For a prime numberp1,⋯,pk, let’s express a natural numbern as follows. The arithmetic functionμ defined as such is called the Möbius function.
μ(n):=⎩⎨⎧1(−1)k0,n=1,a1=⋯=ak=1,otherwise
Basic Properties
[1] Möbius series: It’s an identityI. In other words,
d∣n∑μ(d)=I(n)
[2] Multiplicativity: For all m,n∈N that satisfy gcd(m,n)=1, μ(mn)=μ(m)μ(n)
Description
nμ(n)∑d∣nμ(d)1112−103−104005−106107−108009001010
The Möbius function, simply put, is a function that only cares about numbers that are not multiplied by the same prime number 2 or more times. If a prime number is not multiplied more than 2 times, its sign only changes depending on whether the prime numbers are used an even or odd number of times. However, this is just an explanation of the definition. The Möbius function itself may not have an intuitive meaning, but it is a major function that appears endlessly throughout number theory, especially in analytic number theory.
Proof
[1]
Let’s say n=p1a1⋯pkak>1, then according to the binomial theorem,
====d∣n∑μ(d)μ(1)+μ(p1)+⋯+μ(pk)+μ(p1p2)+⋯+μ(pk−1pk)⋮+μ(p1p2⋯pk−1pk)1+k(−1)+⋯+(−1)+k(k−1)/21+⋯+1⋮+(−1)k1+(1k)(−1)+(2k)(−1)2+⋯+(kk)(−1)k[1+(−1)]k
■
[2]
If either m or n has a prime square as a divisor, then from the definition of μ, μ(mn)=0 and therefore μ(m)μ(n)=0, then
μ(mn)=μ(m)μ(n)
Hence, let’s assume m and n are in the form where each prime is multiplied only once.
m=p1⋯psn=q1⋯qt
Then,
μ(mn)=(−1)s+t=(−1)s(−1)t=μ(m)μ(n)
■
Apostol. (1976). Introduction to Analytic Number Theory: p24. ↩︎