1 December 2016
by IMA

Smallest positive integer

[vc_row][vc_column][vc_column_text]$1^2+2^2=5$, which is not a square. $1^2+2^2+3^2=14$, which is not a square. What is the smallest positive integer value of $n, n>1,$ such that $1^2+2^2+ \dots +n^2$ is a square number? Are there any larger possible values of $n$?[latexpage][/vc_column_text][lvca_accordion][lvca_panel]

To show that $\sum\limits_{r=1}^{n}{{{r}^{2}}}={{N}^{2}}$ has a unique solution in integers for $n>1$.

Since $\operatorname{hcf}\left( m,n \right)=\operatorname{hcf}\left( m,m+n \right)$, with $m=n+1$ we have that $n,n+1$ and $2n+1$ are co-prime. For  $\sum\limits_{r=1}^{n}{{{r}^{2}}}=\frac{n\left( n+1 \right)\left( 2n+1 \right)}{6}$ to be square the (co-prime) terms in the numerator must be square numbers multiplied by coefficients with product 6.
$n$ $n+1$ $2n+1$ Reason why not possible
$6{{a}^{2}}$ ${{b}^{2}}$ ${{c}^{2}}$
${{a}^{2}}$ $6{{b}^{2}}$ ${{c}^{2}}$ $n+1=6{{b}^{2}}\Rightarrow 2n+1=12{{b}^{2}}-1$ which can't be square since congruent to 11(mod12).
${{a}^{2}}$ ${{b}^{2}}$ $6{{c}^{2}}$ $2n+1$ is an odd number but $6{{c}^{2}}$ is even.
${{a}^{2}}$ $2{{b}^{2}}$ $3{{c}^{2}}$ $n+1=2{{b}^{2}}\Rightarrow 2n+1=4{{b}^{2}}-1$ which can't be square since congruent to 3(mod4).
${{a}^{2}}$ $3{{b}^{2}}$ $2{{c}^{2}}$ $2n+1$ is an odd number but $2{{c}^{2}}$ is even.
$2{{a}^{2}}$ ${{b}^{2}}$ $3{{c}^{2}}$ $n=2{{a}^{2}}\Rightarrow 2n+1=4{{a}^{2}}+1$ but this is not divisible by 3: $4{{\left( 3k\pm 1 \right)}^{2}}+1\equiv 2\left( \bmod 3 \right)$ and $4{{\left( 3k \right)}^{2}}+1\equiv 1\left( \bmod 3 \right)$.
$2{{a}^{2}}$ $3{{b}^{2}}$ ${{c}^{2}}$ $n+1=3{{b}^{2}}\Rightarrow 2n+1=6{{b}^{2}}-1$ which can't be square since congruent to 5(mod6).
$3{{a}^{2}}$ ${{b}^{2}}$ $2{{c}^{2}}$ $2n+1$ is an odd number but $2{{c}^{2}}$ is even.
$3{{a}^{2}}$ $2{{b}^{2}}$ ${{c}^{2}}$ $n+1=2{{b}^{2}}\Rightarrow 2n+1=4{{b}^{2}}-1$ which can't be square since congruent to 3(mod4).
This leaves only $n=6{{a}^{2}}, n+1={{b}^{2}}, 2n+1={{c}^{2}}$. Since $\operatorname{hcf}\left( n+1,2n+1 \right)=1$ then $\operatorname{hcf}\left( {{b}^{2}},{{c}^{2}} \right)=1$ and so $\operatorname{hcf}\left( b,c \right)=1$ and $\operatorname{hcf}\left( b,b+c \right)=1.$ $\operatorname{hcf}\left( b-c,b+c \right)=\operatorname{hcf}\left( 2b,b+c \right)$ and by the line above this is 1 or 2. ${{c}^{2}}-{{b}^{2}}=\left( c-b \right)\left( c+b \right)=6{{a}^{2}}$. Note $a=1$ does not lead to a solution (since $6\times {{1}^{2}}+1$ is not square)  so the factors could be
$c-b$ $c+b$ Reason why not possible
1 $6{{a}^{2}}$ $2c=1+6{{a}^{2}}$ would not give an integer for $c$.
2 $3{{a}^{2}}$
3 $2{{a}^{2}}$ $2c=3+2{{a}^{2}}$ would not give an integer for $c$.
6 ${{a}^{2}}$ ${{c}^{2}}={{\left( b+6 \right)}^{2}}={{b}^{2}}+12b+36$. This leads to $n=12b+36$ and so ${{b}^{2}}-1=12b+36$ which does not have integer solutions.
${{a}^{2}}$ 6 $c>b>1$ . Only possibility is $c=4,b=2$ but doesn't lead to a solution.
This leaves only $c-b=2,c+b=3{{a}^{2}}$ . Substitute $c=b+2$ in $n+1={{b}^{2}},2n+1={{c}^{2}}$ leads to ${{b}^{2}}-4b-5=0$. Then $b=5,c=7,a=2$ giving $n=24,n+1=25,2n+1=49$. \[\sum\limits_{r=1}^{24}{{{r}^{2}}}={{\left( 2\times 5\times 7 \right)}^{2}}={{70}^{2}}.\]   For N>1 the next solution is N=24 (sum=4900 ie 70x70) It is worth noting that N=1, N=24 are the only solutions. An extensive proof can be found by Prof. G.N. Watson in ”Messenger of Mathematics” 1918, Volume 48, Pages 1-22 (an electronic version can be seen at: http://archive.org/stream/messengerofmathe4849cambuoft#page/n9/mode/2up) [/lvca_panel][/lvca_accordion][/vc_column][/vc_row][vc_row][vc_column][vc_column_text]
Problem Page Coordinator: Stephen Lee CMath MIMA – Mathematics in Education and Industry Acknowledgement: The IMA are indebted to MEI for sourcing and supplying Mathematics Today with these well-known puzzles.
First published in Mathematics Today December 2016
Image credit: University Club Entryway Pilaster Letter N (New York, NY) by takomabibelot / Flickr / CC-BY-PD-1.0
[/vc_column_text][/vc_column][/vc_row]

Related topics