Abstract
The Nekrasov–Shatashvili limit of the SU(2) pure gauge (deformed) super Yang–Mills theory encodes the information about the spectrum of the Mathieu operator. On the other hand, the Mathieu equation emerges entirely within the frame of twodimensional conformal field theory ( CFT) as the classical limit of the null vector decoupling equation for some degenerate irregular block. Therefore, it seems to be possible to investigate the spectrum of the Mathieu operator employing the techniques of CFT. To exploit this strategy, a full correspondence between the Mathieu equation and its realization within CFT has to be established. In our previous paper [1], we have found that the expression of the Mathieu eigenvalue given in terms of the classical irregular block exactly coincides with the well known weak coupling expansion of this eigenvalue in the case in which the auxiliary parameter is the noninteger Floquet exponent. In the present work we verify that the formula for the corresponding eigenfunction obtained from the irregular block reproduces the socalled Mathieu exponent from which the noninteger order elliptic cosine and sine functions may be constructed. The derivation of the Mathieu equation within the formalism of CFT is based on conjectures concerning the asymptotic behaviour of irregular blocks in the classical limit. A proof of these hypotheses is sketched. Finally, we speculate on how it could be possible to use the methods of CFT in order to get from the irregular block the eigenvalues of the Mathieu operator in other regions of the coupling constant.
Classical limit of irregular blocks and Mathieu functions
Marcin Piatek
Institute of Physics and CASA*, University of Szczecin
ul. Wielkopolska 15, 70451 Szczecin, Poland
Institute of Theoretical Physics
University of Wrocław
pl. M. Borna 9, 50204 Wrocław, Poland
Bogoliubov Laboratory of Theoretical Physics,
Joint Institute for Nuclear Research, 141980 Dubna, Russia
Contents:
1 Introduction
In a last few years much attention was paid to the study of the connections among
twodimensional conformal field theory ( CFT), supersymmetric gauge
theories and integrable systems, cf. e.g. [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22].
The AGT correspondence states that the Liouville field theory (LFT) correlators on the Riemann surface with genus and punctures can be identified with the partition functions of a class of fourdimensional supersymmetric SU(2) quiver gauge theories:
(1.1) 
Let us recall that for a given pant decomposition of the Riemann surface , both sides of the equation above have an integral representation. Indeed, LFT correlators can be factorized according to the pattern given by the pant decomposition of and written as an integral over a continuous spectrum of the Liouville theory in which, for each pant decomposition , the integrand is built out of the holomorphic and the antiholomorphic Virasoro conformal blocks and multiplied by the DOZZ 3point functions [31, 32]. The Virasoro conformal block on depends on the following quantities: the cross ratios of the vertex operators locations denoted symbolically by , the external conformal weights , the intermediate conformal weights and the central charge .
On the other hand, the partition function can be written as the integral over the holomorphic times the antiholomorphic Nekrasov partition functions [33, 34]:
where is some appropriate measure. The Nekrasov partition function can be written as a product of three factors . The first two factors describe the contribution coming from perturbative calculations. Supersymmetry implies that there are contributions to only at the tree () and 1loop () levels. is the instanton contribution. The Nekrasov partition function depends on the set of parameters: , , , , . The components of are the gluing parameters associated with the pant decomposition of , where the are the complexified gauge couplings. The multiplet contains the mass parameters. Moreover, , where the ’s are the vacuum expectation values of the scalar fields in the vector multiplets. Finally, , represent the complex background parameters.
Comparing the integral representations of both sides of eq. (1.1) it is possible, thanks to AGT hypothesis, to identify separately in the holomorphic and antiholomorphic sectors the Virasoro conformal blocks on and the instanton sectors of the Nekrasov partition functions for the super Yang–Mills theories .
Soon after its discovery, the AGT conjecture has been extended to the conformal Toda/ SU(N) gauge theories correspondence [35, 36], and to the socalled ‘nonconformal’ cases [37, 38, 39] (see also [22, 40, 41, 42]), which will be of main interest in the present work.
The AGT correspondence works at the level of the quantum Liouville field theory. It is intriguing to ask, however, what happens if we proceed to the semiclassical limit of the Liouville correlation functions. This is the limit in which the central charge , the external and intermediate conformal weights tend to infinity in such a way that their ratios are fixed , cf. [32]. For the standard parametrization of the central charge , where and for heavy weights with , the classical limit corresponds to . It is commonly believed that in the classical limit the conformal blocks behave exponentially with respect to :
The function is known as the classical conformal block.
The AGT correspondence dictionary says that . Therefore, the semiclassical limit of the conformal blocks corresponds to the socalled Nekrasov–Shatashvili limit ( being kept finite) of the Nekrasov partition functions. In [25] it was observed that in the limit the Nekrasov partition functions have the following asymptotical behavior:
(1.2) 
where is the effective twisted superpotential of the corresponding twodimensional gauge theories restricted to the twodimensional background.
The twisted superpotentials play a pivotal role in the already mentioned Bethe/gauge correspondence [25, 26, 27] which maps supersymmetric vacua of the theories to Bethe states of quantum integrable systems (QIS’s). A result of that duality is that the twisted superpotentials are identified with the Yang–Yang (YY) functions [43] which describe the spectra of some QIS’s. Therefore, combining both the classical/Nekrasov–Shatashvili limit of the AGT duality and the Bethe/gauge correspondence one thus gets a triple correspondence which connects the classical blocks with the twisted superpotentials and then with the Yang–Yang functions (cf. Fig.1).
For example, the twisted superpotentials
for the SU(N) (pure gauge) and the
SU(N) SYM theories determine respectively the spectra
of the N–particle periodic Toda (pToda) and the
elliptic Calogero–Moser (eCM) models [25]. In the case of the SU(2) gauge group these
QIS’s are simply quantum–mechanical systems whose dynamics is described by some
Schrödinger equations. Concretely, for the 2–particle pToda
and eCM models these Schrödinger equations correspond to
the celebrated Mathieu and Lamé equations
with energy eigenvalues expressed in terms of the twisted superpotentials.
This correspondence can be used to investigate
nonperturbative effects in the Mathieu and Lamé quantum–mechanical systems, cf. [44].
On the other hand, the Mathieu and Lamé equations emerge entirely within the framework of
CFT as the classical limit of the null vector decoupling (NVD) equations for
the 3–point degenerate irregular block and for the 2–point block (projected 2–point function)
on the torus with one degenerate light operator [1, 15, 45].
It turns out that the classical irregular block and the classical
1–point block on the torus determine the spectra of the Mathieu and Lamé
operators in the same way as their gauge theory counterparts,
i.e.: and .
Therefore, it seems that there is a way to study the spectrum of the Mathieu and Lamé operators
using twodimensional conformal field theory methods.
The organization of the paper is as follows. In section 2 the necessary tools of CFT are introduced. In section 3 the simplest irregular blocks are defined and some of their properties are described. In particular, an exponentiation of the pure gauge irregular block within the classical limit is proved at the leading order. After that, the NVD equations for certain degenerate irregular blocks are derived. Section 4 is devoted to the derivation of the Mathieu equation within the formalism of CFT. The calculation presented there provides formulas for the Mathieu eigenvalue and the related eigenfunction in terms of the classical limit of irregular blocks. It is shown that these formulas reproduce the well known noninteger order weak coupling expansion of the Mathieu eigenvalue and the corresponding Mathieu function. In subsection 4.2 a factorization property of the degenerate irregular block with the light operator and its representation in the classical limit as a product of light and heavy parts is proved at the leading order. This factorization property is crucial for deriving the Mathieu equation. Section 5 contains our conclusions. In particular, the problems that are still open and the possible extensions of the present work are discussed.
2 Conformal blocks in the operator formalism
2.1 Chiral vertex operators
Starting from the Belavin–Polyakov–Zamolodchikov axioms [46], Moore and Seiberg
[47]
have constructed formalism of the socalled rational conformal field theories
(RCFT’s),

