You are on page 1of 24

Series solutions of confluent Heun equations in terms of incomplete Gamma-functions

A.M. Ishkhanyan

Institute for Physical Research, NAS of Armenia, 0203 Ashtarak, Armenia

Abstract. We present a simple systematic algorithm for construction of expansions of the
solutions of ordinary differential equations with rational coefficients in terms of mathematical
functions having indefinite integral representation. The approach employs an auxiliary
equation involving only the derivatives of a solution of the equation under consideration.
Using power-series expansions of the solutions of this auxiliary equation, we construct
several expansions of the four confluent Heun equations' solutions in terms of the incomplete
Gamma-functions. In the cases of single- and double-confluent Heun equations the
coefficients of the expansions obey four-term recurrence relations, while for the bi- and tri-
confluent Heun equations the recurrence relations in general involve five terms. Other
expansions for which the expansion coefficients obey recurrence relations involving more
terms are also possible. The particular cases when these relations reduce to ones involving
less number of terms are identified. The conditions for deriving closed-form finite-sum
solutions via right-hand side termination of the constructed series are discussed.

PACS numbers: 02.30.Gp Special functions, 02.30.Hq Ordinary differential equations,
02.30.Mv Approximations and expansions

MSC numbers: 33E30 Other functions coming from differential, difference and integral
equations, 34B30 Special equations (Mathieu, Hill, Bessel, etc.), 30Bxx Series expansions

Keywords: Confluent Heun equations, special functions, series expansions, multi-term
recurrence relations

Introduction
The expansions of the solutions of complicated differential equations in terms of
simpler mathematical functions have played a notable role in developing the theory of these
equations, as well as in their numerous applications in mathematics, physics, engineering,
chemistry, biology, etc. Among many well-appreciated examples are the general Heun
equation [1] and its four confluent cases [2-4], which are currently widely applied in
numerous branches of contemporary physics ranging from atomic, molecular and optical
physics to condensed state physics, nuclear physics, astrophysics, cosmology and general
relativity [2-4]. The solutions of the Heun equations generalize many known mathematical
functions including the hypergeometric, Bessel, Mathieu, Lamé functions, etc., and for this
reason the special functions emerging from these equations are supposed to constitute a part
of the next generation of mathematical functions [2-4].

1

A useful step in developing the theory of the Heun equations was the development of
expansions of the solutions of these equations in terms of functions more advanced than mere
powers. Such expansions have been initiated by Svartholm [5], Erdélyi [6] and continued by
Schmidt [7], who used the Gauss hypergeometric functions to construct solutions of the
general Heun equation having a wider convergence region as compared with those suggested
by the power-series solutions. This useful technique has later been developed to cover many
other equations, including the confluent equations of the Heun class [8-25]. Besides, not only
the Gauss hypergeometric functions, but a number of other expansion functions have been
applied, e.g., the Kummer and Tricomi confluent hypergeometric functions [11-14], Coulomb
wave functions [15-17], Bessel and Hankel functions [18], incomplete Beta functions [19-
22], Goursat and Appell generalized hypergeometric functions [23-25], and other known
special functions.
In the present paper, we consider a simple systematic algorithm for construction of
expansions of the solutions u (z ) of an ordinary differential equation L [u ( z )]  0 in terms of
certain functions  n (z ) , which have indefinite integral representation. The basic idea is the

following. We consider an auxiliary equation involving only the derivatives d k u / d z k of a
solution u (z ) of the equation L [u ]  0 , but not the function u (z ) itself. Next, we introduce a
new dependent variable v(z ) – weighted first derivative of u (z ) according to
1 du
v( z )  (1)
( z ) d z
with an auxiliary function  (z ) . Considering now an expansion of v(z ) in terms of certain
expansion functions wn :

v( z )   c n wn ( z ) , (2)
n 0

and term-by-term integrating Eq. (1) multiplied by  (z ) , we arrive at an expansion

u ( z )  C 0   c n n ( z ) , C 0  const , (3)
n 0

where  n ( z )   ( z ) wn ( z ) dz . (4)

There are many choices for the weight function  ( z ) along with the expansion
functions wn ( z ) , for which these integrals are expressed in terms of known special functions.

For instance, this takes place if  ( z )  z  and all wn ( z ) belong to the class of (generalized)

2

B (a. it is appropriate to make some notes of general character. is that when passing to an equation for the new dependent variable v( z ) (i. when constructing the expansions. and if wn ( z ) is a Gauss hypergeometric function.g. if wn ( z ) is a Kummer confluent hypergeometric function. z )   e t t a 1d t . Before proceeding to particular developments. rather. one arrives at expansions in terms of Goursat generalized hypergeometric functions 2 F2 [23]. after integration as described above. Furthermore. b. rather it refers to a solution which is defined with accuracy up to an arbitrary constant factor. a useful observation which turns out to be rather helpful in practice. it should be appropriately chosen in order to achieve a consistent solution (see the details below). such expansions. the simplest possibility is suggested by the Frobenius-type power-series expansion of v(z ) in the neighborhood of point z1 of the complex z -plane:  v( z )   c n ( z  z1 )   n .. Since these extra singularities turn out to be regular. the functions  n (z ) turn into the incomplete Gamma- or incomplete Beta-functions. In the present paper. the result is an expansion in terms of the Clausen generalized hypergeometric functions 3 F2 [23]. [26. lead to new expansion functions 3 . e. respectively. Examples of application of this approach for construction of expansions of the solutions of the general and bi-confluent Heun equations in terms of the incomplete Beta-functions are presented in [25. z )   t a 1 (1  t ) b 1 dt . weighted first derivative of u ( z ) according to Eq. one can employ the Frobenius solutions of v -equations in the neighborhood of these singularities. 22]. below we assume that the notation u ( z ) does not refer to a solution normalized to unity at the origin. we note that the integration constant C0 appearing in the expansion (3) is not arbitrary. as well as on the allowed variation region of the variable z (see. b . if ( z ) is chosen as e  z or z  . Finally. In turn. (5) n 0 Then. (1)) one usually encounters an equation which possesses additional singularities as compared with the starting equation for the function u (z ) . (6) z 0 with appropriate restrictions imposed on the involved parameters a.hypergeometric functions. we apply this approach to construct solutions of the four confluent Heun equations in terms of the incomplete Gamma-functions.27]). First. In this way. These functions are defined as  z (a..e. Probably.

