωp
(2.284)
∂
δ B ϕ† (x) = −iλi xi ϕ† (x)
∂t
3
( √
ωp ∂
Z
d p 1 1 ∂[ωp α] ωp
− iλi ζ ωp − √ ζ bp e−i(ωp t−px)
2 ωp ∂pi ζωp ∂pi
p
2ωp ωp
√ )
ωp

1 1 ∂[S(ωp )α] ∂ S(ωp )
− S(ωp ) − √ ζ a†p e−i(S(ωp )t−S(p)x)
2 S(ωp ) ∂S(p)i ζS(ωp ) ∂S(p)i ωp
(2.285)
To understand these additional terms, let us use eq. (2.202), and let us assume that
we want to impose
ξ(p) : α(p) = 1. (2.286)
This is a natural choice, since it eliminates this factor from the charges. In this case,
we have
1 p4 |p+ |3

−2
ζ(p) = ωp − S(ωp ) (2.287)
2ωp κ κ3
and one can show that, to the leading order in κ1 , the boost transformation becomes
(
3p pi m2
Z
∂ d
B
δ ϕ(x) = iλi x i
ϕ(x) + iλi − 2 ap e−i(ωp t−px)
κ ωp2
p
∂t 2ωp
)
m2

pi 5
+ − b†p e−i(S(ωp )t−S(p)x) (2.288)
κ 2 2ωp2
(
3p pi m2
Z
B † ∂ † d
δ ϕ (x) = −iλi xi
ϕ(x) + iλ i
− 2 bp e−i(ωp t−px)
κ ωp2
p
∂t 2ωp
)
m2

pi 5
+ − a†p e−i(S(ωp )t−S(p)x) (2.289)
κ 2 2ωp2
We see that particles and antiparticles, upon boost, get an additional translation.
Furthermore, notice also that using the convention in eq. (2.287), this additional
translation is the same for both a† and b† . This is already apparent looking at eq.
(2.274), (2.276), since if α = 1 the quantities α and S(α) disappear from the formulae4 .
If we used instead the convention
|p+ |3 κ

2
ζ(p) : ζ (p) 1 + =1 (2.290)
κ p4
then one gets different translations for a† and b† , as is again apparent from eq. (2.274),
(2.276). In this case, in fact, we would have S(α) ̸= α.
4
The antipode does not act on objects independent of momenta, so that S(1) = 1.
Repeating the same calculations as before, in this case one gets the following boost
transformation for the fields.
(
3p
Z
∂ d pi 19
B
δ ϕ(x) = iλi x i
ϕ(x) + iλi p − √ ap e−i(ωp t−px)
∂t 2ωp κ 16 2
)
4m2

pi 3
+ √ 9+ 2 b†p e−i(S(ωp )t−S(p)x) (2.291)
κ 16 2 ωp
(
3p
Z
B † ∂ † d pi 11
δ ϕ (x) = −iλi x i
ϕ(x) + iλ i
p − √ bp e−i(ωp t−px)
∂t 2ωp κ 16 2
! )
pi 19 3m2
+ √ + √ a†p e−i(S(ωp )t−S(p)x) . (2.292)
κ 16 2 4 2ωp2
In this case, the additional translations for a† and b† are different.
2.14.2.3.2 The (very) peculiar features of boosts We already saw in eq.

(1.133) that a single particle state gets boosted canonically. This is not just an as-
sumption, since it can be proven by using the algebra of the charges. Anticipating a
little the results of section 2.15, we will show that all the charges that we compute
in this section satisfy the canonical, non-deformed Poincaré algebra. This means in
particular that [Ni , Pj ] = −iηij P0 . We now define single-particle states as usual,
namely given the vacuum state |0⟩ we have
a†p |0⟩ = |p⟩a b†p |0⟩ = |p⟩b . (2.293)
Since we already computed the translation charges (see eq. (2.259), (2.260)), we can
use them to obtain what is the eigenvalue of the above momentum eigenstates. For
example, recalling eq. (2.208), (2.209), we have
Z
P0 |q⟩a = d3 p α[−S(ωp )a†p ap ]a†q |0⟩
Z
1
= d3 p α[−S(ωp ) δ(p − q)]a†p |0⟩
α
= −S(ωp )|p⟩a (2.294)
and in the same way one obtains
Pi |p⟩a = −S(p)i |p⟩a Pi |p⟩b = pi |p⟩b (2.295)

P0 |p⟩a = −S(ωp )|p⟩a P0 |p⟩b = ωp |p⟩b . (2.296)
We can now use the commutation relation [Ni , Pj ] = −iηij P0 to show that
−iλj Nj Pi |p⟩a = iS(p)i λj Nj |p⟩a = −iPi λj Nj |p⟩a + λi S(ωp )|p⟩a (2.297)

j j j i
−iλ Nj P0 |p⟩a = iS(ωp )λ Nj |p⟩a = −iP0 λ Nj |p⟩a + λ S(p)i |p⟩a (2.298)
which immediately imply
−iλj Nj |p⟩a = |p + λωp ⟩a (2.299)

which is the canonical action of an infinitesimal boosts. One can then obtain the finite
boost, which is therefore the canonical one.
If we now define a wave packet using the fields in eq. (2.115), (2.116) (we can
concentrate on particles created by a† , since the same considerations will be valid
about those created by b† ), we have
d3 p
Z
†
ϕ |0⟩ = p ζ(p) a†p e−i(S(ωp )−S(p)x) |0⟩
2ωp
d3 p
Z
= p ζ(p)e−i(S(ωp )−S(p)x) |p⟩
2ωp
= |φ(x)⟩ (2.300)
and acting now with a boost, since single particle states behave canonically, we will
get a canonically boosted single particle wave packet. The φ in this equation describes
the packet obtained by an appropriate definition of ζ(p). On the other hand, if we
first boost the field (obtaining eq. (2.289) or (2.292), or any other expression based
on different choices of ζ(p)), and only then apply the boosted field to the vacuum
state, we get a wave packet with a different distribution due to the presence of the
additional factors in eq. (2.289), (2.292).
This is a puzzling property of boosts which needs to be more deeply understood.
One of the possible solutions would be to understand whether there are differences
between active and passive boosts. Indeed, one can consider the boost acting on the
state |φ(x)⟩ as a passive boost (the wave packet has already been created, and only
later a boost is applied, which changes the coordinates describing the distribution).
On the other hand, eq. (2.289), (2.292) may be considered as active boosts (the field
is boosted before any particle is created). In a passive boost, the particle does not
need to exchange energy with the environment, since we (the observers) are moving
at a different speed, describing what we see from a different point of view. On the
other hand, to physically create an active boost on a field one needs to interact with
it in some way, for example an electron can be actively boosted by switching on an
electric field in order to accelerate it, and then switching it off, obtaining an electron
in uniform rectilinear motion. In canonical, non-deformed QFT, the net result is the
same and indeed active and passive transformations are completely equivalent. It is
however possible that in the deformed context there may be different contributions in
the two cases, leading to the difference just described. However in this context we are
only considering free particles, so that such considerations would be premature. We
will leave the investigation of this puzzle to forthcoming publications.
2.14.2.3.3 The role of ζ(p) Notice that in order to get a better idea for the boost
of a field in the deformed context we needed to assume a specific value for ζ(p). In the
above discussions, we used two natural definitions, namely eq. (2.287) and (2.290). At
first sight, therefore, it would seem that the physical results of our model depend on
our choice of the arbitrary factor ζ(p), which is of course absurd.
This is however not the case. Notice, in fact, that the factor ζ(p) can be found
inside α in eq. (2.202), which is in turn present in the charges (up to now we only have
the translation charges in eq. (2.259), (2.260), and the boost charge in eq. (2.279),
but we will see that it also appears in the rotation charge), symplectic form (see eq.
(2.203)), and creation/annihilation operators algebra (eq. (2.208), (2.209)). However,
it also appears in the off-shell action in momentum space in eq. (2.156). This means
that if we sue the convention for ζ(p) in eq. (2.287), then the generators of the algebra
are much simpler, but of course the off-shell action now has a global prefactor in the
integrand given by
−1
|p+ |3 1 |p+ |3

p4
1+ 3 3
ωp − S(ωp ) . (2.301)
κ 2ωp κ κ
On the other hand, with convention (2.290), the global prefactor in the momentum-
space off-shell action in eq. (2.156) is 1, but the charges now contain a global prefactor
in the integrand
−1
|p+ |3 1 |p+ |3

p4
1+ 3 3
ωp − S(ωp ) . (2.302)
κ 2ωp κ κ
Therefore, there is a do ut des scenario between the algebra generators and the off-shell
momentum-space action, which is important because, as we will see in chapter 3, it is
used in the definition of the propagator by using the path integral in the κ-deformed
context.
In other words, if we are in the situation described by eq. (2.287), then after
a boost particles and antiparticles receive the same additional translation, but their
propagator is modified by some non-trivial prefactor. On the other hand, we can choose
to work with convention given by eq. (2.290), in which case after a boost particles
and antiparticles receive a different additional translation, but the propagator will be
the traditional one. One may very well choose another convention, in which case the
behaviour of particles and antiparticles under boosts and their propagator have mixed
features, in between the two extremal case discussed above.
2.14.2.4 Rotations
For rotations, we can proceed in the same manner as for boost, and we only need to
understand how to write δR a and δR a† (the case of b and b† will be analogous). We
start from
∂ ∂
δ R ap = iϵijk ρi pj ap δ R a†p = iϵijk ρi S(p)j a†
∂pk ∂S(p)k p
∂ ∂
δ R bp = iϵijk ρi pj k bp δ R b†p = iϵijk ρi S(p)j b† (2.303)
∂p ∂S(p)k p
and we have
Z
−δR ⌟ Ω = i d3 p α (δ∂R a†p δap + δa†p S(δ∂R )ap
− δ∂R bp δb†p − δbp S(δ∂R )b†p ) (2.304)

∂a†p
Z ∂ap
= iϵij
d3 p α S(p)[i j]
δap − δa†p S(p)[i
∂S(p) ∂S(p)j]
∂bp ∂b†p
− p[i j] δb†p + δbp p[i j] (2.305)
∂p ∂p
∂a†p
Z ∂ap
= iϵij δ d3 p α S(p)[i ap − a†p S(p)[i
∂S(p)j] ∂S(p)j]
∂bp ∂b†p
− p[i j] b†p + bp p[i j] (2.306)
∂p ∂p
which give the rotation charge

!
∂a†q ∂b†q
Z
1 † ∂aq ∂bq
Mi = −ϵi jk d3 q S(q)j k
aq − aq S(q)j k
+ bq q j k
− qj k b†q .
8 ∂S(q) ∂S(q) ∂q ∂q
(2.307)
The considerations in this case are the same as for the boost charges.
2.14.3 Is it δA ∧ δB = −δB ∧ δA, or δA ∧ δB = S(δB ∧ δA)?

We are at point vii)b) of our roadmap in section 2.1. We have introduced the anti-
pode in the contraction rule in eq. (2.252), but so far we have not discussed whether
this inclusion is consistent with other rules concerning forms, vector fields, and their
interaction.
The first thing to tackle is the one reported in the title of this subsection. Indeed, in
obtaining the symplectic form in eq. (2.203), we used the fact that δA∧δB = −δB∧δA
(here A, B represent the quantities a, a† , b, b† ). Furthermore, we used it in two ways.
The first one was when we were taking the exterior derivative of an object like (δB)A.
Indeed canonically we have
δ[(δB)A] = −δB ∧ δA (2.308)
but we did not address whether this should instead be substituted by
δ[(δB)A] = S(δB ∧ δA). (2.309)
The second time was when we added the two symplectic forms δθ1 and δθ2 in eq.
(2.188) and (2.199) coming from the two action orderings, since one of the two had a
global minus sign to compensate a global ordering change in the wedge products. In
other words, to go from eq. (2.188) and (2.199) to eq. (2.203) we used the fact that
δA ∧ δB = −δB ∧ δA. (2.310)
Notice however that these two relations are not uncorrelated. Indeed eq. (2.308) can
be considered as a consequence of eq. (2.310) because the exterior derivative only acts
on the first object, and then we switch places using eq. (2.310). In other words, we
have
δ(δBA) = δ(AδB) = δA ∧ δB = −δB ∧ δA. (2.311)
In what follows, we will consider what happens when we try to obtain a symplectic
form using the assumption (2.309). We will show that the only way to obtain an
explicitly time-independent symplectic form is by imposing
S(δA ∧ δB) = −δA ∧ δB. (2.312)
In other words, the antipode acting on wedge products (without any contraction)
behaves like a canonical minus sign. We stress that the time independence of the sym-
plectic form is a necessary requirement if one wants time-independent commutators
between creation/annihilation operators, as well as time independent charges.
Indeed, using eq. (2.309), and keeping in mind eq. (2.177) and (2.178), one can
easily see that the time dependent coefficients in both δθ1 and δθ2 are of the type5
[ωq δbq ∧ δap − S(ωp )S(δbq ∧ δap )] ei(q⊕p)x e−i(ωq ⊕ωp )t . (2.313)
We know from eq. (2.193) that the Dirac delta which comes out of the exponential
imposes ωq = −S(ωp ), so that the time-dependent term becomes
−S(ωp ) δbS(p) ∧ δap + S(δbS(p) ∧ δap ) e−i([−S(ωp )]⊕ωp )t . (2.314)

Notice that, without the assumption in eq. (2.309), this term would be identically
zero. If instead we use the assumption in eq. (2.309), then this time-dependent term6
would remain, unless
S(δaq ∧ δbS(q) ) = −δaq ∧ δbS(q) (2.315)
i.e. if the antipode acting on a wedge product is exactly the same as the canonical
minus acting on it, which is what we wanted to show.
2.14.4 And what if we add contractions with vector fields?

In the previous subsections we have seen that
δB ∧ δA = S(δA ∧ δB) = −δA ∧ δB. (2.316)
Eq. (2.316) however, only deals with wedge products in the absence of contraction
with vector fields. In case we need to contract with some vector fields, because of the
antipode in eq. (2.252), one needs to take care of whether to use the antipode or the
canonical minus sign when changing the order of the wedge product involved in a
contraction.
Given any two functions f, g in phase space, their Poisson bracket is defined as
(91)
{f, g} = ω(Xf , Xg ) (2.317)
where Xf and Xg are the vector fields generated by f and g respectively, i.e.
Xf ⌟ ω = −δf Xg ⌟ ω = −δg (2.318)
and where ω is the symplectic 2-form (compare the above equations with eq. (2.210)).
In our case, we will use the symplectic form (2.203). In this case the functions on
momentum space would be the creation/annihilation operators themselves.
For simplicity, we first concentrate on the a, a† part. In this case, recalling the
definition (2.252) it is easy to see that, assuming the symplectic form in eq. (2.203),
we have
1 ∂ 1 ∂
Xaq = i Xa† = i (2.319)
S(α(q)) ∂a†q q α(q) ∂aq
5
Here we discuss about only one of these time-dependent terms, but for all the others the same
considerations hold.
6
Crucially [−S(ωp )] ⊕ ωp ̸= 0, this can be easily realized by noticing that the κ → ∞ of this sum
is ωp + ωp ̸= 0.
In fact we have
Z " #
1 ∂ †
3
Xaq ⌟ Ω = i d p α(p) i ⌟ (δap ∧ δap ) (2.320)
S(α(q)) ∂a†q
Z " ! #
1 ∂
3
= i d p α(p) 0 + δap S i a†p (2.321)
S(α(q)) ∂a†q
∂a†p
Z
1
= − d3 p α(p) δap (2.322)
α(q) ∂a†q
|{z}
=δ(p−q)
= −δaq (2.323)
Z
1 ∂ †
Xa† ⌟ Ω = i 3
d p α(p) i ⌟ (δap ∧ δap ) (2.324)
q α(q) ∂aq
 
Z  1 ∂ap 
=i d3 p α(p) i δa†p + 0 (2.325)
 
 α(q) ∂aq 
|{z}
=δ(p−q)
= −δa†q (2.326)
Notice that here we used the assumption S(a) = a and S(a† ) = a† , so that the
antipode does not act on the single creation/annihilation operator but only on their
wedge product. There are some subtleties that need to be sorted out.
First of all, notice that the vector field Xaq is defined in such a way that it requires
the application of the antipode to ‘work as intended’. For this reason, in order to
get the correct value for the commutator {a, a† } we must first contract Ω with Xaq
and only after this we can contract with Xa† , and not vice versa. In this way, we
q
immediately get
i
{ap , a†q } = Xa† ⌟ (Xap ⌟ Ω) = − δ(p − q) (2.327)
q α
which is in accordance with the canonical result (see for example eq. (2.206)). It is
important to notice that at first sight one has
Xa† ⌟ (Xap ⌟ Ω) ̸= −Xap ⌟ (Xa† ⌟ Ω) (2.328)

q q
since, as can be seen from eq. (2.324), we have
i
Xap ⌟ (Xa† ⌟ Ω) = − δ(p − q). (2.329)
q S(α)
However, as explained above, intuitively speaking one can say that in this case the
vector field Xaq is not ‘working in the way it was built to work’, since it requires the
application of an antipode. More technically, keeping in mind eq. (2.252), notice that
when we are contracting with vector fields, we are not free in general to switch the
order of the wedge products unless we use the antipode. An example of this fact is
given by the computation of the charges using the symplectic form. Take for example
the time translation charge (the computations of the charges are done in section 2.14,
here we just use an example for the argument’s sake). The symplectic form is given by
(2.203), and the creation/annihilation operator transformations are given by (2.244),

and at the end we get the charge in eq. (2.259).
However, if one wants to change the order of the wedge products involved in a
contraction with an external vector field (this point is crucial), then one gets exactly
the same result if one uses eq. (2.316) (with the antipode). Indeed, using the symplectic
form (consider just the a, a† case and ignoring the factor α for simplicity)
Z
Ω = −i d3 p α S(δap ∧ δa†p ) (2.330)
then the charge becomes

Z
−δ∂0 ⌟ Ω = i d3 p α S(δ∂0 )ap δap + δap S(S(δ∂0 ))a†p (2.331)
Z
=i d3 p α S(δ∂0 )ap δap + δap δ∂0 a†p (2.332)
Z
= −ϵ d3 p α S(ωp )ap δap + δap S(ωp )a†p (2.333)
Z
= ϵδ d3 p α [−S(ωp )ap a†p ] (2.334)
which gives exactly the same result. Notice that, in the presence of a contraction with
a vector field, the prescription in eq. (2.316) needs means that the antipode is applied
last, meaning that
X ⌟ (δA ∧ δB) = S(X ⌟ (δB ∧ δA)). (2.335)
In other words, given the contraction
X ⌟ (δA ∧ δB) = (XA)δB + δA(S(X)B) (2.336)
we define
S(X ⌟ (δB ∧ δA)) = (S(X)A)δB + δA(XB) . (2.337)
Notice that, under the assumption that S does not act on A, δA, B, δB individually
(in this section they represent a, b, a† , b† ), eq. (2.335) explains the above computations
regarding charges. Alternatively, one could drop the assumption that S(·) does not
act on A, δA, B, δB individually, and simply define
X ⌟ (δA ∧ δB) = S(X) ⌟ (δB ∧ δA) . (2.338)
In light of these considerations, and also keeping in mind that eq. (2.316) implies
that
δA ∧ δB = −S(δA ∧ δB) (2.339)
and, if we apply the same point of view that antipodes should be applied last as before,
we see that there are several ways to define the contraction of a 2-form with two vector
fields. In particular, since Xaq has been designed to account for the presence of an
antipode, and since as shown in eq. (2.335) the antipode is applied last, notice that
we have
i i
−S(Xap ⌟ (Xa† ⌟ Ω)) = δ(p − q) = δ(p − q) (2.340)
q S(S(α)) α
which gives the correct value for the Poisson bracket defined as
i
{a†q , ap } = −S(Xap ⌟ (Xa† ⌟ Ω)) = δ(p − q). (2.341)
q α
Therefore, once a prescription for the contraction of two vector fields with a 2-form
is given, when switching the order of contractions one needs to impose an additional
global −S to the computations. Notice that we indeed recover the property
{a†q , ap } = −{ap , a†q } (2.342)
Of course, if instead of starting from the symplectic form in eq. (2.203) we started
from the one in eq. (2.253), then the vector fields Xaq , Xa† would now be given by
q
1 ∂ 1 ∂
Xaq = −i Xa† = −i (2.343)
α(q) ∂a†q q S(α(q)) ∂aq
In fact we would have

Z " #
1 ∂
Xaq ⌟ Ω = −i d3 p α(p) −i ⌟ (δa†p ∧ δap ) (2.344)
α(q) ∂a†q
∂a†p
Z
1
=− d3 p α(p) δap (2.345)
α(q) ∂a†q
|{z}
=δ(p−q)
= −δaq (2.346)
Z
1 ∂ †
3
Xa† ⌟ Ω = −i d p α(p) −i ⌟ (δap ∧ δap ) (2.347)
q S(α(q)) ∂aq
Z
1 ∂
= −i d3 p α(p) 0 + δa†p S −i δap (2.348)
S(α(q)) ∂aq
 
Z  1 ∂ap 
= −i d3 p α(p) 0 − i δa†p  (2.349)
 
 α(q) ∂aq 
|{z}
=δ(p−q)
= −δa†q (2.350)
Once again, following the fact that this time Xa† has been defined to accommodate
q
for the antipode, it must be contracted first, so that we end up with
i
{a†p , aq } = Xaq ⌟ (Xa† ⌟ Ω) = δ(p − q) (2.351)
p α
which is in perfect agreement with eq. (2.341). Once again, if we switch the order of
contraction, we need to add a global −S, and indeed one has
i
{ap , a†q } = −S(Xa† ⌟ (Xap ⌟ Ω)) = − δ(p − q) (2.352)
q α
which is in accordance with eq. (2.327). Of course, the same reasoning gives the correct
propagators for b, b† , and all the cross-commutators involving mixing of a, a† with b, b†
are zero. Furthermore, one also gets {a, a} = {b, b} = 0.
To clarify and summarize the above discussion, we write the main steps below:
1) We start from some symplectic form ω and from two quantities f, g of which we
want to compute the Poisson brackets. For the moment, we restrict our attention
to the case when f, g are some creation/annihilation operators which are part
of the phase space of the theory (and therefore δf, δg appear somewhere in the
definition of ω);
2) Use the definitions Xf ⌟ ω = −δf and Xg ⌟ ω = −δg to obtain the vector fields
Xg , Xf ;
3) The definition of the vector fields themselves will imply that there is one correct
order of contraction of Xg , Xf with ω. Using this contraction, one defines the
relevant Poisson bracket, say for example {f, g} = Xg ⌟ (Xf ⌟ Ω);
4) We now want to relate the brackets {f, g} and {g, f }. Notice that, because of
the previous point, only {f, g} contains the appropriate contractions of Xf , Xg
with ω. In order to correctly define {g, f }, therefore, we make use of a global −S
to be added to the reversed contraction, defining {g, f } = −S(Xf ⌟ (Xg ⌟ Ω)),
i.e.
{f, g} = Xg ⌟ (Xf ⌟ Ω) ⇔ {g, f } = −S(Xf ⌟ (Xg ⌟ Ω)) . (2.353)
Notice of course that this gives the correct κ → ∞ limit.
5) When f, g are creation/annihilation operators, the above procedure ensures that

one gets the crucial property {f, g} = −{g, f }, as well as the correct value for
the Poisson brackets.
2.15 The algebra of the κ-deformed charges

We are now at point viii) of our roadmap in section 2.1. We now verify that the
translation, rotation, and boost charges, which are respectively eq. (2.259), (2.260),
(2.307), (2.279), satisfy the canonical, non-deformed Poincaré algebra.
[Mi , Pj ] = iϵijk Pk [Mi , P0 ] = 0 [Ni , Pj ] = −iηij P0 [Ni , P0 ] = −iPi

(2.354)
[Mi , Mj ] = iϵijk Mk [Mi , Nj ] = iϵijk Nk [Ni , Nj ] = −iϵijk Mk (2.355)
Recall that we are using the convention η = diag(+ − −−).