the operator algebra of local fields contains purely holomorphic subalgebra called chiral or vertex algebra;

the Hilbert space of states of the theory is a direct sum of irreducible representations of the algebra :
(2.1)
In RCFT’s the summation in (2.1) is over a discrete finite set. However, one can generalize and successfully apply the Moore–Seiberg formalism to the case of twodimensional conformal field theories with continuous spectrum, cf. e.g. [49, 50]. In such a case the direct sum in eq. (2.1) becomes a direct integral.
In any 2d CFT there exist at least two chiral fields, i.e., the identity operator and its descendant — the holomorphic component of the energymomentum tensor . Therefore, each chiral algebra contains as a subalgebra the Virasoro algebra ,
(2.2) 
In the Moore–Seiberg formalism the ‘physical’ fields of [46] are built out of more fundamental objects — the socalled chiral vertex operators (CVO’s). These are intertwining operators acting between representations of the vertex algebra. In the present paper we confine ourselves to the simplest case when and define CVO’s as operators acting between Verma modules.
Let be the free vector space generated by all vectors of the form
(2.3) 
where is an ordered () sequence of positive integers of the length , and is the highest weight vector:
(2.4) 
The graded representation of the Virasoro algebra determined on the space:
by the relations (2.2) and (2.4) is called the Verma module of the central charge and the highest weight . The dimension of the subspace of all homogeneous elements of degree is given by the number of partitions of (with the convention ). It is an eigenspace of with the eigenvalue .
On there exists the symmetric bilinear form uniquely defined by the relations
The Gram matrix of the form is blockdiagonal in the basis with blocks
In particular, one finds

