IPhO 1998 · Theory problem 3

Faster than light?

In this problem we analyze and interpret measurements made in 1994 on radio wave emission from a compound source within our galaxy.

The receiver was tuned to a broad band of radio waves of wavelengths of several centimeters. Figure 3.1 shows a series of images recorded at different times. The contours indicate constant radiation strength in much the same way as do altitude contours on a geographical map. In the figure the two maxima are interpreted as showing two objects moving away from a common center shown by crosses in the images. (The center, which is assumed to be fixed in space, is also a strong radiation emitter but mainly at other wavelengths). The measurements conducted on the various dates were made at the same time of day.

The scale of the figure is given by a line segment showing one arc second (as). (1 as = 1/3600 of a degree). The distance to the celestial body at the center of the figure, indicated by crosses, is estimated to be \(R=12.5\ \mathrm{kpc}\). A kiloparsec (kpc) equals \(3.09\cdot10^{19}\ \mathrm{m}\). The speed of light is \(c=3.00\cdot10^8\ \mathrm{m/s}\). Error calculations are not required in the solution.

a) (2 points) We denote the angular positions of the two ejected radio emitters, relative to the common center, by \(\theta_1(t)\) and \(\theta_2(t)\), where the subscripts 1 and 2 refer to the left and right hand ones, respectively, and \(t\) is the time of observation. The angular speeds, as seen from the Earth, are \(\omega_1\) and \(\omega_2\). The corresponding apparent transverse linear speeds of the two sources are denoted by \(v'_{1,\perp}\) and \(v'_{2,\perp}\).

Using Figure 3.1, make a graph to find the numerical values of \(\omega_1\) and \(\omega_2\) in milli-arc-seconds per day (mas/d). Also determine the numerical values of \(v'_{1,\perp}\) and \(v'_{2,\perp}\), and write all answers on the answer sheet. (You may be puzzled by some of the results).

b) (3 points) In order to resolve the puzzle arising in part (a), consider a light-source moving with velocity \(\vec v\) at an angle \(\phi\) (\(0\leq\phi\leq\pi\)) to the direction towards a distant observer \(O\) (Figure 3.2). The speed may be written as \(v=\beta c\), where \(c\) is the speed of light. The distance to the source, as measured by the observer, is \(R\). The angular speed of the source, as seen from the observer, is \(\omega\), and the apparent linear speed perpendicular to the line of sight is \(v'_\perp\).

Find \(\omega\) and \(v'_\perp\) in terms of \(\beta\), \(R\) and \(\phi\) and write your answer on the answer sheet.

c) (1 point) We assume that the two ejected objects, described in the introduction and in part (a), are moving in opposite directions with equal speeds \(v=\beta c\). Then the results of part (b) make it possible to calculate \(\beta\) and \(\phi\) from the angular speeds \(\omega_1\) and \(\omega_2\) and the distance \(R\). Here \(\phi\) is the angle defined in part (b), for the left hand object, corresponding to subscript 1 in part (a).

Derive formulas for \(\beta\) and \(\phi\) in terms of known quantities and determine their numerical values from the data in part (a). Write your answers in the designated fields on the answer sheet.

d) (2 points) In the one-body situation of part (b), find the condition for the apparent perpendicular speed \(v'_\perp\) to be larger than the speed of light \(c\).

\(^{7}\) Authors: Einar Gudmundsson, Knútur Árnason and Thorsteinn Vilhjálmsson

Figure 1

Figure 3.1: Radio emission from a source in our galaxy.

Figure 2

Figure 3.2: The observer is at \(O\) and the original position of the light source is at \(A\). The velocity vector is \(\vec v\).

Write the condition in the form \(\beta>f(\phi)\) and provide an analytic expression for the function \(f\) on the answer sheet.

Draw on the graph answer sheet the physically relevant region of the \((\beta,\phi)\)-plane. Show by shading in which part of this region the condition \(v'_\perp>c\) holds.

e) (1 point) Still in the one-body situation of part (b), find an expression for the maximum value \((v'_\perp)_{\max}\) of the apparent perpendicular speed \(v'_\perp\) for a given \(\beta\) and write it in the designated field on the answer sheet. Note that this speed increases without limit when \(\beta\to1\).

f) (1 point) The estimate for \(R\) given in the introduction is not very reliable. Scientists have therefore started speculating on a better and more direct method of determining \(R\). One idea for this goes as follows. Assume that we can identify and measure the Doppler shifted wavelengths \(\lambda_1\) and \(\lambda_2\) of radiation from the two ejected objects, corresponding to the same known original wavelength \(\lambda_0\) in the rest frames of the objects.

