The Nature of Light

Caroline H Thompson

Web site: http://www.aber.ac.uk/~cat

17:2:99

Introduction

I present here some of my thoughts on the nature of light. They seem to me to explain several "quantum mysteries" -- in particular, the apparent "action at a distance" that is supposed to have been actually demonstrated in "Bell test" experiments -- and to eliminate unnecessary ideas such as uncertainty as to the direction of time, associated with modern ways of teaching classical electromagnetism. (See my Web site for my various papers on "EPR" or "Bell test" experiments.)

I am aware that they are speculative. I know that in places I have may have laid myself open to ridicule. This is inevitable, as I am working on my own, apart from very valued exchanges over the internet. One person, learning mainly by studying published papers, whose meaning and real achievements are rarely transparent, cannot know everything in such a wide field. Something that is quite clear, though, is that the formalisms of quantum theory (QT) and relativity theory have, for the best part of a century now, been acting as straitjackets, preventing real understanding. Some of the concepts deserved to last this long, but not the particular formalisms, devised to account in clumsy, incomplete ways for phenomena whose detailed nature was but dimly perceived, whose complexities were yet to be revealed. As Werner Heisenberg, one of the Founding Fathers of quantum theory implied, Niels Bohr's version of QT was not designed to last for ever:

" ... the taboo [on asking about the true nature of things] need not really upset us. There will always be young people enough to think about the wider context, if only because they want to be absolutely honest in all things." (Heisenberg, "Physics and Beyond", George Allen and Unwin, 1971, p246.)

The problem of the 20^th century has been that fundamental physics has become so large, in terms of the human and financial resources tied up in it. An enormous resistance has evolved to new ideas, for these are developed by individuals and small groups, who are inevitably a tiny minority. Proponents of the ruling theories have acquired almost absolute power, controlling the text books, the funding, even, to a much larger extent than is healthy, the media. They have the power to destroy the livelihood of the "young person" who steps out of line. For myself, though, I have no livelihood. So long as I speak only the truth, without libel, the freedom of thought that we are all supposed to share is for me a reality!

Against the "Photon"

How did we come to think of light as "photons"? A lot more historical research to really find out, but I suspect that it went something like this.

In the 19^th century, nobody seriously questioned the purely wave nature of light. Newton's corpuscle ideas had been firmly banished after Maxwell and his contemporaries rejected them. Perhaps it was Hertz and the discovery of radio waves that clinched the argument. It seemed quite easy to account for them by assuming the fields associated with oscillating electric currents propagated outwards at the speed of light, and that a radio receiver was able to resonate to the oscillations. How could there be quantisation here? Maxwell had shown that the behaviour of light was consistent with the idea that it was just a shorter-wavelength kind of radio wave, so why should it come in particles?

History reveals that there was at least one person who hung out till the end against the idea: Hendrik Lorentz, a great but, these days, little known Dutch physicist. As it happens, his book, "Problems of Modern Physics" (Dover, 1927, based on a series of lectures delivered in Berkeley, California), was the starting-point of my current interest in physics. I read it in 1993. It seems that not many of the problems have yet been solved.

One of the crucial problems he covered was the matter of the nature of radiation. Lorentz said:

p48: " ... we can certainly say that the waves with which we operate in wireless telegraphy are produced by rapidly alternating currents in conductors ... therefore if it is true that a current in a metal consists in a motion of negative electrons ... there is a case of radiation by moving electrons."

Continuity, not quantisation, seems to be the name of the game here.

He discussed the photoelectric effect, and its apparent conflict with Fresnel's interference experiments. Fresnel's light and dark fringes were best interpreted as showing light interfering in just the same kind of way as water waves, whilst the photoelectric effect was interpreted as showing it behaving as particles, with individual "photons" donating their energy to electrons. Lorentz gave no solution. Possibly it would have helped him to have considered the presence of the "zero point field" (ZPF) of Stochastic Electrodynamics (SED) (See Trevor Marshall's ideas at http://homepages.tesco.net/~trevor.marshall and Luis de la Peña and Ana Maria Cetto's inspirational book, "The Quantum Dice: an introduction to Stochastic Electrodynamics", Kluwer, 1996). It seems that this really does exist. In a way, it is the same thing as the "quantum vacuum fluctuations" that feature in many quantum theory explanations these days. Under SED, the ZPF is taken as having intensity about half that of a "photon" at any given frequency. Its existence means that what a "photomultiplier" (a light detector) is really measuring is never just the stream of "photons" we thought we were looking at, as the ZPF is always added. This fact has never been obvious as the instruments only detect a small fraction of the incoming signals anyway. How do we know what they really do?

