Maths Olympiad Prep

Track / Stage 3 / 48 of 260 #48 of 1964

Problem 48

AMC 10/12, early questions
Number theory Difficulty 3.2 Find the answer fermat

For how many odd integers kk between 0 and 100 does the equation 24m2+2m2n2+4=2k+4+23m2+n2+k2^{4m^{2}}+2^{m^{2}-n^{2}+4}=2^{k+4}+2^{3m^{2}+n^{2}+k} have exactly two pairs of positive integers (m,n)(m, n) that are solutions?

A number or a short expression. Fractions can be typed as 3/2, and spacing doesn't matter.

Official solution

Step 1: Using parity and properties of powers of 2 to simplify the equation. We note that if 2x=2y2^{x}=2^{y} for some real numbers xx and yy, then x=yx=y. We examine equations of the form 2a+2b=2c+2d2^{a}+2^{b}=2^{c}+2^{d} where a,b,ca, b, c, and dd are integers. We may assume without loss of generality that aba \leq b and cdc \leq d and aca \leq c. We factor the equation as 2a(1+2ba)=2c(1+2dc)2^{a}(1+2^{b-a})=2^{c}(1+2^{d-c}), and then divide both sides by 2a2^{a} to obtain 1+2ba=2ca(1+2dc)1+2^{b-a}=2^{c-a}(1+2^{d-c}). We show that c=ac=a by contradiction: If cac \neq a, then cac \geq a gives c>ac>a. If c>ac>a, then ca>0c-a>0, so ca1c-a \geq 1, since cac-a is an integer. Therefore, the right side has a factor of 2ca2^{c-a}, so the right side is even. Thus, the left side is even too, which means that 2ba2^{b-a} must be an odd integer. For 2ba2^{b-a} to be an odd integer, we must have 2ba=12^{b-a}=1 and so ba=0b-a=0 or b=ab=a. In this case, the left side equals 2 and the right side is greater than 2, since 2ca22^{c-a} \geq 2 and 1+2dc>11+2^{d-c}>1. This is a contradiction. Therefore, c=ac=a. Since a=ca=c, then 2a+2b=2c+2d2^{a}+2^{b}=2^{c}+2^{d} becomes 2b=2d2^{b}=2^{d} and so b=db=d. Therefore, if 2a+2b=2c+2d2^{a}+2^{b}=2^{c}+2^{d} with a,b,c,da, b, c, d integers, then either a=b=c=da=b=c=d or a=ca=c and b=db=d (with aba \neq b) or a=da=d and b=cb=c (with ab)a \neq b). We examine these three possibilities in the given equation, noting that m,nm, n and kk are all positive integers: Case 1: 4m2=m2n2+4=k+4=3m2+n2+k4m^{2}=m^{2}-n^{2}+4=k+4=3m^{2}+n^{2}+k. From the last equality, we obtain 3m2+n2=43m^{2}+n^{2}=4. Since m,nm, n are positive integers, then m21m^{2} \geq 1 and n21n^{2} \geq 1. Since 3m2+n2=43m^{2}+n^{2}=4, then it must be that m=n=1m=n=1. Thus, 4m2=k+44m^{2}=k+4 implies 4=k+44=k+4 or k=0k=0. But k>0k>0, so this case is not possible. Case 2: 4m2=k+44m^{2}=k+4 and m2n2+4=3m2+n2+km^{2}-n^{2}+4=3m^{2}+n^{2}+k and 4m2m2n2+44m^{2} \neq m^{2}-n^{2}+4. From the second equality, we obtain 2m2+2n2+k=42m^{2}+2n^{2}+k=4, which is not possible since m,n,k>0m, n, k>0, and so 2m2+2n2+k52m^{2}+2n^{2}+k \geq 5. Therefore, this case is not possible. Case 3: 4m2=3m2+n2+k4m^{2}=3m^{2}+n^{2}+k and m2n2+4=k+4m^{2}-n^{2}+4=k+4 and 4m2m2n2+44m^{2} \neq m^{2}-n^{2}+4. The first equality rearranges to m2n2=km^{2}-n^{2}=k. The second equality also rearranges to m2n2=km^{2}-n^{2}=k. The last statement is equivalent to 3m2+n243m^{2}+n^{2} \neq 4. As we saw in Case 1, this means that (m,n)(m, n) cannot be the pair (1,1)(1,1), which is consistent with m2n2=km^{2}-n^{2}=k and k>0k>0. Therefore, having examined all of the cases, we have reduced the original problem to finding the number of odd integers kk between 0 and 100 for which the equation m2n2=km^{2}-n^{2}=k has exactly two pairs of positive integers (m,n)(m, n) that are solutions. Step 2: Connecting solutions to m2n2=km^{2}-n^{2}=k with factorizations of kk. We can factor the left side of this equation to give (m+n)(mn)=k(m+n)(m-n)=k. Since m,nm, n and kk are positive integers, then m+n>0m+n>0 and k>0k>0 so mn>0m-n>0, or m>nm>n. Since kk is odd and each of m+nm+n and mnm-n is an integer, then each of m+nm+n and mnm-n is odd (since if either was even, then their product would be even). Also, we note that m+n>mnm+n>m-n since n>0n>0. Suppose that (m,n)(m, n) is a solution of the equation m2n2=km^{2}-n^{2}=k with m+n=am+n=a and mn=bm-n=b for some odd positive integers aa and bb