betaFeedback

WorksPhysics and Astronomy

Construction of holomorphic SCFTs with minimal index

Abstract

It has been an open question whether there exist holomorphic VOAs with N=1 supersymmetry and Witten index 1 for any positive central charge c. The existence of such a theory with c=288 is closely related to the 576-periodicity of the TMF spectrum via the Segal-Stolz-Teichner conjecture. Here we report the construction of thousands of such theories by extending the approach initiated by Gaiotto and Johnson-Freyd via ternary codes. In addition, for every positive integer k, we also construct holomorphic N=1 VOAs of central charge c=12k with index 24/gcd(24,k) – conjectured to be the minimum – in a computer-aided search carried out by AI coding agents. We overcome the exponential-time barrier by using chains of smaller cells to build ternary codes, whose indices can then be evaluated via a transfer-matrix technique, and we use a symmetric orbifold construction to reduce the problem of k=12 and 24 to that of k=6. Along the way, we also establish several results on ternary codes and their indices. Using an observation relating codes and Landau-Ginzburg orbifold models, we argue that the non-chiral doubles of the ternary code VOAs can always be gapped supersymmetrically.

Introduction

One surprising prediction of the Segal–Stolz–Teichner conjecture [1, 2, 3] is that the elliptic genus of a 2d2d (0,1)(0,1) theory cannot be an arbitrary modular form, but has to live in the image of a ring homomorphism from topological modular forms,

π∗(TMF)⟶MF∗/2.(1)\pi_{*}(\mathrm{TMF}) \longrightarrow\mathrm{MF}_{*}/2. \tag*{(1)}

See [4, 5, 6, 7, 8, 9] for a small collection of works on tests of this constraint on physical theories and its applications, and [10, 11] for an overview as well as a comprehensive review of the subject of topological modular forms.

In particular, it predicts that a holomorphic SCFT TT of central charge c=12kc = 12k must have Witten index I(T)\mathcal{I}(T) satisfying

k⋅I(T)≡0(mod24).(2)k \cdot\mathcal{I}(T) \equiv0 \pmod{24}. \tag*{(2)}

Equivalently, the index I\mathcal{I} must be divisible by 24/gcd⁡(24,k)24/\operatorname{gcd}(24,k).

It has been an open question, first asked in [4], whether the minimum,

Imin⁡=24gcd⁡(24,k),(3)\mathcal{I}_{\min} = \frac{24}{\operatorname{gcd}(24,k)}, \tag*{(3)}

can always be attained by a holomorphic SCFT for arbitrary kk.1 One particularly interesting case is when k∈24Zk \in24\mathbb{Z}, where the minimum positive value of I\mathcal{I} is 1. In this work, we answer this question in the affirmative by explicitly constructing such SCFTs.

Status and strategy

The case of k=0k = 0 is trivial, where one can take TT to be the trivial theory with a single bosonic massive vacuum. For k=1k = 1, the super-Moonshine module [12] has index 24, which is minimal. For k=2k = 2 through 5, the theories with minimal indices have been found by Gaiotto and Johnson-Freyd in the foundational work [4]. Their approach is based on the construction of holomorphic SCFTs via self-dual ternary codes (reviewed in § 2) and then gauging (a.k.a. orbifolding) an internal discrete symmetry of the theory, as illustrated below

self-dual ternary codes with automorphism →3-framed lattice VOA\xrightarrow{\text{3-framed lattice VOA}} holomorphic SCFTs with symmetry →orbifolding\xrightarrow{\text{orbifolding}} holomorphic SCFTs with smaller index {#eq:1.4}

The approach taken in [4] to find self-dual ternary codes with a desired automorphism is via a computer-aided random search that scales exponentially with kk. Therefore, it would seem unrealistic to hope that one can eventually reach k=24k = 24 with Imin⁡=1\mathcal{I}_{\min} = 1—or even k=12k = 12 with Imin⁡=2\mathcal{I}_{\min} = 2—by brute force.

Our solution to the k=24k = 24 problem is based on the observation that, in order to obtain such a theory with index 11, one just needs a theory with k=6k = 6 (i.e. c=72c = 72) and index ±4\pm4, to which one applies a “double symmetric square.”

Symmetric squares

Given a holomorphic SCFT TT with index I\mathcal{I}, one can construct a symmetric square of the theory

Sym⁡2(T):=(T⊗T)/Z2(4)\operatorname{Sym}^{2}(T) := (T \otimes T)/\mathbb{Z}_{2} \tag*{(4)}

where Z2\mathbb{Z}_{2} acts by exchanging the two copies. The index of the new theory is

I(Sym⁡2(T))=(I+12)+I=I(I+3)2.(5)\mathcal{I}(\operatorname{Sym}^{2}(T)) = \binom{\mathcal{I}+1}{2} + \mathcal{I} = \frac{\mathcal{I}(\mathcal{I}+3)}{2}. \tag*{(5)}

The two terms in the middle are indices of the untwisted and twisted sectors.2

This simple-looking formula has a surprising consequence: if one starts with a theory TT whose I=−4\mathcal{I} = -4, then T′:=Sym⁡2TT' := \operatorname{Sym}^{2}T will have index

I(T′)=−4⋅(−4+3)2=2,(6)\mathcal{I}(T') = \frac{-4 \cdot(-4+3)}{2} = 2, \tag*{(6)}

and its own symmetric square after stacking with the Arf theory, T′′=Sym⁡2(T′⊗Arf⁡)T'' = \operatorname{Sym}^{2}(T' \otimes\operatorname{Arf}), has index