Since the computations are long but straightforward, we will only explicitly verify
two of these commutators, but the rest of them can be verified using the same exact
procedures.
2.15. The algebra of the κ-deformed charges 75
2.15.0.1 [Ni , Pj ] = −iηij P0

The commutator [Ni , Pj ] contains two components, i.e.
(" # )
∂a†p † † ∂ap †
−iα(p)α(q)S(ωp )S(q)j ap , aq aq − ap , a aq (2.356)
∂S(p)i ∂S(p)i q
(" # )
∂b†p ∂bp †
i α(p)α(q)ωp qj bp i , bq b†q − b , bq b†q (2.357)
∂p ∂pi p
Starting from the first one, we have

" #
∂a†p
− iα(p)α(q)S(ωp )S(q)j ap , a†q aq (2.358)
∂S(p)i
( " # )
∂a†p ∂a†
p
= −iα(p)α(q)S(ωp )S(q)j [ap , a† ] aq + a†q , aq ap (2.359)
∂S(p)i | {z q} ∂S(p)i
1
=α δ(p−q) | {z }
1 ∂
=− α δ(q−p)
∂S(p)i
∂a†p
= −iα(p)S(ωp )S(q)j δ(p − q)aq
∂S(p)i

† ∂
+ iα(p)S(ωp )S(q)j aq δ(q − p) ap (2.360)
∂S(p)i
Integrating by parts the last term in the last passage above we get
∂ ∂(α(p)S(ωp ))
−iα(p)S(ωp )S(q)j a†q δ(q − p) i
ap − i S(q)j a†q δ(q − p)ap
∂S(p) ∂S(p)i
(2.361)
and substituting back in eq. (2.360) we get (we can now safely apply the Dirac delta
since no derivative acts on it)

∂ † ∂(α(p)S(ωp )) †
−iα(p)S(ωp )S(p)j (a ap ) − i S(p)j ap ap δ(q − p) (2.362)
∂S(p)i p ∂S(p)i
∂ ∂
= −i (α(p)S(ωp )S(p)j a†p ap ) + iS(p)j a†p ap (α(p)S(ωp ))
∂S(p)i ∂S(p)i
!
∂(α(p)S(ω p ))
+ iα(p)S(ωp )ηij a†p ap − i S(p)j a†p ap δ(q − p) (2.363)
∂S(p)i
= iα(p)S(ωp )ηij a†p ap δ(q − p) (2.364)
†
where in the last passage we ignored ∂S(p)
∂
i (α(p)S(ωp )S(p)j ap ap ) since it is a surface
term after the application of the Dirac delta.

We can analogously proceed for the second term, as follows
" #
∂b†p
i α(p)α(q)ωp qj bp i , bq b†q (2.365)
∂p
( " # )
∂b†p † † ∂b†p
= i α(p)α(q)ωp qj bp , bq bq + 0 + 0 + bq [bp , bq ] (2.366)
∂pi | {z } ∂pi
1
| {z } =α δ(p−q)
1 ∂
=− α δ(q−p)
∂pi
∂b†p ∂bp
= +i α(p)ωp qj bq δ(p − q) i
+ i α(p)ωp qj i δ(q − p)b†q
∂p ∂p
∂(α(p)ωp )
+ i qj bp δ(q − p)b†q (2.367)
∂pi

∂(α(p)ωp ) † ∂ †
= i pj bp bp + i α(p)ωp qj i (bp bp ) δ(p − q) (2.368)
∂pi ∂p
∂ ∂
= i i
(pj α(p)ωp bp b†p ) − i pj α(p)ωp i (bp b†p ) − i ηij α(p)ωp bp b†p
∂p ∂p
∂
+ i α(p)ωp qj i (bp b†p ) δ(p − q) (2.369)
∂p
= −i ηij α(p)ωp bp b†p δ(p − q). (2.370)
In summary, we have shown that

Z
[Ni , Pj ] = d3 p d3 q iα(p)S(ωp )ηij a†p ap − i ηij α(p)ωp bp b†p δ(q − p) (2.371)
= −iηij P0 (2.372)
which is what we wanted to show.
2.15.0.2 [Ni , Nj ] = −iϵijk Mk

The commutator [Ni , Nj ] contains two parts, the a, a† part and the b, b† part. Starting
with the one concerning the a’s we have
"( ) ( )#
∂a†p † ∂ap ∂a†q † ∂aq
− α(p)α(q)S(ωp )S(ωq ) ap − ap , aq − aq
∂S(p)i ∂S(p)i ∂S(q)j ∂S(q)j
(2.373)
(" # " #
∂a†p ∂a†q ∂a†p ∂aq
= − α(p)α(q)S(ωp )S(ωq ) i
ap , j
a q + i
ap , −a†q
∂S(p) ∂S(q) ∂S(p) ∂S(q)j
" # " #)
† ∂ap ∂a†q † ∂ap † ∂aq
+ − ap , aq + − ap , −aq (2.374)
∂S(p)i ∂S(q)j ∂S(p)i ∂S(q)j
We will compute these commutators one by one starting from the first. Once again,
we use
[AB, CD] = A[B, C]D + [A, C]BD + CA[B, D] + C[A, D]B (2.375)
and we get
" #
∂a†p ∂a†q
− α(p)α(q)S(ωp )S(ωq ) ap , aq (2.376)
∂S(p)i ∂S(q)j
( " # " # )
∂a†p ∂a†q ∂a†q ∂a†p
= −α(p)α(q)S(ωp )S(ωq ) ap , aq + , aq ap
∂S(p)i ∂S(q)j ∂S(q)j ∂S(p)i
(2.377)
( " # " # )
∂a†p ∂a†q ∂a†q ∂a†p
= −α(p)α(q)S(ωp )S(ωq ) ap , aq − aq , ap
(2.378)
Now we notice that the second term is just the first one with p ↔ q and i ↔ j, but
we can just rename the integration variables in such a way that the second one and
the first are the same except with i ↔ j, and therefore the above integral becomes
" #
∂a†p ∂a†q
−α(p)α(q)S(ωp )S(ωq ) ap , aq (2.379)
∂S(p)[i ∂S(q)j]
where we just used the antisymmetrized index to shorten the notation. Now we have
a small subtlety related to the presence of the α1 in front of the delta coming from the
commutator. In fact, recall from eq. (2.208) that
h i 1
ap , a†q = δ(p − q). (2.380)
α
However we have two choices, one where the α1 on the RHS depends on p and the
other where it depends on q. We are in the following situation
" #
∂a†q

∂ 1
f (p, q)α(p)α(q) ap , = f (p, q)α(p)α(q) δ(p − q) (2.381)
∂S(q) ∂S(q) α
Now, if the α in the round brackets depends on p, then the above formula reduces to

∂ int. by parts and appl.δ ∂
f (p, q)α(q) δ(p − q) −−−−−−−−−−−−−−→ − f (q)α(q) δ(p − q)
∂S(q) ∂S(q)
(2.382)
However, if the α in the round brackets depends on q, then we have

∂ 1
f (p, q)α(p)α(q) δ(p − q) (2.383)
∂S(q) α
1 ∂α(q) ∂
= −f (p, q)α(p)α(q) 2 δ(p − q) + f (p, q)α(p) δ(p − q) (2.384)
α (q) ∂S(q) ∂S(q)
∂α(q) ∂
= −f (q) δ(p − q) + f (p, q)α(p) δ(p − q) (2.385)
∂S(q) ∂S(q)

∂α(q) ∂
= −f (q) δ(p − q) − f (p, q) α(p)δ(p − q) (2.386)
∂S(q) ∂S(q)

∂
=− f (q)α(q) δ(p − q) (2.387)
∂S(q)
where in eq. (2.385) we applied the Dirac delta in the first term (because it is not acted
upon by any derivatives) and in eq. (2.386) we integrated by parts the second term
on the RHS. One sees that both approaches lead to the same final result. Therefore
for simplicity we will always consider the α1 in the definition of [a, a† ] to depend on
the variable that is not acted upon by the derivation inside the commutator.
Coming back to eq. (2.379), we have
" #
∂a†p ∂a†q
− α(p)α(q)S(ωp )S(ωq ) ap , aq (2.388)
∂S(p)[i ∂S(q)j]
∂a†p ∂
= −α(q)S(ωp )S(ωq ) aq δ(p − q) (2.389)
∂S(p)[i ∂S(q)j]
†
∂α(q) 2 ∂aq ∂a†q †
2 ∂aq ∂aq
= [j
S(ωq ) i]
aq + α(q)S(q)[j i]
aq + α(q)S(ωq )
∂S(q) ∂S(q) ∂S(q) ∂S(q) ∂S(q)j]
[i
(2.390)
We can now integrate by parts the first term of the last step, which (ignoring surface
terms) will give us
†
∂α(q) 2 ∂aq
S(ωq ) aq (2.391)
∂S(q)[j ∂S(q)i]
∂S(ωq ) ∂a†q †
2 ∂aq ∂aq
= −α(q)2 [j
S(ω q ) i]
a q − α(q)S(ωq ) (2.392)
∂S(q) ∂S(q) ∂S(q) ∂S(q)j]
[i
∂a†q †
2 ∂aq ∂aq
= −2α(q)S(q)[j i]
a q − α(q)S(ωq ) . (2.393)
∂S(q) ∂S(q) ∂S(q)j]
[i
Notice that there is no term with wo derivatives because the antisymmetrization sends
it to zero. Substituting this back into eq. (2.390) we finally obtain that all the terms
go away with the exception of one of them, and we get
" #
∂a†p ∂a†q ∂a†q
−α(p)α(q)S(ωp )S(ωq ) ap , aq = −α(q)S(q)[j aq (2.394)
∂S(p)i ∂S(q)j ∂S(q)i]
∂a†q
= α(q)S(q)[i aq (2.395)
∂S(q)j]
which is the expected result.

We now pass to the second and third commutator. We treat them together because
it is easy to show that their sum is zero. In fact, we have
(" # " #)
∂a†p ∂aq ∂ap ∂a†
q
ap , −a†q + − a†p , aq (2.396)
∂S(p)i ∂S(q)j ∂S(p)i ∂S(q)j
(" # " #)
∂a†p † ∂aq ∂a†q † ∂ap
= ap , −aq − aq , −ap (2.397)
and we notice that the second commutator is exactly the same as the first one but
with p ↔ q. However, as was done previously, since we are integrating in p and q, we
can just rename them and therefore switch them. Therefore, the sum of the second
and third term will be
(" # " #)
∂a†p † ∂aq ∂a†p † ∂aq
ap , −aq − ap , −aq =0 (2.398)
∂S(p)i ∂S(q)j ∂S(p)j ∂S(q)i
which is what we wanted to show.

We can finally deal with the last term. We have
" #
† ∂ap † ∂aq
− α(p)α(q)S(ωp )S(ωq ) ap ,a (2.399)
∂S(p)i q ∂S(q)j
( )
∂ap ∂aq ∂aq ∂ap
= −α(p)α(q)S(ωp )S(ωq ) a†p ,a† † †
+ aq ap ,
∂S(p)i q ∂S(q)j ∂S(q)j ∂S(p)i
(2.400)
( )
∂ap ∂aq ∂aq ∂ap
− α(p)α(q)S(ωp )S(ωq ) a†p , a† − a†q , a†
∂S(p)i q ∂S(q)j ∂S(q)j p ∂S(p)i
(2.401)
Once again, we can without problem switch p ↔ q on the second term, obtaining
finally

† ∂ap † ∂aq
− α(p)α(q)S(ωp )S(ωq )ap [i
, aq (2.402)
∂S(p) ∂S(q)j]
Proceeding as before, we have

† ∂ap † ∂aq
− α(p)α(q)S(ωp )S(ωq )ap [i
, aq (2.403)
∂S(p) ∂S(q)j]
∂aq ∂
= −α(p)S(ωp )S(ωq )a†p δ(p − q) (2.404)
∂S(q) ∂S(p)i]
[j
†
∂α(q) 2 † ∂aq † ∂aq 2 ∂aq ∂aq
= [i
S(ωq ) aq j]
+ α(q)S(q)[i aq j]
+ α(q)S(ωq )
∂S(q) ∂S(q) ∂S(q) ∂S(q) ∂S(q)j]
[i
(2.405)
Once again, we integrate by parts the first term, obtaining

†
∂α(q) 2 † ∂aq † ∂aq 2 ∂aq ∂aq
[i
S(ω q ) a q j]
= −2α(q)S(q) a
[i q j]
− α(q)S(ωq )
∂S(q) ∂S(q) ∂S(q) ∂S(q) ∂S(q)j]
[i
(2.406)
and substituting back we get the final result

† ∂ap † ∂aq ∂aq
−α(p)α(q)S(ωp )S(ωq )ap [i
, aq j]
= −α(q)S(q)[i a†q (2.407)
∂S(p) ∂S(q) ∂S(q)j]
Finally, the result of the commutator coming from the a, a† part is given by
∂a†q
Z
1 ∂aq
d3 p d3 q α(q)S(q)[i j]
aq − α(q)S(q)[i a†q (2.408)
4 ∂S(q) ∂S(q)j]
which is what was expected.

For the commutators coming from the b, b† part, we have
(" # " #
∂b†p ∂b†q ∂b†p ∂bq †
−α(p)α(q)ωp ωq bp i , bq j + bp i , − j bq (2.409)
∂p ∂q ∂p ∂q
" # )
∂bp † ∂b†q ∂bp † ∂bq †
+ − i bp , bq j + − i bp , − j bq (2.410)
∂p ∂q ∂p ∂q
The second and third term go away for the same reasons as before, and only the first
and last term remain.
Notice that if one (formally) switches b → ã† and b̃† → a, one gets exactly the
same commutators as the a, a† case7 but with an additional minus sign. This minus
sign comes from the fact that sending b → ã† and b̃† → a also sends δ = [b, b† ] →
[ã† , ã] = −δ. Therefore, one can do the exact same computations as for a, with the
additional minus sign coming from the deltas. Therefore the final result will be
∂ã†q ∂ãq
−α(q)q[i j]
ãq + α(q)q[i ã†q j] (2.411)
∂q ∂q
or, in other words,
∂bq † ∂b†q
−α(q)q[i b + α(q)q b
[i q (2.412)
∂qj] q ∂qj]
One can also double check the above result by direct computations. Therefore, the
b, b† contribution to the commutator is given by
!
∂b†q
Z
1 ∂bq †
3 3
d p d q α bq q[i j] − q[i j] bq (2.413)
4 ∂q ∂q
Finally, from these results, one can immediately say that

!
∂a†q ∂b†q
Z
1 † ∂aq ∂bq
[Ni , Nj ] = d3 q α(q) S(q)[i j]
a q − a q S(q) [i j]
+ bq q [i j]
− q[i j] b†q
4 ∂S(q) ∂S(q) ∂q ∂q
(2.414)
Calling the RHS Mij and noticing that Mk = 12 ϵkij Mij we get the result.
2.16 C and its commutator with the boosts

We are now at point ix) of our roadmap in section 2.1. The C operator in our context
is given by
Z
C= d3 p α(b†p ap + a†p bp ) (2.415)
In fact, it is easy to show that this operator satisfies the canonical C property
Ca† |0⟩ = b† |0⟩ Cb† |0⟩ = a† |0⟩ (2.416)
where we used the fact that C −1 |0⟩ = |0⟩. In fact, using eq. (2.208) and (2.209) we
have
Z
†
Cap |0⟩ = d3 q α(b†q aq + a†q bq )a†p |0⟩ (2.417)
Z
= d3 q αb†q aq a†p |0⟩ + 0 (2.418)
Z
1
= d3 q αb†q δ(p − q) + a†p aq |0⟩ (2.419)
α
= b†p |0⟩ (2.420)
7
The fact that the derivatives and the prefactors do not have the antipode is irrelevant in this
context.
2.16. C and its commutator with the boosts 81
and the computation for the verification of the relation Cb† |0⟩ = a† |0⟩ proceed in the
same way. Notice that C = C † and C 2 = 1. In fact we have
Z
C ak |0⟩ = d3 p d3 q α(p)α(q)(b†p ap + a†p bp )(b†q aq + a†q bq )a†k |0⟩
2 †
(2.421)
Z
= d3 p α(p)(b†p ap + a†p bp )b†k |0⟩ (2.422)
= a†k |0⟩ (2.423)
and the same if we started from b†q |0⟩. Finally, notice that the C operator in eq. (2.415)
is a symmetry of the off-shell action, since acting by conjugation we only exchange
a(†) ↔ b(†) .
We now compute the commutator between the charge conjugation operator and
the boost operator. Using the boost charge in eq. (2.279) we have
( " # " #
Z † †
i ∂a p ∂ap
[Ni , C] = d3 p d3 q α(p)α(q) S(ωp ) ap , b†q aq + S(ωp ) ap , a†q bq
2 ∂S(p)i ∂S(p)i
(2.424)

∂ap ∂ap
+ S(ωp ) −a†p , b†q aq + S(ωp ) −a†p , a† bq (2.425)
∂S(p) i ∂S(p)i q
" # " #
∂b†p † ∂b†p †
+ ωp bp i , bq aq + ωp bp i , aq bq (2.426)
∂p ∂p
)
∂bp † † ∂bp † †
+ ωp − i bp , bq aq + ωp − i bp , aq bq (2.427)
∂p ∂p
We compute these commutators one by one using the commutator property
[AB, C] = A[B, C] + [A, C]B (2.428)
We have the following.
• First
" #
∂a†p †
α(p)α(q)S(ωp ) ap , bq aq (2.429)
∂S(p)i
" #
∂a†p
= 0 + α(p)α(q)S(ωp ) , aq ap b†q (2.430)
∂S(p)i
∂
= −α(p)S(ωp )ap b†q δ(p − q) (2.431)
∂S(p)i
( )
∂α S(p)i ∂ap †
= S(ωp ) + α ap b†q + α(p)S(ωp ) b δ(p − q)
∂S(p)i S(ωp ) ∂S(p)i q
(2.432)
∂b†p
= −α(p)S(ωp )ap (2.433)
∂S(p)i
where in the last passage we integrated once again by parts.

• Second
" #
∂a†p †
α(p)α(q)S(ωp ) ap , aq bq (2.434)
∂S(p)i
∂a†p h †
i
= α(p)α(q)S(ωp ) a p q bq + 0
, a (2.435)
∂S(p)i
∂a†p
= α(p)S(ωp ) bp (2.436)
∂S(p)i
• Third

∂ap
α(p)α(q)S(ωp ) −a†p †
, b aq (2.437)
∂S(p)i q
h i ∂a
p
= 0 + α(p)α(q)S(ωp ) −a†p , aq b† (2.438)
∂S(p)i q
∂ap †
= α(p)S(ωp ) b (2.439)
∂S(p)i p
• Fourth

† ∂ap †
α(p)α(q)S(ωp ) −ap , a bq (2.440)
∂S(p)i q

† ∂ap †
= −α(p)α(q)S(ωp )ap , a bq + 0 (2.441)
∂S(p)i q
∂
= −α(p)S(ωp )a†p bq δ(p − q) (2.442)
∂S(p)i
( )
∂ap†

∂α S(p)i
= i
S(ωp ) + α a†p bq + α(p)S(ωp ) bq δ(p − q)
∂S(p) S(ωp ) ∂S(p)i
(2.443)
∂bp
= −α(p)S(ωp )a†p (2.444)
∂S(p)i
• Fifth
" #
∂b†p † h i ∂b†
p
α(p)α(q)ωp bp i , bq aq = 0 + α(p)α(q)ωp bp , b†q aq (2.445)
∂p ∂pi
∂b†p
= α(p)ωp ap (2.446)
∂pi
• Sixth
" #
∂b†p †
α(p)α(q)ωp bp i , aq bq (2.447)
∂p
" #
∂b†p
= α(p)α(q)ωp bp , bq a†q + 0 (2.448)
∂pi
∂
= −α(p)ωp bp a†q δ(p − q) (2.449)
∂pi
2.16. C and its commutator with the boosts 83
( )
∂α pi ∂bp †
= ωp + α bp a†q + α(p)ωp i aq δ(p − q) (2.450)
∂pi ωp ∂p
∂a†p
= −α(p)ωp bp i (2.451)
∂p
• Seventh

∂bp
α(p)α(q)ωp − i b†p , b†q aq (2.452)
∂p

∂bp † †
= 0 + α(p)α(q)ωp − i , bq bp aq (2.453)
∂p
∂
= −α(p)ωp b†p aq i δ(p − q) (2.454)
∂p
( )
∂b†p

∂α pi †
= ωp + α b aq + α(p)ωp i aq δ(p − q) (2.455)
∂pi ωp p ∂p
∂ap
= −α(p)ωp b†p (2.456)
∂pi
• Eighth

∂bp † † ∂bp h i
α(p)α(q)ωp − i bp , aq bq = −α(p)α(q)ωp i b†p , bq a†q (2.457)
∂p ∂p
∂bp †
= α(p)ωp i ap (2.458)
∂p
Putting all these computations together, we are left with

( " #
Z † †
i ∂ap ∂bp ∂ap ∂bp
[Ni , C] = d3 p α(p) S(ωp ) b† − ap + bp − a†p
2 ∂S(p)i p ∂S(p)i ∂S(p)i ∂S(p)i
" #)
∂b†p † ∂ap ∂bp † ∂a†p
+ ωp a p − bp + a − bp (2.459)
∂pi ∂pi ∂pi p ∂pi
which (as expected) is different than zero.

