MS-C1300 Complex Analysis

5 Power series

5.1 Uniform convergence

Definition 5.1 Uniform convergence [Palka1991, Sec. VII.1.1]

Let \((f_n)_{n \in \mathbb {N}}\) be a sequence of functions \(f_n \colon X \to \mathbb {C}\), and let \(f \colon X \to \mathbb {C}\) also be a such function. We say that the sequence \((f_n)_{n \in \mathbb {N}}\) converges uniformly to \(f\) (on \(X\)) if for every \(\varepsilon {\gt} 0\) there exists an \(N \in \mathbb {N}\) such that for all \(n \ge N\) we have

\begin{align*} \big| f_n(x) - f(x) \big| {\lt} \varepsilon \qquad \text{ for all } x \in X . \end{align*}
Lemma 5.2 Cauchy criterion for uniform convergence [Palka1991, Thm VII.1.2]

Let \(f_n \colon A \to \mathbb {C}\), \(n \in \mathbb {N}\), be complex-valued functions defined on the same set \(A\). Then the sequence \((f_n)_{n \in \mathbb {N}}\) converges uniformly on \(A\) if and only if for every \(\varepsilon {\gt} 0\) there exists an \(N \in \mathbb {N}\) such that for all \(m,n \ge N\) and all \(z \in A\) we have \(|f_n(z) - f_m(z)| {\lt} \varepsilon \).

(When \((f_n)_{n \in \mathbb {N}}\) satisfies the condition above, it could be called a uniform Cauchy sequence on \(A\).)

Proof

Lemma 5.3 Continuity is preserved in uniform limits [Palka1991, Thm VII.1.1]

Let \(X\) be a metric space (e.g., \(\mathbb {R}\), \(\mathbb {C}\), or a subset of these). If a sequence \((f_n)_{n \in \mathbb {N}}\) of continuous functions \(f_n \colon X \to \mathbb {C}\) converges uniformly to a function \(f \colon X \to \mathbb {C}\), then \(f\) is continuous.

Proof

See MS-C1541 Metric Spaces.

Lemma 5.4 Integration commutes with uniform limits [Palka1991, Thm VII.1.1]

If a sequence \((f_n)_{n \in \mathbb {N}}\) of continuous functions \(f_n \colon [a,b] \to \mathbb {C}\) on a closed interval \([a,b] \subset \mathbb {R}\) converges uniformly to a function \(f \colon [a,b] \to \mathbb {C}\), then we have

\begin{align*} \lim _{n \to \infty } \int _a^b f_n(x) \, \mathrm{d}x = \int _a^b f(x) \, \mathrm{d}x . \end{align*}
Proof

Corollary 5.5 Contour integration commutes with uniform limits [Palka1991, Thm VII.1.1]

If a sequence \((f_n)_{n \in \mathbb {N}}\) of continuous functions \(f_n \colon A \to \mathbb {C}\) on a subset \(A \subset \mathbb {C}\) of the complex plane converges uniformly to a function \(f \colon A \to \mathbb {C}\), then for any piecewise smooth path \(\gamma \) in \(A\) we have

\begin{align*} \lim _{n \to \infty } \int _\gamma f_n(z) \, \mathrm{d}z = \int _\gamma f(z) \, \mathrm{d}z . \end{align*}
Proof

This follows straightforwardly from the definition of contour integration and Lemma 5.4 above.

Definition 5.6 Convergence uniformly on compacts [Palka1991, Sec. VII.1.2]

Let \((f_n)_{n \in \mathbb {N}}\) be a sequence of functions \(f_n \colon A \to \mathbb {C}\) on \(A \subset \mathbb {C}\), and let \(f \colon A \to \mathbb {C}\) also be a such function. We say that the sequence \((f_n)_{n \in \mathbb {N}}\) converges uniformly on compacts (UOC) to \(f\) if for every compact subset \(K \subset A\) the restrictions \(f_n|_K \colon K \to \mathbb {C}\) converge uniformly on \(K\) to \(f|_K \colon K \to \mathbb {C}\). We then write

\begin{align*} f_n \xrightarrow {\mathrm{UOC}} f \qquad \text{ as } n \to \infty . \end{align*}

(This notion is also called by the alternative names locally uniform convergence and normal convergence.)

