----- noc19-ph02 Lecture 53 ---- old

IIT Kanpur July 201
20 followers
1 Video View·Dec 3, 2023
We have calculated and discussed the density of states of phonons and we have also
discussed the average energy of a mode. And we are now ready to combine the two to get
the specific heat through the route of calculating total energy of a system where
these phonons are there or a crystal with vibration both are equivalent.
So let us see what we did we said if I have an optical mode in this take density of states
to be independent of omega and that means the vibrations are all at the same frequency
and this we call this the Einstein model. And we also had Debye model in which case
we assumed that omega as a function of k is always equal to V sound times k that is a
linear dispersion and in that case the density of states d omega came out to be V over 2
pi square omega square over V s cubed. So this changes with the frequency and then
we had also calculated energy of a mode with frequency omega this I will call average energy
of a mode with frequency omega at temperature T and this came out to be let us call it E
average came out to be h cross omega over E raised to h cross omega over k Boltzmann
T minus 1. And on the side it also meant that the number of phonons is equal to 1 over E
raised to h cross omega over kB T minus 1 number of phonons times the energy of each
phonon h cross omega gives you the average energy.
So we now combine all this and get the specific heat so let us first do Einstein's model
and this is applicable optical phonons in a good way because optical phonons have a
very flat omega k versus k curve. So in this case the density of states is equal to N delta
omega minus omega 0 where omega 0 is the frequency of oscillation and where N is the
number of atoms or number of cells over which the boundary condition is applied. This is
for one particular mode this is an optical mode we are talking about. So this is about
one particular mode it could be a transverse mode or a longitudinal mode total number of
modes is going to be 3N. So in this case when I calculate the total energy is going to be equal
to integration over all omegas average energy of this mode times d omega 0 to infinity and
this comes out to be so let us write this again E omega will come out to be 0 to infinity d
omega h cross omega over e raised to h cross omega over kB T minus 1 d omega now is N times
delta omega minus omega 0. So this comes out to be h cross omega 0 over e raised to h cross omega
0 over kB T minus 1 times N. Now for counting all modes because remember this is one mode over
which I applied the periodic boundary condition there are two other modes. So there are going
to be two transverse and one longitudinal mode I have to multiply this number by 3.
You can think of it another way these N cells if there are N atoms or N units all of them
together are going to have 3N degrees of freedom and we have calculated this energy only for one
degree of freedom. So actual energy is going to be E equal to 3N h cross omega 0 over e raised
to h cross omega 0 over kB T minus 1. So if all atoms vibrate with the same frequency the total
energy E comes out to be N times 3 h cross omega 0 where omega 0 is the frequency of vibration e
raised to h cross omega 0 over kB T minus 1. Now this is a little complicated expression
let us learn about it more by taking different limits. So if I take T going to infinity very
large then e raised to h cross omega 0 over kB T minus 1 is approximately equal to h cross omega 0
over kB T and therefore total energy E is going to be 3N h cross omega 0 over h cross omega 0
over kB T and this term cancels and I get total energy to be 3N kB T and therefore Cv which is
dE dt at constant volume constant volume because then the frequency would remain the same is 3N
kB which is 3R per unit mole which is the Durand-Pettit law experimentally you observe
that at high temperatures the specific heat is 3R or 24 joules per mole. So I have E equals 3N
h cross omega 0 over e raised to h cross omega 0 over kB T minus 1 as T goes to 0 e raised to
h cross omega 0 over kB T goes to infinity it becomes very large. So I can neglect that 1 and
the total energy E as 3N h cross omega 0 e raised to minus h cross omega 0 over kB T and therefore
Cv which is dE dt at constant volume is going to be equal to 3N h cross omega 0 times h cross
omega 0 over kB T square e raised to minus h cross omega 0 over kB T.
So this is nothing but 3N h cross omega 0 square over kB T square e raised to minus h cross omega
0 over kB T. As T tends to 0 this exponential factor goes down to 0 very fast much faster than
any polynomial and therefore this 1 over T square in front does not really matter and the heat
capacity Cv will tend to 0 as T tends to 0. If I plot it therefore T versus Cv I am going to get
a curve this fellow goes to 0 exponentially and as T goes to large value it goes to 3R roughly
24 joules per Kelvin per mole and it goes to 0. So this was the first calculation that was done
by Einstein to show that quantum mechanically if I treat the vibrations of a crystal quantum
mechanically that means the energy excitation is going to be with h cross omega the Cv indeed
goes to 0 as T goes to 0 and this was a very nice discovery by him and proposed that these
vibrations should be treated quantum mechanically. Although the actual experimental test it was
found that the behavior is not quite as fast going to 0 as proposed by Einstein it goes more as T
cubed. So we have to explain that and that comes through Debye model. So next we are going to
discuss now is Debye model. So remember in Debye model what we have is omega k is equal to velocity
of sound times density of states I am going to calculate now again it does not hurt if I just
repeat myself is d omega which is calculated through counting the number of states in k
space. So let me just do it quickly in k space the total volume of sphere that contains all the
modes and these are going to be acoustic modes let us say is kd then this sphere of radius kd
has volume 4pi by 3 kd cubed and the total number of modes is therefore going to be V over 8pi by
3 cubed times this volume and this is equal to the total number of modes.
