Abstract
The degeneracies of 1/4 BPS states with unit torsion in heterotic string theory compactified on a six torus are given in terms of the Fourier coefficients of the reciprocal of the Igusa cusp Siegel modular form \(\Phi _{10}\) of weight 10. We use the symplectic symmetries of the latter to construct a fine-grained Rademacher-type expansion which expresses these BPS degeneracies as a regularized sum over residues of the poles of \(1/\Phi _{10}\). The construction uses two distinct \(\textrm{SL}(2, {\mathbb {Z}})\) subgroups of \(\textrm{Sp}(2, {\mathbb {Z}})\) which encode multiplier systems, Kloosterman sums and Eichler integrals appearing therein. Additionally, it shows how the polar data are explicitly built from the Fourier coefficients of \(1/\eta ^{24}\) by means of a continued fraction structure.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Automorphic forms have played an uncannily effective role in the counting of BPS states in superstring theory wherein they arise as generating functions of the degeneracies of said states. More precisely, the integral degeneracies in these cases arise as coefficients of a Fourier expansion of these automorphic forms. Prominent examples include modular forms, Jacobi forms and Siegel modular forms in the context of BPS state counting in toroidal compactifications of type II/heterotic string theory to four dimensions, namely [1,2,3,4,5]
Modular and Jacobi forms, as shown in the first two examples, have automorphic properties under \(\textrm{SL}(2,{\mathbb {Z}})\) and are defined in the \(\sigma \) upper half plane. Both of these forms have a finite number of Fourier coefficients associated with a negative power of q. These coefficients are referred to as a polar coefficients. The modular symmetries of these functions are powerful enough to constrain the Fourier coefficients so that each non-polar coefficient can be exactly expressed in terms of the polar ones in a Rademacher expansion [6]. As an illustrative example, we present the well-known Rademacher expansion for \(1/\eta ^{24}\), which has been used to extract asymptotic degeneracies for 1/2 BPS states in toroidally compactified heterotic string theory [7],
We see that the polar coefficient \(d(-1)\) is sufficient to reconstruct the modular form.
In this note we will focus on the third of the above examples, the reciprocal of the Igusa cusp Siegel modular form \(\Phi _{10}\) of weight 10. It is defined on the Siegel upper half plane, with each of its Fourier coefficients defined in terms of three integers corresponding to the three variables defining the Siegel upper half plane. Further, each of its polar terms satisfies \(\Delta < 0\), where \(\Delta = 4 m n - \ell ^2\), and is determined in terms of the coefficients of \(1/\eta ^{24}\) and the continued fraction representation of \(\ell /2m\) [8,9,10]. Using the automorphic properties of \(\Phi _{10}\) with respect to \(\textrm{Sp}(2, {\mathbb {Z}})\),Footnote 1 we will demonstrate that this symmetry group constraints the Fourier coefficients of \(1/\Phi _{10}\) even more powerfully than in the \(\textrm{SL}(2, {\mathbb {Z}})\) case, resulting in a fine-grained Rademacher expansion that not only reconstructs each element of the infinite set of its Fourier coefficients that satisfy \(\Delta >0\) from the finite polar data, whose elements satisfy \(\Delta < 0\), but also encodes the continued fraction structure underlying the polar terms. Thus, the result of this paper can be summarised as follows:
We use the \(\textrm{Sp}(2,{\mathbb {Z}})\) symmetries of \(1/\Phi _{10}\) to construct a fine-grained Rademacher expansion which expresses its Fourier coefficients as a regularized sum over residues of its poles. The construction uses two distinct \(\textrm{SL}(2, {\mathbb {Z}})\) subgroups of \(\textrm{Sp}(2, {\mathbb {Z}})\) which encode multiplier systems, Kloosterman sums and Eichler integrals appearing therein. Additionally, it shows how the polar data are explicitly built from the Fourier coefficients of \(1/\eta ^{24}\) by means of a continued fraction structure.
The Rademacher expansion for \(d(m,n,\ell )\) was derived in [11] by viewing \(d(m,n,\ell )\) as Fourier coefficients of a Mock Jacobi form \(\psi _m^F\) [12, 13] using the mixed Mock Rademacher expansion developed in [14]. However, just as in the case of modular forms, the Rademacher expansion in [11] does not contain information about the explicit values of the polar coefficients of \(\psi _m^F\), as computed in [9, 10].
2 Set Up of the Calculation and Outline of the Paper
In order to define a Rademacher expansion for the reciprocal of the Igusa cusp \(\Phi _{10}\) of weight 10, we adopt the following sequential program elucidated below:
-
1.
Identify polar data in \(1/\Phi _{10}\). In contrast to the modular and Jacobi forms displayed above, the Siegel modular form \(\Phi _{10}\) has a countably infinite set of zeroes defined in the Siegel upper half plane by the loci [15],
$$\begin{aligned}{} & {} D(n_2, n_1,m_1,m_2, j): \quad n_2 (\rho \sigma - v^2) + j v + n_1 \sigma - m_1 \rho + m_2 = 0 \nonumber \\{} & {} m_1, n_1, m_2, n_2 \in {\mathbb {Z}}, \; j \in 2 {\mathbb {Z}} +1, \;\;\; \;\; m_1 n_1 + m_2 n_2 = \frac{1-j^2}{4}. \end{aligned}$$(2.1)It has been shown [8,9,10] that for \(\Delta < 0\),
$$\begin{aligned} d(m, n, \ell ) \equiv (-1)^{\ell +1} \, g(m, n, \ell ) \end{aligned}$$(2.2)can be constructed from the residues of \(1/\Phi _{10}\) corresponding to the linear poles (\(n_2=0\)), where \(g(m, n, \ell )\) are the Fourier coefficients of \(1/\Phi _{10}\) (cf. (1.1)). Therefore, the polar data in \(1/\Phi _{10}\) are determined by its linear pole residues. Viewing non-polar data as the complementary information to its polar counterpart, one is motivated to identify the non-polar data in \(1/\Phi _{10}\) as the contribution to \(g(m,n,\ell )\) with \(\Delta \ge 0\). Generically, these Fourier coefficients receive contributions from both linear and quadratic poles.
-
2.
Extracting non-polar data from \(d(m,n,\ell )\) with \(\Delta \ge 0\). From (1.1), we can define,
$$\begin{aligned} d(m, n, \ell ) = (-1)^{\ell +1} \, \int \limits _0^1 \textrm{d} \sigma _1 \int \limits _0^1 \textrm{d} v_1 \int \limits _0^1 \textrm{d} \rho _1 \, \frac{1}{\Phi _{10} (\rho , \sigma , v)} \, e^{-2 \pi i \left( m \rho + n \sigma + \ell v \right) } , \nonumber \\ \end{aligned}$$(2.3)where \(\rho = \rho _1 + i \rho _2, \sigma = \sigma _1 + i \sigma _2, v = v_1 + i v_2\), and where the imaginary parts of \(\rho \), \(\sigma \) and v are held fixed in the Siegel upper half plane, which is defined by the conditions \(\rho _2> 0, \, \sigma _2 > 0\) and \(\rho _2 \sigma _2 - v_2^2 > 0\). The \(n_2 =0\) poles of \(1/\Phi _{10}\) correspond to co-dimension one surfaces in the Siegel upper half plane. Therefore, in order to identify the non-polar data in terms of the Fourier coefficients, we first need to isolate a chamber in the Siegel upper half plane where we perform a Fourier expansion. Following [8] we will define this region, which is referred to as the \({{\mathcal {R}}}\)-chamber, by
$$\begin{aligned} \frac{\rho _2}{\sigma _2} \gg 1\;\;,\;\; \frac{v_2}{\sigma _2} = -\frac{ \ell }{2m} \;\;,\;\;\; \frac{\ell }{2m} \in [0, 1). \end{aligned}$$(2.4)We will write down the Rademacher expansion for the Fourier coefficients with \(\Delta \ge 0\) in the \({{\mathcal {R}}}\)-chamber.
-
3.
Setting up the Rademacher expansion. Viewing (2.3) as a \(\rho \) integral, we define:
$$\begin{aligned}{} & {} \textrm{d}(m,n,\ell )|_{\Delta \ge 0} \nonumber \\{} & {} \quad =(-1)^{\ell +1} \! \sum _{\begin{array}{c} D \\ n_2 \ne 0 \end{array}}\! \left( \int \limits _{\Gamma _{\sigma }(D) } \textrm{d}\sigma \!\int \limits _{\Gamma _v (D)} \!\textrm{d} v \, \textrm{Res} \!\left( \frac{1}{\Phi _{10} (\rho , \sigma , v)} \, e^{-2 \pi i \left( m \rho + n \sigma + \ell v \right) }\right) \right) \Big \vert _\textrm{reg},\nonumber \\ \end{aligned}$$(2.5)where ‘reg’ refers to a regularized sum of residues,
-
1.
\(\textrm{Res} f \) is the residue of f at an \(n_2 \ne 0\) pole D in the \(\rho \) plane;
-
2.
\(\Gamma _v(D) \) is the integration contour of v with \(v_1 \in [0,1)\), subject to the condition \(\frac{v_2}{\sigma _2}=-\frac{\ell }{2m}\). Further, the condition that D lies in the Siegel upper half plane constraints the range of \(v_1\);
-
3.
\(\Gamma _{\sigma }(D)\) lies in the projection of D in the \(\sigma \) upper half plane with the restriction that \(\sigma _1 \in [0,1)\).
This defines our starting point. Our objective is to express the right-hand side of (2.5) in terms of the polar data. This implies that the sum over residues of quadratic poles must be rewritten as one over linear pole contributions.
-
1.
In Sect. 3 we use the \(\textrm{Sp}(2, {\mathbb {Z}})\) symmetries to map each quadratic pole to the simplest linear pole \(v=0\). We will reparametrize the five numbers defining the quadratic poles by entries of the \(\textrm{Sp}(2, {\mathbb {Z}})\) matrix that performs this map and consequently rewrite the summation on the right-hand side of (2.5) as a sum over these matrix entries. In Sects. 4 and 5 we explicitly evaluate the summands by performing the v and \(\sigma \) integrals, respectively. We isolate multiplier systems, Kloosterman sums and error functions giving rise to Eichler integrals and to the continued fraction structure. These elements serve as building blocks for the Rademacher expansion. We identify a symmetry that is crucial to obtain the Rademacher expansion by enabling massive cancellations between various terms. The resulting expansion is given by (5.126) in the case when \(\Delta > 0\). Additionally, we write down the Rademacher expansion for the case when \(\Delta = m = \ell =0\) in (6.35). In Sect. 7 we conclude with comments on the implications of this expansion for defining exact quantum entropy functions from a gravity path integral. In the appendices we review and discuss various useful relations and calculations.
