You are on page 1of 14
410 16.3 16 Special relativity - a study in invariance In the above example, we have arranged that x’ = 0, x = 0, /' = 0, f = 0. This event, the coincidence of the origins of the two sets of axes in Fig. 16.1, is simultaneous frames of reference. For an event at any other valuc of x, the observers in 5 and 5” about the time at which the event occurs, since et =y(ct—Vx/e), (16.28), and henee if x #0, ¢’ # f. In other words, although observers in § and S’ can agree on simulianeity at one point in space-time, they will disagree at all other points. This is the origin of the phenomena of time dilation, length contraction, the twin paradox ete. It is the fundamental difference between Galilean and special relativity. In Galilean relativity, observers in S and S' always agree everywhere about simultaneity, This is clear from the Newtonian limit of (16.28). If V/e — 0, y — Land ¢’ =f everywhere. Three-vectors and four-vectors ‘Three-vectors provide a compact notation for writing the laws of physics in a form which is independent of the choice of reference frame. No matter how we rotate or translate the frame of reference, or change the system of coordinates used, the vector relations remain form-invariant. For example: (i) Vector addition is preserved: if at+b=c then a’ +a’ (ii) Equally, if a-(b+e) then a’ -(b' +e"). iii) The worm or magnitude of the three-vector is invariant with respect to rotations and displacements: a tay. la? =a} +a} +42 say? +a} (iv) Scalar products are invariant: ab =ayby + arb, + agbs = ay by + ash) + ay}. In other words, these vector relations express truths which are independent of the frame of reference in which we perform our calculation, Our objective is to find quantities similar to three-vectors which remain form-invariant under Lorentz transformation, These faur-vectors are objects which enable us to write down the law of physics in relativistic form. The components of the four-vectors should transform like the components of the primitive four-veetor [c ], Which we met in the previous section; we need to find expressions for these components which relate them to physical quantities measurable in the laboratory. There remains the thorny question of notation, I will use the Lorentz transformations in the form (16.27) and define the components of four-vectors as quantities which transform 6.3. 16.3 Three-vectors and four-vectors 411 like cr and x, y, 2, The components of the four-vectors are written in brackets and | use bold italic capital letters to represent a four-vector. Thus = let.x.) (16.29) To work out the norm of R we write R=|RP = (16.30) that is, there is a plus sign in front of and minus signs in front of x”, y? and in the language of relativity, the signature of the metric is [1, —1, —1, —I]. The time component has a different status from the spatial components, and so | will use the rule that the time component has subscript 0. Thus, if'a four-vector is written as A = [Ap Ay. 42, As] then its norm is Ay — Aj — AR AR (16.31) lar exactly equivalent to x? 4 y? + 2? for three-vectors, The transformation rules for the components of A can be found from (16.27), according to the identifications ct Ay, et > Agi Poedts amedy Vode Ay Pod, 2 Ay In the rest of this section we will be concerned with finding four-vectors for kinematic quantities, that is, quantities wl describe motion. We have already introduced our prim- itive four-veetor R = [er, x, Let us proceed now to the displacement fowr-vector. Displacement four-vector The four-veetor [er, x, transforms between inertial frames of reference accord- ing to the Lorentz transformation and therefore so does the four-vector [er + cAt, x + Ax, y + Ay,z+ Az]. From the linearity of the Lorentz transformations, it follows that [eAr, Ax, Ay, Az] also transforms like [er, x, y, z]. We therefore define the quantity = [cAt, Ax, Ay, Az] (16.32) to be the displacement four-vector. To express it another way, it must be a four-vector by virtue of being the difference of two four-vectors. Evidently, the time interval between two events depends upon the inertial frame of reference in which the observer is located. Of special importance is the case in which two events take place at the same spatial location in some frame S$’, that is, for which Ax’ = Ay’ = Az’ = 0. The proper tite At is the time interval between events in this special frame of re ference, It is straightforward to show that it is the shortest time interval measured in any