Personally, I have never felt the photoelectric effect to be evidence of quantisation at all. I have always thought it was just the way our instruments (or photosensitive substances) worked, producing effects that we could measure only when circumstances were right and some threshold happened to be exceeded. ( Now that I know more, I am even more convinced. See later.) Mind you, I am aware of the fact that high-energy light does seem sometimes to come in strong pulses, each reliably able to cause a "detection". These strong pulses were presumably one of the reasons for the initial acceptance of the photon idea, but I propose to concentrate on low-energy light, as this is what is used in the EPR experiments, the area I have studied most intensely. (Incidentally, the same kind of qualitative difference is seen with sound: high energy "ultra-sound" has different properties to ordinary sound, for example being able to form narrow pencil-like beams.)

But let us return for minute to history.

We have Planck, and his famous explanation of the "black body spectrum", in which the idea of quantisation appeared, as I understand, for the first time. Now Planck himself always thought of light as purely wave. It was not he who made the leap from the mathematical need for the assumption of quantisation to the existence of indivisible "particles", each with energy proportional to frequency. This astounding step was due to an imaginative young man named Einstein. And at the time there as very strong resistance to it. As I discovered only recently, in a small book by John Hendry ("The Creation of Quantum Mechanics and the Bohr-Pauli Dialogue", first published in 1952), there was much correspondence between Bohr and Einstein, and Bohr repeatedly made statements such as:

"[The hypothesis of light quanta] presents insuperable difficulties when applies to the explanation of the phenomena of interference ... [it] excludes in principle the possibility of a rational definition of the conception of a frequency ..." (1921 Solvay congress)

And Lorentz, in that 1927 book, reports some fascinating facts. For when you calculate the radiation from an oscillating dipole, what you actually assume is that the electron emits spherical waves continuously. These are necessarily longitudinal (for some reason or other, they would be described today as "scalar"). From the patterns in these waves, allowing for interference, you can deduce the form of the resultant emitted "radiation" -- the transverse pattern associated with the emission. There were problems -- I must re-read the book -- but they seem to me to be not too important.

The point is that, mathematically, transverse patterns can arise from longitudinal ones, and I can see no good reason why this particular bit of mathematics should not be truly representing reality. And what this is leading up to is the fact that this same mathematics forms, I believe, the basis for the QT prediction for the emission from an oscillating dipole. Only now, instead of a pattern of varying characteristics in the different directions, we are supposed to believe that photons are emitted with varying probabilities! (This is the usual way in which QT is linked to classical theory: whenever classical ideas predict a high intensity, QT predicts high probability of finding a photon.)

And the QT picture is strange (Maxwell wave theory as taught these days is no better in this respect!). It considers only the transverse component and ignores the longitudinal one. I have an idea that when it comes to similar reasoning with relation to synchrotron radiation, this causes an absurdity. For the QT accelerating electron should not radiate in the forwards direction, and in order to explain the fact that it does theorists have to plunge into the murky waters of Special Relativity. I know I am on dangerous territory here. I do not know nearly enough facts. Yet I think it is worth voicing my suspicions in the hope that somebody will investigate them. If I am right, it could mean an enormous simplification in our view of things.

To return to safer ground, let us look at the optical "Bell tests" (Einstein-Podolsky-Rosen, or EPR tests). These all involve very low intensity light, usually in the visible range. I have studied the experimental evidence, and found that as soon as you abandon the photon idea a whole new vista opens up. If you think in terms of photons, you are constrained to assume that each is either transmitted or (in more realistic explanations) absorbed by a polariser. If transmitted, it has the same frequency as it started with and hence, according to QT, the same energy. You then have no choice but to assume that the photomultipliers detect a fixed proportion of the incident photons. One single parameter, the "quantum efficiency", characterises them completely.