3 \(\textrm{Sp}(2, {\mathbb {Z}})\) Symmetries, Poles and Residues of \(1/\Phi _{10}\)
In this section, we map each quadratic pole of \(1/\Phi _{10}\) to the simplest linear pole \(v=0\). In order to so, we will use the \(\textrm{Sp}(2, {\mathbb {Z}})\) symmetries of \(\Phi _{10}\) to reparametrize the five numbers defining the quadratic poles by entries of the \(\textrm{Sp}(2, {\mathbb {Z}})\) matrix that performs this map. Subsequently, we rewrite the summation on the right-hand side of (2.5) as a sum over the entries of these mapping matrices.
3.1 \(\textrm{Sp}(2, {\mathbb {Z}})\) Symmetries of \(\Phi _{10}\)
The Igusa cusp form \(\Phi _{10}\) transforms as follows under \(\textrm{Sp}(2, {\mathbb {Z}})\) transformations,
where
The elements of \(\textrm{Sp}(2, {\mathbb {Z}})\) satisfy
\(\Phi _{10}\) is invariant under S-duality transformations, which consist of \(\textrm{SL}(2,{\mathbb {Z}})\) transformations
that operate on \(\Omega \) through the \(\textrm{Sp}(2, {\mathbb {Z}})\) transformations
Thus, under S-duality, we infer the transformation laws
They leave \(\Phi _{10}\) invariant,
Next, we consider a different set of \(\textrm{Sp}(2,{\mathbb {Z}})\) transformations, which we denote by \(\textrm{SL}(2, {\mathbb {Z}})_{\sigma }\). We denote its group elements by
They operate on \(\Omega \) through the \(\textrm{Sp}(2, {\mathbb {Z}})\) transformations
which results in the transformation laws
Under this set of transformations, \(\Phi _{10}\) transforms as
Furthermore, \(\Phi _{10}\) is also invariant under the integer shifts
which implies that it possesses a Fourier expansion.
Performing first an \(\textrm{SL}(2,{\mathbb {Z}})_\sigma \) transformation, then an S-duality transformation and finally an integer shift of v by \(-\Sigma \), \(\Phi _{10}\) changes as
where
The reciprocal of \(\Phi _{10}\) transforms accordingly.
3.2 Reparametrization of the Poles of \(1/\Phi _{10}\)
The Siegel modular form \(\Phi _{10}^{-1}(\rho ,\sigma ,v)\) has the following behaviour near \(v=0\),
Therefore, it has poles at all the \(\textrm{Sp}(2,{\mathbb {Z}})\) images of the divisor \(v=0\) in the Siegel upper half plane. The location of the poles is determined by
The poles are labelled by five integers, \((m_1,m_2,j,n_1,n_2)\) that satisfy the constraint
Since (3.16) and (3.17) are invariant under \((m_1,m_2,j,n_1,n_2) \mapsto (-m_1,- m_2,- j,- n_1,- n_2)\), we may restrict to \(n_2 \ge 0\) [16].
A specific parametrization of these poles was given in [17]. We will use an equivalent parametrization, as follows.
Proposition 3.1
Any pole (3.16) with \(n_2 \ge 1\) can be reparametrized in terms of nine integers as
where eight of the integers can be arranged into two \(\textrm{SL}(2,{\mathbb {Z}})\) matrices,
with entries that satisfy \(a>0,\, c<0, \, \gamma >0,\, \alpha \in {\mathbb {Z}}/\gamma {\mathbb {Z}}\), while \(\Sigma \in {\mathbb {Z}}\).
Moreover, any pole (3.16) with \(n_2 \ge 1\) can be mapped to the pole \(v=0\) by the following sequence of \(\textrm{Sp}(2,{\mathbb {Z}})\) transformations, namely first an \(\textrm{SL}(2,{\mathbb {Z}})_\sigma \) transformation, then an S-duality transformation and finally an integer shift of v.
Proof
By performing \(\textrm{SL}(2,{\mathbb {Z}})_{\sigma }\) transformation (3.9), generic pole (3.16) with \(n_2 \ge 1\) gets mapped to
which equals
in the original variables. Therefore, this transformation maps
leaving j invariant. Let
so that
Note that \(\gamma > 0\). Now define \(\alpha \in {\mathbb {Z}}/\gamma {\mathbb {Z}}\) and \(\beta \) to satisfy \(\alpha \delta -\beta \gamma =1\). Then, inspection of (3.21) shows that the \(\textrm{SL}(2,{\mathbb {Z}})_{\sigma }\) transformation
takes generic pole (3.16) to the linear pole
This is equivalent to starting from (3.16) and mapping
Since the determinant of the matrix on the left-hand side equals \((1-j^2)/4\) by virtue of (3.17), we infer
Now we define \(-c = \gcd (r, (j-1)/2)\) and \(d= \gcd (m_2 \gamma - n_1 \delta , (j+1)/2)\). Then \(\gcd (-c, d) = 1\). Consequently, we write
Note that \(c<0\) and \(a> 0\), as well as \(\gcd (a, (j-1)/2) = 1 = \gcd (b, (j+1)/2)\). It is straightforward to seeFootnote 2 that
Determinant equation (3.28) gives
which, using (3.30), implies \(a = \pm A\) and \(b = \pm B\), where the signs are correlated. We pick the \(+\) sign. Then, pole (3.26) can be written as
This allows us to deploy the \(\textrm{SL}(2, {\mathbb {Z}})\) matrix \(\begin{pmatrix} d &{} - b \\ - c &{} a \end{pmatrix}\) to implement an S-duality transformation on \(\Omega \), as in (3.6), to yield
We finally execute a simple translation operation \(v\rightarrow v + \Sigma \), where \(\Sigma = n_1 \beta - m_2 \alpha \in {\mathbb {Z}}\), to generate the pole \(v=0\). We now solve
If \(\alpha =0\), we infer using \(\beta \gamma = -1\),
On the other hand, if \(\alpha \ne 0\), then we may multiply the first equation with \(\alpha \) and the second equation with \(\gamma \) to get
Inserting this result back into the first equation determines \(m_2\) to equal
These expressions reproduce the values of \(n_1\) and \(m_2\) when \(\alpha =0\).
Thus, we have shown that the five integers \(m_1, m_2, j, n_1, n_2\) can be parametrized as in (3.18), and any pole (3.16) with \(n_2 \ge 1\) can be mapped to the pole \(v=0\) by the sequence of \(\textrm{Sp}(2, {\mathbb {Z}})\) transformations given above. \(\square \)
Using parametrization (3.18), each of poles (3.16) with \(n_2 \ge 1\) corresponds to an element of the set \(P \cup \{ \Sigma \in {\mathbb {Z}} \}\), where P is defined by
Defining the sets of matrices,
we will be summing over elements in \(\Gamma _\infty \backslash S_\Gamma \), \(S_G\), and in \( \{ \Sigma \in {\mathbb {Z}} \}\), where
We can therefore parametrize the sum over poles in the set P as
3.3 Residues at \(v'=0\)
As shown above, any pole (3.16) with \(n_2 \ge 1\) can be mapped to the pole \(v=0\) under the \(\textrm{Sp}(2, {\mathbb {Z}})\) transformation
with \(\rho ', \sigma ',v' \) given by (3.14).
Defining
we infer
From the above we see that for a \(\Lambda \) satisfying the constraint \(\rho = \Lambda \), a translation of \(\Sigma \) in integral units of ac will modify \(\Lambda \) such that it will no longer satisfy the constraint in the given range of \(\rho _1\). Hence we restrict \(\Sigma \) to take values
Next we evaluate the \(\rho \)-integral in (2.3). We only consider the contributions from the residues associated with the poles of \(1/\Phi _{10}\) with \(n_2 \ge 1\). Since any such pole can be mapped to the pole \(v' =0\), we compute the residue associated with \(\rho = \Lambda (\sigma , v) \), for fixed \(\sigma \) and v by noting that in the neighbourhood of \(v=0\), \(1/\Phi _{10}\) behaves as
Therefore, for fixed \(\sigma \) and v, evaluating the residue at \(\rho = \Lambda (\sigma , v) \) using (3.46), we obtain
Using
we obtain
where
4 Integrating over v
Next, we perform the integration over v in (2.5). We will first define the contour \(\Gamma _v (D)\) for a given \(n_2 \ge 1\) pole. For notational simplicity we will henceforth refer to this contour by \(\Gamma _v\).
Recall that we are considering poles with \(n_2 \ge 1\). The pole \(v' _*=0\) specified in (3.50) will be in the Siegel upper half plane provided
Evaluating
and imposing (4.1) results in
Using the value for \(v_2/\sigma _2\) given in (2.4), we obtain
with
The above defines a contour \(\Gamma _v\) of integration for the v-integral that goes from
Since \(n_2 \ge 1\), the range of integration specified by (4.4) lies in the original unit interval length integration contour for \(v_1\), as required.
4.1 The v Integral
The residue associated with the pole \(v_*'=0\) was given in (3.49). We now turn to the v-integral over the sum of the residues associated with the poles that belong to the set \(P \cup \{ \Sigma \in {\mathbb {Z}} \}\) given in (3.38). We follow the prescription given in (2.5),
Since \(\rho '_*\) and \(\sigma '_*\) satisfy (4.1), we may Fourier expand \(1/\eta ^{24}\) and \(E_2\) using
and
Then, (4.7) becomes
Substituting the values for \(\rho '_*\) and \(\sigma '_*\) given in (3.50) results in
where
This leads us to define the combinations
Notice that (4.11) contains the following sum over \(\Sigma \),
which is only non-vanishing provided that the combination L is an integer, in which case this sum equals \(-ac\). In other words, the only contributions to (4.11) will come from poles that satisfy the divisibility condition
Thus, restricting to \(L\in {\mathbb {Z}}\), (4.11) becomes
Using (4.13), we obtain
We note that the triplets \((m,{\tilde{n}},{\tilde{\ell }})\) and (M, N, L) are related by the following \(\textrm{SL}(2,{\mathbb {Z}})\) transformation,
or equivalently,
Therefore, since the triplet (M, N, L) consists of integers, also \({\tilde{n}}\) and \({\tilde{\ell }}\) have to be integers. Thus, we have two triplets of integers that are related by the \(\textrm{SL}(2,{\mathbb {Z}})\) matrices G given in (3.19).
