Critique of the "photon" theory of electromagnetic radiation

Short version:

Interference between two microwave transmitters cannot be
explained with the "photon" theory of electromagnetic
radiation (emr). This theory cannot be a true
representation of Nature, and so should not be used to
evaluate observations or other theories.

Long version:

As best I can tell, the generally accepted definition of a
photon involves:

1 - The photon begins when quantum of energy is lost,
somewhere, in one specific location - such as an
electron dropping an energy level in an atom.
(Actually, I think conventional photon theory may
allow for probability distributions for when and where
the photon began. Also, a highly coherent photon
involves a long time period for the long series of
sinusoidal oscillations to be transmitted.)

2 - The resulting photon contains that exact quantum of
energy. Therefore, all electromagnetic radiation is

3 - The photon propagates as electromagnetic radiation
with a frequency proportional to the quantum of
energy it carries.

4 - Somewhere, the photon delivers this quantum of energy
- typically to a single electron. That's the end of
the story - the photon no longer exists. ("Collapse
of the wave-function" etc.)

5 - So the photon begins at one point and ends at another.

6 - The relationship in time and space between the
starting point and the end point makes it clear that
the photon operates as a wave as it gets from one
place to another.

7 - Photons do not interfere with each other - but the
wavefunction of an individual wavefunction does
interfere with itself. Consequently, the
probabilities of where photons will deliver their
energy in any experiment are not affected by the
number of photons in the system. (Except in non-linear

8 - Photons are real things - not just a handy theoretical
construct. They are the means by which the
electromagnetic force gets from place to place. This
is for changing and oscillating electromagnetic forces
and for "static" forces such as a magnet sticking to
a fridge, electrons on one plate of a capacitor
affecting electrons on the other plate (there's no
fundamental difference between this and sending light
or gamma rays across the Universe for billions of
years) and for all radio, optical, X-ray etc.
interactions. Consequently any proposed mechanism,
such as for non-Doppler redshift, can be discounted if
it is incompatible with what we think we know about

9 - Photons are massless "particles".

10 - A photon carries momentum - in proportion to its

There may well be other common beliefs about "photons",
such as to do with their polarization, spin etc. Can
anyone point to a similarly detailed list of what most
people understand by the term "photon"?

There are a number of other peripheral and consequent
beliefs about photons. For instance, the question of how
the photon "knows" where it will land - perhaps at the
time it is first radiated. This is related to the
perplexing question, photons or not, of momentum deposition
at the source of isotropic light when we believe that in
the whole system a single quantum of energy has been
received at some specific location, quite some time after
it must have emanated "isotropically" from the source.

I understand that the concept of a photon - not by name
originally - came from Einstein. It has been supported by
many greats of the field, not least Richard Feynman. I
understand it is an attractive concept when dealing with
X-rays and emr where we can detect or imply an individual
quantum of energy is involved and/or measure the momentum
of an emitted, absorbed or scattered "photon". But this
doesn't mean that the "photon" model accurately represents

Critics include Willis Lamb, who wrote a paper
"Anti-Photon" in 1995, which I can't find on the Net or
at the journal site. He wrote, in his most recent paper
(2001 - Super classical quantum mechanics: The best
interpretation of nonrelativistic quantum mechanics):

> In case you don't know it already, despite what many
> other people think, THERE ARE NO PHOTONS OF A

(Lamb's capitalisation.)

See also the talking-to he gave the delegates to the 1995
US Army workshop on quantum cryptography and computing:

I expect any emr theory such at that of "photons" to either
apply completely from static situations to gamma rays
(DC to daylight and beyond) or to include some specific
limits beyond which it does not apply, and to point to some
other approach should be used instead. I have never heard
of such a limit, and when I believed in "photons", I used
to think that an electron and a nucleus, or my fridge and
one of its magnets, were continually exchanging "photons".

Before I make my critique, I want to point to some other
possible ways in which the "photon" concept could be shown
to be inadequate.

