OFFSET
1,1
COMMENTS
All terms are coprime to 2, 3, 5, 7, 11. - Robert Israel, Sep 20 2017
If p = lpf(2^m - 1) and A002326((p-1)/2) = m composite, then m is in this sequence. - Thomas Ordowski, Sep 20 2017
Conjecture: there are no numbers k in this sequence such that, for each prime factor q of 2^k - 1, q == 1 (mod k). - Thomas Ordowski, Sep 20 2017
Note: if all prime factors q of 2^k - 1 are q == 1 (mod k), then 2^k - 1 == 1 (mod k), thus 2^k == 2 (mod k), so k is a pseudoprime. The pseudoprime k = a(42) = 4369 = 17*257 is not a counterexample to this conjecture. A pseudoprime k = P*Q such that both 2^P - 1 and 2^Q - 1 are primes would be a counterexample, but the known Mersenne primes do not give such k. - Thomas Ordowski, Oct 02 2017
If lpf(2^n - 1) == 1 (mod n), then gpf(2^n - 1) == 1 (mod n). Cf. A291855. - Thomas Ordowski, Oct 20 2017
Composites m such that lpf(2^m - 1)*gpf(2^m - 1) is a Fermat pseudoprime to base 2, i.e., is in A214305. - Thomas Ordowski, Oct 29 2017
FORMULA
A049479(m) == 1 (mod m).
MATHEMATICA
searchMax = 1000; Complement[Select[Range[searchMax], Mod[FactorInteger[2^# - 1][[1, 1]], #] == 1 &], Prime[Range[PrimePi[searchMax]]]] (* Alonso del Arte, Sep 19 2017 *)
PROG
(PARI) lista(nn) = forcomposite(n=1, nn, sp = factor(2^n-1)[1, 1]; if ((sp % n) == 1, print1(n, ", "))); \\ Michel Marcus, Sep 19 2017
CROSSREFS
KEYWORD
nonn,hard
AUTHOR
Michel Marcus and Thomas Ordowski, Sep 19 2017
EXTENSIONS
a(10)-a(54) from Charles R Greathouse IV, Sep 19 2017
STATUS
approved