frame of reference, as follows, Taking the norms of the displacement four-vector in the frames of reference S and 5’, a Ai? — Ax? — Ay’ 412 In the frame 5°, Ax’ = Ay’ =Oand Af’ = Ato, and so — Az, (16.3) At = c7 At? — Ax? — Ay since Ax”, Ay”, Az? are necessarily positive, Af to must be the minimum time measured in any inertial frame of reference. An important aspect of the proper time interval Avo is that it is the only invariant time in- terval upon which all observers can agree fora pair of events separated by cAr, Ax, Ay, Az Observers in different inertial frames make different measurements of cAr, Ax, Ay, Az and, although they all measure different values of A¢ for the time interval between the same two events, they all agree about the value of Afy when they measure Ax, Av and Az aswell and insert their values into (16.33). As indicated above, the proper time is the only invariant time interval on which they can all agree. Let us work out the relation between the proper time interval Af measured in S” and the ‘time interval between the same two events measured in the laboratory frame of reference S. The four-vectors associated with the proper time interval Af and the laboratory frame are feAty,0,0, 0), [cAt, Ax, Ay, Az]. (16.4) Equating the norms of these four-vectors, Ag = CAP — Ax — A, Ar — Ar’. (16.35) But Ar is just the displacement of the origin of S’ in the time As. In other words, the velocity ¥ of the origin of $’ is Ar/Ar and hence from (16.35) Ag = ar (1 - (16.36) that is, soe’, (1631) ¥ Since y is always greater than unity, this calculation demonstrates that the proper time interval Af is the shortest time between two events. We will use (16.37) repeatedly in what follows. The relation (16.37) explains the observation of muons at level. Cosmic rays are very high energy protons and nuclei which originate in violent explosions of stars and make their way to the Solar System through the interstellar medium. Muons are created in the upper layers of the atmosphere by collisions between cosmic rays, mastly protons, and the nuclei of atoms of the atmospheric gases. A typical interaction is Ptpoant tan tan + ptpte, that is, a shower of positive, negative and neutral pions is created in the collision. In these collisions, roughly equal numbers of the three types of pion are produced. The positive and negative pions have very short lifetimes, r = 2.551 x 107 s, and decay into muons: mes pt ty ey bis 16.3.2 16.3 Three-vectors and four-vectors 413 the v,, and i, are muon neutrinos and antincutrinos respectively. The muons decay after a short lifetime Af x 107° s into electrons, positrons, neutrinos, muon neutrinos and so on. Since they are created at the top of the atmosphere, at a height of about 10 km, according to Galilean relativity they should travel a typical distance of only cAty before decaying, that is, a distance of 3 x 10° x 2.2 x 10-® km = 660 m. Therefore, we would expect very few of them to reach the surface of the Earth. However, intense fluxes of relativistic muons are in fact observed at the surface of the Earth. These high-energy muons: have Lorentz factors y > 20, Because of the effects of time dilation, (16.37), the observer on the Earth measures a decay half-life Ar = y Afo, an effect which is often confusingly referred to as ‘moving clocks run slow’, meaning that Afy is the shortest possible time interval between the muon’s creation and decay, The muon itself acts as a moving clock. According to the observer on the surface of the Earth, the muon has a half life of y Ato, in which time it can travel a distance of 13 km and so easily reach the surface of the Earth. The velocity four-vector To find the velocity four-vector, we need a quantity which transforms like AR, (16.32), and has the form of a velocity. The only time which is Lorentz invariant is the proper time Aly. As we discussed above, an observer in any inertial frame of reference can measure cAt, Ax, Ay, Az for two events and compute from these c7 At = Al? — Ax? ~ Ay? — 2. An observer in any other inertial frame of reference would find exactly the same value, c*Af2. Therefore, let us define the velactty four-vector using the Lorentz-invariant quantity Ato: AR v=—. (16.38) Ato But, from the analysis which led to (16.37), the proper time interval is related to the time interval Af measured in the laboratory frame S by Afy = Ar/y. We ean therefore write Uas walks [ex so =), Ato «Ar Ar’ AG = [ves vise vip, ¥us). =[ye. vel. (16.39) In this relation