OFFSET
1,2
COMMENTS
Whether binomial(a+b*i, n) is a Gaussian integer appears to be periodic for fixed integer b, n (which means that it suffices to check up to a certain integer a before it begins cycling). It appears that all terms in the sequence are products of only ramified primes (2) and inert primes (3, 7, and other primes that are equivalent to 3 modulo 4). It would be great to find a generalized, explicit formula for any n.
LINKS
Eric Weisstein's World of Mathematics, Gaussian Integer.
Rick Zhou, Even and Odd Gaussian Integer Lines
EXAMPLE
For n = 1, b = 1 suffices, as binomial(0 + i, 1) = i is a Gaussian integer.
For n = 2, binomial(0 + 2*i, 2) = -2-i is a Gaussian integer.
For n = 3, the minimal b is 3, as binomial(1 + 3*i, 3) = -5*i is a Gaussian integer and no binomial coefficients with smaller b work.
PROG
(PARI) isok(n, b) = for (a=0, b, my(z=binomial(a+I*b, n)); if ((denominator(real(z))==1) && (denominator(imag(z))==1), return(1)));
a(n) = my(b=1); while (!isok(n, b), b++); b; \\ Michel Marcus, Dec 30 2025
CROSSREFS
KEYWORD
nonn,more
AUTHOR
Rick Zhou, Dec 28 2025
EXTENSIONS
a(9)-a(18) from Michel Marcus, Dec 29 2025
a(19) from Michel Marcus, Dec 30 2025
a(20)-a(23) from Michel Marcus, Dec 31 2025
STATUS
approved