There are many critiques to be made of "photons" in such
low frequency or static situations. There are critiques
to be made in the three polarizer paradox - but photon
proponents get around that by saying a photon at one
polarization can be considered to have a superposition of
polarizations at a range of angles (except 90 degrees),
which the middle polarizer can select from to produce a
polarization which in turn has superposed polarization
states which pass the third polarizer. For instance:

Polarized Light and Quantum Mechanics: An Optical
Analog of the Stern=96Gerlach Experiment
Joseph M. Brom,=86 and Frank Rioux=87 2002

It was my understanding that most people consider that a
polarizing sheet can only absorb a "photon" - that the
sheet can't change the photon's polarization. But if
"photons" are considered to have a superposition of all
polarisation states except those at 90 degrees to their
official state, then I think the "photon" concept can be
made to do pretty much whatever people want. I think this
adds no explanatory or predictive power to the classical
view of emr polarisation - and is far less elegant.

My critique is on familiar ground: interference. I use
two-slit interference, but the same arguments hold for most
system of lenses, mirrors, multiple-slit interference,
holograms, zone-plates etc.

A two-slit interference experiment involves a monochromatic
wave emanating to a barrier and a fraction of the wave
passing through the two slits in the barrier - with a
measuring arrangement for the resulting deposition of
quanta at a distant screen.

Open slit A and we get X quanta per second per unit area
in the part of the screen where the diffracted light from
this slit illuminates the screen reasonably evenly. Close
slit A and open slit B and we get the same pattern, offset
an insignificant amount due to the distance between the

Open both slits and we get overall, twice the number of
quanta per second per unit area at the screen, but the
resulting interference pattern causes some locations to
receive 0 quanta and others to receive 4 times the number
they received with a single slit.

Since we know the quantum of energy which is always
deposited for this particular frequency, we can turn down
the light-source to the point where most of the time there
can't be more than one "photon" in the system - and we
observe the same statistics about where the quanta are
deposited. Therefore, we conclude that if "photons" exist,
then each photon's wave properties involve interference
with itself, and do not seem to be affected by the presence
of other "photons".

With this traditional experiment, there is a single source
of light, and we say that this is the source of all the
photons. The photon model correctly describes the
observations - because there is a single source and a
single end-point for each photon, where the statistics of
the locations of of the endpoints seem to be controlled
by each photon's wave-like properties.

Since photon theory has no explicit frequency or wavelength
limits, we should be able to apply it to radio waves too.

Lets say I have a 10GHz oscillator, with a small antenna.
This produces emr with a 3cm wavelength. According to the
photon theory, this emr is composed of vast numbers of
individual photons, each carrying a small quantum of
energy. 1.243 micron emr deposits quanta of 1 electron
volt, so each 10 GHz "photon" carries 41.4 micro electron
volts of energy. We can't detect such a small individual
quantum of deposited energy, but all the evidence leads us
to believe that emr of this frequency can, or will always,
deposit individual quanta of this energy. According to
the photon theory, each such quanta of deposited energy
arises from a similar quantum which is lost in the
transmit antenna.

At a distant "screen" or observing line, perpendicular to
the path from the transmit antenna, I have a 10GHz receiver
which I can use to measure energy deposition at various
points line. The receiver could be an electronic amplifier
or it could work by detecting temperature changes in an
absorbing screen due to the deposition of quanta.

In the central area of the screen, I observe an energy
density - which I reasonably conclude means X 41.4ueV
quanta per second per unit area.

Now I invent a second identical transmitter and place its
antenna close to the first one. These transmitters have
independent power supplies and oscillators, but they both
run at 10GHz, with an accuracy of (for instance) 0.01 Hz.
I think this accuracy is practical with a phase-locked
loop from an atomic frequency standard.

As with the two-slit experiment, the overall energy
deposition ( =3D number of quanta per second) doubles at the
detecting screen. Likewise, the number of quanta per
second goes down to 0 in some locations and up to 4 times
the single transmitter value in others.

According to my understanding of "photons", this experiment
does not allow a "photon" to interfere with itself, since
a photon can only arise from one transmit antenna or the
other. Yet we observe perfect interference. So I say
this proves that the "photon" theory is wrong. The theory
could only account for the observations if it was re-jigged
to allow a photon to arise simultaneously from two places
at once, or to allow two photons to interfere with each

The second possibility is excluded by turning down the
transmit power to the point where we reasonably conclude
that there is usually no more than one "photon" in the
system at any given time. We still see the interference
pattern. So the only way a "photon" theory could describe
this is to modify it to allow a "photon" to arise from two
or more places at the same time. This would leads to
problems in deciding how much energy was lost at those
sites and would result in a theory very different from the
current one.