Starting from the equations for the relativistic Doppler shift, \(\lambda=\lambda_0(1-\beta\cos\phi)(1-\beta^2)^{-1/2}\), and assuming, as before, that both objects have the same speed, \(v\), show that the unknown \(\beta=v/c\) can be expressed in terms of \(\lambda_0,\lambda_1,\) and \(\lambda_2\) as

\[ \beta=\sqrt{1-\frac{\alpha\lambda_0^2}{(\lambda_1+\lambda_2)^2}}. \tag{3.1} \]

Write the numerical value of the coefficient \(\alpha\) in the designated field on the answer sheet.

You may note that this means that the suggested wavelength measurements will in practice provide a new estimate of the distance.

Official English text of the problem.

Answers

  • a) Representative graphical values are \[ \boxed{\omega_1\simeq17.5\ \mathrm{mas/d},\qquad \omega_2\simeq9.0\ \mathrm{mas/d}.} \] Consequently, \[ \boxed{v'_{1,\perp}\simeq3.8\times10^8\ \mathrm{m/s}\simeq1.3c,\qquad v'_{2,\perp}\simeq2.0\times10^8\ \mathrm{m/s}\simeq0.65c.} \]

  • b) \[ \boxed{\omega=\frac{\beta c\sin\phi}{R(1-\beta\cos\phi)},\qquad v'_\perp=\frac{\beta c\sin\phi}{1-\beta\cos\phi}.} \]

  • c) With angular speeds expressed in radians per unit time, define \[ X=\frac{\omega_1-\omega_2}{\omega_1+\omega_2}, \qquad Y=\frac{2R\omega_1\omega_2}{c(\omega_1+\omega_2)}. \] Then \[ \boxed{\beta=\sqrt{X^2+Y^2}\simeq0.92,\qquad \phi=\arccos\!\left(\frac{X}{\beta}\right)\simeq70^\circ.} \] The left-hand object is approaching.

  • d) The superluminal region is \[ \boxed{0<\phi<\frac{\pi}{2},\qquad \frac{1}{\sin\phi+\cos\phi}<\beta<1.} \] Thus \(f(\phi)=1/(\sin\phi+\cos\phi)\) on the relevant angular interval. Such motion requires \(\beta>1/\sqrt2\).

  • e) \[ \boxed{(v'_\perp)_{\max}=\frac{\beta c}{\sqrt{1-\beta^2}},\qquad \phi_{\max}=\arccos\beta.} \]

  • f) \[ \boxed{\alpha=4,\qquad \beta=\sqrt{1-\frac{4\lambda_0^2}{(\lambda_1+\lambda_2)^2}}.} \]

Solution

a) Angular motions and apparent transverse speeds

Take \(t=0\) on 27 March. Measure each component’s angular distance from the cross in its own image, using the full marked scale length as \(1''=1000\ \mathrm{mas}\). Follow the two outward-moving components indicated by the dashed tracks. We use positive angular distances on both sides of the center.

Approximate readings from the supplied figure are:

Date \(t\) (days) \(\theta_1\) (arcseconds) \(\theta_2\) (arcseconds)
27 March 0 0.12 0.07
3 April 7 0.22 0.14
9 April 13 0.35 0.20
16 April 20 0.46 0.26
23 April 27 0.61 0.31
30 April 34 0.70 0.38

Plot \(t\) horizontally and angular distance vertically. Each set of points lies approximately on a straight line; the left-hand component has the greater slope. Representative lines are

\[ \frac{\theta_1}{\mathrm{arcsec}} \simeq0.11+0.0175\,\frac{t}{\mathrm d}, \qquad \frac{\theta_2}{\mathrm{arcsec}} \simeq0.08+0.0090\,\frac{t}{\mathrm d}. \]

Thus their vertical intercepts are approximately \(0.11''\) and \(0.08''\), and their slopes give

\[ \omega_1\simeq17.5\ \mathrm{mas/d}, \qquad \omega_2\simeq9.0\ \mathrm{mas/d}. \]

The angles are extremely small, so the transverse displacement corresponding to an angular displacement \(\theta\) is \(R\theta\). Consequently,

\[ v'_{i,\perp}=R\omega_i, \]

where \(\omega_i\) must be expressed in radians per second. The required conversions are

\[ R=12.5(3.09\times10^{19}) =3.8625\times10^{20}\ \mathrm m, \]

\[ 1\ \mathrm{mas} =\frac{\pi}{180\times3600\times1000}\ \mathrm{rad} =4.8481\times10^{-9}\ \mathrm{rad}, \]

and therefore

\[ 1\ \mathrm{mas/d} =5.6113\times10^{-14}\ \mathrm{rad/s}. \]

At the stated distance, \(1\ \mathrm{mas/d}\) corresponds to

\[ R(5.6113\times10^{-14}) =2.1674\times10^7\ \mathrm{m/s}. \]

Hence

\[ v'_{1,\perp} \simeq17.5(2.1674\times10^7) \simeq3.8\times10^8\ \mathrm{m/s} \simeq1.3c, \]

\[ v'_{2,\perp} \simeq9.0(2.1674\times10^7) \simeq2.0\times10^8\ \mathrm{m/s} \simeq0.65c. \]

The first value exceeds \(c\). These speeds use the time interval between the arrivals of signals at Earth; that interval differs from the interval between their emissions.

b) Effect of light-travel time

Work in the inertial frame in which the observer and central source are stationary. Let two signals be emitted a time \(\Delta t_{\mathrm e}\) apart in this frame.

