0

I would appreciate some help about the following problem.

Given $z\in\mathbb{C}$ and $$\left\vert {1+z\over 1-z}\right\vert =1 $$ show that $z$ is purely imaginary.
Deduce that the roots of the equation $$2{(1+z)^5}-(1+{\sqrt 3}i){(1-z)}^5=0 $$ are purely imaginary, then write down the five distinct roots of the above equation in $z$ .

Jack D'Aurizio
  • 353,855

3 Answers3

3

A slightly different approach is to notice that $f(z)=\frac{1-z}{1+z}$ is an involutive map: $$\forall z\in\mathbb{C}\setminus\{-1\},\quad f(f(z))=z $$ hence $f(z)$ gives a bijection between the punctured plane and itself. It is simple to notice that $f$ maps the imaginary axis into the unit circle minus the point $-1$ in a bijective way, hence the inverse image of the unit circle minus the point $-1$ is the imaginary axis.

The second equation is equivalent to $$ \left(\frac{1-z}{1+z}\right)^{5} = \exp\left(-\frac{\pi i}{3}\right) $$ and every solution lies on the imaginary axis by the previous lemma. A solution, for instance, is given by $z=\frac{1-\exp(-\pi i/15)}{1+\exp(-\pi i/15)}=i\tan\frac{\pi}{30}$.

Jack D'Aurizio
  • 353,855
  • What about the last part? Roots seem fairly difficult. – sharding4 May 30 '17 at 20:39
  • @sharding4: no, they are not. It is enough to notice that $\frac{1+i\sqrt{3}}{2}=\exp\frac{\pi i}{3}$. – Jack D'Aurizio May 30 '17 at 20:41
  • 1
    Yes. They're much more approachable in trigonometric form than the algebraic surds WolframAlpha kicked out for me. Good example where knowing an answer can really put you off track compared with thinking about how to find an answer. – sharding4 May 30 '17 at 20:45
  • @sharding4: I agree that $\sqrt{7-2 \sqrt{5}-2 \sqrt{3 \left(5-2 \sqrt{5}\right)}}$ is not really pleasant to deal with :D – Jack D'Aurizio May 30 '17 at 20:59
1

Geometric reasoning for the first: $$\left\vert {1+z\over 1-z}\right\vert = 1 \iff |z - 1| = |z - (-1)|,$$ i.e., iff the distance from $z$ to $1$ is equal to the distance from $z$ to $-1$.

Hint for the second: $$2{(1+z)^5}-(1+{\sqrt 3}i){(1-z)}^5=0 \iff \left(\frac{1+z}{1-z}\right)^5 = \frac{1+\sqrt 3 i}2, $$ i.e., $(1+z)/(1-z)$ is a fifth root of $(1+\sqrt 3 i)/2$.

-1

$$|1+z|^2=|1-z|^2$$ $$2Re(z)=-2Re(z)$$ $$Re(z)=0$$ Hence $z$ is purely imaginary.

You can also see this geometrically: If $z$ is equidistant from both $1$ and $-1$, then it has to lie on imaginary axis.

Sahiba Arora
  • 10,847