---
title: Torus One-Point Conformal Blocks
url: https://www.emergentmind.com/topics/torus-one-point-conformal-blocks
type: topic
---

# Torus One-Point Conformal Blocks

Torus one-point conformal blocks are the holomorphic constituents in the conformal-block decomposition of one-point functions on a torus. In the Virasoro setting they depend on the modular parameter \(q\), the central charge \(c\), an external weight, and an intermediate weight, and they encode the contribution of a single Verma module inside the torus trace. They occupy a distinctive position among two-dimensional conformal blocks: the one-point torus case is the unique \(k=1\) necklace-channel block, it admits recursive, integral, probabilistic, and holographic descriptions, and in Liouville theory it is tied to classical uniformization, gauge theory, and isomonodromic systems [0911.2353] [1309.7672].

## 1. Definition, geometry, and basic algebraic structure

The torus is commonly realized either as \(T\simeq \mathbb{R}^2/(\mathbb{Z}+\tau\mathbb{Z})\) or, after exponentiation, by the identification \(z\sim qz\) with \(q=e^{2\pi i\tau}\). In the necklace-channel formulation, the one-point block is the simplest torus block: a closed loop with a single external leg, and for the one-point function there is only one channel [2205.05038].

In Liouville theory the torus one-point correlator with a primary insertion is written as a trace over the Hilbert space,
\[
\langle V_B(1)\rangle_\tau
=
\mathrm{Tr}_{\mathcal H_L}\!\left(
q^{L_0-c/24}\,\bar q^{\bar L_0-c/24}\,V_B(z,\bar z)
\right)\Big|_{z=1},
\qquad q=e^{2\pi i\tau},
\]
and after decomposition into conformal blocks and structure constants one introduces the holomorphic torus one-point block \(\mathcal F_{c,\Delta}^{\lambda}(q)\) [1309.7672]. In the descendant basis \(\nu_{\Delta,M}=L_{-M}\nu_\Delta\), with inverse Gram matrix \(\big[B^n_{c,\Delta}\big]^{MN}\), the block takes the form
\[
\mathcal{F}_{c,\Delta}^{\lambda}(q)
=
q^{\Delta-\frac{c}{24}}
\sum_{n=0}^{\infty} q^n F^{\lambda,n}_{c,\Delta},
\]
with coefficients defined by descendant three-point matrix elements and inverse Gram matrices. The torus one-point function then decomposes as
\[
\langle \phi_{\lambda,\bar\lambda}\rangle
=
\sum_{(\Delta,\bar\Delta)}
\mathcal{F}_{c,\Delta}^{\lambda}(q)\,
\mathcal{F}_{c,\bar\Delta}^{\lambda}(\bar q)\,
C^{\lambda,\bar\lambda}_{\Delta,\bar\Delta}
\]
[0911.2353].

A characteristic simplification is that the torus one-point block is independent of the insertion point. This is emphasized in large-\(c\) analyses, where the one-point torus block is described as a \(q\)-series whose nontrivial dependence is entirely encoded in the modular parameter and conformal data [1612.05891].

## 2. Recursive, elliptic, and integral representations

A central structural result is the Zamolodchikov-type recursive representation of the torus one-point Virasoro block. The coefficients \(F^{\lambda,n}_{c,\Delta}\) are rational in \(c\) and \(\Delta\), polynomial in the external weight, and their poles are controlled by degenerate Virasoro representations. Passing to the elliptic normalization,
\[
\mathcal{H}_{c,\Delta}^{\lambda}(q)
=
q^{\frac{c-1}{24}-\Delta}\,\eta(q)\,\mathcal{F}_{c,\Delta}^{\lambda}(q)
=
\sum_{n=0}^\infty q^n H^{\lambda,n}_{c,\Delta},
\]
one obtains the recursion
\[
H^{\lambda,n}_{c,\Delta}
=
\delta^n_0
+
\sum_{1\le rs\le n}
\frac{
A_{rs}(c)\,
P^{rs}_{c}\!\left[^{\Delta_\lambda}_{\Delta_{rs}+rs}\right]
P^{rs}_{c}\!\left[^{\Delta_\lambda}_{\Delta_{rs}}\right]
}{
\Delta-\Delta_{rs}(c)
}
\,H^{\lambda,n-rs}_{c,\Delta_{rs}+rs}.
\]
This recursion computes the torus elliptic block from lower levels, with residues determined by singular vectors and fusion polynomials. The same framework proves identities conjectured by Poghossian that relate torus one-point elliptic blocks to specially chosen sphere four-point elliptic blocks [0911.2353].

