I have a question which has fluxommed me and my pals for the past few days. Any help or solution is welcome

Show using Binomial theorem that the integer just after $(3^{1/2} + 1)^{2n}$ is divisble by $2^{n+1}$.

here n belongs to natural numbers (positive integers).

  • $\begingroup$ Just for clarification: So you mean $\lceil (3^{1/2} + 1)^{2n} \rceil \mod 2^{n+1}=0$? $\endgroup$ – draks ... Apr 4 '12 at 6:54
  • $\begingroup$ By ‘the integer just after’ I assume that you mean the ceiling, $$\left\lceil(\sqrt3+1)^{2n}\right\rceil\;.$$ $\endgroup$ – Brian M. Scott Apr 4 '12 at 6:54
  • $\begingroup$ By "the integer just after" I mean the following : Suppose n=1 we get 7.46 and the integer just after is 8 which is divisible by 4 $\endgroup$ – Rohan Apr 4 '12 at 6:58
  • 1
    $\begingroup$ Actually, in order for it to work for $n=0$ we have to make it $$\left\lfloor(1+\sqrt3)^{2n}\right\rfloor+1\;.$$ $\endgroup$ – Brian M. Scott Apr 4 '12 at 7:02
  • 2
    $\begingroup$ @Rohan: There are many mathematicians who consider $0$ to be a natural number. So it is better to make it unambiguous by saying positive integer etc. $\endgroup$ – Aryabhata Apr 4 '12 at 7:20

This is a classic, the standard proof is to consider

$$ (\sqrt{3} + 1)^{2n} = x_n + y_n \sqrt{3}$$

for integer $x_n$, $y_n$, (which you can see using Binomial theorem) and show that

$$ (\sqrt{3} - 1)^{2n} = x_n - y_n \sqrt{3}$$

again using Binomial theorem.

Now use the fact that $\sqrt{3} - 1 \lt 1$ and that

$$(\sqrt{3} + 1)^{2n} + (\sqrt{3} - 1)^{2n} = 2x_n$$

Thus the integer you are looking for is $2x_n$.

Get a recurrence for $x_n$ and $y_n$ and use induction.

To get the recurrence:

we have that $(\sqrt{3} + 1)^2 = 4 + 2\sqrt{3}$ and so

$$ x_{n+1} + y_{n+1} = (x_n + y_n \sqrt{3})(4 + 2 \sqrt{3}) = (4x_n + 6y_n) + (2x_n + 4y_n) \sqrt{3}$$


$$ x_{n+1} = 4x_n + 6y_n$$ $$ y_{n+1} = 2x_n + 4y_n$$

Write $$x_{n+2} = 4x_{n+1} + 6y_{n+1}$$ and eliminate $y_n$ and $y_{n+1}$

(I believe it comes to $x_{n+2} = 8x_{n+1} - 4 x_n$)

  • $\begingroup$ well I accept your answer but can you please show a bit details like what you mean by "recurrence for x and y" I am really not as advanced as you are. Thnx for help $\endgroup$ – Rohan Apr 4 '12 at 7:06
  • $\begingroup$ A very big thanks man I had to rack my brains to understand your stuff but I finally got it !!! Could you also help me out with the induction $\endgroup$ – Rohan Apr 4 '12 at 8:03
  • $\begingroup$ @Rohan: I suggest you try it out first. $\endgroup$ – Aryabhata Apr 4 '12 at 15:25
  • $\begingroup$ @Aryabhata Emm! Why do you say this is a classic? Has it appeared before in some famous Math Magazines? If yes, do send me a source :) $\endgroup$ – crskhr Jun 29 '18 at 11:37
  • $\begingroup$ @crskhr: I have seen it in textbooks which were dated early 1900s. Sorry don't have a reference handy. $\endgroup$ – Aryabhata Jun 29 '18 at 21:00

This similar to the Fibonacci sequence and is going to involve a recurrence.

Note that $1+\sqrt{3}$ is a solution of $x^2=2x+2$ and the other solution is $1-\sqrt{3}$.

So you can see that $$a_n=(1+\sqrt{3})^n + (1-\sqrt{3})^n$$ satisfies $$a_{n+2}=2a_{n+1}+2a_n.$$ If this is not instantly obvious then prove it by induction. The sequence $a_n$ starts $2,2,8,20,56,152,416,\ldots$. OEIS A080040

Since $(1-\sqrt{3})^{2n}$ is positive and is less than $1$ for positive $n$, you have $$a_{2n}=\left\lceil(\sqrt3+1)^{2n}\right\rceil\;$$ for positive $n$.

Since $a_0$ and $a_1$ are both divisible by $2$, it is an easy induction from the recurrence that $a_{2n}$ and $a_{2n+1}$ are both divisible by $2^{n+1}$.


This is similar to Henry's answer, but things are simpler if we use $4+2\sqrt{3}=(1+\sqrt{3})^2$ and $4-2\sqrt{3}=(1-\sqrt{3})^2$ instead. These satisfy $x^2-8x+4=0$.

The sequence defined by $$ a_0=2\text{, }a_1=8\text{, and }a_n=8a_{n-1}-4a_{n-2}\tag{1} $$ has the solution $$ \begin{align} a_n &=(4+2\sqrt{3})^n+(4-2\sqrt{3})^n\\ &=(1+\sqrt{3})^{2n}+(1-\sqrt{3})^{2n}\\ &=\left\lceil(\sqrt{3}+1)^{2n}\right\rceil\text{ for }n>0\tag{2} \end{align} $$ Let $b_n=a_n2^{-n-1}$. Then, $(1)$ becomes $$ b_0=1\text{, }b_1=2\text{, and }b_n=4b_{n-1}-b_{n-2}\tag{3} $$ Recursion $(3)$ insures that $b_n\in\mathbb{Z}$ for all $n\ge0$. Thus, for $n>0$, $(2)$ yields $$ \begin{align} \left\lceil(\sqrt{3}+1)^{2n}\right\rceil &=a_n\\ &=2^{n+1}b_n\\ &\in2^{n+1}\mathbb{Z}\tag{4} \end{align} $$


You can prove by induction method and easily get answer.

Obviously, the statement is true for n =1.

Suppose the statement is true for n=k;

i.e, (√3+1)^(2k) is divisible by 2^(k+1).

(√3+1)^(2k) = [2^(k+1)]*q, for some q.

Now we have to show that the statement is true for n=k+1;

(√3+1)^(2(k+1)) = (√3+1)^(2k + 2) 
                = [(√3+1)^(2k)]*[(√3+1)^(2)]
                = [2^(k+1)]*q*[3 + 1 + 2√3]
                = [2^(k+1)]*q*2*(2 +√3)
                = [2^(k+2)]*q*(2 +√3)  

There fore, (√3+1)^(2(k+1)) is divisible by 2^(k+2).

There fore the statement is true for n=k+1.

  • $\begingroup$ Thnx Prasad G I appreciate ur efforts but I require A proof through Binomial Theorem $\endgroup$ – Rohan Apr 4 '12 at 10:05
  • $\begingroup$ @Rohan: Unfortunately, the edit I made to your question makes that unclear. So it is not Prasad's fault. Of course, this answer is incorrect for other reasons... $\endgroup$ – Aryabhata Apr 4 '12 at 15:25
  • $\begingroup$ What Prasad has done is incorrect as although his method his correct n=1 does't satisfy the equation and he was required to prove that the next integer is divisble which has not featured in his solution $\endgroup$ – Rohan Apr 6 '12 at 9:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.