Scribd
Upload
118 views

On Euler'S Number E: Avery I. Mcintosh

1) The document defines the number e as the unique number whose derivative is equal to itself multiplied by the original function. This makes e the "natural" base for exponential functions. 2) It is shown that taking the derivative of any exponential function ax results in the original function multiplied by the logarithm of the base a. 3) e is further defined as the limit of the compound interest formula as the number of compounding periods approaches infinity, which evaluates to approximately 2.718, known as e.

Uploaded by

Frédéric Bambus Moresmau
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
118 views

On Euler'S Number E: Avery I. Mcintosh

1) The document defines the number e as the unique number whose derivative is equal to itself multiplied by the original function. This makes e the "natural" base for exponential functions. 2) It is shown that taking the derivative of any exponential function ax results in the original function multiplied by the logarithm of the base a. 3) e is further defined as the limit of the compound interest formula as the number of compounding periods approaches infinity, which evaluates to approximately 2.718, known as e.

Uploaded by

Frédéric Bambus Moresmau
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

On Eulers Number e

Avery I. McIntosh
[email protected]
The number e, an irrational number whose first digits are 2.7182818284...,
is usually presented to students in precalculus classes as the number toward
which compounded interest approaches as the number of compounding intervals approaches infinity and the interval size approaches 0. This representation has the benefit of being navigable for the student who does not know
any calculus, but as students mathematical knowledge increases, the usual
definition begins to seem awkward, and even arbitrary: what does compound
interest have to do with derivatives and integrals?
This short monograph details a method for deriving e that is both succinct
and comprehensive. I first saw it presented in an MIT lecture by David Jerison, and I believe it is the best method for making this elusive number clear
in the context of single variable calculus.

Suppose we want to know the derivative of some exponential function ax ,


a R+ . Well, by definition the derivative is
ax+x ax
d x
a = lim
x0
dx
x
We are used to thinking of x as the moving variable, but here it is x that
is changing. We can think of x as fixed. Thus, the above derivative simplifies
by the laws of exponents:
n
x
d x
1
ax 1 o
xa
x
a = lim a
=a
lim
x0
x0
dx
x
x

This means that the derivative of ax is proportional to itself, by some constant: the quantity inside the limit. Call that quantity M(a), a function of
the base of the exponential function. Note that this quantity is not dependent
on x. Then we have that
d x
a = ax M (a)
dx
Lets pretend we have never heard of the number e. Define e as that number
such that M(e) = 1. Therefore we have that, by the definition of this new e
number,
d x
d x
e = ex M (e) = ex ,
e |x=0 = 1
dx
dx
All of that is well and good, but the original task, that of finding the derivative of ax , is still not solved. To accomplish this, I introduce the logarithm
with base e. Define w = ln(x). So if y = ex , then ln(y) = x. This function
obeys all the usual laws of a logarithm of any base, i.e. ln(1) = 0, ln(x) +
ln(y) = ln(xy), and so on. Now I solve the derivative of this function using
implicit differentiation.
w := ln(x) ew = eln(x) = x,
and from the chain rule:
d
d w
e
x
dx
dx

d w
dw
e = ew
=1
dx
dx

This last quantity implies that


dw
1
= w
dx
e
But what did we define ew to be? Its x. Thus, the derivative of ln(x) is 1/x.
Now we have the machinery to work out the derivative of ax .
d x
d ln(a) x
d x ln(a)
a =
(e
) =
e
dx
dx
dx
2

and again from the chain rule, the result is


d x ln(a)
e
= ln(a) ex ln(a) = ln(a) ax
dx
For some given a, say a = 3, the derivative of 3x is ln(3) times 3x . This is
one of the reasons that the exponential function with base e is called the
natural exponential function, and it is why the logarithm base e is called
the natural logarithm. It naturally appears in the formula for the derivative
of an exponential, and if we choose the base e for our calculations, we never
need actually compute one of these clunky logarithm functions.

Alright. The correct, natural base has been derived. But what is this
number actually? To find that out, I invoke the common approximation to
the exponential base (which is in fact inferior in almost every way to the
power series definition, but thats another story). This formula is usually
introduced in the context of compounding interest:
lim 1 +

1 n
n

Now take the logarithm of this function, which is equal by the properties
of logarithms to n ln(1 + 1/n). Now define x = 1/n. Note that as n
grows large, x will tend toward 0. Then we can rewrite the logarithm of
the compounding interest formula as
1
ln(1 + x)
x0 x
lim

Now subtract 0 = ln(1) from the above quantity to yield


1
(ln(1 + x) ln(1))
x0 x
lim

But look what this is equal to:


1
d
(ln(1 + x) ln(1)) =
ln(x)|x=1
x0 x
dx
lim

But we know what this derivative is. Its 1/x. So the limit is 1. Now
recall that this entire formula was the logarithm of the compound interest
formula. To recover its actual value, we need to perform the inverse of the
logarithm base e, which is the natural exponentiation:
n
n o
1

limn ln 1+ n
1 n
=e
= e1 = e
lim 1 +
n
n
Now we know roughly what e is. If I take n = 1000, I get that e 2.71.

To recap, we decided to define e as that number such that the derivative


of that number raised to some power is identical to the original number to
some power. In more formal language, the derivative is idempotent and the
function is of differentiability class C , infinitely differentiable, or smooth.
With that definition in hand, we found that any derivative of an exponential
base to a power is just that function times a constant multiple involving the
logarithm. Thus, e is the natural base to use in applications, as it eliminates
the need to calculate a logarithm every time a derivative is taken.
Finally, I showed that this number e is between 2 and 3, and that it also
happens to be the limit of compounding interest as the length of compounding intervals tends toward 0.

The figure below shows the relatively quick convergence of the sequence formula for e. Notice that after only 35 iterations the error is less than one
tenth: about 0.03.

You might also like