(8). in the context of isomonodromic families of Fuchsian equations parametrized by Painlevé VI solutions. here we discuss several relevant equations involving only the derivatives of u (z ) (not the function u (z ) itself). (8). in this way one arrives at expansions in terms of certain generalized hypergeometric functions [28]. not used before. Furthermore. and that the behavior of this function at infinity now interferes with that of the function f (z ) . a direct one for the first derivative v  du / dz . in particular. It is also worth to note that further application of the same procedure. (7) by g ( z ) . 4 . (3). if any. are involved in functions f and g .. regardless of the multiplicity of the roots of the equation g ( z )  0 . Extra singularity of the equation for the derivative of a solution of a Heun equation Let we have a second-order linear ordinary differential equation with polynomial or rational coefficients f (z ) and g (z ) : d 2u du 2  f ( z)  g ( z) u  0 . since it is the logarithmic derivative g z / g  d (ln g ) / d z that immediately stands for the appearance of new singularities at finite points of the complex z -plane.30-33]. this time to Eq. is readily derived by dividing Eq. It is further understood that the additional singularities originate from the zeros of the function g (z ) . (7) dz dz Singularities of this equation. The resultant equation reads  g   g  v zz   f  z  v z   f z  f z  g  v  0 .for the series Eq. It is seen that because of the term g z / g new singularities may emerge in this equation as compared with Eq. The first of such equations.g. (7). The appearance of an extra singularity in equations obeyed by the functions involving the derivatives of the solutions of the Heun equations have been noticed in several cases [23- 25. e. For instance. 2. then differentiating the equation and further multiplying it by g (z ) . As it was already mentioned above. in terms of the Appell hypergeometric function F1 of two variables of the first kind [29]. it is understood that for a polynomial or rational g ( z ) the new singularities all are regular. may or may not reveal further new singularities depending on the properties of the coefficients of Eq. Consider the appearance of these extra regular singular points in more detail. (8)  g   g  where (and hereafter) the subscripts denote differentiation of corresponding order with respect to the indicated variable.

(9). namely. 1. (9). as P( z )  z ( z  1) . It is interesting that. with a cubic polynomial P( z )  z ( z  1)( z  a ) ]. it can be written. all regular. z  q /( ) . (10) presents another general Heun equation with altered parameters [24]. that is from the root of the above function g (z ) which in the case of the general Heun equation is a rational function with a linear in z numerator. a and z   . z  s0  s1 z . this root. (8) reads   1  1  1   ( z ) v zz      v z  v  0. as long as the Fuchsian differential equations having at the utmost three regular singular points are discussed. Four confluent modifications of the Heun equation arise via coalescence of some of the singular points of the general Heun equation [2-4]. (10)  z z 1 z  a   z  q  z ( z  1)( z  a)(  z  q) where  ( z ) is a quadratic polynomial:  ( z )  (1   )(1   ) z (  z  2q)  (q 2  q (  a  a   )  a   ) . Eq. The first such equation is the general Heun equation having four regular singular points. we then have the double- 5 . without loss of generality. we will have the single-confluent Heun equation (SCHE). If the polynomial has a twice multiple root. are z  0. Compared with Eq. as a result. however. For the derivative v ( z )  u z of a solution of this equation. (11) As we see. it has an extra regular singularity located at the (finite or infinite) point z  q /( ) .  z  q . It is indeed understood that this singularity originates from the numerator of the coefficient of the last term in Eq. by shifting the origin and scaling. (12) where P(z ) is at most a second-degree polynomial: P ( z )  p 0  p1 z  p 2 z 2 [note that the general Heun equation itself is also written in this form. when q  0 . These equations can be written as P( z ) u zz  (   z   z 2 ) u z  ( z  q) u  0 . the described procedure does not lead to new singularities. In its canonical form. coincides with one of the already existing singular points of Eq. is changed if Fuchsian equations having more regular singularities are considered. and. namely. this equation is written as [1]      z  q u zz     uz  u  0. however. q   . The situation. If this polynomial has two distinct roots. (10) is a Fuchsian differential equation having in general five singular points. and it is for this reason that only in these four particular cases Eq. (9)  z z 1 z  a  z ( z  1)( z  a ) The singular points of this equation. q  a and   0 . Eq. the latter can be put at the origin: P ( z )  z 2 . Only in four particular cases. (9).

(12).  and F ( z )  (1  z )  z  . DCHE.  . so that this singularity originates from the accessory parameter q of the predecessor general Heun equation (9).. (13) where    . if P ( z ) is a constant. (12) if it is divided by P ( z ) : all the four equations have a constant term in the coefficient of the first-derivative term. The result is written as u ( z )  C1  C 2  e   z F ( z )dz . 2  const . As it is already mentioned above.confluent Heun equation (DCHE).  . again. we arrive at the bi-confluent Heun equation (BCHE). It is immediately seen from Eq. it is useful to examine the solutions of the truncated confluent Heun equations without the last term: P( z ) u zz  (   z   z 2 ) u z  0 . BCHE. the irregularity of the infinity is pronounced in the second term of the confluent Heun equations. respectively.  for SCHE. Thus. e  z / 2 z 3 / 3 2 2 z (14) for SCHE. Then. e  z / 2  . TCHE. (5). we are immediately being led to an incomplete Gamma-function expansion for u (z ) : 6 .  .e. by putting its root at the origin. where the regular singularity of the general Heun equation (9) at z  a has been merged into the regular singularity at infinity. BCHE. It is shown that the characteristic exponents of this singularity are 0 and 2. the four confluent Heun equations also lead to an extra regular singular point if an equation for the derivative is considered. 3. The infinity becomes irregularity already in the (single) confluent Heun equation. If P ( z ) is a linear function. we conclude that the extra singularity always originates from the last term of the Heun equations. (8) that in all cases this extra singularity is located at the point z 0  q /  . while the extra singularity of the equation for the derivative is exceptionally due to the last term of the equations. It is then understood that if we consider a power-series expansion for the function F (z ) of the form of Eq. we have the tri-confluent Heun equation (TCHE). e  / z z  . C1. In notations of Eq. putting P( z )  z . and TCHE. respectively. Heuristic observations leading to incomplete Gamma-function expansions A common feature of the four confluent Heun equations is that for all of them the infinity is an irregular singular point. Finally. DCHE. A result of the irregular nature of the singularity of z   is seen in Eq. this constant is given as    . Thus.  . i. to get an insight into how the irregularity of the infinity acts.