wv is the three-velocity of the particle in the frame S and y is the corresponding Lorentz factor; uy, tty, w:, are the components of #, Notice the procedure we use to work out the components of the four-velocity. In the frame § we measure the three-velocity w and hence y. We then form the quantities ye and yay, yity, yu; and know that they will transform exactly as ct and x, V2. Let us use this procedure to add two velocities relativistically. If the relative velocity of the two frames of reference is in standard configuration and w is the velocity of the particle in the frame S then what is its velocity in the frame 5"? Call the Lorentz factor associated with the relative motion of § and S’ yr. t, we write down the velocity four-vectors for 414 16 Special relativity - a study in invariance the particle in S and 8": 1p in S, Ive. va] = [ve, vite. Yay. ue) y= w/e ins’, [yey ]afy'ey'ty. yy yl. yy = — ue? /e’y ‘We can relate the components of the four-vectors through the following identities: et + y'e, ct > ye, Soy plu, x yits, y= yl, YP ¥ity, zi yu, 2 yi. Therefore, applying the Lorentz transformation (16.27), we find ye = yr (ve — Vue). yu, = yyy — Vy). ne (16.40) yu = yuy, yu = yu. The first relation gives t a (1641) ty ed = Paley’ us yw (= Puy /e?y’ ‘These are the standard expressions for the addition of velocities in special relativity. They have many pleasing features. For example, if any of wy, try or w. is separately equal to ¢ and, necessarily, the others are zero then the magnitude of the total velocity u’ in 5’ also equals c, as required by Einstein's second postulate. Note that the norm of the four-velocity (16.39) is (16.42) U2) ty? = yn? = 7a? — 2 = eC — 7/7) which is an invariant as expected. 16.3.3 The acceleration four-vector ‘We can now repeat the procedure of subsection 16.3.2 to find the acceleration four-vector, First, we form the increment of the four-velocity AU! = [cAy, A(yu)], which is necessarily a four-vector. Then we define the only invariant acceleration-like quantity we can form, by 16.3 Three-vectors and four-vectors 415 dividing AU by the proper time interval Ato: AU Ay oA dy od Aaa] pe yp =|re2 yo 16.43 At [re ir etw| [rz ar \ ) in the limit Ar — 0. This is the acceleration four-vector. Let us convert it into a more useful form for practical applications. First, note how to differentiate y = (1 — 17 /e?)7 2: dy d asa dr de eee (16.44) ee Then Gece cya Si = ( Jat ya and so du au 16.45 a di c 5) Notice what this means. In some frame of reference, say S, we measure at a particular instant the three-velocity w and the thre eleration @ of a particle. From these quantities we form the sealar quantity }-4(u » @)/c and the three-vector p4(w - a/c? )w + y?a. Because Ais a four-vector, we then know from (16.29) that these quantities transform exactly as cr and r between inertial frames of reference. Itis ofien convenient in relativistic calculations to transform into the frame of reference in which the particle is instantaneously at rest. The particle may be accelerated, but this does ‘not matter since there is no dependence upon acceleration in the Lorentz transformations, In the instantaneous rest frame, a =O and y = 1 and hence the four-acceleration in this frame is A=(0,a] =[0, aa]. (16.46) where ap is the proper acceleration of the particle. In the same frame, the four-velocity of the particle is U=([re, yu] = [c. 0]. Therefore, using the definition of the scalar product as for three-vectors, A - U' = 0. Since ions are true in any frame of reference, this means that, no matter which Jar product of the velocity and acceleration four-veetors four-vector rela frame we care to work in, the sc is always zero, that is, the velocity and acceleration four-vectors are orthogonal. If you are picious, it is a useful exereise to take the scalar product of the expressions (16.39) and (16.45) for the four-vectors A and U and show that it is indeed zero, 416 16.4 16.4.7 16.4.2 16 Special relativity — a study in invariance Relativistic dynamics — the momentum and force four-vectors that is, the description of motion, but now we means introducing the concepts of momentum So far, we have been dealing with Ainematics have to tackle the problems of dynamics, Th and force and how they are related. The momentum four-vector First of all, let us go through the purely formal exercise of defining suitable momentum and force four-veetors and then finding out if they make physical sense. Let us introduce a four-vector