Let us do some cancellations this gives me 2pi square and therefore kd comes out to be 6pi square
n over V raised to one third that is for one of the modes and correspondingly I would also have
omega maximum which I will call omega Debye that is what this d stands for which will be equal to
Vs kd up to which all the modes will be exhausted. If you look at it slightly differently what I am
going to have is integral d omega integrated over from 0 to omega d should be equal to total number
of modes which will again give me and I leave this calculation to you is going to give me that
omega Debye is going to be Vs times 6pi square n over V raised to one third.
So let us just estimate this omega d corresponds to what temperature which I will call Debye
temperature. So Debye temperature let me call it theta Debye which is going to be equal to h cross
omega d over KB omega d is equal to Vs times 6pi square n over V raised to one third n over V is
the density of atoms. So if I substitute the numbers Vs is of the order of 1000 so this is
of the order of 1000 6pi square times n over V and n is Avogadro number but I want to write
the units per meter cubed. So it is going to be 6 times 10 raised to 29 raised to one third
roughly equal to 1000 times pi square is roughly 10 so 10 times 10 raised to 29 gives me 10 raised
to 30 so that is going to be 10 raised to 10 and 6 times 6 is 36 so I can write it as 36 one third
of 36 is roughly 3. So this is equal to 3 times 10 raised to 13 radians per second and therefore
theta Debye will be equal to h cross which is 10 raised to minus 34 times 3 times 10 raised to 13
divided by KB which is 10 raised to minus 23 so this comes out to be roughly 300 K.
So let me collect all this and tell you that in Debye model I have omega d which is the omega max
up to which all the modes are exhausted that means by this frequency all the modes are exhausted so
this is the uppermost frequency that I excite that is omega d correspondingly there is a Debye
vector right and there is a temperature theta d omega d is of the order of 10 raised to 13 Kd
would be correspondingly whatever you calculate this I will maybe give it as an assignment and
theta d is of the order of 300 K so let us say 300 to 500 K. Let us calculate the total energy
therefore the total energy of the system is going to be integral 0 to omega Debye V over 2 pi square
omega square over Vs cube that is the density of states times h cross omega over e raised to h
cross omega over KBT minus 1 d omega. So which I can collect different terms right as V h cross
over 2 pi square Vs cubed integral 0 to omega d omega cubed d omega over e raised to h cross
omega over KBT minus 1 that is the expression. So the total energy is V h cross over 2 pi square
Vs cubed integral 0 to omega d omega cubed over e raised to h cross omega over KBT minus 1 d
omega to facilitate the integral I am going to change variables to x equals h cross omega over
KBT. So that I have omega equals KBT over h cross x d omega equals KBT over h cross dx for a given
temperature and the upper limit of the integral is x d which is theta Debye over T and therefore
the total energy can be written as e equals V h cross over 2 pi square Vs cubed 0 to theta Debye
over T omega cubed is going to be KBT cubed T cubed over h cross cubed times d omega which
is going to be KBT over h cross dx there is going to be an x cubed here and 1 over e raised to x
minus 1. Collecting different terms I am going to get from here e equals V h cross over 2 pi
square Vs cubed KB raised to 4 T raised to 4 over h cross raised to 4 integral 0 to theta Debye
divided by T x cubed over e raised to x minus 1 dx which is equal to V KB raised to 4 T raised
to 4 divided by 2 pi square Vs cubed h cross cubed 0 to theta D over T x cubed over e raised
to x minus 1 dx that is the total energy in this volume V. Notice that there is a term here
proportional to T raised to 4 if this theta Debye T was not there this is reminiscent of Stefan's
law for phonons where the energy comes out to be proportional to T raised to 4.