Finally, notice that in the non-deformed case, all the computations will proceed
in exactly the same way but without antipodes (recall that α(p) is a global prefactor
such that α → 1 when κ → ∞). Furthermore, notice also that
∂ κ→∞ ∂ ∂
S(A) −−−→ −A =A . (2.460)
∂S(B) ∂(−B) ∂B
Therefore, the above result in the non-deformed case reduces to

( " #
Z † †
i ∂ap ∂b p ∂ap ∂bp
[Ni , C] = d3 p ωp b† − ap i + bp − a†p i
2 ∂pi p ∂p ∂pi ∂p
" #)
∂b†p † ∂ap ∂bp † ∂a†p
+ ωp ap − bp + a − bp (2.461)
∂pi ∂pi ∂pi p ∂pi
=0 (2.462)
where in the last passage we used the fact that, because of the limit κ → ∞, the two
square brackets are equal and opposite.
2.17 CP T ⇔(
Lorentz?
(
((((
In (92) it was shown that CP T violation implies Lorentz violation and vice versa.
The theorem is obtained in the context of axiomatic QFT, and in particular on the
properties of Wightman functions. The relation between CP T invariance and Lorentz
invariance is still not clear in the non-deformed context (100), (101). In our context, κ-
deformation introduces non-trivialities already at the one-particle level even in absence
of interaction, a fact which can be intuitively understood by noticing that −S(p) ̸= p.
Furthermore, as discussed in the previous section, although the action is manifestly
CP T and invariant under κ-Poincaré, the Noether charges obtained from continuous
symmetries of the action (and in particular the boost charge) have very non-trivial
relations with the discrete symmetries (in this particular case C). A theorem like (92)
is therefore not automatically satisfied in the κ-deformed context. In what follows, we
reproduce the reasoning in the non-deformed case, explicitly performing all the steps
and adding some figures, and at the end we write a discussion on possible departures
from the conclusions of the theorem. We are at point x) of our roadmap in section
2.1.
2.17.1 Greenberg’s argument

The proof of Greenberg’s theorem is based on a theorem by Jost in (95) (a scan of
the original paper in german can be found in (98), see also (96) for a description
of this argument. Greenberg has also described the theorem and its steps in (97)).
First we write down what Jost theorem says without proof, then we write down the
setup for Greenberg’s argument, and finally the argument itself. Since Jost theorem
is fundamental for Greenberg’s proof, we will describe it in the next section. The
description of this theorem provided in this section is solely based on the provided
references, and we will use the same notation and setup.
Jost theorem states the following:
Jost theorem: Given the (Wightman) functions
W (n) (x1 , x2 , . . . , xn ) := ⟨0|ϕ1 (x1 )ϕ2 (x2 ) . . . ϕn (xn )|0⟩ (2.463)
the validity of the weak local commutativity condition at Jost points
W (n) (x1 , x2 , . . . , xn ) = W (n) (xn , xn−1 , . . . , x1 ) (2.464)
is equivalent to the validity of the CP T symmetry. The Jost points {xi } are a set of
points such that the linear combination
X
ci (xi − xi−1 ) (2.465)
i
is always spacelike, for any choice of ci ≥ 0 with

P
i ci > 0.
Setup: We consider only passive transformations. A quantum field theory is con-

sidered Lorentz covariant if it is both in-cone covariant and out-of-cone covariant. A
T ⇔ 85

CP
2.17. Lorentz?

quantum field theory is said to be in-cone covariant if the Wightman functions are co-
variant (notice that the Wightman functions are matrix elements of unordered fields).
On the other hand, a quantum field theory is said to be out-of-cone covariant if the
matrix elements of time-ordered fields τ are covariant. In the notation used in (92),
the τ functions are written as
X
τ= θ(x0σ1 , x0σ2 , . . . , x0σn )W (n) (xσ1 , xσ2 , . . . , xσn ) (2.466)
σ
where the function θ(xσ1 , xσ2 , . . . , xσn ) ensures that x0σ1 ≥ x0σ2 ≥ · · · ≥ x0σn . The sum
goes over all the permutations of n elements.
Argument: The idea is to show that violation of CP T (in the form of violation
of the weak local commutativity condition) implies a violation of the out-of-cone
covariance condition, and since a theory is Lorentz covariant if it is covariant both
in-cone and out-of-cone, we get a violation of the Lorentz covariance of the theory.
Since we want to use Jost theorem, consider points x1 , . . . , xn separated by spa-
celike intervals. For these points, there can be Lorentz transformations that reverse
their time ordering. Therefore, in order for τ to be invariant under Lorentz trans-
formation, the Wightman functions must not depend of the order of the fields in its
argument (when these fields are evaluated at points with spacelike separation). In fact,
say that there exist some Lorentz transformation such that one goes from x0σ1 ≥ x0σ2
to x0σ2 ≥ x0σ1 . Then the function (2.466) gets sent to
!
X
τ = τ̃ = θ(x0σ2 , x0σ1 , . . . , x0σn )W (n) (xσ1 , xσ2 , . . . , xσn ) (2.467)
σ
X
= θ(x0σ1 , x0σ2 , . . . , x0σn )W (n) (xσ2 , xσ1 , . . . , xσn ) (2.468)
σ
and therefore we must have
W (n) (xσ1 , xσ2 , . . . , xσn ) = W (n) (xσ2 , xσ1 , . . . , xσn ). (2.469)
Since the same can be repeated for any choice of points, then indeed for spacelike-
separated points the W (n) functions do not depend on the order of the fields.
Now we can chose the points separated by spacelike intervals to be Jost sets, and
we can chose them in such a way that their successive time differences are all positive.
Graphically speaking, Jost points of this kind can be represented for example as follows
(in red)
Then there exist some Lorentz transformation such that all the successive time diffe-
rences are all negative, i.e. we are in the following situation.
However, because of the invariance of τ under such a transformation, then for these
Jost points we must have
W (n) (x1 , x2 , . . . , xn ) = W (n) (xn , xn−1 , . . . , x1 ) (2.470)
which is the weak local commutativity of Jost theorem. Therefore, Lorentz symmetry
implies weak local commutativity, which by Jost theorem is equivalent to CPT sym-
metry, and vice versa if CPT is violated, then the same is true for Lorentz symmetry.
2.17.2 Jost theorem

We use as a reference Greenberg’s paper (97) where he describes in detail Jost proof.
In this section we will just write down the main steps of the proof, using the same
notation used in (97).
Jost proof follows the following steps:
1) The key idea is to use the group SL(2, C), which is the double covering of
the proper orthochronous Lorentz group L↑+ . One uses SL(2, C) instead of L↑+
because spacetime inversion P T = −1 is connected to the identity in SL(2, C),
while it is not connected in L↑+ .
2) To use this extended Lorentz group, one needs to use the analytic continuation
of the Wightman functions, i.e. of the vacuum expectation values of products
of fields. However, instead of using the canonical W (n) (x1 , . . . , xn ), because of
translation invariance one can use the equivalent function W̃ (x1 −x2 , . . . , xn−1 −
xn ) := W̃ (ξ1 , . . . , ξn−1 ) where ξj = xj − xj+1 . In order to obtain an analytic
continuation of W̃ , we promote each difference ξj to a complex variable, and in
particular
zj = ξj − iηj (2.471)
The introduction of the −iηj transforms the distribution W̃ (ξ1 , . . . , ξn−1 ) into
an analytic function W̃ (z1 , . . . , zn−1 ).
Notice that the use of the differences ξj is a necessary step for the proof. In
fact, the whole argument rests on the existence of real points in the analyticity
domain of the analytic function W̃ (z1 , . . . , zn−1 ). These real points {ξj } are the
Jost points discussed above, and are defined by the condition that the sum
X
s := cj ξj (2.472)
j
is spacelike, where cj ≥ 0 and j cj > 0. Of course, since s must be spacelike

P
for any observers, it must be defined in terms of ξj and not xj .
T ⇔ 87

CP
2.17. Lorentz?

3) The newly defined analytic functions W̃ (z1 , . . . , zn−1 ) are still defined only in a
domain which does not include real points. However, because of Lorentz inva-
riance with respect to the double covering group SL(2, C), we have
W̃ (Λz1 , . . . , Λzn−1 ) = W̃ (z1 , . . . , zn−1 ) (2.473)
This allows us to enlarge the domain, and the new domain does contain the Jost
points. Furthermore, because of the fact that in SL(2, C) there is spacetime
inversion, there is some Λ such that
W̃ (Λz1 , . . . , Λzn−1 ) = (−1)L W̃ (−z1 , . . . , −zn−1 ) (2.474)
for some coefficient L derived in (97). Writing the above relation at the Jost
points, using the vacuum expectation values instead of the Wightman functions
and using eq. (2.473) we get8
⟨0|ϕ1 (x1 ) . . . ϕn (xn )|0⟩ = (−1)L ⟨0|ϕ1 (−x1 ) . . . ϕn (−xn )|0⟩. (2.475)
4) There is now a small but important subtlety in the above equality. In fact,
the equality merely states that the value of the function on the LHS is the
same as the one on the RHS. However (essentially because of the antisymmetric
nature of the T operator) the function (−1)L ⟨0|ϕ1 (−x1 ) . . . ϕn (−xn )|0⟩ has a
different domain of definition than the function ⟨0|ϕ1 (x1 ) . . . ϕn (xn )|0⟩. The way
to solve this is to permute the fields inside the expectation value on the RHS
of eq. (2.475). Of course, in doing so one needs to use the fact that boson
fields commute when evaluated at different spacetime points and fermion fields
anticommute, i.e.
[ϕ(x), ϕ(y)] = [A(x), A(y)] = {ψ(x), ψ(y)} = 0 x ̸= y (2.476)
Upon using these relations, one can write that
⟨0|ϕ1 (x1 ) . . . ϕn (xn )|0⟩ = (i)F ⟨0|ϕn (xn ) . . . ϕn (x1 )|0⟩ (2.477)
where F is the number of fermions in the expectation value. The identity (2.477)
is called ‘weak local commutativity’, and for the sake of the theorem it only needs
to hold at Jost points. Notice that weak local commutativity is not needed for
the theorem if, instead of considering just Wightman functions, one uses the
time ordered products of Wightman functions τ analogous to the one in eq.
(2.466).
5) Using the weak local commutativity, eq. (2.475) becomes
⟨0|ϕ1 (x1 ) . . . ϕn (xn )|0⟩ = (−1)L (i)F ⟨0|ϕn (−xn ) . . . ϕ1 (−x1 )|0⟩. (2.478)
Now both functions in both sides have the same domain of definition, and since
they are analytic and coincide in some open subset of their domain (i.e. the
neighbour(s) of the Jost points) the above equality implies the same equality of
Wightman functions across the whole domain. Therefore we have
W̃ (z1 , . . . , zn ) = (−1)L (i)F W̃ (−zn , . . . , −z1 ). (2.479)

8
The fields ϕi in this expression can be any kind of field, boson or fermion.
At this point, we can also take the limit ηj → 0 and the analytic functions
become once again distributions, and we have
W̃ (ξ1 , . . . , ξn ) = (−1)L (i)F W̃ (−ξn , . . . , −ξ1 ). (2.480)
Notice that we couldn’t take this limit before using the weak local commutativity
because the domains of definitions of the functions were different. Therefore, we
would have been stuck to the complex case, without the possibility of taking the
limit and coming back to the real (physical) case.
6) We can write eq. (2.480) in terms of vacuum expectation value as
⟨0|ϕ1 (x1 ) . . . ϕn (xn )|0⟩ = (−1)L (i)F ⟨0|ϕn (−xn ) . . . ϕ1 (−x1 )|0⟩ (2.481)
Notice that this equation is formally equivalent to eq. (2.478), but while eq.
(2.478) only made sens in a neighbour of Jost points (i.e. the associated Wight-
man function was an analytic function) here everything is real (i.e. the Wight-
man functions are again distributions). One can restore the order of the fields
through hermitian conjugation on the RHS obtaining
⟨0|ϕ1 (x1 ) . . . ϕn (xn )|0⟩ = (−1)L (i)F ⟨0|ϕ†1 (−x1 ) . . . ϕ†n (−xn )|0⟩∗ (2.482)
and from this one can read off the CPT transformation which reads9
θϕ(x)θ† = (−1)l if ϕ† (−x). (2.483)
7) With the points 1) to 6) one has proved that (assuming Lorentz symmetry)
the validity of the weak local commutativity at Jost points (2.477) implies CPT
symmetry. To show the other direction one can just use the same steps in reverse
order, and therefore weak local commutativity in the neighbour of Jost points
is equivalent to CPT symmetry.
2.17.3 Comments on Jost and Greenberg theorems

There are already some doubts about the validity of Greenberg’s argument for cano-
nical quantum field theory (100), (101). Apart from these arguments, one can also say
the following.
• The first problematic step in the whole argument is point 4) of the proof of Jost
theorem. In fact, in the non-deformed context the (anti)commutator of fields
reduces to the (anti)commutators between creation/annihilation operators be-
cause plane waves are just functions and commute with everything (themselves
included). Therefore, the statement about field is just a rewriting of the same
statement about creation/annihilation operators. However, in the deformed con-
text, plane waves do not commute and their product is described in terms of
star operator and antipode. This fact, coupled with the deformed commutators
in (2.208), (2.209), immediately implies that (anti)commutation relations for
(fermion) fields are at best not obvious, and its imposition could very well lead
to the destruction of some κ-deformed effects.
• On a more pragmatical note, the definition of something like ⟨0|ϕ1 (x1 )⋆ϕ2 (x2 )|0⟩
requires the knowledge of the commutator [x01 , x2 ]. Several proposals have been
9
The indices l and f refer to the number of dotted and undotted indices in the field representation
used in (97), but it is not important for the proof.
T ⇔ 89

CP
2.17. Lorentz?

given for this commutator, each with its own pros and cons, highlighting the
fact that an objective, natural definition of such commutator is lacking. As such,
although any of the definition could be taken as assumption, different choices
would result in different mathematical formulations of ⟨0|ϕ1 (x1 ) ⋆ ϕ2 (x2 )|0⟩,
with different physical properties. This issue could be sidestepped by going to
momentum space, where a natural definition of two-point function could be
given using the path integral approach (see e.g. chapter 3). However, if we have
more than two fields in the VEV, one has the issue that the Fock space is not
well defined, see for example (99).
• As stated in (97), one does not need to use any assumption related to the
(anti)commutators between fields if, instead of using just the Wightman func-
tion, one uses the time-ordered product τ of such functions as in eq. (2.466). This
of course is true also in the deformed context, but in light of the discussion on
the previous point it is not clear whether the τ in (2.466) does indeed correspond
to the correct physical n-point function. For example, in (85) it was shown that
the time-ordered vacuum expectation value (VEV) of two fields gives the correct
propagator only in some subset of phase space (for trans-Planckian momenta
the propagator is not the time-ordered VEV). The model considered in (85)
is different from the one discussed in this work, nevertheless it highlights the
fact that general conclusions about the behaviour of time-ordered VEV are in
general groundless, and a lot depends on the considered model. In other words,
one could question whether the correct physical quantities to be considered are
time ordered products of Wightman functions in the first place.
• The weak local commutativity in the proof of Jost theorem needs to hold at
Jost points which are defined in terms of difference of coordinates. However, how
to sum or subtract coordinate vectors is not a trivial matter in the deformed
context.
As a concrete proof of the above considerations, in what follows we will explicitly

show that the two-point function ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ satisfies the CPT invariance
relation, in the form of eq. (2.482), but it is not Lorentz invariant, meaning that eq.
(2.473) (in the real limit) is not satisfied, and therefore both Jost and Greenberg
theorem are not valid.
2.17.3.1 CP T properties of ⟨0|ϕ†1 (x1 ) ⋆ ϕ2 (x2 )|0⟩

We want to show that
⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ = ⟨0|ϕ(−x1 ) ⋆ ϕ† (−x2 )|0⟩∗ (2.484)
In order to do so, we use the fields in eq. (2.115) and (2.116), and we have
⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩

d4 q d4 p
Z
= θ(p0 ) ζ(p)ζ(q) ⟨0|bq b†p |0⟩e−i(ωq t1 −qx1 ) ⋆ e−i(S(ωp )t2 −S(p)x2 ) (2.485)
q4 /κ p4 /κ
⟨0|ϕ(−x1 ) ⋆ ϕ† (−x2 )|0⟩∗ (2.486)

d4 p d4 q
Z
= θ(p0 ) ζ(p)ζ(q) ⟨0|ap a†q |0⟩ei(ωp t2 −px2 ) ⋆ ei(S(ωq )t1 −S(q)x1 ) (2.487)
p4 /κ q4 /κ
Recalling eq. (2.208) and (2.209) we immediately have
⟨0|ap a†q |0⟩ = ⟨0|[ap , a†q ]|0⟩ = ⟨0|[bq , b†p ]|0⟩ = ⟨0|bq b†p |0⟩ (2.488)
so we only need to consider the plane waves. Notice that we are working in different
spacetime points (but with p = q because of the Dirac delta and we are on-shell). We
now use the commutation relation
i j
[x01 , xj2 ] = x . (2.489)
κ 2
We compute the two star products in the bicrossproduct basis for simplicity. We have
(using the time-to-the-right ordering)
k
− κ0
lx) i(k0 x0 +l0 x0 )
êk (x)êl (x) = êk⊕l (x) = ei(kx+e e (2.490)
k0
kx −ik0 x0
êS(k) (x) = e−ie κ e (2.491)
Before proceeding recall that the inverse Weyl map sends (78)

W (êk (x)) = e−i(ωk t−kx) ⇔ W −1 e−i(ωk t−kx) = êk (x) (2.492)

W −1 ei(ωk t−kx) = W −1 e−i(ωk (−t)−k(−x)) = êk (−x) (2.493)
Using this fact we have
ei(ωp t2 −px2 ) ⋆ ei(S(ωp )t1 −S(p)x1 ) 7→ êp (−x1 )êS(p) (−x2 ) (2.494)
and computing the RHS explicitly we have

p p0
− κ0
p)(−x1 )) i(p0 (−x02 )−p0 (−x01 )) 0 0
ei(p(−x2 )+e (−e κ
e = eip(x1 −x2 ) eip0 (x1 −x2 ) (2.495)
and furthermore
e−i(ωp t1 −px1 ) ⋆ e−i(S(ωp )t2 −S(p)x2 ) 7→ êp (x1 )êS(p) (x2 ) (2.496)
and the RHS computed explicitly gives

p p0
− κ0
p)x2 ) i(p0 x01 −p0 x02 ) 0 0
ei(px1 +e (−e κ
e = eip(x1 −x2 ) eip0 (x1 −x2 ) (2.497)
Therefore the two integrands are equal and we have shown that eq. (2.484) holds.
Notice that eq. (2.484) is a particular case of the expression
⟨0|ϕ1 (x1 ) ⋆ ϕ2 (x2 )|0⟩ = ⟨0|ϕ†1 (−x1 ) ⋆ ϕ†2 (−x2 )|0⟩∗ (2.498)
with the choice ϕ1 (x1 ) = ϕ† (x1 ), ϕ2 (x2 ) = ϕ(x2 ). Any other choice leads to zero
in our case, because in each case there would be an operator a or b acting on the
vacuum (recall that [a, b] = [a, b† ] = 0), or terms of the kind ⟨0|a† a† |0⟩ = 0 because
a† a† |0⟩ ̸= |0⟩ and a† a† |0⟩ ⊥ |0⟩. The only other choice that does give a non-zero
contribution is ⟨0|ϕ† (x1 )⋆ϕ(x2 )|0⟩, however in this case the computations go in exactly
T ⇔ 91

CP
2.17. Lorentz?

the same way. Therefore, we have proved that eq. (2.498) holds for any combination
of two scalar fields in the deformed context.
Notice the important observation that, even if we used the commutation relations
(
0, if x1 ̸= x2
0 i
[x1 , x2 ] = i i (2.499)
κ x2 , if x1 = x2
we would have gotten the same result. Indeed, we compute these expectation values
at different points, and if we use the fact that fields at different points behave as if
they were non-deformed, we reach exactly the same result as the canonical case.
2.17.3.2 Lorentz properties of ⟨0|ϕ†1 (x1 ) ⋆ ϕ2 (x2 )|0⟩

Since we are considering products of fields at different spacetime points, in this section
we work in the context of the commutation relations in eq. (2.499). In this way, we get
that the boosts act with a trivial co-product on star products. From the first order
boost transformations in eq. (2.288) and (2.289) we immediately have (up to first
order in the parameter λi )
⟨0|Λϕ† (x1 ) ⋆ Λϕ(x2 )|0⟩

≈ ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩

i ∂ † † i ∂
+ ⟨0| −iλi x1 0 ϕ (x1 ) ⋆ ϕ(x2 )|0⟩ + ⟨0|ϕ (x1 ) ⋆ iλi x2 0 ϕ(x2 ) |0⟩
∂x1 ∂x2
Z 3
d p 3
d q
+ p p ζ(p)ζ(q) ⟨0|bq b†p |0⟩e−i(ωq t1 −qx1 ) ⋆ e−i(S(ωp )t2 −S(p)x2 ) ×
2ωp 2ωq
2
m2

i qi m i pi 5
× ζ(p)iλ − 2 − ζ(q)iλ −
κ ωq2 κ 2 2ωp2
= ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩

∂ ∂
− iλi xi1 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ + iλi xi2 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩
∂x1 ∂x
Z 3
22
d p ip(x1 −x2 ) ip0 (x01 −x02 ) i pi 3m 9
+ e e iλ − (2.500)
2ωp κ 2ωp2 2
were we used the fact that [a, a† ] = [b, b† ] = 1 and the fact that
1 2ωp2 + m2
ζ(p) ≈ 1 − . (2.501)
κ 4ωp
We first consider the terms

∂ ∂
−iλi xi1 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ + iλi xi2 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ (2.502)
∂x1 ∂x2
Using the CPT property (2.484) of the two-point function we can rewrite these two
terms as
∂ ∂
−iλi xi1 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ + iλi xi2 0 ⟨0|ϕ(−x1 ) ⋆ ϕ† (−x2 )|0⟩ (2.503)
∂x1 ∂x2
Using the fields in eq. (2.115) and (2.116) one an now show that (keeping only the
relevant part for the two-point function)
d4 q
Z
i ∂ † † ∂ 0
x 0
ϕ (x) = θ(q0 ) ζ(q) aq −iS(ωq ) e−i(S(ωq )x −S(q)x)
∂x q4 /κ ∂iS(q)i
4
Z
d q ∂ 0
+ θ(q0 ) ζ(q) bq −iωq e−i(ωq x −qx) . (2.504)
q4 /κ ∂iqi
d4 q
Z
∂ † † ∂ 0
xi
ϕ (−x) = θ(q0 ) ζ(q) aq iS(ωq ) ei(S(ωq )x −S(q)x)
∂x0 q4 /κ ∂(−iS(q)i )
4
Z
d q ∂ 0
+ θ(q0 ) ζ(q) bq iωq ei(ωq x −qx) . (2.505)
q4 /κ ∂(−iqi )
Hence we have
∂ † ∂
xi1 0 ϕ (x1 ) = xi2 0 ϕ† (−x2 ) (2.506)
∂x1 ∂x2
and therefore
∂ ∂
xi1 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ = xi2 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ (2.507)
∂x1 ∂x2
Because of this relation, eq. (2.502) is therefore zero
∂ ∂
−iλi xi1 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ + iλi xi2 0 ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ = 0 (2.508)
∂x1 ∂x2
and the boost transformation property for the two point function becomes
⟨0|Λϕ† (x1 ) ⋆ Λϕ(x2 )|0⟩ ≈ ⟨0|ϕ† (x1 ) ⋆ ϕ(x2 )|0⟩ (2.509)

3
Z 2
d p ip(x1 −x2 ) ip0 (x01 −x02 ) pi 3m 9
+ e e iλi 2
− . (2.510)
2ωp κ 2ωp 2
We now treat the second term on the RHS. First of all, recall that we are performing
all the computations at equal times, so that we only need to consider
d3 p ip(x1 −x2 ) pi 3m2 9

Z
e iλi − (2.511)
2ωp κ 2ωp2 2
Calling ∆x = x1 − x2 and assuming without loss of generality that ∆x ∥ z we have
d|p|p2 d cos θ i|p||∆x| cos θ 1 3m2 9

Z
∂
−2π e iλi − (2.512)
∂i∆xi 2ωp κ 2ωp2 2
Performing the angular integral one gets

