Sie sind auf Seite 1von 2

A recursion equation for prime numbers

Joseph B. Keller
Department of Mathematics
Stanford University
Stanford, CA 94305
arXiv:0711.3940v2 [math.NT] 5 Oct 2008

Email: keller@math.stanford.edu
June 24, 2007

Abstract
It is shown that the first n prime numbers p1 , . . . , pn determine the next one by the recursion equation
" n  ∞
 X #−1/s
Y 1 1
pn+1 = lim 1− s −1 .
s→+∞ pk js
k=1 j=1

The upper limit on the sum can be replaced by 2pn − 1, and the result still holds.

The first n prime numbers p1 , p2 , · · · , pn determine the next one by the recursion equation:
 −1/s
n  ∞
 X
Y 1 1
pn+1 = lim  1− s − 1 . (1)
s→+∞ pk j=1 j s
k=1

To prove (1) we replace the sum in it, which is the zeta function ζ(s), by the Euler product[1], to get

n   n   ∞ !−1 ∞
!−1
Y 1 Y 1 Y 1 Y 1
1 − s ζ(s) = 1− s 1− s = 1− s . (2)
pk pk j=1 pj j=n+1
pj
k=1 k=1

As s → +∞, the final product in (2) is asymptotic to 1 + p−sn+1 . When 1 is subtracted from it, the difference is
asymptotic to p−s
n+1 . Taking the −1/s power of this yields p n+1 , which proves (1).
We can obtain another recursion equation by first factoring the product out of the bracketed expression in
(1). Then as s → ∞, the product tends to 1, and so does the −1/s power of it. Thus we obtain the recursion
equation
−1 −1/s
 
∞ n 
X 1 Y 1
pn+1 = lim  s
− 1− s  . (3)
s→+∞
j=1
j p kk=1

Since pn+1 < 2pn , (3) remains valid when the sum is terminated at j = 2pn − 1, so it becomes a finite sum:

−1 −1/s
 
2pn −1 n 
X 1 Y 1
pn+1 = lim  − 1− s  . (4)
s→∞
j=1
js pk
k=1

To show that (1) also holds with the sum terminated at 2pn − 1, we just factor the product out of the
bracketed expression in (4). The −1/s power of the product tends to 1 as s tends to +∞, so the result follows.

1
References
[1] Edwards, Harold. Riemann’s Zeta Function, Academic Press, New York, 1974.