Physics 511: Electrodynamics Problem Set #6 Due Friday March 23 (1) Standing Waves (10 points) Consider the

superposition of two counterpropagating and plane waves with the same frequency, whose electric fields have the same polarization
x k -k y z

E0 cos(kz ! "t ) E1 = x

E0 cos(kz + ! t ) E2 = x

(a) What is the total electric field E 3 = E1 + E2 ? What is the total magnetic field, B3? Sketch E3 ( z, t ) and B3 ( z, t ) as a function of z over one wavelength for
t = 0, T / 4, T / 2, 3T / 4, T , where T is the period of oscillation.

(b) What are the electric and magnetic energy densities as a function of z and t., and time averaged? What is the time average energy flux (intensity) - Explain you answer. Now consider now two the superposition to two counterpropagating traveling planes waves that are cross-polarized and 90! out of phase:

x k -k y

E0 cos(kz ! "t ) E1 = x

E0 sin(kz + ! t ) E2 = y

(Next Page) (c) The polarization of this wave is not inform in space. Show that, relative to the +z-direction the polarization varies over one half wavelength from positive-helicity circular (z=0), to linear (z="/8), to negative-helicity circular (z="/4), to linear (z=3"/8), and back to positive-helicity circular (z="/2).

(d) Show that the total field can be written as a superposition of a positive-helicity circular, and a negative-helicity circular polarized standing wave. You should find that the nodes of one standing wave corresponds to the anti nodes of the other. (e) What is the intensity as a function of position? Note: This field configuration is very important in the study of "laser cooling", whereby laser light can be used to cool a gas of atoms.

Problem 2: Spherical Waves (15 points) Consider the wave equation in three dimensions for a scalar field #(r,t).
!2" # 1 $ 2" = 0. v2 $t 2

We seek solutions for monochromatic wave in spherical coordinates, independent of $ and % - this corresponds, e.g., to waves generated by a point source.

(r ) e" i#t . Show that (a) Let ! (r , t ) = !

% 1 ! # 2 !" = 0 , where k = ! / v . r + k 2" 2 r !r $ ! r &

Show that the most general solution can be written as superpositions of

cos( kr " #t " $ ) cos(kr + "t # $ ) and ! 2 (r , t ) = u0 . r r What is the physical difference between #1 and #2 ?

! 1 ( r , t ) = u0

(b) In free space we know that Maxwells Equations imply that the waves are transverse. of ! A first guess at the vector spherical wave would choose the polarization in the ! cos( kr ! "t ) direction. Show that E = E0 # does not satisfy Maxwells Equations. kr (c) The simplest possible vector spherical wave for the electric field in free space is
# sin ! % ) #1% , ( ) E( r, ! , " ) = E0 $ cos kr ' ( t ' sin( kr ' ( t ) ". + . $ kr & kr & * -

(This corresponds to magnetic dipole radiation as we will see)

Show that E obeys all four Maxwells equation, in vacuum, and find the associated magnetic field. (d) Show that in the limit kr << 1 , the magnetic field has the instantaneous form of a )r " m(t ) E0 3( m(t ) ! r z . static dipole field , where m(t ) = 3 cos !t e B(x , t ) = 3 k r (e) Find the time average Poynting vector; does it point in the expected directions and have the expected fall off with r. (f) Find the flux of energy through a sphere, radius R, centered at the origin, and comment on your result.

Problem 3: The Paraxial Wave Equation (15 Points) Plane waves propagate in a unique direction, determined by the wave vector k. However, these waves are unphysical since they extend over all space and thus contain infinite energy. We know that a laser beam travels nearly unidirectionaly, but has extends only over a finite diameter. The propagation of this pencil like beam is described by the paraxial wave equation, which the subject of this problem.