Z ∞
∂ d|p|p2 1
−i|p||∆x|
1 3m2 9
−2π e −e i|p||∆x|
iλi − (2.513)
∂i∆xi 0 2ωp i|p||∆x| κ 2ωp2 2
which can be equivalently written as

Z ∞
π λi ∂ d|p||p| i|p||∆x| 3m2

e − 9 (2.514)
|∆x| κ ∂i∆xi −∞ 2ωp ωp2
2.18. Phenomenological consequences of deformed CP T transformations 93
This means that we only need to compute the two Fourier transforms
3m2 ∞ |p|
Z
d|p| ei|p||∆x| 2 (2.515)
2 −∞ (m + p2 )3/2
∞
|p|
Z
9
− d|p| ei|p||∆x| (2.516)
2 −∞ (m2 + p2 )1/2
One can easily check that, since m > 0, |∆x| > 0, we have
3m2 ∞ |p|
Z
d|p| ei|p||∆x| 2 = 3im2 |∆x|K0 (m|∆x|) (2.517)
2 −∞ (m + p2 )3/2
∞
|p|
Z
9
− d|p| ei|p||∆x| = −9imK1 (m|∆x|) (2.518)
2 −∞ (m2 + p2 )1/2
where K0 (x), K1 (x) are the modified Bessel functions of the second kind. The addi-
tional term in eq. (2.509) therefore becomes
π λi ∂m|∆x| ∂
3m2 |∆x|K0 (m|∆x|) − 9mK1 (m|∆x|) (2.519)

i
|∆x| κ ∂∆x ∂m|∆x|
Using the fact that

∂ ∂ 1
K0 (x) = −K1 (x) K1 (x) = − (K0 (x) + K2 (x)) (2.520)
∂x ∂x 2
we end up with
m ∆xi 15m

9m
2π λi 2
K0 (m|∆x|) − 3m |∆x|K1 (m|∆x|) + K2 (m|∆x|)) (2.521)
κ |∆x| 2 2
For small values of |∆x| one can show that

15m 9m 9
K0 (m|∆x|) − 3m2 |∆x|K1 (m|∆x|) + K2 (m|∆x|)) ≈ (2.522)
2 2 m|∆x|2
while for large values of |∆x| one has

r
15m 9m 1 π −m|∆x| p
K0 (m|∆x|) − 3m2 |∆x|K1 (m|∆x|) + K2 (m|∆x|)) ≈ −3m 2 e |∆x|.
2 2 2
(2.523)
As a final comment, let us note that if we used the commutation relation in eq. (2.489)
instead of eq. (2.499), then we would have a non-trivial co-product acting on the star
product. However, the additional contributions that one gets can be shown not to
cancel the extra factor in eq. (2.509).
2.18 Phenomenological consequences of deformed CP T trans-

formations
We are now at point xi) of our roadmap in section 2.1. We have already seen the
behaviour of infinitesimal boosts on states in eq. (2.299), from which one can get the
finite (canonical) action given by
|p1 , p2 , p3 ⟩a 7→ | cosh ξ p1 + sinh ξ ωp , p2 , p3 ⟩a (2.524)

|p1 , p2 , p3 ⟩b 7→ | cosh ξ p1 + sinh ξ ωp , p2 , p3 ⟩b . (2.525)
We now consider the case of a particle and an antiparticle originally at rest, and we
put ourselves in their center of mass frame. Notice that we are not considering the two
as being part of a single tensor state, but as two separate single states. As such, we are
ignoring contributions coming from the finite boost of two-particle states described in
section 1.6 of chapter 1. If we boost their initial single-particle states at rest with the
boosts (2.524), (2.525) we have
|p = 0⟩a 7→ |M sinh ξ, 0, 0⟩a

|p = 0⟩b 7→ |M sinh ξ, 0, 0⟩b (2.526)
where
P1 |M sinh ξ, 0, 0⟩a = −S(M sinh ξ)1 |M sinh ξ, 0, 0⟩a (2.527)

P0 |M sinh ξ, 0, 0⟩a = −S(M cosh ξ)|M sinh ξ, 0, 0⟩a (2.528)
P1 |M sinh ξ, 0, 0⟩b = M sinh ξ|M sinh ξ, 0, 0⟩b (2.529)

P0 |M sinh ξ, 0, 0⟩b = M cosh ξ|M sinh ξ, 0, 0⟩b (2.530)
with
κM sinh ξ
−S(M sinh ξ)1 = √ , (2.531)
M cosh ξ + κ2 + M 2
M 2 sinh2 ξ
−S(M cosh ξ) = M cosh ξ − √ (2.532)
M cosh ξ + κ2 + M 2
It is clear that the C operator switches a particle for an antiparticle of a different
momentum due to its non-commutativity with the boost. In order to derive pheno-
menological consequences, it is helpful to first expand at first order in 1/κ eq. (2.531)
and (2.532), obtaining
p2
−S(ωp ) = ωp − + O(1/κ2 ), (2.533)
κ
(p)i ωp
−S(p)i = (p)i − + O(1/κ2 ). (2.534)
κ
The γ factor corresponding to the two boosts will be different because of the presence
of the antipode, and in particular it will be ωp /M for the particle and −S(ωp )/M for
the antiparticle.
We now consider an unstable particle/antiparticle couple. Their evolution is dicta-
ted by a complex energy, where the imaginary part is responsible for the description of
the decay (since it corresponds to an exponential decay of both the amplitude and the
probability density function). In particular, using proper time t > 0 as a parameter,
the amplitudes will be given by

i E
ψpart (t) = A(M, Γ, E) exp −i(E − Γ) t , (2.535)
2 m

i S(E)
ψapart (t) = A[M, Γ, S(E)] exp −i S(E) − Γ t , (2.536)
2 m
where A is a normalization factor. Notice that we used the notation ωp = E − 2i Γ,

where Γ = τ1 describes the reciprocal of the decay time and is called decay width.
Furthermore, we used the fact that
i i
S(ωp ) = S(E) − S(Γ) ≈ S(E) − Γ (2.537)
2 2
i.e. we used the assumption that the decay width is the same for particles and anti-
particles (we are only considering total decay width, so there are no corrections on
considering only some of the total decay channels), so that it is unaffected by the
antipode. This is an a-priori well defined assumption since any correction to Γ (which
is computed in the center of mass frame for both particles and antiparticles) can only
be proportional to m κ or κ , where M is the mass of the decaying particle and m
M
stands for the mass of some decay product. Because we expect κ to be of the order
of magnitude of the Planck energy, such contributions (which are also not enhanced
by boosting since both m and M are the invariant rest masses) are negligible. We
explicitly verify such a-priori estimate by computing the correction to Γ for the decay
of a single particle ϕ of mass M into two particles χ of mass m in the next section
2.18.1.
From the amplitudes, one can get the probability density functions

ΓE E
Ppart (t) = exp −Γ t , (2.538)
m m

ΓS(E) S(E)
Papart (t) = exp −Γ t
m m
p2 p2

E E
= Γ − exp −Γ − t . (2.539)
m κm m κm
For the moment, we will not discuss possible contributions coming from non-zero mass
distribution width coming from loop corrections to the propagator of the decaying
particle. We will return to this issue at the end of the next chapter 3, section 3.5.
In order to measure the effects on decay times due to deformation, the best can-
didate are particle/antiparticle pairs with small mass and high momentum, so that
p2
the quantity M κ is the biggest possible. Natural candidates are therefore µ , µ , for
+ −
which lifetimes are also known with high accuracy. A detailed discussion in this case
can be found in (102). One can also highlight the phenomenological consequences for
the next best candidates (78).
We consider a particle decaying into a particle/antiparticle pair in the center of
mass, which will produce two particles moving back to back (any correction to the
modulus of the spatial momentum due to deformation, and in particular the antipode,
will contribute higher order corrections to eq. (2.538), (2.539), and so it can be igno-
red for the moment). Boosting in the lab frame, and choosing only particles which (in
the center of mass frame) decay orthogonally to the boost direction, we will obtain
particles whose momenta are pointing approximately along the boost direction, with
deviation angle θ of the order of 10−4 rad to 10−6 rad. One can then rotate one mo-
mentum over the other, aligning them. This is equivalent to assuming that any effect
which is not CP T -deformation related is absent. Indeed, for example, possible experi-
mental signatures of anisotropy-induced corrections to CPT and Lorentz symmetries
have been extensively studied, confirming the absence of a preferential direction in
spacetime with better accuracy than we are considering in this section (103), (104),
(105), (106), (107), (108). Notice that, in general, anisotropy-induced corrections to
CPT and Lorentz symmetries require working with angles which are much larger than
the θ we are considering here. With big enough data samples, one could obtain a high
enough accuracy on the average value of θ. Such a control over θ is however outside
current experimental techniques. The candidate decaying particles can be produced in
current particle accelerators (like LHC) or future one (FCC, see (109)), and in Table
2.1 we report several possible decay channels, together with the possible values of
Lorentz boosts, experimental errors, and limits on the value of κ.
2.18.1 Estimation of deformed decay width

In this section we present a simple calculation showing that corrections to the decay
width resulting from κ-deformation are neglectable. For simplicity we consider only
the ϕχ2 interaction. We consider decay of a single ϕ particle of mass M to two χ
particles of the mass m. We have
1
dΓ = dLIPS2 |T |2 (2.540)
2M
where at tree level T = g where g is the coupling constant. Since T is computed
in the rest frame, the corrections might be only of the form m/κ and M/κ. In the
non-deformed theory the Lorentz-invariant phase space factor has the form
d3 k1 d3 k2
dLIPS2 = (2π)4 δ 4 (k1 + k2 − k)
(2π)3 2ωk1 (2π)3 2ωk2
1
= δ 4 (k1 + k2 − k)d3 k1 d3 k2 , (2.541)
4(2π)2 ωk1 ωk2
where k is the incoming momentum, while k1 , k2 are the outgoing ones.

In the deformed case the only difference comes from the fact that instead of the
ordinary delta function we have the one of deformed momentum composition rule
δ 4 (k1 ⊕ k2 ⊕ S(k)) (2.542)
Since here we want to find just the leading order deviation from the non-deformed
theory, we consider only one ordering here; all other orderings will give the same
first-order result.
Recall that to the leading in 1/κ we have
pq p i q0
(p ⊕ q)0 = p0 + q0 + , (p ⊕ q)i = pi + qi +
κ κ
p2 pi p0
S(p)0 = −p0 + , S(p)i = −pi + (2.543)
κ κ
We compute Γ in the reference frame in which the initial particle is at rest so that
k = 0. Then
δ 4 (k1 ⊕ k2 ⊕ S(k))
= δ (ωk1 ⊕ ωk2 ⊕ S(ωk )) δ 3 (k1 ⊕ k2 )

k1 k 2 3 k1 ωk2
= δ ωk1 + ωk2 − ωk + δ k1 + k2 +
κ κ

k1 k2 ωk ωk
= δ ωk1 + ωk2 − ωk + δ 3 k1 + k2 1 − 2 (1 − 3 2 ) (2.544)
κ κ κ
Notice that
ωk k2 ωk2
k1 = −k2 1 − 2 = −k2 + (2.545)
κ κ
and therefore
r s
k2 ωk k22 k22

ωk1 = m2 + k22 − 2 2 2 = ωk2 1−2 ≈ ωk2 1 − (2.546)
κ κωk2 κωk2
which means that

k22

1 1
≈ 1+ (2.547)
ωk1 ωk2 κωk2
Substituting this to (2.541) and integrating we get
g2 d3 k1 d3 k2
Z
k1 k2 3 k1 ωk2
Γ= δ ωk1 + ωk2 − ωk + δ k1 + k2 +
2M 4(2π)2 ωk1 ωk2 κ κ
2 3 2
Z
g k2 d k2 k2 ωk2 k1 k 2
=− 1+ 1−3 δ ωk1 + ωk2 − ωk +
2M 4(2π)2 ωk2 2 κωk2 κ κ
2 Z 2 2 2 2

g k2 dk2 3m + 2k2 2k
=− 2 1− δ 2ωk2 − M − 2 . (2.548)
2M 4πωk2 κωk2 κ
The argument of the delta function is zero for

2
k22 M

k22 +m = 2
+ (2.549)
κ 2
which means that

M2

M
k22 = − m2 1− (2.550)
4 κ
Furthermore, notice that
M2

M
ωk2 2 2
=m + k22 =m + 2
−m 2
1− (2.551)
4 κ
2
2
M M M
= − − m2 (2.552)
4 κ 4

m2

2 M
ωk−1 = 1+ 1−4 (2.553)
2 M 2κ M2
Substituting this result in the above steps we get
g2
Z 2
3m2 + 2k22 2k22

k2 dk2
Γ=− 1 − δ 2ωk2 − M − (2.554)
2M 4πωk2 2 κωk2 κ
g2 1 4 m2
2
M M 2 M
=− 1+ 1−4 2 −m 1− ×
2M 4π M 2 κ M 4 κ
3m2 2 m2

h M
× 1− 1+ 1−4 2
κ M 2κ M
2
2
2 2 M m M M i
− 1+ 1−4 2 −m 2
1− (2.555)
κM 2κ M 4 κ
and keeping only the terms up to leading order in 1/κ and using the fact that
m2
2 2
M 4m4

M M 2 M M 2
1+ 1−4 2 −m 1− ≈ −m +
κ M 4 κ 4 κ M2
(2.556)
we get
g2 k22 dk2 3m2 + 2k22 2k22

Z
Γ≈− 1− δ 2ωk2 − M − (2.557)
2M 4πωk2 2 κωk2 κ
g2 1 4
2
M 4m4 3m2 2 2 2 M2

M 2 2
=− −m + 1− − −m
2M 4π M 2 4 κ M2 κ M κM 4
(2.558)
To simplify things a bit, notice that

2
M 4m4 3m2 2 2 2 M2

M
− m2 + 1 − − − m2
4 κ M2 κ M κM 4
2 2 4 3
M m M 6m M
= − m2 + + − (2.559)
4 2κ Mκ 4κ
so that, calling ΓU the non-deformed Γ we see that
m2 24m4

M
U
Γ=Γ 1+ 2 2+ −1 (2.560)
κ M M4
and we see that the kinematical corrections to the integral (2.548) resulting from the
deformation are of the form m/κ or M/κ . Together with possible corrections to the
coupling constant g, which are of the same order, we conclude that overall corrections
to the decay width Γ are at most m/κ or M/κ, i.e. of order 10−19 , and therefore
completely negligible. Moreover, since even in deformed case the masses of particles
and antiparticles are identical, the corrections to the decay width Γ are the same for
particles and antiparticles.
Tabela 2.1: The following table is taken from (83). It represents the limits on the deformation parameter κ for sets of particle-antiparticle
pairs and energies which may be (or are) produced at LHC and FCC. All the values of decay times with their respective errors, as well
as the particle masses, are taken from (110). The assumed lifetime accuracies were σττ = 10−6 everywhere. The Lorentz boosts γ were
obtained for the assumed energies 6.5 TeV (LHC) and 50 TeV (FCC).
στ p2 p2
Parent Γ τ γ γ κ= κ=
Particle τ [s] M [GeV] M δτ M δτ
resonance M (from PDG) (LHC) (FCC) (LHC) (FCC)
µ± J/ψ, Υ 2.2 × 10−6 0.11 2.8 × 10−18 1 × 10−6 6.1 × 104 4.7 × 105 4 × 1014 2 × 1016
τ± J/ψ, Υ 2.9 × 10−13 1.8 1.3 × 10−12 1.7 × 10−3 3.6 × 103 2.8 × 104 2.5 × 1013 1.5 × 1015
KS , ρ0 , ω 0
π± 0.14
D0 , B 0
2.6 × 10−8 1.8 × 10−16 1.9 × 10−4 4.6 × 104 3.6 × 105 3 × 1014 1.8 × 1016
K± ϕ0 , D 0 , B 0 1.2 × 10−8 0.49 1.1 × 10−12 1.6 × 10−3 1.3 × 104 1.0 × 105 8.5 × 1013 5.1 × 1015
D± ψ, B 0 1.0 × 10−12 1.9 3.4 × 10−13 6.7 × 10−3 3.5 × 103 2.7 × 104 2.2 × 1013 1.3 × 1015
B± Υ 1.6 × 10−15 5.3 0.8 × 10−13 2.4 × 10−3 1.2 × 103 0.9 × 104 0.8 × 1013 0.5 × 1015
2.18. Phenomenological consequences of deformed CP T transformations
99
101
3
κ-deformed propagator, and 1-loop
correction to it
3.1 Introduction
In the previous chapter we studied in details the properties of the complex scalar field,
its charges under continuous transformations and the behaviour of discrete symme-
tries, and the phenomenological consequences of the deformation of CP T transforma-
tions.
In this chapter we will compute the Feynman propagator of our model, as well
as the imaginary part to the 1-loop correction to the propagator. We will begin by
computing the propagator in two ways in section 3.2. We will then give a standard
example of computation of the imaginary part of the 1-loop correction to the propa-
gator in section 3.3. We do this to highlight that each of the tools which is used is
independent from the presence of κ-deformation. As such, we will immediately go to
the deformed case in section 3.4. Because of the peculiar feature of the momentum
sum, we will need to consider 4 different cases, and each of them is treated separately.
Finally, in section 3.5 we discuss possible experimental signatures coming from the
non trivial mass distribution width contribution to the experimental signatures of de-
formed CP T discussed in section 2.18. Everything presented in this chapter (except
for the definitions for the functional derivative in the κ-deformed context, which have
been taken from (85)) is original work.
3.2 Feynman propagator

We now have a well-defined formalism of a complex scalar field in κ-Minkowski spa-
cetime, with a well defined action invariant both under discrete symmetries (C, P, T ),
and under continuous symmetries. We now exploit this model to define the propagator,
and to compute the imaginary part of the 1-loop correction to it.
3.2.1 From the generating functional

The off-shell action (in normal ordering) is given by eq. (2.156). Furthermore, following
(85), we can define the generating functional as
Z
† 1 † i
R † † † †
Z[J, J ] = D[ϕ][ϕ† ]eiSfree [ϕ,ϕ ]+ 2 [ϕ ⋆J+J⋆ϕ +J ⋆ϕ+ϕ⋆J ] (3.1)
Z[0, 0]
102 Rozdział 3. κ-deformed propagator, and 1-loop correction to it
Notice that, differently from (85), we include all the possible orderings for the star
product between the sources J, J † and the fields ϕ, ϕ† .
The first step that we have to do is to rewrite the exponent in the generating
functional in momentum space, because in this way we will be able to complete the
squares and then easily compute the propagator. The action has already been written
in eq. (2.156). Furthermore, in analogy with the fields in eq. (2.137), (2.138), we define
J, J † as follows
d4 p d4 p
Z Z
J(x) = θ(p0 ) ζ(p) Jp e−i(ωp t−px) + θ(p0 ) ζ(p) Jp† e−i(S(ωp )t−S(p)x)
p4 /κ p4 /κ
(3.2)
Z 4
d p
Z 4
d p
J † (x) = θ(p0 ) ζ(p) Jp† e−i(S(ωp )t−S(p)x) + θ(p0 ) ζ(p) Jp e−i(ωp t−px) .
p4 /κ p4 /κ
(3.3)
If we perform the same steps as for the computation of the momentum space action,
everything remains the same except for the absence of the terms (pµ pµ − m2 ) and
(S(p)µ S(p)µ − m2 ) (recall that the mixed terms containing ca and d† b† all go away
because they are multiplied by δ(ω ⊕ ω) or δ(S(ω) ⊕ S(ω)) whose arguments can never
be zero). Therefore, the final result (applying normal ordering) would be
d4 p |p+ |3
Z
† †
ϕ ⋆J +J ⋆ϕ = 2
θ(p0 ) ζ (p) 1 + 3 [a†p Jp + Jp† bp ] (3.4)
p24 /κ2 κ
d4 p |p+ |3
Z
† †
ϕ⋆J +J ⋆ϕ= 2
θ(p0 ) ζ (p) 1 + 3 [Jp† ap + b†p Jp ] (3.5)
p24 /κ2 κ
For simplicity, we now concentrate only on the a, a† part of the above action, the
computations for the b, b† parts are the same. Therefore, we only consider
d4 p |p+ |3
Z
2
θ(p0 ) ζ (p) 1 + 3 [(pµ pµ − m2 )a†p ap + a†p Jp + Jp† ap ] (3.6)
p24 /κ2 κ
and we complete the squares using the fact that

! 2
Jp
q
µ
[(pµ p − m 2
)a†p ap + a†p Jp + Jp† ap ] = pµ pµ − m2 ap + p
pµ pµ − m2
Jp† Jp
− (3.7)
pµ pµ − m2
The generating functional in momentum space can be written as

|p |3

d4 p
iSfree [ϕ̃,ϕ̃† ]+ 2i θ(p0 ) ζ 2 (p) 1+ +3 [a†p Jp +Jp† bp +Jp† ap +b†p Jp ]
Z R
1 p2
Z[J, J † ] = D[ϕ̃]D[ϕ̃† ]e
2 κ
4 /κ
Z[0, 0]
(3.8)
where D[ϕ̃]D[ϕ̃† ] is a shorthand notation for D[ap ]D[a†p ]D[bp∗ ]D[b†p∗ ]. Using the fact
that the measure is invariant under constant shifts like
Jp
ap → ap − (3.9)
pµ pµ − m2
3.2. Feynman propagator 103
and recalling that the same exact computations can be reproduced for the b, b† ope-
rators, the generating functional reduces to
" Z #
d4 p |p+ |3 Jp† Jp

†
Z[J, J ] = exp −i 2
θ(p0 ) ζ (p) 1 + 3 (3.10)
p24 /κ2 κ pµ pµ − m2 + iϵ
where the remaining contribution

|p |3

d4 p
i
θ(p0 ) ζ 2 (p) 1+ +3 (pµ pµ −m2 )[a†p ap +b†p bp ]
Z R
p2
D[ϕ̃]D[ϕ̃† ]e (3.11)
2 2 κ
4 /κ
gets cancelled by the 1

Z[0,0] factor. Notice that there is no prefactor 2 in eq. (3.10)
1
because one has two equal factors coming from the computations of a, a† and b, b† . At
this point, following (85) we can define the the functional derivative as follows
δZ[J, J † ] Z[Jp + εδ(S(p) ⊕ q), Jp† ] − Z[J, J † ]

= lim (3.12)
δJq ε→0 ε
δZ[J, J † ] Z[Jp , Jp† + εδ(S(p) ⊕ S(q))] − Z[J, J † ]
= lim (3.13)
δJq† ε→0 ε
Notice that these expressions reduce to the canonical ones in the limit κ → ∞. Using
these we can compute the Feynman propagator ∆ ˜ κ (p, q) as follows.
! !
δ δ † δ δ
˜
i∆κ (p, q) = i † −i Z[J, J ] † = †
Z[J, J † ] †
δJl δJ q J,J =0 δJl δJ q J,J =0
(3.14)
We have1
( " Z #
4p 3 †
|p |

δ d + J p Jp
Z[J, J † ] = lim exp −i θ(p0 ) ζ 2 (p) 1 + 3 ×
δJq ε→0 p24 /κ2 κ pµ pµ − m2 + iϵ
†
R d4 p 2 |p+ |3 Jp δ(S(p)⊕q)
exp −iε p2 /κ2 θ(p0 ) ζ (p) 1 + κ3 pµ pµ −m2 +iϵ
4
×
ε
Jp† Jp