Their method concentrates on these singularities. the equation for a weighted derivative function has at least one regular and one irregular singular points.  ( z  z1 )  .e. The presented approach consists of two steps. Then. (12) as a perturbation.  z   1  n. which allows a "natural" expansion in terms of the incomplete Gamma-functions as follows [27]: a  (b  a ) n 1 F1 a. see examples below). the situation is somewhat similar to the approach by Svartholm [5] and Schmidt [7]. another regular singular point. Hence. if available. the structure of the equations does not matter much. In this sense. we pass to an equation obeyed by the weighted derivative function v  e  z u z and construct a power-series solution of this equation in the neighborhood of its extra (regular) singular point z 0  q /  (or. i. This is sufficient for construction of the incomplete Gamma-function expansions. the method of variation of constants. it is readily seen that we again come to similar expansions in terms of the incomplete Gamma-functions. irrespective of the equations’ structure outside the mentioned region containing the two regular singularities. Because of this. Applying. who proposed a regular method for construction of series solutions in terms of the Gauss hypergeometric functions for equations having only two regular singular points in a certain region of the complex z -plane. (15)  n 0   n 0   One more step is to look what happens if one treats the last term of Eq. the approach is more general and can be applied to more general equations having an irregular singularity at infinity and potent to produce an additional regular singularity in the equations obeyed by the derivatives of their solutions. z   1   1  n. b.      c e   z1  u  C1  C 2   c n  e  z ( z  z1 )   n dz   C1  C 2   n1 n   (1  n   .. Below we show that such expansions are constructed for any set of the involved parameters. there are several cases indicating the incomplete Gamma-function expansions of the solutions of the four confluent Heun equations. 0 . for example. (16) b n0 (b  1) n n! Thus. the integration produces an incomplete Gamma-function expansion. The above approach necessarily employs the fact that for the confluent Heun equations the infinity is an irregular singular point. A few remarks are now appropriate. First. A further argument supplementing our speculations concerning the appearance of incomplete Gamma-function expansions of the solutions for all the four confluent Heun equations comes from the observation that for some particular specifications of the involved parameters all these equations have particular solutions written in terms of the Kummer confluent hypergeometric function. 7 .

27] (a. Discussion In Appendices 1-4 we present several incomplete Gamma-function expansions of the solutions of the four confluent Heun equations.1  a. for which the involved expansion functions differently depend on the summation index (see the examples below). (19) dz where (z ) is a linear. 5. according to the above- described approach. To construct the incomplete Gamma-function expansions. In a certain neighborhood of a point z  z1 of the complex z -plane the solution of this equation is expanded into a power series:  v( z )   c n( z1 ) ( z  z1 )   n . (20) n 0 where the coefficients cn( z1 ) are supposed to be zeros for negative n . The difference is well seen if one recalls the following representation of the incomplete Gamma-function through the Kummer confluent hypergeometric function [26] za (a. z )  e  zU (1  a. A similar approach has later been applied by Kurth and Schmidt. (17) a or the representation through the Tricomi confluent hypergeometric function [26. and the point z1 is a center. around which the expansion is developed. all these forms are readily obtained from those used here by straightforward simple specifications of the involved parameters. (18) A concluding remark is that in several cases it is possible to construct incomplete Gamma- function expansions. a  1. Incomplete Gamma-function expansions of types I and II. To this end.  z ) . z ) . z )  (a )  1 F1 ( a. it seems appropriate to make a comparison of the incomplete Gamma-functions used in the present paper with the confluent hypergeometric functions applied in [34]. Note that the forms of the Heun equations adopted here slightly differ from those used in various papers as well as in [2-4]. quadratic or cubic polynomial depending on the particular equation at hand. 8 . However. who developed a global representation of the solutions of second-order linear differential equations having an irregular singularity of rank one at infinity by series in terms of confluent hypergeometric functions [34]. then the difference becomes more pronounced. we first write a second-order differential equation for a weighted first derivative of the form du v ( z)  e ( z) .

Such expansions are referred to as of type II. as u 1 z q  P( z ) v z  (   z   z 2 ) v  (21) with P( z )  z ( z  1). (19) and subsequent integration results in an incomplete Gamma-function expansion. respectively. (20) into Eq. Acknowledgments This research has been conducted within the scope of the International Associated Laboratory (CNRS-France and SCS-Armenia) IRMAS. The latter point is readily understood if the solution of a confluent equation of the Heun class is rewritten. We refer to such expansions as of type I. If  (z ) is a quadratic or cubic polynomial. of at least two hypergeometric functions. 9 . We conclude by noting that the examples of expansions presented below show that incomplete Gamma-function expansions of the solutions of the confluent Heun equations are constructed for any set of the involved parameters with proviso that the parameter  standing for the irregular singularity of equations (12) is not zero. The recurrence relations for the coefficients of the expansions turn out to be of four- or five-term form. 295025 – IPERA. double-. 13RB-052). bi-. substitution of v(z ) of Eq. If this is a linear function of z . The research has received funding from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement No. s ( z  z1 )) depends on n as a n  a 0  n ( a 0  const ). z 2 . The work has been supported by the Armenian State Committee of Science (SCS Grant No. as a linear combination. the parameter a of the resultant incomplete Gamma-functions (a n . using Eq. If. and tri-confluent cases. then the expansion coefficients are explicitly calculated in terms of the ordinary Gamma-functions and then the solution of the v -equation is constructed in terms of the hypergeometric functions. with rational coefficients. it is possible to reduce the relations to two-term ones. In some particular cases the recurrence relations are reduced to ones involving less number of terms. respectively. In these cases the solution of the starting confluent Heun equation is finally written either as a quadrature. by a specification of the involved parameters. z . Now. the dependence becomes a n  a 0  n / 2 or a n  a 0  n / 3 . or. The particular form of the dependence of the involved incomplete Gamma-functions on the summation index n depends on  (z ) . equivalently. (12). 1 for single-.

