An entire function is a function that is complex differentiable at every point of the complex plane. Entire functions form a central class in complex analysis: they have no singularities at any finite point and can be represented globally by convergent power series. Every polynomial is entire, as are the exponential, sine, and cosine functions. Entire functions that are not polynomials are called transcendental entire functions.
Definition and power-series characterization
A function is entire precisely when it is holomorphic on all of . Here, holomorphic means that the complex derivative
exists at each point, with the limit independent of the direction through which the complex increment approaches zero.
Complex differentiability is substantially more restrictive than differentiability of functions of a real variable. It implies the existence of derivatives of every order and a local convergent Taylor series. For an entire function, the Taylor series about any center converges throughout the plane:
Conversely, any power series
with infinite radius of convergence defines an entire function. The coefficient criterion is
The series converges uniformly on every compact subset of the plane, and it may be differentiated or integrated term by term there. This local uniform convergence is important: an entire function's Taylor series need not converge uniformly on the whole unbounded plane.
In this setting, “entire” is a global condition, whereas analytic describes a local property. A function may be analytic on a disk or a half-plane without being entire.
Examples and nonexamples
Every polynomial
is entire. The basic transcendental examples include the exponential function
and the trigonometric functions
All three series converge for every complex argument.
A formula with an apparent singularity may still define an entire function after that singularity is removed. For example,
is entire, as its power series is
By contrast, is not entire because it has a pole at zero. The function is smooth on the real axis but has complex poles at and . Thus, smoothness or real analyticity on the entire real line does not establish that a function extends to an entire function.
No single-valued holomorphic logarithm exists on all of , much less on the whole plane. Branches of logarithms and noninteger powers therefore are not entire.
Algebraic and convergence properties
Sums, products, and compositions of entire functions are entire. Their derivatives are entire as well. Every entire function also has an entire antiderivative, since the complex plane is simply connected. Consequently, its integral around every closed piecewise smooth contour is zero.
A quotient of two entire functions need not be entire: zeros of the denominator can produce poles. It is entire precisely when every apparent singularity at a denominator zero is removable, assuming the denominator is not identically zero.
An entire function with no zeros has an entire logarithm. More explicitly, if for every , then there is an entire such that
One construction uses an antiderivative of the entire function .
A sequence of entire functions that converges uniformly on every compact subset has an entire limit, and its derivatives converge locally uniformly to the corresponding derivatives of that limit. Mere pointwise convergence is not enough to guarantee an entire limit.
Rigidity and behavior at infinity
Entire functions are strongly constrained by their global behavior.
Liouville's theorem and polynomial growth
Liouville's theorem states that every bounded entire function is constant. In particular, an entire function cannot be both nonconstant and uniformly bounded over the complex plane.
The theorem follows from Cauchy's derivative estimates. If
then
If remains bounded as , the estimate forces every derivative of positive order at zero to vanish.
More generally, if
for a nonnegative integer , then is a polynomial of degree at most . Thus, a transcendental entire function cannot satisfy a global polynomial growth bound.
Liouville's theorem also gives a short proof of the fundamental theorem of algebra: if a nonconstant polynomial had no zeros, its reciprocal would be a bounded entire function.
The point at infinity
The behavior of an entire function at infinity can be studied through near :
- Infinity is a removable singularity exactly when is constant.
- Infinity is a pole exactly when is a nonconstant polynomial.
- Infinity is an essential singularity exactly when is transcendental entire.
According to Picard's little theorem, a nonconstant entire function takes every complex value, with at most one exception. The exponential function demonstrates that an exception is possible: it never equals zero.
Zeros and product representations
The zeros of a nonzero entire function are isolated. Each zero has a finite multiplicity , meaning that locally
An accumulation of zeros at a finite point forces the entire function to vanish identically. This is a consequence of the identity theorem.
Although finite accumulation is impossible, a transcendental entire function may have infinitely many zeros tending to infinity. The Weierstrass factorization theorem describes entire functions by products built from their zeros. For a nonzero entire function, a representation has the form
where are its nonzero zeros, repeated according to multiplicity, is entire, and
for , with . The nonnegative integers are chosen so that the product converges locally uniformly.
The exponential corrections are essential: simply multiplying the linear factors associated with infinitely many zeros need not produce a convergent product. A classical example of a successful product representation is
Conversely, any prescribed collection of zeros with positive integer multiplicities and no finite accumulation point can be realized by a nonzero entire function.
Growth, order, and type
The maximum modulus measures the largest value of on a circle of radius . For a nonconstant entire function, its order is
The order may be any nonnegative real number or infinity. It distinguishes major growth scales:
- Nonconstant polynomials have order .
- , , and have order .
- has order for every positive integer .
- has infinite order.
Order zero does not imply that a function is a polynomial; transcendental entire functions of order zero also exist.
For , the type relative to order is
Type provides a finer comparison among functions of the same order. Entire functions of exponential type satisfy a bound
for some constants .
Growth and zeros are linked. Hadamard's factorization theorem strengthens Weierstrass factorization for entire functions of finite order: suitable canonical factors can be chosen so that the remaining exponential factor is the exponential of a polynomial. This makes growth restrictions a powerful means of identifying functions from their zeros.
Historical development and applications
The theory developed from nineteenth-century work on complex integration, power series, and analytic continuation. Augustin-Louis Cauchy's integral methods supplied fundamental estimates, while Karl Weierstrass's work established a systematic approach through power series and infinite products. Later developments connected the distribution of zeros with growth, producing the classical theory of entire functions.
Entire functions occur in several major areas:
Number theory. The completed Riemann zeta function
extends to an entire function. Its zeros are the nontrivial zeros of , making entire-function methods relevant to the study of prime numbers.
Fourier analysis. A compactly supported integrable function has a Fourier transform that extends to an entire function. If its support lies in , then, under the convention
one has
Paley–Wiener theorems give precise relationships between support restrictions and entire extensions of controlled growth.
Differential equations and special functions. Many distinguished solutions of differential equations, including Airy functions and Bessel functions of integer order, are entire in their argument.
Approximation and computation. Taylor polynomials approximate an entire function uniformly on every fixed closed disk. Error estimates depend on its growth on larger circles. Infinite radius of convergence does not, by itself, guarantee fast numerical convergence or prevent loss of precision through cancellation.
These applications exploit both sides of entire-function theory: the flexibility to construct functions with prescribed zeros or approximation properties, and the rigidity imposed by global analyticity and growth bounds.