Now with this photon picture, it might just be possible to account for some of the published results - ones that can be explained by failure of the "fair sampling" assumption -- but I do not believe you can explain Aspect's anomalous effects reported in his thesis, such as the inequality of N+- and N-+. I do not believe you can explain, either, all the unpublished results of the trial runs needed in order to find the "proper" experimental parameters to get the "right" result!

I believe that it would be perfectly feasible to investigate my assertions! All that is needed is, for example, to repeat Aspect's experiments with a range of different photomultiplier settings and coincidence windows, and with different emission rates. The whole picture that would then emerge would be explainable only if we admit that photomultipliers do not, in fact, have a linear response to "number of photons" (i.e. to intensity). If it is not linear, then the photon concept is wrong.

Or could the EPR results be explained by a theory such as Paul Wesley's ("Classical Quantum Theory", 1996), in which each of Alain Aspect's "photons" is assumed in fact to comprise a great number? If we take just the EPR experiments, this is possible, but the idea forces us to confront another range of problems. How many photons are there? How do we count them? As Aspect himself was aware (see his paper with Grangier and Roger: Europhysics Letters 1, 173-179, 1986), in practice we generally estimate the "number of photons" by measuring the energy in some way and then assuming that there is one "photon" per unit of energy, the unit depending on the frequency according to Planck's rule. It is probably only in a few experiments, the ones that are supposed to demonstrate "quantum entanglement", that special sources are used that quantum theorists consider to emit "photons" exactly one pair at a time. Under Wesley's theory this is denied: each emission comprises a whole bunch of coherent "photons". The model gives a satisfactory explanation of the EPR experimental results, and it is interesting in that it does force us to think about what really is being emitted, whether in "atomic cascades" or by "parametric down-conversion" in "non-linear crystals", but it offers no advantages that I can see over a pure wave picture.

And I think it can be proved that Wesley's is not actually the "true" solution. Ironically, it can be shown to be false by some experiments designed by David Chalmers (not the one who is famous in the area of consciousness theory, but one who does actual optical experiments in a back room), who himself has a corpuscle theory but one that he is not yet ready to release to the public.

Chalmers' experiments involve considering the light that seems to originate at a point where there is apparently destructive interference. Under Wesley's theory, which associates groups of photons with the places in an interference pattern where light is visible, the dark places in the pattern should contribute less than the visible ones, however you measure them. But Chalmers has shown that if you mask out a dark region, it has as much effect on the total light at a later point as masking out a bright one. The only explanation I can see is that though the region appeared dark, this fact was of no more significance than the regions of zero motion in a pattern of interfering water waves. The light was at this point in fact moving in two or more separate but superimposed waves, crossing at a slight angle. It just does not make physical sense to "dissect" a wave in this way, as its energy is distributed over many wavelengths longitudinally, and some finite distance laterally.

(Chalmers' own explanation is different, and he prefers to describe the experiment as selecting either bright or dark fringes then checking that the dark fringe contains the same energy as the bright one. In his view, the actual experiment is equivalent to this, but I am not myself so sure.)

This masking experiment is the only way I know so far of trying to distinguish experimentally between Wesley's "mini-photons" and the classical continuous wave idea, but in my view continuous waves have the overall advantage in many other respects. Any "photon" model seems to pose such unnecessary difficulties in visualising how light is generated - and it can be generated in many different ways. When weak light is measured, for example from distant stars, it is so much simpler to imagine the light spread out thinly, more-or-less uniformly weak, than to imagine it in little photons, each as strong as when it was born.

But let us return from our digression into "mini-photon" ideas. It has left us with an unsolved problem: fact that we cannot have both the mini-photons and the standard ones with the same energy!