Next, we define
Note that \({\tilde{\Delta }}\) is invariant under the \(\textrm{SL}(2,{\mathbb {Z}})\) transformation G given in (3.19),
Now we recall that we are considering BPS dyons with \(\Delta = 4\,m n - \ell ^2> 0\), and hence \(m>0\), in which case we may perform the following rewriting of the exponent in (4.16),
Using this, (4.16) can be written as
4.1.1 T-Shifts
In the following, we rewrite expression (4.23) by recasting the sum over \(b\in {\mathbb {Z}}\) as a sum over a new integer T, as follows. Let \(T \in {\mathbb {Z}}\) and consider the matrix \(\begin{pmatrix} 1 &{} T \\ 0 &{} 1 \end{pmatrix}\). First we note that the operation
induces the change
while leaving a and c invariant. Therefore, expressions such as L that only depend on a and c are invariant under this change. Hence, also \({{\tilde{\Delta }}}\) is invariant under this change. On the other hand, \({\tilde{\ell }}\) and \({\tilde{n}}\) transform as follows,
Next, note that \({\tilde{\ell }} \) can be written as
and hence the condition \(b\in {\mathbb {Z}}/a{\mathbb {Z}}\) translates to \({\tilde{\ell }} \in {\mathbb {Z}}/2m{\mathbb {Z}}\). Thus we can express (4.23) as
Using (4.25), integration contour (4.6) goes from
and the range of integration specified by (4.4) requires restricting the values of T to
4.1.2 Performing the v-Integral
We now perform the v-integral in (4.28) along the contour \(\Gamma _v\) specified by (4.29). To do so, we assume the legitimacy of interchanging the integration with the summation over M, N in (4.28). Note that the dependence on v is contained in the last line of (4.28), only,
Using the expression for the error function
we obtain for (4.31),
where we take the principal branch of the square roots.
Hence, integrating (4.28) over v results in
4.1.3 \(\textrm{SL}(2,{\mathbb {Z}})\) Multiplier System
Inspection of (4.34) reveals the presence of the multiplier system associated with the \(\textrm{SL}(2,{\mathbb {Z}})\) matrix \(\Gamma = \begin{pmatrix} \alpha &{} \beta \\ \gamma &{} \delta \end{pmatrix}\) [18, 19],
It has the property \(\psi _{(\ell + 2\,m k) j}(\Gamma ) = \psi _{ {\ell } j}(\Gamma )\) (with \(k \in {\mathbb {Z}}\)) [20]. This multiplier system arises when considering the transformation law of the standard Jacobi theta function \(\vartheta _{m,\ell }(\sigma ,v)\) of weight 1/2 and index m under modular transformations [13],
The multiplier system that enters in (4.34) is \( \psi (\Gamma )_{{\tilde{\ell }}\ell }\).
4.1.4 Splitting the Error Functions
The error functions in (4.34) can be split into three different terms. We refer to Appendix A for the details on this. Here we briefly summarise its salient features.
Using the property \(\text {Erf}(-x) = -\text {Erf}(x)\), we write the difference of the error functions in (4.34) as
Next we define the following functions,
and
We introduce
and we note that since \(-1/ac >0\), X and Y cannot be both negative. We now consider the product
Depending on the sign of XY, the expression for sum (4.37) will take a different form, as follows. If \(X Y > 0\), (4.37) becomes
If \(X Y = 0\), (4.37) has the form
where we used that when \(X=0\), Y is given by \(Y=- 1/ac\), and when \(Y=0\), X is given by \(X=-1/ac\). Since \(- 1/ac > 0\), we can re-express the condition \(X Y \ge 0\) as
This condition [10] is satisfied by all the convergents of the continued fraction of
Finally, when \(X Y < 0\), (4.37) takes the form
The above shows that the error functions in (4.34) give rise to three distinct types of terms, namely constants, terms involving \(E_{m,\gamma (\gamma \sigma +\delta )}\) and terms involving \(I_{m,\gamma (\gamma \sigma +\delta )}\). As we will see below, upon performing the \(\sigma \)-integral, these three distinct contributions will give rise to terms involving the Bessel function \(I_{23/2}\), the Bessel function \(I_{12}\) and the integral of the Bessel function \(I_{25/2}\), respectively.
4.2 A Symmetry
Expression (4.34) possesses a symmetry that can be used to simplify the terms that involve \(E_{m,\gamma (\gamma \sigma +\delta )}(x)\) and \(I_{m,\gamma (\gamma \sigma +\delta )}(x)\), as follows. Let us first consider the case when \(X Y >0\) or \(XY<0\). Then, (4.34) will contain terms of form (c.f. (4.42) and (4.46))
We observe that under the mapping,
the property \(a',-c'>0\) as well as the unit determinant property are preserved. Further, the following quantities remain invariant:
while \({\tilde{\ell }}, X\) and Y are mappedFootnote 3
Thus, we can write (4.47) as
and hence we can express the combination
which occurs in (4.34) when \(X Y >0\) or \(XY < 0\), as
On the other hand, when \(XY=0\), (4.34) will contain terms of form (4.43), which, using mapping (4.48), can be brought to form (4.51). Hence, the combined contributions from \(X=0\) (in which case \(Y=-1/ac\)) and from \(Y= 0\) (in which case \(X=-1/ac\)) take again form (4.53). Thus, irrespective of whether \(XY >0, XY=0\) or \(XY<0\), the contributions to (4.34) from the terms involving \(E_{m,\gamma (\gamma \sigma +\delta )}(x)\) and \(I_{m,\gamma (\gamma \sigma +\delta )}(x)\) take the form
5 Integrating over \(\sigma \)
Next, we turn to the \(\sigma \)-integral of (4.34). We first explicate our choice of integration contour \(\Gamma _\sigma (D)\) in (2.5) for a given \(n_2 \ge 1\) pole D in \(P \cup \{\Sigma \in {\mathbb {Z}} \}\).
5.1 Contour of Integration
The contour \(\Gamma _\sigma (D)\) defined in the \(\sigma \) upper half plane restricted to \(\sigma _1 \in [0,1)\) crosses locus (3.16) associated with quadratic poles. Writing out the real and imaginary parts of (3.16) gives
Solving the second equation for \(\rho _1\) yields
Now we recall that \(v_1\) lies in range (4.3), which we write as
where \(x\in (0,-1/ac\gamma )\). Inserting (5.3) as well as (5.2) into the first equation of (5.1) gives
This describes an ellipse in the \((\sigma _1, x)\)-plane provided that the right-hand side of this equation is non-vanishing and positive,
where we set
Combining the \(\sigma _2\)-direction with (5.4) results in an ellipsoid in three dimensions, with \(\sigma _2\) taking values in the range specified by (5.5).
On the other hand, we note that the point \((\sigma _1 = - \delta /\gamma , x, \sigma _2 =0)\) lies in locus (5.4). This point, which lies on the boundary of the Siegel upper half plane, is the anchoring point of a curve in the complex \(\sigma \)-plane at fixed \(x\in (0,-1/ac\gamma )\), as follows. Since \(\sigma _2 >0\), we may divide (5.4) to arrive at
where we defined
Note that \(X(x)> 0\) for \(x\in (0,-1/ac\gamma )\), and that the positivity of \(\lambda \) can be enforced by taking \(\rho _2\) to be sufficiently large. At fixed x, above equation (5.7) describes a circle in the complex \(\sigma \)-plane anchored at \((\sigma _1 = - \delta /\gamma , \sigma _2 =0)\), provided that \(\lambda \) is kept fixed. The latter is compatible with condition (5.5), as follows. We write (5.5) as
The bound on the right-hand side is precisely saturated when \(\sigma _2 = X_{\max } /\lambda \), where \(X_{\max } \) is the maximal value of X, which is attained for \(x = - 1/(2 a c \gamma )\). The associated point on the circle is \((\sigma _1 = - \delta /\gamma , \sigma _2 = X_{\max } /\lambda )\), which is the point on the circle that intersects the line \(\sigma _1 = - \delta /\gamma \) in the complex \(\sigma \)-plane. Thus, we see that keeping \(\lambda \) fixed is compatible with restricting the range of \(\sigma _2\) to
At fixed \(\lambda \), the circle described by (5.7) is homotopic to a Ford circle \({\mathcal {C}}( -\delta , \gamma )\) in the complex \(\sigma \)-plane anchored on the real axis at \(\sigma _1 = - \delta /\gamma \) (see Appendix C.1 for details). Note that since the homotopy is between circles, the leading behaviour of the integrand in (5.11) when approaching the point \((\sigma _1 = - \delta /\gamma , \sigma _2 =0)\) along any of these two circles in the same. The chosen range \(\sigma _1 \in [0,1)\) constrains the poles contributing to (4.34) to those associated with \(0 \le -\frac{\delta }{\gamma }< 1\). Since this holds for any \(x\in (0,-1/ac\gamma )\), our integration contour over \(\sigma \) for a given pole is \(\Gamma _\sigma (D) = {\mathcal {C}}( -\delta , \gamma ) \), which for notational simplicity we will denote by \(\Gamma _{\sigma }\).
The interpretation of this construction is the one given in [21]. When \(\sigma _2\) is large, the integration contour does not intersect the ellipsoid described above. When lowering the value of \(\sigma _2\), the integration contour will cross some of the poles in the Siegel upper half plane described by (5.7). This will cease to be the case when \(\sigma _2\) reaches the boundary \(\sigma _2 = 0\) of the Siegel upper half plane. Fixing the value of \(\lambda \) to be large enough, we note that as we decrease \(\sigma _2\) we continue to remain in the \({{\mathcal {R}}}\)-chamber, ensuring that the integration contour does not cross any \(n_2=0\) pole.
