Sep 23-27

**Definition:** An integer \(p>1\) is called *prime* if its only positive divisors are \(1\) and \(p\). Otherwise it is called *composite.*

**Proposition:** Every integer greater than \(1\) can be written as a product of prime numbers (including the case where the integer is a product of just one prime number.)

**Lemma:** Let \(N>1\) be an integer. Let \(d>1\) be the smallest divisor of \(N\) greater than \(1\). Then \(d\) is prime.

**Proof:** By contradiction. If \(d\) is not prime, it has a divisor \(r\) greater than \(1\) and smaller than \(d\). Since \(r|d\), and \(d|N\), \(r\) is a divisor of \(N\). (Proposition 2.1 (i)). This contradicts the fact that \(d\) is the smallest divisor of \(N\) greater than \(1\). Therefore \(d\) is prime.

**Proof of the Proposition:** Let \(S\) be the set of integers greater than one that are not a product of prime numbers. If \(S\) is not empty, it has a smallest element, Call that element \(N\). Let \(d\) be the smallest divisor of \(N\) greater than \(1\). If \(d=N\), then \(N\) is prime by the Lemma, so it is a product of prime numbers, which is a contradiction of \(N\in S\). If \(d<N\), then \(d\) is a prime number by the lemma, and \(N/d<N\). Since \(N/d<N\), \(N/d\not\in S\), so \(N/d\) is a product of prime numbers. But then \(N=d(N/d)\) so \(N\) is also a product of prime numbers. Therefore \(S\) must be empty, and every integer is a product of primes.

Implicit in this result is an algorithm for writing \(N\) as a product of primes. Given \(N\), start with \(2\) and try dividing \(N\) by \(2\). Do this until it’s not divisible by \(2\) any more. Then do that by \(3\), and \(4\), and so on.

**Theorem:** There are infinitely many primes.

**Proof:** We will show that, given any prime number \(P\), there is a prime number \(Q\) that is greater than \(P\). Given \(P\), let \(M\) be the product of all the prime numbers less than or equal to \(P\), and let \(H=M+1\). Notice that if \(L\le P\) is a prime number, then \(L|M\), so \(L\not|H\) (Proposition 2.1(ii)). Let \(Q\) be the smallest divisor of \(H\) that is greater than \(1\). By the lemma, \(H\) is prime. By the preceeding remark, \(Q\) cannot be less than or equal to \(P\). Therefore \(Q\) is a prime number greater than \(P\), as desired.

% Math 2710 % Sep 23-27

**Theorem:** Let \(N>0\) be an integer and let \(b>0\) be another integer. Then there exists an integer \(n\) and exactly one set of integers \(r_0,\ldots, r_n\), with \(r_n\not=0\) and all \(0\le r_{i}<b\), so that \[
N = r_{n}b^{n}+r_{n-1}b^{n-1}+\ldots+r_{0}.
\]

**Proof:** One proof of this is given on pages 42-43 of the text. Here is a slightly different one. First we prove that, for every positive integer, there is at least one set \(r_{0},\ldots, r_{n}\) such that \[
N = r_{n}b^{n}+r_{n-1}b^{n-1}+\ldots+r_{0}.
\]

Then we will show that there is only one such set. Let \(S\) be the set of positive integers for which there DO NOT exist an integer \(n\) and at least one sequence \(r_0,\ldots, r_n\) as in the theorem. We will show \(S\) is empty by contradiction. So if \(S\) is not empty, by well-ordering it has a smallest element. Call that element \(M\). By the division algorithm, we can write \(M=Ab+r\) with \(0\le r<b\). Since \(A<M\), and \(M\) is the first number not of the desired form, we can write \[
A = r_{m}b^{m}+\cdots+r_{0}
\] But then \[
M=Ab+r=r_{m}b^{m+1}+\cdots+r_{0}b+r
\] which *IS* of the desired form. This contradicts the assertion that \(S\) is non-empty so \(S\) must be empty.

To show that there is only one sequence that works, suppose we have two such sequences so that \[ N = r_{n}b^{n}+r_{n-1}b^{n-1}+\ldots+r_{0}. \]

and also

\[ N = s_{n}b^{n}+s_{n-1}b^{n-1}+\ldots+s_{0}. \]

If the two representations are different, there must be a smallest integer \(j\) such that \(r_j\not=s_j\). Subtracting the two representations, all of the terms involving \(b^{i}\) for \(i<j\) would cancel out, so we would have

\[ N-N=0=b^{j}(Ab+(r_j-s_j)) \]

and so \(Ab+(r_j-s_j)=0\). Since \(b\) divides \(Ab\), we must have \(b\) divides \(r_j-s_j\), and since both are between \(0\) and \(b\) , this means \(r_j-s_j=0\). This contradicts the assumption that there was a \(j\) where \(r_j\) and \(s_j\) were different so they must all be the same

**Proposition:** If \(p\) is prime, and \(p\) divides a product \(ab\), then \(p|a\) or \(p|b\).

**Proof:** We know that \(\mathrm{gcd}(a,p)=1\) or \(\mathrm{gcd}(a,p)=p\). In the second case, \(p|a\). In the first, case, we can apply Proposition 2.28 to see that \(p|b\).

**Theorem:** Let \(N>0\) be a positive integer. Then there is one and only one way to write \(N\) as a product of primes written in non-decreasing order.

**Proof:** Assume the result is false and Let \(N\) be the smallest integer that has two such representations \[
p_1p_2\cdots p_k=q_1q_2\cdots q_k.
\] Then \(p_1|q_1q_2\cdots q_k\). If \(p_1=q_1\), we could cancel \(p_1\) from the two representations and get a smaller integer \(N/p_1\) with two representations, so we must have \(p_1\not=q_1\). Therefore \(p_1|q_2\cdots q_k\). By the same argument, \(p_1\not=q_2\) so \(p_1|q_3\cdots q_k\). Continuing in this way we eventually get \(p_1|q_k\). Since \(p_1\not=1\), we have \(p_1=q_k\). This means we can cancel \(p_1=q_k\) from the two representations to get a smaller integer with two representations; that’s a contradiction since \(N\) was the smallest such. Therefore the representation is unique.

**Definition:** Let \(\mathrm{ord}_p(n)\) be the power of \(p\) that occurs in the prime factorization of \(n\).

**Proposition:** If \(m\) and \(n\) are two integers and \(\mathrm{ord}_p(n)=\mathrm{ord}_{p}(m)\) for all primes \(p\), then \(n=\pm m\).

**Proposition:** If \(d\) and \(n\) are two integers, then \(d|n\) if and only if \(\mathrm{ord}_p(d)\le \mathrm{ord}_p(n)\) for all primes \(p\).

**Proposition:**

- \(\mathrm{ord}_p(\mathrm{gcd}(a,b))=\mathrm{min}(\mathrm{ord}_p(a),\mathrm{ord}_p(b))\) for all primes \(p\).
- \(\mathrm{ord}_p(\mathrm{gcd}(a,b))=\mathrm{max}(\mathrm{ord}_p(a),\mathrm{ord}_p(b))\) for all primes \(p\).