They are the uppermost limit because the number of modes is infinity so that
proportionality is exactly T raised to 4. Here there is slightly more complication but in
appropriate limits you will see that limits the law becomes exactly like Stefan's law. So,
if I take the limit of T becoming very small in that case e becomes V KB raised to 4 over 2 pi
square Vs cubed h cubed 0 to infinity and there is a T raised to 4 x cubed over e raised to x
minus 1 dx and the dependence on temperature for the total energy becomes exactly T raised to 4
which is like the photon energy. Notice that this integral is a number it does not depend on T
anymore. So, this is exactly like photon dependence on the other hand if the temperature
becomes very large then the upper limit becomes very small. So, in the limit of T going to 0 I
have e which is V over 2 pi square Vs cube h cross cube KB raised to 4 T raised to 4 integral 0 to
infinity x cubed e raised to x minus 1 dx. This number is well known the value of the integral
is pi raised to 4 over 15 but anyway this is a constant. So, if I take Cv which is dE by dt at
constant volume is going to be come out to be 4 V KB raised to 4 T cubed over 2 pi square Vs cubed
h cross cubed integral 0 to infinity x cubed over e raised to x minus 1 dx. So, it is proportional
to T cubed. How small this temperature? We are talking about temperature which is much,
much less than theta dy. So, we are really talking about temperature T which is much,
much less than say 100 to 500 depending on the material. So, we are really talking about
temperature in the range of may be 1 degree, 2 degree, 10 degrees or so on. In that case
the specific heat is going as T cubed. Now, let us look at the other limit if I take T
becoming very large. In that case I have E equals V over 2 pi square Vs cubed h cross
cubed KB raised to 4 T raised to 4 integral 0 to very small number theta d over T x cubed
over e raised to x minus 1 dx and now theta d over T is a very small number. In fact much,
much less than 1 and therefore I can write the integrand x cubed over e raised to x minus 1
as roughly equal to x square by expanding e raised to x minus 1 as x and we get therefore
E equals V over 2 pi square Vs cubed KB raised to 4 T raised to 4 times theta d cubed over 3 T
cubed. Now, I can cancel certain terms. This T cubed goes with this T raised to 4. This becomes
T and I get energy proportional to T and this implies Cv is equal to a constant.
Now, notice that all this I have done taking only 1 mode into account. Now, I should be
actually writing E taking all 3 modes of 3 dimensional feature or 3 N degrees of freedom
into account as 3 V over 2 pi square Vs cubed h cross cubed KB raised to 4 T raised to 4 0 to
theta d by over T x cubed over e raised to x minus 1 dx. So, I have to multiply by this factor 3.
So, as T goes to infinity, this E goes to 3 V over 2 pi square Vs cubed KB raised to 4 T raised
to 4 over h cross cubed 0 to very small number x cubed over e raised to x minus 1 dx and e raised
to x minus 1, the integrand now is only for small values can be approximated as 1 plus x minus 1
which is equal to x. This is because I have the theta d over T to be much, much, much less than
1 and this immediately tells me that the integrand now is nothing but this integral is 0 to theta d
over T x square dx. Therefore, I can write the total energy E as 3 V over 2 pi square Vs cubed
KB raised to 4 T raised to 4 over h cross cubed the times theta d cubed over T cubed cancelling
this T cubed with T raised to 4 and getting T here, I get E is proportional to T.
So, energy in the limit of large temperature comes out to be proportional to T. Let us see
exactly what the expression is. So, I get E equals 3 V over 2 pi square Vs cubed h cross
cubed times KB raised to 4 T theta d cubed over again cancelling this 3, we get total energy
equals V KB raised to 4 theta d by cubed over 2 pi square Vs cubed h cross cubed times T.
Therefore, Cv which is dE dt at constant volume is going to be V KB raised to 4 theta d by cubed
over 2 pi square Vs cubed h cross. I leave it as an exercise for you to show that by substituting
theta d by and all those things together, this actually is 3 N KB and therefore, per mole it
comes out to be 3 R per mole. So, let me summarize this lecture by pointing out that we have
calculated total heat energy if you like. So, which I mean because of the temperature in a
crystal by doing this integral 0 to infinity d omega h cross omega over E raised to h cross
omega over KB T minus 1 d omega. And in the limit of large and small temperatures we have
calculated this in Einstein model where the heat capacity goes to 0 as T goes to 0, but
and we have calculated this in Debye model where we had an approximation for omega K.
So, here omega is a constant, here omega K depends on the material on K as omega K equals V sk and
we have found that Cv as a function of temperature goes as exponentially down in Einstein model and
T cubed in the large limit both goes to 3 R as T cubed in Debye model and this matches.
As a supplement to this lecture I will send you some experimental figures that will show
you the real graphs for real systems where you will see this T cubed behavior.