We will now perform the \(\sigma \)-integral of (4.34) over the Ford circle \(\Gamma _{\sigma }\) described above, following the prescription given in (2.5), which results in
To perform this integral, we will use the decomposition of the error functions given above.
5.2 Bessel Function \(I_{23/2}\)
We first focus on the constant terms in decompositions (4.42) and (4.43).
Firstly we show that the two cases corresponding to \(X=0\) and \(Y=0\) give rise to the same contribution, as follows. The condition \(X=0\) yields \(m -a^2\,M = - c^2 N\), and hence \(L = 2 c N / a\). We will show later that only terms with \({{{\tilde{\Delta }}}} <0\) contribute. Therefore combining \(L = 2 c N / a\) with \({{{\tilde{\Delta }}}} <0\) results in \(L > 0\), which in turn implies \(N=-1\). The latter implies a|2. Consequently \({{{\tilde{\ell }}}} = - b \, 2\,m /a = k m, k \in {\mathbb {Z}}\). Therefore \({{{\tilde{\ell }}}} = - {{{\tilde{\ell }}}} \, \text {mod} \, 2\,m \) and hence \( \psi (\Gamma )_{-{\tilde{\ell }}\ell } = \psi (\Gamma )_{{\tilde{\ell }}\ell } \). Using mapping (4.48), this shows that the contribution from the sector \(X=0\) equals the one from the sector \(Y=0\). Therefore, the combined contribution from the sectors \(X Y \ge 0\) can be expressed as follows,
where continued fraction condition (4.44) now takes the form given in (5.12). To proceed, we interchange the integration with the summation over M, N. This is allowed by the following arguments.
First we note that the condition \(0 \le \frac{b}{a}+\frac{{\tilde{\ell }}}{2\,m} < -\frac{1}{ac}\) in the summation above can be written as
Then, using the expression for L given in (4.13), we write out \({\tilde{\Delta }} = 4MN-L^2\) and obtain
Now let us consider terms that satisfy \({\tilde{\Delta }}\le 0\), in which case we obtain from (5.14),
Combining this with (5.13) we infer
Then, by combining this last inequality with (5.13) we obtain the bounds
Therefore, for a given a, c there is only a finite set of values M, N which satisfy \({\tilde{\Delta }} \le 0 \) as well as continued fraction condition (4.44) and the condition \(L \in {\mathbb {Z}}\). Thus, in this case, we can interchange the integration with the summation over M, N.
Next, let us consider the terms with \({\tilde{\Delta }} >0 \). For large values of M, the Fourier coefficients of \(1/\eta ^{24}\) grow exponentially as
From (5.14), and using (5.13), we infer that for large M, N, \({\tilde{\Delta }} \) behaves, schematically, as \({\tilde{\Delta }} \sim M + N\), and hence becomes large. Parametrizing the Ford circle \(\Gamma _{\sigma }\) in (5.12) by
or, equivalently, by
where \(\theta \in [0,\pi ) \cup (\pi , 2 \pi )\), we infer that on the contour \(\Gamma _{\sigma }\),
Since \(d(M) \, d(N) \, e^{- \frac{2 \pi }{4\,m} {{{\tilde{\Delta }}}}} \) is exponentially suppressed for large M, N, the sum over M, N in (5.12) is uniformly convergent on \(\Gamma \) by the Weierstrass M test, and since each summand is integrable, we conclude that interchanging the integration with the summation over M, N is justified also when \({\tilde{\Delta }} >0 \).
Thus, interchanging the integration with the summation over M, N results in
5.2.1 Bessel Integral
We perform the \(\sigma \)-integration over the Ford circle \(\Gamma _{\sigma }\) that skirts the point \(- \delta /\gamma \). This Ford circle has radius \(1/(2 \gamma ^2)\), is centred at \(\sigma = - \frac{\delta }{\gamma } + i \frac{1}{2 \gamma ^2}\),
and is oriented counter clockwise. Then, (5.22) becomes replaced by
where \(P'\) denotes the set P, but with \(\delta \) restricted to lie in the range \(0 \le -\delta < \gamma \). We change the integration variable to
The essential singularity is now located at the origin \({{{\tilde{\sigma }}}} = 0\). We choose the branch cut, which originates at \({{{\tilde{\sigma }}}} = 0\), to lie along the negative imaginary axis of the \({\hat{\sigma }}\)-plane. Next, we change the integration variable once more,
Now the branch cut originates at \(w = 0\) and lies along the negative real axis of the w-plane. The integration contour now runs along a line parallel to the imaginary axis,
with \({{{\tilde{\epsilon }}}} >0\). Now recall that \(\Delta >0\). When \({{{\tilde{\Delta }}} } \ge 0\), the coefficient \({{\tilde{\Delta }} \Delta } / (4m \gamma )\) in the exponent is \(\ge 0\), and hence the integration contour can be closed in the half plane \(\textrm{Re} \, w > 0\), where the integrand is analytic and hence the integral vanishes. Thus, we now take \({{{\tilde{\Delta }}}} < 0\).
Then, performing the redefinition
and defining \(z = \frac{\pi }{m\gamma }\sqrt{|{{\tilde{\Delta }}}|\Delta }\), we obtain for integral (5.27), with \(\epsilon > 0\),
where \(I_\nu (z)\) denotes the modified Bessel function of first kind of index \(\nu \),
where \(\epsilon > 0\). Then, (5.22) becomes
Note that in (5.31) the dependence on \(\alpha \) and \(\delta \) is entirely encoded in multiplier system \(\psi (\Gamma )_{{\tilde{\ell }}\ell }\) and in the phase
Since \(0\le - \delta < \gamma \) and \(\alpha \in {\mathbb {Z}}/\gamma {\mathbb {Z}}\), and since \(\alpha \) is the modular inverse of \(\delta \), i.e. \(\alpha \delta = 1 \text { mod } \gamma \), each \(\delta \) uniquely specifies one \(\alpha \). Thus, the sum over \(\delta \) yields the generalized Kloosterman sum \(\textrm{Kl}( \frac{\Delta }{4\,m}, \frac{{\tilde{\Delta }}}{4\,m};\gamma ,\psi )_{\ell {\tilde{\ell }}}\),
Thus, (5.31) can be written as
where \({\tilde{\Delta }} = 4 M N - L^2\), with L given in (4.13). Note that the sum over the allowed M, N is finite.
Next, using (4.17), we express the triplet (N, L, M) in terms of the triplet \(({{{\tilde{n}}}}, m, {{{\tilde{\ell }}}})\). We then trade the sum over M, N for a sum over \({{{\tilde{n}}}}, {{{\tilde{\ell }}}}\). In Sect. 5.3 we will show that \({{{\tilde{n}}}}\) is bounded by \({{{\tilde{n}}}} \ge -1\). Writing \({{{\tilde{\Delta }}}}\) as \({\tilde{\Delta }} = 4\,m {{{\tilde{n}}}} - \ell ^2\), we rewrite (5.34) as
which we write as
where \(c_m^F({\tilde{n}},{\tilde{\ell }}) \) is defined by
Note that the above sum includes two subsets of matrices in \(S_G\). The first subset contains matrices satisfying \({\tilde{\ell }}/2\,m=-b/a\), while the second subset contains matrices that correspond to the continued fraction expansion of \({\tilde{\ell }}/2\,m\). The latter subset is finite by definition, while the bounds \(M,N \ge -1\) and (5.13) can be used to show the finiteness of the first subset. This is consistent with the proofs of finiteness of [9, 10].
5.3 Lower Bound on \({\tilde{n}}\)
We next show that \({\tilde{n}} \ge -1\) whenever the condition
is true. To this end, we first recall that (5.38) can be written as
and further \(M, N \ge -1\). Then beginning with
we study three cases:
-
1.
\(d=0\): Then, (5.40) becomes
$$\begin{aligned} {{{\tilde{n}}}} = - \frac{b}{c} M = \frac{M}{c^2}, \end{aligned}$$(5.41)which satisfies \({{{\tilde{n}}}} \ge -1\), since \(M \ge -1\) and \(c^2 \ge 1\).
-
2.
\(bc<0\): In this case, we obtain,
$$\begin{aligned} (ac)^2 {{{\tilde{n}}}}= & {} a b c d m - b c a^2 M + d a c^2 N \nonumber \\= & {} a b c d m + c^2 N - b c (a^2 M - c^2 N ) \nonumber \\= & {} a b c d m + c^2 N + | b c | (a^2 M - c^2 N ). \end{aligned}$$(5.42)Using the lower bound in (5.39), we get,
$$\begin{aligned} (ac)^2 {{{\tilde{n}}}} > a b c d m + c^2 N - | b c | m= & {} c^2 N + m b c \left( a d + 1 \right) \nonumber \\= & {} c^2 N + m \left( (a d)^2 - 1 \right) \ge c^2 N, \end{aligned}$$(5.43)Since \((ad)^2 \ge 1\),
$$\begin{aligned} {\tilde{n}} > \frac{N}{a^2} \ge -1. \end{aligned}$$(5.44) -
3.
\(bc \ge 0\): Here we have, using the upper bound in (5.39),
$$\begin{aligned} (ac)^2 {{{\tilde{n}}}}= & {} a b c d m - b c a^2 M + d a c^2 N \nonumber \\= & {} a b c d m + c^2 N - b c (a^2 M - c^2 N ) \nonumber \\\ge & {} a b c d m + c^2 N - b c m \nonumber \\= & {} c^2 N + b c m (a d -1) = c^2 N + (b c)^2 m \ge c^2 N. \end{aligned}$$(5.45)Hence,
$$\begin{aligned} {\tilde{n}} \ge \frac{N}{a^2} \ge -1. \end{aligned}$$(5.46)
5.4 Bessel Function \(I_{12}\)
Now we focus on the terms \(E_{m,\gamma (\gamma \sigma + \delta )}\) in decompositions (4.42), (4.43) and (4.46). In Sect. 5.5, we will show that these terms only give a non-vanishing contribution when \(X Y \ne 0\), whereas when \(X Y =0\) they cancel out. Thus, we will assume \(X Y \ne 0\) in the following.