R d4 p 2 |p+ |3
exp −i p2 /κ2 θ(p0 ) ζ (p) 1 + κ3 pµ pµ −m2 +iϵ )
4
− (3.15)
ε
" #
†
d4 p |p |3
Z
+ Jp Jp
= exp −i θ(p0 ) ζ 2 (p) 1 + 3 ×
( Z † )
d4 p |p |3 ⊕

+ Jp δ(S(p) q)
× −i θ(p0 ) ζ 2 (p) 1 + 3 (3.16)
1
Notice that we are now dealing with canonical expressions of the variables p, q, and we don’t
need to use the star product because there are no mixed functions of spacetime and momentum space.
Furthermore J, J † are not dynamical quantities, and therefore their Poisson brackets are trivial.
" #
d4 p |p+ |3 Jp† Jp
Z
2
= exp −i θ(p0 ) ζ (p) 1 + 3 ×
( )
|q+ |3 Jq†

κ 2
× −i ζ (q) 1 + 3 (3.17)
p4 κ qµ q µ − m2 + iϵ
where we expanded the exponential containing ϵ and in the last passage we used the
fact that (85)2
d4 p
Z
δ(S(p) ⊕ q) = δ(S(q) ⊕ p) δ(S(q) ⊕ p)f (p) = f (q). (3.18)
p4 /κ
Hence (considering only the relevant term which do not go away when J, J † = 0)
!
|q+ |3 εδ(S(q) ⊕ S(p))

δ δ † 1 κ 2
Z[J, J ] = lim −i ζ (q) 1 + 3
δJ † l
δJq †
ε→0 ε
J,J =0
(3.19)
|q+ |3 δ(S(q) ⊕ S(l))

κ 2
= −i ζ (q) 1 + 3 (3.20)
Notice that, if we were to invert the order of the functional derivatives in eq. (3.14) we
will not be able to immediately apply the first Dirac delta, but this does not matter
because at the end of the computations of both derivatives we are left with an integral
containing the product of Dirac deltas δ(S(p)⊕q)δ(S(p)⊕S(l)), which using eq. (3.18)
can be rewritten as δ(S(q) ⊕ p)δ(S(p) ⊕ S(l)), and applying the integration in p to
the first Dirac delta we still get a final δ(S(q) ⊕ S(l)), like in the previous order of the
functional derivatives.
The only remaining fact which is left to clarify is the presence of the additional
factor
|q+ |3

κ 2
ζ (q) 1 + 3 (3.21)
p4 κ
in front of the propagator. We already discussed about this factor in section 2.14.2.3.3,
and for simplicity we will be using the convention expressed in eq. (2.290) from now
onwards.
3.2.2 As inverse of the field operator

We know that the equation of motion for the field in momentum space is given by
the inverse of the field operator, which in momentum space is given by p2 − m2 (any
additional factor goes away because of the convention in eq. (2.290)). Therefore, we
have
1
(p2 − m2 )(−i∆κ ) = 1 =⇒ −i∆κ = (3.22)
p2 − m2 + iϵ
2
In particular ibid. eq. (59),(60), and discussion below eq. (60). Notice that the measure dµ̄(q)
4
in eq. (59) in (85) is the left-invariant measure, which can be equivalently written as pd4 /κ
p
θ(p0 ) (see
eq. (48) and (51) in (85)).
3.3. Example of loop correction to the propagator: non-deformed case 105
3.3 Example of loop correction to the propagator: non-

deformed case
3.3.1 Using renormalization
In the non-deformed context, we consider the following diagram
iΠ(p2 ) =
k−p
2
Assuming a ψϕ2 theory with interaction vertex gψ ϕ2! , the vertex is given by ig (times
a Dirac delta enforcing momentum conservation) and the propagator is given by
i
(3.23)
p2 − m2 + iϵ
We have therefore
d4 k
Z
i i
iΠ(p2 ) = (ig)2 (3.24)
(2π) k − m + iϵ (p − k) − m2 + iϵ
4 2 2 2
| {z }| {z }
B A
We now use the so called Feynman trick, which says that

 
n
Z 1Y n
1 X (n − 1)!
= dxi δ  xj − 1 Pn (3.25)
A1 . . . An 0 ( l=1 xl Al )n
i=1 j=1
and we have
A + (B − A)x = [k − p(1 − x)]2 + p2 x(1 − x) − m2 +iϵ (3.26)

| {z }
−∆(x)
Notice that ∆(x) ≥ 0. Thus, after using the Feynman parameters, we have
1
(ig)2
Z Z
1
2
iΠ(p ) = − 4
d k dx (3.27)
(2π)4 0 {k 2 − ∆(x) + iϵ}2
where we performed the constant linear shift k 7→ k + p(1 − x) which leave the metric
invariant. The poles can be represented as follows in the plane below. We can therefore
use the integration contour in red in the same picture.
The integral computed in the red circuit Γ is zero, and if R → ∞ the contribution from
the circular sections goes to zero, meaning that we can use the change of variables
k 0 7→ ikE
0
(3.28)
which is called Wick rotation. In this way one has
k 2 = (k 0 )2 − k2 7→ −kE
2 0 2
= −(kE ) − k2 . (3.29)
The amplitude reduces to

1
d4 k
Z Z
1
2
iΠ(p ) = −i(ig) 2
dx 4 2 (3.30)
0 (2π) (kE + ∆)2
where there is no additional iϵ because there are no more singularities in the integrand.
The momentum space integral is a known integral since
dd k (k 2 )a Γ(b − a − 12 d)Γ(a + 12 d) −(b−a−d/2) 4−d

Z
µ 4−d
= ∆ µ (3.31)
(2π)4 (k 2 + ∆)b (4π)d/2 Γ(b)Γ( 12 d)
Notice that [µ] = 1, and that the term µ4−d has been added so that the integral still
has the same mass dimension as before dimensional regularization (i.e. 4), regardless of
or choice of d. In our case, using dimensional regularization, we can assume d = 4 − ϵ,
and in our case we have a = 0 and b = 2. Therefore the momentum space integral is
Γ( 2ϵ ) ϵ ϵ
2
∆− 2 (µ2 ) 2 (3.32)
(4π) Γ(2)
One can use the expansions

ϵ ϵ ϵ ∆
∆− 2 (µ2 ) 2 = 1 − log 2 + O(ϵ2 ) (3.33)
2 µ
Furthermore, we have
π2 z2 γπ 2

1 z
Γ(z) = − γ + 2
γ + − 3
γ + + 2ζ(3) + . . . (3.34)
z 2 6 6 2
Therefore we have
π2 ϵ2 γπ 2

ϵ −ϵ 2 ϵ 2 ϵ 2 3 ϵ ∆
Γ( )∆ 2 (µ ) 2 = −γ+ γ + − γ + + 2ζ(3) (1 − log 2 )
2 ϵ 4 6 24 2 2 µ
(3.35)
2 ∆
≈ − log 2 − γ + O(ϵ) (3.36)
ϵ µ
Putting everything together we end up with
Z 1
(ig)2

2 ∆
2
Π(p ) = − dx − log 2 − γ + O(ϵ) (3.37)
(4π)2 Γ(2) 0 ϵ µ
2 Z 1
(ig) 2 ∆
=− dx − log 2 − γ + O(ϵ) (3.38)
(4π)2 0 ϵ µ
To renormalize this amplitude we add the counterterms Ak 2 + Bm2 to the amplitude,

we then chose A, B in such a way that the divergence is eliminated leaving behind
3.3. Example of loop correction to the propagator: non-deformed case 107
only a finite contribution, and finally we fix these finite contributions by imposing the
canonical restrictions Π(m2 ) = Π′ (m2 ) = 0 where Π′ = dp dΠ
2 . We have
1
(ig)2
Z
2 ∆
Π(p2 ) = − dx − log 2 − γ + O(ϵ) + Ap2 + Bm2 (3.39)
(4π)2 0 ϵ µ
(ig)2 2
so that choosing A = κA , B = κB + (4π)2 m2 ϵ
we have
1
(ig)2
Z
∆
2
Π(p ) = dx log + κA p2 + κB m2 (3.40)
(4π)2 0 µ2
At this point, notice that the logarithm may have an imaginary part. It is sufficient to
notice that the argument of the logarithm is positive unless p2 x(1 − x) > m2 (recall
that ∆ = m2 − p2 x(1 − x)), in which case the numerator becomes negative, and
therefore the whole argument of the logarithm becomes negative. Notice that this is
indeed possible since p refers to the momentum of the external particle, meaning that
p2 = m2ψ > m2 where m = mϕ is the mass appearing in the propagators in the loop.
q
2
Therefore, as long as |x| < 12 ± 21 1 − 4 m p2
and p2 > 4m2 , the amplitude gains a
small imaginary part. However, notice that this imaginary part cannot be cancelled
(or in general dealt with) by the counterterms, since they are real parameters coming
from a Hermitian Lagrangian. Therefore, we can at best deal with the real part of
the amplitude. In other words, instead of simply imposing Π(p2 = m2ψ ) = 0, we are
actually imposing ℜ(Π(p2 = m2ψ )) = 0, and the same for Π′ .
The renormalized amplitude reads
(ig)2 1 (ig)2
Z
∆ 2π
2
Π(p ) = dx log 0 + 2 √ − 1 (p2 − m2 ) (3.41)
(4π)2 0 |∆ | m (4π)2 3 3
Notice that this final expression is dimensionally correct.

We can now deal with the imaginary part of the amplitude. We can obtaining it
by simply assuming that the argument of the logarithm becomes negative.
In this case ℑ log |∆∆0 | = −π, where the minus arises from the fact that the quantity
∆ − iϵ (which is found in the denominator (k 2 − (∆(x) − iϵ))2 after the Feynman
trick) has a negative imaginary part. Therefore, recalling that the only part of the x
integration which contributes is the one such that
s
1 1 m2
|x| < ± 1−4 2 (3.42)
2 2 p
we have
s
g2π g2 m2
Z
ℑΠ(p2 ) = dx = 1−4 . (3.43)
(4π)2 16π p2
Notice the very important fact that the only real role of the imposition of the con-
ditions Π(m2 ) = Π′ (m2 ) = 0 (for what concerns the imaginary part contribution)
is to substitute µ2 ↔ |∆0 | inside the argument of the logarithm, but both µ2 and
|∆0 | are strictly positive, which means that they are irrelevant in the determination
of the imaginary part of the logarithm. In other words, one can get the imaginary
contribution of the 1-loop correction directly from eq. (3.37).
Furthermore, the reason we showed in such detail the non-deformed computation
was to show that all the tools that are conventionally employed in non-deformed QFT
can be used as well in the deformed context. In other words, we have all the tools to
approach the same calculations in the κ-deformed context.
3.4 Example of loop correction to the propagator: κ- de-

formed case
We can consider the case of one incoming particle (with momentum S(p)) and two
outgoing ones (let us call them p2 and p3 ). The conservation of momentum then
implies that
S(p) ⊕ p2 ⊕ p3 = 0 (3.44)
This identity is immediately satisfied if p2 ⊕p3 = p, and we can use one of the following
parametrizations for p2 and p3
p2 = k p3 = S(k) ⊕ p (3.45)
p2 = p ⊕ S(k) p3 = k (3.46)
p2 = S(k) p3 = k ⊕ p (3.47)
p2 = p ⊕ k p3 = S(k) (3.48)
We can now perform the computation as in the non-deformed case, keeping in mind
to add all the above contributions.
3.4.1 Dimensional regularization and cut-off in the κ-deformed con-

text
We start from the same amplitude as before, namely
iΠ(p2 ) =
S(k) ⊕ p
d4 k
Z
i i
2
iΠ(p ) = (ig) 2
(3.49)
(2π) k − m + iϵ (S(k) ⊕ p) − m2 + iϵ
4 2 2 2
Notice that, apart from the deformed conservation of momentum, we are assuming
that each vertex only contributes ig.
From the relations
1 Q0 P⃗ · Q
⃗
(P ⊕ Q)0 = P0 Q+ + κ + (3.50)
κ P+ P+
1
(P ⊕ Q)i = Pi Q+ + Qi (3.51)
κ
3.4. Example of loop correction to the propagator: κ- deformed case 109
1 Q0 P⃗ · Q
⃗
(P ⊕ Q)4 = P4 Q+ − κ − (3.52)
κ P+ P+
q
P+ = P0 +P4 = P0 + κ2 + P02 − P⃗ 2 (3.53)
we have
1 p0 S(k)p
(S(k) ⊕ p)0 = S(k)0 p+ + κ + (3.54)
κ S(k)+ S(k)+
2

p+ k p0 k+ κkp k+
= −k0 + + − (3.55)
κ k+ κ k+ κ2
k0 p+ p+ k2 p0 k+ kp
=− + + − (3.56)
κ κk+ κ κ
1
(S(k) ⊕ p)i = S(k)i p+ + pi (3.57)
κ
ki
= − p+ + pi (3.58)
k+
from which we get

2
k0 p+ p+ k2 p0 k+ kp

(S(k) ⊕ p)20 = − + + − (3.59)
κ κk+ κ κ
2
p+ k p0 2
p+ k (kp) p0 k+ (kp)
+2 −2 −2 (3.60)
κ2 κ2 k+ κ2
k2 p2+ (kp)p+
(S(k) ⊕ p)2 = 2 + p2 − 2 (3.61)
k+ k+
It is useful to obtain the following approximated relations, valid up to second order

in κ1 .
pq 1
+ 2 p20 q0 + p0 q02 + q0 p2 − p0 q2 − 2p0 pq (3.62)

(p ⊕ q)0 ≈ p0 + q0 +
κ 2κ
pi q0 pi
+ 2 q02 − q2 (3.63)

(p ⊕ q)i = pi + qi +
κ 2κ
pq p2 1
+ 2 p20 q0 − p0 q02 + q0 p2 + p0 q2 (3.64)

(S(p) ⊕ q)0 ≈ −p0 + q0 − +
κ κ 2κ
pi q0 pi p0 pi
− 2 p20 − 2p0 q0 + q02 + p2 − q2 (3.65)

(S(p) ⊕ q)i = −pi + qi − +
κ κ 2κ
pq q2
(p ⊕ S(q))0 ≈ p0 − q0 − +
κ κ
1
−p20 q0 + p0 q02 − q0 p2 + 2p0 pq + 2q0 pq − p0 q2 (3.66)

+ 2
2κ
pi q0 qi q0 1
+ 2 q02 pi − q02 qi + q2 pi − q2 qi (3.67)

(p ⊕ S(q))i = pi − qi − +
κ κ 2κ
Notice that we are going up to second order because, as will be shown below, the
first order contribution to the propagator identically vanish, so that we are reduced
to the non-deformed case. Since we want to study the effects of deformation, we need
therefore to consider terms up to second order.
We first show that the first order contribution in κ1 to (S(k)⊕p)i2 −m2 +iϵ vanish.
(S(k) ⊕ q)2 = (S(k) ⊕ q)20 − (S(k) ⊕ q)2 (3.68)

2
kq k2 ki q0 ki k0 2

≈ −k0 + q0 − + − −ki + qi − + (3.69)
κ κ κ κ
2

kq k
≈ (q − k)2 + 2(−k0 + q0 ) − +
κ κ

ki q0 ki k0
− 2(−ki + qi ) − + (3.70)
κ κ
kq k2 kq k2
= (q − k)2 + 2k0 − 2k0 − 2q0 + 2q0
κ κ κ κ
kq k2 kq k2
− 2k0 + 2k0 + 2q0 − 2q0 (3.71)
κ κ κ κ
= (q − k)2 . (3.72)
We need therefore to go to second order contributions. We have (once again using

Mathematica) using eq. (3.64) and (3.65)
(S(k) ⊕ q)2 = (S(k) ⊕ q)20 − (S(k) ⊕ q)2 (3.73)
1
≈ (q − k)2 + − k03 q0 + (−k2 + kq)(q02 + kq − q2 )
κ2
+ k02 (2q02 − 2k2 + kq − q2 ) + k0 q0 (−q02 + 3k2 − 2kq + q2 ) (3.74)
∆2 [(S(k) ⊕ q)2 ]
:= (q − k)2 + (3.75)
κ2
where we introduce the short-hand notation
1 d2

2
∆2 [(S(k) ⊕ p) ] := 2 2
[(S(k) ⊕ p) − (p − k) ] (3.76)
2 d(1/κ)2 1/κ=0

3 2 2 2
= − k0 p0 + (−k + kp)(p0 + kp − p )

+ k02 (2p20 − 2k2 + kp − p2 ) + k0 p0 (−p20 + 3k2 − 2kp + p2 ) (3.77)
where the index 2 in ∆2 means that it is the coefficient of the second order of the
expansion of the quantity inside square brackets. Therefore, we have
i i
≈ ∆2 [(S(k)⊕p)2 ]
(3.78)
(S(k) ⊕ p)2 − m2 + iϵ (p − k)2 + − m2 + iϵ
κ2
i 1
= ∆2 [(S(k)⊕p)2 ]
(3.79)
1+ 1 (p − k)2 − m2 + iϵ
κ2 (p−k)2 −m2 +iϵ
i 1 i∆2 [(S(k) ⊕ p)2 ]

= − (3.80)
(p − k)2 − m2 + iϵ κ2 [(p − k)2 − m2 + iϵ]2
The leading contributions in the 1/κ expansion of the amplitude in eq. (3.49) is given
by
d4 k
Z
i i
2
iΠ(p ) = (ig)2
(3.81)
(2π)4 k 2 − m2 + iϵ (p − k)2 − m2 + iϵ
(ig)2 d4 k i∆2 [(S(k) ⊕ p)2 ]
Z
i
− 2 (3.82)
κ (2π)4 k 2 − m2 + iϵ [(p − k)2 − m2 + iϵ]2
g2 d4 k ∆2 [(S(k) ⊕ p)2 ]
Z
1
= iΠU (p2 ) − . (3.83)
κ2 (2π)4 k 2 − m2 + iϵ [(p − k)2 − m2 + iϵ]2
Considering only the integral, after treating the denominator in the same as the non-
deformed case using eq. (3.25) we are reduced to
g2 1
d4 k ∆shif t
[(S(k) ⊕ p)2 ]
Z Z
− dx 2
(3.84)
κ2 0 (2π)4 [k 2 − ∆(x) + iϵ]3
where we have used the generalized Feynman trick
Γ(m + n) ∞ λm−1 dλ
Z
1
= (3.85)
Am B n Γ(m)Γ(n) 0 [λA + B]n+m
which can be rewritten in our particular case as

Z 1 Z 1
1 2δ(x + y − 1) 2
= dx dy = dy
AB 2 0 [xA + yB] 3
0 [A + (B − A)y]3
Z 1
2
= dx (3.86)
0 [A + (B − A)x]3
and therefore
Z 1
1 1 2
= dx (3.87)
k − m + iϵ [(p − k) − m2 + iϵ]2
2 2 2
0 {[k − p(1 − x)]2 − ∆(x) + iϵ}3
−∆(x) = p2 x(1 − x) − m2 . (3.88)
Since the poles are in the same position as before, we can now do a Wick rotation
obtaining
g2 1
d4 k ∆shif t+W ick
[(S(k) ⊕ p)2 ]
Z Z
i 2 dx 2
(3.89)
κ 0 (2π)4 [k 2 + ∆(x)]3
where the iϵ has gone away and where, as we computed before, we have
∆shif
2
+W ick
[(S(k) ⊕ p)2 ]
= ik03 p0 + k02 −3p20 x + p20 + 2k2 − 4kpx + 3kp + 2p2 x2 − 3p2 x + 2p2

+ k0 − 3ip30 x2 + 2ip30 x + 4ip0 k2 x − ip0 k2 − 8ip0 kpx2 + 8ip0 kpx

− 2ip0 kp + 4ip0 p2 x3 − 7ip0 p2 x2 + 6ip0 p2 x − 2ip0 p2

+ p40 x3 − p40 x2 − 2p20 k2 x2 + p20 k2 x + 4p20 kpx3 − 5p20 kpx2
+ 2p20 kpx − 2p20 p2 x4 + 4p20 p2 x3 − 4p20 p2 x2 + 2p20 p2 x − k3 p

+ k2 p2 x + (kp)2 (2x − 1) − 3kp3 x2 + 2kp3 x + p4 x3 − p4 x2 (3.90)
Terms proportional to k0 or k elevated to an odd power do not contribute to the final

integral. The numerator can therefore be simplified as
∆shif +W ick
[(S(k) ⊕ p)2 ] = k02 −3p20 x + p20 + 2k2 + 2p2 x2 − 3p2 x + 2p2

2

+ p40 x3 − p40 x2 − 2p20 k2 x2 + p20 k2 x − 2p20 p2 x4
+ 4p20 p2 x3 − 4p20 p2 x2 + 2p20 p2 x

+ k2 p2 x + (kp)2 (2x − 1) + p4 x3 − p4 x2 (3.91)
Since we are integrating in the Euclidean space, each time we encounter k2 alone we
can substitute it with 3k02 . Furthermore, since the terms linear in k do not contribute,
the mixed terms in (kp)2 do not contribute, so that we also have
X X
(kp)2 = ki pi kj pj 7→ k2i (pi )2 7→ k02 p2 (3.92)
i,j i
Therefore, the above numerator can be simplified to
∆shif +W ick
[(S(k) ⊕ p)2 ] = k02 p20 1 − 6x2 + p2 (2x(x + 1) + 1) + 2k02 k2 (3.93)

2
+ (x − 1)x p40 x − 2p20 p2 ((x − 1)x + 1) + p4 x

Notice that there is one obvious exception to the substitution done above, and this
exception is obtained once we have terms like k02 k2 , which explain the remaining k2
in the expression above.
We now have three types of integral:
1)
d4 k k02
Z
(3.94)
(2π)4 (k 2 + ∆)3
This integral can be computed with the help of eq. (3.31)Considering the case
a = 1, the above integral reduces to
d
dd k ki2 dd k ki2
X Z Z
4−d
µ =d
(2π)4 (k 2 + ∆)b (2π)4 (k 2 + ∆)b
i=1
Γ(b − 1 − 21 d)Γ(1 + 12 d) −(b−1−d/2) 4−d
= ∆ µ
(4π)d/2 Γ(b)Γ( 12 d)
(3.95)
which means that
dd k k02 1 Γ(b − 1 − 12 d)Γ(1 + 12 d) −(b−1−d/2) 4−d

Z
µ 4−d
= ∆ µ ; (3.96)
(2π)4 (k 2 + ∆)b d (4π)d/2 Γ(b)Γ( 21 d)
2)
d4 k
Z
1
(3.97)
(2π) (k + ∆)3
4 2
This integral is just an example of eq. (3.31) with a = 0. We have therefore
dd k Γ(b − 12 d)Γ( 12 d) −(b−d/2) 4−d

Z
1
µ 4−d
= ∆ µ ; (3.98)
(2π)4 (k 2 + ∆)b (4π)d/2 Γ(b)Γ( 21 d)
3)
d4 k ki2 kj2
Z
i ̸= j (3.99)
(2π)4 (k 2 + m2 )3
To compute this in dimensional regularization we cannot simply use the relation

