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”).
%I #14 May 25 2021 08:07:20
%S 1,3,597,609,1791,2035,3403,3701,4263,27515,27955
%N Odd numbers k such that Sum_{j=1..(k-1)/2, gcd(j,k)=1} 1/j == -2*q_2(k) + k*q_2(k)^2 (mod k^3), where q_2(k) = (2^phi(k) - 1)/k is the Euler quotient of k to base 2.
%C Emma Lehmer proved that Sum_{j=1..(p-1)/2} 1/j == -2*q_2(p) + p*q_2(p)^2 (mod p^2) for all odd primes p.
%C Tianxin Cai generalized Lehmer's congruence and proved that Sum_{j=1..(k-1)/2, gcd(j,k)=1} 1/j == -2*q_2(k) + k*q_2(k)^2 (mod k^2) for all odd numbers k.
%C This sequence includes the odd numbers k for which the congruence is still valid when (mod k^2) is being replaced with (mod k^3).
%C The prime terms are 3, 3701, ...
%C No more terms below 147000.
%H Tianxin Cai, <a href="https://doi.org/10.4064/aa103-4-1">A congruence involving the quotients of Euler and its applications (I)</a>, Acta Arithmetica, Vol. 103, No. 4 (2002), pp. 313-320.
%H Tianxin Cai, <a href="https://doi.org/10.1007/978-1-4757-3621-2_5">A Generalization of E. Lehmer's Congruence and Its Applications</a>, in: ChaohuaJia and Kohji Matsumoto (eds.), Analytic Number Theory, Springer, Boston, MA, 2002, pp. 93-98.
%H Emma Lehmer, <a href="https://doi.org/10.2307/1968791">On congruences involving Bernoulli numbers and the quotients of Fermat and Wilson</a>, Annals of Mathematics, Second Series, Vol. 39, No. 2 (1938), pp. 350-360, <a href="https://web.archive.org/web/20190509083052/http://gradelle.educanet2.ch/christian.aebi/.ws_gen/14/Emma_Lehmer_1938.pdf">alternative link</a>.
%t q[n_] := (2^EulerPhi[n] - 1)/n; Select[Range[1, 2100, 2], Divisible[Numerator[Sum[Boole @ CoprimeQ[j, #]/j, {j, 1, (# - 1)/2}] + 2*q[#] - #*q[#]^2], #^3] &]
%Y Cf. A000010, A001226.
%K nonn,more
%O 1,2
%A _Amiram Eldar_, Feb 28 2020