We proceed as in Sect. 5.2.1, namely, we restrict the sum over \(\delta \) to the range \(0 \le -\delta < \gamma \) and perform the integration over the Ford circle \(\Gamma _{\sigma }\) in (5.23). Collecting the terms proportional \(E_{m,\gamma (\gamma \sigma + \delta )}\) in (5.11), and changing the integration variable to \({{{\tilde{\sigma }}}}\) given in (5.25), we obtain
where L and \({{{\tilde{\ell }}}}\) are given in (4.13), and where the generalized Kloosterman sum \(\textrm{Kl} ( \frac{\Delta }{4\,m}, \frac{{\tilde{\Delta }}}{4\,m};\gamma ,\psi )_{\ell {\tilde{\ell }}}\) is given in (5.33). The set \(P_0\) is given by
The integration contour \({{{\tilde{\Gamma }}}}\) denotes a Ford circle centred at \({{{\tilde{\sigma }}}} = i/2\) that skirts the origin \({\tilde{\sigma }} =0\).
Using symmetry property (4.54), we write (5.47) as
Using the expression for \(E_{m,{\tilde{\sigma }}}\) given in (4.38) and the relation
we obtain
Note that the integrand does not exhibit a branch cut. Performing the variable change given in (5.26), the integral over the Ford circle \({{\tilde{\Gamma }}}\) takes a form similar to (5.27), with \({{{\tilde{\Delta }}}}/4m\) replaced by \(N/a^2\). The integral will be non-vanishing provided \(N/a^2 <0\). This in turn implies \(N=-1\). As shown in Sect. 5.5, the only contributions to the sum come from the terms in set the \(P_0\) satisfying \(a=1\) and \(M=m\). Using the expression for L and \({{{\tilde{\ell }}}}\) given in (4.13), we infer that in this case,
in which case
so that (5.51) yields (using \(d(-1) = 1\))
where the set \(P_{0}'\) is the set \(P_0\) subject to the restrictions \(a=1,N=-1,M=m\).
From (5.52) we infer
in which case
Using (5.52), we can write the Kloosterman sums in (5.54) as (recall that \(b \in {\mathbb {Z}}/a{\mathbb {Z}} \) with \(a=1\) fixes b to a single value)
Using
and using the multiplier system property \(\psi (\Gamma )_{(j + 2\,m k) {\ell }} = \psi (\Gamma )_{j {\ell }}\) (with \(k \in {\mathbb {Z}}\)) [20], we write (5.57) as
Using (5.56), this becomes
Next we focus on the sum over c in this expression,
which we write as
where we made use of the symmetry \(j \rightarrow - j\) to obtain an expression with the symmetry \(c \rightarrow -c\).
Expression (5.62) is divergent. Below we will discuss a regularization procedure to extract a finite part of this expression.
5.4.1 Regularization
We now regard the sum over \(c \in {\mathbb {Z}} \backslash \{0\} \) in (5.62),
as
where \( \vartheta _{m,j}(\tau ) \) denotes the standard weight 1/2 index m Jacobi theta function. Using this, we then regard (5.62) as
We focus on the combination
which is one of the combinations contained in (5.65). Using that \(\psi (\Gamma )_{j {\ell }}\) are the components of a unitary matrix, i.e.
we infer the property
where
Now we recall the transformation property under \(\textrm{SL}(2,{\mathbb {Z}})\) transformations,
Setting \(v=0\) and choosing the \(\textrm{SL}(2,{\mathbb {Z}})\) transformation \({{{\tilde{\Gamma }}}}\), this becomes
Then, by combining (5.66) with (5.68) and (5.71), we obtain
Now we study this equation in the limit \(\tau \rightarrow -\alpha /\gamma \). Using
we infer that in the limit \(\tau \rightarrow -\alpha /\gamma \), the right-hand side of (5.72) tends to \(- \psi (\Gamma )_{0 {\ell }}\) when \(\ell \ne 0\), whereas when \(\ell = 0\) it diverges and behaves as \((\gamma \tau +\alpha )^{-1/2} - \psi (\Gamma )_{0 0}\). We note that the divergent term \((\gamma \tau +\alpha )^{-1/2}\) that arises when \(\ell =0\), is due to the presence of a constant term in \(\vartheta _{m,0}(\tau )\). Then, by subtracting this constant term we obtain the following regularized expression,
A similar reasoning applies to the other combination, proportional to \(\psi (\Gamma )_{(-j) {\ell }}\), contained in (5.65).
Thus, we are led to the following regularized expression for (5.65),
Using this in (5.60), we arrive at our proposal for the regularized expression for (5.60),
To summarize, the regularization procedure described above removes one divergent contribution that arises when \(\ell =0\). At present we do not have neither a physics nor a mathematics justification for using precisely this regulator.
5.4.2 Bessel Integral
Now we return to (5.54), which we regularize using expression (5.76),
Changing the integration variable to
we get
Using (5.30) we obtain,
5.5 Isolating Non-vanishing Contributions
We return to (4.54) show that the sum exhibits cancellations between various terms, thereby identifying non-vanishing contributions. Taking into account the form of \( E_{m,\gamma (\gamma \sigma +\delta )} \left( X\right) \) and \(I_{m,\gamma (\gamma \sigma +\delta )}\left( X \right) \) given in (4.38) and (4.39), we write (4.54) as
where \(F_{m,\gamma (\gamma \sigma +\delta )}\) is a function that only depends on \(X^2\). Consider changing the variable \(\sigma \) to \({{{\tilde{\sigma }}}} = \gamma (\gamma \sigma + \delta )\), and integrating along a Ford circle \({{{\tilde{\Gamma }}}}\), centred at \({{{\tilde{\sigma }}}} = i/2\), that skirts the origin \({{{\tilde{\sigma }}}} =0\). Then, as described in Sect. 5.4, only when \(N=-1\) do we get a non-vanishing contribution. This was shown in Sect. 5.4 for the case of the contribution proportional to \( E_{m,\gamma (\gamma \sigma +\delta )} \left( X\right) \), but the same argument also applies to \(I_{m,\gamma (\gamma \sigma +\delta )}\left( X \right) \). Thus, in the following, we set \(N=-1\).
We evaluate
We recall that the condition \(L\in {\mathbb {Z}}\) in (5.81) implies the divisibility condition \(ac \mid m -a^2\,M +c^2\). Hence, if we fix M and a, the sum over c runs only over those \(c<0\) that satisfy
We now fix M and a. There are three cases to be considered. First, consider the case when
Let \(c<0\) be an integer that divides \(m-a^2M\) and contributes amount (5.82). For any such integer there exists another integer \(c' = -\frac{m-a^2M}{c}<0\), which also divides \(m-a^2M\) and contributes the amount
Thus, in (5.81), for a fixed M and a, any summand \(c<0\) with an associated value L comes accompanied by a summand \(c'<0\) with an associated value \(L' = -L\). Both these summands give rise to the same value \({\tilde{\Delta }} = 4MN-L^2 = - 4\,M -L^2\). Since both c and \(c'\) have the same value \(\frac{b}{a}+\frac{{\tilde{\ell }}}{2m}\), we infer that the value \({\tilde{\ell }}' \) associated with \(c'\) is
We now show that it is possible to pick \(b=b'\) in the above as follows. Observing that \(a|(m+c^2)\), we write
and hence
Hence the operation
yields a new matrix with the same a and \(b'=b\) and with the required \(c'\). Hence we can choose \(b'=b\), in which case \({\tilde{\ell }}' = {\tilde{\ell }}\). Since the multiplier systems are the same, we get a cancellation between the contributions from c and \(c'\).
Next, let us consider the case when
Let \(c<0\) be an integer that divides \(m-a^2M\) and contributes amount (5.82). For any such integer there exists another integer \(c' = \frac{m-a^2M}{c}<0\), which also divides \(m-a^2M\) and contributes the amount
Note that now \(L' = L\), while
Using (5.87), we obtain
We then choose
which satisfies \(a' d' - b' c' = 1\). The shift by j returns \(b'\) to \(b' \in {\mathbb {Z}}/a {\mathbb {Z}}\). Then, from (5.92), we infer
Using the property \(\psi (\Gamma )_{(-{{{\tilde{\ell }}}} - 2\,m j) \ell } = \psi (\Gamma )_{- {{{\tilde{\ell }}}} \ell }\), and recalling that (5.81) uses the two multiplier systems \(\psi (\Gamma )_{ {\tilde{\ell }}\ell }\) and \(\psi (\Gamma )_{-{\tilde{\ell }}\ell }\), also in this case we get a cancellation between the contributions from c and \(c'\).
Finally, let us consider the case when \(m-a^2M = 0\). Then, the divisibility condition \(ac \mid m -a^2M +c^2\) becomes
and since \(\gcd (a,c)=1\), this imposes
and therefore
Thus, when \(N=-1\) the only terms that contribute satisfy \(M=m\) and \(a=1\).
Finally, we note that when \(N=-1, M=m\) and \(a=1\), then
Since \(c<0\), we infer \(\frac{{\tilde{\ell }}}{2\,m}+\frac{b}{a} \notin [0, - \frac{1}{ac}]\), and hence \(X \ne 0, - \frac{1}{ac}\).
Thus, we conclude that only when \(X Y \ne 0\), do we get non-vanishing contributions in (5.81).
5.6 Bessel Function \(I_{25/2}\)
Finally, we focus on the terms \(I_{m,\gamma (\gamma \sigma + \delta )}\) in decompositions (4.42), (4.43) and (4.46). In the previous subsection we showed that these terms only give a non-vanishing contribution when \(X Y \ne 0\). Thus, we will assume \(X Y \ne 0\) in the following.
We proceed as in Sect. 5.2.1, namely, we restrict the sum over \(\delta \) to the range \(0 \le -\delta < \gamma \) and perform the integration over the Ford circle \(\Gamma _{\sigma }\) in (5.23). Collecting the terms proportional \(I_{m,\gamma (\gamma \sigma + \delta )}\) in (5.11), and changing the integration variable to \({{{\tilde{\sigma }}}}\) given in (5.25), we obtain
where L and \({{{\tilde{\ell }}}}\) are given in (4.13), and where the generalized Kloosterman sum \(\textrm{Kl} ( \frac{\Delta }{4\,m}, \frac{{\tilde{\Delta }}}{4\,m};\gamma ,\psi )_{\ell {\tilde{\ell }}}\) is given in (5.33). The set \(P_0\) is given in (5.48). The integration contour \({{{\tilde{\Gamma }}}}\) denotes a Ford circle centred at \({{{\tilde{\sigma }}}} = i/2\) that skirts the origin \({{{\tilde{\sigma }}}} =0\).
