Diffusion and Dispersion of

Pollutants
Sohani Afroja

Diffusion and Dispersion of Pollutants 1

Diffusion
• Diffusion occurs when a molecule changes place with an adjacent
molecule. Typically, the molecules of the pollutant released into a
carrier medium change places with adjacent molecules of the carrier
medium.

Diffusion and Dispersion of Pollutants 2

Diffusion
• Diffusion is a physical process that refers to the net movement of
molecules from a region of high concentration to one of lower
concentration. The material that diffuses could be a solid, liquid or
gas.

Diffusion and Dispersion of Pollutants 3

Diffusion
• The movement of particles in a solid from an area of high
concentration to an area of low concentration, resulting in the
uniform distribution of the element.
• Diffusion is process which is NOT due to the action of a force, but a
result of the random movements of atoms

Example: Two chambers, each containing a different gas, separated by a

removable barrier; when the barrier is pulled away, interdiffusion occurs.

Diffusion and Dispersion of Pollutants 4

Rates of Diffusion
• The rate of diffusion depends on the nature of interaction between
the medium and material. For instance, a gas diffuses very quickly in
another gas.

• An example of this is the way the harmful smell of ammonia gas

spreads in air. Similarly, if a container of liquid nitrogen leaks a little,
nitrogen gas that escapes would quickly diffuse into the atmosphere.
The same gas would diffuse slightly more slowly in a liquid such as
water and slowest in a solid.

Diffusion and Dispersion of Pollutants 5

 Rates of Diffusion
• Similarly, two mixable
liquids will also diffuse
into each other to form a
uniform solution.

• For example, when water

is mixed with glycerol,
over time the two liquids
diffuse radially into each
other.

Diffusion and Dispersion of Pollutants 6

 Wedge Machine
• Consider a physical system consisting of an array of
wedges in uniform rows.
• The wedges might be constructed of pieces of wood,
for example, glued to a plywood back panel with a sheet
of glass as the face.
• Suppose that many small particles—say, grains of
sand—can be released from a “source” at the very top
of the structure and that they eventually fall by gravity
through the array to the bottom of the structure.
• As we shall see, the movement of particles downward
through this “wedge machine” is a mechanical analog of
the dispersion and diffusion of pollutants in the
environment. Diffusion and Dispersion of Pollutants 7
Movements of the Particles
• The major movements of the particles can be separated into two
components:
(1) downward motion (drift) due to gravity, and
(2) side-by-side movements due to encounters with the wedges.

Diffusion and Dispersion of Pollutants 8

 Distribution with Respect to Space (downward
motion (drift) due to gravity) 𝟏
𝟏 𝑷 𝑩𝟏 =
𝑷 𝑨𝟏 = 𝟐
𝟐

𝑷 𝑩𝟐 = 𝑷 𝑨𝟐 ȁ𝑨𝟏 . 𝑷 𝑨𝟏 + 𝑷 𝑩𝟐 ȁ𝑩𝟏 . 𝑷 𝑩𝟏
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
= ∗ + ∗ = + = .
𝟐 𝟐 𝟐 𝟐 𝟒 𝟒 𝟐
𝟏 𝟏 𝟏
𝑷 𝑨𝟐 = 𝑷 𝑨𝟐 ȁ𝑨𝟏 . 𝑷 𝑨𝟏 = ∗ = .
𝟐 𝟐 𝟒 𝟏 𝟏 𝟏
𝑷 𝑪𝟐 = 𝑷 𝑪𝟐 ȁ𝑩𝟏 . 𝑷 𝑩𝟏 = ∗ = .
𝟐 𝟐 𝟒

Diffusion and Dispersion of Pollutants 9

Distribution with Respect to Space (downward
motion (drift) due to gravity)

Diffusion and Dispersion of Pollutants 10

Distribution with Respect to Space (downward
motion (drift) due to gravity)

Diffusion and Dispersion of Pollutants 11

 Distribution with Respect to Space (downward
motion (drift) due to gravity)
• Consider a single particle that is released from the
source at the top of the array and falls downward.

• When it reaches the upward tip of the wedge in row 1,

assume that it has probability 𝒒 = 𝟏/𝟐 of falling down
the left side of this wedge (route 𝑨𝟏 ) and probability
𝒑 = 𝟏/𝟐 of falling down the right side of this wedge
(route 𝑩𝟏 ). ). Suppose that this particular particle
happens to travel down route 𝑨𝟏 .