This is important. How much energy does "the photon" have? The question brings me round full circle, to how we came to invent it. For it must have been fairly obvious even at the time that there could have been other explanations for the black body spectrum (see Steven Rado's brilliant book, advertised at http://www.aethro-kinematics.com, which contains a lot more solid physics than the blurb might suggest). The physics community could easily have decided that Planck's mathematics was no reason at all to assume that light really did come in lumps of a fixed size. (Another digression: Was the real reason for accepting the photon Einstein's sudden rise to god-like status with the exaggerated reaction of the New York Times to his correct prediction of the bending of light, observed during the 1919 solar eclipse?)

Yet the word was invented, and it stuck! But did it really make any more sense than declaring, for example, that because we had invented the word "gram", a gram must be a particle? It is only because we are dealing with the unobservably small that we have been led astray by language! What we really invented was just a unit for use with radiation energy. It is most exceedingly unfortunate that we felt it necessary to invent a different unit for each frequency!

Each unit became a particle. But why? It just does not make sense.

Waves, but of what?

I believe that light is purely wave (in fact, I would go further and say the whole Universe is purely wave, and it may happen that I need to tell the beginnings of that story as well), but waves of what? Current teaching of Maxwell wave theory tries to persuade us that we do not need to know what the waves are made of. Under the same kind of philosophical view that enabled Niels Bohr to declare that the aim of physics was not to understand but merely to predict observables, we have been told that the mathematical formulae for light waves are all we need. These waves simply exist, with no need for a medium!

There were some fairly good reasons, of course, for deciding there was no medium. They amounted to the fact that the experimental results of trials such as the famous Michelson-Morley ones were confusing. (There is a lot of new evidence available now -- see, for example, http://www.geocities.com/Athens/2740/ -- but it is still confusing!)

They seemed to show that there could not be a fixed universal aether for the waves to travel in, and they did not think it possible for the aether to get dragged along with the earth, so they just couldn't come up with a consistent idea of its properties. It might, in hindsight, have been a good idea to declare that they did not know how the aether behaved on the large scale, but they did have consistent ideas so far as laboratory experiments were concerned, but this was not to be.

Again, more historical research is needed, but one of the major stumbling blocks to understanding light and the aether must, I think, have been the apparent "fact" that light is a "transverse" wave. As mentioned above in relation to Lorentz' ideas, light could be, and I think is in fact, just a transverse pattern carried by longitudinal waves. I call the longitudinal waves "matter waves", or "phi-waves". In my view, this is all there is in the whole Universe.

Once you have this idea, it becomes much easier to envisage the possibility that the aether is partly dragged along with the Earth, as longitudinal waves can exist in liquids, in gases, in anything. (Purely) transverse ones can only exist in solids. Transverse ones travelling at the speed of light can only exist in solids of enormously great rigidity! Thus I think it is clear that forgetting them helps our imagination, as there is no problem in thinking about the Earth moving through a highly compressible gas, or, possibly, a gas that partly flows through it, as if the Earth were just a mesh of atoms with spaces in between large enough for our gas particles. Or even, as Lorentz himself suggested, an aether that forms the atoms themselves as well as the space in between -- an aether such that a moving atom takes some of it into itself at the front and releases back into the "vacuum" some from behind, with the aether itself staying still. But this was the kind of static aether that it was thought the Michelson-Morley results ruled out. This is yet another area for more research, as new evidence is conflicting.

Fortunately, this essay is about light, and does not need to concern itself over much with matter. We can leave open, therefore, questions of how phi-waves or light interact with matter, whether or not the aether flows through atoms or gets dragged with them, and a host of other fascinating questions.

First, just a little clarification of terminology. I assume that the whole Universe is made of aether, and that matter or phi-waves are patterns formed by some kind of change of state of the aether. Sometimes one may refer to a vacuum, but this is always to be taken as full of aether and of phi-waves, comprising, in the language of Stochastic Electrodynamics (SED) , the zero point field (ZPF). It may or may not be possible to identify my phi-waves with QT's basic waves, at the "Compton wavelength". I suspect that we do not know their usual wavelength, though it must be very short.

Now phi-waves have a natural tendency to travel at speed c in the vacuum of "open space", though there seems to be evidence that they go slower than this is the vicinity of condensed matter. An important characteristic is that they always have a specific direction. This is in contrast to the current teaching on Maxwell waves, where, for given a "wave equation" there are always two solutions, one forwards and one backwards. This teaching leads to ambiguous ideas about the "arrow of time". There is no ambiguity with phi-waves. The waves are what we talk about, not their mathematical representation!