Using symmetry property (4.54), we write (5.100) as
Using the expression for \(I_{m,{\tilde{\sigma }}}\) given in (4.39) and relation (5.50), we obtain
Now consider the integral
which we rewrite using [22]
where \(s \in {\mathbb {R}} \backslash \{0\}\), and where \(t \in {\mathbb {C}}\) with \(\textrm{Re} \, t > 0\). Setting \(u = - i m z\) and \(t = im /{{{\tilde{\sigma }}}}\) (note that \(\textrm{Re} \, t = m \sigma _2/|{{{\tilde{\sigma }}}}|^2 > 0\)), integral (5.103) can be expressed as
Then, collecting the exponential terms in (5.102) with \(1/{\tilde{\sigma }}\) in the exponent gives
Interchanging the two integrations in (5.102), and performing the integration over \({{{\tilde{\sigma }}}}\) along the Ford circle \({{{\tilde{\Gamma }}}}\) as described in Sect. 5.4, the latter will only be non-vanishing provided that \(\left( \frac{N}{a^2}+mx^2 \right) < 0\), which in turn implies \(N=-1\). Then, as shown in the previous subsection, the only non-vanishing contribution to (5.102) stems from \(N=-1, a=1, M=m\).
Then, setting \(N=-1, a=1, M=m\), we have
Using (5.58) and the multiplier system property \(\psi _{(j + 2\,m k) {\ell }}(\Gamma ) = \psi _{j {\ell }}(\Gamma )\) (with \(k \in {\mathbb {Z}}\)) [20], we write (5.102) as
Noticing that the dependence on c is quadratic, we write
where we made use of the symmetry \(j \rightarrow - j\) to obtain an expression with the symmetry \(c \rightarrow -c\). We now re-express the sum over c as [11]
Then, the c depending part of Kloosterman sum (5.33) gets re-expressed as
while the exponential term in the dz-integral becomes
Thus we rewrite (5.108) as
Note that the sum over p and g builds up the standard weight 1/2 index m Jacobi theta function, with the exception of the term \(2m\gamma p+ g =0\),
Next, we follow [11, 14]. Using the results reviewed in Appendix B, we rewrite (5.113) as
where
We now turn our attention to the \({\tilde{\sigma }}\) and \(x'\) integrals
We perform the change of variables
and obtain (\({{{\tilde{\epsilon }}}} > 0\))
Interchanging the two integrations and performing the integration over w first, we only get a non-vanishing Bessel integral provided that we restrict the range of integration over \(x'\) to \(1- mx'^2 > 0\), that is to
Performing another change of variables,
integral (5.119) becomes (\(\epsilon >0\))
which, using (5.30), equals
Thus, (5.115) becomes
The above expression involves two Kloosterman sums. Using the property \(f_{\gamma , - g,m}(- x') = f_{\gamma ,g,m}(x')\), we can show that both Kloosterman sums give rise to the same contribution, because the sign change of j in the multiplier system can be compensated by a sign change in g and \(x'\). Thus, we arrive at the expression
5.7 Complete Result for \(\Delta > 0\)
Combining contributions (5.36), (5.80) and (5.125), and relabelling j by \({{\tilde{\ell }}}\), we obtain the following expression for the degeneracy \(d(m,n,\ell )\) defined in (2.5):
Theorem 5.1
with
Note that the continued fraction structure is encoded in a subset of \(S_G\), while the Kloosterman sums \(\textrm{Kl}\) and Bessel functions are built up from \(\Gamma _{\infty } \backslash S_{\Gamma }\) and \(S_G\).
Expression (5.126), without the continued fraction structure of \(c^F_m({\tilde{n}},{\tilde{\ell }})\), was first derived in [11] by viewing \(d(m, n, \ell )\) as Fourier coefficients of a mixed Mock Jacobi form \(\psi _m^F\) [13]. The above result is an exact expression for \(d(m,n,\ell )\) which can be viewed as the non-perturbative completion of previous results in [16, 17, 23].
6 The Case \(\Delta = m= \ell = 0\)
Now we consider the case \(m= \ell = 0\), which implies \(\Delta =0\).
We return to (4.16) and set \(m=\ell =0\),
where the integers \(L, {{{\tilde{\ell }}}}\) and \({{{\tilde{n}}}}\) are given by
We apply T-shift transformation (4.25) to b and to d, which leaves \(L, {\tilde{\ell }}\) invariant and changes \({{\tilde{n}}}\) into
Then, as before, we split the sum over b into a sum over \(b\in {\mathbb {Z}}/a{\mathbb {Z}}\) and a sum over \(T \in {\mathbb {Z}}\).
We pick the following integration contour \(\Gamma _v\). We set \(v_2 = 0\). This choice is motivated by noting that setting \(\ell =0\) in (2.4) one obtains \(v_2 =0\). Then, setting \(v_2 = 0\) in expressions (4.2), we infer that the range of \(v_1\) is restricted to
where we have made use of the T-shift. Demanding that the range of integration over \(v_1\) is contained in an interval of length 1 constrains T to take values in \(T\in {\mathbb {Z}}/\gamma {\mathbb {Z}}\). Then, (6.1) takes the form
We interchange the integration over \(v_1\) with the summation over M, N. The integration over \(v_1\) will produce a different result depending on whether \({{{\tilde{\ell }}}}\) vanishes or not. Therefore, we discuss both cases separately.
6.1 \({\tilde{\ell }} = 0\)
Setting \({{{\tilde{\ell }}}} = 0\) implies
The integral over \(v_1\) is trivial and independent of T. The sum over T yields a factor \(\gamma \). Thus, we obtain for (6.5),
Using
as well as (6.6), we get
Next, we integrate over \(\sigma _1\). We proceed as in Sect. 5.2.1, namely, we restrict the sum over \(\delta \) to the range \(0 \le -\delta < \gamma \) and perform the integration over the Ford circle \(\Gamma _{\sigma }\) in (5.23). We interchange the integration over \(\sigma \) with the sum over M and N. The resulting integral is similar to the one in (5.24), but without a branch cut, and will only be non-vanishing provided that \(N <0\), i.e. when \(N=-1\). Inserting \(N=-1\) into the expression for \(L - {{{\tilde{\ell }}}}\) given in (6.2), and recalling that \(\gcd (a, c)=1\), fixes \(a=1\). Then, from (6.6), we infer \(M = -c^2\), which in turn implies \(M=-1\) and \(c=-1\), and hence \(L=2\). Moreover, since \(b\in {\mathbb {Z}}/a{\mathbb {Z}} \) with \(a=1\), this fixes b to a single value. Thus, integrating (6.9) over \(\sigma \), we obtain, using \(d(-1)=1\),
Performing the change of variables
the Ford circle \(\Gamma _{\sigma }\) becomes the Ford circle \({\tilde{\Gamma }}\) of radius 1/2 and of centre \({\tilde{\sigma }} = i/2\),
Performing a further change of variables,
we obtain, with \(\epsilon > 0\),
Using (5.30) we obtain,
which equals
when written in terms of the classical Kloosterman sum
6.2 \({\tilde{\ell }} \ne 0\)
Now we take \({\tilde{\ell }}\ne 0\) in (6.5) and obtain
Using
and the expressions for \({\tilde{n}},{\tilde{\ell }}\) given in (6.2), we get
We will integrate over \(\sigma _1\) following the procedure described below (6.9). The integration over the Ford circle \(\Gamma \) will select the values \(M=-1\) and \(N=-1\) in (6.20), respectively. Let us then focus on these terms in (6.20). We first consider the contribution from \(M=-1\). Inserting \(M=-1\) into the expression for L and \({{{\tilde{\ell }}}} \in {\mathbb {Z}}\) given in (6.2), we see immediately that the integrality of these two variables, combined with \(\gcd (a, c)=1\), fixes \(c=-1\). Then, we obtain the contribution
Using (6.2) we infer
where \(a | 1-b\), since \(a d - b c = a d + b = 1\). We perform the decomposition
where \(A \in {\mathbb {Z}}/ \gamma {\mathbb {Z}}\) and \(k \in {\mathbb {Z}}\). Then, the exponent in (6.21) becomes
Since both A and T lie in \({\mathbb {Z}}/ \gamma {\mathbb {Z}}\), we can absorb A into T, thereby arriving at
Thus, we may replace the exponent in (6.21) with exponent (6.25), which is independent of \({{\tilde{n}}}\). This we use in the cancellation argument that we now present.
Let us consider
where a|N, since \(L, {{{\tilde{\ell }}}} \in {\mathbb {Z}}\). If \(N>0\), then for each divisor a of N, there will be another divisor \(a' = \frac{N}{a}>0\) which contributes
Since \({\tilde{\ell }}' = {\tilde{\ell }}\), this contribution will cancel against contribution (6.26) in (6.21). Thus, (6.21) does not receive contributions from \(N >0\).
When \(N=0\), we have \(L/{\tilde{\ell }} = 1\), \({{{\tilde{\ell }}}} = a\), \({{{\tilde{n}}}} = - b\). Now we recall that T-transformation (4.25) can be viewed as either imposing the restriction \(b\in {\mathbb {Z}}/a{\mathbb {Z}}\) or \(d\in {\mathbb {Z}}/(-c{\mathbb {Z}})\). We choose the latter viewpoint, in which case we may set \(d=0\), since \(c=-1\). Then \(a d - bc =1\) implies \(b=1\), and we obtain from (6.21),
Summing over T restricts a to
Therefore, only when \(a=k \,\gamma \), with \(k \in {\mathbb {N}}\), do we get a contribution to (6.28),
The sum over \(k>0\) is divergent. We regularize this sum by replacing it with
where \(\zeta (s)\) denotes the Riemann zeta function.
Next, performing the \(\sigma \)-integral over the Ford circle \(\Gamma _{\sigma }\) following the procedure described below (6.9), we obtain
which, when written in terms of classical Kloosterman sum (6.17) and modified Bessel function (5.30), equals
where we have used \(d(0) = 24\).