Lets tweak the experiment. Firstly, lets imagine that one
transmitter is running 1Hz faster than the other. We
would still observe interference, but the pattern would be
moving by one fringe per second. Only when the detuning
and therefore the fringe movement got so high as to prevent
us from detecting the interference pattern would we say
that there might be no interference. However, this is a
purely practical matter. If we had a 10 GHz transmitter
interfering with an 11 GHz transmitter and a set of
receivers we could move fast enough (or create a suitably
moving detection pattern) to keep up with the billion
fringes per second movement of the pattern, we would be
able to detect it just as well.

Secondly, lets increase the frequency of the transmitters.

I am not sure how fast electronics go these days, but I
guess up to frequencies of 100 GHz or so - 3mm wavelength.
This experiment could be made to work fine at those
frequencies too. Now the quanta being deposited are 4.14
milli electron volts - but we probably still don't have a
way of detecting them individually.

We can detect near visible infra-red quanta - with
photographic film and silicon photodiodes - but this
involves frequencies where we need lasers, rather than
electronic oscillators, as transmitters.

I imagine it is possible to lock the frequency of a laser
to an electronic reference. Maybe it would be possible to
have lasers interfering detectably when they are locked to
a single lower frequency clock, or to two clocks derived
from two independent atomic references.

Two lasers synchronised from a third would interfere fine,
but then we get into arguments about a single source of a
"photon" and how this leads to the stimulated emission of
other identical "photons".

So far, we have considered monochromatic sinusoidal emr.
I think the two microwave transmitter experiment shows
that the photon theory is invalid, because the theory
insists that each photon must originate from a single
quantum at a single location.

Here are some other critiques.

Photon's can't be "massless". Saying they are "massless at
rest" is pointless, since they cannot be at rest.

We have a sealed tank sitting on some scales and inside the
tank is a battery operated xenon, laser or LED flash
arrangement which puts out a bunch of light much shorter
than the length of the tank. Chemical energy in the
battery (in the form of electromagnetic energy - perhaps
"photons" - in the bonds and electron energy levels) is
converted into electronic energy in a capacitor (more em
energy, perhaps "photons") and then into light which
traverses (as emr, perhaps as "photons") from the left
end of the tank to the right end, where it is absorbed,
heating the material there (kinetic energy of atomic
movement and "photons" of greater forces as these atoms
jostle each other more).

While the light is in flight, it is temporarily decoupled
from the body of the tank, but as soon as it is absorbed,
the gravitational effects on that light would result in
it falling downwards, leading to a vertical force when
the light is absorbed by the tank. So the overall
weight/mass of the tank would not change, although there
might be a temporary loss and then a short-term gain as
deviations from average, just as if someone had thrown a
ball from one end of the tank to the other. We believe
that while the light was in transit, the battery and
capacitor had a lower mass then before it was emitted,
and the receiving wall on the right had a lower mass than
after the light was absorbed. No energy or matter has
escaped from the tank - so the only reasonable
explanation is that the light has mass.

So how can anyone maintain that light or other emr has no
mass? It doesn't matter whether or not the emr is is
considered to be photons. My understanding of the photon
theory is that "photons" are the sole mediating mechanism
for the electromagnetic force, so all chemical energy, and
all the energy of a spring etc. is composed entirely of
electromagnetic forces, in the form of "photons". A taut
spring or a charged battery has more energy and as far as
we know (it may not be measurable) it has a
correspondingly greater mass. So the idea that "photons"
have no mass is provably wrong. (I believe that if a beam
of light is shone past the Sun and so is bent by the Sun's
gravity, then the Sun itself is attracted and slightly
moved towards the beam.)

I understand that the traditional view of photons is that
each one has a specific wavelength / frequency - and
therefore quantum of energy. Are there any exceptions to
this view? Do some theories involving photons allow for
a probability distribution for the photon's wavelength?

