1. Given Equation and Initial Setup:
We start with the equation xy−z2=1 for positive integers x,y,z. We aim to show that there exist non-negative integers a,b,c,d such that:
x=a2+b2,y=c2+d2,z=ac+bd
2. **Factorization in Z[i]:**
Consider the factorization in the Gaussian integers Z[i]:
xy=(z+i)(z−i)
By the unique factorization property in Z[i], we can write:
z+i=u1v2,z−i=u2v1
for some non-zero u1,u2,v1,v2∈Z[i].
3. **Expressing x and y:**
Since x and y are real, we can write:
x=u1u1,y=u2u2
where u1 and u2 are the conjugates of u1 and u2, respectively.
4. **Rational Numbers q1 and q2:**
We express v1 and v2 in terms of u1 and u2:
v1=u1x=N(u1)xu1=q1u1,v2=u2y=N(u2)yu2=q2u2
where q1 and q2 are positive rational numbers.
5. **Equating z+i and z−i:**
Substituting these into the equations for z gives:
z+i=q2u1u2,z−i=q1u1u2
Since z+i and z−i are conjugates, the right-hand sides must also be conjugates. Given that q1 and q2 are real, we must have q1=q2.
6. **Common Value k/ℓ:**
Let the common value be ℓk for coprime positive integers k and ℓ. Then:
ℓz+ℓi=kRe(u1u2)+ikIm(u1u2)
From the imaginary parts, we get that k divides ℓ, so k=1. Also, ℓ=Im(u1u2)∣Re(u1u2).
7. **Writing u1 and u2:**
Write u1=a0+b0i and u2=c0+d0i. Then:
ℓ=b0c0−a0d0∣a0c0+b0d0=zℓ
and:
x=ℓu1u1=ℓa02+b02
Thus:
v1=u1x=ℓa0−b0i
Since v1∈Z[i], it follows that ℓ∣a0,b0; similarly, ℓ∣c0,d0.
8. **Substituting a0,b0,c0,d0:**
Write a0=ℓa, b0=ℓb, c0=ℓc, d0=ℓd. Then:
ℓ=ℓ2(bc−ad)⟹ℓ(bc−ad)=1⟹ℓ=1
This implies:
x=a2+b2,y=c2+d2,z=Re(u1u2)=ac+bd
9. Condition Verification:
The condition xy=z2+1 insists on:
(ad−bc)2=xy−z2=1⟹∣ad−bc∣=1
Conversely, any a,b,c,d with ∣ad−bc∣=1 give rise to a solution x=a2+b2, y=c2+d2, z=ac+bd to xy=z2+1.