: ,

: ,

: ,
The Verma module is irreducible if and only if the form is nondegenerate. The criterion for irreducibility is vanishing of the determinant of the Gram matrix, known as the Kac determinant, given by the formula [51, 52, 53, 54, 55, 56]:
(2.5) 
In the equation above is a constant and
The Kac determinant vanishes for
or
For these values of and the representations or are reducible.
The set of the degenerate conformal weights can be parametrized as follows
(2.6) 
where
Sometimes, it is also convenient to use the alternative parametrization:
(2.7) 
for which the central charge is given by with .
The nonzero element of degree is called a null vector if , and , . Hence, is the highest weight state which generates its own Verma module , which is a submodule of . One can prove that each submodule of the Verma module is generated by a null vector. Then, the module is irreducible if and only if it does not contain null vectors with positive degree.
For nondegenerate values of , i.e. for , there exists in the ‘dual’ basis whose elements are defined by the relation for all . The dual basis vectors have the following representation in the standard basis
where is the inverse of the Gram matrix .
Let be the Verma module with the highest weight state . The chiral vertex operator is the linear map
such that for all the operator
satisfies the following conditions
(2.8)  
(2.9)  
(2.10)  
(2.11)  
and
The commutation relation (2.8) defines the primary vertex operator corresponding to the highest weight state . Eqs. (2.9)–(2.11) characterize the decendant CVO’s.
2.2 The 3point block
For a given triple of conformal weights we define the trilinear map
induced by the matrix element of a single chiral vertex operator
The form is uniquely determined by the conditions (2.8)(2.11). In particular,

for eingenstates
^{8} one gets(2.12) 
for basis vectors , one finds
(2.13) where for a given partition ,
(2.14)
In terms of the trilinear form (3point block) one can spell out
an important result known as the
null vector decoupling theorem (Feigin–Fuchs [57]):
Let be chosen such that , , . Let us assume that

, (cf. parametrization (2.6)) and

the vector lies in the singular submodule generated by the null vector , i.e.:
Then, if and only if
satisfy the fusion rules , where and .
3 Quantum and classical zero flavor irregular blocks
3.1 Definition and basic properties
To begin with, let us consider the following (coherent) vector in the
Verma module discovered by D. Gaiotto in
[37] and constructed by A. Marshakov, A. Mironov and A. Morozov
in [38]:
(3.1)  
The summation in eq. (3.1) runs over all partitions or equivalently over their pictorial representations — Young diagrams. The symbol in eq. (3.1) denotes a single–row Young diagram, where the total number of boxes equals the number of columns , i.e. .
The zero flavor qunatum irregular block is defined as the inner product of the Gaiotto state [37, 38]:
(3.3)  
(3.4)  
(3.5) 
In fact, there are much more
Gaiotto’s states and therefore irregular blocks.
Let denotes a Riemann surface with genus and punctures. Let be the modular parameter of the 4punctured Riemann sphere . Then, the channel conformal block on is defined as the following formal expansion:
(3.6) 
where
(3.7)  
Let be the elliptic variable on the torus with modular parameter , then the conformal block on is given by the following formal series:
where
The irregular block (3.3) can be recovered from the conformal blocks on the torus and on the sphere in a properly defined decoupling limit of the external conformal weights [38, 39]. Indeed, employing the AGT inspired parametrization of the external weights , and the central charge , i.e.:
and introducing the dimensionless expansion parameter it is possible to prove the following limits [38, 39]:
(3.8) 
Due to the ‘nonconformal’ AGT relation, the irregular block can be expressed through the SU(2) pure gauge Nekrasov instanton partition function [37, 40, 22, 42]:
(3.9) 
The identity (3.9), which in particular is understood as term by term equality between the coefficients of the expansions of both sides, holds for
(3.10) 
where
(3.11) 
In [25] it was observed that in the limit the Nekrasov partition functions behave exponentially. In particular, for the instantonic sector we have
(3.12) 
Therefore, taking into account the AGT relation (3.9), the fact that and the Nekrasov–Shatashvili limit (3.12) of the instanton function, one can expect that the irregular block has the following exponential behavior in the limit :