Find argument of a complex number

Aion

Junior Member
Joined
May 8, 2018
Messages
214
Hi all, I’m trying to understand a solution involving complex numbers in trigonometric form. The problem is:

Find the argument of the complex number z1=z2zz_1=z^2-z if z=cos(φ)+isin(φ)z=cos(\varphi)+isin(\varphi).

The solution proceeds as follows:

z1=(cos(φ)+isin(φ))2(cos(φ)+isin(φ))z_1=\left( cos(\varphi)+isin(\varphi)\right)^2-(cos(\varphi)+isin(\varphi))z1=(cos(2φ)cos(φ))+i(sin(2φ)sin(φ))z_1=\left( cos(2\varphi)-cos(\varphi)\right)+i\left( sin(2\varphi)-sin(\varphi)\right)z1=2sin(φ2)(sin(3φ2)+icos(3φ2))z_1=2sin\left(\frac{\varphi}{2}\right)\left( -sin\left(\frac{3\varphi}{2} \right)+icos\left( \frac{3\varphi}{2}\right)\right)
Thus the modulus is

z1=4sin2(φ2)(sin2(3φ2)+cos2(3φ2))=2sin(φ2)\left| z_1\right|=\sqrt{4sin^2\left(\frac{\varphi}{2}\right)\left(sin^2\left( \frac{3\varphi}{2}\right) + cos^2\left( \frac{3\varphi}{2}\right)\right)}=2 \left| sin\left(\frac{\varphi}{2}\right)\right|
In accordance with the definition of absolute value, we have to consider three cases:

(a) If sin(φ2)=0sin\left(\frac{\varphi}{2} \right)=0, that is φ=2kπ\varphi=2k\pi, kk is any integer, the argument of z1z_1 is not defined.

(b) If sin(φ2)>0sin\left(\frac{\varphi}{2} \right)>0, which occurs when 4πk<φ<(4k+2)π4\pi k<\varphi<(4k+2)\pi then z1=2sin(φ2)\left| z_1\right|=2sin\left(\frac{\varphi}{2} \right) and the trigonometric form of the complex number z1z_1 is as follows:

z1=2sin(φ2)(sin(3φ2)+icos(3φ2))=2sin(φ2)(cos(π+3φ2)+isin(π+3φ2))z_1=2sin\left(\frac{\varphi}{2}\right)\left( -sin\left(\frac{3\varphi}{2} \right)+icos\left( \frac{3\varphi}{2}\right)\right)=2sin\left(\frac{\varphi}{2}\right)\left( cos\left(\frac{\pi+3\varphi}{2} \right)+isin\left( \frac{\pi+3\varphi}{2}\right)\right)
Consequently, if φ\varphi satisfy the above condition, then arg(z1)=π+3φ2arg(z_1)=\frac{\pi+3\varphi}{2}
(c) If sin(φ2)<0sin\left(\frac{\varphi}{2} \right)<0, that is (4k+2)π<φ<(4k+4)π(4k+2)\pi<\varphi<(4k+4)\pi, kk any integer then z1=2sin(φ2)\left| z_1\right|=-2sin(\frac{\varphi}{2})

z1=2sin(φ2)(sin(3φ2)+icos(3φ2))=2sin(φ2)(sin(3φ2)icos(3φ2))=2sin(φ2)(cos(3π+3φ2)+isin(3π+3φ2))z_1=2sin\left(\frac{\varphi}{2}\right)\left( -sin\left(\frac{3\varphi}{2} \right)+icos\left( \frac{3\varphi}{2}\right)\right)=-2sin\left(\frac{\varphi}{2}\right)\left( sin\left(\frac{3\varphi}{2} \right)-icos\left( \frac{3\varphi}{2}\right)\right)=-2sin\left(\frac{\varphi}{2}\right)\left( cos\left(\frac{3\pi+3\varphi}{2} \right)+isin\left( \frac{3\pi+3\varphi}{2}\right)\right)
Hence, if φ\varphi satisfies the condition above, then arg(z1)=3π+3φ2arg(z_1)=\frac{3\pi+3\varphi}{2}
I understand most of this derivation, but I’m having trouble fully grasping the step where the expression inside the parentheses is rewritten using the angle identities. Is there an easy way to see that
sin(3φ2)icos(3φ2)=cos(3π+3φ2)+isin(3π+3φ2) sin\left(\frac{3\varphi}{2} \right)-icos\left( \frac{3\varphi}{2}\right)= cos\left(\frac{3\pi+3\varphi}{2} \right)+isin\left( \frac{3\pi+3\varphi}{2}\right)I can see that the result checks out using the addition formulas, but I don’t quite understand how to derive it on my own from the beginning.
 
