We say a complex number is complex integer or Gaussian integer if both , the set of real integers. The set of all Gaussian integers is denoted by . Intuitively, the notation means adjoining to real integers. What are complex prime numbers?
Recall that a real integer p is prime if and is divisible only by and (itself). Let us extend the deﬁnition of primes to complex numbers. Suppose we deﬁne a Gaussian integer to be a complex prime if and is divisible only by and (itself). Note that, with such a deﬁnition, no Gaussian integer can be a complex prime.
- Note that is a factor of in and hence will divide all Gaussian integer . This motivates the deﬁnition of unit. An unit is a Gaussian integer which is a factor of in . There are exactly four units in , viz., . This is an outcome of the fact that, in real integers, there are exactly two units, viz., and . It can also be shown that units are, precisely, those elements whose modulus is exactly .
- Even if units are included as factors in the deﬁnition of complex prime, we still notice that no Gaussian integer can be a complex prime. This is because if is a Gaussian integer such that then all its multiples by units, viz., and are all factors of . This motivates the deﬁnition of associate. An associate of a Gaussian integer is a multiple of the units in .
A simple observation from the deﬁnition is that if a Gaussian integer is a Gaussian prime then all of its associates are also Gaussian prime. Further, the conjugate of a Gaussian prime is also a Gaussian prime because if is a factor of then is a factor of .
We deﬁne the map as where . Note the following properties of :
- . The map is multiplicative.
If divides in , then divides in because , for some , and .
Let be a Gaussian integer such that either or and the non-zero, say , satisﬁes , i.e., or . Without loss of generality, we shall assume . When is prime? Equivalently, what are the positive real integers which are Gaussian prime? Obviously, every positive real composite is not a Gaussian prime. Therefore, the question reduces to asking if every real prime is a Gaussian prime. A real prime can fail to be a Gaussian prime only if there is a non-zero, non-real Gaussian integer that divides , i.e., . Thus, a real prime fails to be a Gaussian prime only if it is sum of two squares. For instance, the ﬁrst real prime is not a Gaussian prime because .
Is every odd prime expressible as sum of squares of integers? For instance, is not expressible as sum of squares of two integers. Therefore, is a Gaussian prime. What are the odd primes which can be expressed as sum of two squares? Observe that any odd prime is either modulo or modulo .
Proof. Let us suppose , for some . being odd is either modulo or modulo . Any perfect square is either modulo or modulo . Thus is either , or modulo . Therefore is modulo . The converse part of the proof is quite involved, so we skip it. □
Therefore, we conclude that any odd prime that is modulo is a Gaussian prime. For example, are all Gaussian primes. Alternately, any odd prime that is modulo is not a Gaussian prime. For example, are all not Gaussian primes. So, what are the complex primes other than these real primes?
Proof. Let be a Gaussian prime. Then is also a Gaussian prime. If is not a prime in then it has a positive divisor which also divides either or because contradicting the Gaussian primality of . Conversely, let be a prime in . Suppose is not a Gaussian prime. Then it has non-trivial factors and and, hence, and are non-trivial factors of . Thus, and are non-trivial real factors of , a contradiction. □
Now that we have understood complex primes, we ask: is any Gaussian integer decomposable as product of Gaussian primes? In other words, we seek the analogues of the prime factorization theorem and the fundamental theorem of arithmetic for positive real integers.
The decomposition in to product of primes is unique and is called the fundamental result of arithmetic.
Do we have similar results for Gaussian integers? The answer is in aﬃrmative, as long as, we bear in mind that unique factorisation is uniqueness up to multiplication with units . The unique prime factorization is obvious for Gaussian integers on real and imaginary line. For instance, any real non-prime positive integer has the unique factorisation . If any of the real prime s are not Gaussian prime then they have a Gaussian prime decomposition as discussed above and this factorization is unique up to multiplication by units. For negative integers with , the factorization is the unit multiplied with the factorization of . For with it is still a multiple of the unit or . The unique factorisation can also be shown for any Gaussian integer. Therefore, we have the general result:
The discussion in this article are protoypes of algebraic entites like rings, prime ideals, unique factorization domains, class number etc. Class number corresponds to unique factorization property (UFP). is also related to Pythagorean triples. Similar to , one may consider other quadratic ring , for . Note that corresponds to and corresponds to . The quadratic rings are useful in the study of Pell’s equation
The solution to Pell’s equation are the unit norm elements od . For , and, for , We know, for both and the corresponding ring of integers satisﬁes UFP. Thus, one is tempted to conjecture that satisﬁes the UFP, for all . But the answer is in negation. For instance, do not satisfy UFP. The number has two diﬀerent prime decomposition, viz., . Check that each of the four factors is a prime in . In fact, it has been shown that there are exactly nine negative integers, viz., for which the UFP holds true. These are called Heegner numbers. It is an open problem whether there are ﬁnite number of positive integers for which UFP is true in .