login

Year-end appeal: Please make a donation to the OEIS Foundation to support ongoing development and maintenance of the OEIS. We are now in our 61st year, we have over 378,000 sequences, and we’ve reached 11,000 citations (which often say “discovered thanks to the OEIS”).

A036236
Least inverse of A015910: smallest integer k > 0 such that 2^k mod k = n, or 0 if no such k exists.
68
1, 0, 3, 4700063497, 6, 19147, 10669, 25, 9, 2228071, 18, 262279, 3763, 95, 1010, 481, 20, 45, 35, 2873, 2951, 3175999, 42, 555, 50, 95921, 27, 174934013, 36, 777, 49, 140039, 56, 2463240427, 110, 477, 697, 91, 578, 623, 156, 2453, 540923, 55, 70, 345119, 287
OFFSET
0,3
COMMENTS
a(1) = 0, that is, no n exists with 2^n mod n = 1. Proof: Assume that there exists such n > 1. Consider its smallest prime divisor p. Then 2^n == 1 (mod p) implying that the multiplicative order ord_p(2) divides n. However, since ord_p(2) < p and p is the smallest divisor of n, we have ord_p(2) = 1, that is, p divides 2^1 - 1 = 1 which is impossible. - Max Alekseyev
Labos Elemer asked on Sep 27 2001 if all numbers > 1 eventually appear in A015910, that is, if a(n) > 0 for n > 1.
Ron Graham's conjecture from 1960 states that for any n > 1 there are infinitely many solutions to 2^k mod k = n. - Max Alekseyev, Oct 19 2024
Obviously k > n. - Daniel Forgues, Jul 06 2015
REFERENCES
P. Erdős and R. L. Graham, Old and new problems and results in combinatorial number theory, Monographies de L'Enseignement Mathematique, 28, 1980.
R. K. Guy, Unsolved Problems in Number Theory, Section F10.
LINKS
David W. Wilson, Table of n, a(n) for n = 0..1026 (from the Havermann file)
Joe K. Crump, 2^n mod n
Nick Hobson, Table of n, a(n) for n = 1..10000 with -1 for those entries where a(n) is unknown. [With 101 new terms and 2 corrections, this supersedes the earlier table from Hans Havermann. The corrections are to terms a(799) and a(881), where smaller values of k were found.]
Nick Hobson, C program
Eric Weisstein's World of Mathematics, 2
FORMULA
It's obvious that for each k, a(k) > k and we can easily prove that 2^(3^n) = 3^n-1 (mod 3^n). So 3^n is the least k with 2^k mod k = 3^n-1. Hence for each n, a(3^n-1) = 3^n. - Farideh Firoozbakht, Nov 14 2006
EXAMPLE
n = 0: 2^1 mod 1 = 0, a(0) = 1;
n = 1: 2^k mod k = 1, no such k exists, so a(1) = 0;
n = 2: 2^3 mod 3 = 2, a(2) = 3;
n = 3: 2^4700063497 mod 4700063497 = 3, a(3) = 4700063497.
MATHEMATICA
a = Table[0, {75} ]; Do[ b = PowerMod[2, n, n]; If[b < 76 && a[[b]] == 0, a[[b]] = n], {n, 1, 5*10^9} ]; a
(* Second program: *)
t = Table[0, {1000} ]; k = 1; While[ k < 6500000000, b = PowerMod[2, k, k]; If[b < 1001 && t[[b]] == 0, t[[b]] = k]; k++ ]; t
nk[n_] := Module[ {k}, k = 1; While[PowerMod[2, k, k] != n, k++]; k]
Join[{1, 0}, Table[nk[i], {i, 2, 46}]] (* Robert Price, Oct 11 2018 *)
PROG
(PARI) a(n)=if(n==1, return(0)); my(k=n); while(lift(Mod(2, k)^k)!=n, k++); k \\ Charles R Greathouse IV, Oct 12 2011
KEYWORD
nonn,nice
EXTENSIONS
a(3) was first computed by the Lehmers.
More terms from Joe K. Crump (joecr(AT)carolina.rr.com), Sep 04 2000
a(69) = 887817490061261 = 29 * 37 * 12967 * 63809371. - Hagen von Eitzen, Jul 26 2009
Edited by Max Alekseyev, Jul 29 2011
STATUS
approved