Section 3.

4 Exponential Growth and Decay

Many natural systems grow or decay over time. For example, population, radioactivity, cooling,
heating, chemical reactions, and money.

Let y = f (t)
some function that represents the number of something with respect to time.

If we want to think about how something changes – as in a “rate of change” what do we look at?

A derivative: y' = f ' (t)

But wouldn't you agree that it is reasonable that this rate of change (of say, population) must be
somehow related to the original f (t)?

In fact, most of the time, in growth & decay problems:

f ' (t) = k f (t) where k is called a “proportionality constant”

Normally, this is written as:

dy
= k y which is a differential equation
dt

If k
0  called “law of natural growth”

If k  0  called “law of natural decay”

This particular differential equation is quite easy to solve because you are looking for a function y
whose derivative is a constant multiple of itself. Do we know any such functions? Yes and only one:
et

dy
Theorem: The only solutions of = k y are the exponential functions y t  = y 0ekt .
dt

How can we use this information? To solve practical problems related to growth and decay! You (the
class) should review the exercises given as examples on pg. 168 – 173. I will do different ones in the
notes so you can see as many as possible.
Examples

1. A common inhabitant of human intestines is the bacterium E. coli. A cell of this bacterium in a
nutrient-broth medium divides into two cells every 20 minutes. The initial population of a
culture is 60 cells.
a. Find the relative growth rate. (This means, what is k?)

dy
= k y where y t  = y 0ekt (note k is the same in both equations)
dt

Let y (t) = population of bacteria

t = time, in hours (choose your own scale)
y(0) = initial population

What do we already know?

y(0) = 60
1 1
at t = 20 min = hr, y   = 2 60  = 120
3 3
Use the formula:
y t  = y 0 ekt

1
y   = 60 e
k 
1
3
3

120 = 60 e
 k
1
3

2=e
 k
1
3
solve for k using logarithms

1
ln 2  = k
3
3 ln 2 = k
ln  23  = k
k = ln 8

b. Find an expression for the number of cells after t hours. (That means put k into your form
and keep general t)

y t  = y 0ekt
y t  = 60e t ln 8
So: t
y t  = 60eln 8
y t  = 608 t

c. Find the number of cells after 8 hours. (This means use your formula when t = 8)

y 8  = 6088  1,006,632,960 (wow!)

d. Find the rate of growth after 8 hours.

 dy
Just like when we did rate of change problems, rate of growth is .
dt
dy
We know = ky
dt
So, substitute in what we know when t = 8.

dy 8
dt 
8 = ln 
8608  2.09 billion
k y(8)

e. When will the population reach 20,000 cells? (This means, find t when y (t) = 20,000)

So, use your formula:

y t  = y 0ekt
20,000 = 608 t
20,000
= 8t
60
1,000
= 8t
3
log8  1,000
3 
= log8 8
t

t = log 8 
1,000
3 
t=
ln  1,000
3 
(leave in this form)
ln 8
t  2.8 hrs
2. The table gives the population of the United States, from census figures in millions, for the
years 1900 to 2000.

Year Population
1900 76
1910 92
1920 106
1930 123
1940 131
1950 150
1960 179
1970 203
1980 227
1990 250
2000 275

a. Use an exponential model and the census figures for 1900 to 1910 to predict the population
in 2000. Compare to the actual figure and try to explain the discrepancy.

dy
= ky where y t  = y 0ekt
dt

Let y (t) = population of bacteria

t = time, in hours
y(0) = initial population

What do we already know?

In 1900 y = 76 Let 1900 be t = 0
In 1910 y = 92 So, 1910 is t = 10

y(0) = 76
at t = 10 years, y(10) = 92
Use the formula to find k (the relative growth rate)
y t  = y 0ekt
k 10
92 = 76e
92
= e10k
76
ln
23
19 
= 10 k

k=

1
10
ln  
23
19
 0.0191

Leave like this

 Put back into our formula

y t  = 76 e
1
10  
ln
23
19
t

What we really want is population estimate when it's the year 2000.
What is t then? t = 100 yrs

Use your formula:

y 100 = 76 e
1
10
ln  
23
19
100

 
10 ln
23
19

y 100 = 76 e
Leave like this
y 100  513.5 million

Our census said 275 million. Why are we so far off? The formula is based on what has
happened in 1900 to 1910 – it doesn't account for outside circumstances. Perhaps declining
birth rate, less immigration, etc.

b. Use an exponential model and the census figures for 1980 to 1990 to predict population in
2000.

In 1980 y = 227 Let 1980 be t = 0

In 1990 y = 250 So, 1990 is t = 10

y(0) = 227
at t = 10 years, y(10) = 250

Follow the same process: Find k

y t  = y 0ekt
k 10
250 = 227e
250
= e10k
227
ln  
250
227
= 10 k

k=

1
10
ln  
250
227
 0.00965

Leave like this

Put back into our formula

y t  = 227 e
1
10
ln  
250
227
t

Population estimate when it's the year 2000. t = 20 yrs

Use your formula:
 y 20 = 227 e
1
10
ln
250
227 
200

y 20  = 227 e
2 ln  
250
227
2

y 20 = 227 e
ln  
250
227

250 2
y 20 = 227

227  
 275.3 million

Leave like this

As compared to 275 million, not bad at all.

3. Bismuth-210 has a half-life of 5.0 days.

a. A sample originally has a mass of 800 mg. Find a formula for the mass remaining after t-
days.

dy
= ky where y t  = y 0ekt
dt

Let y(t) = mass of Bismuth-210, in mg

t = time, in days
y(0) = initial mass, in mg

What do we already know?

y(0) = 800
when t = 5 days, y(5) = 400
Note, half-life is the amount of time for ½ of the material to decay (or be removed)

Use formula to find k.

y t  = y 0ekt
k 5
400 = 800e
400
= e5k
800
ln
1
2 
= ln  e5k 

ln
1
2 
= 5k

1
k = ln
5
1
2  5
1
=  ln 2  =
1
5
ln 2

Put back into formula

 1
ln  2 t
5
y t  = 800 e More common way to write
t
ln  2
5
y t  = 800 e Re-arrange
t
5
y t  = 800 eln 2 Power Law
t
5
y t  = 8002 Simplify

b. Find the mass after 30 days (Find y(t) when t = 30)

30
5
y 30 = 8002
= 8002 6
1
= 800
64
100 25
= = = 12.5 mg
8 2

c. When is the mass reduced to 1 mg? (Find t when y(t) = 1)

t
5
1 = 8002
t
1 5
=2
800

log2  
1
800
= log2 2 5   t

ln  1
800
=
t
ln 2 5

t=
5 ln  
1
800
 48 days
 ln 2
Leave like this
 Additional Selected Homework Problems
Note: These are scans of hand-work. You may have difficulty viewing them on some web browsers.