This is the contribution stemming from the \(M=-1, N>-1\) terms. The contribution from the \(M>-1, N=-1\) terms is exactly the same, so that in total we obtain
6.3 Complete Result for \(\Delta = m = \ell =0\)
Adding up contributions (6.16) and (6.34) from the two sectors \({\tilde{\ell }}=0\) and \({\tilde{\ell }}\ne 0\), we get for the degeneracy d(0, n, 0) of BPS dyons with charge bilinears \((m=0, n, \ell =0)\),
This is in agreement with the expression for the degeneracy of immortal BPS dyons generated by the quasi-modular form [10, 13, 24]
Its Rademacher expansion, which we review in Appendix C, can be obtained by means of the quasi-modular transformation property
and reads
This precisely agrees with (6.35),
Finally, we note that the extra sign stemming from zeta function regularization (6.31) is crucial for obtaining agreement with the expression of Rademacher expansion (6.38).
7 Conclusions
By using two distinct \(\textrm{SL}(2, {\mathbb {Z}})\) subgroups of \(\textrm{Sp}(2, {\mathbb {Z}})\) we have obtained a Rademacher-type expansion for the exact degeneracies of \(\frac{1}{4}\) BPS states with unit torsion in heterotic string theory compactified on a six torus. This expansion is obtained by summing over the contributions from quadratic poles (3.16) of \(1/\Phi _{10}\). The resulting expansion is given by (5.126) in the case when \(\Delta > 0\), and by (6.35) in the case when \(\Delta = m = \ell =0\). When \(\Delta > 0\), the exact expression for the degeneracies exhibits a fine-grained structure that is tied to the presence of the two sets \(S_G\) and \(S_{\Gamma }\) of \(\textrm{SL}(2, {\mathbb {Z}})\) matrices. Result (5.126) reproduces the expression that was obtained in [11] by a very different approach, namely by the Rademacher expansion of the Fourier coefficients of the mixed Mock Jacobi forms \(\psi _m^F\). The latter were shown in [13] to encode the coefficients \(d(m, n, \ell )\) computed in a region [25] where there are no linear pole contributions when \(n>m\). These coefficients are called immortal degeneracies. The Rademacher expansion in this case involved information about the shadow which repairs the modular behaviour of \(\psi _m^F\). The Mock behaviour manifests itself through the \(I_{12}\) and the \(I_{25/2}\) terms in (5.126), both of which would be absent in the case of a modular form. In our approach, the v-integral yields error functions, each of which can be split into three distinct types of terms, namely a constant, a term involving \(E_{m,\gamma (\gamma \sigma +\delta )}\) and a term involving \(I_{m,\gamma (\gamma \sigma +\delta )}\) (c.f. (4.38) and (4.39)). These, respectively, yield the \(I_{23/2}\) terms, the \(I_{12}\) term and the \(I_{25/2}\) term. Further, the sum over \(S_G\) matrices which have a non-vanishing contribution to the \(I_{25/2}\) term yields an integral of a Jacobi theta function minus a constant term. This, when added to the regularized contribution from the \(E_{m,\gamma (\gamma \sigma +\delta )}\) terms, yields precisely the Eichler integral of the shadow for \(\psi _m^F\). Further, the restriction on the range of \(v_1\), which arises by demanding that the poles lie in the \(\sigma \) upper half plane, coupled with the constraint \(v_2/\sigma _2 = - \ell /2m\), reveals the continued fraction structure underlying the polar coefficients \(c^F_m({\tilde{n}},{\tilde{\ell }})\). This structure is not revealed in the approach of [11].
If microscopic result (5.126) is to be obtained from a suitably defined gravity path integral, such as the quantum entropy function [26], then the latter has to make use of the two \(\textrm{SL}(2, {\mathbb {Z}})\) subsets \(S_G\) and \(S_{\Gamma }\). How these two subsets are built into the quantum entropy function is an interesting research question worth exploring [17]. Earlier work [16, 27] on the quantum entropy function for this heterotic model did not identify the distinct roles played by \(S_G\) and \(S_{\Gamma }\). Moreover, inspection of (5.126) shows that the series associated with the Bessel function \(I_{23/2}\) is organized in powers of \(1/\gamma \), and not in powers of \(1/n_2\) as in the semi-classical asymptotic expansion [16, 23, 28] (we recall that \(n_2 = - a c \gamma \)). The dependence on ac in the argument of \(I_{23/2}\) is encoded in \({\tilde{\Delta }}\) through expression (5.14). The results in this paper constitute the non-perturbative completion of the semi-classical results that exist in the literature. Understanding how the Rademacher expansion can be used to rigorously write down the quantum entropy function for immortal degeneracies associated with single centre \(\frac{1}{4}\) BPS black holes in this heterotic model is a challenging question.
Notes
This group is also referred to as \(\textrm{Sp}(4, {\mathbb {Z}})\).
A common prime factor p of a and B must divide \((j-1)/2\). Hence, \(p | (\gcd (a, (j-1)/2) = 1 )\) implying \((a, B) =1\). Similarly \((b, A)=1\).
Note that this mapping might need to be supplemented by a \(\Gamma _{\infty }\) operation on the right to ensure that we are still in the chosen gauge for \(b \in {\mathbb {Z}}/a {\mathbb {Z}}\).
The integral should be understood as an improper integral with both endpoints tending to \(z=0\). The integrand has an essential singularity at \(z=0\). However, the remarkable property of the Rademacher contour is that the integrand is bounded on this path of integration, and therefore the integral over the circle \(C_{1/2}\) is well defined.
References
Dabholkar, A., Harvey, J.A.: Nonrenormalization of the superstring tension. Phys. Rev. Lett. 63, 478 (1989)
Dabholkar, A.: Exact counting of black hole microstates. Phys. Rev. Lett. 94, 241301 (2005). [arXiv:hep-th/0409148]
Maldacena, J.M., Moore, G.W., Strominger, A.: Counting BPS black holes in toroidal Type II string theory, arXiv:hep-th/9903163
Dijkgraaf, R., Verlinde, E.P., Verlinde, H.L.: Counting dyons in N=4 string theory. Nucl. Phys. B 484, 543–561 (1997). [arXiv:hep-th/9607026]
Shih, D., Strominger, A., Yin, X.: Recounting Dyons in N=4 string theory. JHEP 10, 087 (2006). [arXiv:hep-th/0505094]
Rademacher, H.: On the partition function \(p(n)\). Proc. London Math. Soc. S2–43, 1–241 (1938)
Dabholkar, A., Denef, F., Moore, G.W., Pioline, B.: Precision counting of small black holes. JHEP 10, 096 (2005). [arXiv:hep-th/0507014]
Sen, A.: Negative discriminant states in N=4 supersymmetric string theories. JHEP 10, 073 (2011). [arXiv:1104.1498]
Chowdhury, A., Kidambi, A., Murthy, S., Reys, V., Wrase, T.: Dyonic black hole degeneracies in \({\cal{N} } = 4\) string theory from Dabholkar-Harvey degeneracies. JHEP 10, 184 (2020). [arXiv:1912.06562]
Cardoso, G.L., Nampuri, S., Rosselló, M.: Arithmetic of decay walls through continued fractions: a new exact dyon counting solution in \( {\cal{N} } \) = 4 CHL models. JHEP 03, 154 (2021). [arXiv:2007.10302]
Ferrari, F., Reys, V.: Mixed Rademacher and BPS Black Holes. JHEP 07, 094 (2017). [arXiv: 1702.02755]
Zwegers, S.: Mock theta functions, arXiv:0807.4834
Dabholkar, A., Murthy, S., Zagier, D.: Quantum Black holes, wall crossing, and mock modular forms, arXiv:1208.4074
Bringmann, K., Manschot, J.: From sheaves on \(P^2\) to a generalization of the Rademacher expansion. Am. J. Math. 135(4), 1039–1065 (2013). [arXiv:1006.0915]
Borcherds, R.E.: Automorphic forms on \(O_{s+2,2} ({\mathbb{R} })\) and infinite products. Invent. Math. 120, 161 (1995)
Banerjee, N., Jatkar, D.P., Sen, A.: Asymptotic expansion of the N=4 Dyon degeneracy. JHEP 05, 121 (2009). [arXiv:0810.3472]
Murthy, S., Pioline, B.: A Farey tale for N=4 dyons. JHEP 09, 022 (2009). [arXiv:0904.4253]
Zagier, D., Skoruppa, N.-P.: A trace formula for Jacobi forms. J. fuer die Reine Angew. Math. 393, 168–198 (1989)
Gomes, J.: Generalized Kloosterman Sums from M2-branes, arXiv:1705.04348
Kloosterman, H.D.: The behavior of general theta functions under the modular group and the characters of binary modular congruence groups. Ann. Math. 47, 3–317 (1946)
Sen, A.: Black Hole entropy function, attractors and precision counting of microstates. Gen. Rel. Grav. 40, 2249–2431 (2008). [arXiv:0708.1270]
Bringmann, K., Lovejoy, J.: Dyson’s Rank, overpartitions, and weak Maass forms, Int. Math. Res. Not.19 (2007) [arXiv:0708.0692]
David, J.R., Sen, A.: CHL Dyons and statistical entropy function from D1–D5 system. JHEP 11, 072 (2006). [arXiv:hep-th/0605210]
Bossard, G., Cosnier-Horeau, C., Pioline, B.: Exact effective interactions and 1/4-BPS dyons in heterotic CHL orbifolds. SciPost Phys. 7(3), 028 (2019). arXiv:1806.03330
Cheng, M.C.N., Verlinde, E.: Dying dyons don’t count. JHEP 09, 070 (2007). [arXiv: 0706.2363]
Sen, A.: Quantum entropy function from AdS(2)/CFT(1) correspondence. Int. J. Mod. Phys. A 24, 4225–4244 (2009). [arXiv:0809.3304]
Murthy, S., Reys, V.: Single-centered black hole microstate degeneracies from instantons in supergravity. JHEP 04, 052 (2016). [arXiv:1512.01553]
Cardoso, G.L., de Wit, B., Kappeli, J., Mohaupt, T.: Asymptotic degeneracy of dyonic N = 4 string states and black hole entropy. JHEP 12, 075 (2004). [arXiv:hep-th/0412287]
Rademacher, H.: Lectures on Analytic Number Theory. Tata Institute of Fundamental Research, 1954-55
Apostol, T.: Modular functions and dirichlet series in number theory. Springer, Berlin (1990)
Dijkgraaf, R., Maldacena, J.M., Moore, G.W., Verlinde, E.P.: A Black hole Farey tail, arXiv:hep-th/0005003
Acknowledgements
We would like to thank Sergei Alexandrov, Abhiram Kidambi, Boris Pioline and Valentin Reys for valuable discussions. This work was partially supported by FCT/Portugal through CAMGSD, IST-ID, projects UIDB/04459/2020 and UIDP/04459/2020, and through the LisMath PhD fellowship PD/BD/135527/2018 (M. Rosselló).
