1 Denitions: Inertial Frame of Reference (IF)

Download as ps, pdf, or txt
Download as ps, pdf, or txt
You are on page 1of 6

Special Relativity: Lecture 1

1 De nitions
ton's laws hold. The standard example is a frame of reference in empty space moving with uniform velocity wrt the xed stars. An IF moving with constant velocity wrt another IF is also an IF. Invariant, a quantity that does not change its value under a transformation, e.g. the length of a vector under a coordinate rotation. Covariant equations have the same form under a transformation, the physical content is the same, e.g. Newton's Laws of motion are covariant wrt a coordinate rotation. Special Relativity deals with physical phenomena true in all IFs, e.g. mechanics and electromagnetism. General Relativity deals with physical phenomena in accelerated frames of reference as well as IFs and is a theory of gravitation. Standard IFs S and S 0 will be de ned to be RH coordinate systems with S at rest and S 0 moving with a velocity v directed along the positive x-axis in S . We shall assume that S and S 0 coincide at t = 0 so that t = t0 = 0 and x = x0 = 0.
Inertial Frame of Reference (IF) is an unaccelerated frame of reference, in which New-

2 Galilean Relativity
Newton's laws of motion are covariant wrt coordinate rotations. They are also covariant wrt the Galilean transformation between IFs S and S 0 with constant relative velocity. For the standard IFs S and S 0 de ned above, the Galilean transformation is: x0 = x vt; y0 = y; z0 = z; t0 = t: By di erentiation we have that; dx0 = dx0 = dx v; dt0 dt dt

which gives the Newtonian law for relative velocity u0 = u v. Acceleration is unchanged since d2x0 = d2x : dt02 dt2 In Newtonian mechanics the mass of a body, m, is an invariant, hence 2x F = md dt2 in S; becomes 2 x0 0 F0 = m d dt02 in S : Thus F0 = F and Newton's law is covariant wrt the Galilean transformation. Since the laws of conservation of energy and momentum follow from Newton's laws they are also Galilean covariant. Notice also that the spatial and temporal intervals between two events r and t are separately invariant under the Galilean transformation.

3 The Problem
Although Newtonian mechanics is a Galilean covariant theory, electromagnetism is not. For example the force between two charges at rest is not the same as the force between the same two charges in uniform motion. Also the Galilean transformation predicts that the velocity of light will depend on the velocity of the source, viz. c0 = c v, and this is at variance with Maxwell's equations. At the end of the 19th Century there were three possible solutions to the puzzle: 1. Galilean relativity correct. Newtonian mechanics correct. EM in need of modi cation. 2. Galilean relativity correct for mechanics. EM correct but in a preferred frame of reference at rest wrt the aether. 3. No preferred frames. Maxwell's equations correct and satisfying a principle of relativity (not Galilean). Newtonian mechanics in need of modi cation. For a time the second proposal was the favourite but the failure to nd any experimental evidence for the aether led to Einstein's theory of special relativity (1905). Einstein based his theory on two fundamental postulates: The laws of physics are identical in all IFs | there are no preferred frames and there is no absolute motion. The speed of light in a vacuum is a universal constant c with the same value in any IF | c does not change with the motion of source or observer. All the physics of Special Relativity follows just from these two postulates.

4 Evidence for the constancy of c


a) Michelson-Morley Experiment: High precision optical interference experiment to detect a change in the speed of light due to motion of the apparatus wrt the aether (as expected in Galilean relativity). The setup is that of a standard Michelson interferometer. Light from a source S is split into two beams by a half-silvered mirror at A. The beams are re ected from mirrors B and C and the nal beam is viewed by telescope T.

Assume that the aether is moving wrt the apparatus with a velocity v parallel to AC as shown and that Galilean relativity is correct. Then: 2`C c : C Time A ! C ! A : tC = c `C v + c ` = 2 +v (c v 2 ) During the time t it takes for light to traverse the distance AB, the mirror B will 22 move a distance vt hence, c2t2 = `2 B + v t . The same is also true for the return path so: `B : Time A ! B ! A : tB = p 2 c2 v2 Hence the time di erence t is given by (for small v2=c2):
2 (I:1) t = tC tB = 2(`C c `B ) + v c3 (2`C `B ): The whole apparatus is now rotated through 90o so that v is now parallel to AB: 2 t0 = t0C t0B = 2(`C c `B ) + v c3 (`C 2`B ): Thus t 6= t0 and there should be a fringe shift n in the interference pattern, 0 2 (`C + `B ) n = c( t t ) = v : (I:2) c2

In the original Michelson-Morley (1887) experiment `B = `C = 11 m and = 600 nm. No shift was observed, giving a limit of n < 0:01 or v < 5 km/s to be compared, for example, with the Earth's orbital velocity of 30 km/s. The experiment has been repeated many times with null results of improving precision, some of which are shown in the Table.

