8

Because the function $\frac{1}{z}$ has a pole at $z=0$, the integral $\int_{0}^{1} \frac{dx}{x} $ doesn't converge.

On the other hand, the integral $\int_{0}^{1} \sin \left(\frac{1}{x} \right) \, dx $ converges even though the function $\sin \left(\frac{1}{z} \right)$ has an essential singularity at $z=0$.

Can an essential singularity be a weaker singularity than an pole?

EDIT:

Originally I had $\int_{0}^{1} \sin \left(\frac{1}{x^{2}} \right) \, dx$ (which converges despite the fact that $\sin \left(\frac{1}{z^{2}} \right)$ has an essential singularity at $z=0$).

1 Answers1

10

Essential singularities have the property of being both small and large near the singularity (see the Picard Theorems or, as Martin comments, the Casorati-Weierstrass Theorem). In this case, you've chosen a path to the singularity where its size is not big.

If instead, you chose to approach along the line $z=(1+i)t$ ($t\in\mathbb{R}$), then $$ \sin\left(\frac1{z^2}\right)=-i\sinh\left(\frac1{2t^2}\right) $$ which grows very fast near $z=0$, faster than any power of $\left|\frac1z\right|$.


For the modified question

Approach along the line $z=it$ ($t\in\mathbb{R}$), then $$ \sin\left(\frac1z\right)=-i\sinh\left(\frac1t\right) $$ which grows very fast near $z=0$, faster than any power of $\left|\frac1z\right|$.

robjohn
  • 345,667