Number theoryDifficulty 6.0National olympiadProve itSouth Africa
Suppose that a is an integer, and that n!+a divides (2n)! for infinitely many positive integers n. Prove that a=0.
Solutions — 2
Solution 1
(2n)!=(n2n)⋅n!2≡(n2n)⋅(−a)2mod(n!+a), so if n!+a divides (2n)!, then it also divides a2(n2n). We will show that when n is large, n!+a is greater than a2(n2n) and therefore does not divide it (unless a=0). Assume in the following that a=0 and n>1245a2. First of all, by the binomial theorem, (n2n)≤k=0∑2n(k2n)=22n=4n. If a>0, we have 0<n!+aa2(n2n)<n!a2(n2n)≤n!a24n=(5⋅6⋯(n−1)4n−5)24n45a2<1. If a<0, then we observe that n!>n>2∣a∣, so that 0<n!+aa2(n2n)=n!−∣a∣a2(n2n)<21n!a2(n2n)≤n!2a24n=(5⋅6⋯(n−1)4n−5)12n45a2<1.
Thus n!+aa2(n2n) is not an integer for all such n, which contradicts our assumption that n!+a divides (2n)! (and thus a2(n2n)) for infinitely many n. So a=0 is the only possibility (and indeed n! divides (2n)! for all n).
Solution 2
We show that n!+a cannot divide (2n)! if a=0, n≥2∣a∣ and n≥9. In this case, an! is an integer. If n!+a divides (2n)!, then so does an!+1. Next note that every prime p≤n divides an!: if p∤a, this is obvious (since p∤n!), and if p∣a, then we see that p divides 2∣a∣, which occurs as a factor of an!. In either case, we find that an!+1 is not divisible by p. So an!+1 does not have any prime factors ≤n, which means that it is coprime to all numbers ≤n as well as all the even numbers ≤2n. Therefore an!+1 must divide the product of all odd numbers greater than n and less than 2n. By our assumption, an!+1≥an⋅(n−1)!−1≥2(n−1)!−1. We show that the product of the odd numbers between n and 2n is less than 2(n−1)!−1 for n≥9, which yields the desired contradiction. For n=9 and n=10, we have the two inequalities 2⋅8!−1=80639>36465=11⋅13⋅15⋅17 and 2⋅9!−1=725759>692835=11⋅13⋅15⋅17⋅19 respectively. Now we proceed by induction. Assume that 2(n−1)!−1>n<k<2nk odd∏k. If n>10 is even, we use the induction hypothesis to obtain n<k<2nk odd∏k=n−1(2n−3)(2n−1)n−2<k<2n−4k odd∏k<2(2n−1)n−2<k<2n−4k odd∏k<(4n−2)(2(n−3)!−1)<(n−1)(n−2)(2(n−3)!−1)=2(n−1)!−(n−1)(n−2)<2(n−1)!−1, and if n>10 is odd, we get analogously n<k<2nk odd∏k=n(2n−3)(2n−1)n−2<k<2n−4k odd∏k<2(2n−3)n−2<k<2n−4k odd∏k<(4n−6)(2(n−3)!−1)<(n−1)(n−2)(2(n−3)!−1)=2(n−1)!−(n−1)(n−2)<2(n−1)!−1.
This completes the induction and thus the proof.
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 and solution reproduced as published; topic and difficulty added by this site.