Some Aether Drift Experiments:


Year Experiment 1887 1903 1924 1924 1927 1930 1963 1963 Michelson & Morley Morley & Miller Miller (sunlight) Tomaschek (starlight) Illingworth Joos Jaseja et al Champeney et al Method Interferometer Interferometer Interferometer Interferometer Interferometer Interferometer He-Ne Maser Mossbauer

` (m) vmax (m/s)


11 32.2 32 8.6 2 21 5000 380 360 2026 173 81 30 5

On the basis of the electrodynamics of moving objects, Fitzgerald and Lorentz proposed that the length q of an object moving parallel q to the aether wind would be contracted by an amount 1 v2=c2 so that `C ! `C 1 v2=c2 and 2 t = tC tB = 2(`C c `B ) + v (I:3) c3 (`C `B ): Thus for equal arms t = 0, similarly t0 = 0, and no fringe shift would be expected. This idea was disproved by the experiment of Kennedy & Thorndike (1932) in which an interferometer with unequal arms was used. Assuming the contraction hypothesis, 2 2 1=2 (I:4) t ' 2 c (`C `B )(1 v =c ) : The apparatus itself was not rotated but a shift in fringe pattern would have been expected over a period of months as the Earth revolved around the Sun. None was however observed. b) Mossbauer Experiment: The Mossbauer e ect is the recoilless emission and absorption of photons from atoms locked into a crystal lattice. The experiment of Champeney, Isaak & Khan (1963) used the photons from the 14.4 keV excited state of 57Fe with a natural line width !=! 10 12 . The idea of the experiment is to look for a frequency shift between a photon source and a resonant absorber placed at opposite ends of a high speed rotator. We discuss this experiment in detail since it provides the most sensitive test of aether drift to date. The phase = k:x !t of an EM wave is an invariant wrt to the Galilean transformation (it is proportional to the number of wave crests passing a xed point in a xed time). Considering our standard IFs S and S 0 we have: k0:x0 !0t0 = k:x !t; or

k0:(x vt) !0t = k:x !t; hence :v ): k0 = k; !0 = !(1 nc (I:5) The shift in phase velocity is c0 = c n:v, where n is the unit vector along k. In addition the direction of the wave packet in the moving frame will be shifted to a direction m as seen from the xed frame (see panel (a) in gure above), cn v : m = jc n vj
To rst order in v=c this can be inverted to give: (I:6) n = 1 v:cm m + v c: If uS;A are the lab velocities of the source and absorber respectively and v is the aether velocity the corresponding velocities relative to the aether are vS;A = uS;A v. Using eq (I.5), we get ! !S !A = ! ( v A vS ):n = (uA uS ):n: c c Next refer n to the lab frame direction m to get, (I:7) !S !A = ! c (uA uS ):(m + v=c): In the experiment source and absorber were at the opposite ends of a rigid rotator (see panel (b) above) so (uA uS ):m = 0 hence ! = v:(u u )=c2 = 2 R sin( t)jv j; (I:8) A S ? ! c2 using variables de ned in panel (b) above and jv?j as the component of the aether velocity perpendicular to the axis of rotation. Two rotation speeds of 1257 and 7728 rads/s were used with R = 4 cm. The observed fractional width was 2 10 12 and no diurnal change was seen to within 1% giving jv?j = 1:6 2:8 m/s. This is clearly consistent with v?, i.e. null evidence for an aether wind.

5 Independence of c from speed of the source


1) X-ray Binary Stars The problem is to nd sources moving at su ciently high speeds to produce observable e ects and for this reason binary star systems are, in principle, attractive. If the velocity of light depends on the velocity of the source (say as c0 = c + kvsource) then there will be deviations from the sinusoidal variation in the apparent velocity of the rotating star as measured by the Doppler e ect. However interstellar space is not completely empty but contains a plasma with electron density ne 0:05 cm 3, which will induce dispersive e ects in a length = =2 j n 1j, where n is the refractive index. In a plasma, this is = ( nea0) 1 where a0 = e2=(4 0mc2). Thus for an optical wavelength of 500 nm, 2 light years in interstellar space, i.e. less than the distance to even the nearest star. However for 70 keV X-rays, 5 104 light years, the extent of the Galaxy. Using the X-ray binary Her X-1 at 1:9 104 light years, Brecher (1977) found a limit of k < 2 10 9 . (You may wish to attempt to
0

repeat his analysis as described in Problem 4, Set 1.)

2)

decay The best terrestrial limit comes from the experiment of Alvager et al (1964) in which photons from 0 decay were timed over an 80 m path length. The 0 velocity was 0.99975c and k was found to be ( 3 13) 10 5 , quite consistent with zero.

6 Independence of c from frequency


1) Crab Pulsar The Crab pulsar (NP0532) is at a distance of 6 103 light years and has a pulse duration of 3 10 3 s. Observations from frequencies of 4 108 Hz through the optical upto 1 MeV photons gives c=c < 10 14. 2) SLAC Experiment Brown et al (1973) used the time of ight technique over a distance of 1300 m to compare the velocity of propagation of pulses of eV (optical) and 7 GeV photons and found c=c = (1:8 6) 10 6 . (They also compared the velocity of 11 GeV electrons with that of light, nding (c ve)=c = ( 1:3 2:7) 10 6 , consistent with although not sensitive enough to test the special relativity expectation of (c ve)=c ' 2 9 m2 e =2Ee 10 .) 3) Supernova 1987A Neutrinos of energy 10 MeV were detected from the core collapse of a massive star 1:6 105 light years away in the Large Magellanic Cloud at 7:35:40 UT on 23 February 1987. Optical photons from the resulting supernova were rst seen at 10:38 UT. (This delay is expected since the neutrinos are emitted in the rst few seconds of the collapse while the outburst of light begins only after the stellar envelope is blown away after several hours by the outgoing shock wave.) Thus the fractional transit time di erence of 2 10 9 gives an upper limit on the fractional di erence in the velocity of propagation of light and neutrinos (which are very nearly massless).

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy