Upload
sreetama-das
View
229
Download
1
Embed Size (px)
Citation preview
8/14/2019 6 Liquid Helium 4
1/66
Statistical and Low Temperature Physics (PHYS393)
6. Liquid Helium-4
Kai Hock
2010 - 2011
University of Liverpool
8/14/2019 6 Liquid Helium 4
2/66
Topics to cover
1. Fritz Londons explanation of superfluidity in liquid helium-4
using Bose Einstein condensate (BEC).
2. Theory and experimental search for BEC.
3. Laszlo Tiszas 2-fluid model theory. Experiments.
4. Lev Landaus critical velocity. Experiments.
5. Quantised vortices.
Liquid Helium-4 1
8/14/2019 6 Liquid Helium 4
3/66
Explaining superfluidity.
Fritz London suggested using Bose Einstein condensate (BEC)to explain superfluidity. This is possible because:
1. Helium-4 is a boson, so it may undergo Bose Einstein
condensation.
2. The liquid would then become a single wavefunction, so it
would show no viscosity unless it flows too quickly.
Liquid Helium-4 2
8/14/2019 6 Liquid Helium 4
4/66
In order to test his theory, London calculated the transition
temperature and heat capacity of the BEC (left figure). He
showed that there is some agreement with liquid helium-4
(right figure), at least in trend.
Left: Enss and Hunklinger, Low Temperature Physics, page 8, 2005.
Right: London, Physical Review, volume 54, page 947, 1938
Liquid Helium-4 3
8/14/2019 6 Liquid Helium 4
5/66
BEC heat capacity.
We shall look at a simplified version of Londons treatment to
calculate the BEC heat capacity.
This involves finding:
1. the temperature at which condensation starts,
2. the change in number of atoms with temperature, and
3. the heat capacity below condensation temperature.
Liquid Helium-4 4
8/14/2019 6 Liquid Helium 4
6/66
Using the Bose-Einstein distribution that we have learnt, we can
calculate the temperature at which bosons condense into BEC.
Recall the expression we first derived for the Bose-Einstein
distribution:nigi
= 1
exp(1 2i) 1
When applying this to phonons, 1 was left out because thenumber of phonons is not fixed.
Now that we are dealing with atoms, which are real bosons, the
number is fixed and we have to keep 1. This is usually written
in terms of as follows:
nigi
= 1
exp(( )/kBT) 1This is usually denoted f(). We call it occupancy, and think of
it as the average number of particle that occupies a state at
energy .
Liquid Helium-4 5
8/14/2019 6 Liquid Helium 4
7/66
As with the ideal gas, electrons, photons and phonons, we
write ni in terms of the number density:
ni=n()d,
and gi in terms of the density of states:
gi=g()d.
Then we have the relations
n()d=f()g()d
and
n()d= g()d
exp(( )/kBT) 1
Liquid Helium-4 6
8/14/2019 6 Liquid Helium 4
8/66
The density of states
We have seen a number of different density of states.
For electrons, we need to multiply by 2 because of the spin
states. For phonons, we multiply by 3 for the three possible
polarisations. For photons, we multiply by 2 for the two
polarisation states.
So what do we do for atoms that are bosons?
The answer is: we use the same density of state for the ideal
gas:
g() =4mVh3
(2m)1/2
This is the very first one that we have seen, before we have to
include the additional effects of spin and polarisation.
Liquid Helium-4 7
8/14/2019 6 Liquid Helium 4
9/66
The chemical potential
is called the chemical potential. It is determined by the totalnumber of particles. We get this by integrating:
N=
0
g()d
exp(( )/kBT) 1I have assumed that the ground state energy is zero, which is
why I integrated from zero.
This is not easy to solve. Instead, we shall make use of some
approximations at low temperature.
For a start, note that could depend on temperature, sincethe above equation contains T.
Liquid Helium-4 8
8/14/2019 6 Liquid Helium 4
10/66
The chemical potential
f() = 1
exp(( )/kBT) 1If is positive, f() is negative for some energies.
Then some states have a negative number of particles, which is
not possible.
Liquid Helium-4 9
8/14/2019 6 Liquid Helium 4
11/66
The chemical potential
So must be negative.
The graph to the right of 0 contributes to the total particle
number N.
Liquid Helium-4 10
8/14/2019 6 Liquid Helium 4
12/66
The chemical potential
If temperature falls, f() = 1exp(()/kBT)1
decreases.
Then particle number decreases, which is not possible.
Liquid Helium-4 11
8/14/2019 6 Liquid Helium 4
13/66
The chemical potential
The particle number can remain fixed if moves closer to 0:
Then f() increases and particle number can remain the same.
Liquid Helium-4 12
8/14/2019 6 Liquid Helium 4
14/66
The occupation number
We focus on the occupation number in the integral:
f() = 1
exp(( )/kBT) 1First, note that cannot be positive, or else for small energy
the exponential function would be less than one. Then the
denominator would be negative. This means negative
occupation number, which is not physical.
Next, we know that when temperature is low enough, a large
number of atoms would go into the ground state = 0. The
number would be as large as N, which is of the order of 1 mole
( 1024
).
This means that f(0) would be very large, so that must be
very close to zero. Then the exponential function would be
close to 1 and the denominator would be close to zero.
Liquid Helium-4 13
8/14/2019 6 Liquid Helium 4
15/66
Number of atoms in the excited state
So at very low temperature, it is safe to assume that = 0.
We can then write
Nex=
0
g()d
exp(/kBT) 1I have added a subscript ex to N. The reason will become clear
in a moment. First, integrate this with the help of the Table of
Integrals:
Nex=
2mkBT
h2
3/22.612V
Notice that T is in the denominator. This suggests that the
total number of atoms decreases with temperature - that they
are disappearing !
In fact, the above integral includes only atoms above the
ground state = 0. The missing atoms are going into the
ground state.
Liquid Helium-4 14
8/14/2019 6 Liquid Helium 4
16/66
The condensation temperature
The formula for Nex is only able to compute the number of
particles above the ground state. This is because the density of
states g() in the integral is zero when = 0.
When a substantial fraction of the particles start going into the
ground state, this has to be added separately. So Nex is the
number of atoms in the excited states. Hence the subscript ex.
We know that as soon as Nex becomes less than the original
number N, then (N Nex) atoms start going into the groundstate.
Therefore, condensation takes place when Nex=N.
Substituting this and solving for T, we get
TBE= h2
2mkB
N
2.612V
2/3.
Liquid Helium-4 15
8/14/2019 6 Liquid Helium 4
17/66
Transition temperature
Substituting the mass of helium atom and the molar volume
into the formula for the condensation temperature TBE, we find
3.13 K.
This is fairly close to the lambda point temperature of 2.18 K,
at which liquid helium-4 changes to a superfluid,
and provides some support for the idea that the superfluid is a
BEC.
Liquid Helium-4 16
8/14/2019 6 Liquid Helium 4
18/66
Number of atoms in the ground state
Above TBE, there are few atoms in the ground state.
Below TBE, this number increases until all atoms fall into theground state.
Glazer and Wark, Statistical mechanics, p. 101
Liquid Helium-4 17
8/14/2019 6 Liquid Helium 4
19/66
Heat capacity.
Next, to calculate the heat capacity, recall the expression for
the number of excited particles:
Nex=
0
g()d
exp(/kBT) 1.
To find the heat capacity, we first find the energy U. The above
integral is a sum over particle number in every energy interval.
To find U, we need to multiply by the energy at each interval
d:
U =
0
g()d
exp(/kBT) 1.
Integrating using the table of integrals gives
U = 0.7704kBNT5/2
T3/2BE
.
Liquid Helium-4 18
8/14/2019 6 Liquid Helium 4
20/66
Below condensation temperature.
The heat capacity below the condensation temperature is then
obtained by differentiating with respect to T:
C= 1.926kBN
T
TBE
3/2.
Note that the peak value is 1.926kBN. This is obtained by
setting T =TBE, when Nex=N (all atoms are just excited).
Liquid Helium-4 19
8/14/2019 6 Liquid Helium 4
21/66
Above condensation temperature.
Above condensation temperature TBE, we cannot use the sameformula for the heat capacity. This is because it is derived
assuming that the chemical potential 0, which is only truewhen condensation starts.
Above TBE, changes. Albert Einstein has published a formulafor this in 1924 (right half of curve):
At high temperatures, the heat capacity reaches the ideal gasvalue of 3NkB/2.
Liquid Helium-4 20
8/14/2019 6 Liquid Helium 4
22/66
The search for condensate.
After Londons work, the search for BEC in liquid helium wason. Calculations and experiments were carried to determine if
BEC really existed in the superfluid.
Measuring the amount of BEC is difficult. The method is to
use neutron scattering to determine the velocities of the heliumatoms.
In a BEC, the helium atoms are in the ground state - so their
velocities are all zero. If the measurement can show that a
significant fraction of the superfluid contains atoms with zerovelocity, then we have found the condensate.
Liquid Helium-4 21
8/14/2019 6 Liquid Helium 4
23/66
Neutron scattering.
A quick look at how neutrons can be used to determine the
velocities of atoms would be helpful.
We cannot see the movement of the atoms in the superfluid,
but we can measure the energy and direction of neutrons
scattered from the superfluid.
Using momentum conservation, we can deduce the velocities of
the helium atoms that the neutrons must have collided with.
Liquid Helium-4 22
8/14/2019 6 Liquid Helium 4
24/66
Helium-4 condensate
The figure shows the theory and experimental results in the
1970s and 1980s. The condensate has indeed be found.
Tilley and Tilley, Superfluidity and Superconductivity, (2003) p. 64
However, even at 0 K, the condensate is only 10% of the
superfluid. The reason is attributed to the interaction between
atoms - it can only be 100% if the helium is an ideal gas at 0 K.
Liquid Helium-4 23
8/14/2019 6 Liquid Helium 4
25/66
Helium-4 superfluid
Recall that London has proposed BEC as an explanation for
superfluidity, because it behaves as a single wavefunction.
If BEC is only 10% of the superfluid, that means only 10% of
the superfluid is a single wavefunction.
If there is only 10% BEC, why does all of the liquid helium-4
behave as a superfluid at 0 K?
This problem is still not fully understood.
Liquid Helium-4 24
8/14/2019 6 Liquid Helium 4
26/66
Two-fluid model.
Even without a full understanding, it is still possible to make
models that are useful for predicting the behaviour. This would
in turn provide some understanding.
One month after London proposed the BEC theory for liquid
helium in 1938, Laszlo Tisza proposed a two-fluid model for
the superfluid.
The idea is that below the 2.18 K transition temperature,
normal and superfluid coexist. As temperature falls, the normal
fluid decreases. At 0 K, the superfluid increases to 100%.
Liquid Helium-4 25
8/14/2019 6 Liquid Helium 4
27/66
Connection with BEC.
The two-fluid model is closely related to the BEC idea. Recall
the left figure, showing the condensate going to 100% as
temperature falls from TBE to 0 K.
Right: http://www.yutopian.com/Yuan/TFM.html
If we include the graph for the excited (not condensate)
fraction, it would look like the figure on the right. This is just
what the two-fluid model would suggest.
Liquid Helium-4 26
8/14/2019 6 Liquid Helium 4
28/66
Two-fluid model.
According to this model, helium II behaves as if it were amixture of two completely interpenetrating fluids with different
properties, although in reality this is not the case.
To avoid any misunderstanding, it must be clearly stated at the
outset that the two fluids cannot be physically separated; it is
not permissible even to regard some atoms as belonging to the
normal fluid and the remainder to the superfluid component,
since all 4He atoms are identical.
Enss and Hunklinger, Low-Temperature Physics (2005) p. 24
Liquid Helium-4 27
8/14/2019 6 Liquid Helium 4
29/66
Two-fluid model.
In this model, the fluids are supposed to have the following
properties:
Guenault, Basic Superfluids (2003) p. 29
Note that we are not supposed to think that we can physically
separate normal fluid from superfluid, since the helium-4 atomsare not distinguishable.
This makes it look rather unrealistic. However, following these
assumptions, it has been possible both to explain many
experiments.
Liquid Helium-4 28
8/14/2019 6 Liquid Helium 4
30/66
Andronikashvilis experiment.
In 1948, Elepter Andronikashvili measured the fraction of
normal fluid in liquid helium using a stack of rotating discs
(left). The small spacings between discs means that any normalfluid would be dragged along and slow down the oscillation.
Right: Lifshitz and Andronikashvili, A Supplement to Helium (1959)
In the experiment, the oscillation period changes a lot below
2.2 K. By attributing this to a decrease in normal fluid stuck in
the spacing, the fraction of normal fluid can be deduced (right).
Liquid Helium-4 29
8/14/2019 6 Liquid Helium 4
31/66
Viscosity measurement.
Viscosity can be measured by allowing liquid helium to flow
through a very narrow capillary tube. From the velocity of flow,
the viscosity can be calculated.
J.F. Allen, A.D. Misener, Proc. R. Soc. A172, 467 (1939)K.R. Atkins, Philos. Mag. Supp. 1, 169 (1952)
Notice that viscosity drops to zero below 2.2 K.
Liquid Helium-4 30
8/14/2019 6 Liquid Helium 4
32/66
Boiling helium.
When pressure is reduced below vapour pressure, superfluidhelium boils (even if temperature is below 2.2 K).
When a liquid like water boils, we expect to see bubbles
forming the liquid moving.
When superfluid helium boils, we cannot see anything
happening!
The helium atoms leave the surface of the liquid. Somehow,
there is no bubble and no movement at all.
Liquid Helium-4 31
8/14/2019 6 Liquid Helium 4
33/66
Fountain effect.
In the setup below (left), liquid helium in a flask and in a bath
is separated by compressed powder which is porous.
Left: Enss and Hunklinger, Low-Temperature Physics (2005) p. 21
Right: J.F. Allen, photographed 1971, unpublished
The temperature is about 1 K. When the liquid in the flask is
warmed very slightly with a heater, it shoots out of the flask
like a fountain (right).
Liquid Helium-4 32
8/14/2019 6 Liquid Helium 4
34/66
Viscosity experiments explained.
We have seen 2 viscosity experiments.
In the rotating discs experiment by Andronikashvili, viscosity
decreases gradually as temperature falls below 2.2 K. This
shows that the superfluid fraction increases gradually.
In the capillary tube experiment, the viscosity drops suddenly tozero. This suggests that the liquid suddenly becomes 100%
superfluid.
The reason is that the capillary tube is too narrow. The normal
fluid would experience too much viscosity trying to go through.So only the superfluid fraction passes through, and this
experiences zero viscosity.
Liquid Helium-4 33
8/14/2019 6 Liquid Helium 4
35/66
Fountain effect explained.
In the fountain experiment, when the liquid helium in the flask
is warmed, the liquid shoots out like a fountain.
Using the two-fluid model, when it is warmed, the superfluid
fraction changes to normal fluid. There is now a difference in
concentration of superfluid between the flask and the helium
bath below.
Since they are separated by compacted powder, superfluid can
pass. Normal fluid cannot because of viscosity.
Since the concentration of superfluid in the flask decreased,superfluid will diffuse through the pores in the compacted
powder from the bath. The volume of liquid in the fask then
increases quickly and shoots out like a fountain.
Liquid Helium-4 34
8/14/2019 6 Liquid Helium 4
36/66
Silent boiling explained.
When water boils, there is plenty of bubbles and movements.
This happens because certain spots in the water gets a lot
hotter than other parts. Boiling then starts a these hotspots,
where bubbles are formed.
If bubbles are not formed in boiling helium, it must be because
there are no hotspots. This can only be possible if the liquid
has extremely good thermal conductivity. Any heat around
potential hotspots must be conducted away immediately.
Liquid Helium-4 35
8/14/2019 6 Liquid Helium 4
37/66
Silent boiling explained.
If any spot is hotter than the surrounding, it could convert the
nearby superfluid into normal fluid. The concentration of
superfluid decreases. Superfluid from further away flows in to
the hotspot, and the normal fluid flows away.
This process is very efficient because the superfluid has noviscosity. The flow is limited only by the viscosity of the normal
fraction.
As a result, heat conduction is extremely fast, and no hotspot
can form. This is why no bubble is produced and the superfluidhelium boils silently.
Liquid Helium-4 36
8/14/2019 6 Liquid Helium 4
38/66
Landau critical velocity.
Landaus dispersion curve is used to deduce a critical velocity
below which liquid helium is superfluid.
Henshaw and Woods, Physical Review, volume 121, p. 1266 (1961)
The above graph is measured using neutron scattering. Let us
now see if the critical velocity calculated from this is indeed
real.
Liquid Helium-4 37
8/14/2019 6 Liquid Helium 4
39/66
Experiment.
The idea is to accelerate a body through superfluid He-4, and
see if it experiences a drag when the critical velocity is reached.
Allum, et al, Philosophical Transactions of the Royal Society A, vol. 284
(1977), p. 179
IN one experiment, a helium ion is injected into the superfluid.
This ion would attract surrounding atoms into a ball with
about 100 atoms. It is then accelerated using an electric field.
Liquid Helium-4 38
8/14/2019 6 Liquid Helium 4
40/66
Critical velocity.
The drag measured is plotted against velocity of the ion.
Allum, et al, Philosophical Transactions of the Royal Society A, vol. 284 (1977), p. 179
The result shows that the critical velocity is about 45 m/s.
Liquid Helium-4 39
8/14/2019 6 Liquid Helium 4
41/66
From the measured dispersion,
the following quantities are obtained:
excitation:
kB= 8.65 K,
momentum: p0
= 19.1 nm1.
Liquid Helium-4 40
8/14/2019 6 Liquid Helium 4
42/66
The Landau critical velocity is then given approximately by
vL=p0
= 58 m s1.
This is fairly close to the velocity of 45 m/s measured using
the helium ion, and provides support for Landaus theory.
Liquid Helium-4 41
8/14/2019 6 Liquid Helium 4
43/66
Rotational motion.
We now look at what the idea of a macroscopic wavefunction
means for rotational motion in a superfluid.
Consider a simple case of uniform motion in one direction. The
wavefunction may be represented by a plane wave:
=eikx.
To see that the velocity is indeed constant, we use the
equation in quantum mechanics for finding momentum:
iddx
=px.
Substituting the wavefunction and differentiating, we find
px= k.
This is the familiar relation for momentum in quantummechanics. Dividing by mass of the helium atom m gives thevelocity:
v = k
m.
Liquid Helium-4 42
8/14/2019 6 Liquid Helium 4
44/66
Phase of a wavefunction.
In this simple wavefunction,
=eikx
,kx is the phase. In other wavefunctions, the phase can depend
in a more complex way on the position x. Let the phase be
(x).
The phase is useful to help explain the behaviour of the
superfluid.
First, note that the simple wavefunction, as we move along x,
the phase changes. A change of 2 corresponds to one
wavelength.
Next, let us redo the calculation on the previous slide in terms
in terms of the phase.
Liquid Helium-4 43
8/14/2019 6 Liquid Helium 4
45/66
First, replace the phase in eikx by the more general function
(x): =ei(x).
Then, applying the equation in quantum mechanics for
momentum:
id
dx =p.
we find
p= d
dx.
Dividing by mass of the helium atom m gives the velocity:
v =
m
d
dx.
Next, consider what happens to the phase in a circular motion.
Liquid Helium-4 44
8/14/2019 6 Liquid Helium 4
46/66
Rotational motion.
Suppose the superfluid is flowing round in a circle. Imagine
following a path along this circle. When we return to the same
point, the wavefuncton
=ei(x)
must return to the same value. Then the phase must either
change by 0, or by a multiple of 2.
Let us now apply this idea to the velocity equation:
v =
m
d
dx.
When we integrate along the circular path, we get v.dl=
m
where is the change in phase, and dl is like a very short
vector along the path.
Liquid Helium-4 45
8/14/2019 6 Liquid Helium 4
47/66
In this equation,
v.d
l=
m.the dot product is used on the left hand side. This is because
we can think of the wavefunction like waves on the sea.
Only moving along direction v of the wave causes phase
change. Moving along the phase front perpendicular to v gives
no change in phase. The dot product picks out the component
along v.
We have seen that must be 0 or a multiple of 2, so the
equation becomes v.dl=
m2n =n
h
m,
where n is an integer.
Liquid Helium-4 46
8/14/2019 6 Liquid Helium 4
48/66
Rotation is quantised.
In this equation
v.dl=nh
mthe left hand side is called the circulation. Since n is an integer,this means that the circulation is quantised.
Note the circulation is a measure of the amount of rotation. Tosee this, start with the simplest case when there is no rotation.
Then the velocity v is uniform, with the same magnitude anddirection in the whole fluid. Integrating along a circle, we findthat half the time, we move in the direction of flow, the otherhalf we move opposite to the flow. Since there is a dot productin the circulation integral, that means the integral would be
zero.
Next, consider uniform rotation of the superfluid. When weintegrate along a circle, v and dl are always in the samedirection. Then the integral would be nonzero. Also, if v islarger, then the integral is larger.
Liquid Helium-4 47
8/14/2019 6 Liquid Helium 4
49/66
Quantised vortex.
When water rotates in a bucket, is looks like this:
We call this a vortex. It is difficult to imagine a vortex like this
can only rotate at certain speeds. Yet, this is exactly what
London predicted with the wavefunction, if we have a bucket of
superfluid helium.
Subsequently experiments have shown that this is indeed true.
Liquid Helium-4 48
M i i
8/14/2019 6 Liquid Helium 4
50/66
Measuring vortices.
One way to measure circulation is to look at the vibration of a
wire in a superfluid. When the circulation changes, the
vibration changes. Using this method, the circulation is plottedagainst angular velocity:
Karn, et al, Physical Review B, vol. 21 (1980), p. 1797.
The distinct steps show that the circulation can only take
discrete values, i.e. it is quantised.
Liquid Helium-4 49
Wh t h l t f
8/14/2019 6 Liquid Helium 4
51/66
What we have learnt so far.
1. Below 2.2 K, liquid helium-4 loses all viscosity. This new
phase is called a superfluid.
2. London used Bose-Einstein condensation to explain this.
He calculated a transition temperature of 3.13 K.
3. Theory and experimental search show that a condensate
really does exist in superfluid helium-4, but only at 10%.
4. The two-fluid model can be used to explain observations
like silent boiling, fountain effect, and zero viscosity.
5. Landaus prediction of a critical velocity is directly verified
by experiments of moving ions.
6. Rotational motion in superfluid helium is quantised.
Liquid Helium-4 50
E i
8/14/2019 6 Liquid Helium 4
52/66
Exercises
Exercise 1
Bose-Einstein condensation is achieved with a gas of rubidium
atoms, which have an atomic weight of 85.47. The
condensation temperature is 1.7 107 K. What was thenumber density of the condensate?
[ Atomic mass unit u is 1.6605 1027 kg.]
Liquid Helium-4 51
E i
8/14/2019 6 Liquid Helium 4
53/66
Exercises
Solution
Recall the formula for the condensation temperature:
TBE= h2
2mkB
N
2.612V
3/2
We can rearrange this to get the number density:
NV
= 2.612
2mkBTBEh2
3/2
The condensation temperature is given as
TBE= 1.7 107 K.
The mass is given by
m= 85.47u= 85.47 (1.6605 1027) kg.
Substituting these into the above equation for the number
density, we get 2.74 1019 m3.
Liquid Helium-4 52
Exercises
8/14/2019 6 Liquid Helium 4
54/66
Exercises
Exercise 2
Show that below the condensation temperature, TBE, the heat
capacity of a gas obeying Bose-Einstein statistics is given by
CV
= 1.93NkB TTBE
3/2
.
[You are given that
0
x3/2
ex 1dx= 1.78.
]
Liquid Helium-4 53
Exercises
8/14/2019 6 Liquid Helium 4
55/66
Exercises
Solution
Recall that at very low temperatures, the chemical potential is approximately zero, and the number of particles above the
ground state is given by:
Nex=
0
g()d
exp(/kBT) 1.
To obtain the energy, we insert into the integral:
U =
0
g()d
exp(/kBT) 1.
Substituting the density of states
g() =4mV
h3 (2m)1/2
gives
U =4mV
h3
0
2m3/2d
exp(/kBT) 1.
Liquid Helium-4 54
Exercises
8/14/2019 6 Liquid Helium 4
56/66
Exercises
Define
x=
kBT
and substitute into the integral for energy:
U =4mV
h3
2m(kBT)
5/2
0
x3/2dx
ex 1.
We can now use the given result:0
x3/2
ex 1dx= 1.78.
This gives:
U =
4mV
h3
2m(kBT)
5/2
1.78.The heat capacity can then be obtained by differentating this
with respect to temperature. First, we express this in terms of
the condensation temperature.
Liquid Helium-4 55
Exercises
8/14/2019 6 Liquid Helium 4
57/66
Exercises
The condensation temperature is given by:
TBE= h2
2mkB
N
2.612V
3/2
Make V the subject
V = 1
2.61
h2
2mkBTBE
3/2
and substitute into the above result for U:
U = 2
()
1.78
2.61NkB
T5/2
T5/2BE
Then differentiating with respect to T:
CV = 1.93NkB
T
TBE
3/2
we get the heat capacity for the Bose-Einstein condensate.
Liquid Helium-4 56
Exercises
8/14/2019 6 Liquid Helium 4
58/66
Exercises
Exercise 3
(a) A helium-4 atom is circulating around a vortex line. Show
that its angular momentum is a multiple of .
(b) When a bucket is rotated slowly, the water inside can
rotate with it as a whole, like a solid. If the water is replacedby superfluid helium-4, the rotation breaks up into a number of
quantised vortices. Estimate how many quanta of circulation a
scientist in Liverpool (lattitude 53.25N) would find in a bucketof the superfluid, diameter 10 cm, as a result of the rotation of
the Earth.
(Modified from Tony Guenault, Basic Superfluids (2003))
Liquid Helium-4 57
8/14/2019 6 Liquid Helium 4
59/66
Solution
(a) The circulation is given by v.dl= n
h
m.
Using a circular path of radius r, we find
v.2r =nh
m.
Then the angular momentum is
mvr =n
2=n.
So it is quantised in units of
(b) Assume that atoms at the wall of the bucket rotate with
the wall. This gives a circulation along the wall of v.2r, where
r is the radius of the bucket and v is the velocity of the wall.
Liquid Helium-4 58
8/14/2019 6 Liquid Helium 4
60/66
Let be the angular velocity of the Earth and be the
lattitude at Liverpool. Then v =r cos .
So the circulation is v.2r =r cos 2r = 2r2 cos . Thecirculation is quantised according to 2r2cos =nh/m.
The number of quanta isn=m/h 2r2 cos =m/h 2r2 2/T cos , where T isthe time of 1 day
Substituting, you will find n = 6.85.
Since n is an integer, it is likely to be 6, with the wall moving a
bit faster than the atoms next to it.
Liquid Helium-4 59
Exercises
8/14/2019 6 Liquid Helium 4
61/66
Exercises
Exercise 4
A modern Andronikashvili experiment uses a long, thin wire
vibrating in liquid helium-4.
This is a very thin metal wire, coated with a layer of aerogel.
The aerogel is a very porous material that is completely
penetrated by the liquid. As the pores are very small, normal
liquid helium in the pores would remain trapped and not be ableto move much. Superfluid helium would of course flow freely.
The metal wire has a diameter of 0.10 mm, and a density of
7000 kg m3. The aerogel layer has a diameter of 1.5 mm. It is98% empty space. The remaining 2% is silica, with a density of
2200 kg m3. The liquid helium has a density of 145 kg m3.
The wire vibrates at 240 Hz at 2.5 K. Estimate its frequency at
1 K.
(Modified from Tony Guenault, Basic Superfluids (2003))
Liquid Helium-4 60
8/14/2019 6 Liquid Helium 4
62/66
Solution
We make two simplifying assumptions:
1. The metal and the aerogel form a long, thin cylinder of
length L.
2. The relation between frequency f and mass m is the sameas for a mass on a spring, i.e. f 1/m.
The mass of the wire is mW = 0.12 7000 L= 70L.
The mass of the aerogel ismA= (1.5
2 0.12) 2200 .02 L= 99L.
The mass of the normal liquid helium is
mH= (1.52 0.12) 145 L 0.98 = 318L.
Liquid Helium-4 61
8/14/2019 6 Liquid Helium 4
63/66
At 2.5 K, the helium is still normal liquid and remains stuck tothe aerogel. So the total mass that moves with the wire is
mW+mA+mH.
At 1 K, the helium is still superfluid and the aerogel moves as if
in empty space. So the total mass that moves with the wire ismW+mA.
Using f 1/m, the frequency at 1 K ismW+mA+mH/
mW+mA
240 Hz = 408 Hz.
Liquid Helium-4 62
Exercises
8/14/2019 6 Liquid Helium 4
64/66
Exercises
Exercise 5
What are the experiments that support the two-fluid model in
superfluid helium-4?
How can this model, which says that there are two different
fluids, be consistent with the fact that all helium-4 atoms are
identical?
(Modified from Tony Guenault, Basic Superfluids (2003))
Liquid Helium-4 63
Exercises
8/14/2019 6 Liquid Helium 4
65/66
Exercises
Exercise 6
Calculate how far the temperature must fall below TB (= 3 K)
in an ideal Bose gas containing 1023 atoms in order that the
lowest energy level should be populated by 1020 atoms.
(Tilley and Tilley, Superfluidity and Superconductivity, 3rd. ed. (2003), p.
70.)
Liquid Helium-4 64
Exercises
8/14/2019 6 Liquid Helium 4
66/66
Exercises
Exercise 7
Calculate the de Broglie wavelength of a thermal neutron
(energy = 0.04 eV), and compare it with the average
interatomic spacing in liquid He II (density = 145 kg m3).
(Tilley and Tilley, Superfluidity and Superconductivity, 3rd. ed. (2003), p.
70.)