quantity P, which has the dimensions of momentum: P = mol =[ymoc, ymonl, (16.47) where U is the velocity four-vector and mq is the mass of the particle, which is taken to be a scalar invariant — it will be identified with the particle's rest mass. We note the following immediate consequences: (i) mioW is certainly a four-vector since mg is an invariant; (ii) the space components of P reduce to the Newtonian formula for momentum if w < e¢, that is, mya —> mow asu — 0. Therefore, if this agrees with experiment, we have found a suitable form for the momentum four-vector. Notice that we can also define a relativistic three-momentum p = yuigu from the spatial components of the four-vector. The quantity a: = ymy is defined to be the relativistic inertial mass. As Will be discussed below, I do not like this term, but it is commonly found in the literature. We need to show that these definitions result in a self-consistent set of dynamics, Let us find a relation between the relativi momentum and the relativistic mass of the particle. Equating the norms of the momentum four-vector in the laboratory frame, P = [ymoe, ymoc, and that in the rest frame of the particle, P = [moc, 0], we abtain 22 22 = yi mg. -p', or are? = mac’. (16.48) We will rewrite this expression in terms of energies in Section 16.5. The force four-vector Following the logic of the previous sections, the natural four-vector generalisation of Newton's nd law of motion is ap © dig” (16.49) 16.4 Relativistic dynamics - the momentum and force four-vectors a7 where F is the four-force and dig is the differential of proper time. We now relate the force which we measure in the laboratory to the four-force, The best approach is through the quantity we have called the relativistic threc-momentum, Why did we adopt the abave definition? Considera collision between two particles which initially have momentum four- vectors P; and P2, After the collision, these become P{ and P5. The conservation of four-momentum can then be written P+ Pi=Pi +P; (16,50) In terms of the components of the four-vectors, this equation implies ny bm =m +s, (16.51) Pit pr =P, +P, where the m's are the relativistic masses. Thus, implicit in this formalism is the requirement that the relativistic three-momentum is conserved and so, for relativistic particles, ymou plays the role of momentum, The corresponding force equation is suggested by Newton's second law: dp I" % = < tymon, (16.52) where f is the normal three-foree of Newtonian dynamics Are these definitions self-consistent? We have to be a bit careful, On the one hand, we could just argue, “Let's look at experiment and see if it works’, and in many ways this is more or less all that the argument amounts to, We can doa little better, In point collisions of particles, the relativistic generalisation of Newton's third law should apply, thatis, f = —f. We can only consider point collisions or else we would get into trouble with action at a distance in relativity — recall the relativity of simultaneity in different frames of reference. Fora point collision, ¢ = —f is true if we adopt the definition of f given above, because we have already argued that relativistic three-momentum is conserved. That is, Ap =-ap, t= ¢ made the correct choice without I problem when we tried to However, we cannot be absolutely sure that we appealing to experiment. We faced the same sort of logi understand the meaning of Newton's laws of motion (Section 7.1). They ended up being a set of definitions which give results consistent with experiment. Similarly, relativistic namics cannot come out of pure thought but can be put into a logically self-consistent mathematical structure which is consistent with experiment We therefore adopt the definition (16.52) of f as the three-foree, int me sense as in but now the particle may be moving relativistically and the relativistic ‘ea number of Newtonian dynamics, three-momentum should be used for p, Within this framework, we can d 418 16.4.3 16.4.4 16 Special relativity - a study in invariance F=mA This follows directly from the definition of F, (16.53) di = mg—— = mod, fo dig deg Inaddition, because A- U = 0, it follows that F-U=0; that is, the force and velocity four-veetors are orthogonal. The relativistic generalisation of f = dp/dt Let us write out the four-vector form of Newton's second law in terms of its components: a AE = 16.54) dig (Cd F=Lh. dorntnc) | dp dr” : fa, fi) dr where p = ymytt. Since we have argued that the relativistic