OFFSET
1,1
COMMENTS
Positive integers k such that both 2*k + 1 and 4*k^2 + 1 are composite. Proof: If 2*k + 1 is composite, then for every odd m >= 1, (2*k)^m + 1 is divisible by 2*k + 1 (since for odd t, x^t + 1 is divisible by x + 1). If m is even, write m = 2^r*v with r >= 1 and v odd. Then (2*k)^m + 1 = ((2*k)^(2^r))^v + 1. If v >= 3 (i.e., m not a power of 2), this is composite by the same odd-exponent factorization. If v = 1 (so m = 2^r), then for r = 1 we get (2*k)^2 + 1 = 4*k^2 + 1 (composite by assumption); for r >= 2, we are in the x^2 + 1 case with m >= 4. Thus only powers of 2 with value >= 4 can be exceptions.
If m is a power of 2 with value >= 4, any prime divisor p must satisfy p == 1 (mod 4), since we are in the x^2 + 1 form.
The terms (2*k)^(2^j) + 1 are composite for all j >= 2 if either (a) 2*k is a perfect power with an odd exponent t >= 3, in which case (2*k)^(2^j) + 1 = (b^(2^j))^t + 1 is divisible by b^(2^j) + 1; or (b) 2*k = 2^r with r > 2 and r not a power of 2, so r has an odd factor s >= 3 and (2*k)^(2^j) + 1 = (2^(r/s))^(s*2^j) + 1 is divisible by (2^(r/s))^(2^j) + 1.
For k in this sequence that are not in A391762: For some such k (e.g., 17, 22, 24, ...), a small j >= 2 making (2*k)^(2^j) + 1 prime is easy to find. For others (e.g., 19, 25, 31, ...), finding such a j may be difficult, or no such j may exist.
LINKS
Felix Huber, Table of n, a(n) for n = 1..10000
EXAMPLE
4 is a term, since 2*4 + 1 = 9 = 3^2 and 4*4^2 + 1 = 65 = 5*13 are both composite.
17 is a term, since 2*17 + 1 = 35 = 5*7 and 4*17^2 + 1 = 1157 = 13*89 are both composite. Note that (2*17)^4 + 1 = 1336337 is prime; however, m = 4 is a power of 2 with value >= 4, so a prime at that exponent does not disqualify 17.
MAPLE
CROSSREFS
KEYWORD
nonn,easy
AUTHOR
Felix Huber, Apr 03 2026
STATUS
approved
