0% found this document useful (0 votes)
93 views13 pages

Lec 14

1. The document discusses several examples of "tricky potentials" including central potentials, dissipative media, rapidly oscillating fields, and the Lorentz force. 2. For a charged particle constrained to move along a helical wire in a uniform magnetic field, the potential energy due to the Lorentz force does not change the particle's dynamics. This is because the velocity is always perpendicular to the magnetic field, so no work is done. 3. For a charged particle free to move in the y-z plane with a magnetic field in the x direction, the Lagrangian includes a term for the magnetic vector potential. The forces on the particle include a Lorentz force term.

Uploaded by

semabay
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
93 views13 pages

Lec 14

1. The document discusses several examples of "tricky potentials" including central potentials, dissipative media, rapidly oscillating fields, and the Lorentz force. 2. For a charged particle constrained to move along a helical wire in a uniform magnetic field, the potential energy due to the Lorentz force does not change the particle's dynamics. This is because the velocity is always perpendicular to the magnetic field, so no work is done. 3. For a charged particle free to move in the y-z plane with a magnetic field in the x direction, the Lagrangian includes a term for the magnetic vector potential. The forces on the particle include a Lorentz force term.

Uploaded by

semabay
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

(14) TrickyPotentials

1 Kaptiza Example
Let’stryaslightlydifferentpendulumsystemthistimeforournextexample.

Kapitza

1 � 2 2
 1
T = m ẋm + ẏm , U = mgym , D = bφ̇2
2 2
xm = l sin φ , ym = yd − l cos φ
ẋm = l cos φφ̇ , ẏm = ẏd + l sin φφ̇

1
TheLagrangianforthissystem,afterdroppingtermswhichcontainonlyy d ,which
isanexplicitfunctionoftime,is

1 2 2 
L = m l φ̇ + 2l sin φ φ̇ẏd + mgl cos φ
  2
d ∂L  
¨
= ml lφ + cos φφ̇ẏd + sin φÿd
dt ∂φ̇
∂L ∂D  
− = ml cos φφ̇ẏd − g sin φ − bφ̇
∂φ ∂φ̇
whichgivesustheequationofmotion.
A vertically drivenpendulum is a bit of astrangething; itdoesn’t seem to work
asadriver!

b ˙
lφ¨ + φ + (g + ÿd ) sin φ = 0
lm
b
wherethedampingtermis lm φ̇.
Instead, the drive appears to modify gravity. This makes sense, due to the
equivalenceprinciple.Interestingly,thislet’susexploreparametricresonance...
Noticehowthependulumbecomesexcitedwithadriveattwicetheresonance
frequency.

We won’t cover parametric resonance further, but LL27 does. Methods for
understandingnon-linear/anharmonicbehaviorare alsocovered in LL28-29, but I
found the mathunenelightening, so I won’t try toreproduce ithere.

2
Wealsoseestrangebehaviorforahighfrequency drive. Dampingisnotim-
portantforthis,solet’s operatewithb = 0. We canunderstandthis by noticingthat
thependulum’smotionconsistsofahighfrequency part(atthedrivefrequency)
andalowfrequency part(swingingaround).

φ (t) ' φ1 (t) + φ2 (t)


whereφ 1 correspondstoslowoscillations,andφ 2 to fast
lφ¨ + (g + ÿd ) sin φ = 0

 
¨ ¨
l φ1 + φ2 + (g + ÿd ) sin (φ1 + φ2 ) = 0
assumeφ 1 ∼ const,andφ 2  1
lφ¨2 + (g + ÿd ) (sin φ1 + cos φ1 φ2 ) = 0

fory d = ad cos (ωt),


ÿd = −ad ω 2 cos ωt = −ω 2 yd

Fastoscillationterms,tofirst order, are

lφ¨2 + g cos φ1 φ2 = ad ω 2 sin φ1 cos ωt


drivenresponse:
2
⇒ φ2 ' l(ωa2dcos
ω sin φ1
φ1 −ω 2 )
cos ωt ' − sin φ1 yld for ω  ω 0
0

Graphicallythisresultis

3
HighFrequency Drive

Returning toourequations ofmotion, but keepingφ 2  1 and φ¨2 = − sin φ1 ÿld

�  � 
lφ̈1 + ω 2 sin φ1 yd + g − ω 2 yd sin φ1 1 − cos φ1 yld = 0
� 
lφ¨1 + g sin φ1 + sin 2φ1 ω 2 y 2 − gyd = 0
2l d

Wearelookingfortheslowbehavior,solet’s average overthe fastdriveperiod.


