FB Ptop

Buoyant force

The volume of gas (liquid) is

at rest with respect to its
surrounding gas (liquid): the
force of gravity is balanced by
the buoyant force
FB  ( Pbottom  Ptop ) A
Pbottom h
Vg
Note that the buoyant force does not care what’s
inside this volume (a brick, a gas, or vacuum): it
depends only on the volume and the density of the
outside gas (liquid).

Ptop Pbottom P
 Lecture 3. Transport Phenomena (Ch.1)
Lecture 2 – various processes in macro systems near the state of
equilibrium can be described by a handful of macro parameters. Quasi-
static processes – sufficiently slow processes, at any moment the system
is almost in equilibrium.
It is important to know how much time it takes for a system to approach an
equilibrium state. A system is not in equilibrium when the macroscopic
parameters (T, P, etc.) are not constant throughout the system. To
approach equilibrium, these non-uniformities have to be ironed out through
the transport of energy, momentum, and mass from one part of the
system to another. The mechanism of transport is molecular collisions. Our
goal - to estimate the characteristic rates of approaching equilibrium, and,
thus, to impose limitations on the rates of “quasi-static” processes.

1. Transfers of Q (“Heat” Conduction)

2. Transfers of Mass (Diffusion)

One-dimensional (1D) case:

n(x,t)
T(x,t)
x
 The Mean Free Path of Molecules
Transports energy, momentum, mass – due to random
thermal motion of molecules in gases and liquids.
The mean free path l - the average distance traveled
by a molecule btw two successive collisions.

An estimate: one molecule is moving with a constant 4r

speed v, the other molecules are fixed. Model of hard
spheres, the radius of molecule r ~ 110-10 m.

l
The av. distance traversed by a molecule until the 1st collision is the
distance in which the av. # of molecules in this cylinder is 1.
N 1
 2 r l  1 V 1
2
1  l  Maxwell: l
V 4 r 2 N  n 2 n
n = N/V
– the density of molecules  = 4r2 – the cross section

The average time interval between successive collisions - the collision time:
l
 v - the most probable speed of a molecule
v
 Some Numbers:
1 1 T
l  for an ideal gas: PV  Nk BT P  nk BT  l 
n n P

1 V k B T 1.38 10 23 J/K  300K  26

air at norm. conditions:     4  10 m 3

n N P 105 Pa
V
the intermol. distance d 3 ~ 3 10 9 m
N
P = 105 Pa: l ~ 10-7 m - 30 times greater than d

P = 10-2 Pa (10-4mbar): l ~ 1 m (size of a typical vacuum chamber)

l
- at this P, there are still ~2.5 10 molecule/cm (!)
12 3  n 2 / 3  P 2 / 3
d
The collision time at norm. conditions:  ~ 10-7m / 500m/s = 2·10-10 s

For H2 gas in interstellar space, where the density is ~ 1 molecule/ cm3,

l ~ 1013 m - ~ 100 times greater than the Sun-Earth distance (1.5 1011 m)
 l Transport in Gases (Liquids)
Box 1 Box 2 Simplified approach: consider the “ballistic”
molecule exchange between two “boxes” within the
gas (thickness of each box should be comparable to
the mean free path of molecules, l). During the
average time between molecular collisions, ,
roughly half the molecules in Box 1 will move to the
x=-l x=0 x=l right in Box 2, while roughly half the molecules in
Box 2 will move to the left in Box 1.
Each molecule “carries” some quantity  (mass, kin. energy, etc.), within
each box -  = N  = A l n . E.g., the flux of the number of molecules
across the border per unit area of the border, Jx:
N 1 1  n  1 n n
Jn   v  n x  l   n x  l    v   2l    v l   D
At 6 6  x  3 x x
in a 3D case, on average 1/6 of the the diffusion constant
molecules have a velocity along +x or –x
“-” - if n/x is negative, the flux is in the positive x direction
(the current flows from high density to low density)
  n(x,t) Jx
In a 3D case, J n   Dn JU   K thT
x
 n1 J n2
Diffusion
J Diffusion – the flow of randomly moving particles
caused by variations of the concentration of
particles. Example: a mixture of two gases, the total
concentration n = n1+n2 =const over the volume (P
= const).
1 n n 1
Fick’s Law: Jx   l v  D D  lv
3 x x 3
- the diffusion coefficient
(numerical pre-factor depends on the dimensionality: 3D – 1/3; 2D – 1/2)

1
D l v its dimensions [L]2 [t]-1, its units m2 s-1
3
Typically, at normal conditions, l ~10-7 m, v ~300 m/s  D ~ 10-5 m2 s-1
(in liquids, D is much smaller, ~10-10 m2 s-1)
For electrons in well-ordered semiconductor heterostructures at low T:
l ~10-5 m, v ~105 m/s  D ~ 1 m2 s-1
 Diffusion Coefficient of an Ideal Gas ( Pr. 1.70 )

