|
%I #33 Jan 11 2020 04:21:31
%S 30,2046,245760,301056,450560,1171456,1351680,3514368,14515200,
%T 16760832,19611648,77220864,159373824,357291648,391444480,477216768,
%U 555714432,754928640,765414240,1006602240,1761500160,2330913312,4314834944,8369053056,20449394784,37949317120
%N Numbers k such that the denominator of sigma(sigma(k))/k is equal to 2.
%C Although the definition here is similar to the one in A019278, it appears that this sequence does not have the same nice features as A019278.
%C Otherwise said: sigma(sigma(k))/k is half-integer, or: sigma(sigma(k)) is an odd multiple of k/2. This also implies that all terms are even. - _M. F. Hasler_, Jan 06 2020
%H Giovanni Resta, <a href="/A330598/b330598.txt">Table of n, a(n) for n = 1..38</a> (terms < 10^13)
%H Michel Marcus, <a href="/A330598/a330598_4.txt">Unexhaustive list of terms</a>
%e sigma(sigma(30))/30 = sigma(72)/30 = 195/30 = 13/2 so 30 is a term.
%o (PARI) isok(n) = denominator(sigma(sigma(n))/n) == 2;
%Y Cf. A019278 (denominator is 1), A051027 (sigma(sigma)).
%Y Cf. A000203 (sigma), A159907 (hemiperfect numbers).
%K nonn
%O 1,1
%A _Michel Marcus_, Dec 19 2019
%E a(22)-a(26) from _Giovanni Resta_, Dec 20 2019
| 