Lemma 5.7 UOC limit of analytic functions [Palka1991, Thm VII.3.1]

Suppose that functions \(f_1, f_2, \ldots \colon U \to \mathbb {C}\) are analytic functions on an open set \(U \subset \mathbb {C}\) and the sequence \((f_n)_{n \in \mathbb {N}}\) converges uniformly on compacts to a function \(f\). Then \(f\) is analytic on \(U\). Moreover, for any \(k \in \mathbb {N}\), the sequence \((f^{(k)}_n)_{n \in \mathbb {N}}\) of \(k\)th derivatives converges uniformly on compacts to \(f^{(k)}\).

Proof

5.2 Complex series

Definition 5.8 Complex series [Palka1991, Sec. VII.2.1]

Let \(z_1, z_2, z_3, \ldots \in \mathbb {C}\) be complex numbers. For \(N \in \mathbb {N}\), define the \(N\)th partial sum of these as

\begin{align*} S_N \; = \; \sum _{n=1}^N z_n \; = \; z_1 + z_2 + \cdots + z_N \, . \end{align*}

We say that the series \(\sum _{n=1}^\infty z_n\) converges if the sequence \((S_N)_{N \in \mathbb {N}}\) of partial sums has a limit, and we then denote

\begin{align*} \sum _{n=1}^\infty z_n = \lim _{N \to \infty } \; \sum _{n=1}^N z_n \, . \end{align*}

(Obvious modifications to the above definition are made if the terms’ indexing starts from \(n=0\) or some other index, and the notation is correspondingly changed to, e.g., \(\sum _{n=0}^\infty \).)

Lemma 5.9 Terms of a convergent series tend to zero

If a complex series \(\sum _{n=1}^\infty z_n\) converges, then we have

\begin{align*} \lim _{n \to \infty } z_n = 0 . \end{align*}
Proof

Lemma 5.10 Geometric series

The geometric series

\begin{align*} \sum _{n=0}^\infty z^n \; = \; 1 + z + z^2 + z^3 + \cdots , \end{align*}

with ratio \(z \in \mathbb {C}\) converges if and only if \(|z| {\lt} 1\). In that case its sum is

\begin{align*} \sum _{n=0}^\infty z^n \; = \; \frac{1}{1-z} . \end{align*}
Proof

Definition 5.11 Absolute convergence of a complex series

A complex series \(\sum _{n=1}^\infty z_n\) is said to converge absolutely if the series of absolute values \(\sum _{n=1}^\infty |z_n|\) converges.

Lemma 5.12 Absolute convergence implies convergence

If a complex series converges absolutely, then it converges.

Proof

Lemma 5.13 D’Alembert’s ratio test

Suppose that \(\sum _{n=1}^\infty z_n\) is a complex series such that the limit

\begin{align*} r = \lim _{n \to \infty } \frac{|z_{n+1}|}{|z_n|} \end{align*}

exists. Then:

  • If \(r {\lt} 1\), then the series \(\sum _{n=1}^\infty z_n\) converges absolutely.

  • If \(r {\gt} 1\), then the series \(\sum _{n=1}^\infty z_n\) does not converge.

Proof

5.3 Series of functions

Definition 5.14 Series of functions [Palka1991, Sec. VII.2.2]

Let \(f_1, f_2, f_3, \ldots \) be complex-valued functions on a set \(A\). For \(N \in \mathbb {N}\), define their \(N\)th partial sum function \(F_N \colon A \to \mathbb {C}\) by

\begin{align*} F_N(z) \; = \; \sum _{n=1}^N f_n(z) \; = \; f_1(z) + \cdots + f_N(z) \, . \end{align*}

We say that the function series \(\sum _{n=1}^\infty f_n\) converges pointwise if the sequence \(\big(F_N(z)\big)_{N \in \mathbb {N}}\) of partial sums has a limit at every \(z \in A\). We say that the function series \(\sum _{n=1}^\infty f_n\) converges uniformly on \(A\) if the sequence \((F_N)_{N \in \mathbb {N}}\) of partial sum functions converges uniformly on \(A\). We say that the function series \(\sum _{n=1}^\infty f_n\) converges uniformly on compacts if the sequence \((F_N)_{N \in \mathbb {N}}\) of partial sum functions converges uniformly on compacts.