(1.. if c N( 0)  0 and c N( 0)1  c N( 0) 2  c N( 0) 3  0 .9) For left-hand side termination of the series at n  0 should be S 0  0 . it is 10 .   0 or    .   (1.7) Qn  2q  (q   )    q    (   )  (q   )  (n   )   (n   ) 2 .8) Pn   (   (n       )) . Note also that in the latter case Eq.3) Note that this equation applies also to the limiting case   0 . so that just one additional restriction is imposed on the involved parameters. Then. even though  ( z ) then becomes a linear function of z . (1. Indeed. i. For non-zero  this is the case if    ( N       ) . (1.5) where S n  q (n   ) (n     ) . are then expected to impose two more restrictions on the parameters of the confluent Heun equation. i.6) Rn  q 2    q         (1   )  q (1       ) (n   )  (q   )(n   ) 2 .4) n 0  For z1  0 the coefficients of the expansion obey the four-term recurrence relation S n c n( 0)  Rn1c n( 0)1  Qn 2 c n( 0)2  Pn 3 c n( 0)3  0 . . (1.1)  z z 1  z ( z  1) The weighted first derivative v ( z )  e  z u z obeys the equation   1  1 1  ( z) v zz       v z  v  0. 2.. for a non-zero  we obtain an expansion of type I:  c n( z1 ) u ( z )  C 0  e  z1  1 n   (1  n   . c N( 0)1  0 and c N( 0) 2  0 .e. that in this particular case the two equations lead to a single condition.  ( z  z1 ) . (1. 2.2)  z z 1 z  z0  z ( z  1)( z  q) where z 0  q /  and  ( z ) is a quadratic polynomial: ( z )  q 2   (  z   z 2 (  (   )  )  q       z (2  (   )   .Appendix 1: (Single) Confluent Heun equation (SCHE) This equation has two regular singularities at z  0 and z  1 and an irregular singularity of rank 1 at z   :    z q u zz      uz  u  0. (1. (1. (1. (1. .2) presents another single-confluent Heun equation with altered parameters. however. It turns out. From the equation c N( 0) 3  0 we find that the termination is possible if PN  0 . (1.e. The series will terminate from the right-hand side if three successive coefficients vanish for some N  1.10) for some N  1.   0. The remaining two equations.

13) Qn    z 0 (   )       (1  2 z 0 ) (n   )  (n   ) 2 .11) is simplified to a three-term one.1).12) Rn    z 0 (   )  1  2 z 0    z 0 (   )  ( z 0  1) z 0  (n   )  (2 z 0  1)(n   ) 2 . ensure the termination of the series.11) multiplied by  2 ).6) is not reduced to a two-term one. the recurrence relation (1.1) vanishes. For z1  z 0 the coefficients c n( z0 ) again obey a four-term recurrence relation: S n c n( z0 )  Rn1c n( z01)  Qn 2 c n( z02)  Pn3 c n( z03)  0 .   if q  0 and   1. when the last term in Eq. (1. If z 0  0 or z 0  1 . (1. the characteristic exponent  should fulfill the condition R0  0 .   2 . (1. It is checked that apart from the trivial case   q  0 .14) Pn     (n       ) . (1. for termination.   if q   .e. For a consistent power series. (1. for left-hand side termination of the series at n  0 should be S 0  0 . (1. The series is terminated from the right-hand side for some N  1. In two particular cases. 1 . this is achieved only if   2 (the exponent   0 leads to a logarithmic solution). This leads to   1. we note that in all three-term cases further reductions to two-term relations are not possible for non-zero  . (1. i. This is readily understood if we recall that in this case Eq. 11 . if z 0   (technically. if the extra singularity of the v -equation coincides with one of the singularities of the starting confluent Heun equation (1.checked that some of the roots of the equation c N( 0)1 (q )  0 fulfill the second equation too. the remaining two equations ( c N( z0 1)  c N( z0 )2  0 ) should also be satisfied by an appropriate choice of the parameters of the confluent Heun equation. (1.  if c N( z0 1)  c N( z0 )2  0 and PN  0 .11) where S n  z 0 ( z 0  1) (n   )(n    2) . the recurrence relation (1.15) If z 0  0. the four-term recurrence relation (1. namely if   0 or if q  0 . Finally. in this limit the coefficient Pn identically vanishes if one considers Eq. i. hence.e.6) is reduced to a three-term one. Since then the higher-order coefficient S n identically vanishes for all n .2) presents another single-confluent Heun equation with altered parameters. For non-zero  the last condition is fulfilled if    (N      ) . We conclude this appendix by noting that another three-term reduction of the recurrence relation is achieved in the limit   0 . 2.16) Besides. (1.

hence.1) is written in terms of the confluent hypergeometric functions (this time. the regular singularities of the general Heun equation (9) at z  1 and z  a are merged separately into the other two to form irregular singularities at.5) Rn  q (q      )  a  (  q  q )(n   )  q (n   ) . one may fix one of the parameters  .7) Pn      (n     )  .8) If q  0 . z  s 0 z .4) where S n   q ( n   ) .2) z z z  z0  z ( z  q ) where z 0  q /  and  ( z ) is the quadratic polynomial   ( z )  q 2  q(     z (   2 ))   z 2 (    )  z    .  .  (2. Also. (2. each of rank 1 :    zq u zz   2     u z  u 0. The differential equation for the weighted first derivative v( z )  e z u z is written as    2 1  ( z) v zz   2     v zz  2 v  0.Appendix 2: Double-confluent Heun equation (DCHE) For the double-confluent Heun equation. (2. The series is terminated from the right-hand side for some N  1. Hence.  if c N( 0)1  c N( 0) 2  0 and PN  0 . Using a scaling transformation. (2. 2 (2. (2.  to an arbitrary value.1) vanishes so that the general solution is readily written in quadratures. (2. this equation has four irreducible parameters. (2.3) Accordingly. a trivial case is   q  0 when the last term in Eq.6) Qn   (q   )  2q  (  q )(n   )   (n   ) 2 . According to the general theory [2-4]. for left-hand side termination of the series at n  0 should be S 0  0 . the only choice is   0 .1) z z  z2 Note that in the case   0 this equation is readily reduced to the confluent hypergeometric equation by the simple change of the dependent variable u  z s w( z ) . Another known case when the solution of Eq. respectively.3) n 0  For z1  0 the coefficients cn( 0) of this expansion obey the four-term recurrence relation S n c n( 0)  Rn 1c n( 0)1  Qn 2 c n( 0)2  Pn3 c n( 0)3  0 . (2. the last condition is fulfilled if 12 .  ( z  z1 ) . For non-zero  . (2. of the argument  / z ) is the case     0 . z  0 and z   . for a non-zero  we obtain an expansion of type I:  c n( z1 ) u ( z )  C 0  e  z1  1 n   (1  n   . (2. we suppose   0 . 2.

