Maths Olympiad Prep

Library / /15 of 22

Number theory Difficulty 5.9 AIME, harder Prove it Romania

Let aa, bb, cc, nn be four integers, where n2n \ge 2, and let pp be a prime dividing both a2+ab+b2a^2 + ab + b^2 and an+bn+cna^n + b^n + c^n, but not a+b+ca + b + c; for instance, ab1(mod3)a \equiv b \equiv -1 \pmod{3}, c1(mod3)c \equiv 1 \pmod{3}, nn a positive even integer, and p=3p = 3 or a=4a = 4, b=7b = 7, c=13c = -13, n=5n = 5, and p=31p = 31 satisfy these conditions. Show that nn and p1p-1 are not coprime.

Solution

Throughout the proof congruences are taken modulo pp. Begin by ruling out the case p=2p = 2. If p=2p = 2, then a2+ab+b2a^2 + ab + b^2 and an+bn+cna^n + b^n + c^n are both even, and a+b+ca + b + c is odd. The first condition forces both aa and bb even, so cc is also even by the second, contradicting the third. Consequently, pp must be odd, and the conclusion follows unless nn is odd.

Henceforth assume nn odd. It is easily seen from the conditions in the statement that a≢0a \not\equiv 0, so a1a^{-1} exists modulo pp, and the hypotheses yield B2+B+10B^2 + B + 1 \equiv 0, Bn+Cn+10B^n + C^n + 1 \equiv 0 and B+C+1≢0B + C + 1 \not\equiv 0, where B=a1bB = a^{-1}b and C=a1cC = a^{-1}c. The first congruence yields B31B^3 \equiv 1, where B≢1B \not\equiv 1. (Otherwise, 303 \equiv 0, so p=3p = 3, Cn1C^n \equiv 1 and C≢1C \not\equiv 1 which is impossible since nn is odd.) Hence 33 is a factor of p1p-1; in particular, p7p \ge 7 and p1p-1 is divisible by 66, so the conclusion follows unless n=6m±1n = 6m \pm 1.

Let n=6m±1n = 6m \pm 1 and recall that B31B^3 \equiv 1 to deduce that B2n+Bn+1B±2+B±1+10B^{2n} + B^n + 1 \equiv B^{\pm 2} + B^{\pm 1} + 1 \equiv 0, so CnB2n(B1)nC^n \equiv B^{2n} \equiv (-B-1)^n; that is, (C(B+1)1)n1(-C(B+1)^{-1})^n \equiv 1, since B+1≢0B+1 \not\equiv 0. The condition B+C+1≢0B + C + 1 \not\equiv 0 shows that C(B+1)1≢1-C(B+1)^{-1} \not\equiv 1, so the multiplicative order of C(B+1)1-C(B+1)^{-1} in Zp\mathbb{Z}_p^* is a divisor dd of nn, greater than 11. Since dd is also a divisor of p1p-1, the conclusion follows.

Want a route through all this instead of an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement reproduced verbatim; metadata (topic, difficulty) added by this project.