(3.31), but we need to compute it from scratch. We use d dimensions and we
go in spherical coordinates. Using the relations ki = |k| cos(αi ) (where αi is the
angle which the vector k forms with respect to the i-th axis) we have
dd k ki2 kj2 ∞
k d−1 k 4
Z Z Z
1
d 2 b
= 2 2
dΩ cos αi cos αj dk (3.100)
(2π) (k + ∆) (2π)d 0 (k 2 + ∆)b
We can now can just define the constant

Z
Θ(i, 2, j, 2) := dΩ cos2 αi cos2 αj (3.101)
which is a strictly positive quantity bounded from above by the surface Sd of a

d-dimensional sphere. The remaining integral can be evaluated with the help of
the beta functions
Z ∞
Γ(α)Γ(γ)
B(α, γ) = = dy y α−1 (1 + y)−α−γ (3.102)
Γ(α + γ) 0
In our case we have

Z ∞ Z ∞ 2 −b
−b −b d+3 k
d+3 2
dk k (k + ∆) = ∆ dk k +1 (3.103)
0 ∆
Z0 ∞
1 1 −1
= ∆−b ∆ 2 y 2 dy ( ∆y)d+3 (y + 1)−b (3.104)
p
0 2
1 −b+ 1 + d + 3 ∞
Z
d d d
= ∆ 2 2 2 dy y 2 +2−1 (1 + y) 2 +2−b−( 2 +2)
2 0
(3.105)
Γ(2 + 12 d)Γ(b − 2 − 12 d) −b+2+ d
= ∆ 2 (3.106)
2Γ(b)
where we used the change of coordinates
k2 k ∆ 1 1 1
y= =⇒ dy = 2 dk =⇒ dk = √ dy = ∆ 2 y − 2 dy (3.107)
∆ ∆ 2 y∆ 2
which corresponds to the result we would have gotten from eq. (3.31) but with
Θ(i,j)
the factors (4π)d/21 Γ( 1 d) exchanged with the factor 2(2π)d . Therefore we have
2
dd k ki2 kj2 Θ(i, 2, j, 2) Γ(2 + 21 d)Γ(b − 2 − 12 d) −b+2+ d 4−d

Z
4−d
µ = ∆ 2µ
(2π)d (k 2 + ∆)b (2π)d 2Γ(b)
(3.108)
where at the end we wrote once again the term µd−4 for dimensional reasons.
We can now explicitly perform the integral. We have to compute

(Z
1
g2 d4 k k02
Z
dx p20 1 − 6x2 + p2 (2x(x + 1) + 1) (3.109)

i 2
κ 0 | {z } (2π)4 [k 2 + ∆(x)]3
:=F (p,x)
1
d4 k k02 k2
Z Z
+2 dx (3.110)
0 (2π)4 [k 2 + ∆(x)]3
)
1
d4 k
Z Z
1
p40 x 2p20 p2 ((x 4

+ dx (x − 1)x − − 1)x + 1) + p x
0 | {z } (2π) [k + ∆(x)]3
4 2
:=G(p,x)
(3.111)
Using the equations obtained above, and using the canonical choice d = 4 − ϵ which
is used in dimensional regularization, we have
d4 k k02 1 Γ(b − 1 − 12 d)Γ(1 + 12 d) −(b−1−d/2) 4−d

Z
= ∆ µ
(2π)4 [k 2 + ∆(x)]3 d (4π)d/2 Γ(b)Γ( 12 d)
b=3
1 Γ(3 − 1 − 2 + 2ϵ )Γ(1 + 2) −(3−1−2+ ϵ ) ϵ
= ∆ 2 µ
4 (4π)2 Γ(3)Γ(2)
− ϵ
Γ( 2ϵ ) ∆ 2
= (3.112)
4(4π)2 µ2
d4 k k02 k2 Θ(0, 2, j, 2) Γ(2 + 12 d)Γ(b − 2 − 21 d) −b+2+ d 4−d

Z
= 3 ∆ 2µ
(2π)4 [k 2 + ∆(x)]3 (2π)d 2Γ(b)
b=3
ϵ − ϵ
Θ(0, 2, j, 2) Γ(4)Γ(3 − 2 − 2 + 2 ) −3+2+2 ∆ 2
=3 ∆
(2π)4 2Γ(3) µ2
− ϵ
Θ(0, 2, j, 2) ϵ ∆ 2
=9 4
∆Γ(−1 + ) 2
(3.113)
2(2π) 2 µ
d4 k Γ(b − 21 d)Γ( 12 d) −(b−d/2) 4−d

Z
1
= ∆ µ
(2π)4 [k 2 + ∆(x)]3 (4π)d/2 Γ(b)Γ( 12 d)
b=3
Γ(1)
= ∆−1
(4π)2 Γ(3)
1
= (3.114)
2(4π)2 ∆
We now need the following relations

n
!
(−1)n 1 X 1
Γ(−n + ϵ) = −γ+ + O(ϵ) (3.115)
n! ϵ k
k=1
ϵ ϵ ϵ 2 ∆
(3.35) → Γ( )∆− 2 (µ2 ) 2 ≈ − log 2 − γ + O(ϵ) (3.116)
2 ϵ µ
In this way, the amplitude becomes
( Z 1
g2

1 2 ∆
U
Π(p) = Π (p) + 2 dxF (p, x) − log − γ (3.117)
κ 4(4π)2 0 ϵ µ2
Θ(0, 2, j, 2) 1
Z
2 ϵ ∆
−9 dx∆ −γ+1 1 − log 2 (3.118)
(2π)4 0 ϵ 2 µ
Z 1 )
1 1
+ dxG(p, x) (3.119)
2(4π)2 0 ∆
Once again, notice that the amplitude contains a real and an imaginary part, because
the ∆ is exactly the same as the one in the non-deformed case. As such, all the
renormalization procedure is intended to be valid only for the real part ℜ(Π) of it.
Like in the non-deformed case, the only contribution comes when the argument of
the logarithm becomes negative (notice, after imposing the renormalization conditions,
the µ2 in the argument of the logarithms will be replaced by ∆0 as previously defined,
which is always positive for x ∈ [0, 1]), and therefore ℑ log ∆∆0 = −π. Hence, recalling
that the integration domain on the x for the imaginary part is
s
1 1 m2
|x| < ± 1−4 2 (3.120)
2 2 p
the imaginary part of the integral in the amplitude is given by

( Z L2 )
g2 Θ(0, 2, j, 2) L2
Z
π
dxF (p, x) − 9π dx∆ (3.121)
κ2 4(4π)2 L1 (2π)4 L1
where we called
s s
1 1 m2 1 1 m2
L1 = − 1−4 2 L2 = + 1−4 . (3.122)
2 2 p 2 2 p2
Computing everything explicitly, and recalling that F (p, x) and G(p, x) are defined
respectively in eq. (3.109) and (3.111), we get (recall that we are using the metric
convention + − −−)
Z L2 Z L2
p20 1 − 6x2 + p2 (2x(x + 1) + 1) (3.123)

F (p, x)dx =
L1 L1
s
m2 m2

8 2 5 2 2 2
= 1−4 2 p − p0 − m − 4 2 p20 (3.124)
p 3 3 3 p
Z L2 Z L2
∆dx = (m2 − p2 x(1 − x))dx (3.125)
L1 L1
s
m2

2 2 1 2
= 1−4 2 − m + p (3.126)
p 3 6
(notice that we switched p2 = p20 − p2 ). To get a fully numerical computation, we need

to compute
Z
I = Θ(0, 2, j, 2) = cos2 α0 cos2 αj dΩ (3.127)
The cosines of αj are called directional cosines. We have
kj
cos αj = p 2 (3.128)
k0 + k1 + k22 + k32
2
so that by definition we have

3
X
cos2 αj = 1 (3.129)
i=0
Since j is fixed, we can assume without loss

P of generality that j = 1. Using the above
relation we can substitute cos2 α0 by 1 − 3i=1 cos2 αi so that
Z Z
I = cos2 α1 dΩ − cos4 α1 dΩ − 2I (3.130)
| {z }
I2
where we used the fact that the integral of the product of any two different directional
cosines does not depend on which pair we choose, because Euclidean space is isotropic.
In the same way, we can write
Z Z
2
I2 = cos α1 dΩ = dΩ − 3I2 = S4 − 3I2 (3.131)
so that
S4
I2 = (3.132)
4
where S4 is the surface of a n = 4 dimensional sphere. We are therefore reduced to
Z
S4 1
I= − cos4 α1 dΩ (3.133)
12 3
Now we only have one integral left, and this we can compute directly. In fact, the
sperical surface element in n dimensions in Euclidean space and in hyperspherical
coordinates is given by (we are considering a fixed radius r = 1, since we already
integrated over the radius in the formula where Θ was introduced in the first place)
dΩ = sinn−2 (ϕ1 ) sinn−3 (ϕ2 ) . . . sin(ϕn−2 )dϕ1 . . . dϕn−1 (3.134)

where ϕ1 , . . . , ϕn−2 ∈ [0, π] and ϕn−1 ∈ [0, 2π]. In our case n = 4 and therefore we
have
dΩ = sin2 (ω) sin(θ)dωdθdϕ (3.135)
Such an explicit choice of coordinate brakes the symmetry of Euclidean space, since
we are now choosing a preferred axis. In particular, the directional cosine should be
taken with respect to this preferred axis, in the same way in which in three dimension
the only directional cosine is the one taken with respect to the preferential axis, which
in three dimension is the z axis. Therefore we have
π2
Z
cos4 (ω) sin2 (ω) sin(θ)dωdθdϕ = (3.136)
4
Furthermore, recalling that
d
2π 2
Sd = =⇒ S4 = 2π 2 (3.137)
Γ( d2 )
we have
π2 π2 π2
I= − = (3.138)
6 12 12
Therefore, the imaginary part of the integral becomes
s ( )
g2 m2 2 m2

5 2 1 1 1
1 − 4 2 p0 − − 2 + p2 − + m2 − + (3.139)
16πκ2 p 12 p 3 8 6 2
s ( )
g2 m2 2 m2

5 13 1
= 1 − 4 p0 − − + p2 + m2 (3.140)
16πκ2 p2 12 p2 24 3
q
g2 2
Recalling from eq. (3.43) that ℑΠU (p2 ) = 16π 1 − 4m
p2
we can write
m2

1 5 13 2 1 2
ℑΠ (1) 2 U 2 2
(p ) = ℑΠ (p ) 1 + 2 p0 − − 2 + p + m (3.141)
κ 12 p 24 3
where the index (1) just means that this is the amplitude corresponding to the first
possibility of the choice of momenta.
We could perform the same integrals using a hard cutoff instead of dimensional
regularization. In fact, from a principled point of view, we are expanding in powers
of 1/κ, which means that momenta should have a momentum smaller than κ. This
in turn implies that we should not be able to integrate up to infinity in momentum
space when integrating virtual momenta. Of course, the finite part of any integrand
does not depend on the choice of regularization, but limiting ourself to a hard cutoff
makes the approximations used in the expansion in powers of 1/κ explicit at all steps.
In particular, if we assume that our momenta are organized according to the following
relation
p, k ≪ Λ ≪ κ (3.142)
where Λ is the cutoff scale, then our expansion in powers of 1/κ is fully justified.
Therefore, we compute the same integral using a cutoff in momentum space. We

have what follows:
1)
d4 k k02
Z
(3.143)
(2π)4 (k 2 + ∆)3
If we go to spherical coordinates this integral becomes (using Mathematica)

Λ
k5
Z
1 S4 1 h1
dk = (−3 − 2 log ∆)
4 (2π)4 0 (k 2 + ∆)3 32π 2 4
1 ∆(3∆ + 4Λ2 )
i
+ 2
+ 2 log(∆ + Λ ) (3.144)
4 (∆ + Λ2 )2
2)
d4 k
Z
1
(3.145)
(2π) (k + ∆)3
4 2
In this case going to spherical coordinates we have (again with Mathematica)

Λ
k3 Λ4
Z
S4 1
dk = (3.146)
(2π)4 0 (k 2 + ∆)3 8π 2 4∆(∆ + Λ2 )2
3)
d4 k ki2 kj2
Z
i ̸= j (3.147)
(2π)4 (k 2 + m2 )3
To compute this with a cutoff we must proceed as before, and we go in spherical

coordinates. Using the relations ki = |k| cos(αi ) (where αi is the angle which
the vector k forms with respect to the i-th axis) we have
d4 k ki2 kj2 Λ
k7
Z Z Z
1
4 2 3
= 2 2
dΩ cos αi cos αj dk (3.148)
(2π) (k + ∆) (2π)4 0 (k 2 + ∆)3
We know from eq. (3.138) that the angular integral corresponds to Θ(0, 2, j, 2) =
π2
12 , and the radial integral can be computed explicitly obtaining
Λ
k7
Z
1
dk 2 3
= ∆(5 + 6 log ∆)
0 (k + ∆) 4
−5∆3 − 4∆2 Λ2 + 4∆Λ4 + 2Λ6 − 6∆(∆ + Λ2 )2 log(∆ + Λ2 )
+
4(∆ + Λ2 )2
(3.149)
As in the dimensional regularization case, the divergencies can be absorbed in the

counterterms, but here we are only interested in the parts which can contribute an
imaginary part to the integral, which are once again the logarithms. Notice here
that the logarithms of the form log(∆ + Λ2 ) cannot contribute any immaginary part
because, although ∆ can be negative, the sum ∆ + Λ2 cannot (Λ2 is positive and
very large). Therefore, limiting ourselves to the logarithms, the contribution from the
above integrals become respectively
d4 k k02
Z
1
4 2 3
→− log ∆ (3.150)
(2π) (k + ∆) 4(4π)2
d4 k
Z
1
→0 (3.151)
(2π)4 (k 2 + ∆)3
d4 k k02 k2 π2 1
Z
2 → 9∆ log ∆ (3.152)
(2π)4 (k 2 + m2 )3 12 (2π)4
We know from the previous equations, and in particular from eq. (3.121) and (3.138),
that these are the same coefficients as before. We obtain therefore the same imaginary
part, as was expected.
We now pass to the second possibility, namely
iΠ(p2 ) =
p ⊕ S(k)
d4 k
Z
i i
2
iΠ(p ) = (ig) 2
(3.153)
(2π)4 k 2 − m2 + iϵ (p ⊕ S(k))2 − m2 + iϵ
Every consideration goes exactly as before, but this time we need to compute
∆shif
1
+W ick
[(S(k) ⊕ p)2 ], ∆2shif +W ick [(S(k) ⊕ p)2 ] (3.154)
In fact, contrary to the previous case, this time there is a 1/κ contribution to the
expansion of the denominator. The explicit expression is given by
iΠ(p2 ) = iΠU (p2 ) (3.155)

g2 d4 k ∆1 [(p ⊕ S(k))2 ]
Z
1
− (3.156)
κ (2π)4 k 2 − m2 + iϵ [(p − k)2 − m2 + iϵ]2
g2 d4 k ∆2 [(p ⊕ S(k))2 ]
Z
1
− 2 (3.157)
κ (2π)4 k 2 − m2 + iϵ [(p − k)2 − m2 + iϵ]2
We follow the same steps as before. We have
∆1 [(p ⊕ S(k))2 ] = 2(k − p)(p0 k − k0 p) (3.158)

∆2 [(p ⊕ S(k))2 ] = k03 (−p0 ) + k02 2p20 − 2k2 + 2kp − p2

− k0 p0 p20 − k2 + p2 − p20 k(k − 2p) (3.159)

Since the denominators are always the same as before, we can treat them in the same
way, obtaining
g2 1
d4 k ∆shif t
[(p ⊕ S(k))2 ]
Z Z
− dx 1
(3.160)
κ 0 (2π)4 [k 2 − ∆(x) + iϵ]3
g2 1
d4 k ∆shif t
[(p ⊕ S(k))2 ]
Z Z
− 2 dx 2
(3.161)
κ 0 (2π)4 [k 2 − ∆(x) + iϵ]3
The shift is again k 7→ k + p(1 − x) and the numerators become
∆shif
1
t
[(p ⊕ S(k))2 ] = 2(k + p(1 − x) − p)(p0 (k + p(1 − x))
− p(k0 + p0 (1 − x))) (3.162)
∆shif
2
t
[(p ⊕ S(k))2 ]
= (k0 + p0 (1 − x))2 2p20 − 2(k + p(1 − x))2 + 2p(k + p(1 − x)) − p2

− p0 (k0 + p0 (1 − x)) p20 − (k + p(1 − x))2 + p2

− p0 (k0 + p0 (1 − x))3 − p20 (k + p(1 − x))(k + p(1 − x) − 2p) (3.163)
We can now perform the Wick rotation. The deformed part of the amplitude becomes
g2 1
[(p ⊕ S(k))2 ]
Z Z
i dx 1
(3.164)
κ 0 (2π)4 [k 2 + ∆(x)]3
g2 1 d4 k ∆shif t+W ick
[(p ⊕ S(k))2 ]
Z Z
+i 2 dx 2
(3.165)
κ 0 (2π)4 [k 2 + ∆(x)]3
Notice that the iϵ has gone away and there is a global sign change coming from
(−k 2 − ∆)3 = −(k 2 + ∆)3 . The additional i comes from the metric. We also remove
from both ∆1shif t+W ick [(p ⊕ S(k))2 ] and ∆shif
2
t+W ick
[(p ⊕ S(k))2 ] the irrelevant factors
(proportional to an odd power component of k), and using the isotropy of Euclidean
space. We have
∆shif
1
t+W ick
[(p ⊕ S(k))2 ] = +6p0 k02 (3.166)
∆2shif t+W ick [(p ⊕ S(k))2 ]

= k02 [p20 (−6x2 + 6x − 5) + p2 (2x2 − 2x + 1)] + 2k02 k2
+ p40 x3 − p40 x2 − 2p20 p2 x4 + 5p20 p2 x3 − 5p20 p2 x2 + 2p20 p2 x (3.167)
We now use the same integrals as before in the cutoff regularization, eq. (3.150),
(3.151), (3.152). Applying these we have
g2 1
d4 k ∆shif t=W ick
[(p ⊕ S(k))2 ]
Z Z
1
i dx
κ 0 (2π)4 [k 2 + ∆(x)]3
1
g2 d4 k k02
Z Z
=i dx6p0 (3.168)
κ 0 (2π) [k + ∆(x)]3
4 2
L2
g2 1
Z
∆
= −i 2
dx 6p0 log 0 (3.169)
κ 64π L1 |∆ |
g2 1
[(p ⊕ S(k))2 ]
Z Z
i dx 2
(3.170)
κ2 0 (2π)4 [k 2 + ∆(x)]3
L2
1 g2
Z
∆
=0− i dx[p20 (−6x2 + 6x − 5) + p2 (2x2 − 2x + 1)] log (3.171)
64π 2 κ2 L1 |∆0 |
L2
π2 9 g2
Z
∆
+i dx ∆ log (3.172)
12 (2π)4 κ2 L1 |∆0 |
This of course means that

L2
g2 1
Z
∆
Π(2) (p2 ) = ΠU (p2 ) − dx 6p0 log (3.173)
κ 64π 2 L1 |∆0 |
L2
1 g2
Z
∆
− dx[p20 (−6x2 + 6x − 5) + p2 (2x2 − 2x + 1)] log (3.174)
64π 2 κ2 L1 |∆0 |
L2
g2
Z
3 ∆
+ dx ∆ log (3.175)
64π 2 κ2 L1 |∆0 |
Once again, each logarithm will contribute −π to the imaginary part, and computing
the integral explicitly we end up with
s
g 2 3p m2
0
ℑΠ(2) (p2 ) = ℑΠU (p2 ) + 1−4 2
πκ 32 p
s
g2 m2 1 2 m2

1 2 1 2
+ 1 − 4 2 p0 − + + p + m (3.176)
16πκ2 p 6 3 p2 24 2
q
g2 2
p2
we finally obtain
1 2 m2

(2) 2 1
U 2 3 1 1 2 1 2
ℑΠ (p ) = ℑΠ (p ) 1 + p0 + 2 p0 − + + p + m
κ 2 κ 6 3 p2 24 2
(3.177)
We now pass to the third option. In this case we have
S(k)
iΠ(p2 ) =
k⊕p
d4 k
Z
i i
2
iΠ(p ) = (ig) 2
(3.178)
(2π) k − m + iϵ (k ⊕ p) − m2 + iϵ
4 2 2 2
which expanded to the first two non-trivial orders in powers of 1/κ is
iΠ(p2 ) = iΠU (p2 ) (3.179)

g2 d4 k ∆1 [(k ⊕ p)2 ]
Z
1
− (3.180)
κ (2π)4 k 2 − m2 + iϵ [(p + k)2 − m2 + iϵ]2
g2 d4 k ∆2 [(k ⊕ p)2 ]
Z
1
− 2 (3.181)
κ (2π)4 k 2 − m2 + iϵ [(p + k)2 − m2 + iϵ]2
Notice the presence of (p + k)2 in the denominator. This has no impact on the non-
deformed amplitude, because we can simply switch k 7→ −k, the metric remains
invariant as well as the integration domain, and we recover the usual integral. In the
deformed terms, however, we have to keep in mind that we must send k 7→ −k also
in the numerator before doing the Feynman trick, so that we can once again use the
same steps af before.
Apart from this, every consideration goes exactly as before, and this time we need
to compute
∆shif
1
+W ick
[(k ⊕ p)2 ], ∆shif
2
+W ick
[(k ⊕ p)2 ] (3.182)
once again. We have
∆1 [(k ⊕ p)2 ] = 2k(k0 p − p0 k) (3.183)
∆2 [(k ⊕ p)2 ] = (k0 + p0 ) p0 k0 (k0 + p0 ) + k2 − 2k0 kp − k0 p2

k 2 p2
− p20 k2 + k p2 − p20 (k + p) + (3.184)

3
where we used the identity
X X 1
(kp)2 = ki pi kj pj 7→ k2i (pi )2 7→ k02 p2 7→ k2 p2 . (3.185)
3
i,j i
We now send k 7→ −k. It is obvious that ∆1 [(k ⊕ p)2 ] is not affected by such a
change. On the other hand ∆2 [(k ⊕ p)2 ] is modified. We have
∆1 [(k ⊕ p)2 ]new = 2k(k0 p − p0 k) (3.186)
∆2 [(k ⊕ p)2 ]new = (p0 − k0 ) p0 k2 − k0 (p0 − k0 ) − 2k0 kp + k0 p2

k 2 p2
− p20 k2 − k p2 − p20 (p − k) + (3.187)

3
Since now the denominators are again the same as before, we can treat them in the
same way, obtaining
g2 1
d4 k ∆shif t
[(k ⊕ p)2 ]new
Z Z
− dx 1
(3.188)
κ 0 (2π)4 [k 2 − ∆(x) + iϵ]3
g2 1
d4 k ∆shif t
[(k ⊕ p)2 ]new
Z Z
− dx 2
(3.189)
κ2 0 (2π)4 [k 2 − ∆(x) + iϵ]3
The shift is again k 7→ k + p(1 − x), and performing the Wick rotation the amplitude
becomes