My view is that various processes, including quite often
the loss of a specific quantum of energy, give rise to
electromagnetic radiation. Then, this emr is subject to a
number of processes as it travels in vacuum or other
media. It can spread out and so become less concentrated.
It can be focussed with a lens, mirror or diffraction
device such as a zone-plate. It can be split into two
beams by passing into a block of material with a different
refractive index. It can be partially absorbed. It can
be divided and parts of delayed in time and then later
mixed together with other time-delayed parts of the emr.
It can be distorted in the electrostatic / magnetic
dimension. It can add and subtract from other sources of
emr. (There are polarization processes too.)

If we consider emr which arises from a specific quantum of
energy, a-la photon, it seems to me that this emr carries
with it either the whole amount of energy and momentum, or
the probability of this energy and momentum. But when the
emr is mixed with other emr the resulting patterns of
energy deposition depend on all the relevant sources
and how they interfere at each point which could absorb

We observe that there is a direct relationship between the
frequency spectrum of the emr and the distribution of
energies in the quanta it deposits. A sine-wave leads to
only one energy level. A single arbitrary shaped pulse
creates a particular and easily predicted probability
distribution for the relationship between quanta energy
levels and the number of such quanta deposited.

So I think the quanta which may have given rise to the emr
do not necessarily have anything to do directly with the
quanta which are deposited.

We know that two very low energy independent microwave
transmitters result in quanta being deposited in some
places at 4 times the rate which results from a single

Just because we usually observe things by amplification
mechanisms which round off energies to exact quanta doesn't
mean that the real nature of emr is also quantitised.
Maybe it just transfers probability from one place to
another, and this probability carries, on average and in
general the momentum and energy. I can't see any other
explanation. Sine the source of the emr can be destroyed
and the emr be in transit well before its destination
absorber has been brought into existence, the energy and
momentum clearly resides solely in the emr. In the case of
a short impulse with spectral qualities the same as black-
body radiation from the Sun, the whole wavefront may be
only a few microns deep. This carries the energy and
momentum, and it exists independent of anything else, in
this small area of space, moving at c.

Maybe when the coupling between the source and all possible
destinations is very close (a small fraction of the
wavelength - which covers all "static" electromagnetic
situations, in which the wavelength is "infinite") there is
a direct 1:1 relationship between a quanta being lost and
the quanta being deposited. But in situations like cyclic
or pulse emr, the quanta (or multiple quanta) loses its
energy to space and this radiates away, carrying a
probability of energy and momentum. These emr waves can
interfere with the emr produced by the loss of other
quanta and result in quanta being deposited which result
from combination of the probabilities / energies /
momenta of the various emr waves interfering at that spot.

So how does a really dim emr signal (from one source or
from the interference of multiple sources) occasionally
result in the deposition of a full-blooded quantum of
energy, when as far as we know it is the same dim signal
which has been illuminating that atom, electron etc. for a
long time in the same way without any such deposition? Why
can we expose the film for a very short time indeed, such
as 0.001 second, and occasionally get (as far as we know) a
full-blooded quanta which changes an electron's energy in
an atom in a silver iodide crystal - even when on average
such a quanta would be deposited only every 1000 seconds?

Stochastic Electrodynamics (SED) invokes a "vacuum energy"
to explain this. I am still trying to understand QM but it
seems that some conventional QM theories involve vacuum
energy too. This has its problems, but it seems like a
better approach than my understanding of conventional
photon-based quantum mechanics: that all emr consists of
independent, totally quantitised "photons" and that each
photon spreads out potentially over the entire Universe,
and magically disappears everywhere else once it deposits
is singular quantum of energy somewhere, sometime, in
exactly one spot.

I see no problem with classical emr and interference
doubling the amplitude of two signals in some locations,
leading to twice the current at twice the voltage into the
load, resulting in quadrupling the energy deposition. The
challenge is to reconcile this this with the observation
that emr of a particular frequency always (as far as we
know) results in deposition of specific quanta of energy.

Since particles (electrons, neutrons, helium atoms and
now Buckyballs) exhibit spatial patterns in their
probability functions just like emr (as evidenced by
interference experiments) - and (I understand) with exactly
the same relationship between the spacing of these
patterns in a vacuum and their momentum I suspect this
"frequency ~ energy-level ~ momentum" stuff is a property
of space, not of matter or electromagnetism.

- Robin