form of Newton's second law in three-vector form is f = dp/dr, it follows that we must write the four-force in the form d(ynige) — d, F=[fyf= [rege P| = mid, (1635) Equating the spatial components of [ fo, yf] = mod and using (16.45), we find faa? (- 5 =) + mipya. (16.56) This is the relativistic generalisation of the Newtonian expression f = soa. Now let us analyse the ‘time’ component of the force four-vector, that is, of dp dormoc) od aE | yaya From (16.44), we can write d(vyatc) gfe SAY ay! (88). or, in terms of the relativistic mass m, dats fata é arm (=). (16.57) Now inspect the quantity (f'- u)/c?, Substituting for f from (16.56), few pu ay i? ua = moy ( z )3 + may ( ). apa 1 = ow (4) ( afta = z 16.58 nay ( 5 ) (16.58) 16.5 The relativistic equations describing motion 419 ‘Therefore, (16.57) becomes (16.59) or d(me?) — d(ymgc?) ade This is one of the amazing results of Einstein's paper of 1905, The quantity f - w is just the rate at which work is done on the particle, that is, its rate of increase of energy, Thus me? is identified with the total energy of the particle. This is the formal proof of perhaps the most famous equation in physics, = fon. (16.60) E=ae. (16.61) Notice the profound implication of equation (16.61): there is a certain amount of inertial mass associated with the energy produced when work is done. [t does not matter what the form of the energy is — clectrostatic, magnetic, kinetic, elastic, etc. All energies are the same thing as inertial mass, Likewise, reading the equation backwards, inertial mass is energy. Nuclear power stations and nuclear explosions are vivid demonstrations of the identity of inertial mass and energy. 16.4.5 A mild polemic Sometimes references are made to /ongitudinal and transverse masses in textbooks. These arise from an unhappy use of the relativistic form of Newton's second law of motion. If u || @ then the force law (16.56) reduces to = pimoa. The quantity y*mo is called the Jongitudinal mass. However, if u L a then we find that f= ymiog. and yvtg is called the rransverse mass. I dislike this introduction of different sorts of mass along with ‘relativistic’ forms of Newton’s second law of motion, in the form f = ‘ma. It is preferable to stick with the correct generalisation of Newton's second law in the form f =dp/di, where p = ymou, and abolish all masses except the rest mass sg, for the reasons discussed in the next seetion. The relativistic equations describing motion Consider a particle of rest mass my travelling at velocity ua in some inertial frame of reference S. The partiele has Lorentz factor y given by the standard expression y = (1 — u2/e?)7'? Then: © The relativistic three-momentum of the particle is ymou. © Its total energy is E = me? = ymoc. 420 16 Special relativity - a study in invariance © Its rest mass energy is Eg = nigc*. netic energy is the difference between its total and rest mass energies, Ekin = E — Bo = pme® — moe? = (y — Lyme’. (16.62) The last definition follows from the calculation which led to (16.60), in which we showed that, when we do work ona particle, we increase the quantity yunge?. For ¢, (16.62) reduces to the Newtonian non-relativistic expr n for the kinetic energy, “1a - 1 mage? * Largn?. (16.63) Wvistic form of Newton's second law of motion is pele, (16.64) + The conservation laws of energy and momentum are subsumed into the single law of conservation of four-momentum, Prt Pr =P, + Ph, (16.65) where the components of the four vectors are given by P=([po. Pe Py. [ymoc, ymottc, Yatatty, Ymou-). (16.66) * Equating the spatial components in the equation of conservation of four-momentum, we find the law of conservation of relativistic three-momenttm, Pit Pr = Pi + Py (16.67) * From the ‘time’ component of the equation of conservation of four-momentum, we find the relativistic kaw of conservation of energy; we can express this in three ways: Por + Por = Por + Pare E\+2= Ei +£i, (16.68) vimion + omar = yyrtor + yymo2 where the £’s are the total energies of the particles. 