The recursive representation also supports direct numerical work. One application is a modular-invariance test of the Liouville torus one-point function, carried out by evaluating the elliptic block through recursion truncated at finite level and comparing the modular transform numerically for \(c=2\), \(\lambda=i\), and \(\tau\in[0.2i,5i]\) with truncations \(n=4,5,6\) [0911.2353].

Integral realizations provide a complementary description. A Dotsenko–Fateev or elliptic \(\beta\)-ensemble representation on the torus was proposed with a crucial A-cycle screening contour and a genus-one shift of the potential. In that representation the torus partition function \(Z_{\rm DF}(q)\) reproduces the torus one-point conformal block after inclusion of a universal \(U(1)\)-type prefactor,
\[
\prod_{m\ge 1}(1-q^m)^{\nu} \left(1+j_1 q+j_2 q^2+\cdots\right)
=
1+B_1 q+B_2 q^2+\cdots,
\]
with explicit parameter identification between \((A,\beta,N)\) and \((A_{\rm ext},\Delta,c)\). The agreement was checked explicitly through the first two levels, including several finite-\(N\) cases such as \((N,\beta)=(1,1),(2,1),(3,1),(1,2),(2,2),(3,2),(1,3),(2,3)\) [1010.1734].

## 3. Global blocks and large-\(c\) asymptotics

In the large-\(c\) regime the Virasoro algebra truncates to the global \(\mathfrak{sl}(2,\mathbb{R})\) subalgebra, and the torus one-point block acquires exact hypergeometric expressions. In one normalization the global block is
\[
F(\Delta,\tilde\Delta\mid q)
=
(1-q)^{\Delta-1}\,{}_2F_1(2\tilde\Delta-\Delta,1-\Delta,2\tilde\Delta\mid q),
\]
while in shadow-formalism conventions the one-point torus global block factorizes as
\[
F^{h}(q,z)=z^{-h}(1-q)^{1-h}\,
q\; {}_2F_1\!\left(
\begin{array}{c}
h,\; h+1\\[2pt]
2h
\end{array}\Bigg|\, q \right).
\]
For \(h=0\), the latter reduces to \(q/(1-q)\), identified as the \(sl(2,\mathbb R)\) character of the relevant Verma module [1612.05891] [2307.12061].

The global block is also characterized by a second-order Casimir equation. In the thermal-trace normalization of the torus one-point global block \(\mathcal F(h_\alpha,h_1;q)\), the quadratic \(\mathfrak{sl}(2)\) Casimir yields
\[
\Big[
q(1-q)^2\partial_q^2
-2q(1-q)\partial_q
-h_\alpha(h_\alpha-1)\frac{(1-q)^2}{q}
-h_1(h_1-1)
\Big]\mathcal{F}(h_\alpha,h_1;q)=0,
\]
together with the small-\(q\) asymptotic behavior \(\mathcal F(h_\alpha,h_1;q)\sim q^{h_\alpha}\) [1706.00047].

Large-\(c\) analysis on the torus distinguishes several non-identical limits. One work separates the global block, the light block, the heavy-light block, and the linearized classical block. In particular, the torus light block differs from the global block by the inverse Euler function,
\[
\mathcal{L}(\Delta,\tilde\Delta\mid q)=\frac{1}{p(q)}\,F(\Delta,\tilde\Delta\mid q),
\qquad
p(q)=\prod_{n=1}^\infty(1-q^n)^{-1},
\]
and the heavy-light block reduces to
\[
H(q)=\frac{1}{p(q)}=\prod_{n=1}^\infty(1-q^n).
\]
The same work relates these blocks to different Inönü–Wigner contractions of the Virasoro algebra [1612.05891].

At the genuinely semiclassical level, where \(c\to\infty\) and all relevant weights scale linearly with \(c\), torus one-point blocks exponentiate. This had been established for the torus one-point case before the general higher-genus theorem, and a later theorem subsumed it into the statement that arbitrary Virasoro blocks on arbitrary Riemann surfaces exponentiate as formal power series,
\[
F(c,c\vec g_\mathrm{ext},c\vec g_\mathrm{int};\vec x)
\sim
\exp\!\left(-\frac{c}{6}I(\vec g_\mathrm{ext},\vec g_\mathrm{int};\vec x)\right)
\]
[2606.18345].

