Does this prime sum converge?

It is well-known that n = 1 1 n = 1 + 1 2 + 1 3 + 1 4 + . . . \displaystyle \sum_{n=1}^{\infty} \frac{1}{n} = 1 + \frac{1}{2} + \frac{1}{3} + \frac{1}{4} + ... diverges. But how about the sum below?

p prime 1 p = 1 2 + 1 3 + 1 5 + 1 7 + . . . \displaystyle \sum_{p \text{ prime}} \frac{1}{p} = \frac{1}{2} + \frac{1}{3} + \frac{1}{5} + \frac{1}{7} + ...

It converges It diverges

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Mark Hennings
Aug 13, 2019

List the primes in ascending order p 1 , p 2 , p 3 , . . . p_1,p_2,p_3,... . Then any integer positive integer N 2 N \ge 2 can certainly be factorised in terms of the first N N (at most) primes, and hence n = 1 N 1 n j = 1 N ( 1 1 p j ) 1 \sum_{n=1}^N \frac{1}{n} \; \le \; \prod_{j=1}^N \left(1 - \tfrac{1}{p_j}\right)^{-1} and hence ln ( n = 1 N 1 n ) j = 1 N ln ( 1 1 p j ) 1 = j = 1 N ln ( 1 1 p j ) = j = 1 N k = 1 1 k p j k = j = 1 N 1 p j + j = 1 N k = 2 1 k p j k j = 1 N 1 p j + j = 1 N k = 2 1 p j 2 × 1 k 2 k 2 j = 1 N 1 p j + n = 1 1 n 2 × k = 2 1 2 k 1 = j = 1 N 1 p j + 1 6 π 2 \begin{aligned} \ln\left( \sum_{n=1}^N \frac{1}{n} \right) & \le \; \sum_{j=1}^N \ln\left(1 - \frac{1}{p_j}\right)^{-1} \; = \; -\sum_{j=1}^N \ln\left(1 - \frac{1}{p_j}\right) \; = \; \sum_{j=1}^N \sum_{k=1}^\infty \frac{1}{kp_j^k} \; = \; \sum_{j=1}^N \frac{1}{p_j} + \sum_{j=1}^N \sum_{k=2}^\infty \frac{1}{kp_j^k} \\ & \le \; \sum_{j=1}^N \frac{1}{p_j} + \sum_{j=1}^N \sum_{k=2}^\infty \frac{1}{p_j^2} \times \frac{1}{k2^{k-2}} \; \le \; \sum_{j=1}^N \frac{1}{p_j} + \sum_{n=1}^\infty \frac{1}{n^2} \times \sum_{k=2}^\infty \frac{1}{2^{k-1}} \; = \; \sum_{j=1}^N \frac{1}{p_j} + \tfrac16\pi^2 \end{aligned} Since n n 1 \sum_n n^{-1} diverges, we deduce that p p 1 \sum_p p^{-1} diverges.

Where can I learn about summation like this and more number theory? Is brilliant.org good for number theory or is it not deep enough for these type of questions?

Raghu Alluri - 1 year, 7 months ago

I saw a theorem early in a number theory book that was similar to this but using the inequality e^(x^2 + x) > 1/(1-x) for 0<=x<=1/2, was able to show that the sum of 1/p for all p <=N is greater than ln(ln(N))-π^2/6.

Casey Appleton - 1 year, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...