Sincey d = ad cos (ωt)

SlowBehavior

4
¨ g 1  ad ω 2
⇒ φ1 + sin φ1 + sin (2φ1 ) = 0
l 4 l

Ifweareclosetoφ 1 ' π (pointingup)

forφ 1 = π + ε withε  1 ⇒ sin φ 1 ' −ε, sin 2φ1 ' 2ε


 �  
1 ad ω 2 g
⇒ ε̈ + 2 l − l ε=0

� 2 g
⇒ oscillatorwith ω 02 = 21 adlω − l
stableifa ω
2 2
d > gl

So,aswehaveseen,the Kapitzapendulumisstablearoundφ ∼ π (i.e. inverted)


given asufficiently fastdrive.
Generally,whentreatingmotioninarapidlyoscillatingfield,wecandefinean
effectivepotential

Uef f = U + T¯ = U (qslow ) + 12 mdotq ¯2


fast
whereq fast isthe fastpart ofq(t) = q slow (t) + qfast (t)

forus,thiswouldbe

 2
1 2
E= 2 ml φ̇1 + φ̇2 + mg (yd − l cos(φ1 + φ2 ))
averageoverfastoscillations(y¯ d = 0)
1 1
⇒ E = m(lφ̇1 )2 + m(lφ̇2 )2 − mgl cos(φ1 )
|2 {z } |2 {z }
T Ueff

5
1
Uef f = −mgl cos φ1 + m(− sin φ1 y˙d )2
2  2
1 1
= −mgl cos φ1 + m sin2 φ1 ad ω
2 2

Note:

∂Uef f � 2
= mgl sin φ1 + m sin (2φ) 21 ad ω
∂φ
dividebyml togetourequationofmotionforφ 1

2 Tricky Potentials
Alreadyinthiscoursewehaveseenafewtricky potentials.
Forcentralpotentials,angularmomentugivesusan effective potentialforr.

Central
    1  ˙ 2
˙ ˙
Uef f ~q, ~q = Uef f r, φ = U (r) + µ rφ
2
L2z
⇒ Uef f (r, Lz ) = U (r) +
2µr2
A homogeneousdissipative medium,whichconvertskineticenergyofthemacro-
scopic to kinetic energy of the microscopic (i.e. heat) can also be treated as a
velocitydependentpotential

Dissipative
Z
Udiss (q, q̇) = Ucon (q) − D (q̇) dt
 
d ∂Lcon ∂Lcon ∂D
⇒ = −
dt ∂q̇ ∂q ∂q̇

6
And last time we saw another sort of tricky potential for rapidly oscillating
fieldsordrives

RapidDrive
Uef f (q, q̇) = U (q) + 12 mq̇f2 ast
forq(t) = q slow (t) + qf ast (t)

Theseareallscalars. NeedtonotethatU ef f comesfromincludingthesameT ,


whichcanbewrittenasT (q),inU togetU ef f = U (q) + T (q).

3 Lorentz Force
Todayweencounteranothertricky potential,thisonefrom8.02. Theforceona
chargedparticlemovinginE andB fieldsis(asyoumayrecall)

Lorentzforceonparticlewithchargee (notq!)
 
~ ~
F = e E + ~v × B ~

whereE~ istheelectricfieldandB ~ isthemagneticfield.

LorentzForce

Manyeverydayobjectsuseelectricmotorsand/orgenerators,allofwhichde-
pendontheLorentzforce.
I’mnotgoingtouseq heretoavoidconfusion. e issomechargeand~r isthe
CartesianCoordinateofthatcharge.

7
Let’s say I have a small test charge constrained to move along a wire in an
externalB-field.TheB-fieldcanbealongẑ,andthewirecanbeahelicalcoil,like
aspring.

x = R cos φ ẋ = −R sin φφ̇


y = R sin φ ẏ = R cos φφ̇
z = αφ ż = αφ̇

Howdoesthechargemove? Ideas? Let’sfindout... weneedapotentialforthe


Lorentzforce

   
˙ ~ ˙
UL ~r, ~r = e Φ − A · ~r

~ = −rΦ (~r, t) − ∂ A
E ~ (~r, t)
∂t
~
B = r×A~

8
Φ ≡ electricscalarpotential
~ ≡ magneticvectorpotential
A
1
NB:forconstant,uniformB-field,A =~ ~r × B ~
2

So,forourtestchargewehave

~ = Bẑ ⇒ A
B ~ = B {−y, x, 0}
2
~ · ~r˙ = mgz + eB (yẋ − xẏ)
U = mgz − A
2
2 � 
eBR
= mgαφ + − sin2 φ − cos2 φ φ˙
2

�  � 
T = 12 m ẋ2 + ẏ 2 + ż 2 = 12 m R2 + α2 φ˙ 2
�  eBR2
⇒ L = 21 m R2 + α2 φ̇2 − mgαφ + φ̇
2
| {z }
drop!

