0

Consider the sequence $a_n$ of the infinite decimal expansion of ${\pi}$. If I prove that the sequence $\sin(a_n)$ converges to $\sin({\pi})$ and similarly consider sequence $b_n$ of the infinite decimal expansion of ${e}$ , then $\sin(b_n)$ converges to $\sin({e})$ .

As both sub-sequence converge to different limits, does this imply the divergence of $\sin(n)$ for $n$ belonging to real numbers?

Sumanta
  • 9,534

2 Answers2

1

Suppose that the sequence $(\sin n )$ converges to a finite limit $g.$ Then the subsequence $(\sin (2n) )$ also converges to $g. $ Hence the sequence $$(|\sin (2n) -\sin (2n +2)|)$$ converges to $0.$ Moreover the sequence $(|\cos n |)=(\sqrt{1-\sin^2 n})$ converges to $\sqrt{1-g^2 }.$ Therefore $$2\sin 1\sqrt{1-g^2}=2\sin1 \lim_{n\to\infty} \cos (2n +1) =\lim_{n\to\infty } |\sin (2n) -\sin (2n +2)| =0.$$ Therefore the sequence $(\cos (2n +1))$ converges to $0$ and hence the sequence the sequence $(|\cos n |) $converges to $0.$ Moreover the sequence $$(|\cos (2n+1) -\cos (2n +3)|)=(2\sin 1 \sin (2n+2) ) $$ is also convergent to $0$ and this implies that $(\sin n)$ is convergent to $0.$ But $$1=\sin^2 n +\cos^2 n \to 0^2 +0^2 =0.$$

Contradiction.

1

If we're interested in proving $\sin n$ diverges, note that $e^{in}$ belongs to $\{e^{it}: \pi/4<t< 3\pi/4\}$ for infinitely many $n,$ and belongs to $\{e^{it}: 5\pi/4<t< 7\pi/4\}$ for infinitely many $n.$ Thus $\sin n >1/\sqrt 2$ for infinitely many $n,$ and $\sin n <-1/\sqrt 2$ for infinitely many $n.$

zhw.
  • 105,693