4

Given the equation $x^5=i$, I need to show by both algebraic and trigonometrical approaches that $$\cos18^{\circ}=\frac{\sqrt{5+2\sqrt5}}{\sqrt[5]{176+80\sqrt5}}$$ $$\sin18^{\circ}=\dfrac1{\sqrt[5]{176+80\sqrt5}}$$ Trying by trigonometric approach, $x^5$ = i $\;\;\;\;$ -- eqn. (a)

=> x = $i\sin(\dfrac{\pi}{2} +2k\pi)$ => $i\sin(\pi\dfrac{4k + 1}{2}) $

Taking the value of k=0, for getting the principal root of 18$^{\circ}$, have x = $i\sin(\dfrac{\pi}{10}) $

Solving algebraically, the solution approach is : $(a+bi)^5$ = i $\;\;\;\;$ -- eqn. (b)

=> $a^5 + 5ia^4b -10a^3b^2 -10ia^2b^3 +5ab^4 +ib^5$

Separating the real & imaginary parts:

$a^5 -10a^3b^2 +5ab^4=0$$\;\;\;$ -- eqn. (c); $\;\;\;\;$$5a^4b -10ia^2b^3+b^5=1$$\;\;\;$ -- eqn. (d)

Solving (c), we have : $a(a^4 -10a^2b^2 +5b^4)=0$$\;\;\;$ -- eqn. (c);

Either $a$ = $0$, or $(a^4 -10a^2b^2 +5b^4)=0$$\;\;$ -- eqn. (c'),

dividing both sides by $b^4$, and having c = a/b, $(c^4 -10c^2 +5)=0$$\;\;$ -- eqn. (c''),

having d = $c^2$, get : $(d^2 -10d +5)=0$$\;\;$ -- eqn. (c'''), with factors as : d =$5\pm 2\sqrt5$

finding value of c for the two values, get square roots of the two values for d.

//Unable to proceed any further with (c''').

Only root of significance, from eqn. (c) is $a = 0$.

Taking eqn.(d), and substituting $a = 0$, we get:$\;\;\;b^5$=1 => $b =1$

//Unable to prove any of the two values for $\sin18^{\circ}$, or $\cos18^{\circ}$

jiten
  • 4,524
  • 1
    The right approach is to solve the equation $x^{5}=1$ which is easy. And then use trigonometric identities. Note that one solution is $$x=\cos 72^{\circ}+i\sin 72^{\circ}=\sin 18^{\circ}+i\cos 18^{\circ}$$ – Paramanand Singh Sep 26 '17 at 13:08
  • The possible angle are : $18^{\circ}, 90^{\circ}, 162^{\circ}, 234^{\circ}, 306^{\circ}$. the angle of $72^{\circ}$ occurs nowhere. May be, I am unable to understand the significance of using $cos(90^{\circ}-\theta) = \sin(\theta)$, and vice-versa (i.e, for $\sin$) – jiten Sep 26 '17 at 13:16
  • @jiten This question might be useful: https://math.stackexchange.com/questions/438362/evaluate-cos-18-circ-without-using-the-calculator – jonsno Sep 26 '17 at 13:54
  • Are you permitted use of the multiple-angle formulas? – Eric Towers Sep 26 '17 at 14:11
  • @EricTowers: No condition is stated. Just the two approaches: algebraic, trigonometric are needed. – jiten Sep 27 '17 at 00:01

4 Answers4

3

Let's start with equation $x^{5}-1=0$ whose one root is $x=\cos(2\pi/5)+i\sin(2\pi/5)$. The equation can be written as $$(x-1)(x^{4}+x^{3}+x^{2}+x+1)=0$$ The first factor gives the root $x=1$ and the second factor leads to the equation $$x^{2}+x^{-2}+x+x^{-1}+1=0$$ Putting $y=x+x^{-1}$ we get $$y^{2}+y-1=0$$ or $$y=\frac{-1\pm\sqrt{5}}{2}$$ Choosing the positive root we get $$2\cos(2\pi/5)=y=\frac{\sqrt{5}-1}{2}$$ Thus the value of $\sin 18^{\circ}$ is obtained as $(\sqrt{5}-1)/4$. Similarly we can find the value of $\cos 18^{\circ}$.