During this interval, the emitter moves a transverse distance

\[ \Delta x_\perp=v\sin\phi\,\Delta t_{\mathrm e}. \]

Its displacement toward the observer is

\[ \Delta x_\parallel=v\cos\phi\,\Delta t_{\mathrm e}. \]

For a distant observer, the second signal’s path is shorter by approximately \(\Delta x_\parallel\). The arrival-time interval is therefore

\[ \Delta t_{\mathrm{obs}} =\Delta t_{\mathrm e}-\frac{\Delta x_\parallel}{c} =(1-\beta\cos\phi)\Delta t_{\mathrm e}. \]

The same expression covers a receding source: then \(\cos\phi<0\), so the second signal travels farther.

The angular displacement is, in the small-angle approximation,

\[ \Delta\theta\simeq\frac{\Delta x_\perp}{R}. \]

Thus

\[ \boxed{ \omega=\frac{\Delta\theta}{\Delta t_{\mathrm{obs}}} =\frac{\beta c\sin\phi}{R(1-\beta\cos\phi)} } \]

and

\[ \boxed{ v'_\perp=R\omega =\frac{\beta c\sin\phi}{1-\beta\cos\phi}. } \]

Here the source’s displacement is negligible compared with \(R\), allowing \(R\) and \(\phi\) to be treated as constant over the observations.

For an approaching emitter, successive signals arrive closer together than they were emitted. Dividing the transverse displacement by this shortened interval can give \(v'_\perp>c\), although the actual emitter speed remains \(v=\beta c<c\).

c) Recovering the true speed and direction

The two directions make angles \(\phi\) and \(\pi-\phi\) with the direction toward the observer. Their transverse velocity magnitudes are equal, while their radial components have opposite signs. Therefore,

\[ \frac{R\omega_1}{c} =\frac{\beta\sin\phi}{1-\beta\cos\phi}, \qquad \frac{R\omega_2}{c} =\frac{\beta\sin\phi}{1+\beta\cos\phi}. \]

Set

\[ X=\beta\cos\phi,\qquad Y=\beta\sin\phi. \]

Then

\[ \omega_1(1-X)=\omega_2(1+X)=\frac{cY}{R}. \]

Equating the first two expressions gives

\[ \boxed{ X=\beta\cos\phi =\frac{\omega_1-\omega_2}{\omega_1+\omega_2}. } \]

Substitution into either angular-speed equation gives

\[ \boxed{ Y=\beta\sin\phi =\frac{2R\omega_1\omega_2}{c(\omega_1+\omega_2)}. } \]

Consequently,

\[ \boxed{ \beta= \frac{\sqrt{(\omega_1-\omega_2)^2+ \left(2R\omega_1\omega_2/c\right)^2}} {\omega_1+\omega_2} } \]

and an expression valid throughout \(0\leq\phi\leq\pi\) is

\[ \boxed{ \phi=\arccos\!\left[ \frac{\omega_1-\omega_2} {\beta(\omega_1+\omega_2)} \right]. } \]

For these observations, \(\omega_1>\omega_2\), so \(\phi\) is acute. We can also use

\[ \tan\phi=\frac{2R\omega_1\omega_2}{c(\omega_1-\omega_2)}. \]

Using the representative rates from (a),

\[ X\simeq\frac{17.5-9.0}{17.5+9.0} \simeq0.321, \]

\[ Y\simeq0.859. \]

Thus

\[ \beta\simeq\sqrt{0.321^2+0.859^2} \simeq0.917, \qquad \phi\simeq69.5^\circ. \]

To the precision justified by the graph,

\[ \boxed{\beta\simeq0.92,\qquad \phi\simeq70^\circ.} \]

The true speed is approximately \(2.8\times10^8\ \mathrm{m/s}\). Since \(\beta\cos\phi>0\), the left-hand component is approaching and the right-hand component is receding.

d) Condition and region for apparent superluminal motion

For a material emitter, \(0\leq\beta<1\). Hence \(1-\beta\cos\phi>0\), and the condition \(v'_\perp>c\) becomes

\[ \frac{\beta\sin\phi}{1-\beta\cos\phi}>1, \]

or

\[ \boxed{\beta(\sin\phi+\cos\phi)>1.} \]

For \(\phi\geq\pi/2\), the denominator of \(v'_\perp/c\) is at least one, while its numerator is at most \(\beta<1\). Superluminal apparent motion is therefore impossible for a receding or exactly transverse source. It is also impossible at \(\phi=0\), where the transverse motion vanishes.

