Applications of CFT determinant formulas in number theory
D. Levin &, H.-G. Shin †, A. Zuevsky ∗ · 2025
Read Applications of CFT determinant formulas in number theory by D. Levin &, H.-G. Shin †, A. Zuevsky ∗ (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-01301.txt.
Introduction
Conformal field theory gives determinant formulas for correlation functions, and the three authors show those formulas have number-theoretic payoffs. At genus one, they compare fermionic and bosonic readings of a twisted torus partition function. With an elliptic determinant identity, that yields determinant formulas for powers of the classical eta-function, built from a twisted Szegö kernel. For certain exponents these echo the classical discriminant formulas of Garvan and Milne, though they do not recover them exactly.
At genus two, a similar comparison on a self-sewn torus gives a determinant formula for another eta power range in terms of genus two data. Which modular form that case specializes to is left open. The note is a bridge between field theory machinery and classical modular form identities. The head title is Determinant representations for Garvan formulas, though the catalog files it under a broader title.
Catalog record
| CREATOR | D. Levin &, H.-G. Shin †, A. Zuevsky ∗ |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 20 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-01301
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Applications of CFT determinant formulas in number theory D. Levin &, H.-G. Shin †, A. Zuevsky ∗ Abstract . In this note we show how to use the determinant representations for correlation functions in CFT to derive new determinant formulas for powers of the modular discriminant expressed via deformed elliptic functions with parameters. In particular, we obtain counterparts of Garvan’s formulas for the modular discriminant corresponding to the genus two Riemann surface case. 1991 Mathematics Subject Classification 17B69, 30F10, 32A25, 11F03 The authors state that: 1.) The paper does not contain any potential conflicts of interests. 2.) The paper does not use any datasets. No dataset were generated during and/or analysed during the current study. 3.) The paper includes all data generated or analysed during this study. 4.) Data sharing is not applicable to this article as no datasets were generated or analysed during the current study. 5.) The data of the paper can be shared openly. 6.) No AI was used to write this paper. 1. Introduction Many identities in number theory have been originated from computations of correlation functions of conformal field theory [ 5 , 15 , 1 , 8 , 12 , 13 , 14 ] . In particular, one finds relations among modular forms, fundamental kernels, and q -series from the comparison of bosonic and fermionic pictures. The most interesting example in this direction is represented by the twisted counterpart of Jacobi triple identity [ 5 , 8 ] . In computations of vertex operator algebra correlation functions on the torus, the classical Weierstrass functions and Eisenstein series naturally occur. Their two-parameter natural generalizations were introduced in [ 1 ] and further developed in [ 8 ] . For higher correlation functions on a genus one Riemann surface, computations come up elliptic versions of the Fay’s trisecant identity [ 2 ] known from algebraic geometry. Various identities for powers of the \eta(\tau) -function appear naturally in studies of affine Lie algebras [ 5 ] . In the genus two case, the bosonization