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 #23 May 23 2024 09:12:50
%S 14,21,26,28,35,38,39,42,52,56,57,62,63,65,70,74,76,77,78,82,84,86,93,
%T 95,98,99,104,105,111,112,114,117,119,122,124,126,129,130,133,134,140,
%U 143,146,148,152,154,155,156,158,161,166,168,171,172,175,182,183,185,186,189,190,194,195,198,201,203,206,208,209
%N Numbers k for which sigma(k^2) and psi(k^2) share a factor, where sigma is the sum of divisors, A000203, and psi is the Dedekind psi function, A001615.
%C Numbers k for which A344695(k^2) > 1.
%C It can be shown that sigma(m) and psi(m) share a factor if m is nonsquare. (See A344695 for more detail.) So here we consider only square numbers, m = k^2.
%C For prime p, sigma(p^2) and psi(p^2) are coprime, since sigma(p^2) = p^2 + p + 1 = psi(p^2) + 1. So all terms are composite. We can say more, since for prime p and positive integer e, psi(p^(2*e)) = p^(2*e-1) * (p+1), whereas sigma(p^(2*e)) can be shown to be congruent to 1 modulo p and to 1 modulo p+1, so shares no factors with p^(2*e-1) * (p+1). So all terms are divisible by more than one prime.
%C If k is in the sequence, m*k is also present for any positive integer m coprime to k.
%H Robert Israel, <a href="/A344703/b344703.txt">Table of n, a(n) for n = 1..10000</a>
%e Sigma (A000203) and the Dedekind psi function (A001615) are both multiplicative, so we gain insight by showing the values of these functions using their multiplicative properties:-
%e sigma(14^2) = sigma(2^2) * sigma(7^2) = 7 * 57 = 7 * (3*19).
%e psi(14^2) = psi(2^2) * psi(7^2) = 6 * 56 = (2*3) * (2^3*7).
%e So sigma(14^2) and psi(14^2) share factors 3 and 7, so 14 is in the sequence.
%e Looking in particular at the shared factor 3, we see it is present in sigma(7^2) and psi(2^2). For prime p, sigma(p^2) and psi(p^2) equate to polynomials in p, so we deduce 3 divides sigma(p^2) for p congruent to 7 modulo 3, and divides psi(p^2) for p congruent to 2 modulo 3. From this we see all products of a prime of the form 3m+1 and a prime of the form 3m+2 are in the sequence; for instance (3*4+1) * (3*1+2) = 13 * 5 = 65.
%p filter:= proc(k) local n,F, sig, psi, t;
%p n:= k^2;
%p F:= map(t -> [t[1],2*t[2]], ifactors(k)[2]);
%p sig:= mul((t[1]^(1+t[2])-1)/(t[1]-1),t=F);
%p psi:= n*mul(1+1/t[1],t=F);
%p igcd(sig,psi) > 1
%p end proc:
%p select(filter, [$1..300]); # _Robert Israel_, Jan 06 2024
%t filter[k_] := Module[{n, F, sig, psi},
%t n = k^2;
%t F = {#[[1]], 2 #[[2]]}& /@ FactorInteger[k];
%t sig = Product[(t[[1]]^(1 + t[[2]]) - 1)/(t[[1]] - 1), {t, F}];
%t psi = n*Product[1 + 1/t[[1]], {t, F}];
%t GCD[sig, psi] > 1];
%t Select[Range[300], filter] // Quiet (* _Jean-François Alcover_, May 23 2024, after _Robert Israel_ *)
%o (PARI)
%o A001615(n) = if(1==n,n, my(f=factor(n)); prod(i=1, #f~, f[i, 1]^f[i, 2] + f[i, 1]^(f[i, 2]-1))); \\ After code in A001615
%o A344695(n) = gcd(sigma(n), A001615(n));
%o isA344703(n) = (A344695(n^2)>1);
%Y Cf. A000203, A000290, A001615, A344695.
%Y Subsequences: A344872.
%K nonn
%O 1,1
%A _Antti Karttunen_ and _Peter Munn_, May 27 2021