Is there an easy way to see that
sin(3φ2)icos(3φ2)=cos(3π+3φ2)+isin(3π+3φ2) sin\left(\frac{3\varphi}{2} \right)-icos\left( \frac{3\varphi}{2}\right)= cos\left(\frac{3\pi+3\varphi}{2} \right)+isin\left( \frac{3\pi+3\varphi}{2}\right)I can see that the result checks out using the addition formulas, but I don’t quite understand how to derive it on my own from the beginning.
First, I assume you see WHY they want to do this: to put it in the form cos( ) + i sin( ), so they can read off the argument from the expression. In particular, they want that angle to be within whatever range they consider appropriate for the argument (presumably (π,π](-\pi,\pi]).

So you want to rewrite sin(3φ2)\sin\left(\frac{3\varphi}{2} \right) as a cosine.

The natural way to do this is the cofunction identity sin(θ)=cos(π2θ)\sin(\theta)=\cos(\frac{\pi}{2}-\theta). Try doing that, and then manipulate it further to get the argument in range. (You may find it helpful to use the fact that sine is an odd function, or that cosine is an even function, to get rid of a subtraction.)

Then, of course, repeat the process with the imaginary part.
 
First, I assume you see WHY they want to do this: to put it in the form cos( ) + i sin( ), so they can read off the argument from the expression. In particular, they want that angle to be within whatever range they consider appropriate for the argument (presumably (π,π](-\pi,\pi]).

So you want to rewrite sin(3φ2)\sin\left(\frac{3\varphi}{2} \right) as a cosine.

The natural way to do this is the cofunction identity sin(θ)=cos(π2θ)\sin(\theta)=\cos(\frac{\pi}{2}-\theta). Try doing that, and then manipulate it further to get the argument in range. (You may find it helpful to use the fact that sine is an odd function, or that cosine is an even function, to get rid of a subtraction.)

Then, of course, repeat the process with the imaginary part.
We have one system of equations to solve.

1. cos(x)=sin(3φ2)cos(x)=sin\left(\frac{3\varphi}{2}\right)
2. sin(x)=cos(3φ2)sin(x)=-cos\left(\frac{3\varphi}{2}\right)

We first solve equation 1.
cos(x)=sin(3φ2)=cos(π23φ2)cos(x)=sin\left(\frac{3\varphi}{2}\right)=cos\left(\frac{\pi}{2}-\frac{3\varphi}{2} \right)x=±(π23φ2)+2πkx=\pm \left( \frac{\pi}{2}-\frac{3\varphi}{2}\right)+2\pi kx1=π23φ2+2πkx_1=\frac{\pi}{2}-\frac{3\varphi}{2}+2\pi kx2=3φ2π2+2πkx_2=\frac{3\varphi}{2}-\frac{\pi}{2}+2\pi k
Then we solve equation 2.
sin(x)=cos(3φ2)=sin(π23φ2)=sin(3φ2π2)sin(x)=-cos\left(\frac{3\varphi}{2}\right)=-sin\left(\frac{\pi}{2}-\frac{3\varphi}{2} \right)=sin\left(\frac{3\varphi}{2}-\frac{\pi}{2} \right)x1=3φ2π2+2πkx_1=\frac{3\varphi}{2}-\frac{\pi}{2}+2\pi kx2=3π23φ2+2πkx_2=\frac{3\pi}{2}-\frac{3\varphi}{2}+2\pi k
Comparing the two solutions for xx. One value appears in both equations:
x=3φ2π2+2πkx=\frac{3\varphi}{2}-\frac{\pi}{2}+2\pi kIf we let k=1k=1, then we have
x=3φ+3π2x=\frac{3\varphi+3\pi}{2}
 
I am too lazy to draw a diagram, but here is a geometric approach:
  1. z=z2=1|z| = |z^2| = 1, which means that 0, zz and z2z^2 form an isoscless triangle.
  2. In such triangles the base, i.e. z2zz^2-z, is orthogonal to the bisector of the apex angle,
  3. The argument of the bisector is 12(ϕ+2ϕ)=3ϕ2\frac{1}{2}(\phi+2\phi) = \frac{3\phi}{2}
  4. The argument of base is thus 3ϕ2+π2\frac{3\phi}{2}+\frac{\pi}{2}.
The reason I like this geometric approach is that it works for arbitrary powers znzmz^n - z^m.
 
Find the argument of the complex number z1=z2zz_1=z^2-z if z=cos(φ)+isin(φ)z=cos(\varphi)+isin(\varphi).
I like to solve this problem by using polar form of Euler's formula and the property of multiplying two arguments.