Diffusion and Dispersion of Pollutants 12

 Distribution with Respect to Space
• In row 2, where a similar decision must be made. Once again, it has probability 𝒑 =
𝟏/𝟐 of traveling down the right side of the wedge (route 𝑩𝟐 ) and probability 𝒒 =
𝟏/𝟐 of traveling down the left side of the wedge (route 𝑨𝟐 ).

• The probability that the particle will travel down route, given that it has traveled
down route route 𝑨𝟏 can be written as: 𝑷[𝑨𝟐 ȁ𝑨𝟏 ] = 𝟏/𝟐.

• Then, the probability that a particle initially released from the source actually ends
up traveling down route 𝑨𝟐 will be 𝟏the𝟏 product 𝟏
of the probabilities at the two
wedges: 𝑷 𝑨𝟐 = 𝑷 𝑨𝟐 ȁ𝑨𝟏 . 𝑷 𝑨𝟏 = ∗ = .
𝟐 𝟐 𝟒

• If we now consider route 𝑩𝟐 , in the middle of the two wedges in row 2, we see that a
particle can reach this point by two alternative routes.
• It can travel down the left side of the top wedge or it can travel down the right side
of the top wedge. Assuming that 𝒑 = 𝒒 = 𝟏/𝟐 for all wedges in the structure, the
probability that the particle actually travels down route 𝑩𝟐 will be given by the union
of the two possible paths:
𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
• 𝑷 𝑩𝟐 = 𝑷 𝑨𝟐 ȁ𝑨𝟏 . 𝑷 𝑨𝟏 + 𝑷 𝑩𝟐 ȁ𝑩𝟏 . 𝑷 𝑩𝟏 = ∗ + ∗ = + = .
𝟐 𝟐 𝟐 𝟐 𝟒 𝟒 𝟐
Diffusion and Dispersion of Pollutants 13
Distribution with Respect to Space
• Finally, the probability that a particle travels down route 𝑪𝟐 will be
computed as the intersection of the probabilities for the two relevant
decision points:
𝟏 𝟏 𝟏
𝑷 𝑪𝟐 = 𝑷 𝑪𝟐 ȁ𝑩𝟏 . 𝑷 𝑩𝟏 = ∗ = .
𝟐 𝟐 𝟒

• As the probabilities for each row are computed in the manner described
above, it becomes evident that each row can be represented as a Bernoulli
process. The particle faces three choices in row 2: {𝑨𝟐 , 𝑩𝟐 , 𝑪𝟐 }. The
corresponding values for 𝐾 will be {0,1,2}, and the probability distribution
𝟏
for 𝐾 will be 𝑩(𝟐, ), giving the following PMF:
𝟐
𝟐
𝑷𝑲=𝒌 = 𝟏/𝟐 𝒌 𝟏/𝟐 𝟐−𝒌 ; 𝒇𝒐𝒓 𝒌 = 𝟎, 𝟏, 𝟐.
𝒌 Diffusion and Dispersion of Pollutants 14
Distribution with Respect to Space
• The probability distribution for row 3, which contains four possible
routes, will be 𝑩(𝟑, 𝒍/𝟐).

• Using similar reasoning, and considering the possible routes for each
particle, it can be seen that the probability distribution for row 4 will be
𝑩(𝟒, 𝒍/𝟐),

• The probability distribution for row 5 will be 𝑩(𝟓, 𝒍/𝟐).

• In general, if the array extended to any row m, the probability distribution

for the number of arrivals would be 𝑩(𝒎, 𝒍/𝟐). As m becomes larger, the
heights of the probability bars, when plotted, increasingly resemble the
familiar bell-shaped curve of the normal distribution.