procedure provides us with genus two counterparts of Jacobi triple identities in germs of determinants of reproduction kernels corresponding to the type of vertex operator algebra used. In this note, using the relations mentioned above, we derive new formulas for the modular discriminant generalizing Garvan identity for elliptic Eisenstein series to the case of genus two. 1.1. Torus correlation functions For an automorphism g twisted module \mathcal{V} for a vertex operator algebra V we find closed formulas for correlation functions of vertex operators {\mathcal{Y}} on the torus, q=e^{2\pi i\tau} , with local coordinates z_{i} , v_{i}\in V , 1\leq i\leq n , [ 8 ] : Z^{(1)}_{V}\left[\begin{array}[]{c}f\\ g\end{array}\right](v_{1},z_{1},\ldots,v_{n},z_{n};q)={\rm STr}_{\mathcal{V}}\left(f{\mathcal{Y}}(v_{1},z_{1})\ldots{\mathcal{Y}}(v_{n},z_{n})\;q^{L(0)-C/24}\right), where L(0) is the Virasoro algebra generator, and C is central charge. The formal parameter is associated to a complex parameter on the torus. Final expressions are given by determinants of matrices with elements being coefficients in the expansions of the regular parts of corresponding differentials: Bergman (bosons) or Szegő (fermions) kernels [ 8 ] . In this note we derive some new genus two generalizations of the fundamental formulas for powers of the \eta -function in terms of deformed versions [ 1 , 8 ] of Weierstrass functions and Eisenstein series. In particular, we find that powers of the modular discriminant are expressed (up to theta-functions multipliers) via determinants of finite matrices containing combinations of deformed modular functions. In the proof we use the generalized elliptic version of the Fay’s trisecant identity for a vertex operator superalgebra. 2. Modular discriminant and Eisenstein series The modular discriminant is defined by \Delta(\tau)=\eta(\tau)^{24} , where \eta(\tau) is the Dedekind eta-function, q=e^{2\pi i\tau} , \eta(\tau)=q^{1/24}\prod_{n=1}^{\infty}(1-q^{n}) . Recall that the Eisenstein series [ 9 , 10 ] E^{(1)}_{n}(\tau) is equal to 0 for n odd, and for n even E^{(1)}_{n}(\tau)=-\frac{B_{n}(0)}{n!}+\frac{2}{(n-1)!}\sum\limits_{r\geq 1}\frac{r^{n-1}q^{r}}{1-q^{r}}, where B_{n}(0) is the n -th Bernoulli number \frac{q_{z}^{\lambda}}{q_{z}-1}=\frac{1}{z}+\sum\limits_{n\geq 1}\frac{B_{n}(\lambda)}{n!}z^{n-1}. One finds [ 9 ] the relations: E^{(1)}_{8}=\left(E^{(1)}_{4}(\tau)\right)^{2} , E^{(1)}_{10}(\tau)=E^{(1)}_{4}(\tau)E^{(1)}_{6}(\tau) , E^{(1)}_{12}(\tau)=\tfrac{441}{691}\left(E^{(1)}_{4}(\tau)\right)^{3}+\tfrac{250}{691}\left(E^{(1)}_{6}(\tau)\right)^{2} . 2.1. Classical Gavan formula Then the fundamental classical formulas for the modular discriminant follow \Delta(\tau)=\frac{1}{1728}\left(\left(E^{(1)}_{4}(\tau)\right)^{3}-\left(E^{(1)}_{6}(\tau)\right)^{2}\right)=\frac{1}{1728}\det\left[\begin{array}[]{cc}\sqrt{7/3}E^{(1)}_{4}(\tau)&E^{(1)}_{6}(\tau)\\ E^{(1)}_{6}(\tau)&\sqrt{7/3}E^{(1)}_{8}(\tau)\end{array}\right]. The next formula is due to F. Garvan [ 3 ] \displaystyle\Delta^{2}(\tau)=-\frac{691}{250\;(1728)^{2}}\;\det\left[\begin{array}[]{ccc}E^{(1)}_{4}(\tau)&E^{(1)}_{6}(\tau)&E^{(1)}_{8}(\tau)\\ E^{(1)}_{6}(\tau)&E^{(1)}_{8}(\tau)&E^{(1)}_{10}(\tau)\\ E^{(1)}_{8}(\tau)&E^{(1)}_{10}(\tau)&E^{(1)}_{12}(\tau)\end{array}\right], which was then proved and generalized in [ 6 ] . In this talk we give various generalizations for higher powers of the modular discriminant computed as a determinant of matrices containing deformed Weierstrass functions [ 1 , 8 ] . 