1 Statement of the formula

The Cauchy–Hadamard formula gives the radius of convergence of a power series in terms of the size of its coefficients. It is one of the standard results used to determine where a series converges absolutely and where it cannot converge.

1.1 Power series notation

A power series is commonly written as

\[ \sum_{n=0}^{\infty} a_n (x-x_0)^n, \]

where \(a_n\) are coefficients, \(x_0\) is the center, and \(x\) is the variable. In complex analysis, the same form is used with \(x\) replaced by a complex variable \(z\).

1.2 Radius of convergence

There is a number \(R \in [0,\infty]\), called the radius of convergence, such that the series converges absolutely when \(x-x_0<R\) and diverges when \(x-x_0>R\). The behavior at \(x-x_0=R\) must be checked separately.

1.3 Expression using limit superior

The formula states that

\[

\frac{1}{R}=\limsup_{n\to\infty}a_n^{1/n},

\]

with the conventions that \(1/0=\infty\) and \(1/\infty=0\). Thus, rapid growth of the coefficients leads to a smaller radius, while slow growth leads to a larger one.

1.4 Special cases

If \(\limsupa_n^{1/n}=0\), then \(R=\infty\), so the series converges for every value of the variable. If \(\limsupa_n^{1/n}=\infty\), then \(R=0\), so only the center point can lie in the region of convergence. If the limsup is a finite positive number, then the radius is its reciprocal.

2 Interpretation and meaning

The formula connects the coefficient growth of a series with the size of the domain on which the series defines a convergent function. It turns an analytic question into a numerical estimate on the coefficients.

2.1 Growth rate of coefficients

The quantity \(a_n^{1/n}\) measures the effective exponential growth of the coefficients. Even if the coefficients fluctuate, the limsup captures the largest asymptotic growth rate that persists along subsequences.

2.2 Convergence inside the disk or interval

When \(x-x_0\) is strictly smaller than \(R\), the factor \((x-x_0)^n\) decays quickly enough to overcome the coefficient growth. As a result, the series converges absolutely and typically uniformly on compact subsets of the open disk or interval.

2.3 Behavior on the boundary

At points where \(x-x_0=R\), the formula gives no direct conclusion. Some series converge there, some diverge, and others converge only conditionally. This boundary is often where the most delicate behavior appears.

3 Proofs

Several standard proofs of the formula rely on the root test and on basic properties of the limsup. These arguments are short but capture the essential mechanism behind the result.

3.1 Direct proof from the root test

Apply the root test to the terms \(a_n(x-x_0)^n\). Their nth roots are

\[

a_n^{1/n}x-x_0.

\]

If the limsup of this expression is less than 1, the series converges absolutely; if it is greater than 1, the series diverges. Solving the inequality in terms of \(x-x_0\) yields the stated radius.

3.2 Proof via limsup properties

Using the definition of limsup, one shows that for any \(\varepsilon>0\), the coefficients are eventually bounded by \((L+\varepsilon)^n\), where \(L=\limsupa_n^{1/n}\). This gives convergence whenever \(x-x_0<1/L\). A complementary argument shows divergence when \(x-x_0>1/L\).

3.3 Alternative derivations

The formula can also be derived from Cauchy estimates in complex analysis, from logarithmic growth considerations, or from the behavior of the generating function associated with a sequence. These approaches are useful in broader analytic contexts, though the root-test proof remains the most direct.

The Cauchy–Hadamard formula is closely connected with the standard tests for series convergence. It may be viewed as a specialized form of the root test adapted to power series.

4.1 Root test

The root test examines the limsup of the nth roots of the absolute values of the terms in a series. For a power series, this test naturally separates the variable part from the coefficient growth and leads immediately to the radius of convergence.

4.2 Ratio test

When the ratio \(a_{n+1}/a_n\) has a limit, the ratio test often gives the radius directly. In that case, the radius is the reciprocal of the limiting ratio, provided the limit exists and is positive.

4.3 Comparison with other criteria

Other tests, such as comparison or condensation arguments, may determine convergence in specific examples, but they do not provide a universal formula for the radius. The Cauchy–Hadamard result is more general because it depends only on the asymptotic growth of the coefficients.

5 Examples

Concrete examples illustrate how the formula translates coefficient behavior into a convergence domain.

5.1 Geometric series

For \(\sum_{n=0}^\infty x^n\), the coefficients are all 1, so \(a_n^{1/n}=1\). Hence \(R=1\), which matches the familiar convergence condition \(x<1\).

5.2 Exponential function series

For the exponential series \(\sum_{n=0}^\infty x^n/n!\), the coefficients decay very rapidly. Since \((1/n!)^{1/n}\to 0\), the radius is infinite, and the series converges for all real or complex values.

5.3 Factorial coefficient examples

For \(\sum_{n=0}^\infty n! \, x^n\), the coefficients grow faster than any exponential rate. The nth roots of \(n!\) tend to infinity, so the radius of convergence is 0.

5.4 Series with sparse nonzero coefficients

If \(a_n\) is nonzero only on a sparse set of indices, the radius is determined by the largest asymptotic growth among the nonzero coefficients. A series may therefore have a large radius even when many coefficients vanish, as long as the remaining ones do not grow too quickly.

6 Extensions and generalizations

The Cauchy–Hadamard formula extends naturally beyond ordinary real power series. Its underlying idea remains the same: coefficient growth controls the size of the convergence region.

6.1 Complex power series

For power series in a complex variable, the formula describes the disk of absolute convergence centered at the expansion point. Within this disk, the series defines a holomorphic function.

6.2 Multivariable power series

For series in several variables, analogous results describe regions of convergence in terms of multi-index coefficients. The geometry becomes more intricate, since the convergence set is no longer determined by a single radius.

6.3 Formal power series context

In algebra, a formal power series is studied without regard to convergence. Even there, coefficient growth may later be interpreted analytically when the formal series is assigned a variable in a numeric or complex setting.

7 Applications

The formula is a basic tool wherever power series are used to represent functions or solve equations. It gives a quick first estimate of where a series expansion is meaningful.

7.1 Analytic function theory

In complex analysis, the radius of convergence describes the largest disk on which a holomorphic function can be recovered from its Taylor series at a point. This makes the formula central to local analytic study.

7.2 Solving differential equations by series

Many differential equations are solved by assuming a power series form for the solution. The Cauchy–Hadamard formula then helps determine the range over which the resulting series is expected to converge.

7.3 Estimating domains of analyticity

When coefficients are known or estimated, the formula provides a practical way to bound the domain of analyticity. This is especially useful in asymptotic analysis, approximation theory, and generating-function methods.

8 History and naming

The result is named after Augustin-Louis Cauchy and Jacques Hadamard, who contributed to the development of convergence theory and the study of power series.

8.1 Augustin-Louis Cauchy

Cauchy helped establish rigorous foundations for convergence and complex-function theory. His work laid much of the groundwork for systematic results on infinite series.

8.2 Jacques Hadamard

Hadamard refined and popularized many tools in analysis, including results on series growth and convergence. His name is attached to the formula in recognition of later contributions to its modern form.

8.3 Development of the convergence theory

The Cauchy–Hadamard formula emerged from broader efforts to classify infinite series by their asymptotic behavior. It became a standard theorem because it gives a clean, general answer to a fundamental question: how far does a power series converge?