Given behaviours at small and large values of a parameter, how can one deduce there is a singular (non-analytic) point?

35 Views Asked by At

A reoccurring statement in condensed matter physics is vaguely the following: Let $f:\mathbb{R}^n\times\mathbb{R}\rightarrow\mathbb{C}$ be some function. $\mathbb{R}^n$ are the spatial coordinates and $\mathbb{R}$ represents temperature. This function is shown to behave very differently at $T\rightarrow 0$ and $T\rightarrow \infty$. For example, it may hava a finite value for low $T$ but is identically $0$ at large enough $T$. Another example is that $f \sim \mid{r}\mid ^k$ for low $T$ and $f \sim e^{-m\mid r\mid}$ for large enough $T$. By $\sim$ I mean that the function converges to these functions at these limits.

Now, it is claimed that these types of functions can not be analytic with respect to $T$ for all $T$: There is a point in which the behavior is singular. My questions are the following:

  1. How does one prove this for the cases above?
  2. What type of arguments and from which fields are used for these types of problems? Particular theorems will be very welcomed
  3. A proof for similar cases will also be a good answer, as my goal is understanding the formal arguments