## 4. Classical torus blocks, Liouville theory, and the Lamé accessory parameter

In Liouville theory the torus one-point block has a classical limit obtained by
\[
b\to 0,\qquad c=1+6Q^2,\qquad Q=b+b^{-1},
\]
with heavy weights scaled as \(\Delta=b^{-2}\delta\) and \(\Delta_B=b^{-2}\delta_B\). In this limit the torus block exponentiates,
\[
\mathcal F_{c,\Delta}^{\Delta_B}(q)
\sim
\exp\!\left[\frac{1}{b^2}\, f_\delta^{\delta_B}(q)\right],
\]
and \(f_\delta^{\delta_B}(q)\) is the classical torus conformal block [1309.7672].

For the one-punctured torus, the classical Liouville action is proposed in saddle-point form as the torus analog of Zamolodchikov’s classical factorization formula. The saddle-point intermediate weight is
\[
\delta_*=\frac14+p_*^2,
\]
and the classical action is obtained by evaluating the classical torus block at the saddle-point momentum,
\[
S_{\text{torus}}(\delta;q)
=
S^{(3)}(2p_*,-i\delta,2p_*)-2\,\mathrm{Re}\,f_{1/4+p_*^2}(q).
\]
This classical block determines the accessory parameter of the Lamé equation associated with uniformization of the one-punctured torus [1309.7672].

The relevant differential equation is the Lamé equation in Weierstrass form,
\[
\frac{d^2\Psi}{dz^2} - \bigl[\kappa\,\wp(z)+B\bigr]\Psi=0,
\]
and the accessory parameter is extracted from the classical action or, equivalently, from the classical torus block. One explicit formula is
\[
B(\tau)=4\pi^2\left[
p_*^2(q,\delta)+\frac14(1-2\delta)E_2(\tau)
\right].
\]
In the parabolic case \(\delta=1\), the algebraic accessory parameter satisfies \(C=-4B\) [1309.7672].

A probabilistic proof of the torus analog of Zamolodchikov’s conjecture later established the semiclassical exponential structure rigorously for Liouville conformal blocks on the one-punctured torus. Under the scaling \(\alpha=\alpha_0/\gamma\), \(P=P_0/\gamma\), the probabilistic block \(\mathcal G_{\gamma,P}^{\alpha}(q)\) satisfies
\[
\lim_{\gamma \rightarrow 0} \gamma^2 \log \mathcal{G}_{\gamma, P}^{\alpha}(q)
=
\phi(\alpha_0, P_0, q)+ \widetilde{f}(\alpha_0,P_0),
\]
with a positive radius of convergence in \(q\). In the degenerate limit with \(\chi=\gamma/2\), the deformed block produces a Lamé equation,
\[
\left(\partial_{z}^2
- \frac{\alpha_0}{4}\left(\frac{\alpha_0}{4}-1\right)\wp(z)
+ \frac{\pi }{2}\partial_\tau {\phi}(q)\right)\widetilde{\Gamma}=0,
\]
so that the accessory parameter is \(\frac{\pi}{2}\partial_\tau\phi(q,\alpha_0,P_0)\) [2407.05839].

## 5. Gauge-theoretic, isomonodromic, and holographic correspondences

A major correspondence identifies the torus one-point Virasoro block with instanton data of four-dimensional \(\mathcal N=2^*\), \(U(2)\) supersymmetric gauge theory. In the Nekrasov–Shatashvili limit \(\epsilon_2\to 0\) with \(\epsilon_1\) fixed,
\[
Z^{\mathcal N=2^*,U(2)}_{\rm inst}
\sim
\exp\!\left(
\frac{1}{\epsilon_2}\,\mathcal W_{\rm inst}(q,a,\mu,\epsilon_1)
\right),
\]
and the classical torus block is related to the effective twisted superpotential by
\[
f_\delta^\delta(q)=(1-2\delta)\log q -2\,\mathcal W_{\rm inst}(q,a,\mu,\epsilon_1),
\]
up to normalization conventions. The same paper further expresses the accessory parameter through the sum of rescaled column lengths \(\omega_{u,i}(q)\) of critical Young diagrams extremizing the instanton free energy, and rewrites that sum as a contour integral built from the special functions \(Y(z)\) and \(Y_0(z)\) [1309.7672].

The torus one-point block also enters the Painlevé/CFT correspondence. For a torus \(C_{1,1}\) with one simple pole, the isomonodromic tau-function of the non-autonomous elliptic Calogero–Moser system admits a Fredholm determinant representation, and the determinant expansion reorganizes into a discrete Fourier series of torus one-point conformal blocks. In that framework,
\[
\mathrm{tr}_{\mathcal{V}_{a+n/2}}\!\left(q^{L_0}V_m(0)\right)=\mathcal B(a+n/2,m,q),
\]
and the tau-function is expressed as a sum over shifted internal momenta \(a+n/2\), with coefficients controlled by monodromy data \(\nu\) and \(\rho\) [2305.04240].