[(k ⊕ p)2 ]new
Z Z
1
i dx
κ 0 (2π)4 [k 2 + ∆(x) + iϵ]3
[(k ⊕ p)2 ]new
Z Z
+i 2 dx 2
(3.190)
κ 0 (2π)4 [k 2 + ∆(x) + iϵ]3
The numerators are
∆shif
1
t+W ick
[(k ⊕ p)2 ]new = −2p0 k2 (3.191)
∆shif
2
t+W ick
[(k ⊕ p)2 ]new = −3k02 p20 x + k02 p20 + 2k02 p2 x − k02 p2
+ p40 x3 − p40 x2 + p20 k2 x − 2p20 k2
4k2 p2
− p20 p2 x3 − p20 p2 x2 + 3p20 p2 x − p20 p2 +
3
4p4 x2 5p4 x p4
+ − + (3.192)
3 3 3
Once again, being in the Eulidean space, we can switch any isolated k2 with 3k02 ,
obtaining
∆1shif t+W ick [(k ⊕ p)2 ]new = −6p0 k02 (3.193)
∆shif
2
t+W ick
[(k ⊕ p)2 ]new = k02 [−5p20 + p2 (2x + 3)]
+ p40 x3 − p40 x2 − p20 p2 x3 − p20 p2 x2 + 3p20 p2 x − p20 p2
4p4 x2 5p4 x p4
+ − + (3.194)
3 3 3
Notice that the 1/κ contribution is numerically equivalent to the previous case, but
with a sign difference, so we don’t need to compute it. For what remains, since we
are only interested in the logarithmic contribution, using eq. (3.150), (3.151), (3.152),
since each log contributes a −π, and since
Z L2 s
m2
dx[−5p20 + p2 (2x + 3)] = 1 − 4 2 p20 + 4p2 (3.195)

L1 p
we end up with
s s
g2 m2 3p0 g2 m2 2
ℑΠ(3) = ℑΠU (p2 ) − p0 + 4p2 (3.196)

1−4 2 + 1−4
πκ p 32 64πκ2 p 2
q
g2 2
p2
we finally obtain

1 3 1 1 2
ℑΠ (3) U 2
= ℑΠ (p ) 1 − p0 + 2 p +p 2
(3.197)
κ 2 κ 4 0
Finally, we consider the last graph.
S(k)
iΠ(p2 ) =
p⊕k
d4 k
Z
i i
2
iΠ(p ) = (ig) 2
(3.198)
(2π) k − m + iϵ (p ⊕ k) − m2 + iϵ
4 2 2 2
The computations for this graph are exactly the same with the previous with k ↔ p
up to the shift, so we have
iΠ(p2 ) = iΠU (p2 ) (3.199)

g2 d4 k ∆1 [(p ⊕ k)2 ]
Z
1
− (3.200)
κ (2π)4 k 2 − m2 + iϵ [(p + k)2 − m2 + iϵ]2
g2 d4 k ∆2 [(p ⊕ k)2 ]
Z
1
− 2 (3.201)
κ (2π)4 k 2 − m2 + iϵ [(p + k)2 − m2 + iϵ]2
and
∆1 [(k ⊕ p)2 ]new = −2p(k0 p − p0 k) (3.202)
∆2 [(k ⊕ p)2 ]new = −p k2 − k02 (k − p) − k02 p2

k 2 p2
+ (k0 − p0 ) k0 p2 − p0 (k0 − p0 ) + p0 k2 − 2p0 kp + (3.203)

3
Since now the denominators are again the same as before, we can treat them in the
same way, obtaining
g2 1
d4 k ∆shif t
[(k ⊕ p)2 ]new
Z Z
− dx 1
(3.204)
κ 0 (2π)4 [k 2 − ∆(x) + iϵ]3
g2 1
d4 k ∆shif t
[(k ⊕ p)2 ]new
Z Z
− 2 dx 2
(3.205)
κ 0 (2π)4 [k 2 − ∆(x) + iϵ]3
The shift is again k 7→ k + p(1 − x), and performing then the Wick rotations the
amplitude becomes
g2 1
[(p ⊕ k)2 ]new
Z Z
i dx 1
(3.206)
κ 0 (2π)4 [k 2 + ∆(x) + iϵ]3
[(p ⊕ k)2 ]new
Z Z
+i 2 dx 2
(3.207)
κ 0 (2π)4 [k 2 + ∆(x) + iϵ]3
and the numerators (after the same simplifications) become
∆shif
1
t+W ick
[(k ⊕ p)2 ]new = 0 (3.208)
∆2shif t+W ick [(k ⊕ p)2 ]new = k02 [p20 (−6x + 1) + p2 (10x − 5)]
+ p40 x3 − p40 x2 − 2p20 p2 x3 + 2p20 p2 x2 + p20 p2 x
5p4 x2 p4 x p4
− p20 p2 + p4 x3 − + + (3.209)
3 3 3
Recalling once more eq. (3.150), (3.151), (3.152), since each log contributes a −π and
since
Z L2 s
m2
dx[p20 (−6x + 1) + p2 (10x − 5)] = 2 1 − 4 2 p20 (3.210)
L1 p
we end up with
s
g2 m2

1
ℑΠ(4) (p2 ) = ℑΠU (p2 ) + 1−4 2 p0 (3.211)
πκ2 p 32
q
g2 2
p2
we finally obtain

1 1 2
ℑΠ (4) U 2
= ℑΠ (p ) 1 + 2 p (3.212)
κ 2 0
3.4.1.1 Summary of the individual diagrams and total 1-loop amplitude

We considered the following diagrams
iΠ(1) (p2 ) =
S(k) ⊕ p
iΠ(2) (p2 ) =
p ⊕ S(k)
S(k)
iΠ(3) (p2 ) =
k⊕p
S(k)
iΠ(4) (p2 ) =
p⊕k
Each of them gives the imaginary part to the amplitude respectively in eq. (3.141),
(3.177), (3.197), (3.212). To get the complete contribution, we just sum the four dif-
ferent results
1 (1) 2
ℑΠT OT (p2 ) = ℑΠ (p ) + ℑΠ(2) (p2 ) + ℑΠ(3) (p2 ) + ℑΠ(4) (p2 ) (3.213)
4
where the factor 1/4 is necessary to get the correct κ → ∞ limit. We get
1 m2

1 1 2 19 5 2
ℑΠT OT (p2 ) = ℑΠU (p2 ) 1 − 2 p20 + − p − m (3.214)
κ 48 12 p2 48 24
Notice the remarkable fact that, although some of the diagrams contain a contribution
proportional to 1/κ, this contribution goes away after we sum all the contributions
from every diagram, leaving a leading contribution proportional to 1/κ2 .
Notice the important point that we chose our initial particle to have momentum
S(p), but we may just as well have decided to take a particle with initial momentum
−p. In this case momentum conservation would amount to
(−p) ⊕ p2 ⊕ p3 (3.215)
which is immediately satisfied if p2 ⊕ p3 = S(−p) (in order to avoid confusion, it is

understood that the equivalent computations using the momenta p∗ would amount
to having a particle with momentum p∗ , so that p2 ⊕ p3 = S(p∗ )). It is therefore
sufficient to switch every p in eq. (3.45), (3.46), (3.47), (3.48) with S(−p) and perform
the same computations again. Since we are expanding up to second order in 1/κ, the
final result will be the same as before plus some additional contribution. Since we have
already shown in detail the previous computations, here we limit ourselves to present
the additional contribution to the numerator of the sum of the four diagrams after all
the shifts, Wick rotation, and Euclidean simplifications have been performed. We call
δ∆shif
2
t+W ick
the additional contribution to the sum of the numerators. We have
8k02 p2 8p2 p20 −x2 + x − 2 + p2 (x − 1)x

shif t+W ick
δ∆2 = + (3.216)
κ2 κ2
where we included that 1/κ powers to show that also in this case we have no 1/κ
contribution after summing all four diagrams contributions. Notice that, because of
eq. (3.151), the term with no k in the numerator does not contribute to the imaginary
part, since it does not generate a log term upon integrating over k. The only term
8k2 p2
which contributes is the term κ02 , and we know from eq. (3.150) that it contributes
to the total amplitude as
1
−8p2 log ∆ (3.217)
64π 2
and since the log contributes as −π to the imaginary part we only need to compute
the quantity
Z L2
g2 1
2
p2

p0
− 2 2 U 2
8p dx = −ℑΠ (p ) 2 2 − 2 2 . (3.218)
κ 64π L1 κ κ
In other words, calling ℑΠS(p) the total amplitude ℑΠT OT (p2 ) in eq. (3.214) cor-
responding to an initial on-shell particle of momentum S(p), and calling ℑΠ−p the
equivalent quantity with momentum −p, we have
2
p2

p0
U 2
ℑΠS(p) − ℑΠ−p = ℑΠ (p ) 2 2 − 2 2 . (3.219)
κ κ
2 m2 p2
The quantity κp2 = κ2ψ is invariant, but the quantity κ02 is sensible to the particles
energy, which means that this difference can be experimentally highlighted by going
to higher energies.
3.5. Mass distribution width 127
3.5 Mass distribution width

In section 2.18 in chapter 2 we addressed possible phenomenological consequences of
the deformation of CP T symmetry, but we did not consider for the moment possi-
ble contributions coming from the higher order corrections to the propagator of the
decaying particle. Such corrections would induce a widening of the mass distribution,
and such contributions need to be taken into account.
More formally, given a mass distribution function ω(m; M, Γ), where m is the
resonant mass, and with mean mass M and decay width Γ, one can describe the time
and momentum dependent decay amplitude with the expression
Z ∞ √
2 2
a(t, p) = dm ω(m; M, Γ) e−it m +p . (3.220)
−∞
The decay probability is then given by the usual formula P = |a(t, p)|2 . Since we are
interested in ultra-relativistic particles (we need high momenta and energies in order
to counterbalance the large value of κ in the denominator), we consider the relativistic
Breit-Wigner distribution
f (M, Γ)
ω(m; M, Γ) = , (3.221)
(m2− M 2 )2 + M 2 Γ2
√ p
2 2 M Γ M 2 (M 2 + Γ2 )
f (M, Γ) = . (3.222)
π M 2 + M 2 (M 2 + Γ2 )1/2
p
Notice that f (M, Γ) is independent of m. Since p we want to understand the effect

of deformation, we now switch the energy E = m2 + p2 at the exponent for its
antipode S(E) (see eq. (2.533)). Furthermore, one can show that
Z
f (M, Γ) √
−it( m2 +p2 −p2 /κ)
a(t, p) = dm e
(m2 − M 2 )2 + M 2 Γ2
√ 2
2 2
= e−it( M +p +iM Γ−p /κ) . (3.223)
The new, corrected decay width Γ̃ (which takes into consideration the non-trivial mass
distribution width) can be now computed by taking into account the real part of the
exponent in eq. (3.223) (which will describe the exponential decay of the amplitude).
2
However, the only effect of deformation comes into play in the term −it pκ , i.e. in
the imaginary part, so no correction to Γ can be obtained from a non-trivial mass
distribution width. Indeed, one explicitly has
p p2
Γ̃ = 2 ℑ M 2 + p2 + iM Γ −
κ
√ hp i1/2
= 2 (M 2 + p2 )2 + M 2 Γ2 − (M 2 + Γ2 ) . (3.224)
so that the only corrections are the same that would be present in the canonical,
non-deformed context.
3.6 Summary
In this work, we have presented our model of scalar field theory, and analysed its
features and the possible phenomenological consequences coming from our results.
We first introduced the framework of κ-Minkowski spacetime and κ-Poincaré Hopf

algebra, which are the fundamental ingredients for any other construction. Before
going to the complex scalar field, we tackled the interesting problem of the boost of
a two-particle state system. The determination of the finite boost of a two-particle
system allows us to draw interesting phenomenological consequences for any type of
particle, since the results are purely of kinematical nature.
We then passed to the complex scalar field. After some preliminary mathematical
work concerning the Weyl maps, the action of derivatives on star products, and the
integration-by-parts relations, we introduced the complex scalar field action and com-
puted the EoM (which turned out to be the canonical Klein-Gordon equations). This
allowed us to derive an explicit formula for the on-shel fields ϕ, ϕ† . Furthermore, using
the surface terms obtained during the integration-by-parts, we obtained the Noether
charges coming from the continuous translation symmetry of the action.
To compute the remaining Noether charges we used the symplectic form, which
was also important for the fact that it allowed us to obtain the creation/annihilation
operators commutation relation. Together with the charges, these allowed us to check
that the algebra satisfied by the charges is the canonical Poincaré algebra. Notice
that, despite the fact that the EoM are the canonical Klein-Gordon equations, and
the fact that the charges satisfy the canonical Poincaré algebra, there are still effect of
deformation on single-particle states which are made manifest by looking at the non-
trivial interaction between discrete and continuous transformations. Here we explicitly
checked that, in the deformed case, we have [Ni , C] ̸= 0. Furthermore, κ-deformation
affects in non-trivial way the features vacuum expectation values of fields. In this work,
we discussed in detail Greenberg’s theorem, showing that it does not hold in general
in the κ-deformed context. We then explored some phenomenological consequences of
the fact that [Ni , C] ̸= 0, by boosting a particle and an antiparticle, both originally
at rest.
We then proceeded to compute the propagator and the imaginary part of the
1-loop correction to it. To do so, we exploit the path integral formalism for the de-
termination of the tree-level propagation, and we assume a deformed conservation of
momenta at each vertex. Because of this, a single diagram in the non-deformed context
is translated into four different diagrams in the deformed context. We computed all of
them for initial momentum of a particle given by S(p) and −p, obtaining a difference
in the imaginary part proportional to p20 /κ2 . We conclude with some comments on
the relevance of a non-trivial mass distribution width on the decay times of boosted
particles and antiparticles.
129
Bibliografia
[1] Ought, Elicit: The AI Research Assistant https://elicit.org

[2] A. Einstein, “Näherungsweise Integration der Feldgleichungen der Gravitation,”
Preussische Akademie der Wissenschaften (Berlin) Sitzungsberichte (1916), pp.
99-108
[3] C. Rovelli, “Notes for a brief history of quantum gravity,” [arXiv:gr-qc/0006061
[gr-qc]].
[4] M. Bronstein, “Quantum theory of weak gravitational fields,” Gen. Rel. Grav. 44
(2012), 267-283, reprint doi:10.1007/s10714-011-1285-4
[5] F. Dyson, “Is a graviton detectable?,” Int. J. Mod. Phys. A 28 (2013), 1330041
doi:10.1142/S0217751X1330041X
[6] R. Jackiw, “Noncommuting fields and nonAbelian fluids,” Nucl. Phys. B
Proc. Suppl. 127 (2004), 53-62 doi:10.1016/S0920-5632(03)02400-9 [arXiv:hep-
th/0305027 [hep-th]].
[7] H. S. Snyder, “Quantized Space-Time,” Phys. Rev. 71 (1947), 38-41
doi:10.1103/PhysRev.71.38
[8] G. Amelino-Camelia, “Are we at the dawn of quantum gravity phenomenology?,”
Lect. Notes Phys. 541 (2000), 1-49 [arXiv:gr-qc/9910089 [gr-qc]].
[9] J. R. Ellis, N. E. Mavromatos and D. V. Nanopoulos, “Quantum gravitational
diffusion and stochastic fluctuations in the velocity of light,” Gen. Rel. Grav. 32
(2000), 127-144 doi:10.1023/A:1001852601248 [arXiv:gr-qc/9904068 [gr-qc]].
[10] J. R. Ellis, N. E. Mavromatos and D. V. Nanopoulos, “Search for quantum
gravity,” Gen. Rel. Grav. 31 (1999), 1257-1262 doi:10.1023/A:1026720723556
[arXiv:gr-qc/9905048 [gr-qc]].
[11] J. R. Ellis, K. Farakos, N. E. Mavromatos, V. A. Mitsou and D. V. Nanopoulos,
“Astrophysical probes of the constancy of the velocity of light,” Astrophys. J. 535
(2000), 139-151 doi:10.1086/308825 [arXiv:astro-ph/9907340 [astro-ph]].
[12] G. Amelino-Camelia, A. Marciano, M. Matassa and G. Rosati, “Testing
quantum-spacetime relativity with gamma-ray telescopes,” [arXiv:1006.0007
[astro-ph.HE]].
[13] G. Amelino-Camelia, N. Loret and G. Rosati, “Speed of particles and a relativity
of locality in κ-Minkowski quantum spacetime,” Phys. Lett. B 700 (2011), 150-
156 doi:10.1016/j.physletb.2011.04.054 [arXiv:1102.4637 [hep-th]].
[14] G. Amelino-Camelia, M. Arzano, J. Kowalski-Glikman, G. Rosati and G. Trevi-
san, “Relative-locality distant observers and the phenomenology of momentum-
space geometry,” Class. Quant. Grav. 29 (2012), 075007 doi:10.1088/0264-
9381/29/7/075007 [arXiv:1107.1724 [hep-th]].
130 BIBLIOGRAFIA
[15] N. E. Mavromatos, A. Meregaglia, A. Rubbia, A. Sakharov and S. Sar-

kar, “Quantum-Gravity Decoherence Effects in Neutrino Oscillations: Expec-
ted Constraints From CNGS and J-PARC,” Phys. Rev. D 77 (2008), 053014
doi:10.1103/PhysRevD.77.053014 [arXiv:0801.0872 [hep-ph]].
[16] G. Amelino-Camelia, L. Barcaroli, G. D’Amico, N. Loret and G. Rosati, “Ice-

Cube and GRB neutrinos propagating in quantum spacetime,” Phys. Lett. B 761
(2016), 318-325 doi:10.1016/j.physletb.2016.07.075 [arXiv:1605.00496 [gr-qc]].
[17] G. Amelino-Camelia, L. Barcaroli, G. D’Amico, N. Loret and G. Ro-

sati, “Quantum-gravity-induced dual lensing and IceCube neutrinos,” Int.
J. Mod. Phys. D 26 (2017) no.08, 1750076 doi:10.1142/S0218271817500766
[arXiv:1609.03982 [gr-qc]].
[18] G. Amelino-Camelia, G. D’Amico, G. Rosati and N. Loret, “In-vacuo-dispersion

features for GRB neutrinos and photons,” Nature Astron. 1 (2017), 0139
doi:10.1038/s41550-017-0139 [arXiv:1612.02765 [astro-ph.HE]].
[19] G. Rosati, “Investigating Quantum Spacetime Through Ice Cube Astrophy-

sical Neutrinos Observation,” Acta Phys. Polon. Supp. 10 (2017) no.2, 325
doi:10.5506/APhysPolBSupp.10.325
[20] G. Amelino-Camelia, M. G. Di Luca, G. Gubitosi, G. Rosati and G. D’Amico,

“Could quantum gravity slow down neutrinos?,” Nature Astron. 7 (2023) no.8,
996-1001 doi:10.1038/s41550-023-01993-z [arXiv:2209.13726 [gr-qc]].
[21] J. Lukierski, A. Nowicki and H. Ruegg, Phys. Lett. B 293 (1992), 344-352
doi:10.1016/0370-2693(92)90894-A
[22] J. Lukierski and H. Ruegg, “Quantum kappa Poincare in any dimension,”

Phys. Lett. B 329 (1994), 189-194 doi:10.1016/0370-2693(94)90759-5 [arXiv:hep-
th/9310117 [hep-th]].
[23] J. Lukierski, H. Ruegg and W. J. Zakrzewski, “Classical quantum me-

chanics of free κ relativistic systems,” Annals Phys. 243 (1995), 90-116
doi:10.1006/aphy.1995.1092 [arXiv:hep-th/9312153 [hep-th]].
[24] P. Kosinski, J. Lukierski, P. Maslanka and J. Sobczyk, “k deformation of Poin-

care superalgebra with classical Lorentz subalgebra and its graded bicrossproduct
structure,” J. Phys. A 28 (1995), 2255-2264 doi:10.1088/0305-4470/28/8/017
[arXiv:hep-th/9411115 [hep-th]].
[25] P. Kosinski, J. Lukierski, P. Maslanka and J. Sobczyk, “The Classical ba-

sis for kappa deformed Poincare (super)algebra and the second kappa de-
formed supersymmetric Casimir,” Mod. Phys. Lett. A 10 (1995), 2599-2606
doi:10.1142/S0217732395002738 [arXiv:hep-th/9412114 [hep-th]].
[26] J. Lukierski, “Quantum deformations of space-time symmetries with mass - like

deformation parameter,” [arXiv:hep-th/9610230 [hep-th]].
[27] G. Amelino-Camelia, “Relativity in space-times with short distance structure

governed by an observer independent (Planckian) length scale,” Int. J. Mod.
Phys. D 11 (2002), 35-60 doi:10.1142/S0218271802001330 [arXiv:gr-qc/0012051
[gr-qc]].
BIBLIOGRAFIA 131
[28] G. Amelino-Camelia, “Doubly special relativity,” Nature 418 (2002), 34-35

doi:10.1038/418034a [arXiv:gr-qc/0207049 [gr-qc]].
[29] G. Amelino-Camelia, “Doubly special relativity: First results and key open pro-
blems,” Int. J. Mod. Phys. D 11 (2002), 1643 doi:10.1142/S021827180200302X
[arXiv:gr-qc/0210063 [gr-qc]].
[30] J. Kowalski-Glikman, “Observer independent quantum of mass,” Phys. Lett. A

286 (2001), 391-394 doi:10.1016/S0375-9601(01)00465-0 [arXiv:hep-th/0102098
[hep-th]].
[31] G. Amelino-Camelia, “Quantum-Spacetime Phenomenology,” Living Rev. Rel. 16

(2013), 5 doi:10.12942/lrr-2013-5 [arXiv:0806.0339 [gr-qc]].
[32] A. Addazi, J. Alvarez-Muniz, R. Alves Batista, G. Amelino-Camelia, V. Antonelli,

M. Arzano, M. Asorey, J. L. Atteia, S. Bahamonde and F. Bajardi, et al. “Quan-
tum gravity phenomenology at the dawn of the multi-messenger era—A review,”
Prog. Part. Nucl. Phys. 125 (2022), 103948 doi:10.1016/j.ppnp.2022.103948
[arXiv:2111.05659 [hep-ph]].
[33] G. Amelino-Camelia, L. Freidel, J. Kowalski-Glikman and L. Smolin,

“The principle of relative locality,” Phys. Rev. D 84 (2011), 084010
doi:10.1103/PhysRevD.84.084010 [arXiv:1101.0931 [hep-th]].
[34] G. Amelino-Camelia, L. Freidel, J. Kowalski-Glikman and L. Smolin, “Relative

locality: A deepening of the relativity principle,” Gen. Rel. Grav. 43 (2011),
2547-2553 doi:10.1142/S0218271811020743 [arXiv:1106.0313 [hep-th]].
[35] G. Amelino-Camelia, M. Arzano, J. Kowalski-Glikman, G. Rosati and G. Trevi-

san, “Relative-locality distant observers and the phenomenology of momentum-
space geometry,” Class. Quant. Grav. 29 (2012), 075007 doi:10.1088/0264-
9381/29/7/075007 [arXiv:1107.1724 [hep-th]].
[36] J. Kowalski-Glikman and S. Nowak, “Doubly special relativity theories as dif-

ferent bases of kappa Poincare algebra,” Phys. Lett. B 539 (2002), 126-132
doi:10.1016/S0370-2693(02)02063-4 [arXiv:hep-th/0203040 [hep-th]].
[37] J. Kowalski-Glikman, “A short introduction to κ-deformation,” Int. J. Mod. Phys.

