Theoretical Question 1 / Solutions Page 1/12

Theoretical Question 1: Ping-Pong Resistor

1. Answers

1 V2
(a) FR = − πR 2ε 0 2
2 d

πr 2
(b) χ = −ε 0
d

2mgd
(c) Vth =
χ

(d) v s = αV 2 + β

 η 2  2 χ   η2 
α =  
2  , β =  (2 gd )
2 
 1 − η  m  1+η 

1+η χ3
(e) γ=
1 −η 2md 2

1 −η 2 mgd 2η 1 − η 2
(f) Vc = , Ic = g mχ
1+η 2 χ (1 + η )(1 + η 2 )

I
I ~ γV 2

Ic

0 zc V/ Vth
1
 Theoretical Question 1 / Solutions Page 2/12

2. Solutions
(a) [1.2 points]
The charge Q induced by the external bias voltage V can be obtained by applying
the Gauss law:
r r
ε 0 ∫ E ⋅ ds = Q (a1)

V 
Q = ε 0 E ⋅ (πR 2 ) = ε 0   ⋅ (πR 2 ) , (a2)
d 
where V = Ed .
The energy stored in the capacitor:

V ′ 
V V 2
1 V
U = ∫ Q(V ′)dV ′ = ∫ ε 0πR 2  dV ′ = ε 0 πR 2 . (a3)
0 0 d  2 d

The force acting on the plate, when the bias voltage V is kept constant:
∂U 1 V2
∴ FR = + = − ε 0πR 2 2 . (a4)
∂d 2 d

[An alternative solution:]

Since the electric field E ' acting on one plate should be generated by the other plate
and its magnitude is
1 V
E' = E= , (a5)
2 2d
the force acting on the plate can be obtained by
FR = QE ' . (a6)

(b) [0.8 points]

The charge q on the small disk can also be calculated by applying the Gauss law:
r r
ε 0 ∫ E ⋅ ds = q . (b1)

Since one side of the small disk is in contact with the plate,
πr 2
q = −ε 0 E ⋅ (πr 2 ) = −ε 0 V = χV . (b2)
d
 Theoretical Question 1 / Solutions Page 3/12

 πr 2 
Alternatively, one may use the area ratio for q = − 2 Q .
 πR 

πr 2
∴ χ = −ε 0 . (b3)
d

(c) [0.5 points]

The net force, Fnet , acting on the small disk should be a sum of the gravitational and
electrostatic forces:

Fnet = Fg + Fe . (c1)

The gravitational force: Fg = − mg .

The electrostatic force can be derived from the result of (a) above:
1 πr 2 χ 2
Fe = ε 0 2 V 2 = V . (c2)
2 d 2d

In order for the disk to be lifted, one requires Fnet > 0 :

χ
V 2 − mg > 0 . (c3)
2d

2mgd
∴Vth = . (c4)
χ

(d) [2.3 points]

Let v s be the steady velocity of the small disk just after its collision with the bottom
plate. Then the steady-state kinetic energy K s of the disk just above the bottom plate
is given by
1
K s = mv s2 . (d1)
2
For each round trip, the disk gains electrostatic energy by
∆U = 2qV . (d2)
For each inelastic collision, the disk lose its kinetic energy by
 1 
∆K loss = K before − K after = (1 − η 2 ) K before =  2 − 1 K after . (d3)
η 
Since K s is the energy after the collision at the bottom plate and ( K s + qV − mgd ) is
 Theoretical Question 1 / Solutions Page 4/12

the energy before the collision at the top plate, the total energy loss during the round trip
can be written in terms of K s :
 1 
∆K tot =  2 − 1 K s + (1 − η 2 )( K s + qV − mgd ) . (d4)
η 
In its steady state, ∆U should be compensated by ∆K tot .
 1 
2qV =  2 − 1 K s + (1 − η 2 )( K s + qV − mgd ) . (d5)
η 
Rearranging Eq. (d5), we have
η2
Ks =
1 −η 4
[(1 + η 2 )qV + (1 − η 2 )mgd ]