3. The generalized Garvan formulas Computations of the twisted partition function Z_{V}^{(1)}\left[\begin{array}[]{c}f\\ g\end{array}\right](\tau) for the free fermion vertex operator superalgebra leads to two alternative expressions (see, e.g., [ 5 , 8 ] ) as expansion over a basis: Z_{V}^{(1)}\left[\begin{array}[]{c}f\\ g\end{array}\right](\tau)=q^{\kappa^{2}/2-1/24}\prod_{l\geq 1}(1-\theta^{-1}q^{l-\frac{1}{2}-\kappa})(1-\theta q^{l-\frac{1}{2}+\kappa}), and \displaystyle Z_{V}^{(1)}\left[\begin{array}[]{c}f\\ g\end{array}\right](\tau)=\frac{e^{2\pi i(\alpha+1/2)(\beta+1/2)}}{\eta(\tau)}\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}-\beta\\ \frac{1}{2}+\alpha\end{array}\right](0,\tau), in terms of the torus theta series with characteristics: \vartheta^{(1)}\left[\begin{array}[]{c}a\\ b\end{array}\right](z,\tau)=\sum\limits_{n\in\mathbb{Z}}\exp\left(i\pi(n+a)^{2}\tau+(n+a)(z+2\pi ib)\right). Here we define f=e^{2\pi i\alpha a(0)} , g=e^{2\pi i\beta a(0)} , with some parameters \alpha , \beta\in\mathbb{R} , and a(0) is the zero mode of a Heisenberg subalgebra in the rank two free fermionic vertex operator superalgebra [ 8 ] . We also define \phi=e^{-2\pi i\beta} and \theta=e^{-2\pi i\alpha} . Note that Z_{V}^{(1)}\left[\begin{array}[]{c}f\\ g\end{array}\right](\tau)=0 for (\theta,\phi)=(1,1) , i.e., (\alpha,\beta)\equiv(0,0)\ (\mbox{mod}\ \mathbb{Z} ). Comparing two representations we obtain Jacobi triple product formula [ 5 ] which can be rewritten in the form: \eta(\tau)=\frac{q^{-\kappa^{2}/2+1/24}e^{2\pi i(\alpha+1/2)(\beta+1/2)}\;\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}-\beta\\ \frac{1}{2}+\alpha\end{array}\right](0,\tau)}{\det\left(I-T^{(1)}\right)}, where the determinant \det\left(I-T^{(1)}\right)=\prod_{l\geq 1}\left(1-\theta^{-1}q^{l-\frac{1}{2}-\kappa}\right)\left(1-\theta q^{l-\frac{1}{2}+\kappa}\right), corresponds to sphere self-sewing to form a torus [ 11 ] . Thus we get the identity for the first power of the \eta -function. In [ 1 , 8 ] the deformed Weierstrass functions (which can be expressed via deformed Eisenstein series) were defined and studied: P_{1}^{(1)}\left[\begin{array}[]{c}\theta\\ \phi\end{array}\right](z,\tau)=-\sum^{\prime}\limits_{n\in\mathbb{Z}+\lambda}\frac{q_{z}^{n}}{1-\theta^{-1}q^{n}}=\frac{1}{z}-\sum\limits_{n\geq 1}\frac{1}{n}E^{(1)}_{n}\left[{\begin{array}[]{c}\theta\\ \phi\end{array}}\right](\tau)z^{n-1}, for q=e^{2\pi i\tau} , and where \sum\limits^{\prime} means we omit n=0 if (\theta,\phi)=(1,1) , and \displaystyle E^{(1)}_{n}\left[{\begin{array}[]{c}\theta\\ \phi\end{array}}\right](\tau) \displaystyle= \displaystyle-\frac{B_{n}(\lambda)}{n!}+\frac{1}{(n-1)!}\sum\limits_{r\geq 0}^{\prime}\frac{(r+\lambda)^{n-1}\theta^{-1}q^{r+\lambda}}{1-\theta^{-1}q^{r+\lambda}} \displaystyle+\frac{(-1)^{n}}{(n-1)!}\sum\limits_{r\geq 1}\frac{(r-\lambda)^{n-1}\theta q^{r-\lambda}}{1-\theta q^{r-\lambda}}. In addition to that, for integers m_{i} , n_{j}\geq 0 , satisfying \sum_{i=1}^{r}m_{i}=\sum_{j=1}^{s}n_{j} , let us introduce the notation \displaystyle\Theta^{(1)}_{r,s,(m_{i},n_{i})}(x,y,\tau) \displaystyle=\frac{\prod\limits_{1\leq i\leq r,1\leq j\leq s}\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}\\ \frac{1}{2}\end{array}\right](x_{i}-y_{j},\tau)^{m_{i}n_{j}}}{\prod\limits_{1\leq i<k\leq r}\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}\\ \frac{1}{2}\end{array}\right](x_{i}-x_{k},\tau)^{m_{i}m_{k}}\prod\limits_{1\leq j<l\leq s}\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}\\ \frac{1}{2}\end{array}\right](y_{j}-y_{l},\tau)^{n_{j}n_{l}}}. Let us introduce {\bf P}_{n}(\theta,\phi) , a the n\times n matrix, 1\leq i , j\leq n , {\bf x}=\left(x_{1},\ldots,x_{n}\right) , {\bf y}=\left(y_{1},\ldots,y_{n}\right) , S^{(1)}({\bf x},{\bf y})={\bf P}_{n}(\theta,\phi)=\left[P^{(1)}_{1}\left[\begin{array}[]{c}\theta\\ \phi\end{array}\right](x_{i}-y_{j},\tau)\right], and another (n+1)\times(n+1) matrix {\bf Q}_{n}(\tau) : {\bf Q}_{n}(\tau)=\left[\begin{array}[]{cc}&1\\ {\bf P}_{n}(1,1)&\vdots\\ &1\\ 1\ldots 1&0\end{array}\right]=\left[\begin{array}[]{cccc}P^{(1)}_{1}(x_{1}-y_{1},\tau)&\ldots&P^{(1)}_{1}(x_{1}-y_{n},\tau)&1\\ \vdots&\ddots&&\vdots\\ P^{(1)}_{1}(x_{n}-y_{1},\tau)&&P^{(1)}_{1}(x_{n}-y_{n},\tau)&1\\ 1&\ldots&1&0\end{array}\right]. Proposition 1 . Generalizing Garvan’s formula, or (\theta,\phi)\neq(1,1) one has (5) \Delta^{n}(\tau)=-\frac{\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}\\ \frac{1}{2}\end{array}\right](0,\tau)\;\Theta^{(1)}_{8n,8n,(1,1)}(x,y,\tau).}{\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}-\beta\\ \frac{1}{2}+\alpha\end{array}\right]\left(\sum\limits_{i=1}^{8n}(x_{i}-y_{i}),\tau\right)}\;\mathrm{\det}\;{\bf P}_{8n}(\theta,\phi), for (\theta,\phi)=(1,1) , (6) \Delta^{n}(\tau)=i\;\frac{\Theta^{(1)}_{8n+1,8n+1,(1,1)}(x,y,\tau)}{\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}\\ \frac{1}{2}\end{array}\right]\left(\sum\limits_{i=1}^{8n+1}(x_{i}-y_{i}),\tau\right)}\;\mathrm{\det}\;{\bf Q}_{8n+1}. Formulas ( 5 )–( 6 ) can be also expressed in terms of deformed Eisenstein series by substitution of the definition of P^{(1)}_{1}\left[\begin{array}[]{c}\theta\\ \phi\end{array}\right](z,\tau) in terms of E^{(1)}_{n}\left[\begin{array}[]{c}\theta\\ \phi\end{array}\right](\tau) leading to a quite involved formula which we do not give here. Let us give a proof. Recall the genus one prime form [ 7 , 2 ] K^{(1)}(x-y,\tau)=\frac{\vartheta^{(1)}\left[{{\theta}\atop{\phi}}\right]\left(\int_{y}^{x}\nu,\tau\right)}{\zeta(x)^{\frac{1}{2}}\zeta(y)^{\frac{1}{2}}} , where \zeta^{(1)}(x)=\sum_{i=1}^{g}\partial_{z_{i}}\vartheta^{(1)}\left[{\gamma}\atop{\delta}\right](0,\Omega)\nu^{(1)}_{i}(x) which gives K^{(1)}(z,\tau)=-\frac{i}{\eta^{3}(\tau)}\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}\\ \frac{1}{2}\end{array}\right](z,\tau) . In particular, for \alpha=\beta=1/2 one has Z^{(1)}\left[{f_{1/2}\atop g_{1/2}}\right]\left(\tau\right)=K^{(1)}(z,\tau)/\eta^{2}(\tau) . In [ 8 ] the elliptic function version of the Fay’s generalized trisecant identity [ 2 ] was derived. Introduce K^{(1)}_{n}({\bf x},{\bf y})=\frac{\prod\limits_{1\leq i<j\leq n}K^{(1)}(x_{i}-x_{j},\tau)K^{(1)}(y_{i}-y_{j},\tau)}{\prod\limits_{1\leq i,j\leq n}K^{(1)}(x_{i}-y_{j},\tau)} . For (\theta,\phi)\neq(1,1) one has \displaystyle\mathrm{\det}\;{\bf P}_{n}(\theta,\phi)=\frac{\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}-\beta\\ \frac{1}{2}+\alpha\end{array}\right]\left(\sum\limits_{i=1}^{n}(x_{i}-y_{i}),\tau\right)}{\vartheta^{(1)}\left[\begin{array}[]{c}\frac{1}{2}-\beta\\ \frac{1}{2}+\alpha\end{array}\right](0,\tau)}K^{(1)}_{n}({\bf x},{\bf y}), and similarly for (\theta,\phi)=(1,1) , \det\;{\bf Q}_{n}=-K^{(1)}\left(\sum\limits_{i=1}^{n}(x_{i}-y_{i}),\tau\right)K^{(1)}_{n}({\bf x},{\bf y}). Then Proposition follows. 3.1. Higher power formulas There exist also the analytic expansion [ 8 ] , for k , l\geq 1 , \displaystyle P^{(1)}_{1}\left[{\begin{array}[]{c}\theta\\ \phi\end{array}}\right](z+z_{1}-z_{2},\tau)