Diffusion and Dispersion of Pollutants 15

Distribution with Respect to Space
• Now suppose that a large number of particles are released from the
source at the top of the array.
• Let 𝑁𝐴 be a random variable denoting the final number of particles
that arrive in channel A at the bottom of the array.
• Let 𝒏𝟎 = 𝟏𝟎𝟎𝟎 particles to be released at the source and assume
that they are released one at a time so that they do not interact with
each other in any manner as they fall down the array.
• Since the probability that a given particle will arrive in channel A is:
𝑷{𝑨} = 𝟏/𝟑𝟐 = 𝟎. 𝟎𝟑𝟏𝟐𝟓,
• The expected number of arrivals in channel A will be:
𝑬 𝑵𝑨 = 𝒏𝟎 . 𝑷{𝑨} = (𝟏𝟎𝟎𝟎)(𝟎. 𝟎𝟑𝟏𝟐𝟓) = 𝟑𝟏. 𝟐𝟓 particles.
Diffusion and Dispersion of Pollutants 16
Distribution with Respect to Space
• The expected number of arrivals for each channel is computed in a similar
manner:
• 𝐸[𝑁𝐴 ] = 31.25
• 𝐸[𝑁𝐵 ] = 156.25
• 𝐸[𝑁𝐶 ] = 312.50
• £[𝑁𝐷 ] = 312.50
• 𝐸[𝑁𝐸 ] = 156.25
• 𝐸[𝑁𝐹 ] = 31.25
• 𝑇𝑜𝑡𝑎𝑙 = 1000.00 𝑝𝑎𝑟𝑡𝑖𝑐𝑙𝑒𝑠

Diffusion and Dispersion of Pollutants 17

 Distribution with Respect to Space (side-by-side movements )
• When a particle encounters a wedge, it experiences a displacement either to
the left or to the right, with probability
𝒑 = 𝒒 = 𝟏/𝟐.
• Horizontal position is determined by the algebraic sum of the left and right
horizontal displacements it has experienced.

• Consider an experiment in which some object begins at an origin and

experiences “jumps,” or fixed displacements of length d, in one dimension
until it reaches a final position.

• In the figure, the object begins at position 𝑲 = 𝟎, experiences four

displacements, and ends at position 𝑲 = 𝟒.

• The object moves to the rightDiffusion

by displacement
and Dispersion of Pollutants
𝒅. 18
Distribution with Respect to Space (side-by-side movements )
• Suppose that a fair coin is flipped on m trials, and the following rule is
adopted based on the outcome of each trial:

• If a “head” appears, the object moves to the right by displacement d;

• If a “tail” appears, the object remains where it is.

• Object moves to the right with probability 𝒑 = 𝟏/𝟐, or it remains

stationary with 𝒑𝒓𝒐𝒃𝒂𝒃𝒊𝒍𝒊𝒕𝒚 𝒒 = 𝟏/𝟐.

Diffusion and Dispersion of Pollutants 19

Distribution with Respect to Space (side-by-
side movements )
• After 𝒎 displacements, the object can occupy any position
𝐾 = 0,1,2, … , 𝑚 .

• The number of displacements will be a Bernoulli process, and the

position of the object will have a binomial distribution.

• The expected value of this distribution will be 𝑬[𝑲] = 𝒎𝒑 = 𝒎/𝟐.

Therefore, the distribution will be symmetrical about the midpoint of its
range 𝟎, 𝒎 .

Diffusion and Dispersion of Pollutants 20

Distribution with Respect to Space (side-by-side movements )
• Consider a slightly different case in which the object can move either to
the left or right.

• If a “head” appears, it moves to the right by displacement d;

• if a “tail” appears, it moves to the left by displacement -d.

• Thus, if, on 5 trials, 3 “heads” and 2 “tails” occur, the object will be at
position Kd = 3d - 2d = d.

• The distribution is symmetrical about the origin, K = 0 for rightward-

leftward movements.
Diffusion and Dispersion of Pollutants 21
Distribution with Respect to Time
• Suppose that particles are released from the source continuously, one at a time with
rate, 𝒓 = 𝟏 particle per minute.

• Then the number of particles released from the source during any time period t will be
𝒏𝟎 𝒕 = 𝒓𝒕.

• If 10-minute periods are considered, the number of particles released will be n = 10 and
the expected number of arrivals in channel A during any 10-minute period will be
𝟏
E[𝑵𝑨 ] = 𝟏𝟎𝑷 𝑨 = 𝟏𝟎 = 𝟎. 𝟑𝟏𝟐𝟓 particles.
𝟑𝟐

• If 100 minutes are considered, the expected number of arrivals in channel A will be
E[𝑵𝑨 ] = 𝟏𝟎𝟎𝑷{𝑨} = 𝟏𝟎𝟎(𝟏/𝟑𝟐) = 3.125 particles.

