ProofComplete

Diophantine Equations - Key Proof

We present complete proofs for the parametrization of Pythagorean triples and Fermat's proof that x4+y4=z2x^4 + y^4 = z^2 has no positive integer solutions. These classical results showcase elegant number-theoretic techniques.

ProofParametrization of Pythagorean Triples

We prove that all primitive Pythagorean triples (a,b,c)(a, b, c) with aa odd are given by a=m2βˆ’n2,b=2mn,c=m2+n2a = m^2 - n^2, b = 2mn, c = m^2 + n^2 where m>n>0m > n > 0, gcd⁑(m,n)=1\gcd(m, n) = 1, and m≑̸n(mod2)m \not\equiv n \pmod{2}.

Step 1: Verify the parametrization gives Pythagorean triples.

(m2βˆ’n2)2+(2mn)2=m4βˆ’2m2n2+n4+4m2n2=m4+2m2n2+n4=(m2+n2)2(m^2 - n^2)^2 + (2mn)^2 = m^4 - 2m^2n^2 + n^4 + 4m^2n^2 = m^4 + 2m^2n^2 + n^4 = (m^2 + n^2)^2 βœ“

Step 2: Show every primitive triple arises this way.

Let (a,b,c)(a, b, c) be a primitive Pythagorean triple with aa odd. Then bb is even (if both a,ba, b were odd, a2+b2≑2(mod4)a^2 + b^2 \equiv 2 \pmod{4}, but squares are 00 or 1β€Šmodβ€Š41 \bmod 4, impossible for c2c^2).

From a2=c2βˆ’b2=(cβˆ’b)(c+b)a^2 = c^2 - b^2 = (c-b)(c+b), write b=2bβ€²b = 2b'. Then: a2+4bβ€²2=c2β€…β€ŠβŸΉβ€…β€Ša2=(cβˆ’2bβ€²)(c+2bβ€²)a^2 + 4b'^2 = c^2 \implies a^2 = (c - 2b')(c + 2b')

Since gcd⁑(a,b,c)=1\gcd(a, b, c) = 1 and aa is odd, both cβˆ’2bβ€²c - 2b' and c+2bβ€²c + 2b' are odd.

Claim: gcd⁑(cβˆ’2bβ€²,c+2bβ€²)=1\gcd(c - 2b', c + 2b') = 1.

Proof: Any common divisor dd divides their sum 2c2c and difference 4bβ€²4b'. Since gcd⁑(b,c)=1\gcd(b, c) = 1 and dd is odd, we have d=1d = 1.

Therefore, cβˆ’2bβ€²c - 2b' and c+2bβ€²c + 2b' are both perfect squares: cβˆ’2bβ€²=n2,c+2bβ€²=m2c - 2b' = n^2, c + 2b' = m^2 with m>n>0m > n > 0.

Solving: c=m2+n22,bβ€²=m2βˆ’n24c = \frac{m^2 + n^2}{2}, b' = \frac{m^2 - n^2}{4}.

Wait, let me reconsider. From b2=c2βˆ’a2b^2 = c^2 - a^2:

b2=(cβˆ’a)(c+a)b^2 = (c - a)(c + a)

Since gcd⁑(a,c)=1\gcd(a, c) = 1 (primitive), we have gcd⁑(cβˆ’a,c+a)≀2\gcd(c - a, c + a) \leq 2.

Since cβˆ’ac - a and c+ac + a have the same parity and their product is b2b^2 (even squared), both must be even. Write cβˆ’a=2u,c+a=2vc - a = 2u, c + a = 2v with gcd⁑(u,v)=1\gcd(u, v) = 1.

Then b2=4uvb^2 = 4uv, so b=2uvb = 2\sqrt{uv}. For bb to be an integer, uvuv must be a perfect square. Since gcd⁑(u,v)=1\gcd(u, v) = 1, both uu and vv are perfect squares: u=n2,v=m2u = n^2, v = m^2.

Therefore: c=u+v=n2+m2,a=vβˆ’u=m2βˆ’n2,b=2uv=2mnc = u + v = n^2 + m^2, a = v - u = m^2 - n^2, b = 2\sqrt{uv} = 2mn.

β– 
ProofFermat's Infinite Descent for $x^4 + y^4 = z^2$

We prove no positive integers satisfy x4+y4=z2x^4 + y^4 = z^2 using Fermat's method of infinite descent.

Suppose (x,y,z)(x, y, z) is a solution with x,y,z>0x, y, z > 0 and minimal zz. We may assume gcd⁑(x,y)=1\gcd(x, y) = 1 (otherwise divide by gcd⁑(x,y)4\gcd(x, y)^4).

Then (x2,y2,z)(x^2, y^2, z) is a Pythagorean triple. By the parametrization: x2=m2βˆ’n2,y2=2mn,z=m2+n2x^2 = m^2 - n^2, \quad y^2 = 2mn, \quad z = m^2 + n^2 for some m>n>0m > n > 0 with gcd⁑(m,n)=1\gcd(m, n) = 1 and m≑̸n(mod2)m \not\equiv n \pmod{2}.

From y2=2mny^2 = 2mn: since gcd⁑(m,n)=1\gcd(m, n) = 1 and their product (times 22) is a square, we have two cases:

Case 1: m=2s2,n=t2m = 2s^2, n = t^2 with gcd⁑(s,t)=1\gcd(s, t) = 1.

Then x2=m2βˆ’n2=4s4βˆ’t4x^2 = m^2 - n^2 = 4s^4 - t^4. This gives (t2,2s2,x)(t^2, 2s^2, x) as a Pythagorean triple: t4+4s4=x2t^4 + 4s^4 = x^2

Applying parametrization again leads to a smaller solution.

Case 2: Similar analysis.

In both cases, we construct a Pythagorean triple (a2,b2,c)(a^2, b^2, c) with c<zc < z, contradicting minimality of zz.

Therefore, no positive integer solution exists.

β– 
Remark

The method of infinite descent is remarkably powerful: by showing that every solution generates a strictly smaller solution, we obtain a contradiction from the well-ordering principle. Fermat used this technique to prove many results about Diophantine equations.

ExampleDescent in Action

If (x,y,z)=(2,3,97)(x, y, z) = (2, 3, \sqrt{97}) were a solution (it's not, since 97\sqrt{97} isn't an integer), the descent would produce (xβ€²,yβ€²,zβ€²)(x', y', z') with 0<zβ€²<970 < z' < \sqrt{97}.

Continuing this process indefinitely is impossible for positive integers, proving no solution exists.

These proofs demonstrate the elegance and power of elementary number-theoretic techniques in solving classical Diophantine equations.