Observe that apart from $x=1$ there are $4$ distinct values of $x$ out of which we have to choose only one namely $x=\cos(2\pi/5)+i\sin(2\pi/5)$. The problem of choice is simplified considerably by using $y=x+x^{-1}$ which satisfies a quadratic equation and therefore has only two values. For our desired value of $x$ the expression $y>0$ and hence the positive root $y$ is chosen. And in reality we are interested in the value $\cos(2\pi/5)=(x+x^{-1})/2=y/2$ so the choice of $y$ completes our work.

The method can be generalized (thanks to Gauss) to solve higher degree equations of type $x^{n} =1$. For example we can solve $x^{17}=1$ and get the value of $\cos(2\pi/17)$ as $$\frac{-1 + \sqrt{17} + \sqrt{34 - 2 \sqrt{17}} + 2\sqrt{17 + 3 \sqrt{17} - \sqrt{34 - 2 \sqrt{17}} - 2\sqrt{34 + 2 \sqrt{17}}}}{16}$$ (see this post for more details).

  • @jiten: that's just algebraic manipulation. But I must say whoever obtained those formats must have done very messy work to obtain those ugly expressions. Mathematics should be simple as well as beautiful. – Paramanand Singh Sep 26 '17 at 13:35
  • 1
    @jiten To free the radical on denominator, consider the expression inside square root, ie $176 + 80\sqrt{5} = (x+y)^5$. By some trial, you can deduce that it must be $(1+\sqrt{5})^5$ – jonsno Sep 26 '17 at 13:44
  • +1 Paramanand Nice answer! I too think someone has tried to make the simple result look ugly.. – jonsno Sep 26 '17 at 13:48
  • @samjoe: I think the idea comes from the equation $x^{5}=i$ given in question. If $x=a+ib$ then one value is $a=\cos(\pi/10),b=\sin(\pi/10)$ and then $b^{5}$ is evaluated as $1/(176+80\sqrt{5})$. – Paramanand Singh Sep 26 '17 at 13:53
  • @ParamanandSingh: i am unable to gather how if x = $\cos(\theta)$, then y = x + 1/x => y = $2\cos(2\theta)$. – jiten Sep 27 '17 at 00:06
  • @jiten: note that $x=\cos\theta +i\sin\theta $ (see beginning of my answer). And then $x^{-1}=\cos\theta-i\sin\theta$ and therefore $y=x+x^{-1}=2\cos\theta$. In my answer $\theta =2\pi/5=72^{\circ}$. – Paramanand Singh Sep 27 '17 at 02:59
  • @jiten: I have added some explanation which shows how the desired value of $x$ is chosen out of all the roots available. – Paramanand Singh Sep 27 '17 at 03:22
  • @ParamanandSingh: Thanks a lot. But, would have liked more an answer that can be used as algorithm (sort of) for that kind of problems. Your answer, (may be I need to spend much time more) may need some sort of tricks (like choosing suitable value for x, that enable sum of x + 1/x to be of good use, for the problem at hand) that may vary. It might be that you have kept into view the final answer into account; but the needed form is still different. I need study your method for a lot of time to find its use as an algorithmic technique (to be used alone, or in conjunction with other approaches). – jiten Sep 27 '17 at 13:58
  • 1
    @jiten: the proper way to deal with values of circular functions for rational multiples of $\pi$ is the use of the identity $(\cos \theta+i\sin\theta) ^{n} =\cos n\theta+i\sin n\theta$ and it ultimately boils down to the solution of $x^{n} =1$ and the substitution $y=x+(1/x)$. This is the systematic approach without any tricks to deal with such problems. I would suggest you study more on "cyclotomic polynomials" which are the key here. The needed form in your question is messy and can only be verified. It can not be obtained via some procedure. – Paramanand Singh Sep 27 '17 at 17:12
0