1 k BT
for an ideal gas: l  3/ 2
n P T T
D  T 1/ 2 
from the equipartition theorem: v  T 1/ 2 P P

1
therefore, at a const. temperature: D
P
and at a const. pressure: D  T 3/ 2
 The Diffusion Equation
flow in flow out
change of n inside:
J x  x  At n(x,t) J x  x  x  At n J
 x
t x
n
x x+x combining with J x  D
x
we’ll get the equation that describes one-dimensional diffusion:

n  2n
D 2 the diffusion equation
t x
t1 =0
the solution which corresponds to an initial condition
that all particles are at x =0 at t =0:
t2 =t C  x2 
n  x, t   exp   C is a normalization factor
Dt  4 Dt 

x =0 the rms displacement of particles: x 2  Dt

 Brownian Motion (self-diffusion)
Historical background:

The experiment by the botanist R. Brown concerning the

drifting of tiny (~ 1m) specks in liquids and gases, had been
known since 1827.
Brownian motion was in focus of the struggle for and against
the atomic structure of matter, which went on during the
second half of the 19th century and involved many prominent
physicists.

Ernst Mach: “If the belief in the existence of atoms is so crucial in your eyes, I
hereby withdraw from the physicist’s way of thought...”

Albert Einstein explained the phenomenon on the basis of the kinetic theory
(1905), connected in a quantitative manner the Brownian motion and such
macroscopic quantities as the coefficients of mobility and viscosity – and
brought the debate to a conclusion in a short time.

Observing the Brownian motion under a microscope, Jean Perrin measured

the Boltzman constant and Avogadro number in 1908 (Nobel 1926).
 Gaussian C  x2 
x n x, t   exp  
distribution Dt  4 Dt 
Brownian
Motion
x2  Dt
(cont.) a 1D random walk
t

of a drunk
the rms displacement
t
A body that participates in a random walk, or a subject of random collisions with the
gas molecules. Its average displacement is zero. However, the average square
distance grows linearly with time: 
    n - a randomly
after N steps, the position is RN 1 RN 1  RN 1  l n oriented unit

  vector
2   2 2 2  
RN 1  RN  l n  RN  l  2l n  RN
 2 2 2
after averaging ( RN 1  0 ): RN 1  RN  l 2  RN  N l 2  t

For air at normal conditions ( l  107 m v  500 m/s D  1.7 105 m 2 /s ) , it takes
L2
t  ~ 105 s for a molecule to “diffuse” over 1m: odor spreads by convection
D
For electrons in metals at 300K ( l  10 7 m v  10 6 m/s D  3 10 2 m 2 /s ), it takes
L2
t  ~ 30 s to “diffuse” over 1m. For the electron gas in metals, convection can be ignored:
D the electron velocities are randomized by impurity/phonon scattering.
 Static Energy Flow by “Heat” Conduction
In general, the energy transport due T 1 J U K th  2T T
to molecular motion is described by   J U   K th
the equation of heat conduction: t C x C x 2 x

Thus, in principle, if you know the initial conditions, e.g. T(x,t=t0), you can describe
the process by solving the equation. Often, you are asked to consider a different
situation: a static flow of energy from a “hot” object to a “cold” one. (At what rate
the internal energy is transferred between two systems with T1  T2 or between
parts of a non-equilibrium system (if one can introduce Ti) ?) The temperature
distribution in this formulation is time-independent, and we need to calculate the
flux of thermal energy JU due to the heat conduction (diffusion/intermixing of
particles with different energies, interactions between the particles that vibrate but
do not move “translationally”).
area A Heat conduction ( static heat flow, T = const)
T(x)
T1 T2
T1 JU

T2
x x
 Fourier Heat Conduction Law
Tt Q T
Q  A  JU    K th A “-” - if T increases from left to right,
x t x energy flows from right to left

Kth [W/K·m] – the thermal conductivity (material-specific)

Pr. 1.56 For a window glass (Kth =0.8W/mK, 3 mm thick, A=1m2) and T = 20K:

Q 20 K ~ 10 times greater than in reality, a thin

 (0.8 W/m  K )(1 m 2 )  5300 W layer of still air
t 0.003 m must contribute to thermal insulation.
A
JU  power   GT , G  K th
T1
G
T2 x
G – the thermal conductivity [W/K]
R =1/G – the thermal resistivity

Electricity Thermal Physics Connection in series (Pr. 1.57):

What “flows” Charge Q Th. Energy, Q T1 T2 Rtot = R1 + R2

Flux Currant dQ/dt Power Q/dt Connection in parallel:

El.-stat. pot. Temperature Rtot-1 = R1-1 + R2-1
Driving “force” difference difference
T1 T2
“Resistance” El. resistance R Th. resistance R
 Relaxation Time due to Thermal Conductivity
the heat capacity (specific heat)
(a rough estimate)
U CT C
~   T1  T2  T1  
Q / dt G T G

G the thermal conductivity

U = CT1 environment
the thermal A
T2 G  K th
conductivity x

Problem 1.60: A frying pan is quickly heated on the stovetop to 2000C. It has
an iron handle that is 20 cm long. Estimate how much time should pass before
the end of the handle is too hot to grab (the density of iron  = 7.9 g/cm3, its
specific heat c = 0.45 J/g·K, the thermal conductivity Kth=80 W/m·K).

C  c A L  c L2 7900 kg  m 3  450 J  kg 1  K 1   0.1m 

2
    1 1 1
~ 400s
G K A K th 80 J  s  m  K
th
L
 Thermal Conductivity of an Ideal Gas

Energy “flow”, t ~  : Q 1 U1  U 2  1 CV  T1  T2  1 CV l dT 1 dT

    CV v
 2  2  2  dx 2 dx
the time between two l
consecutive collisions

v
l Q T 1 CV v 1 CV l v 1 CV
  K th A K th    lv
 x 2 A 2 Al 2 V
Box 1 Box 2
f
Nk B
the specific CV 2 f
heat capacity
cV    n kB
T1 T2 V V 2

f
K th  n kB l v
T 4

The thermal conductivity of air at norm. conditions:

f 5 W
K th  n k B l v   2.4 1025 m 3  1.38 10 23 J/K 107 m  500m/s  0.02
4 4 m K
(exp. value – 0.026 W/m·K)
 Thermal Conductivity of Gases (cont.)

f T T
 l   n , v 
1
K ht  n k B l v  K th 
4 m m
1. K th  1 / m

- an argon-filled window helps to reduce Q

2. Thermal conductivity of an ideal gas

is independent of the gas density!
(at higher densities, more
molecules participate in the
energy transfer, but they
Dewar carry the energy over a
shorter distance)

This conclusion holds only if L >> l .

For L < l , Kth  n
 Sate-of-the-art Bolometers
(direct detectors of e.-m. radiation)
Q
 power   RI 2  Ge ph  T  T  Ge ph  T  Te  Tph 
dt
9
10 specific heat of electrons
HEDDs
meanders

  VTc Ge ph
8
10

G, W /K m
3
Nb
7
10

6
10
Ti
5 1000
10

ħ 4
10

0.1 1

Time constant (µs)

T, K 100

Te electrons
Ge-ph= Ce/e-ph
BT85-4
BT87-1

Ge  ph  G
10
BT100-1

Tph phonons BT121-1

BT121-4

G = Cph/es  ~  e  ph 1

T heat sink 10 100

Temperature (mK)
1000
 Momentum Transfer, Viscosity

Drag – transfer of the momentum in the

direction perpendicular to velocity. z ux

 px A   u x , top  u x , bottom 
 Fx  Laminar flow of a gas (fluid) between two
t z surfaces moving with respect to each other.

Fx d ux Fx – the viscous drag force,  - the coefficient of viscosity


A z Fx/A – shear stress

Box 2 ux (z2)
Viscosity of an ideal gas ( Pr. 1.66 ): z ~ l
Box 1 ux (z1)
1 1 1
p z  Nmu x ( z1 )  Nmu x ( z2 )   Nm u x
2 2 2
Fx 1 p z Nmu x v Fx  1  du 1
      v l x   vl  T1/2
A A  2Al A 2 dz 2
 Effusion of an Ideal Gas
- the process of a gas escaping through a small hole (a << l)
into a vacuum (Pr. 1.22) – the collisionless regime.
The opposite limit of a very large hole ( a >> l ) – the hydrodynamic regime.

The number of molecules that escape through

a hole of area A in 1 sec, Nh, in terms of P(t ), T
(how is T changing in the process?)
p 1 2m v x 1 P A t 1 1 k BT
P  Nh  Nh Nh  vx 
2
m v x  k BT , vx  vx
2

t A t A 2m v x 2 2 m

Nh = - N, where N is the total # of molecules in a system

P A t m Nk BT At m ANt k BT N A k BT 1
 N     N  N
2m k BT V 2m k BT 2V m t 2V m 

 t 2V m
N (t )  N (0) exp   ,  
  A k BT

Depressurizing of a space ship,

V - 50m3, A of a hole in a wall – 10-4 m2 2  50 m 3 30 1.7 10 27 kg
  4 2  23
 106 10  2  0.3 s  3000 s
10 m 1.38 10 J/K  30 K
(clearly, a << l does not apply)