hence.   2 (2. (2. ensure the termination of the series.15) and c N( z0 1)  c N( z0 )2  0 . the characteristic exponent  should fulfill the condition R0  0 . the characteristic exponent  should satisfy the equation S 0  0 . (2. i. for which the recurrence relation is further reduced to a two-term one. some of the roots of the equation c N( 0)1 (q )  0 fulfill the equation c N( 0) 2 (q )  0 too. (2. if the extra singularity of the v -equation coincides with the (irregular) singularity of the starting double-confluent Heun equation (2. (2. i.e. if q  0 . This leads to the only possible exponent   1 for this case. If z 0  0 .    (n     ) .10) to a three-term one is achieved in the limit   0 ( z 0  ) . Again. A consistent power series is achieved only for the greater exponent   2 . (2. For z1  z 0 the coefficients c n( z0 ) also obey a four-term recurrence relation: S n c n( z0 )  Rn1c n( z01)  Qn 2 c n( z02)  Pn3 c n( z03)  0 .1). namely. Finally. Apart from the degenerate case   0 . as in the case of SCHE. the series terminates from the right-hand side if    (n     ) .11)   Rn  (  z 0 )    z 0 (  2)  z 02 (n   )  2 z 0 (n   ) 2 . we note that in all above three-term cases further reductions to two-term relations are not possible for non-zero  . the four-term recurrence relation (2. which impose two more restrictions on the parameters of the double- confluent Heun equation (2. As in the case of the single-confluent Heun equation.4) is turned into a three-term one if   0 or q  0 .9) Interestingly. (2.10) where S n  z 02 (n   )(n    2) . Since then the higher-order coefficient S n identically vanishes.12) Qn  (  z 0 )  (  2 z 0  )(n   )  (n   ) 2 . it turns out that in this case also the remaining two equations lead to a single condition. hence. 13 .10) is simplified to a three-term one. For non-zero  . the recurrence relation (2. another reduction of the recurrence relation (2.   0 or   2 . there are no nontrivial cases.13) Pn     (n     ) .   0 .14) If z 0  0 .1).e.

Appendix 3: Bi-confluent Heun equation (BCHE) This equation has a regular singularity at z  0 and an irregular singularity of rank 2 at z   :   zq u zz       z  u z  u 0.1) vanishes so that the general solution is readily written in quadratures. Note that.6) 14 .  ( z  z1 ) . (3.31)). (3. (3.3) (    q (2  2     2 )) z  (q    (       2 )) z 2    z 3 .4) n 0  For both z1  0 and z1  z 0 the coefficients cn( z1 ) of this expansion obey a five-term recurrence relation: Tn c n( z1 )  S n1c n( z11)  Rn 2 c n( z12)  Qn3 c n( z13)  Pn 4 c n( z14)  0 . of the argument   z 2 / 2 ) is the case   q  0 (see below Eq.1) z  z As it is immediately seen.5) If z1  0 . if   0 this equation is always reduced to the confluent hypergeometric equation by the simple change of the dependent variable u  e  z w( s z ) . in a sense trivial is the case   q  0 when the last term in Eq. (3. The expansions of the first type are constructed if one considers the differential equation for the weighted first derivative v( z )  e  ( z  z1 ) u z :   1 1  ( z) v zz      2   z   v z  v  0. Accordingly. as desired. Finally. (3. this equation turns into the Kummer confluent hypergeometric equation if   0 and   0 . which can be specified. The solutions of the biconfluent Heun equation allow incomplete Gamma-function expansions of both type I and type II. here  is an arbitrary non-zero constant. (3. for a non-zero  we obtain an expansion of type I:  c n( z1 ) u ( z )  C 0  e  z1  1 n   (1  n   . In fact.2)  z z  z0  z ( z  q ) where z 0  q /  and  (z ) is the cubic polynomial  ( z )  q 2  q (      )     (3. afterwards. the coefficients of this relation are written as Tn  q (n   ) (n     ) . unlike the above single-confluent and double-confluent cases. (3. Another known case when the solution is written in terms of the confluent hypergeometric functions (this time.

Finally. 15 . (3.   . however.8) that another four-term reduction.11) S n   z 0  (   z 0   z 02 )    1  z 0 (  z 0  2 )  (n   )  (n   ) 2 . we note that since Pn does not depend on n .5) is reduced to one involving four successive terms if q  0 or    0 (   0 ). Again. involving non-successive terms. (3. (3. for the coefficients of the recurrence relation (3. 2 .5) is reduced to one involving four successive terms if q  0 or   0 (   0 ). In this case also the recurrence relation (3. in the case   0 it is possible to specify the constant  so that the relation is reduced to a three-term one: 2       0 . Eq.5). is achieved if z 0  2 /  and one puts   z 0    / 2 ( Rn  0 ).12) Rn  2 z 0   (   z 0   z 02 )  (  2 z 0   2 )(n   ) . The recurrence relation (3.8) Qn   2  q     (   )    (1  n   ) . (16). (3. at   0 the biconfluent Heun equation is reduced to the confluent hypergeometric equation. Since Pn does not depend on n .14) Pn    . (3. (3. (3.7) Rn  2q(   )   (q     q )  (   q  2 )(n   ) . involving non-successive terms. by choosing    / 2 the relation is reduced to one involving three (non-successive terms). however. It is seen from Eq.13) Qn       (  2 z 0   )   (n   ) . If z1  z 0 . in the case   0 the recurrence relation is reduced to a three-term one by choosing  so that 2       0 . if q    0 . Here the situation is much similar to the previous case. Besides. (3.9) Pn     . This is of course an expected result since. for nonzero    the series cannot terminate from the right-hand side.15) where generally   0.10) where generally   0. (3. (3. S n  q 2    q        (     2q  q )(n   )   (n   ) 2 . as it was mentioned above. Another four-term reduction of the recurrence relation (3.5) we have Tn  z 0 (n   ) (n    2) . for nonzero   0 the series cannot terminate from the right-hand side. one then recovers the above-mentioned expansion of the Kummer hypergeometric function in terms of the incomplete Gamma-functions. is achieved if one chooses q and  so that Rn  0 for all n . Similarly. Performing the calculations.

