Physics 210A PS 7 Solutions: Fourier Transform & Diffusion Eq. by A. Karmis, Lab Reports of Physics

Solutions to problem 7 in physics 210a by anthony karmis, focusing on the application of fourier transforms to the diffusion equation. The definition of the fourier transform, the solution to the diffusion equation using the fourier transform, and the inverse transform. The document also discusses the green function and its relation to the solution.

Typology: Lab Reports

Pre 2010

Uploaded on 09/17/2009

koofers-user-7xr
koofers-user-7xr 🇺🇸

4

(1)

10 documents

1 / 9

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Physics 210A PS 7 Solutions
Anthony Karmis
March 9, 2009
Jackson 6.3
Part (a)
Defining the Fourier transform by:
~
A(~x, t) = 1
(2π)3Z~
A(~
k, t)e~
k·~xd3k
Therefore, the diffusion equation becomes:
t~
A=k2
µσ ~
A
This is solved by ~
A(~
k, t) = ~
A(~
k, 0)ek2t/µσ .
The inverse transform gives us:
~
A(~x, t) = 1
(2π)3Z~
A(~
k, 0)ek2t/µσ ei~
k·~xd3k
where
~
A(~
k, 0) = Z~
A(~x,0)ei~
k·~xd3x
The result is:
~
A(~x, 0) = 1
(2π)3Z Z ~
A(~x,0)ek2t/µσei~
k·(~x ~x )d3kd3x
=ZG(~x ~x, t)~
A(~x,0)d3x
With the Green function:
G(~x ~x, t) = 1
(2π)3Zek2t/µσei~
k·(~x ~x )d3k
1
pf3
pf4
pf5
pf8
pf9

Partial preview of the text

Download Physics 210A PS 7 Solutions: Fourier Transform & Diffusion Eq. by A. Karmis and more Lab Reports Physics in PDF only on Docsity!

Physics 210A PS 7 Solutions

Anthony Karmis

March 9, 2009

Jackson 6.

Part (a)

Defining the Fourier transform by:

A(~x, t) =

(2π)

3

A(

k, t)e

−~k·~x

d

3

k

Therefore, the diffusion equation becomes:

t

A = −

k

2

μσ

A

This is solved by

A(

k, t) =

A(

k, 0)e

−k

2 t/μσ .

The inverse transform gives us:

A(~x, t) =

(2π)

3

A(

k, 0)e

−k

2 t/μσ e

i~k·~x d

3 k

where

A(

k, 0) =

A(~x

′ , 0)e

−i

~ k·~x

d

3 x

The result is:

A(~x, 0) =

(2π)

3

A(~x

′ , 0)e

−k

2 t/μσ e

i

~ k·(~x−~x

′ ) d

3 kd

3 x

G(~x − ~x

′ , t)

A(~x

′ , 0)d

3 x

With the Green function:

G(~x − ~x

′ , t) =

(2π)

3

e

−k

2 t/μσ e

i

~ k·(~x−~x

′ ) d

3 k

Part (b)

Introducing the Fourier transform

G(~x, t) =

(2π)

3

G(

k, ω)e

i(

~ k·~x−ωt) d

3 kdω

The inhomogeneous equation in Jackson becomes:

(−iω)

2 − |ik|

2 /μσ

G = e

−i

~ k·~x

Which gives the Green function:

G(

k, ω) =

e

−i

~ k~x

k

2 /μσ − iω

We may invert the transform by performing the ω integral. We have:

G(

k, t) =

2 π

−∞

G(

k, ω)e

−iωt dω =

ie

−i

~ k·~x

2 π

e

−iωt

ω + ik

2 /μσ

We can compute this integral by contour integration. For t > 0, Jordan’s

lemma tells us to close the contour in the lower-half plane. As a result, we pick

up the residue at ω = −ik

2 /μω. The result is:

G(

k, t) = (− 2 πi)

ie

−i

~ k·~x

2 π

e

−k

2 t/μσ = e

−k

2 t/μσ e

−i

~ k·~x

However, for t < 0, we close the contour in the upper-half plane. Since there

are no enclosed poles, we end up with G = 0. Therefore, we have:

G(~x − ~x

′ , t) =

Θ(t)

(2π)

3

e

−k

2 t/μσ e

i

~ k·(~x−~x

′ ) d

3 k

Part (c)

Taking μ and σ to be uniform in all space, we may perform the above integral

by completing the square:

G(~x − ~x

′ , t) =

Θ(t)

(2π)

3

e

−μσ|~x−~x

′ |

2 / 4 t

e

−t|

~ k−iμσ(~x−~x

′ )/ 2 t|

2 /μσ d

3 k

Θ(t)

(2π)

3

πμσ

t

3 / 2

e

−μσ|~x−~x

′ |

2 / 4 t

= Θ(t)

μσ

4 πt

3 / 2

e

−μσ|~x−~x

′ |

2 / 4 t

Jackson 6.