On the holographic side, the global torus one-point conformal block has an exact thermal-AdS description. A Witten-diagram representation is
\[
W_1(h_\alpha,h_1;q)
=
\int_{\text{Thermal AdS}} d^3x\,\sqrt{g}\;
G_b^{(h_1)}(x,w)\, G_1^{(h_\alpha)}(x,q),
\]
where the internal propagator is restricted to a single thermal winding. The same Casimir operator that defines the CFT block acts on the bulk propagator, and the Witten diagram obeys the same differential equation and small-\(q\) asymptotics as the global block [1706.00047]. In a semiclassical linearization, the torus one-point classical block is instead represented by a tadpole geodesic graph in thermal AdS, with one loop on the non-contractible cycle and one radial leg attached to the boundary insertion [1603.08440].

## 6. Extensions, variants, and applications

A rigorous non-perturbative construction of the one-point toric Virasoro conformal block was developed using Gaussian multiplicative chaos for Liouville CFT with \(c>25\). In that construction a normalized GMC expectation \(G^\alpha_{\gamma,P}(q)\) is analytic near \(q=0\), and its \(q\)-series coincides with the standard torus Virasoro block. For \(\alpha\in[0,Q)\), the analytic continuation extends to an open set containing \(\{z\in\mathbb C:|z|<1/2\}\cup[0,1)\), implying radius of convergence at least \(1\) [2003.03802].

Global and higher-rank generalizations have been developed through shadow formalisms. For \(\mathfrak{sl}_2\), shadow projectors reproduce the known torus one-point global block; for \(\mathfrak{sl}_3\), the same method yields an exact integral representation for the one-point torus block of \(\mathcal W_3\) theory in the large-\(c\) limit, under a restriction on the external field chosen to avoid fusion multiplicities [2401.02580]. In the \(A_{n-1}\) Toda setting, the light asymptotic limit of the one-point torus \(W_n\) conformal block is obtained through AGT as an explicit sum over \(n\) Young diagrams, with a selection rule stating that only boxes with specific arm lengths contribute in the \(b\to 0\) limit [2604.01804].

Supersymmetric analogues display a parallel but richer structure. In the \(N=1\) Neveu–Schwarz sector, large-\(c\) one-point torus superblocks split into lower and upper components, satisfy \(\operatorname{osp}(1|2)\) super-Casimir equations reducible to Heun equations, and organize into global, light, and heavy-light superblocks related by contractions of the NS superalgebra [1805.12585]. A later shadow-formalism derivation showed that the torus one-point superconformal block in the large-\(c\) or global \(\mathfrak{osp}(1|2)\) limit is an explicit linear combination of two ordinary \(\mathfrak{sl}(2)\) torus blocks with internal weights \(\Delta\) and \(\Delta+\tfrac12\) [2408.07684].

Several applications use torus one-point conformal blocks as computational input rather than as formal objects in isolation. In critical loop models, torus one-point functions are expressed as infinite linear combinations of torus one-point conformal blocks, and a sphere–torus relation maps them to sphere four-point functions at a different central charge, so that sphere crossing symmetry implies modular covariance on the torus [2604.24491]. In two-dimensional string cosmology, torus one-point Liouville correlators in the \(c=1\) and \(c=25\) sectors are decomposed into torus one-point Virasoro conformal blocks, recursively truncated up to \(q^8\), and integrated over the torus fundamental domain; the resulting torus one-point diagram is numerically fit by
\[
\Psi^{g=1}_{\text{1-pt}}(\omega)
=
\frac{g_s C_{T^2}}{24}\,\omega(1+2\omega^2),
\]
with reported maximum discrepancy \(10^{-5}\%\) [2304.13043].

Taken together, these developments show that torus one-point conformal blocks form a nexus between algebraic representation theory, elliptic recursion, semiclassical Liouville theory, supersymmetric gauge theory, isomonodromy, thermal AdS holography, and mathematically rigorous probabilistic constructions. A plausible implication is that the one-point torus block serves not only as the simplest nontrivial genus-one conformal block, but also as a test case in which several deep correspondences can be formulated in exact, computable form.

Source: https://www.emergentmind.com/topics/torus-one-point-conformal-blocks