# §31.11(i) Introduction

The formulas in this section are given in Svartholm (1939) and Erdélyi (1942b, 1944).

The series of Type I (§31.11(iii)) are useful since they represent the functions in large domains. Series of Type II (§31.11(iv)) are expansions in orthogonal polynomials, which are useful in calculations of normalization integrals for Heun functions; see Erdélyi (1944) and §31.9(i).

For other expansions see §31.16(ii).

# §31.11(ii) General Form

Let $w(z)$ be any Fuchs–Frobenius solution of Heun’s equation. Expand

 31.11.1 $w(z)=\sum_{j=0}^{\infty}c_{j}P_{j},$

where (§15.11(i))

 31.11.2 $P_{j}=\mathop{P\/}\nolimits\!\begin{Bmatrix}0&1&\infty&\\ 0&0&\lambda+j&z\\ 1-\gamma&1-\delta&\mu-j&\end{Bmatrix},$

with

 31.11.3 $\lambda+\mu=\gamma+\delta-1=\alpha+\beta-\epsilon.$ Symbols: $\gamma$: real or complex parameter, $\delta$: real or complex parameter, $\epsilon$: real or complex parameter, $\alpha$: real or complex parameter, $\beta$: real or complex parameter, $\lambda$ and $\mu$ Permalink: http://dlmf.nist.gov/31.11.E3 Encodings: TeX, pMML, png

The coefficients $c_{j}$ satisfy the equations

 31.11.4 $L_{0}c_{0}+M_{0}c_{1}=0,$
 31.11.5 $K_{j}c_{j-1}+L_{j}c_{j}+M_{j}c_{j+1}=0,$ $j=1,2,\dots$,

where

 31.11.6 $\displaystyle K_{j}$ $\displaystyle=-\frac{(j+\alpha-\mu-1)(j+\beta-\mu-1)(j+\gamma-\mu-1)(j+\lambda% -1)}{(2j+\lambda-\mu-1)(2j+\lambda-\mu-2)},$ 31.11.7 $\displaystyle L_{j}$ $\displaystyle=a(\lambda+j)(\mu-j)-q+\frac{(j+\alpha-\mu)(j+\beta-\mu)(j+\gamma% -\mu)(j+\lambda)}{(2j+\lambda-\mu)(2j+\lambda-\mu+1)}+\frac{(j-\alpha+\lambda)% (j-\beta+\lambda)(j-\gamma+\lambda)(j-\mu)}{(2j+\lambda-\mu)(2j+\lambda-\mu-1)},$ 31.11.8 $\displaystyle M_{j}$ $\displaystyle=-\frac{(j-\alpha+\lambda+1)(j-\beta+\lambda+1)(j-\gamma+\lambda+% 1)(j-\mu+1)}{(2j+\lambda-\mu+1)(2j+\lambda-\mu+2)}.$

$\lambda$, $\mu$ must also satisfy the condition

 31.11.9 $M_{-1}P_{-1}=0.$ Symbols: $P_{j}$ and $M_{j}$: coefficients Referenced by: §31.11(iii) Permalink: http://dlmf.nist.gov/31.11.E9 Encodings: TeX, pMML, png

# §31.11(iii) Type I

Here

 31.11.10 $\displaystyle\lambda$ $\displaystyle=\alpha,$ $\displaystyle\mu$ $\displaystyle=\beta-\epsilon,$

or

 31.11.11 $\displaystyle\lambda$ $\displaystyle=\beta,$ $\displaystyle\mu$ $\displaystyle=\alpha-\epsilon.$

Then condition (31.11.9) is satisfied.

Every Fuchs–Frobenius solution of Heun’s equation (31.2.1) can be represented by a series of Type I. For instance, choose (31.11.10). Then the Fuchs–Frobenius solution at $\infty$ belonging to the exponent $\alpha$ has the expansion (31.11.1) with

 31.11.12 $P_{j}=\frac{\mathop{\Gamma\/}\nolimits\!\left(\alpha+j\right)\mathop{\Gamma\/}% \nolimits\!\left(1-\gamma+\alpha+j\right)}{\mathop{\Gamma\/}\nolimits\!\left(1% +\alpha-\beta+\epsilon+2j\right)}z^{-\alpha-j}\*\mathop{{{}_{2}F_{1}}\/}% \nolimits\!\left({\alpha+j,1-\gamma+\alpha+j\atop 1+\alpha-\beta+\epsilon+2j};% \frac{1}{z}\right),$

and (31.11.1) converges outside the ellipse $\mathcal{E}$ in the $z$-plane with foci at 0, 1, and passing through the third finite singularity at $z=a$.

Every Heun function (§31.4) can be represented by a series of Type I convergent in the whole plane cut along a line joining the two singularities of the Heun function.

For example, consider the Heun function which is analytic at $z=a$ and has exponent $\alpha$ at $\infty$. The expansion (31.11.1) with (31.11.12) is convergent in the plane cut along the line joining the two singularities $z=0$ and $z=1$. In this case the accessory parameter $q$ is a root of the continued-fraction equation

 31.11.13 $\left(L_{0}/M_{0}\right)-\cfrac{K_{1}/M_{1}}{L_{1}/M_{1}-\cfrac{K_{2}/M_{2}}{L% _{2}/M_{2}-\cdots}}=0.$ Symbols: $K_{j}$: coefficients, $L_{j}$: coefficients and $M_{j}$: coefficients Referenced by: §31.18 Permalink: http://dlmf.nist.gov/31.11.E13 Encodings: TeX, pMML, png

The case $\alpha=-n$ for nonnegative integer $n$ corresponds to the Heun polynomial $\mathop{\mathit{Hp}_{n,m}\/}\nolimits\!\left(z\right)$.

The expansion (31.11.1) for a Heun function that is associated with any branch of (31.11.2)—other than a multiple of the right-hand side of (31.11.12)—is convergent inside the ellipse $\mathcal{E}$.

# §31.11(iv) Type II

Here one of the following four pairs of conditions is satisfied:

 31.11.14 $\displaystyle\lambda$ $\displaystyle=\gamma+\delta-1,$ $\displaystyle\mu$ $\displaystyle=0,$ 31.11.15 $\displaystyle\lambda$ $\displaystyle=\gamma,$ $\displaystyle\mu$ $\displaystyle=\delta-1,$ 31.11.16 $\displaystyle\lambda$ $\displaystyle=\delta,$ $\displaystyle\mu$ $\displaystyle=\gamma-1,$ 31.11.17 $\displaystyle\lambda$ $\displaystyle=1,$ $\displaystyle\mu$ $\displaystyle=\gamma+\delta-2.$

In each case $P_{j}$ can be expressed in terms of a Jacobi polynomial (§18.3). Such series diverge for Fuchs–Frobenius solutions. For Heun functions they are convergent inside the ellipse $\mathcal{E}$. Every Heun function can be represented by a series of Type II.

# §31.11(v) Doubly-Infinite Series

Schmidt (1979) gives expansions of path-multiplicative solutions (§31.6) in terms of doubly-infinite series of hypergeometric functions.