Within the relevant interval \(0<\phi<\pi/2\),

\[ \boxed{\beta>f(\phi),\qquad f(\phi)=\frac{1}{\sin\phi+\cos\phi}.} \]

The complete physical condition is thus

\[ \boxed{ 0<\phi<\frac{\pi}{2}, \qquad \frac{1}{\sin\phi+\cos\phi}<\beta<1. } \]

For the requested diagram, the physically relevant region has

\[ 0\leq\beta<1,\qquad 0\leq\phi\leq\pi. \]

With \(\phi\) horizontal and \(\beta\) vertical, draw the threshold curve over \(0\leq\phi\leq\pi/2\). Its useful points are:

\(\phi\) \(f(\phi)\)
\(0\) \(1\)
\(\pi/6\) \(0.732\)
\(\pi/4\) \(1/\sqrt2\simeq0.707\)
\(\pi/3\) \(0.732\)
\(\pi/2\) \(1\)

The curve is symmetric about \(\phi=\pi/4\). It decreases from \(1\) to its minimum \(1/\sqrt2\), then rises to \(1\). Shade the region strictly above this curve and strictly below \(\beta=1\). The threshold curve itself has \(v'_\perp=c\), so it is excluded from the shading. No region at \(\phi\geq\pi/2\) is shaded.

Since

\[ \sin\phi+\cos\phi =\sqrt2\cos\!\left(\phi-\frac{\pi}{4}\right)\leq\sqrt2, \]

such a region exists only when

\[ \boxed{\beta>\frac{1}{\sqrt2}.} \]

Equivalently, for any fixed \(\beta>1/\sqrt2\), its angular extent is

\[ \left|\phi-\frac{\pi}{4}\right| < \arccos\!\left(\frac{1}{\sqrt2\,\beta}\right). \]

e) Maximum apparent transverse speed

At fixed \(0<\beta<1\), differentiate the result of (b):

\[ \frac{\mathrm d}{\mathrm d\phi} \left(\frac{v'_\perp}{c}\right) = \frac{\mathrm d}{\mathrm d\phi} \left(\frac{\beta\sin\phi}{1-\beta\cos\phi}\right) = \frac{\beta(\cos\phi-\beta)} {(1-\beta\cos\phi)^2}. \]

The derivative changes from positive to negative where

\[ \cos\phi=\beta. \]

Because the apparent transverse speed vanishes at both \(\phi=0\) and \(\phi=\pi\), this stationary point is the global maximum. At that angle,

\[ \sin\phi=\sqrt{1-\beta^2}, \qquad 1-\beta\cos\phi=1-\beta^2. \]

Therefore,

\[ \boxed{ (v'_\perp)_{\max} =\frac{\beta c}{\sqrt{1-\beta^2}}, \qquad \phi_{\max}=\arccos\beta. } \]

As \(\beta\to1^{-}\), the denominator tends to zero and the maximum apparent speed increases without limit. For \(\beta=0\), the apparent speed is zero at every angle.

f) Determination of \(\beta\) from the Doppler shifts

Let

\[ \gamma=\frac{1}{\sqrt{1-\beta^2}}. \]

For the first object, the given Doppler formula is

\[ \lambda_1=\lambda_0\gamma(1-\beta\cos\phi). \]

For the oppositely moving object, \(\cos(\pi-\phi)=-\cos\phi\), giving

\[ \lambda_2=\lambda_0\gamma(1+\beta\cos\phi). \]

Adding these expressions eliminates the direction:

\[ \lambda_1+\lambda_2 =2\lambda_0\gamma =\frac{2\lambda_0}{\sqrt{1-\beta^2}}. \]

Thus

\[ 1-\beta^2 =\frac{4\lambda_0^2}{(\lambda_1+\lambda_2)^2}, \]

and, taking the nonnegative speed,

\[ \boxed{ \beta=\sqrt{1-\frac{4\lambda_0^2}{(\lambda_1+\lambda_2)^2}}, \qquad \alpha=4. } \]

This also explains how the distance could be obtained. The angular motions determine

\[ X=\beta\cos\phi =\frac{\omega_1-\omega_2}{\omega_1+\omega_2}, \]

while the wavelengths determine \(\beta\). Therefore \(\beta\sin\phi=\sqrt{\beta^2-X^2}\), and the relation derived in (c) gives

\[ R= \frac{c(\omega_1+\omega_2)}{2\omega_1\omega_2} \sqrt{\beta^2-X^2}. \]

Thus the combined angular and wavelength measurements determine \(R\) without adopting its initial estimate.

The solution was checked against the official solution.