logo

Proof of the Gaussian Prime Theorem 📂Number Theory

Proof of the Gaussian Prime Theorem

Theorem 1

A Gaussian integer πZ[i]\pi \in \mathbb{Z}[i] is called a Gaussian prime if it satisfies one of the following conditions.

  • (i): π=1+i\pi = 1 + i
  • (ii): For a prime pZp \in \mathbb{Z}, p3(mod4)p \equiv 3 \pmod{4} for π=p\pi = p
  • (iii): For a prime pZp \in \mathbb{Z}, when p1(mod4)p \equiv 1 \pmod{4}, p=u2+v2p = u^2 + v^2 for π=u+iv\pi = u + iv
  • (iv): For π\pi corresponding to (i)~(iii), ikπ i^{k} \pi obtained by multiplying the unit 1,1,i,i1,-1,i,-i of Z[i]\mathbb{Z}[i]
  • (iv): For π\pi corresponding to (i)~(iii), π\overline{\pi} obtained by taking the conjugate

Description

When talking about Gaussian integers, π\pi usually represents a Gaussian prime, not pi. This is to prevent confusion, as the primes of ordinary integers are often referred to as Natural Primes pp. Such an extension of primes makes the study of Gaussian integers truly number-theoretic.

(i)

(1+i)(1 + i) serves as a number that replaces the traditional 2=1(i)2=(1+i)(1i)2 = 1 - (i)^2 = (1 + i)(1-i), implying that there’s no ‘smaller’ Gaussian prime than this.

(ii)

For instance, 77 cannot be factorized into primes even with Gaussian integers. It’s pointless to check if it’s really impossible, so it’s recommended to start with the proof.

(iii), (iv), (v)

As an example, 55 can be factorized since 5=(2+i)(2i)=(1+i2)(1i2)5 = (2 + i)(2 - i) = (1 + i2)(1 - i2). It’s problematic to think that there are two factorizations for Z[i]\mathbb{Z}[i], a UFD, since it can be expressed as the product of units as shown in 2+i=i(12i)2+i = i(1 - 2i). Meanwhile, (2+i)(2+i) and (2i)(2-i), and (1+i2)(1 + i2) and 1i21 - i2 form conjugates, proving they are Gaussian primes.

Proof

Strategy: (Z[i],N)(Z[i] , N) is a Gaussian ring defined by the norm as N(x+iy)=x2+y2N(x+iy) = x^2 + y^2. (Where x,yZx, y \in \mathbb{Z}) The proof of Gaussian prime theorem itself is nothing more than combining its algebraic properties with various results from elementary number theory, but understanding these properties and results is the challenging part.


Part 0. If N(π)=p| N( \pi ) | = p is a prime, then π\pi is a Gaussian prime.

Multiplicative Norm Property: Let’s say pZ p \in \mathbb{Z} is a prime.

  • [1]: If the multiplicative norm NN is defined as N(1)=1N(1) = 1, then for all units uDu \in D, N(u)=1| N ( u ) | = 1
  • [2]: If all αD\alpha \in D satisfying N(α)=1| N ( \alpha )| =1 are units in DD, then πD\pi \in D satisfying N(π)=p| N ( \pi ) | = p is an irreducible element in DD.

Properties of Gaussian Ring:

  • [3]: The only units in Z[i]\mathbb{Z}[i] are 1,1,i,i1,-1,i,-i.

According to [3], the only units in Z[i]\mathbb{Z}[i] are 1,1,i,i1,-1,i,-i and by [1] N(1)=N(1)=N(i)=N(i)=1 N(1) = N(-1) = N(i) = N(-i) = 1 . Since all α\alpha satisfying N(α)=1| N ( \alpha )| =1 were units in Z[i]\mathbb{Z}[i], by [2], π\pi satisfying N(π)=p| N( \pi ) | = p becomes an irreducible element in Z[i]\mathbb{Z}[i]. In other words, if N(π)=p| N( \pi ) | = p is a prime, then π\pi is a Gaussian prime.


Part (i). π=1+i\pi = 1 + i

If π=1+i\pi = 1 + i then N(π)=12+12=2N(\pi) = 1^2 + 1^2 = 2 is a prime, so by Part 0, π=1+i\pi = 1 + i is a Gaussian prime.


Part (ii). π3(mod4)\pi \equiv 3 \pmod{4}

Let’s assume there exists a factorization like π=(a+ib)(c+id)\pi = ( a + ib )( c + id ) for a prime π=p\pi = p satisfying p3(mod4)p \equiv 3 \pmod{4} that is not a Gaussian prime of Z[i]\mathbb{Z}[i]. According to the multiplicative property of NN, p2=π2+02=N(π+i0)=N(a+ib)N(c+id)=(a2+b2)(c2+d2) \begin{align*} p^2 =& \pi^2 + 0^2 \\ =& N ( \pi + i 0 ) \\ =& N ( a + ib ) N ( c + id ) \\ =& (a^2 + b^2) (c^2 + d^2) \end{align*} Then, p2=(a2+b2)(c2+d2)p^2 = (a^2 + b^2) (c^2 + d^2), but since pZp \in \mathbb{Z} is a prime, there must exist a solution satisfying {a2+b2=pc2+d2=p\begin{cases} a^2 + b^2 = p \\ c^2 + d^2 = p \end{cases}.

Necessary and Sufficient Condition for a Prime to be Congruent to 1 mod 4: Let’s say p2p \ne 2 is a prime. When p1(mod4)p \equiv 1 \pmod{4}, for some a,bZa,b \in \mathbb{Z}, p=a2+b2p = a^2 + b^2

However, since p3(mod4)p \equiv 3 \pmod{4}, according to the Necessary and Sufficient Condition for a Prime to be Congruent to 1 mod 4, there’s no solution that satisfies {a2+b2=pc2+d2=p\begin{cases} a^2 + b^2 = p \\ c^2 + d^2 = p \end{cases}, leading to a contradiction, hence π3(mod4)\pi \equiv 3 \pmod{4} is a Gaussian prime.


Part (iii). π=u+iv\pi = u + iv

For a prime pZp \in \mathbb{Z}, since p1(mod4)p \equiv 1 \pmod{4}, according to the Necessary and Sufficient Condition for a Prime to be Congruent to 1 mod 4, N(π)=N(u+iv)=u2+v2=p N (\pi) = N ( u+ iv) = u^2 + v^2 = p π=u+iv\pi = u + iv that satisfies these conditions is a Gaussian prime due to Part 0.


Part (iv). ikπ i^{k} \pi

Since in Part 0, if N(π)=p| N( \pi ) | = p is a prime, then π\pi is a Gaussian prime, therefore for kZk \in \mathbb{Z}, N(ikπ)=N(ik)N(π)=1N(π)=p \begin{align*} N( i^{k} \pi ) =& N( i^{k} ) N (\pi ) \\ =& 1 \cdot N (\pi ) \\ =& p \end{align*} ikπi^{k} \pi fulfilling these are also Gaussian primes.


Part (v). π\overline{\pi}

Since in Part 0, if N(π)=x2+y2=p| N( \pi ) | = x^2 + y^2 = p is a prime, then π=x+iy\pi = x + i y is a Gaussian prime, N(π)=N(xiy)=x2+(y)2=p \begin{align*} N( \overline{\pi} ) =& N( x - i y ) \\ =& x^2 + (-y)^2 \\ =& p \end{align*} thus, π\overline{\pi} fulfilling these conditions are also Gaussian primes.


  1. Silverman. (2012). A Friendly Introduction to Number Theory (4th Edition): p275. ↩︎