1). (3.17) Using a Frobenius power-series solution of this equation in the neighborhood of its regular singularity z  0 :  v( z )   c n( 0) z   n . (3.20) is reduced to one involving four successive terms if q   0 (   0 ). If q  0 and   0 simultaneously. The coefficients cn( 0) obey the five-term recurrence relation Tn c n( 0)  S n 1c n( 0)1  Rn 2 c n( 0)2  Qn 3 c n( 0)3  Pn 4 c n( 0)4  0 . (3. the recurrence relation becomes two-term: S n 1c n( 0)1  Qn3 c n( 0)3  0 .22) Rn  2q  q   (   q )(n   ) .  (3. Now we present the second type of incomplete Gamma-function expansions of the solutions of the bi-confluent Heun equation (3. Several such expansions can be suggested applying the differential equation obeyed by the weighted first derivative of the form v( z )  z  e  z  z 2 /2 u z . Below we present two examples.21) S n  q 2    q  (     q )(n   )   (n   ) 2 .27) 16 . (3.24) Pn     . (3. The equation for v(z ) reads   1   ( z) v zz     z   v z  v  0. Qn     (n     ) . (3.  1   and (   0 ) S n  (n    1)(n     ) .   . (3.26) with   0.16)  z z q z ( z  q) where  (z ) is the cubic polynomial  ( z )  q 2  q     (q   2q ) z  (q    (    )) z 2    z 3 .23) Qn   2  q     (n     ) . (3. (3.25) where generally   0.18) n 0 we get the expansion  c n( 0) 1 n    z 2  u( z)  C0    (1 n   ) / 2   . (3.20) where Tn  q (n   ) (n     ) . . The recurrence relation (3. The first example is constructed if v  e  z 2 /2 u z .19) n 0 2( / 2)  2 2  which applies if   0 . (3.

34) n 0 we get the expansion  c n( z0 )  1  n    ( z  z0 ) 2  u( z )  C0    (1 n   ) / 2  . (3. (3.35) n 0 2( / 2)  2 2  which applies if   0 and   0 (since z 0 should be a finite point of the complex plane).32)  z z  z0  z( z  z0 ) where      z 0 and  (z ) is the cubic polynomial  ( z )    z 0   z 02 (    )  z 0 (2  2   z 0   ) z  (     2 z 0   ) z 2   z 3 . (3. v (z ) is expressed in terms of the confluent hypergeometric functions:   1  3    z2    1   z2  v ( z )  C1 z1 F1    . 2  . (3.  (3. another four-term reduction of the recurrence relation (3. is achieved if one chooses    2  0 and    q  0 so that Rn  0 for all n . 17 . (3. the coefficients cn( 0) of the expansion (3. using a Frobenius solution of this equation in the neighborhood of its regular singular point z  z 0 :  v( z )   c n( z0 ) ( z  z 0 )   n . a k  . Using the Pochhammer symbol.29) 2 1   / 2k 1  (1     ) / 2k Correspondingly.20). . we would like to mention the three-term reduction achieved at     0 . Eq. . A second example of type II expansions can be constructed by means of employing the differential equation obeyed by the function v  e  ( z  z0 ) 2 /2 uz :   1 1  ( z) v zz       z0   z   v z  v  0.   C 2 z  1 F1   . .18) are explicitly calculated in terms of the Gamma functions.31)   2 2   2 2 2   2 2 Comparing this solution with the expansion (3.19). we recover the above mentioned expansion of the Kummer confluent hypergeometric function in terms of the incomplete Gamma-functions.Then.30)     This leads to the following general solution of the bi-confluent Heun equation for   q  0 :   1   z2    1  3    z 2  u ( z )  C1 1 F1  . the result reads 1  (1) n1 ( / 2) k (1       /  ) / 2 k cn  a( n1) / 2 .33) Now. Apart from the above obvious cases q   0 . Finally. (16). (3. involving non-successive terms. after some algebra. .   C 2 z 1 1 F1   .  . however.  2 2 2 2   2 2 .

in the three-term case one should impose one more restriction ( c N( z0 1)  0 ). we have the two- term particular case considered above. Appendix 4: Tri-confluent Heun equation Here. (3. (3. Since z 0  q /  .42) In addition.1) If   0 . (3. necessarily should hold Q N  0 for some N  1. below we suppose   0 . (3. We note also another known case when Eq. For this reason. If   0 and   0 .40) Pn   (   z 0 ) . The coefficients c n( z0 ) obey the five-term recurrence relation Tn c n( z0 )  S n1c n( z01)  Rn 2 c n( z02)  Qn 3 c n( z03)  Pn 4 c n( z04)  0 . Finally. and in the four-term case two such restrictions ( c N( z0 1)  0 and c N( z0 )2  0 ) should be fulfilled.38) Rn   ( n   ) .36) reduces to a four-term one if z 0 (   z 0 )  0 . (4. Since Pn does not depend on n . this equation is always reduced to simpler equations.36) where Tn  z 0 (n   ) (n    2) . (4. a of the general Heun equation have coalesced into that at infinity to form an irregular singularity of rank 3 at z   : u zz  (   z   z 2 ) u z  ( z  q) u  0 . the series cannot terminate from the right-hand side unless    z 0  0 (   0 is forbidden). that is    (N     ) . (3. and it is reduced to the Airy equation if     0 . it is reduced to the confluent hypergeometric equation. which may terminate from the right-hand side. 2 . . 1. For termination. 2. It is readily seen that the recurrence relation (3. we note that if z 02   2  0 but    z 0  0 ( Pn  0 ).39) Qn     (n       z 0   z 02 ) .41) where generally   0. however.   2 .1) is reduced to the confluent hypergeometric equation corresponding to the 18 . it is understood that further reduction to a three-term one is possible only if  and q vanish simultaneously. we have a series (ruled by a recurrence relation for the coefficients involving three or four terms). the singularities z  0.37) S n  (n    1)(n       z 0   z 02 ) . In this case.   0 . (3. (3.

