I have to find where the Cauchy-Reimann equations $\left( \frac{\partial u}{\partial x} = \frac{\partial v}{\partial y} \text{ and } \frac{\partial u}{\partial y} = - \frac{\partial v}{\partial x}\right)$ are satisfied for the function $$f(x+iy)= \sqrt{\frac{\sqrt{x^2+y^2}+x}{2}}+isign(y)\sqrt{\frac{\sqrt{x^2+y^2}-x}{2}}.$$

I found that $$\frac{\partial u}{\partial x} =\frac{1}{2}\left(\frac{(x^2+y^2)^{1/2}+x}{2}\right)^{-1/2}\frac{1}{2}\left( \frac{1}{2}\left(x^2+y^2\right)^{-1/2}(2x)+1\right),$$

$$\frac{\partial v}{\partial y}= sign(y)\frac{1}{2}\left(\frac{(x^2+y^2)^{1/2}-x}{2}\right)^{-1/2}\frac{1}{2}\left( \frac{1}{2}\left(x^2+y^2\right)^{-1/2}(2y)\right),$$

$$\frac{\partial u}{\partial y} = \frac{1}{2}\left(\frac{(x^2+y^2)^{1/2}+x}{2}\right)^{-1/2}\frac{1}{2}\left( \frac{1}{2}\left(x^2+y^2\right)^{-1/2}(2y)\right),$$


$$\frac{\partial v}{\partial x} = sign(y)\frac{1}{2}\left(\frac{(x^2+y^2)^{1/2}-x}{2}\right)^{-1/2}\frac{1}{2}\left( \frac{1}{2}\left(x^2+y^2\right)^{-1/2}(2x)-1\right).$$

Of course, these can be simplified a bit (wolfram alpha will give something more concise), and when I put them in the Cauchy-Riemann equations, some things cancel, but I still am left with something too complicated (for me) to figure out any conditions on $x$ and $y$ that make these hold. I've also tried adding the equations, as well as setting $$\frac{\partial u}{\partial x} - \frac{\partial v}{\partial y} + \frac{\partial u}{\partial y} + \frac{\partial v}{\partial x} = 0,$$

but couldn't deduce anything from this either.

My instructor hinted to note that $\sqrt{x^2+y^2}\geq |x|,$ but I'm not seeing how to use this fact. This hint made me think that I should look at this function as $$f(z) = \sqrt{\frac{\sqrt{z\bar{z}}+RE(z)}{2}} + isign(IM(z))\sqrt{\frac{\sqrt{z\bar{z}}-RE(z)}{2}}.$$

But this doesn't seem to provide me with anymore insight. I was thinking of going this route because we proved in class that a function is analytic iff you can write the function in terms of only $z.$

I have a hunch that the hint provided will lead to the function not being analytic anywhere, but am not sure if this is true because I don't really have a good sense of what this function does or how it acts on the complex plane.

Thanks in advance for any help / hints you can provide.

  • 1
    $\begingroup$ Write $z=x+iy$. What is $(f(z))^2$? I think that could help a little. Also, I think that your instructor's hint was intended to stress that $f$ is given by a pair of real valued functions (because the root is well defined) $\endgroup$ – Brandon Mar 8 '14 at 6:26
  • $\begingroup$ Thanks Brandon. Looking at what you suggested, I found $$f(x+iy)^2 = x + isign(y)\left(\frac{y}{\sqrt{2}}\right).$$ Since this function is not analytic anywhere (because there are no values for which the C-R equations hold), then the original function must not be analytic, correct? $\endgroup$ – Jonathan Mar 8 '14 at 14:41
  • $\begingroup$ You have miscalculated the square, $$f(x+iy)^2 = x + 2i\operatorname{sign}y \sqrt{\frac{y^2}{4}} = x+iy.$$ $\endgroup$ – Daniel Fischer Mar 8 '14 at 14:47

You can simplify the derivatives a bit more. It is maybe easier to see if we write $r = \sqrt{x^2+y^2}$ and note $\frac{\partial r}{\partial x} = \frac{x}{r},\; \frac{\partial r}{\partial y} = \frac{y}{r}$ when $r > 0$. Then for generic points ($x\neq 0 \neq y$) we find

$$\begin{align} \frac{\partial u}{\partial x} &= \frac{1}{2u}\cdot \frac{1}{2}\left(\frac{\partial r}{\partial x} + \frac{\partial x}{\partial x}\right) = \frac{1}{4u}\left(\frac{x}{r}+1\right) = \frac{x+r}{4r\sqrt{\frac{r+x}{2}}} = \frac{\sqrt{r+x}}{2\sqrt{2}\cdot r},\\ \frac{\partial v}{\partial y} &= \frac{1}{2v}\cdot \frac{1}{2}\left(\frac{\partial r}{\partial y} - \frac{\partial x}{\partial y}\right) = \frac{1}{4v}\frac{y}{r} = \frac{y}{4r\operatorname{sign} y\sqrt{\frac{r-x}{2}}} = \frac{\sqrt{r^2-x^2}}{2\sqrt{2}\cdot r\sqrt{r-x}}, \end{align}$$

from which it is not hard to see that $\frac{\partial u}{\partial x} = \frac{\partial v}{\partial y}$. The cases $x = 0$ or $y =0$, but $r > 0$ need to be handled separately (further distinguishing between positive and negative $x$ resp. $y$; noting also that $v$ isn't continuous everywhere).

The treatment of $\frac{\partial u}{\partial y}$ and $\frac{\partial v}{\partial x}$ is similar.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.