6-25-2008

Divisor Functions
Denition. The sum of divisors function is given by (n) =
d|n

d.

The number of divisors function is given by (n) =

d|n

1.

Example. Recall that a number is perfect if its equal to the sum of its divisors other than itself. It follows that a number n is perfect if (n) = 2n.

Example. (15) = 1 + 3 + 5 + 15 = 24 and (15) = 4.

I want to show that and are multiplicative. I can do most of the work in the following lemma. Lemma. The divisor sum of a multiplicative function is multiplicative. Proof. Suppose f is multiplicative, and let D(f ) be the divisor sum of f . Suppose (m, n) = 1. Then [D(f )](m) =
a|m

f (a) and

[D(f )](n) =
b|n

f (b).

Then

Now (m, n) = 1, so if a | m and b | n, then (a, b) = 1. Therefore, multiplicativity of f implies [D(f )](m) [D(f )](n) =
a|m b|n

[D(f )](m) [D(f )](n) =

a|m

f ( a)

b|n

f (b) =

f (a)f (b).
a|m b|n

f (ab).

Now every divisor d of mn can be written as d = ab, where a | m and b | n. Going the other way, if a | m and b | n then ab | mn. So I may set d = ab, where d | mn, and replace the double sum with a single sum: f (ab) = [D(f )](mn). [D(f )](m) [D(f )](n) =
d|mn

This proves that D(f ) is multiplicative.

Example. The identity function id(x) = x is multiplicative: id(mn) = mn = id(m) id(n) for all m, n (so a fortiori for (m, n) = 1). Therefore, the divisor sum of id is multiplicative. But [D(id)](n) =
d|n

id(d) =
d|n

d = (n).

Hence, the sum of divisors function is multiplicative.

Example. The constant function I (n) = 1 is multiplicative: I (mn) = mn = I (m) I (n) for all m, n (so a fortiori for (m, n) = 1). Therefore, the divisor sum of I is multiplicative. But [D(I )](n) =
d|n

I (d) =
d|n

1 = (n).

Hence, the number of divisors function is multiplicative.

Ill use multiplicativity to obtain formulas for (n) and (n) in terms of their prime factorizations (as I did with ). First, Ill get the formulas in the case where n is a power of a prime. Lemma. Let p be prime. Then: (pk ) = pk+1 1 p1

(pk ) = k + 1 Proof. The divisors of pk are 1, p, p2 , . . . , pk . So the sum of the divisors is (pk ) = 1 + p + p2 + + pk = pk+1 1 . p1

And since the divisors of pk are 1, p, p2 , . . . , pk , there are k + 1 of them, and (pk ) = k + 1.
rk 1 Theorem. Let n = pr 1 pk , where the ps are distinct primes and ri 1 for all i. Then: r1 +1 p1 1 p1 1
k +1 pr 1 k pk 1

(n) =

(n) = (r1 + 1) (rk + 1) Proof. These results follow from the preceding lemma, the fact that and are multiplicative, and the fact i that the prime power factors pr i are pairwise relatively prime. Here is a graph of (n) for 1 n 1000.
3000 2500 2000 1500 1000 500 200 400 600 800 1000

Note that if p is prime, p = p + 1. This gives the point (p, p + 1), which lies on the line y = x + 1. This is the line that you see bounding the dots below. Here is a graph of (n) for 1 n 1000.
30 25 20 15 10 5 200 400 600 800 1000

If p is prime, (p) = 2. Thus, repeatedly returns to the horizontal line y = 2, which you can see bounding the dots below.

Example. 720 = 24 32 5, so (720) = 25 1 21 33 1 31 52 1 51 = 2418,

(720) = (4 + 1)(2 + 1)(1 + 1) = 30.

Example. For each n, there are only nitely many numbers k whose divisors sum to n: (k) = n. For k divides itself, so n = (k) = (junk) + k > k. This says that k must be less than n. So if Im looking for numbers whose divisors sum to n, I only need to look at numbers less than n. For example, if I want to nd all numbers whose divisors sum to 42, I only need to look at {1, 2, . . . , 41}.

c 2008 by Bruce Ikenaga