Square-root of a complex number

We define the set of complex numbers, C, as the set of pairs of real numbers, R2, equipped with addition and multiplication. However, our original motive for complex numbers was to solve the quadratic equation in general. How do they relate to each other? 



Theorem (square root of a complex number)

Suppose that a,bR. The square roots of a+ib are ±(x+iy) where
x=a+a2+b22,y=sign(b)a+a2+b22
where
sign(b)={1(if b<0),1(otherwise).

Remark. As useful as it is, memorizing this formula is not recommended. Why? Because this "formula" is so complicated that you will most likely forget it! Try to derive the square root of a+ib every time by solving
(x+iy)2=a+ib
for x and y. □
Proof. The result is nearly trivial if b=0 (exercise!). So let's assume b0. Let z=a+ib and w=x+iy where a,b,x,yR such that z=w2 (i.e., w is a square root of z). Thus,
z=w2=(x+iy)2=(x2y2)+2xyi
so (z)=(w2) and (z)=(w2) give
a=x2y2,b=2xy.
Since b0, x0 and y0. So we have y=b/(2x). Substituting this into the first equation, we have
a=x2(b2x)2
from which we obtain
4(x2)24ax2b2=0.
This is a ``quadratic'' equation of x2. Solving this for x2 gives
x2=a±a2+b22.
However, note that a2+b2>|a| since b0. This means (aa2+b2)/2<0, so this cannot be the square of the real number x. Hence we are left with
x2=a+a2+b22
and
x=±a+a2+b22.
From x2y2=a, we have
y2=x2a=a+a2+b22a=a+a2+b22.
If b>0, we need to have b=2xy>0. This can be accomplished if we choose both + or both for x and y. If b<0, then b=2xy<0, which can be achieved if we choose opposite signs for x and y. All these can be done consistently by putting sign(b) in y.

Finally, by reversing the argument (or by directly verifying (x+iy)2=a+ib), we confirm that thus obtained x+iy is indeed the square root of a+ib. ■

Example. The square root of 1 is ±i
Example. Let us find i. Since i=0+1i, i=±(x+iy) where
x=0+02+122=12y=sign(1)0+02+122=12.
Thus i=±1+i2. To check this,
(±1+i2)2=1+2i+i22=i.
Example. Let us find i without using the formula. Let i=x+iy where x,yR. Thus,
(x+iy)2=x2+2ixy+i2y2=(x2y2)+(2xy)i=i.
By comparing the real and imaginary parts, we have
x2y2=0,2xy=1.
From the second equation, y=1/(2x). Substituting this into the first, we have x4=1/4, so x=±1/2. If x=1/2, then y=1/(2x)=1/2. If x=1/2, y=1/2. Therefore,
i=±1+i2.

Linear, quadratic, cubic, quartic, equations

To solve a linear equation with real coefficients,
ax+b=0, a,bR,a0,
we only need real numbers. That is, the solution is always real.

To solve a quadratic equation with real coefficients,
ax2+bx+c=0,  a,b,cR,a0,
we needed to extend the field R to C. That is, the solutions are generally complex.

Now consider a cubic equation with complex coefficients,
ax3+bx2+cx+d=0,  a,b,c,dC,a0.
Do we need to extend the field C to something ``bigger''? The answer is negative. In fact, we have the following theorem.

Theorem (Fundamental Theorem of Algebra)

A polynomial equation of degree n with coefficients in C,
anxn+an1xn1++a1x+a0=0,  a0,a1,,anC,an0,
has exactly n (possibly redundant) solutions in C. 
Proof. Omitted. ■

Comments

Popular posts from this blog

Birth process

Branching processes: Mean and variance

Informal introduction to formal logic