The limit function is then denoted by \(\sum _{n=1}^\infty f_n\).

(Obvious modifications to the above are made if the terms’ indexing starts from \(n=0\) or some other index, and the notation is correspondingly changed to, e.g., \(\sum _{n=0}^\infty \).)

Lemma 5.15 Weierstrass M-test [Palka1991, Thm VII.2.2]

Suppose that \(M_1,M_2,\ldots \ge 0\) are nonnegative numbers such that the series \(\sum _{n=1}^\infty M_n\) converges. Suppose also that for each \(n \in \mathbb {N}\), \(f_n \colon X \to \mathbb {C}\) is a function on \(X\) such that \(|f_n(x)| \le M_n\) for all \(x \in X\). Then the series \(\sum _{n=1}^\infty f_n\) converges absolutely and uniformly on \(X\).

Proof

Lemma 5.16 Series of analytic functions [Palka1991, Thm VII.3.2]

Suppose that functions \(f_1, f_2, \ldots \colon U \to \mathbb {C}\) are analytic functions on an open set \(U \subset \mathbb {C}\) such that the series \(\sum _{n=1}^\infty f_n\) converges uniformly on compacts to a function \(f \colon U \to \mathbb {C}\). Then \(f\) is analytic on \(U\). Moreover, for any \(k \in \mathbb {N}\), the series \(\sum _{n=1}^\infty f^{(k)}_n\) of \(k\)th derivatives converges uniformly on compacts to \(f^{(k)}\).

Proof

5.4 Power series

Definition 5.17 Power series [Palka1991, Sec. VII.3.3]

Let \(z_0 \in \mathbb {C}\) be a point in the complex plane and let \(a_0,a_1,a_2\ldots \in \mathbb {C}\) be coefficients. A function series of the form

\begin{align*} \sum _{n=0}^\infty a_n \, (z-z_0)^n = a_0 + a_1 \, (z - z_0) + a_2 \, (z - z_0)^2 + \cdots \end{align*}

is called a power series centered at \(z_0\).

Lemma 5.18 Abel’s theorem

If a power series

\begin{align*} \sum _{n=0}^\infty a_n \, (z-z_0)^n \end{align*}

converges at \(z = w \in \mathbb {C}\), then it converges absolutely for all \(z \in \mathbb {C}\) such that \(|z-z_0| {\lt} |w-z_0|\).

Proof

Corollary 5.19 Abel’s theorem in the contrapositive

If a power series

\begin{align*} \sum _{n=0}^\infty a_n \, (z-z_0)^n \end{align*}

does not converge at \(z = w \in \mathbb {C}\), then it does not converge at any \(z \in \mathbb {C}\) such that \(|z-z_0| {\gt} |w-z_0|\).

Proof

Definition 5.20 Radius of convergence

The radius of convergence of a power series

\begin{align*} \sum _{n=0}^\infty a_n \, (z-z_0)^n \end{align*}

is defined as

\begin{align*} R \; := \; \sup \bigg\{ |z-z_0| \; \bigg| \; \sum _{n=0}^\infty a_n \, (z-z_0)^n \, \text{ converges} \bigg\} . \end{align*}

From Lemma 5.18 and Corollary 5.19 it follows that the power series \(\sum _{n=0}^\infty a_n \, (z-z_0)^n\) converges for all \(z \in \mathbb {C}\) such that \(|z-z_0| {\lt} R\) and diverges for all \(z \in \mathbb {C}\) such that \(|z-z_0| {\gt} R\). The disk \(\mathcal{B}(z_0; R)\) is called the disk of convergence of the power series \(\sum _{n=0}^\infty a_n \, (z-z_0)^n\).