While we are on this subject, criticising the current formulations of Maxwell theory, there is another little experiment of David Chalmers' that is relevant. It shows, I think, that Maxwell's equations are not enough. The experiment goes as follows. Direct a coherent laser beam onto a beamsplitter set at 45 degrees to its path. Reflect the transmitted light using a plane mirror perpendicular to the path. Observe light that has returned to the beamsplitter then been reflected. Under Maxwell theory, as I understand it, the direct laser light would be producing a transverse oscillation in the electric field at the far side of the beamsplitter, and this would interfere with the similar oscillation caused by the returning beam from the mirror. Therefore the observed light should show interference, changing from high to low intensity as you change the exact position of the mirror. It does not.

To my mind, this confirms that the laser light has a definite direction and cannot be properly modelled by equations that do not recognise this. At the beamsplitter, the transmitted fraction does not set up a stationary oscillation in the field. It is an oscillation with an associated momentum -- that of the underlying phi-waves -- and this momentum is straight on, not towards the viewer. The outgoing laser light and the returning light pass through each other with no interaction.

Mostly, this is typical of the behaviour of phi-waves, at least in open space. It seems that the beamsplitter of Chalmers' experiment was not too different from open space, as phi-waves can behave differently in other circumstances. When they interact with matter more fully, they can alter the local "static" electric field. Also, it is possible that when two waves are superposed and in almost the same direction, they may merge if they are almost in phase. Or again, waves of extremely high amplitude may behave as if almost solid, as when two laser beams interact, or when lasers interact with molecules in "optical tweezers" and "spanners" (See Miles Padget and Les Allen, Optical Tweezers and Spanners, Physics World, Sept 1997, p35).

Thus light, to me, is just a flow of phi-waves that happens to be modulated so as to carry an oscillating pattern. The amount of "energy" in it is determined in some sense by the strength of the pattern -- its amplitude, frequency, and coherence properties. The aether is full of these phi-waves, so that when two beams "interfere destructively" and seem to leave no visible trace, they in fact have reduced themselves to raw phi-waves that do not carry enough pattern to be detected. "Radiation", in this view, is not a very definite concept at all. It is not surprising, therefore, to find that the results of experiments involving very low intensities can depend strongly on our instruments. This is an accepted fact of "quantum" phenomena!

Let us return to how Maxwell envisaged things. You may have noticed that I do not refer at all to the magnetic field, only to the electric one (and for light in a vacuum, I have even suggested that there is no electric one worthy of the name!). I feel that Maxwell, and more especially those who tidied up his equations later, made too many assumptions about being able to carry over results established in the neighbourhood of solid matter, where there are static or near-static fields, to open space. In open space, the magnetic field of light can be regarded as just another derived variable, another way of looking at the pattern of phi-waves. I think Maxwell made complications for himself in thinking that the electric and magnetic fields interacted, sort-of crawling over each other in endless cycles, to produce the effect of light. The truth, for open space, might be much simpler. Matter waves just flow, not interacting with anything unless circumstances become extreme.

One often sees pretty pictures of light propagating. Circularly polarised light is supposed to move with the electric field rotating to form a spiral. In a sense it does, but remember, in open space there is no "electric field" as such. This only materialises when light interacts with solid bodies. In open space, there is just a pattern of compression and rarefaction of the underlying tiny phi-waves. One can think of the compressed regions as being like the material of a screw, or perhaps better, of a spring. But instead of thinking of the spring as twisting as it moves longitudinally, you need to visualise it simply moving rigidly along. No twisting. If you are at rest in its path, standing, say, at a point along its axis, you will perceive it rotating around you, but this will be slightly illusory. Each "bit" of the spiral is just a little region of compressed phi-waves, and it is travelling in a straight line!

