2232.11 – From a Progression to a Quadratic


Consider the following list of integers:

P0 =41, P1 =43, P2 =47, P3 =53,P_0 = 41, P_1 = 43, P_2 = 47, P_3 = 53, \ldots

in which PnP_n is obtained by adding 2n2n to Pn1P_{n-1}. It so happens that there is a quadratic function F(x)F(x) with the property that Pn = F(n)P_n = F(n) for all non-negative nn. Find FF.

Once you have FF, consider this question: Is F(n)F(n) a prime number for every non-negative value of nn?


Solution

The trick is to rewrite each integer in terms of the first value (41) and look for a pattern that resembles a quadratic function:

P0 =41=41+0=41+0P1 =43=41+2=41+12 +1P2 =47=41+6=41+22 +2P3 =53=41+12=41+32 +3P4 =61=41+20=41+42 +4\begin{aligned}P_0 &= 41 &&= 41 + 0 &&= 41 + 0 \\P_1 &= 43 &&= 41 + 2 &&= 41 + 1^2 + 1 \\P_2 &= 47 &&= 41 + 6 &&= 41 + 2^2 + 2 \\P_3 &= 53 &&= 41 + 12 &&= 41 + 3^2 + 3 \\P_4 &= 61 &&= 41 + 20 &&= 41 + 4^2 + 4\end{aligned}

Thus,

Pn=41+n2+n=n2+n+41P_n = 41 + n^2 + n = n^2 + n + 41

Although the instances we're seeing are all prime, when we reach n =41n = 41,

P41 =412 +41+41=41(41+2)=4143P_{41} = 41^2 + 41 + 41 = 41(41 + 2) = 41 \cdot 43

the result is not prime.

This sequence is famous because for a sequence with a simple formula it generates primes for a long time. It was Euler (in 1772) who first noticed it.