Funding
Open access funding provided by FCT|FCCN (b-on).
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Ruben Minasian.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A: Error Function Relations
We summarise various useful relations involving error functions.
The error function \(\text {Erf}\) and the complementary error function \(\text {Erfc}\) are defined by
with \(x \in {\mathbb {C}}\). Note that \(\text {Erf}(-x) = -\text {Erf}(x)\). When \(\textrm{Re} \, x > 0\), the complementary error function can be brought to the form
by performing the change of variables \(t' = t^2\) and subsequently integrating by parts.
Using (3.10) we infer
since \(\textrm{Im} \, \sigma ' > 0\). Then
where we take the principal branch of the square root. Defining
the quantities \(x = \sqrt{\frac{2\pi i m}{\gamma (\gamma \sigma +\delta )}} \, X\) and \(x = \sqrt{\frac{2\pi i m}{\gamma (\gamma \sigma +\delta )}} \, Y\) have positive real part when \(X, Y >0\).
Let us then set \(x = \sqrt{\frac{2\pi i m}{\gamma (\gamma \sigma +\delta )}} X \) with \(X >0\), so that \(\textrm{Re} \, x > 0\). Performing the change of variables
complementary error function (A.2) becomes expressed as
On the other hand, when \(X<0\), we use the relations
We define
and
so that for \(X>0\)
Appendix B: Deriving (5.116)
We briefly describe how to obtain (5.116), following [11, 22].
Performing the change of variables \(z = iz'\) in (5.113), we obtain
Using (5.104) and setting
we infer
Performing another change of variables, \(x = \sqrt{m}x'\), this becomes
Next, we perform the sum over \(p \in {\mathbb {Z}}\) in (5.113). Noting that p only enters in the denominator of (B.4), we evaluate
Using the Mittag-Leffler formula
we infer, in the case when \(g \ne 0 \text { mod } 2m\gamma \),
while when \(g = 0 \text { mod } 2m\gamma \), there is one term in the sum that gets excluded and hence
Appendix C: Rademacher expansion for \(2E_2(\sigma )/\eta ^{24}(\sigma )\)
The Fourier coefficients \(c_0^F(n)\) of
have the following well-known exact expression. Namely, using
the Fourier coefficients \(c_0^F(n)\) can be written as
where d(n) are the Fourier coefficients of \(\eta ^{-24}(\sigma )\). The latter admit the following Rademacher expansion in terms of the modified Bessel function \(I_{13}\),
and hence
Using the recurrence relation for the modified Bessel functions
we infer
and hence, (C.5) may also be written as
Rademacher expression (C.8) can also be obtained by deforming the integration contour along Ford circles and subsequently subjecting (C.1) to a modular transformation to re-express the integrand on each Ford circle. In the following, we will use this approach to obtain expression (C.8).
1.1 Appendix C.1: Ford circles and Farey fractions: the Rademacher contour
As is well known, the function \(E_2(\sigma )\) transforms as a quasi-modular form under \(\textrm{SL}(2, {\mathbb {Z}})\) transformations,
Therefore,
To obtain the Rademacher expansion for its Fourier coefficients,
we will apply Rademacher’s method (see [29,30,31]), which was originally developed for modular forms, to this quasi-modular form, as follows. The contour \({\mathcal {C}}\) is any contour in the \(\sigma \) upper half plane that starts at some point, \(\sigma _0\), and ends at \(\sigma _0+1\). Rademacher’s approach consists in deforming the contour \({\mathcal {C}}\) to a new contour that is the union of upper arcs along Ford circles defined by the Farey sequence of order N, and subsequently use the modular properties of the integrand to re-express it on each of these arcs. Then, in the limit \(N\rightarrow +\infty \), the integral becomes expressed as an infinite sum of integrals over Ford circles, yielding a convergent series called the Rademacher expansion. We now review this construction following [29,30,31].
Definition
The set of Farey fractions of order N, denoted by \({\mathcal {F}}_N\), is the set of reduced fractions in the interval [0, 1] with denominators \(\le N\), listed in increasing order of magnitude.
Definition
For a given rational number \(\delta /\gamma \) with \(\gcd (\delta ,\gamma )=1\), the Ford circle \({\mathcal {C}}(\delta ,\gamma )\) is defined as the circle in the complex upper half plane with radius \(1/(2\gamma ^2)\) that is tangent to the point \(\delta /\gamma \).
Two Ford circles \({\mathcal {C}}(\delta ,\gamma )\) and \({\mathcal {C}}(\delta ',\gamma ')\) are either tangent to each other or they do not intersect. They are tangent if, and only if, \(\gamma \delta '-\gamma '\delta = \pm 1\), which is the same as saying that they are the Ford circles associated with two consecutive Farey fractions. If \(\delta _1/\gamma _1< \delta /\gamma < \delta _2/\gamma _2\) are three consecutive Farey fractions in \({\mathcal {F}}_N\), the points of tangency of \({\mathcal {C}}(\delta ,\gamma )\) with \({\mathcal {C}}(\delta _1,\gamma _2)\) and \({\mathcal {C}}(\delta _2,\gamma _2)\) are
and
Then, for a given \(N\in {\mathbb {N}}\), the contour \({\mathcal {C}}\) in (C.11) that goes horizontally from i to \(i+1\) can be deformed to a contour P(N) which also goes from i to \(i+1\), but follows a different path. This new contour P(N), the Rademacher contour, consists in taking the upper arcs of the Ford circles defined by the Farey sequence \({\mathcal {F}}_N\). If \(\delta _1/\gamma _1<\delta /\gamma <\delta _2/\gamma _2\) are three consecutive Farey fractions, the upper arc of \({\mathcal {C}}(\delta ,\gamma )\), denoted by \({\mathcal {C}}_{\delta ,\gamma ,N}\), is the path that joins the two points of tangency \(\alpha _1(\delta ,\gamma ;N)\) and \(\alpha _2(\delta ,\gamma ;N)\) following the upper arc of \({\mathcal {C}}(\delta ,\gamma )\). For the fractions 0/1 and 1/1 we only take the upper arcs lying in the interval [0, 1]. The union of all the upper arcs in \({\mathcal {F}}_N\) is P(N) (Fig. 1).
Therefore, we can write the integral over \({\mathcal {C}}\) as a sum over integrals over the upper arcs
where from now on we take \(\delta \) to be negative, for later convenience. We introduce the following transformation
which maps the Ford circle \({\mathcal {C}}(-\delta ,\gamma )\) to the circle with centre 1/2 and with radius 1/2, and hence tangent to zero, denoted by \(C_{1/2}\). The points of tangency \(\alpha _1(-\delta ,\gamma ;N),\alpha _2(-\delta ,\gamma ;N)\) of \({\mathcal {C}}(-\delta ,\gamma )\) with \({\mathcal {C}}(-\delta _1,\gamma _1)\) and \({\mathcal {C}}(-\delta _2,\gamma _2)\) with \(-\delta _1/\gamma _1< -\delta /\gamma < -\delta _2/\gamma _2\) become
and
The upper arc joining \(\alpha _1(-\delta ,\gamma ;N)\) and \(\alpha _2(-\delta ,\gamma ;N)\) becomes the arc on the right, namely the one that joins \(z_1(-\delta ,\gamma ;N)\) and \(z_2(-\delta ,\gamma ;N)\) without touching the imaginary axis. If z is on the chord joining the points \(z_1(-\delta ,\gamma ;N)\) and \(z_2(-\delta ,\gamma ;N)\), then \(|z|\le \max (|z_1|,|z_2|)\). Using the property that successive Farey fractions \(-\delta /\gamma ,-\delta '/\gamma '\) in \({{\mathcal {F}}}_N\) satisfy the inequalities
one obtains the bound \(|z_i|< \frac{\sqrt{2}\gamma }{N}\), and hence
for any z on the chord. Moreover, using the triangle inequality, the length of the chord has to be \(\le |z_1|+|z_2|\), and hence it may not exceed
Now we perform the Rademacher expansion. We deform the initial contour in (C.11) to the Rademacher contour P(N), for some \(N\in {\mathbb {N}}\),
Next, we apply the modular transformation
to the integrand. Note that this maps the cusp \(-\delta /\gamma \) to \(+i\infty \).
Using (C.10), we obtain
Next, on each of the arcs we perform the z-change of variables (C.15), using
and obtain
Now we split the sum in the integral into two pieces, namely the polar part, which corresponds to \(m=-1\),
and the non-polar part,
Using the analysis given in [30], it can be shown that the non-polar part does not contribute in the limit \(N\rightarrow +\infty \). The polar part, on the other hand, is non-vanishing in this limit and given byFootnote 4
The sum over \(\delta \) can be written as a classical Kloosterman sum,
where \(\alpha \in {\mathbb {Z}}/\gamma {\mathbb {Z}}\) is the modular inverse of \(\delta \) and hence uniquely specified. By performing the change of variables
the z integral over the circle \(C_{1/2}\) becomes the t integral over the vertical line defined by \(\text {Re}(t) = 2\pi \),
which equals a sum of modified Bessel functions of the first kind,
and which results in
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Cardoso, G.L., Nampuri, S. & Rosselló, M. Rademacher Expansion of a Siegel Modular Form for \({{\mathcal {N}}}= 4\) Counting. Ann. Henri Poincaré (2023). https://doi.org/10.1007/s00023-023-01400-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00023-023-01400-3