with a>ba>b. Then ab=kab=k, so abab is a factorization of kk. Therefore, the solution (m,n)(m, n) corresponds to a specific factorization of kk. Now suppose that we start with a factorization k=ABk=AB where AA and BB are odd positive integers with ABA \geq B. If we try setting m+n=Am+n=A and mn=Bm-n=B, then we can add these equations to give 2m=A+B2m=A+B (or m=12(A+B)m=\frac{1}{2}(A+B)) and subtract them to give 2n=AB2n=A-B (or n=12(AB)n=\frac{1}{2}(A-B)). Note that since n>0n>0, then A>BA>B. Therefore, every factorization of kk as the product of two odd positive integers AA and BB with A>BA>B gives a solution to the equation m2n2=km^{2}-n^{2}=k. Since each solution gives a factorization and each factorization gives a solution, then the number of solutions equals the number of factorizations. Therefore, we have reduced the original problem to finding the number of odd integers kk between 0 and 100 which have exactly two factorizations as the product of distinct odd integers aa and bb with a>ba>b. Step 3: Counting the values of kk. Since kk is odd, then all of its prime factors are odd. Since k<100k<100, then kk cannot have three or more distinct odd prime factors, because the smallest possible product of three distinct odd prime factors is 357=1053 \cdot 5 \cdot 7=105. Thus, kk has two or fewer distinct prime factors. If k=pqk=pq for distinct primes p<qp<q, then the divisors of kk are 1,p,q,pq1, p, q, pq, so kk has exactly two factorizations of the desired type (namely 1pq1 \cdot pq and pqp \cdot q). Since k<100k<100 and p3p \geq 3, then q<1003q<\frac{100}{3}. Since qq is an integer, then q33q \leq 33. The odd primes less than 33 are 3,5,7,11,13,17,19,23,29,313,5,7,11,13,17,19,23,29,31. If p11p \geq 11, then pq>112=121pq>11^{2}=121, which is larger than 100. Therefore, pp can only be 3,5 or 7. If p=3p=3, there are 9 possible values for qq (primes from 5 to 31). If p=5p=5, there are 5 possible values for qq (primes from 7 to 19). If p=7p=7, there are 2 possible values for qq (11 and 13). Thus, there are 9+5+2=169+5+2=16 values of kk of this form that work. If k=prqsk=p^{r}q^{s} with rr and ss positive integers and at least one of rr or ss is larger than 1, then kk will have at least three factorizations. (For example, if r>1r>1, then k=1prqs=ppr1qs=prqsk=1 \cdot p^{r}q^{s}=p \cdot p^{r-1}q^{s}=p^{r} \cdot q^{s} and all of these are distinct.) If k=pk=p or k=p2k=p^{2} with pp an odd prime, then kk has only one factorization as the product of distinct factors (1p1 \cdot p and 1p21 \cdot p^{2}, respectively). Thus, kk cannot be of this form. If k=p3k=p^{3} with pp an odd prime, then the divisors of kk are 1,p,p2,p31, p, p^{2}, p^{3}, so it has exactly two factorizations of the desired type (namely 1p31 \cdot p^{3} and pp2p \cdot p^{2}). Since k<100k<100, then pp can only equal 3 (because 53>1005^{3}>100). Thus, there is 1 value of kk of this form that works. If k=p4k=p^{4} with pp an odd prime, then the divisors of kk are 1,p,p2,p3,p41, p, p^{2}, p^{3}, p^{4}, so it has exactly two factorizations of the desired type (namely 1p41 \cdot p^{4} and pp3p \cdot p^{3}). In this case, kk has a third factorization, but it is of the wrong type since the two factors will be equal. Since k<100k<100, then pp can only equal 3 (because 54>1005^{4}>100). Thus, there is 1 value of kk of this form that works. If kk has more than 4 factors of pp, then kk will have at least three factorizations of the desired type, so kk cannot be of this form. (In particular, if k=pnk=p^{n} and n>4n>4, then k=1pn=ppn1=p2pn2k=1 \cdot p^{n}=p \cdot p^{n-1}=p^{2} \cdot p^{n-2} and these are all distinct since n2>2n-2>2.) Having examined all of the possible forms, we see that there are 16+1+1=1816+1+1=18 values of kk that work, and so there are 18 positive integer solutions to the original equation.

Source: Omni-MATH, licensed Apache-2.0. Statement and solution reproduced as published; topic, difficulty and ordering added by this site.