ThedynamicsareunchangedbytheB-field!Why?
 
~ · ~v = 0 ⇒ noworkdone
F · v = ~v × B

Thiswillbetrueforany 1-Dmotion,solet’stry2-D...
Howaboutachargefreetomoveinthey-zplanewithaB-fieldinthex̂ direc-
tion?

B~ = Bx̂ ⇒ A ~ = B {0, −z, y}


1
� 2 2
 eB2
⇒ L = 2 m ẏ + ż + 2 (yż − zẏ) − mgz

9
∂L eB ∂L eB
Fy = ∂y = 2 z,
˙ py = ∂ẏ = my˙ − 2 z
eB eB
ṗy = Fy ⇒ m¨
y− 2 ż = 2 ż

Fz = −eB
2 ẏ − mg, pz = mż + eB
2 y
⇒ ÿ = βz,
˙ z̈ = −g − βẏ
withβ = eB
m

Theseequationsofmotionaresimpleenoughtosolve,andthesolutionisin-
teresting...let’sseewhathappensforaparticlethatstartsatrest.
Ifyouignoreg, you mightguessthatsincetheLorentzforce is⊥ to~v , the
trajectorymustbeacircle.

y(t) = a sin ωt ⇒ ẏ = −ωz, ÿ = −ω 2 y


z(t) = −a cos ωt ⇒ ż = ωy, z̈ = −ω 2 z

Anotherwaytoseethatthetrajectorymustbeacircleistonoticethat

...
z = −βÿ = −β 2 ż = −ω 2 ż
withβ = ω

whichisthetimederivativeofequationofmotionforaharmonicoscillatorwith
frequencyβ.
Comparingwithourequationsofmotionsuggeststhatω = β, butwehavethis
pesky gravity...noproblem,add− βg t toy(t).Thisdoesn’t changeÿ.

10
g
y(t) = a sin βt − t
β
g
ẏ = aβ cos βt −
β
ż = aβ sin βt

Tostartatrest,weneed

g
ẏ (t = 0) = 0 ⇒ a = −
β2
g
y(t) = − (sin βt + βt)
β2
g
z(t) = cos βt
β2

Sothisparticledoesn’tfall, itmovessideways withaveragevelocity (for βg


generalinitialconditions,yougetsin andcos componentsforbothy andz).
We canquicklyrelatethisresulttoparticleaccelerators,thoughournon-relativistic
physicsisclearlyinadequatetogetagoodanswer...
Foraparticlemovinginaplane ⊥ togravity,andwithourB-fieldpointingup,
wecanreusethepreviousresult

11
forB~ = Bẑ ẍ = βẏ , ÿ = βẋ
x(t) = a sin βt ⇒ ẋ = aβ cos βt
y(t) = −a cos βt ⇒ ẏ = aβ sin βt

Soifwestartaproton withaninitialvelocityc inthex̂ direction...

c mp c
forv~ 0 = cx̂ ⇒ a ==
β eB
forB ' 8 Tesla ⇒ a = 0.4 m

withc = 3 × 10 8 m/s,m p = 1.7 × 10−27 kgande = 1.6 × 10 −19 C.


ButCERNhasaradiusof4.5km!
Ifwereplacem p withtherelativisticmassm r = Ep /c2 wegettherightanswer:

E E
mr = c2p ∼ 104 mp ⇒ ar = eBcp ∼ 4.2 km
forE p ≈ 10 Tev ≈ 1.6 × 10−6 J

12
4 Gauge Invariance
WehavesomefreedominchoosingthemagneticvectorpotentialA ~

Gauge Transformation
0 ∂ ~0 = A~ + rf
Φ = Φ− f A
∂t  
~ 0 ~ 0 ~
B = r × A = r × A + rf = B ~ + r × (rf )
| {z }
thisis 0

~ 0 = −rΦ0 − ∂ A
E ~0 =
 ∂t 
∂ ∂ ~ 
~
= −r Φ − f − A + rf = E
∂t ∂t

WhataboutourLagrangian?

GaugeInvariantEquationofMotion
 0   
0
~ · ~r = L + e
0 ∂ ∂
L =Φ−e Φ −A f + ~ṙ · f
∂t ∂~r
d
= L+e f
dt

ForInteresting physicsassociatedwithmagnetic vectorpotential,googleAharnov-


Bohmeffect.

13

You might also like