applying a Frobenius solution of this equation in the neighborhood of its regular singular point z  z 0 :  v( z )   c n( z0 ) ( z  z 0 )   n . these transformations can be applied to achieve maximal simplifications. For the purposes of the present appendix.   z  z 0 . for a non-zero  .2)  z  z0  z  z0 where  (z ) is the cubic polynomial  ( z )  (     z 0   z 02 )(1   )  (       2 z 0  )  2   3 . in doing this. we use the general form (4. along with fixing the value of  .7) 19 . (4. Different specifications of two of the five parameters involved in Eq.1) were applied by different authors depending on the specific theoretical context or particular problem at hand (see.specification     q  0 . Consider the case   0 . According to the general theory.  ( z  z 0 ) .3) n 0 we derive an expansion of type I:  c n( z0 ) u( z)  C0   1 n   (1  n   .4) n 0  the coefficients cn( z0 ) of which obey the five-term recurrence relation Tn c n( z0 )  S n1c n( z01)  Rn 2 c n( z02)  Qn 3 c n( z03)  Pn 4 c n( z04)  0 . e. one may make the parameter q to adopt any zero or nonzero value. so that z 0  q /  is a finite point of the complex z -plane. We note that in the case of non-zero  one may fix  to any (non-zero) value by scaling z  s1 z . Depending on the particular context of interest.6) S n    z 0 (   z 0 )    (  z 0 (   z 0 )  2 ) (n   ) . However. (4. [2-4]).. as well as the case   q  0 when the last term of the equation vanishes so that the general solution is written in terms of quadratures. Alternatively. instead of being interested in  or  . (4. (4. one may achieve any desired (zero or non-zero) value either for  [4] or  [2-3] by shifting the origin: z  s1 z  s0 . (4. it is advantageous to have q  0 .1) since then other forms can be readily employed by straightforward specifications of the involved parameters. (4.5) with Tn  (n   ) (n    2) . (4.g. (4. the tri-confluent Heun equation has only three irreducible parameters. The differential equation for the function v( z )  e  ( z  z0 ) u z is written as  1  ( z) v zz     2   z   z 2   v z  v  0. in the case of nonzero  .4) Accordingly. Simultaneously.

(4. Rn  (  z 0 (   z 0 )   )   (  2 z 0 ) (n   ) . (4.13) n 0 2( / 2)  2 2  the coefficients cn( z0 ) of which. however.  (4. .12) and applying the Frobenius solution of this equation in the neighborhood of its regular singular point z  z 0 .9) Pn    . (4.16) S n  (  2 z 0  2 ) (n   ) .10) where one should choose the greater exponent   2 .11)  z  z0  z  z0 where  (z ) is the quartic polynomial  ( z )  (   z 0   z 02 )(1   2)  (   )  2 (  2 z 0   )  3   4 . Finally. we now obtain an expansion of type II:  cn( z0 )  1  n    ( z  z0 ) 2  u ( z )  C0    (1 n   ) / 2   .14) with Wn  ( n   ) ( n    2) . (4.18) Qn  (  2 z 0   )  . we note that since Pn does not depend on n . the recurrence relation necessarily involves five terms unless   2 z 0   0 . (4. (4. (4.17) Rn      (  z 0 (   z 0 ))   (n   ) .15) Tn  (  z 0 (   z 0 )) (n    1) .5) involves four terms (however.19) Pn    . The only case when the recurrence relation (4. (4. obey a six-term recurrence relation: Wn c n( z0 )  Tn1c n( z01)  S n 2 c n( z02)  Rn3 c n( z03)  Qn 4 c n( z04)  Pn5 c n( z05)  0 . (4. Considering the differential equation for the function v( z )  e  ( z  z0 ) 2 /2 uz :  1  ( z) v zz     2 z 0   (  2 ) z   z 2  v z  v  0. the series cannot terminate from the right-hand side.   z  z 0 . since Pn does not depend on n . non- successive) is achieved if   2 z 0  0 and one chooses      z 02 . hence. this series cannot terminate from the right-hand side.20) where one should put   2 . (4. since then the coefficient Rn vanishes for all n . (4. (4. As in the previous case. We note that the quantity   2z 0  is not changed by the scaling and shifting transformation z  s1 z  s0 .8) Qn      (  2 z 0 )   (n   ) . 20 .

(4. one can reduce the recurrence relation (4.14) to one involving five terms (however. (4. i.14) may be reduced at most to a five-term one (achieved when   0 . 21 . respectively.23) n  0 3( / 3) (1 n  ) / 3  3 3  the coefficients cn( z0 ) of which obey the six-term recurrence relation Wn c n( z0 )  Tn1c n( z01)  S n 2 c n( z02)  Rn3 c n( z03)  Qn 4 c n( z04)  Pn5 c n( z05)  0 . (4. non-successive) by putting either 2 or  equal to   2 z0 .  (4. As in the previous two cases. (4. we see that if   2 z 0   0 . thus forcing. S n  0 or Qn  0 .30) where generally   0. in general this series cannot terminate from the right-hand side. If in addition z0 is a root of the quadratic equation    z 0   z 02  0 . If. .28) Qn   (  z 0 (   z 0 )) . We conclude by discussing one more expansion constructed when the differential equation for the function v( z )  e  ( z  z0 ) 3 /3 uz :  1  ( z) v zz     2 z 02  (  4 z 0 ) z   z 2  v z  v  0. Tn  0 ).22) Applying the Frobenius solution of this equation in the neighborhood of the regular singular point z  z 0 .24) with Wn  ( n   ) ( n    2) . Inspecting the coefficients S n and Qn . 2 . since  cannot be zero. (4.25) Tn  (  z 0 (   z 0 )) (n    1) .27) Rn     ( n   ) .29) Pn   (  2 z 0 ) . (4.26) S n  (  2 z 0 ) (n   ) .   2 z 0   0 . (4.e. the recurrence relation (4. we obtain another expansion of type II:  c n( z0 )  1  n    ( z  z0 )3  u( z)  C0    . (4. since Pn does not depend on n .21)  z  z0  z  z0 where  (z ) is the quartic polynomial  ( z )  (   z 0   z 02 )(1    3)    2 (  2 z 0  )  4 . (4. however. the coefficient Tn also vanishes and the recurrence relation involves only four terms.   z  z 0 .