I think you're going at this backwards. Quickly, $\arg \mathrm{i} = \pi/2 = 90^\circ$, so fifth roots of $\mathrm{i}$ are points on the unit circle at small multiples of $18^\circ$. If the given values of cosine and sine are to be believed, then \begin{align*} \mathrm{i} &= \left( \cos 18^\circ + \mathrm{i} \sin 18^\circ \right)^5 \\ &= \left( \frac{\sqrt{5+2\sqrt5}}{\sqrt[5]{176+80\sqrt5}} + \mathrm{i} \frac{1}{\sqrt[5]{176+80\sqrt5}} \right)^5 \text{.} \end{align*} Let $x$ be the expression we are raising to the fifth power. Then, squaring twice, (and leaving the denominator largely unchanged) \begin{align*} x^2 &= \frac{2(2 + \sqrt{5}) + 2\mathrm{i}\sqrt{5+2\sqrt{5}}}{(\sqrt[5]{176+80\sqrt{5}})^2} \text{,} \\ x^4 &= \frac{8(2 + \sqrt{5}) + 8\mathrm{i}((2+\sqrt{5})\sqrt{5+2\sqrt{5}})}{(\sqrt[5]{176+80\sqrt{5}})^4} \text{, and} \\ x^5 = x^4 \cdot x &= \frac{16 \mathrm{i}(11+5\sqrt{5})}{176+80\sqrt{5}} = \mathrm{i} \text{.} \end{align*}

This confirms that the given values of cosine and sine are a fifth root of $\mathrm{i}$. The other roots also have unit magnitude and arguments multiples of $360^\circ/5 = 72^\circ$ greater than the given root, so have arguments $90^\circ$, $162^\circ$, $234^\circ$, and $306^\circ$. We see that $x$ is the only one in the first quadrant, so is the desired root.

For a trigonometry method: $$ 0 = \cos(90^\circ) = \cos(5 \cdot 18^\circ) = \cos( 2 \cdot 2 \cdot 18^\circ + 18^\circ) \text{.} $$ Then use sum of angles and half angle, twice. Do the same thing to sine. You'll get a pair of algebraic equations in $\sin 18^\circ$ and $\cos 18^\circ$. Then check that the given sine and cosine values actually satisfy them.

Eric Towers
  • 67,037
  • 1
    Nice, but I feel you meant that 'your' method goes backwards.I mean that substituting back answer to prove it is correct. – jiten Sep 26 '17 at 21:15
  • The evaluation of powers of $x$ is far too complicated compared to showing that $(1+\sqrt{5})^{5}=176+80\sqrt{5}$ via binomial theorem. And this shows that the given value of $\sin(\pi/10)$ is $1/(1+\sqrt{5})=(\sqrt{5}-1)/4$. – Paramanand Singh Sep 27 '17 at 00:04
0

Trigonometrical way: One has $$\sin 36^\circ = \cos 54^\circ = 4\cos^3 18^\circ - 3\cos 18^\circ$$ $$2\sin 18^\circ\cos 18^\circ = 4\cos^3 18^\circ - 3\cos 18^\circ$$ $$2\sin 18^\circ = 4\cos 18^\circ -3 = 1-4\sin 18^\circ$$

So, $\sin 18^\circ$ is positive root of equation $4x^2+2x-1=0$, or $\sin 18^\circ = \frac{-1+\sqrt{5}}{4}$.

GAVD
  • 7,296
  • 1
  • 16
  • 30
  • Nice, but there is still no way to get the stated solution right. There is an attempt by @Eric Towers to prove it backwards that the solution is correct. – jiten Sep 26 '17 at 22:30
0

This problem seems a bit wonky to me because there are simpler expressions for these functions and easier ways to prove them, but let's roll with it.

Once you render $c^2=5\pm2\sqrt5$, take its square roots to get four roots

$c=a/b=\pm\sqrt{5\pm2\sqrt5}$

with the $\pm$ signs independent. Now from DeMoivre's Theorem we know that the complex root we are seeking is $\cos18°+i\sin18°$, so $a/b$ must be the cotangent of $18°$, therefore greater than $1$. Only by choosing both $\pm$ signs as positive can we get that result; $\sqrt{5-2\sqrt5}$ is too small with $2\sqrt5=\sqrt{20}$ lying between $4$ and $5$.

So, $a=(\sqrt{5+2\sqrt5})b$. Plug that into the imaginary part of $(a+bi)^5$ you computed above and equate this to $1$, which gives $1/b^5=176+80\sqrt5$. This will lead easily to the final claims.

Oscar Lanzi
  • 39,403