Modanese - Inertial Mass and Vacuum Fluctuations in Quantum Field Theory (2003), Energy from the Vacuum
[ Pobierz całość w formacie PDF ]
Inertial Mass and Vacuum Fluctuations
in Quantum Field Theory
Giovanni Modanese
California Institute for Physics and Astrophysics
366 Cambridge Ave., Palo Alto, CA 94306
and
University of Bolzano – Industrial Engineering
Via Sernesi 1, 39100 Bolzano, Italy
E-mail address: giovanni.modanese@unibz.it
Abstract
Motivated by recent works on the origin of inertial mass, we revisit the relationship
between the mass of charged particles and zero-point electromagnetic elds. To this
end we rst introduce a simple model comprising a scalar eld coupled to stochastic
or thermal electromagnetic elds. Then we check if it is possible to start from a
zero bare mass in the renormalization process and express the nite physical mass in
terms of a cut-o. In scalar QED this is indeed possible, except for the problem that
all conceivable cut-os correspond to very large masses. For spin-1/2 particles (QED
1
with fermions) the relation between bare mass and renormalized mass is compatible
with the observed electron mass and with a nite cut-o, but only if the bare mass
is not zero; for any value of the cut-o the radiative correction is very small.
PACS: 03.20.+i; 03.50.-k; 03.65.-w; 03.70.+k 95.30 Sf
Key-words: Inertial mass, Quantum Electrodynamics, Radiative corrections
In modern physics each elementary particle is characterised by a few parameters which
dene essentially its symmetry properties. Mass and spin dene the behavior of the particle
wavefunction with respect to spacetime (Poincare) transformations; electric charge, barion
or lepton number etc. dene its behavior with respect to gauge transformations. These
same parameters also determine the (gravitational or gauge) interactions of the particle.
Unlike spin and charge, mass is a continuous parameter which spans several magni-
tude orders in a table of the known elementary particles. In spite of several attempts, there
is no generally accepted way of expressing these masses, or at least their scale, in terms
of fundamental constants. In the standard model particles acquire a mass thanks to the
Higgs eld, but the reproduction of the observed spectrum is only possible by choosing a
dierent coupling for each particle.
Inertia in itself is not really explained by quantum eld theory; rather, it is incor-
porated in its formalism as an automatic consequence of the spacetime invariance of the
classical Lagrangians. In turn, these Lagrangians are essentially a generalization of New-
tonian dynamics. In the equations for quantum elds, like in the wave equations for single
particles or in their classical limits, mass appears as a free parameter which can take zero
or positive values.
Therefore it is not surprising that several works in the last years (for a discussion
and a list of references see for instance [
) have been devoted to the search of a possible
fundamental explanation of the inertial properties of matter. Some of these works look for
the source of inertia in the interaction between charged particles and the electromagnetic
vacuum uctuations, exploring analogies with the dynamical Casimir forces on an acceler-
ated cavity [
or with the unbalanced radiation pressure in the Davies-Unruh thermal bath
seen in accelerated frames [
. The possibility was also investigated, in connection with
astrophysical problems, that Newton law does not hold true for very small acceleration [
.
In this work, we try to clarify whether some of the proposals contained in the men-
tioned papers can be implemented, or at least partially analysed, within the standard
formalism of quantum eld theory–perhaps leading to a more satisfactory inclusion of the
concept of mass. Of course, the idea of dynamical mass generation induced by vacuum
uctuations is already familiar in quantum eld theory [
, but it is usually connected to
phenomena of spontaneous symmetry breaking, where a quantum eld acquires a non-zero
vacuum expectation value. Here, on the other hand, we are interested only into the eects
of the uctuations.
One should also keep in mind that the mass of a particle can come into play, in
quantum eld theory, in dierent equivalent forms, namely as (a) the response to the
coupling with an external eld; (b) a parameter in the dispersion relation
E
(
k
); (c) the
pole in the particle propagator and in its creation/annihilation cross section.
2
The pragmatic attitude of quantum eld theory towards the origin of mass curiously
seems to disappear only at one point, namely when in the mass renormalization procedure
the “bare” mass
m
0
is assumed to be innite. What happens if we introduce nite cut-
os in the eld theoretical expressions for the radiatively induced mass shift
, and set
m
0
= 0? One nds that the result depends much on the spin of the particles. For scalar
particles, it is possible to introduce a cut-o in
, set the bare mass to zero and interpretate
somehow the physical mass as entirely due to vacuum uctuations–except for the problem
that the “natural” cut-os admitted in quantum eld theory (supersymmetry scale, GUT
scale, Planck scale) all correspond to very large masses. For spin-1/2 particles (QED
with fermions) one obtains a relation between bare mass and renormalized mass which is
compatible with the observed electron mass and with a nite cut-o, but only if the bare
mass is not zero. Below we shall give the explicit expressions for the scalar and spinor case.