Ordnung um zwei einfache Singularitäten durch Reihen nach hypergeometrischen Funktionen". Kalnins and W. 4. Duke Math. Olver. E. c1( z0 )  c 2( z0 )  0 reads 1  2 cos(2 n / 3) ( / 3)   /(3 ) n / 3 n/3 c n( z0 )  .   3 3 . A. 2. S. As it is immediately seen. the auxiliary function v (z ) is expressed in terms of the confluent hypergeometric functions as  2  5  ( z  z0 )3    ( z  z 0 ) 3  v ( z )  C1 z 2 1 F1   .   0.33) 3 1 / 3n / 3 1n / 3 Correspondingly. 9.31) with Wn  (n   ) (n    2) . (4. 2010). 1 . "Hypergeometric expansions of Heun polynomials". Erdélyi. 413 (1939). "Die Lösung der linearen Differentialgleichung 2. New York. 1995). Math. "The Fuchsian equation of second order with four singularities". (4. F. 127 (1979). Math. Heun. Miller. R. 161 (1889). Ann. Reine Angew.W. References 1. D. if    z 02   q 2 /  2 and   2 z 0   2 q /  .34)  3 3 3 3   3  This leads to the following general solution of the tri-confluent Heun equation for    q 2 /  2 and   2 q /  :  2  ( z  z 0 ) 3   1 4  ( z  z 0 ) 3  u ( z )  C1 1 F1  . 3. Lozier.J.F. "Certain expansions of solutions of the Heun equation". 309. we have a two-term recurrence relation: Wn c n( z0 )  Rn 3 c n( z03)  0 .  . Q. 2000). The result for   0 and initial conditions c0( z0 )  1. Heun’s Differential Equations (Oxford University Press. 7.Yu. Lay. this solution is readily deduced from the known one for     q  0 .   C 2 1 F1    . Erdélyi.    C z  F 2 1 1   . . Jr. N. 5. . 33. Schmidt. "Zur Theorie der Riemann'schen Functionen Zweiter Ordnung mit Verzweigungspunkten". Math. Oxford. 22 . if   2 z 0   0 or    z 0   z 02  0 . 62 (1944). (4. 116. A.G.).W.35)  3 3 3   3 3 3 3  Of course. 2 . Boisvert. . the recurrence relation involves four terms. 6. If these two equations are satisfied simultaneously. Rn     (n   ) . 48 (1942). and C. 8. J. Svartholm. D. London. Ronveaux (ed.W. J.. NIST Handbook of Mathematical Functions (Cambridge University Press. J. Special functions (Oxford University Press.23) are then explicitly calculated in terms of the Gamma functions. Ann. A.). “Die Lösung der Fuchs'schen Differentialgleichung zweiter Ordnung durch Hypergeometrische Polynome”. Math. Clark (eds. i. (4.32) The coefficients cn( z0 ) of the expansion (4.e. Slavyanov and W. K. (Oxford) 15. (4.

S. (2014). Math.D. Kazakov and S. B. 085203 (2013).W.Yu. Fonctions hypergéométriques et hypersphériques: polynomes d'Hermite (Gauthier-Villars. L81 (2003).Ya. Math.B. 013503 (2007). 25. D. Math.S. Maksimov. Appell. Schmidt.M. Ishkhanyan. Figueiredo. L.J. “Incomplete Beta-function expansions of the solutions to the confluent Heun equation”. and A. Anal. Sci. Th. Slavyanov. 11. Phys. Contemp. 49.M. Ishkhanyan and K. A 37. Paris. Phys.D. 1966). A. Ishkhanyan. B. SIAM J. L591 (2001). Generalized hypergeometric functions (Cambridge University Press. Ishkhanyan. Phys. A. "Confluent Heun equations: convergence of solutions in series of coulomb wave functions". A 38. Heun Functions: Expansions in Series of Hypergeometric Functions. L. 9923 (2004). Physics (Armenian Ac.A. L.2871 [math-ph] 26. arXiv:1405.A.-A. ibid.Yu. El-Jaick.J. Figueiredo. 29. Leroy.nist. M. El-Jaick and B.M. 27. A 46. L. 1926). M. Kuznetsov. 1086 (1986). 19. "Expansions of the solutions of the biconfluent Heun equation in terms of incomplete beta and gamma functions". J. B.B. Pashayan-Leroy. 20. A. IJDEA 13(4). confluent and biconfluent Heun equations and their connection with Abel equations". AIP Advances 4. Melikdzanian. New York. Suominen. 17. "A limit of the confluent Heun equation and the Schrödinger equation for an inverted potential and for an electric dipole". 087132 (2014). V. Cheb-Terrab. L. "Expansions of the solutions to the confluent Heun equation in terms of the Kummer confluent hypergeometric functions". "New solutions of Heun's general equation". and the two-center problem in molecular quantum mechanics". El-Jaick and B.D. Cambridge. Leaver. "Euler integral symmetries for a deformed Heun 23 . I. 24. C.B. J. J. 10. Kampe de Feriet. J. Math. A 36. "The Appell hypergeometric expansions of the solutions of the general Heun equation". A. T. 28. 14. Sokhoyan. Sleeman and V.D. Stegun (Eds..B. 17.M. arXiv:1311. arXiv:1410.SIAM J. 34.J.J. Figueiredo. "New solutions to the confluent Heun equation and quasiexact solvability". 1238 (1986). 48. 15. 18. Phys. A. A.J. 13.S. Shahnazaryan et al. J. "Generalized spheroidal wave equation and limiting cases". "New hypergeometric series solutions to the general Heun equation". Phys. 083508 (2008). "On the global representation of the solutions of second order linear differential equations having an irregular singularity of rank one in by series in terms of confluent hypergeometric functions".11. Abramowitz. Ishkhanyan. Armenian J. http://dlmf. Ishkhanyan. "New relations for the derivative of the confluent Heun function". Y. (2014). Math. Figueiredo. 123511 (2009). P. Phys. 16. 50. Phys. M.8575 [math-ph] 23. 1971).B. 5(3). Slater.D..7677 [math-ph] (2013).B. Anal. R. 30. 21. New York. 1972). Agrest. 22. Theory of incomplete cylindrical functions and their applications (Springer. El-Jaick and B. "Solutions for the general.gov/31. E. 27. "A series solution of the general Heun equation in terms of incomplete Beta functions". "Solutions for confluent and double-confluent Heun equations". J.A. J. 22.M. M. 9. J. Phys. Handbook of Mathematical Functions (Dover.). Phys. Ishkhanyan and A. Figueiredo.). 1450 (1991). L491 (2005). 202 (2014). J.D. Kurth and D. E. 40(6). 1 (2005). 12.M.M. Math. "Solutions to a generalized spheroidal wave equation: Teukolsky's equations in general relativity. Manukyan et al.J. 146 (2012).

Constr. 32. Schmidt. Theor. SIAM J. "Relations between linear equations and Painlevé’s equations". Th. 040 (2009). 342. 1086 (1986). Math. Takemura. "Middle Convolution and Heun’s Equation".equation and symmetries of the Painlevé PVI equation". Math. 24 . "On the global representation of the solutions of second-order linear differential equations having an irregular singularity of rank one in by series in terms of confluent hypergeometric functions". Kurth and D. Anal. 33. S. 75 (2014). Slavyanov. K.Y. "Integral representation of solutions to Fuchsian system and Heun’s equation" J. Phys. 17. 39. Anal. Math. Appl. 155. 34. 31. Takemura. K. Approx. 52 (2008). SIGMA 5. 722 (2008).