(If \(R=+\infty \), we interpret \(\mathcal{B}(z_0; R) = \mathbb {C}\).)

Lemma 5.21 D’Alembert’s ratio test for the radius of convergence

Suppose that for the coefficients of a power series

\begin{align*} \sum _{n=0}^\infty a_n \, (z-z_0)^n \end{align*}

the limit

\begin{align*} \rho = \lim _{n \to \infty } \frac{|a_n|}{|a_{n+1}|} \end{align*}

exists. Then the radius of convergence \(R\) of the power series is \(R = \rho \).

Proof

Theorem 5.22 Hadamard’s formula for the radius of convergence [Palka1991, Thm VII.3.3]

Let \(z_0 \in \mathbb {C}\) be a point in the complex plane and let \(a_0,a_1,a_2\ldots \in \mathbb {C}\) be coefficients. The radius of convergence of a power series

\begin{align*} \sum _{n=0}^\infty a_n \, (z-z_0)^n \end{align*}

is given by the formula

\begin{align*} R = \frac{1}{\limsup _{n \to \infty } \sqrt[n]{|a_n|}} , \end{align*}

with the conventions \(\frac{1}{+\infty }=0\) and \(\frac{1}{0} = +\infty \).

Proof

Lemma 5.23 Analyticity of power series [Palka1991, Thm VII.3.3]

Let \(z_0 \in \mathbb {C}\) be a point in the complex plane and let \(a_0,a_1,a_2\ldots \in \mathbb {C}\) be coefficients. Suppose that the power series

\begin{align*} f(z) = \sum _{n=0}^\infty a_n \, (z-z_0)^n \end{align*}

has radius of convergence \(R {\gt} 0\). Then it defines an analytic function \(f\) on the disk \(\mathcal{B}(z_0; R)\). The derivative of \(f\) is given by the power series

\begin{align*} f’(z) = \sum _{n=1}^\infty n \, a_n \, (z-z_0)^{n-1} . \end{align*}

Moreover, the coefficients \(a_k\) are related to the \(k\)th derivatives of \(f\) at \(z_0\) through the formula

\begin{align*} a_k = \frac{f^{(k)}(z_0)}{k!} . \end{align*}
Proof

Lemma 5.24 Uniqueness of power series representation

Suppose that two power series \(\sum _{n=0}^\infty a_n \, (z-z_0)^n\) and \(\sum _{n=0}^\infty b_n \, (z-z_0)^n\) converge in a disk \(\mathcal{B}(z_0; r)\) of radius \(r {\gt} 0\) and represent the same function

\begin{align*} \sum _{n=0}^\infty a_n \, (z-z_0)^n \; = \; \sum _{n=0}^\infty b_n \, (z-z_0)^n \qquad \text{ for } z \in \mathcal{B}(z_0; r) . \end{align*}

Then their coefficients must be equal: \(a_n = b_n\) for all \(n\).

Proof

5.5 Taylor series and local representation of analytic functions

Theorem 5.25 Taylor series of analytic functions [Palka1991, Thm VII.3.4]

Suppose that \(f \colon U \to \mathbb {C}\) is an analytic function on an open set \(U \subset \mathbb {C}\) which contains a disk \(\mathcal{B}(z_0; r) \subset U\). Then the function \(f\) can be represented in \(\mathcal{B}(z_0; r)\) as a power series

\begin{align*} f(z) = \sum _{n=0}^\infty \frac{f^{(n)}(z_0)}{n!} \, (z-z_0)^n . \end{align*}

Moreover, this is the unique power series centered at \(z_0\) that representats \(f\) in a neighborhood of \(z_0\).

Proof

Theorem 5.26 Equivalent characterizations of analyticity

Let \(f \colon U \to \mathbb {C}\) be a continuous function on an open set \(U \subset \mathbb {C}\). Then the following are equivalent:

  • \(f\) is analytic on \(U\);

  • for any \(z \in U\) there exists a neighborhood of \(z\) in which \(f\) has a primitive;

  • for any \(z \in U\) there exists a neighborhood of \(z\) in which \(f\) can be represented as a convergent power series.