Before that, however, it is useful to consider a semiclassical approximation, which turns
out to contain much of the physics of the problem.
Let us consider charged particles with bare mass
m
0
immersed in a thermal or stochas-
tic background
A
µ
(
x
). For scalar particles described by a quantum eld
, the Lagrangian
density is of the form
L
=
2
(
P
µ
−
eA
µ
)
(
P
µ
−
eA
µ
)−
1
2
m
0
|
|
2
(1)
and contains a term
e
2
A
µ
A
µ
, which after averaging on
A
µ
can be regarded as a mass
term for the eld
. Take, for instance, the Coulomb gauge: The eective squared mass
turns out to be equal to
m
2
=
m
0
+
e
2
h|A(x)|
2
i.
For homogeneous black body radiation at a given temperature
T
, the average is
readily computed. One has
h|A|
2
i=
Z
1
d!
u
!
!
2
(2)
0
k
B
T
(the constant is adimensional and of order
1). This mass shift can be signicant in a hot plasma, but only for spin-zero particles, not
for fermions. In fact, the Dirac Lagrangian is linear with respect to the eld
A
µ
, therefore
it is impossible to obtain a mass term for spinors by averaging over the electromagnetic
eld. One expects that a mass shift for fermions will only appear at one-loop order.
This is in fact conrmed by the full calculation in thermal quantum eld theory [
and
by experimental evidence (no relevant mass shifts are observed in the Sun). Note that
although a second-order formalism for Dirac fermions in QED exists, it has been used until
now for calculations with internal fermion lines only
. The result above seems to conrm
that a proper treatment of on-shell fermions intrinsically requires a rst-order lagrangian.
Eq. (
can also be applied to the Lorentz-invariant frequency spectrum of the zero-
point eld in Stochastic Electrodynamics, namely
u
!
=
!
3
[
. In this case the integral
diverges, unless we introduce either a cut-o, or a resonant coupling of the zero-point eld
to the particle at a certain frequency
!
0
–which therefore denes the mass of the particle
. This could be viewed as an alternative to mass generation by coupling to the Higgs,
but, again, only for scalar particles.
Turning now to scalar QED, one can consider the Feynman mass renormalization
condition
m
0
+
(
m
2
) =
m
2
, set
m
0
= 0, impose a physical cut-o
M
in
and compute
m
as a function of
M
. One nds in this way, as mentioned, that
m
is of the order of the
cut-o. Actually, scalar QED only describes particles like pions or other charged mesons
which are not regarded as fundamental. Better known and physically more relevant is
=
const.
p
3
1
where
u
!
is the Planck spectral energy density. By integrating one nds that the squared
mass shift is given by
m
2
spinor QED, i.e. the quantum electrodynamics of spin 1
/
2 charged particles. Perturbative
expansions in spinor QED lead to some divergences in the radiative corrections, but such
divergences are usually mild ones. Starting in the Sixties, the limit in which spinors
have zero bare mass has been studied in great detail, and some general theorems were
proven. Novel infrared divergences appear in this limit, but as rst shown in
, under
certain physical conditions all transition matrix elements are nite. The zero mass limit is
furthermore important because it corresponds to the limit in which the charged particles
are not massless, but interaction energies are very large compared to the mass scale [
.
This ultraviolet limit is governed by Weinberg theorem [
, which predicts the behaviour
of transition amplitudes in the limit of large external momenta.
The authors of [
wondered if charged particles can have zero mass. From the clas-
sical point of view this looks impossible, because the energy of the electric eld generated
by a particle bears a mass–the well known electromagnetic mass of the electron. In the
quantum theory, however, this is not obvious a priori. An analysis of the renormalization of
radiative corrections to the self mass is required. The rst step in this direction was taken
by the authors of [
, who investigated whether, more generally, a nite electron mass
could be generated by radiative corrections starting from zero bare mass and found that a
full radiatively-generated mass was possible, provided the photon wave function renormal-
ization constant
Z
3
was nite. Later, they established the fact that
Z
3
has, perturbatively,
logarithmic divergences to all orders [
. They concluded that the only way to arrive at
a nite
Z
3
is to make the prefactor of this logarithm vanish, i.e. in modern language, to
nd a nontrivial zero of the QED beta function. Up to now, however, there are no hints
of the existence of such a zero.