• Finally, if 1000-minute time periods are considered, the expected number of arrivals in
channel A will be
E[𝑵𝑨 ] = 𝟏𝟎𝟎𝟎𝑷{𝑨} = 𝟏𝟎𝟎𝟎(𝟏/𝟑𝟐) = 31.25 particles.
Diffusion and Dispersion of Pollutants 22
Distribution with Respect to Time
• Suppose that the average number of particles arriving in channel A is 𝑿𝑨 given by
𝑿𝑨 = 𝑵𝑨 /𝒕.
• For 10-minute time periods, the expected value of the average will be given by the
expected number of arrivals divided by time,
E[𝑿𝑨 ] = 𝑬[𝑵𝑨 ]/𝒕 = 𝟎. 𝟑𝟏𝟐𝟓/𝟏𝟎 = 𝟎. 𝟎𝟑𝟏𝟐𝟓 particles per minute.

• For 100-minute periods, the expected value of the average will be

𝑬[𝑿𝑨 ] = 𝟑. 𝟏𝟐𝟓/𝟏𝟎𝟎 = 𝟎. 𝟎𝟑𝟏𝟐𝟓 particles per minute.

• The expected value of the average for 1000-minute periods will be

𝑬[𝑿𝑨 ] = 𝟑𝟏. 𝟐𝟓/𝟏𝟎𝟎𝟎 = 𝟎. 𝟎𝟑𝟏𝟐𝟓 particles per minute.

• It is equal to the particle generation rate at the source 𝑟 multiplied by the

probability that a particle arrives at channel A; that is,
𝑬[𝑿𝑨 ] = 𝒓𝑷{ 𝑨} = (𝟏)(𝟏/𝟑𝟐) = 𝟎. 𝟎𝟑𝟏𝟐𝟓 particles per minute.
Diffusion and Dispersion of Pollutants 23
Plume Model
• Suppose that the particles of a pollutant are
released from a point source.

• Instead of gravity, assume that the drift now is

caused by the predominant motion of the carrier
medium at a constant speed and direction.

• An air pollution example is the discharge of

particles from a smokestack into a wind moving
with flow in a constant direction.
Diffusion and Dispersion of Pollutants 24
Plume Model
• On the 𝒊𝒕𝒉 collision, the particle will experience horizontal
displacement 𝒅𝒊 .

• If 𝒚(𝒎) denotes the horizontal position of the particle after m collisions,

then 𝒚(𝟎) = 𝟎 is the initial position at the source, and

• The final position of the particle will be the sum of all the horizontal
displacements it has experienced since it left the source:
𝒚(𝒎) = 𝒅𝟏 + 𝒅𝟐 + ⋯ + 𝒅𝒎 .

Diffusion and Dispersion of Pollutants 25

Plume Model
• Let 𝒀 be a random variable denoting the horizontal position of the particle
measured by its positive or negative horizontal distance from the center
line.

• Let 𝑫 be a random variable denoting its horizontal displacement on each

successive collusion.

• Then Y for an individual particle will be the sum of the realizations

{𝒅𝟏 , 𝒅𝟐 , … , 𝒅𝒎 } obtained from the random variable D.

• Assume that the particles are displaced to the left or to the right with
equal probability and that the expected value of the displacement is zero:
𝑬[𝑫] = 𝟎.
Diffusion and Dispersion of Pollutants 26
Plume Model
• Because the displacements are independent, the expected value of Y will
be the sum of the expected values of the displacements by the rules for
combining expected values:
𝑬[𝒀] = 𝑬[𝒅𝟏 ] + 𝑬[𝒅𝟐 ] + ⋯ + 𝑬[𝒅𝒎 ] = 𝑚 𝑬[𝑫] = 𝟎

• Similarly, the variance of Y will be the sum of the variances of the

individual displacements by the rules for combining variances

𝑽[𝒀] = 𝑽[𝒅𝟏 ] + 𝑽[𝒅𝟐 ] + ⋯ + 𝑽[𝒅𝒎 ] = 𝑚 𝑽[𝑫]

• The final position Y of the particle after m collisions will asymptotically

approach a normal distribution whose mean is at the center line and
variance is proportional to m. The probability density
𝟐
function of the
normal distribution with mean 0 and variance 𝝈 = 𝒎 is as follows:

Diffusion and Dispersion of Pollutants 27

Plume Model
• If many particles were
released at once from the
source, then their
“expected arrival density”
(expected number of
particles per unit length)
will follow similar normal
distributions.

• The expected number of

particles arriving in a given
segment of the y-axis will
be given by the area under
the normal curve
Diffusion and Dispersion of Pollutants 28