16.5.1 Massless particles There are real advantages in rewriting the relations we have just derived in terms of the various energies defined above. Let us rewrite the list entirely in terms of energies: energy E = yg"; * rest mass energy Ey = mpc’; * relativistic momentum p = yinqu = + — kinetic energy Eun = E — 16.5 The relativistic equations describing motion 421 Thus, we can write the four-momentum ofa particle as follows: (16.69) P= mW =molye, Yitr, Yuy, yus] = [ In Case Study V, we demonstrated in considerable detail that light consists of massless particles, photons, which travel at the speed of light, Their energies are related to their frequencies v by £ = Av, where f is Planck’s constant, For these particles, the rest mass energy Eo = 0 and w = c. Thus, for massless particles, such as the photon, we find directly and Eyig = Ey (16.70) (16.71) (16.72) It can be scen that there is a real advantage in writing the four-momenta in terms of energies since then we can treat particles and photons on exactly the same basis, Notice also that all masses have been climinated from the expressions. To complete this rewrite of the expressions, let us perform the same change of variable for the expression relating the masses and momenta of the particles in their rest and moving frames (16.48). Equating the norms of the momentum four-vector (16.69) in the laboratory frame and the rest frame of the particle, Cee) =) -(S2) =(= ¢ < © E? — p’c* = E2 = constant. (16.73) This equation is a great help in solving problems in relativity, We can work out the total energy E = 3°, £) and total momentum or 2, p; in some frame of reference and then we know that the quantity £7 — p*c? 7 is an invariant in all inertial frames of reference. ‘This is particularly useful in working out threshold energies for the creation of particles in high-energy collisions, One final note is that, in the extreme high-energy limit E >> Ep, (16.73) becomes. = ple. (16.74) Thus, in the ultrarelativistic limit, particles behave like photons, 16.5.2 16.6 16 Special relativity — a study in invariance Relativistic particle dynamics in a magnetic field Consider next the dynamics of a relativistic particle ina magnetic field. The expression. the Lorentz force is S=eE+ux B) and hence, if E = 0, f =ela x B). Therefore, from (16.56), eu x B) = may? (SS)a+ yaad. (16.78 ‘The left-hand side of this expression and the first term on the right are perpendicular ve because wl (a x B), and hence we require simultaneously that e(ux By=ymoa ands w-a =0, (16.76 that is, the acceleration impressed by the magnetic field is perpendicular to both B and velocity w. This is the origin of the circular or spiral motion of charged particles in magnet fields. The frequency four-vector Finally, let us derive the frequency four-vector from the scalar product rule: namely, if A is a four-vector and A - B is an invariant then B must also be a four-veetor. The simplest approach is to consider the phase of a wave, If we write the expression for the wave in the form exp[i(k +r — c4)] then the quantity A-r — wr is the phase of the wave, in other words, it defines how far through the cycle of the wave from 0° to 360° one happens to be at coordinates [cr, r]. This is an invariant scalar quantity, whatever frame of reference ong happens to be in. Thus, in any inertial frame of reference, A-r — eof = invariam. But [ef, r] is a four-vector and therefore the quantity K =[c/c, ] (16.77) must also be a four-vector; it is called the frequency four-vector. ‘We can also derive this four-vector for photons by deriving the appropriate momentum four-vector, Thus, from (16.71), [" voi. | =/—. i ene ra[5 2 ee 16.7 Lorentz contraction and the origin of magnetic fields 423 where /iv/c is the momentum of the photon and /rv its energy; é; is the unit vector in the direction of propagation of the photon. Writing fy _ ft hv : = nal, e ¢ ¢ ha the momentum four-vector of the photon becomes w pan[2 x], (16.78) that is, PH=hk, 6.7 Lorentz contraction and the origin of magnetic fields We have stated that special relativistic phenomena are outside our normal experience, but there is one remarkable example in which these second-order effects in /care omnipresent in everyday life — they are the origin of magnetic fields, which are associated with electric current, We noted in subsection 16.1.1 the interchangeability of electric and magnetic fields depending upon the inertial frame of reference in which the measurements are made. We need to repeat the analyses of subsection 16.1.1 to work out the transformation laws for E and B according to special relativity. This involves replacing (16.3) by the Lorentz transformations for partial derivatives; these are a as a y a a ae oe) ie) ar ar} (16.79) az \ ae" It is left to the reader to show that Maxwell’s equations are form-invariant under these transformations and that the transformed components of F and B are (16.80) (16.81)

You might also like