Proof

Lemma 5.27 No vanishing of all derivatives at a point [Palka1991, Thm VIII.1.1]

Suppose that \(f \colon \mathcal{D}\to \mathbb {C}\) is an analytic function on a connected open set \(\mathcal{D}\subset \mathbb {C}\). If there exists a point \(z_0 \in \mathcal{D}\) such that \(f^{(n)}(z_0) = 0\) for all \(n \in \mathbb {N}\), then \(f\) is a constant function.

Proof

Theorem 5.28 Factor theorem for analytic functions [Palka1991, Thm VIII.1.2]

Suppose that \(f \colon \mathcal{D}\to \mathbb {C}\) is a non-constant analytic function on a connected open set \(\mathcal{D}\subset \mathbb {C}\), and \(z_0 \in \mathcal{D}\) is a point where \(f(z_0) = 0\). Then \(f\) can be uniquely represented as

\begin{align*} f(z) = (z - z_0)^m \, g(z) \qquad \text{ for } z \in \mathcal{D}, \end{align*}

where \(m \in \mathbb {N}\) and \(g \colon \mathcal{D}\to \mathbb {C}\) is an analytic function such that \(g(z_0) \ne 0\).

Proof

Corollary 5.29 Local representation of analytic functions [Palka1991, Cor VIII.1.3]

Suppose that \(f \colon \mathcal{D}\to \mathbb {C}\) is a non-constant analytic function on a connected open set \(\mathcal{D}\subset \mathbb {C}\). Then for any \(z_0 \in \mathcal{D}\), we can write \(f\) uniquely in the form

\begin{align*} f(z) = f(z_0) + (z - z_0)^m \, g(z) \qquad \text{ for } z \in \mathcal{D}, \end{align*}

where \(m \in \mathbb {N}\) and \(g \colon \mathcal{D}\to \mathbb {C}\) is an analytic function such that \(g(z_0) \ne 0\).

Proof

Apply Theorem 5.28 to the function \(z \mapsto f(z) - f(z_0)\).

Theorem 5.30 L’Hospital’s rule for analytic functions [Palka1991, Thm VIII.1.4]

Let \(f\) and \(g\) be functions that are analytic in a neighborhood of \(z_0\) such that \(f(z_0) = 0\) and \(g(z_0) = 0\). Then we have

\begin{align*} \lim _{z \to z_0} \frac{f(z)}{g(z)} = \lim _{z \to z_0} \frac{f'(z)}{g'(z)} , \end{align*}

understood in the sense that either both limits exist and are equal to each other, or else neither limit exists.

Proof

Theorem 5.31 Discrete mapping theorem [Palka1991, Thm VIII.1.5]

Suppose that \(f \colon \mathcal{D}\to \mathbb {C}\) is a non-constant analytic function on a connected open set \(\mathcal{D}\subset \mathbb {C}\). Then the set of zeros of \(f\) is discrete, i.e., for every \(z_0 \in \mathcal{D}\) such that \(f(z_0)=0\), there exists a \(r{\gt}0\) such that \(f(z) \ne 0\) for all \(z \in \mathcal{B}(z_0; r) \setminus \left\{ z_0 \right\} \).

Proof

Corollary 5.32 Principle of analytic continuation [Palka1991, Cor VIII.1.6]

Let \(f, g \colon \mathcal{D}\to \mathbb {C}\) be two analytic functions on a connected open set \(\mathcal{D}\subset \mathbb {C}\). If \(f(z) = g(z)\) for all \(z\) in some subset of \(\mathcal{D}\) which has an accumulation point in \(\mathcal{D}\), then we have \(f(z) = g(z)\) for all \(z \in \mathcal{D}\).

