logo

Zeros and Poles of Meromorphic Functions 📂Complex Anaylsis

Zeros and Poles of Meromorphic Functions

Theorem 1

In a simple closed path C\mathscr{C}, let an analytic function ff have ZZ zeros and PP poles inside C\mathscr{C}, and on C\mathscr{C}, let it be f(z)0f(z) \ne 0. Then, 12πiCf(z)f(z)dz=ZP {{1} \over {2 \pi i }} \int_{\mathscr{C}} {{f ' (z)} \over {f(z)}} dz = Z - P


  • ZZ and PP are the sums including multiplicities.

Explanation

Analytic Number Theory

If the function f:CCf : \mathbb{C} \to \mathbb{C} has no poles, it would become a formula for counting the number of solutions to the equation f(z)=0f(z) = 0. It’s noteworthy that integers have appeared. At first glance, complex analysis might seem completely unrelated to number theory, yet it’s extensively used in reality. In number theory, there’s even a term ’elementary proof’ referring to proofs that don’t utilize complex analysis.

Log Trick

It might seem pointless to use a function shaped like this, but it’s the derivative form of logf\log f. Such a form appears more often in mathematics than one might think, across various fields.

Derivative of an Arithmetic Function: The derivative or derivative form of an arithmetic function ff is defined as follows. f(n):=f(n)logn,nN f ' (n) := f(n) \log n \qquad , n \in \mathbb{N}

Proof

Strategy: Without loss of generality, assume that there is only one nnth-order zero and one mmth-order pole. Geometrically, they might seem like a single point, but algebraically, they are considered to have multiplicities nn and mm, i.e., multiple points clustered together. Derive a formula for these two points and generalize it by partitioning C\mathscr{C} for multiple points.


Part 1. Zeros

Let α\alpha be a ff’s nnth-order zero. Then, for some function gg, it can be expressed as f(z)=(zα)ng(z)f(z) = (z- \alpha)^{n} g(z), and its derivative f(z)=n(nα)n1g(z)+(zα)ng(z) f ' (z) = n \left( n - \alpha \right)^{n-1} g(z) + (z- \alpha)^{n} g ' (z) when divided by f(z)f(z), can be represented as a logarithmic derivative trick form. f(z)f(z)=nzα+g(z)g(z) {{f ' (z)} \over {f(z)}} = {{n} \over {z - \alpha}} + {{g ' (z)} \over {g(z)}} Since f(z)0f(z) \ne 0 on C\mathscr{C}, logf\log f is analytic over the whole of C\mathscr{C}. Furthermore, since an analytic function’s derivative is also analytic, f/ff ' / f is also analytic, and similarly, since f(z)0f(z) \ne 0 on C\mathscr{C}, 1/(zα)1 / (z - \alpha) is also analytic. Therefore, g(z)g(z)=nzαf(z)f(z)\displaystyle {{g ' (z)} \over {g(z)}} = {{n} \over {z - \alpha}} - {{f ' (z)} \over {f(z)}} is also analytic at α\alpha.

  • Cauchy’s Integral Formula f(α)=12πiCf(z)zαdz f(\alpha) = {{1} \over {2 \pi i }} \int_{\mathscr{C}} {{f(z)} \over { z - \alpha }} dz
  • Cauchy’s Theorem: Cf(z)dz=0 \int_{\mathscr{C}} f(z) dz = 0

In some neighborhood Nα\mathcal{N}_{\alpha} of α\alpha that does not include any other zeros or poles of ff, according to Cauchy’s Integral Formula, Nαnzαdz=2nπi \int_{\mathcal{N}_{\alpha}} {{n} \over {z - \alpha}} dz = 2 n \pi i and by Cauchy’s Theorem, Nαg(z)g(z)dz=0 \int_{\mathcal{N}_{\alpha}} {{g ' (z)} \over {g(z)}} dz = 0 thus obtaining the following result. Nαf(z)f(z)dz=0+2nπi \int_{\mathcal{N}_{\alpha}} {{f ' (z)} \over {f(z)}} dz = 0 + 2 n \pi i


Part 2. Poles

The case for poles is similar to that for zeros. Let β\beta be a ff’s mmth-order pole. Then, for some function hh, it can be expressed as f(z)=h(z)(zβ)m\displaystyle f(z) = {{h(z)} \over {(z - \beta)^m}}.

Meanwhile, near β\beta, f(z)f(z)=h(z)h(z)mzβ\displaystyle {{f ' (z)} \over {f(z)}} = {{h ' (z)} \over {h(z)}} - {{m} \over {z - \beta}} is analytic, so by Cauchy’s Theorem, Nβh(z)h(z)dz=0 \int_{\mathcal{N}_{\beta}} {{h ' (z)} \over {h(z)}} dz = 0 and by Cauchy’s Integral Formula, Nβmzαdz=2mπi \int_{\mathcal{N}_{\beta}} {{m} \over {z - \alpha}} dz = 2 m \pi i


Part 3. Conclusion

By the Generalized Contraction Lemma, repeating the calculations finitely for all zeros and poles yields the following. 12πiCf(z)f(z)dz=nimj=ZP {{1} \over {2 \pi i }} \int_{\mathscr{C}} {{f ' (z)} \over {f(z)}} dz = \sum n_{i} - \sum m_{j} = Z - P


  1. Osborne (1999). Complex variables and their applications: p98. ↩︎