Polar Euler’s formula:\displaystyle \text{Polar Euler's formula:}
eiφ=cosφ+isinφ\displaystyle e^{i\varphi} = \cos \varphi + i\sin \varphi

The argument property of a product:\displaystyle \text{The argument property of a product:}
arg(z1z2)=arg(z1)+arg(z2)\displaystyle \text{arg}(z_1z_2) = \text{arg}(z_1) + \text{arg}(z_2)


We have z1=z2z=z(z1)=eiφ(eiφ1)\displaystyle \large z_1 = z^2 - z = z(z - 1) = e^{i\varphi}(e^{i\varphi} - 1)

Then,

arg(z1)=arg(eiφ(eiφ1))=arg(eiφ)+arg(eiφ1)\displaystyle \large \text{arg}(z_1) = \text{arg}(e^{i\varphi}(e^{i\varphi} - 1)) = \text{arg}(e^{i\varphi}) + \text{arg}{(e^{i\varphi} - 1})

Since eiφ\displaystyle \large e^{i\varphi} is a vector starts at the origin and ends on a point on the unit circle, its argument is:
arg(eiφ)=φ\displaystyle \large \text{arg}(e^{i\varphi}) = \varphi

Since eiφ1\displaystyle \large e^{i\varphi} - 1 is a vector starts at the point (1,0)\displaystyle \large (1, 0) on the unit circle and ends on the point (cosφ,sinφ)\displaystyle \large (\cos \varphi, \sin \varphi) on the unit circle, we can draw two lines from the origin of the unit circle to connect the tail and the head of that vector. This forms an isosceles triangle.

The angle bisector of this isosceles triangle with the chord of that vector is φ2\displaystyle \large \frac{\varphi}{2}. If we rotate this bisector 90\displaystyle 90 degrees (φ2+π2)\displaystyle \large \left(\frac{\varphi}{2} + \frac{\pi}{2}\right), we get the direction of the vector eiφ1\displaystyle \large e^{i\varphi} - 1.

Then,

arg(eiφ1)=φ2+π2\displaystyle \large \text{arg}(e^{i\varphi} - 1) = \frac{\varphi}{2} + \frac{\pi}{2}

Finally, we get:

arg(z1)=arg(eiφ)+arg(eiφ1)=φ+φ2+π2=3φ2+π2\displaystyle \large \text{arg}(z_1) = \text{arg}(e^{i\varphi}) + \text{arg}{(e^{i\varphi} - 1}) = \varphi + \frac{\varphi}{2} + \frac{\pi}{2} = \frac{3\varphi}{2} + \frac{\pi}{2}

💪😼😼
 
I am too lazy to draw a diagram, but here is a geometric approach:
  1. z=z2=1|z| = |z^2| = 1, which means that 0, zz and z2z^2 form an isoscless triangle.
  2. In such triangles the base, i.e. z2zz^2-z, is orthogonal to the bisector of the apex angle,
  3. The argument of the bisector is 12(ϕ+2ϕ)=3ϕ2\frac{1}{2}(\phi+2\phi) = \frac{3\phi}{2}
  4. The argument of base is thus 3ϕ2+π2\frac{3\phi}{2}+\frac{\pi}{2}.
The reason I like this geometric approach is that it works for arbitrary powers znzmz^n - z^m.
Interesting metod. Can we obtain both solutions with this approach, or does it assume that 0<φ<π0<\varphi<\pi? I will attempt to draw a diagram of your method.

Geometry.png
 
Last edited:
Can we obtain both solutions with this approach, or does it assume that 0<φ<π0<\varphi<\pi0<φ<π?
Good point. It becomes 3ϕπ2\frac{3\phi-\pi}{2} for π<ϕ<2π\pi < \phi < 2\pi. Here is an "algebra-geometric" reasoning:
The angle between two complex numbers uu and vv measured from uu to vv is the argument of vu\frac{v}{u}
Bisector's direction is given by z2+zz^2 + z, thus we are interested in the argument of z2zz2+z=z1z+1\frac{z^2-z}{z^2+z} = \frac{z-1}{z+1}.
But:
z1z+1=(z1)(zˉ+1)(z+1)(zˉ+1)=z21+2(z)z+12=(z)2z+12\frac{z-1}{z+1} = \frac{(z-1)(\bar z +1)}{(z+1)(\bar z+1)} = \frac{|z|^2-1 + 2\Im (z)}{|z+1|^2} = \Im(z)\frac{2}{|z+1|^2}Thus the resulting angle is that of (z)\Im(z), i.e. the imaginary part of zz, which is ±π2\pm\frac{\pi}{2} depending on the argument of zz.
 
There is some ambiguity here. Consider ϕ1=π2\phi_1 = -\frac{\pi}{2} and ϕ2=3π2\phi_2 = \frac{3\pi}{2}. For both cases z=iz = -i and θ=arg(z2z)=arg(i1)=3π4\theta = \arg(z^2-z) = \arg(i-1) = \frac{3\pi}{4}. But
θ3ϕ12=3π2π2          and      θ3ϕ12=3π2π2\theta - \frac{3\phi_1}{2} = \frac{3\pi}{2}\equiv -\frac{\pi}{2} \;\;\;\;\;\text{and}\;\;\; \theta - \frac{3\phi_1}{2} = -\frac{3\pi}{2} \equiv\frac{\pi}{2}I.e., if ϕ\phi is defined mod  2π\mod 2\pi then 3ϕ2\frac{3\phi}{2} is defined mod  π\mod \pi :)
 
Top