$ 2 1 #2 ' (a) Start with the wave equation & ! " 2 2 E (x, t ) = 0 . % c #t ( E ( x)e ik0 z! "t (real part in the Assume a monochromatic wave and an Ansatz E( x, t ) = e end). Here E ( x) is known as the envelope function (for a plane wave, E ( x) would be
ik z! "t constant), and e 0 as the carrier wave which propagates in the z-direction (which

will call the longitudinal direction). Under slowly varying envelope approximation (SVEA) where the envelope varies of distances large compared to the carrier wave length,

! 2E ! E ! 2E "2 "2 2 2 << k << " E ! = + , , where is the Transverse Laplacian, 0 T T ! z2 !z ! z2 " x2 " y2
!E 1 2 =" # E : the paraxial wave equation !z 2k0 T ( k, ! ) = E ( k " k e )# (! " c k ) . (b) Show that the Fourier transform of the field is, E 0 z
show that the envelope satisfies i

Let q = k ! k0 e z , be the wave vectors of the envelope. Show that the SVEA translates in the Fourier domain, into the paraxial approximation, i.e., all the wave vectors (rays) make no more than a small angle $ with respect to a main carrier wave vector k 0 e z .
! k0

(c) If the characteristic width of the beam in the transverse direction is w0 , use the uncertainty principle to show that ! ~ 1/ k0 w0 = ! / w0 (in optics we call it the 2 diffraction angle). Argue qualitatively after a distance on order z ~ k0 w0 , the beam width spreads substantially (this is known as the Rayleigh range, or diffraction length). (d) Notice that the paraxial wave equation has the form of a Schrdinger equation for a free particle in two dimensions, where z plays the role of time, and (x,y) are spatial dimensions. Given the transverse field profile at the input plane z=0, E ( x T ,0) , we can propagate the envelope forward to any other plane as we evolve a wave packet in quantum mechanics:

E( x T , z ) = e

iz!2 T

/ (2k0 )

E ( x T ,0) = #

d qT e (2 " )2

$i

qT z 2k0

(q ) e iq %x , E T

( q ) is the Fourier transform of the input beam E ( x ,0) . Given the initial where E T T 2 profile of a Gaussian envelope: E ( x T ,0) = E0 exp(!( x 2 + y 2 ) / w0 ) , show that
E ( x T , z ) = E0 e! i"( z)
2

w0 exp( !( x 2 + y 2 ) / w2 ( z)) exp(ik 0 ( x 2 + y 2 ) / 2 R(z )) w( z)

2 where w (z) = w0 (1 + z / z0 ) , R( z) = z + z0 / z , ! (z) = tan "1( z / z0 ) ,with 2 z0 = k0 w0 / 2 (the Rayleigh Range). Interpret.

(e) Show that for z >> z0 , the wave fronts are approximately spherical. (f) There is a fudge in our derivation. We started with the wave equation which assumed in fact transverse waves. Show that our Ansatz with uniform transverse polarization e violates ! " E = 0 . How can we resolve this dilemma?

Problem 4: Angular momentum in electromagnetic waves (15 points) (From Jackson 3rd Edition: Problem 7.27). The angular momentum of the electromagnetic field is L = the integration is over all space. (a) Eliminate the magnetic field in favor of the vector potential. Show that for field localized to a finite region of space,

1 d3 x x " ( E " B) , where 4!c #

L=

1 d3 x [E " A + Ei (x " # ) Ai ] (sum over components i implicit). $ 4!c

The second term is referent to as the "orbital" angular momentum because of the presence of the orbital angular momentum operator familiar in wave mechanics, L op = !i(x " # ) . The first term relates to the vector nature of the filed itself as is referred to as the "spin" angular momentum. (b) Consider a Fourier decomposition of the vector potential into transverse, circularly polarized plane waves d 3k ik"x #i$t + c.c. , A(x, t ) = & % 3 A (k) e ( k) e = (2! ) . ) are circular polarizations orthogonal to k where e ( k

Show that the time average of the "spin" angular momentum is

Lspin

2 1 d 3k # ( k) 2 & . = k A ( k ) " A 3 + " ( $ ' 2!c ) (2! ) %

(c) Calculate the total energy in the field using the expansion in (b). If we associate, !! according to quantum mechanics, an energy density of per mode (i.e. Fourier V component), what is the "spin" angular momentum of a positive or negative helicity electromagnetic mode, both in direction and magnitude? These are the fundamental characteristics of "photons".