 η2   η2 
=  
2 
qV +  mgd
2 
(d6)
 1 −η   1+η 
1
= mv s2 .
2
Therefore,

 η 2  2 χV 2   η 2 
v s =  
2 
 +  (2 gd ) .
2 
(d7)
 1 − η  m   1 + η 

Comparing with the form:

v s = αV 2 + β , (d8)

 η 2  2 χ   η2 
α =  
2  , β =  (2 gd ) .
2 
(d9)
 1 − η  m  1+η 

[An alternative solution:]

Let v n be the velocity of the small disk just after n-th collision with the bottom plate.
Then the kinetic energy of the disk just above the bottom plate is given by
1
K n = mv 2n . (d10)
2
When it reaches the top plate, the disk gains energy by the increase of potential energy:

∆U up = qV − mgd . (d11)

Thus, the kinetic energy just before its collision with the top plate becomes
1
K n − up = mv 2up = K n + ∆U up . (d12)
2
 Theoretical Question 1 / Solutions Page 5/12

Since η = v after / v before , the kinetic energy after the collision with the top plate becomes
scaled down by a factor of η 2 :

K n′ −up = η 2 ⋅ K n −up . (d13)

Now the potential energy gain by the downward motion is:

∆U down = qV + mgd (d14)
so that the kinetic energy just before it collides with the bottom plate becomes:

K n −down = K n′ − up + ∆U down . (d15)

Again, due to the loss of energy by the collision with the bottom plate, the kinetic
energy after its (n + 1) -th collision can be obtained by
K n+1 = η 2 ⋅ K n−down
= η 2 ( K n′ − up + ∆U down )
= η 2 (η 2 ( K n + ∆U up ) + ∆U down ) (d16)
= η 2 (η 2 ( K n + qV − mgd ) + qV + mgd )
= η 4 K n + η 2 (1 + η 2 )qV + η 2 (1 − η 2 )mgd .
1
As n → ∞ , we expect the velocity v n → v s , that is, K n → K s = mv s2 :
2

Ks =
1
1 −η 4
[
η 2 (1 + η 2 )qV + η 2 (1 − η 2 )mgd ]
 η2   η2 
=  
2 
qV +  mgd
2 
(d17)
1 −η  1 +η 
1
= mv s2
2

(e) [2.2 points]

The amount of charge carried by the disk during its round trip between the plates is
∆Q = 2q , and the time interval ∆t = t + + t − , where t + ( t − ) is the time spent during the
up- (down-) ward motion respectively.
Here t + ( t − ) can be determined by
1
v 0+ t + + a + t +2 = d
2 (e1)
1
v 0− t − + a − t −2 = d
2
where v 0+ ( v 0− ) is the initial velocity at the bottom (top) plate and a + ( a − ) is the up-
 Theoretical Question 1 / Solutions Page 6/12

(down-) ward acceleration respectively.

Since the force acting on the disk is given by
qV
F = ma ± = qE m mg = m mg , (e2)
d
in the limit of mgd << qV , a ± can be approximated by
qV
a0 = a + = a − ≈ , (e3)
md
which implies that the upward and down-ward motion should be symmetric. Thus,
Eq.(e1) can be described by a single equation with t 0 = t + = t − , v s = v 0+ = v 0- , and
a0 = a+ = a− . Moreover, since the speed of the disk just after the collision should be
the same for the top- and bottom-plates, one can deduce the relation:
v s = η (v s + a 0 t 0 ) , (e4)
from which we obtain the time interval ∆t = 2t 0 ,
 1 −η  vs
∆t = 2t 0 = 2  . (e5)
 η  a0
From Eq. (d6), in the limit of mgd << qV , we have