A 32 (2017) no.35, 1730026 doi:10.1142/S0217751X17300265 [arXiv:1711.00665
[hep-th]].
[38] L. Freidel, J. Kowalski-Glikman and L. Smolin, “2+1 gravity and doubly spe-
cial relativity,” Phys. Rev. D 69 (2004), 044001 doi:10.1103/PhysRevD.69.044001
[39] S. Meljanac, A. Samsarov, M. Stojic and K. S. Gupta, “Kappa-Minkowski space-

time and the star product realizations,” Eur. Phys. J. C 53 (2008), 295-309
doi:10.1140/epjc/s10052-007-0450-0 [arXiv:0705.2471 [hep-th]].
[40] M. Daszkiewicz, K. Imilkowska, J. Kowalski-Glikman and S. Nowak, Int. J.

Mod. Phys. A 20 (2005), 4925-4940 doi:10.1142/S0217751X0502238X [arXiv:hep-
th/0410058 [hep-th]].
[41] F. Mercati and M. Sergola, “Pauli-Jordan function and scalar field quantization in
κ-Minkowski noncommutative spacetime,” Phys. Rev. D 98 (2018) no.4, 045017
132 BIBLIOGRAFIA
[42] S. Meljanac, A. Samsarov, J. Trampetic and M. Wohlgenannt, “Noncommuta-

tive kappa-Minkowski phi4 theory: Construction, properties and propagation,”
[arXiv:1107.2369 [hep-th]].
[43] T. R. Govindarajan, K. S. Gupta, E. Harikumar, S. Meljanac and D. Meljanac,

“Deformed Oscillator Algebras and QFT in kappa-Minkowski Spacetime,” Phys.
Rev. D 80 (2009), 025014 doi:10.1103/PhysRevD.80.025014 [arXiv:0903.2355
[hep-th]].
[44] H. C. Kim, Y. Lee, C. Rim and J. H. Yee, “Scalar Field theory in kappa-Minkowski
spacetime from twist,” J. Math. Phys. 50 (2009), 102304 doi:10.1063/1.3250148
[arXiv:0901.0049 [hep-th]].
[45] T. Poulain and J. C. Wallet, “κ-Poincaré invariant orientable field theories at one-
loop,” JHEP 01 (2019), 064 doi:10.1007/JHEP01(2019)064 [arXiv:1808.00350
[hep-th]].
[46] M. Daszkiewicz, J. Lukierski and M. Woronowicz, “Kappa-deformed oscillators,

the choice of star product and free kappa-deformed quantum fields,” J. Phys. A 42
(2009), 355201 doi:10.1088/1751-8113/42/35/355201 [arXiv:0807.1992 [hep-th]].
[47] M. Daszkiewicz, J. Lukierski and M. Woronowicz, “Towards quantum non-

commutative kappa-deformed field theory,” Phys. Rev. D 77 (2008), 105007
[48] M. Arzano and A. Marciano, “Fock space, quantum fields and kappa-Poincare
symmetries,” Phys. Rev. D 76 (2007), 125005 doi:10.1103/PhysRevD.76.125005
[arXiv:0707.1329 [hep-th]].
[49] P. Kosinski, J. Lukierski and P. Maslanka, “kappa deformed Wigner construction

of relativistic wave functions and free fields on kappa-Minkowski space,” Nucl.
Phys. B Proc. Suppl. 102 (2001), 161-168 doi:10.1016/S0920-5632(01)01552-3
[50] P. Kosinski, P. Maslanka, J. Lukierski and A. Sitarz, “Generalized kappa defor-

mations and deformed relativistic scalar fields on noncommutative Minkowski
space,” [arXiv:hep-th/0307038 [hep-th]].
[51] I. P. Lobo and C. Pfeifer, “Reaching the Planck scale with muon
lifetime measurements,” Phys. Rev. D 103 (2021) no.10, 106025
doi:10.1103/PhysRevD.103.106025 [arXiv:2011.10069 [hep-ph]].
[52] G. Amelino-Camelia, J. R. Ellis, N. E. Mavromatos, D. V. Nanopoulos and S. Sar-

kar, “Tests of quantum gravity from observations of gamma-ray bursts,” Nature
393 (1998), 763-765 doi:10.1038/31647 [arXiv:astro-ph/9712103 [astro-ph]].
[53] J. R. Ellis, K. Farakos, N. E. Mavromatos, V. A. Mitsou and D. V. Nanopoulos,

“Astrophysical probes of the constancy of the velocity of light,” Astrophys. J. 535
(2000), 139-151 doi:10.1086/308825 [arXiv:astro-ph/9907340 [astro-ph]].
[54] P. Aschieri, A. Borowiec and A. Pachoł, “Observables and dispersion relations

in κ-Minkowski spacetime,” JHEP 10 (2017), 152 doi:10.1007/JHEP10(2017)152
[arXiv:1703.08726 [hep-th]].
BIBLIOGRAFIA 133
[55] J. R. Ellis, N. E. Mavromatos, D. V. Nanopoulos and G. Volkov, “Gravitational

recoil effects on fermion propagation in space-time foam,” Gen. Rel. Grav. 32
(2000), 1777-1798 doi:10.1023/A:1001980530113 [arXiv:gr-qc/9911055 [gr-qc]].
[56] G. Amelino-Camelia, J. R. Ellis, N. E. Mavromatos and D. V. Nanopo-

ulos, “Distance measurement and wave dispersion in a Liouville string ap-
proach to quantum gravity,” Int. J. Mod. Phys. A 12 (1997), 607-624
doi:10.1142/S0217751X97000566 [arXiv:hep-th/9605211 [hep-th]].
[57] G. Amelino-Camelia, D. Frattulillo, G. Gubitosi, G. Rosati and S. Bedic,

“Phenomenology of DSR-relativistic in-vacuo dispersion in FLRW spacetime,”
[arXiv:2307.05428 [gr-qc]].
[58] L. Barcaroli, L. K. Brunkhorst, G. Gubitosi, N. Loret and C. Pfeifer, “Curved

spacetimes with local κ-Poincaré dispersion relation,” Phys. Rev. D 96 (2017)
no.8, 084010 doi:10.1103/PhysRevD.96.084010 [arXiv:1703.02058 [gr-qc]].
[59] J. Kowalski-Glikman and S. Nowak, “Noncommutative space-time of do-

ubly special relativity theories,” Int. J. Mod. Phys. D 12 (2003), 299-316
doi:10.1142/S0218271803003050 [arXiv:hep-th/0204245 [hep-th]].
[60] M. Arzano and J. Kowalski-Glikman, “A group theoretic description

of the κ-Poincaré Hopf algebra,” Phys. Lett. B 835 (2022), 137535
doi:10.1016/j.physletb.2022.137535 [arXiv:2204.09394 [hep-th]].
[61] S. Majid and H. Ruegg, “Bicrossproduct structure of kappa Poincare group and
noncommutative geometry,” Phys. Lett. B 334 (1994), 348-354 doi:10.1016/0370-
2693(94)90699-8 [arXiv:hep-th/9405107 [hep-th]].
[62] P. Kosinski and P. Maslanka, “The Duality between kappa Poincare algebra and
kappa Poincare group,” [arXiv:hep-th/9411033 [hep-th]].
[63] A. Borowiec and A. Pachol, “kappa-Minkowski spacetime as the re-

sult of Jordanian twist deformation,” Phys. Rev. D 79 (2009), 045012
doi:10.1103/PhysRevD.79.045012 [arXiv:0812.0576 [math-ph]].
[64] A. Borowiec and A. Pachol, “κ-Minkowski spacetimes and DSR algebras: Fresh
look and old problems,” SIGMA 6 (2010), 086 doi:10.3842/SIGMA.2010.086
[arXiv:1005.4429 [math-ph]].
[65] A. Pachol, “κ-Minkowski spacetime: Mathematical formalism and applications in

Planck scale physics,” [arXiv:1112.5366 [math-ph]].
[66] M. Arzano, G. Gubitosi and J. J. Relancio, “Deformed relativistic symmetry

principles,” [arXiv:2211.11684 [hep-th]].
[67] G. Gubitosi and S. Heefer, “Relativistic compatibility of the interacting κ-

Poincaré model and implications for the relative locality framework,” Phys. Rev.
D 99 (2019) no.8, 086019 doi:10.1103/PhysRevD.99.086019 [arXiv:1903.04593
[gr-qc]].
[68] A. Borowiec and A. Pachol, “Unified description for κ-deformations of orthogonal

groups,” Eur. Phys. J. C 74 (2014) no.3, 2812 doi:10.1140/epjc/s10052-014-2812-
8 [arXiv:1311.4499 [math-ph]].
134 BIBLIOGRAFIA
[69] A. Borowiec, J. Lukierski and A. Pachoł, “Twisting and κ-Poincaré,” J. Phys. A

47 (2014) no.40, 405203 doi:10.1088/1751-8113/47/40/405203 [arXiv:1312.7807
[math-ph]].
[70] A. Ballesteros, G. Gubitosi, I. Gutierrez-Sagredo and F. J. Herranz, “κ-Galilean

and κ-Carrollian noncommutative spaces of worldlines,” Phys. Lett. B 838 (2023),
137735 doi:10.1016/j.physletb.2023.137735 [arXiv:2212.01125 [hep-th]].
[71] A. Ballesteros, G. Gubitosi, I. Gutierrez-Sagredo and F. J. Herranz, “The

κ-Newtonian and κ-Carrollian algebras and their noncommutative spaceti-
mes,” Phys. Lett. B 805 (2020), 135461 doi:10.1016/j.physletb.2020.135461
[arXiv:2003.03921 [hep-th]].
[72] A. Ballesteros, G. Gubitosi and F. Mercati, “Interplay between Spacetime Cu-

rvature, Speed of Light and Quantum Deformations of Relativistic Symmetries,”
Symmetry 13 (2021) no.11, 2099 doi:10.3390/sym13112099 [arXiv:2110.04867 [gr-
qc]].
[73] F. Mercati, “A perspective on the theory and phenomenology of quantum space-

time,”
[74] A. Borowiec and A. Pachol, “Classical basis for kappa-Poincare algebra and do-
ubly special relativity theories,” J. Phys. A 43 (2010), 045203 doi:10.1088/1751-
8113/43/4/045203 [arXiv:0903.5251 [hep-th]].
[75] M. Arzano and J. Kowalski-Glikman, “Deformations of Spacetime Symmetries:

Gravity, Group-Valued Momenta, and Non-Commutative Fields,” 2021, ISBN
978-3-662-63095-2, 978-3-662-63097-6 doi:10.1007/978-3-662-63097-6
[76] J. Kowalski-Glikman and A. Bevilacqua, “Doubly special relativity and

relative locality,” PoS CORFU2021 (2022), 322 doi:10.22323/1.406.0322
[arXiv:2203.04091 [physics.gen-ph]].
[77] A. Bevilacqua, J. Kowalski-Glikman and W. Wiślicki, “Finite κ-deformed two-

particle boost,” [arXiv:2305.09180 [hep-th]].
[78] M. Arzano, A. Bevilacqua, J. Kowalski-Glikman, G. Rosati and J. Unger, “κ-

deformed complex fields and discrete symmetries,” Phys. Rev. D 103 (2021)
no.10, 106015 doi:10.1103/PhysRevD.103.106015 [arXiv:2011.09188 [hep-th]].
[79] L. Freidel, J. Kowalski-Glikman and S. Nowak, “Field theory on kappa-Minkowski

space revisited: Noether charges and breaking of Lorentz symmetry,” Int. J. Mod.
Phys. A 23 (2008), 2687-2718 doi:10.1142/S0217751X08040421 [arXiv:0706.3658
[hep-th]].
[80] G. Amelino-Camelia and M. Arzano, “Coproduct and star product in field theories
on Lie algebra noncommutative space-times,” Phys. Rev. D 65 (2002), 084044
doi:10.1103/PhysRevD.65.084044 [arXiv:hep-th/0105120 [hep-th]].
[81] D. Oriti and G. Rosati, “Noncommutative Fourier transform for the Lo-
rentz group via the Duflo map,” Phys. Rev. D 99 (2019) no.10, 106005
[82] C. Guedes, D. Oriti and M. Raasakka, “Quantization maps, algebra representa-

tion and non-commutative Fourier transform for Lie groups,” J. Math. Phys. 54
(2013), 083508 doi:10.1063/1.4818638 [arXiv:1301.7750 [math-ph]].
BIBLIOGRAFIA 135
[83] A. Bevilacqua, J. Kowalski-Glikman and W. Wislicki, “κ-deformed complex sca-

lar field: Conserved charges, symmetries, and their impact on physical observa-
bles,” Phys. Rev. D 105 (2022) no.10, 105004 doi:10.1103/PhysRevD.105.105004
[arXiv:2201.10191 [hep-th]].
[84] A. Bevilacqua and J. Kowalski-Glikman, “Doubly Special Relativity and Relative

Locality,” [arXiv:2203.04091 [physics.gen-ph]].
[85] M. Arzano and L. T. Consoli, “Signal propagation on κ-Minkowski space-

time and nonlocal two-point functions,” Phys. Rev. D 98 (2018) no.10, 106018
[86] L. Freidel, J. Kowalski-Glikman and S. Nowak, “Field theory on kappa-Minkowski

space revisited: Noether charges and breaking of Lorentz symmetry,” Int. J. Mod.
Phys. A 23 (2008), 2687-2718 doi:10.1142/S0217751X08040421 [arXiv:0706.3658
[hep-th]].
[87] D. Harlow and J. Q. Wu, “Covariant phase space with boundaries,” JHEP 10
(2020), 146 doi:10.1007/JHEP10(2020)146 [arXiv:1906.08616 [hep-th]].
[88] R. M. Wald, “Black hole entropy is the Noether charge,” Phys. Rev. D 48 (1993)
no.8, R3427-R3431 doi:10.1103/PhysRevD.48.R3427 [arXiv:gr-qc/9307038 [gr-
qc]].
[89] R. M. Wald, “Black hole entropy is the Noether charge,” Phys. Rev. D 48 (1993)
no.8, R3427-R3431 doi:10.1103/PhysRevD.48.R3427 [arXiv:gr-qc/9307038 [gr-
qc]].
[90] V. Iyer and R. M. Wald, “Some properties of Noether charge and a pro-
posal for dynamical black hole entropy,” Phys. Rev. D 50 (1994), 846-864
doi:10.1103/PhysRevD.50.846 [arXiv:gr-qc/9403028 [gr-qc]].
[91] F.J.E. Dillen, L.C.A. Verstraelen, “Handbook of Differential Geometry” isbn:

9780080461205 Elsevier Science.
[92] O. W. Greenberg, “CPT violation implies violation of Lorentz invariance,” Phys.

Rev. Lett. 89 (2002), 231602 doi:10.1103/PhysRevLett.89.231602 [arXiv:hep-
ph/0201258 [hep-ph]].
[93] M. Duetsch and J. M. Gracia-Bondia, “On the assertion that PCT vio-
lation implies Lorentz non-invariance,” Phys. Lett. B 711 (2012), 428-433
doi:10.1016/j.physletb.2012.04.038 [arXiv:1204.2654 [hep-th]].
[94] M. Chaichian, A. D. Dolgov, V. A. Novikov and A. Tureanu, “CPT Violation Does

Not Lead to Violation of Lorentz Invariance and Vice Versa,” Phys. Lett. B 699
(2011), 177-180 doi:10.1016/j.physletb.2011.03.026 [arXiv:1103.0168 [hep-th]].
[95] R. Jost, “A remark on the C.T.P. theorem,” Helv. Phys. Acta 30 (1957), 409-416
[96] A. Wightman, “Recent Achievements of Axiomatic Field Theory,”
[97] O. W. Greenberg, “Why is CPT fundamental?,” Found. Phys. 36 (2006), 1535-

1553 doi:10.1007/s10701-006-9070-z [arXiv:hep-ph/0309309 [hep-ph]].
[98] R. Jost, “Eine Bemerkung zum CPT Theorem,” https://doi.org/10.5169/

seals-112824
136 BIBLIOGRAFIA
[99] M. Arzano and J. Kowalski-Glikman, “Quantum particles in noncommuta-

tive spacetime: An identity crisis,” Phys. Rev. D 107 (2023) no.6, 065001
[100] M. Duetsch and J. M. Gracia-Bondia, “On the assertion that PCT vio-
lation implies Lorentz non-invariance,” Phys. Lett. B 711 (2012), 428-433,
doi:10.1016/j.physletb.2012.04.038 [arXiv:1204.2654 [hep-th]].
[101] M. Chaichian, A. D. Dolgov, V. A. Novikov and A. Tureanu, “CPT Violation

Does Not Lead to Violation of Lorentz Invariance and Vice Versa,” Phys. Lett. B
699 (2011), 177-180, doi:10.1016/j.physletb.2011.03.026 [arXiv:1103.0168 [hep-
th]].
[102] M. Arzano, J. Kowalski-Glikman and W. Wislicki, “A bound on Planck-scale

deformations of CPT from muon lifetime,” Phys. Lett. B 794 (2019), 41-44
doi:10.1016/j.physletb.2019.05.025 [arXiv:1904.06754 [hep-ph]].
[103] KLOE Collaboration: D. Babusci et al., “Test of CPT and Lorentz symmetry
in entangled neutral kaons with the KLOE experiment,” Phys. Lett. B 730, 89
(2014) doi:10.1016/j.physletb.2014.01.026 [arXiv:1312.6818 [hep-ex]].
[104] KLOE and KLOE-2 Collaborations: D. Babusci et al., “Precision tests of Qu-
antum Mechanics and CPT symmetry with entangled neutral kaons at KLOE,”
submitted to JHEP, [arXiv:2111.0432[hep-ex]
[105] LHCb Collaboration: R. Aaij et al., “Search for violations of Lorentz inva-
riance and CPT symmetry in B(s)
0 mixing,” Phys. Rev. Lett. 116, 241601 (2016),
doi:10.1103/PhysRevLett.116.24160 [1603.04804 [hep-ex]]
[106] BaBar Collaboration: B. Aubert et al., “Search for CPT and Lorentz violation in
B 0 − B̄ 0 oscillations with dilepton events ,” Phys. Rev. Lett. 100, 131802 (2008),
doi:10.1103/PhysRevLett.100.131802 [0711.2713 [hep-ex]]
[107] D0 Collaboration: V.M. Abazov et al., “Search for Violation of CPT and Lorentz
Invariance in B(s)
0 Meson Oscillations,” Phys. Rev. Lett. 115, 161601 (2005), doi:
10.1103/PhysRevLett.115.161601 [1506.04123 [hep-ex]]
[108] FOCUS Collaboration: J.M. Link et al., “Studies of correlations between D

and D̄ mesons in high-energy photoproduction,” Phys. Lett. B566, 51 (2003),
doi:10.1016/S0370-2693(03)00808-6 [hep-ex/0305018 [hep-ex]]
[109] https://home.cern/science/accelerators/future-circular-collider
[110] M. Tanabashi et al. (Particle Data Group), Phys. Rev. D98, 030001 (2019);
http://pdg.lbl.gov

κ-deformed scalar field: National Centre for Nuclear Research

Uploaded by

Copyright:

Available Formats

κ-deformed scalar field: National Centre for Nuclear Research

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

κ-deformed scalar field: National Centre for Nuclear Research

Uploaded by

Copyright:

Available Formats

National Centre for Nuclear Research

κ-deformed scalar field

A thesis submitted in fulfillment of the requirements

• I have acknowledged all main sources of help.

NATIONAL CENTRE FOR NUCLEAR RESEARCH

W niniejszej pracy wprowadzimy i przedyskutujemy w szczegółach model zespolonego,

Declaration of Authorship iii

1 κ-Minkowski and κ-Poincaré 1

2 Complex scalar field on κ-Minkowski spacetime 31

2.12.1 First ordering of the action . . . . . . . . . . . . . . . . . . . . 47

3 κ-deformed propagator, and 1-loop correction to it 101

To A.B. and L.B-B.

1.1 Introduction to κ-deformation

where Ea µ is the tetrad which is related to the momentum-space metric by

Gµν (k, κ) = η ab Ea µ (k, κ)Eb ν (k, κ), (1.3)

One can show that

1.2 Formal definition of Hopf algebra

A co-algebra can basically be defined by reversing all the arrows. In particular, a

A bi-algebra B is therefore both an algebra and a co-algebra, with additional

∆(µ(a, b)) = µ(∆(a), ∆(b)), ϵ(µ(a, b)) = ϵ(a)ϵ(b) (1.8)

for all a, b ∈ B, and where

∆(id) = id ⊗ id ϵ(id) = 1 (1.9)

where 1 ∈ K. Finally, a Hopf algebra H is a bi-algebra with an additional map

S(µ(a, b)) = µ(S(a), S(b)) S(1) = 1 (S ⊗ S) ◦ ∆ = τ ◦ ∆ ◦ S (1.10)

where a, b ∈ H and τ : H ⊗ H → H ⊗ H is the flip map defined by

1.3 From κ-Poincaré to κ-Minkowski

[Mi , kj ] = iϵijl kl , [Mi , k0 ] = 0, (1.12)

[Mi , Mj ] = iϵijk Mk , [Mi , Nj ] = iϵijk Nk [Ni , Nj ] = −iϵijk Mk . (1.14)

The co-algebra sector is given by

∆ki = ki ⊗ 1 + e−k0 /κ ⊗ ki , ∆k0 = k0 ⊗ 1 + 1 ⊗ k0 , (1.15)

and finally the antipode sector is given by

S(k0 ) = −k0 , S(ki ) = −ki ek0 /κ , (1.18)

have the definitions (37)

kµ ▷ xν = −iδµν , Ni ▷ x̂j = ix̂0 δij , Ni ▷ x̂0 = ixi , kµ ▷ 1 = ϵ(kµ ) = 0

Furthermore, one can also obtain the relation

One can therefore deduce

ki ▷ x̂j x̂k − ki ▷ x̂k x̂j = 0 (1.47)

−P02 + P2 + P42 = κ2 P+ := P0 + P4 > 0 P4 > 0. (1.56)

[Mi , Pj ] = iϵijk Pk , [Mi , P0 ] = 0, [Ni , Pj ] = iδij P0 , [Ni , P0 ] = iPi , (1.57)

[Mi , Mj ] = iϵijk Mk , [Mi , Nj ] = iϵijk Nk , [Ni , Nj ] = iϵijk Mk . (1.58)

The co-algebra sector is given by

and the antipodes by

1.4 From κ-Minkowski to κ-Poincaré

Similarly one can check that

Multiplying these two matrices we get

êP (k) O = (P0 , P, P4 )T (1.85)

which allows us to define

ê∗k = êk z (1.87)

S(P ∗ ) = −S(P ) (1.91)

1 (∗) (∗) (∗) P(∗) Q(∗) κ (∗)

1 (∗) (∗) (∗) P(∗) Q(∗) κ (∗)

∆kµ (|p⟩ ⊗ |q⟩) = (p ⊕ q)µ |p⟩ ⊗ |q⟩ (1.95)

where p ⊕ q is define through the group multiplication

êp êq = êp⊕q . (1.96)

which means that since because of eq. (1.48) we have

Using this relation, we get

Furthermore, we can also recover the inverse group element ê−1

We now consider the bicrossproduct coordinate functions on AN (3) defined as (75)

(f1 , f2 )(g) = f1 (g)f2 (g) I(g) = 1 (1.107)