Proof

5.6 Laurent series

Definition 5.33 Doubly infinite series [Palka1991, Sec. VII.2.1]

A doubly infinite series of complex numbers is a series of the form

\begin{align*} \sum _{n=-\infty }^\infty z_n = \cdots + z_{-2} + z_{-1} + z_0 + z_1 + z_2 + \cdots , \end{align*}

where \(\ldots , z_{-2}, z_{-1}, z_0, z_1, z_2, \ldots \in \mathbb {C}\). We say that such a series converges to \(s \in \mathbb {C}\) if for all \(\varepsilon {\gt} 0\) there exists an \(N \in \mathbb {N}\) such that for all \(m_+ \ge N\) and \(m_- \le -N\) we have

\begin{align*} \Big| \sum _{n=m_-}^{m_+} z_n - s \Big| {\lt} \varepsilon . \end{align*}
Lemma 5.34 Convergence of doubly infinite series [Palka1991, Lem VII.2.1]

A doubly infinite series

\begin{align*} \sum _{n=-\infty }^\infty z_n = \cdots + z_{-2} + z_{-1} + z_0 + z_1 + z_2 + \cdots , \end{align*}

of complex numbers converges if and only if both the series \(\sum _{n=0}^\infty z_n\) and \(\sum _{n=1}^\infty z_{-n}\) converge.

Proof

Definition 5.35 Laurent series [Palka1991, Sec. VII.3.4]

A Laurent series centered at \(z_0 \in \mathbb {C}\) is a doubly infinite series of functions of the form

\begin{align*} z \mapsto & \; \sum _{n=-\infty }^\infty a_n (z - z_0)^n \\ = & \; \cdots + \frac{a_{-2}}{(z-z_0)^2} + \frac{a_{-1}}{z-z_0} + a_0 + a_1 (z-z_0) + a_2 (z-z_0)^2 + \cdots . \end{align*}
Lemma 5.36 Annulus of convergence of Laurent power series [Palka1991, Thm VII.3.5]

Consider a Laurent series

\begin{align*} f(z) = & \; \sum _{n=-\infty }^\infty a_n (z - z_0)^n . \end{align*}

Denote

\begin{align*} \rho _- = \limsup _{n \to \infty } \sqrt[n]{|a_{-n}|}, \qquad \rho _+ = \Big( \limsup _{n \to \infty } \sqrt[n]{|a_{n}|} \Big)^{-1} . \end{align*}

Then the series \(\sum _{n=-\infty }^\infty a_n (z - z_0)^n\) converges for all \(z\) is the annulus

\begin{align*} \mathcal{A}_{\rho _-,\rho _+}(z_0) := \Big\{ z \in \mathbb {C}\; \Big| \; \rho _- {\lt} |z-z_0| {\lt} \rho _+ \Big\} . \end{align*}

Moreover, the convergence is uniform on compact subsets of \(\mathcal{A}_{\rho _-,\rho _+}(z_0)\), and the series defines an analytic function \(f(z)\) on the annulus \(\mathcal{A}_{\rho _-,\rho _+}(z_0)\).

Proof

Theorem 5.37 Laurent series for analytic functions [Palka1991, Thm VII.3.6]

Suppose that \(f \colon U \to \mathbb {C}\) is an analytic function on an open set \(U \subset \mathbb {C}\) which contains an annulus

\begin{align*} \mathcal{A}_{r_1,r_2}(z_0) = \Big\{ z \in \mathbb {C}\; \Big| \; r_1 {\lt} |z - z_0| {\lt} r_2 \Big\} \end{align*}

for some \(z_0 \in \mathbb {C}\) and \(0 \le r_1 {\lt} r_2\). Then the function \(f\) can be uniquely represented in \(\mathcal{A}_{r_1,r_2}(z_0)\) as a series

\begin{align*} f(z) = \sum _{n=-\infty }^\infty a_n \, (z-z_0)^n , \end{align*}

where the coefficients \(a_n\), for \(n \in \mathbb {Z}\), are given by

\begin{align*} a_n = \frac{1}{2\pi \mathfrak {i}} \oint _{\partial \mathcal{B}(z_0; r)} \frac{f(z)}{(z - z_0)^{n+1}} \, \mathrm{d}z \qquad \text{ for any } r \in (r_1, r_2) . \end{align*}
Proof