The standard proofs of the irrationality of e make use of the infinite series
representation
or the corresponding alternating series representation for 1/e. (One such
proof is given at the end of this
article.) While these proofs are elementary, they obviously require some
familiarity with infinite series. The
following proof requires only integration-by-parts and some basic properties of
the Riemann integral. The
sum (1) follows as a consequence, thereby making this proof useful as an
introduction to infinite series.
e is irrational.
Proof: Suppose e = a/b, where a and b are positive integers . Choose an
integer n≥max{b, e}. Now
consider the definite integral
This integral is easily evaluated to give 1 − 1/e . On the other hand, repeated
integration-by-parts ( n times ) gives
Upon multiplying both sides by e and isolating the integral, we obtain
Multiplying both sides of (2) by n! gives
Because of the choice of n and the assumption that e is rational , the left
hand side must reduce to an integer .
However the value of the expression on the right is between zero and one. Indeed
This contradiction implies that e must be irrational.
Notice that the integral in (2) approaches zero as
Therefore we obtain (1) as a by- product of the
proof . The series representation (1) was derived in a similar way by Chamberland
in [1] and by Johnson in
[2].
A proof using the series for 1/e ...
Use the fact that
and let Sn denote the nth partial sum of the series:
Notice that Sn is a rational number, and it can be written in the
form M/n!, where M is an integer. By the
alternating series estimation theorem, it follows that
In either case, e−1 is strictly between two
rational numbers of the forms , where a is an
integer. It follows that e−1 cannot be written as a fraction with denominator (n
+ 1)! for any n≥0. Since
any rational number can be written as a fraction with denominator (n+1)!, we
conclude that e−1 cannot be
a rational number . Since 1/e is irrational, it follows that e is irrational.
(This proof is similar to Sondow’s
geometric proof [3].)