Emission and Detection of Light

Here we enter the vast unknown! Quantum theory has almost put a stop to genuine investigation for nearly 100 years. And yet the practical people who we expect to produce the latest lasers, or the wonderful molecules that will mimic the chloroplasts of Nature and convert sunlight cheaply and efficiently into energy for us, know a great deal. I know but a little, and this little has been gleaned mainly from one source: Alain Aspect's PhD thesis. Being in French, and my knowledge of the language being limited, it took me some time to translate as accurately as I could the important parts. They are now available at my Web site.

The light source in Aspect's EPR experiments was an "atomic radiative cascade". The quantum theory behind this is all to do with incoming "pump laser" light shifting the energy levels of ionised atoms to a certain "excited level". From this level, they "relax". This being a two-level cascade, they do so in two stages, emitting first one photon at one frequency, then another, at another frequency. The time of emission of the second photon is supposed to be random, but controlled by a distribution such that the probability of emission decreases exponentially with time, with a "half-life" of about 5 ns. All the probabilities of changes of state are supposed to be covered by ideas about the relative population numbers in the different states -- Einstein's "A" and "B" coefficients.

Now, as I have mentioned in a few of my papers (see, for example, my paper in the proceedings of a conference in Athens in 1997: "Behind the Scenes at the EPR Magic Show" in "Open questions in relativistic physics", Franco Selleri (ed.) (Apeiron, Montreal, 1998)), I think that there is no actual evidence that this weak, visible, light is emitted by individual atoms at all. It could, for all I know, originate from the field between the atoms, possibly with the whole field of the source region (of a fraction of a millimetre in size in each direction) acting in unison. I imagine both "photons" being simply resonances, starting simultaneously and tapering off exponentially in amplitude. The reason they are detected at the times they are, appearing to support the QT model, is the way the detectors work, together with the fact that their intensities follow the same kind of pattern as the QT supposed probability distribution.

So how do the detectors work? Aspect describes the construction and operation of his photomultipliers, "discriminators" and electronics of the coincidence circuitry in some detail. I do not propose to go into it now. I will just say that the picture seems most easily explained if one presumes that some kind of local electromagnetic noise is added to the signal at the "cathode", or elsewhere along the path. The signal itself sets up oscillations in the field of the cathode. These oscillations produce interference peaks here and there. When a peak happens to combine favourably with some other source of noise, the "electrical resistance" of the apparatus breaks down so that a pulse of current flows. (The QT story is in terms of electrons being emitted, then amplified by a series of cascade effects as they hit a succession of "dynodes".) The whole device seems to behave as a kind of capacitor.

The output current is converted to a neat square pulse by the discriminator. This is just ordinary electronics, with ordinary macroscopic currents. The big question is how to decide when you've got a pulse big enough to represent a photon! The decision seems to be very arbitrary. Aspect discussed it at considerable length, saying at one point that the techniques of single photon counting are well known, but at another that there are real difficulties in making this decision.

It seems to me that the whole apparatus has been designed so as to simulate the QT idea of the photon in the best way possible. I suspect that modern photomultipliers do this even more convincingly, so that the experimenter is relieved of the tricky decision! For the use of photomultipliers in this "Geiger mode", for counting single photons, is a very artificial one. In most applications, they are not used like this. They just produce continuous currents, and it is assumed that the currents are proportional to the input light energy.

So is the whole of "quantum optics" an artifact, a result of using these "quantum simulators" for detectors, together with the fact that inconsistencies are unimportant when you are dealing with thousands, if not millions of photons and looking at ("normalised") statistical properties only? I rather think it is!

Topics not yet Covered:

Detection by photographic plates

Lasers

Polarisation: is "unpolarised light" sometimes purely longitudinal?

Hanbury Brown and Twiss effect

Holography

Coherent X-rays

Aspect's statement that the photoelectric effect was not conclusive evidence

Weakness of Grangier, Roger and Aspect's 1986 experiment and other demos of particle nature of light

Absence of refs to photons in books such as Hecht Optics

Clauser and Horne, 1972: EPR effect was just about the only "quantum effect" that could not be explained classically

Effect of belief in "EPR effect" on the powers of reason!

Absence of phrases such as "collapsing wave functions", "relativistic electrons", "Hermitian matrices", "Time dilation", "Lorentz invariance" from discussion!

Ideas for experimental testing