In the modern picture of elementary particles the historical results of spinor QED
are generalized to include those of unied electroweak theory and, in principle, the strong
interactions for hadrons (which, however, do not admit a perturbative treatment except in
special cases). Disregarding quark-gluons interactions and without going into details which
fall outside the scope of this work, we just observe here that the masses of all fermions in
the electroweak theory are subjected to radiative corrections due to the electromagnetic
eld and to the eld of the W and Z vector bosons. Being vector bosons massive, their
radiative corrections are far less divergent; in particular, they are compatible with zero
mass neutrinos.
All the above refers to quantum eld theory renormalized in the usual sense, i.e. in
the limit where all energy cut-os tend to innity. This approach has been highly successful
in explaining the phenomenology of elementary particles. Renormalizability of the theory
implies that phenomena occurring at very high energies do not aect results at the energy
scale of interest. In matters of principle, however–like the origin of inertia considered in
this paper–one can take a dierent attitude. One can admit that an intrinsic high-energy
cut-o for quantum eld theory exists, due to some more general high energy theory or to
quantum gravity (the Planck scale). It is therefore interesting to see what QED predicts in
this case. Let us consider the simplest radiative correction to the self mass, namely the one-
loop self energy diagram, and the corresponding expression for the mass renormalization
condition, that is
h
(
p
2
, M
)
3
4
ln
M
2
1
2
!
p
m
−
m
0
=
=
m
0
+
(3)
p
2
=m
m
0
Inserting
m
0
= 0 we nd that the radiatively corrected mass
m
is zero, too. A charged
massless particle could then exist, under the assumption of a nite cut-o, thanks to the
mild logarithmic divergence. If we set
m
0
>
0 instead, we obtain a radiative correction
4
i
k >
1 (i.e., vacuum uctuations
increase the mass by a factor
k
), solve eq. (
for
and plot the inverse function
k
(
) (Fig.
1). We see that even for very large values of the cut-o, the renormalization eect is quite
moderate.
In conclusion, in the framework of quantum eld theory it is impossible to interpret
the observed mass of the electron as due to radiative corrections. Namely
(1) in the renormalized theory, the radiative corrections to mass are divergent if the
bare mass is zero;
(2) in the presence of a large but nite energy cut-o, the one-loop radiative correction
vanishes for zero mass and is otherwise irrelevant.
Acknowledgments - This work was supported in part by the California Institute for
Physics and Astrophysics via grant CIPA-MG7099. I am grateful to V. Hushwater, V.
Savchenko and A. Rueda for useful discussions. I also thank C. Schubert for bringing
to my attention ref.s [
,
and their meaning in terms of the renormalization group
equation.
References
[1] H. C. Rosu, “Classical and quantum inertia: a matter of principles”, Grav. Cosm. 5
(1999) 81.
[2] M.-T. Jaekel and S. Reynaud, “Quantum uctuations and inertia”, in Electron theory
and Quantum Electrodynamics: 100 Years later, ed. J.P. Dowling (Plenum Press, New
York, 1997), p. 55. H. C. Rosu, “Relativistic quantum eld inertia and vacuum eld
noise spectra”, Int. J. Theor. Phys. 39 (2000) 285.
[3] B. Haisch and A. Rueda, “Inertia as a reaction of the vacuum to accelerated motion”,
Phys. Lett. A 240 (1998) 115; “Contribution to inertial mass by reaction of the
vacuum to accelerated motion”, Found. Phys. 28 (1998) 1057.
[4] M. Milgrom, “Dynamics with a non-standard inertia-acceleration relation: an alter-
native to dark matter in galactic systems”, Ann. Phys. 229 (1994) 384.
[5] S. Coleman and E. Weinberg, “Radiative corrections as the origin of spontaneous
symmetry breaking”, Phys. Rev. D 7 (1973) 1888.
[6] M. Le Bellac, Thermal eld theory (Cambridge Univ. Press, Cambridge, 1996).
[7] A. Morgan, “Second order fermions in gauge theories”, Phys. Lett. B 351 (1995) 249.
[8] P.W. Milonni, The quantum vacuum (Academic Press, New York, 1994).
[9] Y. Dobyns, B. Haisch, and A. Rueda, “Inertial mass and the quantum vacuum elds”,
Annalen Phys. 10 (2001) 393.
[10] T.D. Lee and M. Nauenberg, “Degenerate systems and mass singularities”, Phys. Rev.
133 (1964) B1549.
5
which even for very large values of the cut-o is much smaller than the electron scale. In
fact, let us parametrize the cut-o as
M
= 10
, set
m/m
0
[ Pobierz całość w formacie PDF ]