|
%I #12 Sep 28 2019 03:23:29
%S 15,35,95,119,143,209,287,319,323,357,377,527,559,779,899,923,989,
%T 1007,1045,1189,1199,1343,1349,1763,1919,2159,2261,2507,2639,2759,
%U 2911,3239,3339,3553,3599,3827,4031,4147,4607,5049,5183,5207,5249,5459,5543,6439
%N Composite n such that sigma(n)-phi(n) divides sigma(n)+phi(n).
%C Primes trivially satisfy the defining condition.
%H Amiram Eldar, <a href="/A061367/b061367.txt">Table of n, a(n) for n = 1..10000</a>
%F It seems that a(n) is asymptotic to c*n^2, 2<c<2.5 and that a(n)>2*n^2. - _Benoit Cloitre_, Sep 17 2002
%e sigma(15)-phi(15) = 24-8 = 16 divides sigma(15)-phi(15)=24+8 = 32, so 15 is a term of the sequence.
%t f[n_] := Module[{a = DivisorSigma[1, n], b = EulerPhi[n]}, Mod[(a + b), (a - b)] == 0]; Select[Range[2, 10^4], (f[ # ] && ! PrimeQ[ # ]) &]
%Y Cf. A051612, A065387.
%K nonn
%O 1,1
%A _Joseph L. Pe_, Feb 13 2002
|