1  η2 
K s = mv s2 ≈  qV .
2 
(e6)
2  1 −η 

By substituting the results of Eqs. (e3) and (e6), we get

 1 − η  2η 2 md 2 1 −η 2md 2
∆t = 2  =2 . (e7)
 η  1 −η 1 +η χV 2
2
qV

∆Q 2q
Therefore, from I = = ,
∆t ∆t

2q 1 +η χV 2 1 +η χ3
I= = χV = V 2. (e8)
∆t 1 −η 2md 2
1 −η 2md 2

1+η χ3
∴γ = (e9)
1 −η 2md 2

[Alternative solution #1:]

Starting from Eq. (e3), we can solve the quadratic equation of Eq. (e1) so that
 Theoretical Question 1 / Solutions Page 7/12

v 0±  2da 
t± =  1 + 2 0 − 1 . (e10)
a0  v 0± 
 
When it reaches the steady state, the initial velocities v 0± are given by
v 0+ = v s (e11)

2da0
v 0 − = η ⋅ ( v s + a0 t + ) = ηv s 1 + , (e12)
v s2

where vs can be rewritten by using the result of Eq. (e6),

 η 2  2qV  η 2 
v s2 ≈ αV =  
2 
=  2a0 d .
2 
(e13)
 1 −η  m  1 −η 

1 vs 1 
As a result, we get v 0− ≅ ηv s ⋅ = v s and consequently t ± =  − 1 , which is
η a0 η 
equivalent to Eq. (e4).

[Alternative solution #2:]

The current I can be obtained from
2 q 2qv
I= = , (e14)
∆t d
where v is an average velocity. Since the up and down motions are symmetric with the
same constant acceleration in the limit of mgd << qV ,
1 v 
v =  v s + s  . (e15)
2 η 
Thus, we have
q  1
I=
1 +  v s . (e16)
2d  η 
Inserting the expression (Eq. (e15)) of v s into Eq. (e16), one obtains an expression
identical to Eq. (e8).

(f) [3 points]
The disk will lose its kinetic energy and eventually cease to move when the disk can not
reach the top plate. In other words, the threshold voltage Vc can be determined from
the condition that the velocity v 0- of the disk at the top plate is zero, i.e., v 0- = 0 .

In order for the disk to have v 0- = 0 at the top plate, the kinetic energy K s at the
 Theoretical Question 1 / Solutions Page 8/12

top plate should satisfy the relation:

K s = K s + qVc − mgd = 0 , (f1)

where K s is the steady-state kinetic energy at the bottom plate after the collision.
Therefore, we have

 η2   η2 
 
2 
qV +  mgd + qVc − mgd = 0 ,
2 
(f2)
 1 −η   1+η 
c

or equivalently,

(1 + η 2 )qVc − (1 − η 2 )mgd = 0 . (f3)

1 −η 2
∴ qVc = mgd (f4)
1+η 2
From the relation q = χVc ,

1 −η 2 mgd
∴Vc = . (f5)
1+η 2 χ

In comparison with the threshold voltage Vth of Eq. (c4), we can rewrite Eq. (f5) by
Vc = z cVth (f6)
where z c should be used in the plot of I vs. (V / Vth ) and

1 −η 2
zc = . (f7)
2(1 + η 2 )

[Note that an alternative derivation of Eq. (f1) is possible if one applies the energy
compensation condition of Eq. (d5) or the recursion relation of Eq. (d17) at the top
plate instead of the bottom plate.]
Now we can setup equations to determine the time interval ∆t = t − + t + :
1
v 0− t − + a − t −2 = d (f8)
2
1
v 0+ t + + a + t +2 = d (f9)
2
where the accelerations are given by

qVc 1 − η 2   − 2η 2 
a+ = −g = − 1 g =  g
2 
(f10)
1 + η 1+η 
2
md 
 Theoretical Question 1 / Solutions Page 9/12

qVc 1 − η 2   2 
a− = +g= + 1 g =   g (f11)
1 + η 1+η
2 2
md  

a+
= −η 2 (f12)
a−

Since v 0− = 0 , we have v 0+ = η (a− t − ) and t −2 = 2d / a− .