Part (a)

We assume the sphere is magnetized and spinning along the ˆz-axis. Since the

magnetic moment is m~ =

M V , where V =

4

3

πr

3 is the volume of the sphere, we

see that the magnetization is simply

M = M zˆ. As seen in Jackson, the sphere

has a uniform magnetic induction

B =

2

3

μ 0

M in its interior. In terms of m, we

have:

B =

μ 0 m

2 πR

3

ˆz

Since there will be no current in the frame where the sphere is at rest, Ohm’s

law in the reference frame tells us that σ

E

′ = σ(

E+~v×

B) = 0, where the primed

variables are in the rotating reference frame, and the nonprimed variables are

in the at-rest lab frame. Therefore, we have:

E = −~v ×

B = −(~ω × ~r) ×

B = −

μ 0

2 πR

3

(~r − zˆ(ˆz · ~r))

In cylindrical coordinates, this is:

E = −

μ 0 mωρ

2 πR

3

ρˆ

Using ρ = ǫ 0

E, we have:

ρ = ǫ 0

E = ǫ 0

ρ

∂ρ (ρEρ) = −

πc

2 R

3

Part (b)

First, we note that the interior electric field can be integrated to obtain the

interior electrostatic potential:

Φ(ρ) = −

E · d

l = −

Erhodρ = Φ 0 +

μ 0

mωρ

2

4 πR

3

Converting back to spherical coordinates, we have:

Φ(r, θ) = Φ 0

μ 0

4 πR

3

r

2 sin

2 θ

μ 0

6 πR

3

r

2 [P 0 (cosθ) − P 2 (cosθ)]

where we have converted sin

2 θ into Legendre polynomials. This can be

written explicitly as a Legendre expansion:

Φ(r, θ) =

0

μ 0 mω

6 πR

3

r

2

P

0

(cosθ) −

μ 0 mω

6 πR

3

r

2 P 2

(cosθ)

Since charge neutrality requires that the monopole moment be zero, this sets

0

for us, and we have:

Φ(r, θ) = −

μ 0 mωr

2

6 πR

3

P 2 (cosθ)

This potential clearly only has a quadrupole term:

4 π

μ 0 mωR

2

6 π

Y 2 , 0 (θ, φ)

r

3

Comparing this with the multipole expansion,

4 πǫ 0

l,m

4 π

2 l + 1

ql,m

Y

l,m

(θ, φ)

r

l+

We see we have:

q 2 , 0

= − 4 πǫ 0

4 π

μ 0

mωR

2

6 π

4 π

2 mωR

2

3 c

2

Converting this to cartesian tensors yields:

Q 33 = 2

4 π

q 2 , 0 = −

4 mωR

2

3 c

2

and

Q

11

= Q

22

Q

33

Part (c)

The surface charge may be computed by first obtaining the jump in the normal

component of the electric field at the surface of the sphere. Working in spherical

coordinates, and taking the gradient of the potential above, we have:

E

out

r

μ 0 mωR

2

2 πr

4

P 2 (cosθ)

E

in

r

μ 0 mωr

3 πR

3

[P 0 (cosθ) − P 2 (cosθ)]

The surface charge is then computed as:

σ = ǫ 0

E

out

r

− E

in

r

|r=R

μ 0 ǫ 0 mω

3 πR

2

[

P 2 (cosθ) − (P 0 (cosθ) − P 2 (cosθ))

]

3 πc

2 R

2

[

P 0 (cosθ) −

P 2 (cosθ)

]

Part (d)

Although the sphere is rotating, both the electric and the magnetic field are

static. Hence, the line integral of the electric field gives simply the electrostatic

potential. In this case, we have:

ξ =

pole

equator

E · d

l = Φequator − Φpole = Φ(θ =

π

) − Φ(θ = 0)

Using our results from Part (b), this becomes:

ξ = −

μ 0 mω

6 πR

[P

2

(0) − P

2

(1)] =

μ 0 mω

4 πR

Part (c)

The surface contribution is:

c

2

∇ × (Φ

H) = −

c

2

∇Φ) ×

H +

∇ ×

H

d

3

x

μ 0

c

2

E ×

Bd

3

x +

c

2

E · ~x)

Jd

3

x

The first integral is given by Jackson 5.62, and we saw the second in Part

(b). Therefore, we have:

c

2

∇ × (Φ

H) = −

3 c

2

E × m~

Therefore, we see that:

P =

3 c

2

E(0) × m~

Jackson 6.

Part (a)

The momentum conservation equation reads:

d

dt

P

f ields

P

mech.

j

T

ij

n j

da = −

T ) · ~nda

This tells us that the pressure due to the fields incident on the plane is Tij nj.

If we choose the wave to be propagating in the z-direction and the plane normal

to it, then the pressure is T 33

. Jackson 6.120 yields:

P = T

33

ǫ 0

E

2

  • c

2 B

2

which is the energy density.

Part (b)

The flux is just the rate at which energy is carried through space. Hence, the

flux is Φ = uc, where u is the energy density. Let the mass density of the sail

be δ. Then, the acceleration is:

a =

F

m

P

δ

(1400 W/m

3 )

(3 × 10

8 m/s) (0. 001 kg/m

2 )

≈ 5 × 10

− 3 m/s

2