Basics

This page contains pre-university level elementary results in number theory.

Definitions

Definition 1 (Divisibility)

For all integers \(a,b\), where \(b\neq 0\), \(b\) divides \(a\) iff there exists an integer \(n\) such that \(a=nb\).

Definition 2 (Prime Number)

For all integers \(n\geq 2\), \(n\) is prime iff it has exactly \(1\) and \(n\) itself as its positive divisors.

Definition 3 (Coprime Numbers)

For all integers \(p,q\), where \(p,q\neq 0\),
\(p\) and \(q\) are coprime iff their greatest common divisor \(\gcd(p,q)\) is \(1\).

Propositions

Proposition 1

For all prime numbers \(p\), \(p\) and \(p+1\) are coprime.

Proof. Take any common divisor \(d\) of \(p\) and \(p+1\). By the property of divisibility, \(d\mid p\) and \(d\mid p+1\) implies that \(d\mid (p+1)-p\) or \(d\mid 1\). Since \(d\) is a divisor of \(1\), it must be \(d=1\) only. Therefore, \(p\) and \(p+1\) are coprime.

Proof. Take the greatest common divisor \(d\) of \(p\) and \(p+1\).

Since \(d\mid p\) and \(p\) is prime, \(d\) must be either \(1\) or \(p\):

  • If \(d=1\), then the only common divisor of \(p\) and \(p+1\) is \(1\). Therefore, \(p\) and \(p+1\) are coprime.

  • If \(d=p\), then \(p\mid (p+1)\). This implies that \(p+1=pn\iff p(n-1)=1\) for some integer \(n\). However, this is impossible since \(p\) is prime and thus greater than \(1\).


Proposition 2

For all prime numbers \(p\) and integers \(a\geq 2\), \(p\) divides \(a\) iff \(p\) divides \(a^2\).

Proof. Take any prime numbers \(p\) and integers \(a\geq 2\).

To show \(\implies\), assume \(p\mid a\).
Then there exists an integer \(n\) such that \(a=np\). This implies that \(a^2=n^2p^2\), and thus \(p\mid a^2\).

To show \(\impliedby\), assume \(p\mid a^2\).
Assume for a contradiction that \(p\) does not divide \(a\), meaning \(\gcd(p,a)=1\).
By Bezout’s identity (Theorem 3), there exist integers \(x,y\) such that \(px+ay=\gcd(p,a)=1\). Multiply both sides to obtain \(apx+a^2y=a\). Since \(p\) divides the LHS, it must divide the RHS \(a\), which is a contradiction.

Lemmas

Lemma 1 (Euclid’s Division Lemma)

For all integers \(a,b\), where \(b\neq 0\), there exist unique integers \(q,r\) such that
\(a=bq+r\) and \(0\leq r<|b|\).

Proof. TODO

Theorems

Theorem 2 (Fundamental Theorem of Arithmetic)

For all integers \(n\geq 2\), there exists a unique prime factorization of \(n\).

Proof. TODO


Theorem 3 (Bezout’s Identity)

For all integers \(a,b\), there exist integers \(x,y\) such that \(ax+by=\gcd(a,b)\).

Proof. Take any integers \(a,b\).

Consider the sequence of pairs \((r_i,r_{i+1})\) defined by the Euclidean algorithm,

  • \(r_{-2}=q_0\cdot r_{-1}+r_0\)

  • \(r_{-1}=q_1\cdot r_0+r_1\)

  • \(r_0=q_2\cdot r_1+r_2\)

  • \(r_{n-2}=q_n\cdot r_{n-1}+r_n\)

where \(r_{-2}=a,r_{-1}=b\) and \(r_n=0\).

The proof proceeds by the (reversed) mathematical induction.
For each \(i\), we assert that there exist integers \(x,y\) such that \(r_{i-2}\cdot x+r_{i-1}\cdot y=d\), where \(d=\gcd(a,b)\).
The claim of the theorem holds by letting \(i=0\).

Base case \(i=n\):
Note that \(r_{n-1}=d\) and \(r_n=0\) by the correctness of the Euclidean algorithm.
Hence we have that \(r_{n-2}=q_n\cdot r_{n-1}+r_n\iff r_{n-2}-(q_n-1)\cdot r_{n-1}=d\),
and thus there exist integers \(x=1,y=1- q_n\) such that \(r_{-2}\cdot x+r_{-1}\cdot y=d\).

Inductive step \(i=k\):
Assume by the inductive hypothesis there exist integers \(s,t\) such that \(r_{k-2}\cdot s+r_{k-1}\cdot t=d\).
Note that \(r_{k-3}=q_{k-1}\cdot r_{k-2}+r_{k-1}\) and \(r_{k-1}\cdot t=d-r_{k-2}\cdot s\).
Hence we have \(r_{k-3}=q_{k-1}\cdot r_{k-2}+(d-r_{k-2}\cdot s)/t\).

By elementary algebra we obtain:

\[\begin{split} \begin{align} & r_{k-3}=q_{k-1}\cdot r_{k-2}+(d-r_{k-2}\cdot s)/t \\ \iff\quad& r_{k-3}\cdot t=q_{k-1}\cdot r_{k-2}\cdot t+d-r_{k-2}\cdot s \\ \iff\quad& r_{k-3}\cdot t-r_{k-2}(q_{k-1}\cdot t-s)=d \end{align} \end{split}\]

Hence there exist \(x=t,y=s-q_{k-1}\cdot t\) such that \(r_{k-3}\cdot x+r_{k-2}\cdot y=d\).