2d d 
t− = = (1 + η 2 )  , (f13)
a− g

By using v 02+ = η 2 (2da − ) = −2da + , we can solve the quadratic equation of Eq. (f9):

v 0+  2da  v 2d  1 + η 2  d  t
t+ =  1 + 2 + − 1 = − 0+ = =  2   = − . (f14)
a+  v 0+  a+ a+  η  g  η

 1 d 
∴ ∆t = t − + t + = 1 +  (1 + η 2 )  (f15)
 η g

∆Qc 2q 2χVc 2η 1−η 2

Ic = = = = g mχ . (f16)
∆t ∆t ∆t (1+η)(1+η 2 )

I
I ~ γV 2

Ic

0 zc V/ Vth
1
 Theoretical Question 1 / Solutions Page 10/12

[A more elaborate Solution:]

One may find a general solution for an arbitrary value of V . By solving the quadratic
equations of Eqs. (f8) and (f9), we have

v 0±  2da ± 
t± = − 1 + 1 + 2  . (f17)
a ±  v 0± 

(It is noted that one has to keep the smaller positive root.)

To simplify the notation, we introduce a few variables:

V 2mgd
(i) y = where Vth = ,
Vth χ

1 −η 2
(ii) z c = , which is defined in Eq. (f7),
2(1 + η 2 )

gd d
(iii) w0 = 2η and w1 = 2 ,
1 −η 2
1 −η 2 g ( )
In terms of y , w , and z c ,
qV
a+ = − g = g (2 y 2 − 1) (f18)
md
qV
a− = + g = g (2 y 2 + 1) (f19)
md

v 0+ = v s = w0 y 2 + z c2 (f20)

v 0− = η ( v s + a + t + ) = w0 y 2 − z c2 (f21)

y 2 − z c2 − η y 2 + z c2
t + = w1 (f22)
2 y2 −1

y 2 + z c2 − η y 2 − z c2
t − = w1 (f21)
2y2 +1
 Theoretical Question 1 / Solutions Page 11/12

∆Q 2q y 8mgdχ
I= = = (2 χVth ) = F ( y) (f22)
∆t t + + t − ∆t w1

where
−1
 y 2 − z 2 − η y 2 + z 2 y 2 + z c2 − η y 2 − z c2 
c c
F ( y) = y +  (f23)
 2 y 2
− 1 2y2 +1 
 Theoretical Question 1 / Solutions Page 12/12

3. Mark Distribution
Total Partial
No. Contents
Pt. Pt.
(a) 1.2 0.3 Gauss law, or a formula for the capacitance of a parallel plate
0.5 Total energy of a capacitor at E ' = electrical field by the other
plate
V
0.4 Force from the energy F = QE '
expression
(b) 0.8 0.3 Gauss law Use of area ratio and result of (a)
0.5 Correct answer
(c) 0.5 0.1 Correct lift-up condition with force balance

0.2 Use of area ratio and result of (a)

0.2 Correct answer

(d) 2.3 0.5 Energy conservation and the work done by the field
0.5 Loss of energy due to collisions
0.8 Condition for the steady state: Condition for the steady state:
recursion relation
energy balance equation (loss =
gain)
0.5 Correct answer
(e) 2.2 0.2 ∆Q = 2q per trip
0.5 Acceleration a ± in the limit of qV >> mgd ; a + = a − by
symmetry
0.3 Kinetic equations for d , v , By using the symmetry, derive
a , and t , solutions for t ±
0.4 Expression of v 0± and t ± in the relation (e4)
its steady state
0.4 Solutions of t± in
approximation
0.4 Correct answer
(f) 3.0 0.5 Condition for Vc ; K up = 0 or Using (d8), Recursion relations
v s ,up = 0
0.3 energy balance equation
0.3 Correct answer of Vc
0.7 Kinetic equations for ∆t
0.3 Correct answer of I c
0.9 Distinction between Vth and Vc ,
the asymptotic behavior I = γV 2 in plots
Total 10