We learn by observing. Then we try to understand what we have learnt. Then we try to prove what we have understood. Then we experiment with what we have proved. Kyriakos Kefalas 2025
Hidden Integer Factors in a Radical Expression
Contents 1.0 Notation 1.1 The algebraic expression 1.2 Linear recurrences 1.3 Conclusion
Kyriakos Kefalas - 4 Oct 2026
Abstract : We consider algebraic expressions whose values are integers in a specified domain, that can be factored into algebraic factors whose values are also integers. The results are integer divisors of the original integer, although not necessarily prime divisors.
1.0 Notation
$$\begin{array}{cc}\mathbb{N} & \text{non}\text{–} \text{negative integers} \\ {\mathbb{N}}_{1} & \text{odd positive integers} \\ {\mathbb{N}}_{2} & \text{even non}\text{–} \text{negative integers} \\ {\mathbb{N}}_{2}^{\ast } & \text{even positive integers} \\\end{array}$$
1.1 The algebraic expression
We consider the algebraic expression
$${y}_{n}=\frac{1}{4}\left( 2+\frac{{(1+\sqrt{2})}^{2n+1}-(17+12\sqrt{2}){(1-\sqrt{2})}^{2n+1}}{4+3\sqrt{2}} \right),\,\qquad n\in \mathbb{N} \tag{1}$$
Why does this expression always give an integer—and why is every term with, \(n\geq 3\), composite?
Question : Why does this expression always give an integer—and why is every term with, \(n\geq 3\), composite?
Indeed every term is a positive integer and admits the following factorization into positive integer-valued algebraic expressions
$$\boxed{{y}_{n}=\left\{\begin{array}{lll}\frac{{a}^{n-1}+{b}^{n-1}}{2}\frac{{a}^{n}+{b}^{n}}{a+b} & \, & n\in {\mathbb{N}}_{1} \\ \, & \, & \, \\ \frac{{a}^{n-1}+{b}^{n-1}}{a+b}\frac{{a}^{n}+{b}^{n}}{2} & \, & n\in {\mathbb{N}}_{2} \\\end{array}\right.},\,\qquad \begin{array}{c}a=1+\sqrt{2} \\ b=\sqrt{2}-1 \\\end{array} \tag{2}$$
Proof
1. Let, \(a=1+\sqrt{2}\), \(b=\sqrt{2}-1\), then the terms of this sequence can be written as
$${y}_{n}=\frac{{a}^{2n-1}+{b}^{2n-1}+a+b}{2(a+b)},\,\qquad n\in \mathbb{N} \tag{3}$$
since, \(a\,b=1\), \(a+b=2\sqrt{2}\), \({a}^{4}=17+12\sqrt{2}\), \({a}^{2} = \frac{4+3\sqrt{2}}{\sqrt{2}}\). This expression factors as
$${y}_{n}=\frac{({a}^{n-1}+{b}^{n-1})({a}^{n}+{b}^{n})}{2(a+b)} \tag{4}$$
where the two integer\(-\)valued algebraic factors are
$${y}_{n}=\left\{\begin{array}{lll}\frac{{a}^{n-1}+{b}^{n-1}}{2}\frac{{a}^{n}+{b}^{n}}{a+b} & \, & n\in {\mathbb{N}}_{1} \\ \, & \, & \, \\ \frac{{a}^{n-1}+{b}^{n-1}}{a+b}\frac{{a}^{n}+{b}^{n}}{2} & \, & n\in {\mathbb{N}}_{2} \\\end{array}\right. \tag{5}$$
and in particular
$$\frac{{a}^{-1}+{b}^{-1}}{a+b}=1,\,\quad \frac{{a}^{0}+{b}^{0}}{2}=1 \tag{6}$$
2. We show that the algebraic factors in (5 ) are positive integers. From the binomial expansion we have
$${a}^{k}={A}_{k}+\sqrt{2}{B}_{k},\,\qquad k,\,{A}_{k},\,{B}_{k}\in \mathbb{N} \tag{7}$$
$$\text{Multiplication by a},\,\text{gives}$$
$${A}_{0}=1,\,\quad {B}_{0}=0,\,\quad {A}_{k+1}={A}_{k}+2{B}_{k},\,\quad {B}_{k+1}={A}_{k}+{B}_{k} \tag{8}$$
By conjugation of (7 ) we have
$${b}^{k}={(-1)}^{k}({A}_{k}-\sqrt{2}{B}_{k}) \tag{9}$$
$$\text{hence}$$
$${a}^{k}+{b}^{k}=\left\{\begin{array}{ll}(a+b){B}_{k} & k\in {\mathbb{N}}_{1} \\ 2{A}_{k} & k\in {\mathbb{N}}_{2} \\\end{array}\right. \tag{10}$$
Then from (4 ) we get
$$\boxed{{y}_{n}=\left\{\begin{array}{lll}1 & \, & n=0 \\ {A}_{n-1}{B}_{n} & \, & n\in {\mathbb{N}}_{1} \\ {A}_{n}{B}_{n-1} & \, & n\in {\mathbb{N}}_{2}^{\ast } \\\end{array}\right.} \tag{11}$$
Every factor displayed is a positive integer. This proves that all sequence members are positive integers. ■■
Recurrence (8 ) shows that both sequences, \({A}_{n},\,{B}_{n}\), are non-decreasing. Since, \({A}_{2}=3,\,{B}_{3}=5\), the factor involving \(A\), in (11 ) is at least 3, and the factor involving \(B\), is at least 5, whenever, \(n\geq 3\). Hence every such term \({y}_{n}\), is composite. The first few terms of the sequence are
$${y}_{n}=1,\,1,\,3,\,15,\,85,\,493,\,2871,\,16731,\,97513,\,568345,\,3312555\ldots ,\,\qquad n\in \mathbb{N} \tag{12}$$
This sequence (without the first term), is given in \(\text{A011900}\), in shifted form.
1.2 Linear recurrences
The algebraic expression in (1 ) satisfies the non–homogeneous linear recurrence
$${y}_{n+2}=6{y}_{n+1}-{y}_{n}-2,\,\quad {y}_{0}={y}_{1}=1,\,\quad n\in \mathbb{N} \tag{13}$$
and the homogeneous linear recurrence
$${y}_{n+3}=7{y}_{n+2}-7{y}_{n+1}+{y}_{n},\,\quad \left\{ \begin{array}{c}{y}_{0}={y}_{1}=1 \\ {y}_{2}=3 \\\end{array} \right\},\,\quad n\in \mathbb{N} \tag{14}$$
Moreover the algebraic factors satisfy the relations
$${A}_{n}={B}_{n+1}-{B}_{n},\,\quad {A}_{n}^{2}-2{B}_{n}^{2}={(-1)}^{n} \tag{15}$$
1.3 Conclusion
Conclusion The expression (1 ) yields positive integers and admits an explicit integer-valued algebraic factorization. For every, \(n\geq 3\), both factors exceed 1, proving that \({y}_{n}\), is composite.