Michelson-Morley Experiments Revisited and the Cosmic Background Radiation Preferred Frame
| Scientific Paper | |
|---|---|
| Title | Michelson-Morley Experiments Revisited and the Cosmic Background Radiation Preferred Frame |
| Read in full | Link to paper |
| Author(s) | Reginald T Cahill |
| Keywords | Michelson interferometer, Cosmic Background |
| Published | 2003 |
| Journal | Apeiron |
| Volume | 10 |
| Number | 2 |
| No. of pages | 14 |
| Pages | 104-117 |
Read the full paper here
Abstract
A new information-theoretic physics has given rise to a quantum-foam description of space relative to which absolute motion is meaningful and measurable. In this new physics Michelson interferometers operating in gas mode are capable of revealing absolute motion. We analyse the old results from gas-mode Michelson interferometer experiments which always showed small but significant effects. Analysis of the Illingworth (1927) experimental data, after correcting for the refractive index effect of the helium used, reveals an absolute speed of the Earth of v = 369 ± 123 km/s, while the Miller experiment (1933), after correcting for the refractive index effect of the air, now gives a speed of v = 335 ± 57 km/s, which are in agreement with the speed of v = 365 ± 18 km/s determined from the dipole fit, in 1991, to the NASA COBE satellite Cosmic Background Radiation (CBR) observations. The new physics also implies that vacuum interferometers will give null results, as has been observed many times. These experimental results imply that absolute motion is observable and that there is a preferred foliation of spacetime coinciding with the CBR frame.
Overview
Written with Kirsty Kitto at Flinders University and published in Apeiron in 2003, this paper makes a single, sharply drawn claim: that the classic Michelson–Morley-type interferometers were never null, that their small residual fringe shifts were mis-analysed because the light in them travelled through gas rather than vacuum, and that when the refractive index is properly accounted for the recovered speeds agree with the Earth's motion relative to the cosmic background radiation frame measured by COBE.
The theoretical setting is Cahill's Process Physics, an information-theoretic scheme in which space is an emergent quantum foam and gravity is the inhomogeneous flow of that foam. For the purposes of the interferometer analysis, however, almost all that is used is a Lorentzian reading of relativity: light travels at c relative to the foam, moving observers find it to be c anyway because their rods and clocks are themselves affected, and the Fitzgerald–Lorentz contraction is a real physical shortening of the arm along the direction of motion. The paper's contribution is to notice that in a gas the contraction and the geometrical path difference no longer cancel exactly, leaving a residual that scales with the refractive index — and therefore that vacuum interferometers must give null results while gas-mode ones must not.
The argument
Three theories, one parameter
Michelson's original analysis gave, for the change in propagation time when an arm of length L is rotated through 90°,
Δt = k2Lv2/c3
with k2 = 1, where v is the projection of the apparatus velocity onto its plane. Miller had already introduced k2 as a phenomenological parameter. Cahill and Kitto tabulate the three candidate physics against it: Newtonian physics, with no contraction, gives k2 = n3; Einsteinian physics, in which absolute motion has no meaning, gives k2 = 0 in both gas and vacuum; Process Physics gives k2 = n(n2 − 1), which vanishes only for n = 1. The same table sets out the three theories' treatments of time (geometry, curved geometry, process), space (geometry, curved geometry, quantum foam), gravity (force, curvature, inhomogeneous flow) and the quantum. Each theory is said to subsume the one above it, with Einstein's spacetime emerging as an approximation to Process Physics but with a preferred frame.
Deriving the residual
The derivation is a standard two-arm calculation carried out in the foam frame, with photon states travelling at V = c/n and the parallel arm shortened to L√(1 − v2/c2). For the arm along the motion the out and back times are L||/(V − v) and L||/(V + v); for the transverse arm Pythagoras gives 2L/√(V2 − v2). The difference is exactly zero when v = 0, and also when v ≠ 0 provided V = c — which is why the vacuum experiment nulls. With V = c/n and v << V, expansion gives
Δt = L n(n2 − 1) v2/c3
against the Newtonian Δt = L n3v2/c3. Since the historical speeds vM were extracted using the Newtonian formula, the corrected speed follows from the ratio of the two: v = vM√(n3/(n2 − 1)) ≈ vM/√(n2 − 1) for n close to 1. The typical Δt at stake is of order 10−17 s, a fraction of a fringe.
The numbers
Because n = 1.00029 for air at STP and n = 1.000036 for helium at STP, the factor 1/√(n2 − 1) is large — about forty for air and over a hundred for helium — and it is very different for the two gases. The authors take as input the re-analysed Newtonian speeds of Hector A Munera, who had reviewed these experiments, found systematic errors and invalid inter-session averaging, and concluded that the runs "never were null". Munera's values are vM = 6.22 ± 0.93 km/s for the Michelson–Morley noon sessions and 6.80 ± 2.49 km/s for the 18 h sessions; 3.13 km/s (95% confidence bounds 2.09 and 4.17) for Illingworth's helium-filled instrument; and 8.22 km/s (bounds 6.83 and 9.61) for Dayton C Miller at Mount Wilson.
The apparent conflict between Miller's 8.22 and Illingworth's 3.13 km/s is, on this account, exactly what should be expected: Illingworth used helium to control temperature variations, and helium's much smaller n − 1 means a much smaller signal for the same absolute speed. Applying the correction turns the two disparate figures into 335 ± 57 km/s and 369 ± 123 km/s respectively, to be compared with 365 ± 18 km/s from the COBE dipole. The authors note that unless a search procedure is used, an interferometer whose plane is tilted from the velocity vector must always register a speed less than or equal to the true one, which they say the data respect.
Conclusions drawn
The authors conclude that absolute motion is meaningful and has been measured, that there is a preferred foliation of spacetime coinciding with the CBR frame, and that gas-mode interferometers should be revived as research tools. They are careful to say this does not overturn the relativistic formalism: it "merely indicates the requirement for a re-interpretation". Vacuum experiments — Joos, Kennedy–Thorndike, Brillet–Hall, Braxmaier and colleagues — are said to test only the Lorentz contraction and so to be silent on the question.
Assessment
What is admirable here is the shape of the claim. Cahill and Kitto identify one parameter, k2, on which three theories give three different answers — n3, 0, and n(n2 − 1) — and they name the experiment that discriminates them. They explain, rather than explain away, the otherwise awkward fact that vacuum interferometers null while the older gas experiments did not, and they turn the long-standing embarrassment of Miller's non-null result into a prediction. The internal consistency of the two corrected numbers is genuinely striking: Illingworth and Miller used gases whose refractive indices differ by a factor of eight in n − 1, and yet the same correction brings both onto the COBE value. That is the kind of coincidence that deserves an explanation rather than a shrug, and the paper is right that the data ought to be reconsidered.
The chief technical difficulty is one the paper raises and then sets aside in a single sentence. The gas in these instruments is at rest in the laboratory, not in the foam, and light in a moving medium is subject to Fresnel drag, with dragging coefficient 1 − 1/n2. That coefficient is of order n2 − 1 — precisely the order of the residual effect being claimed. Calling the dragging "a very small effect and not required in the present analysis" is therefore not enough: it is small in the same sense and to the same degree as the signal, and a defensible derivation would have to carry both terms and show what survives. The classical Fizeau result, which measures the drag directly, is the natural place to test that bookkeeping, and it is not mentioned.
The second difficulty is amplification. Dividing by √(n2 − 1) multiplies a signal of a few km/s by forty or a hundred — and it multiplies every systematic error by the same factor. Thermal gradients, mechanical flexure and drift are the classic afflictions of these instruments, and Illingworth's whole reason for using helium was to suppress temperature variation; the correction that rescues his number also inflates whatever thermal residual he failed to suppress. The quoted uncertainties reflect this: 369 ± 123 km/s is consistent with a very wide range of speeds, so the "agreement" with COBE is a good deal weaker than the coincidence of central values suggests.
Third, only speeds are compared. A real velocity through a preferred frame has a direction, and it produces a characteristic sidereal-day and annual modulation whose phase is as diagnostic as its amplitude. Miller's own reported apex lies far from the direction of the CBR dipole, and nothing in this paper addresses that discrepancy; the earlier 30 km/s orbital contribution is acknowledged but not removed from the interferometer results, unlike the COBE figure. Until direction and phase are reproduced, the case rests on four scalar values, two of them not quoted in the abstract at all.
Finally, the information-theoretic framework does no work in the derivation. What is actually used is a Lorentzian aether with a real contraction and a light speed of c/n relative to it — a nineteenth-century model dressed in new vocabulary. The result therefore stands or falls independently of Process Physics, which is fair enough, but it means the paper should not be read as evidence for that wider programme. The proper reply to the paper is the one it invites: build a modern gas-mode interferometer, with temperature control and continuous sidereal coverage, and see. That test remains the strongest thing about the argument, and it has not been independently carried out.