I(T′′)=−2⋅(−2+3)2=−1,(7)\mathcal{I}(T'') = \frac{-2 \cdot(-2+3)}{2} = -1, \tag*{(7)}

which would then achieve our goal. This can be summarized in the diagram below:

T (index −4)→Sym2T′ (index 2)→⊗ Arf−T′ (index −2)→Sym2T′′ (index −1)(8)\begin{aligned} T\ (\text{index }-4) &\xrightarrow{\mathrm{Sym}^2} T'\ (\text{index }2) \xrightarrow{\otimes\,\mathrm{Arf}} -T'\ (\text{index }-2) \xrightarrow{\mathrm{Sym}^2} T''\ (\text{index }-1) \tag*{(8)} \end{aligned}

Alternatively, an index-1 theory can also be achieved by a single twisted D4D_4 orbifold of a theory with index ±4\pm4 as illustrated in Figure 1 and explained in § 3.3.

The sectors of the twisted $D_4$ orbifold of $T^{\otimes 4}$

Figure 1. The sectors of the twisted D4D_4 orbifold of T⊗4T^{\otimes4} that gives the theory T′′T'' with index −1-1, labeled by the conjugacy class of the holonomy g∈D4g \in D_4 around the spatial circle. The four copies of TT (colored dots on the fiber over a point) are permuted by gg, so going once around the circle joins them into the strands shown. The small square below each panel shows the conjugacy class of D4D_4 via its action on a square, with the four theories at the corners. Below each panel is the contribution of these sectors to the Witten index, as a polynomial in I=I(T)\mathcal{I}=\mathcal{I}(T) and its value at I=−4\mathcal{I}=-4. The total is thus 15−24+10−6+4=−115-24+10-6+4=-1.

Here we have used the fact that, in the world of holomorphic SCFTs, there is a Z2\mathbb{Z}_2 involution3

T⟶−T:=T⊗Arf(9)T \longrightarrow-T := T \otimes\mathrm{Arf} \tag*{(9)}

given by stacking with an Arf theory—an invertible spin TQFT generating Ω2Spin\Omega^{\mathrm{Spin}}_2—to flip the parity of Ramond-sector ground states, which then flips the sign of the Witten index. Therefore, getting a theory with index −1-1 is as good as a theory with index 11, and, for the “seed” theory TT, having index 44 is as good as −4-4 as we can simply use −T-T as the seed instead. In the diagram (1.9), we do need to use this operation to flip the middle theory T′⟶−T′T' \longrightarrow-T' as otherwise the index one gets would be 55 instead. Alternatively, one can combine the last two arrows as a twisted orbifold—discrete gauging but with a “theta angle” given by an invertible topological phase (or SPT).

Because of the existence of this double symmetric square, to find a holomorphic SCFT with minimal index for k=12k = 12 and 2424, it is good enough to find such a theory with k=6k = 6 and index I=±4\mathcal{I} = \pm4. As this is just slightly outside of the range studied in [4], it is actually possible via a brute-force random search with a required symmetry. To find more examples and, in particular, conceptually simple ones, we used agentic AI systems including GPT-6 Astra, GPT-6.1 Sol, Opus 5.5, and Fable 5.1. They produced and executed a collection of scripts in Python 3 using NumPy [14] to find thousands of such seed theories within a short time span of a week, as well as the minimal-index representatives for k=8,9,10,11,13,14,15,16,18,21k = 8, 9, 10, 11, 13, 14, 15, 16, 18, 21. For k=8k = 8 and 1616, we use an orbifold construction by a new Q8Q_{8} symmetry.

The number of distinct codes found and the estimated average CPU time per code for each kk are documented in Table 1.

kkcodescell structurecandidatesseconds per code
68,748(12,3)3(12, 3)^32⋅1052 \cdot10^52.3
782,1883×(12,3)+(6,3)3 \times(12, 3) + (6, 3)7⋅1067 \cdot10^60.1
84,8844×(12,4)4 \times(12, 4) with Q8Q_{8} symmetry4⋅1064 \cdot10^60.5
9407[(12,3)+(6,3)]3[(12, 3) + (6, 3)]^32⋅1072 \cdot10^740
105,0735×(12,3)5 \times(12, 3)6⋅1066 \cdot10^61.5
112,5115×(12,3)+(6,3)5 \times(12, 3) + (6, 3)8⋅1068 \cdot10^65
134696×(12,3)+(6,3)6 \times(12, 3) + (6, 3)9⋅1069 \cdot10^640
142337×(12,3)7 \times(12, 3)7⋅1067 \cdot10^670
1553[2×(12,4)+(6,4)]3[2 \times(12, 4) + (6, 4)]^32⋅1072 \cdot10^7∼1000\sim1000
16678×(12,4)8 \times(12, 4) with Q8Q_{8} symmetry6⋅1086 \cdot10^8∼3000\sim3000
1835[3×(12,4)]3[3 \times(12, 4)]^32⋅1082 \cdot10^8∼9000\sim9000
219[3×(12,4)+(6,4)]3[3 \times(12, 4) + (6, 4)]^310810^8∼50000\sim50000

Table 1. Yield of the computer search. The codes column counts the distinct codes found by the search. A “candidate” is a random choice of the cells in a chain (Figure 2), and (a,b)(a, b) denotes a cell with a total and bb boundary spins. “+” and “×\times” join cells chosen independently, while an exponent repeats the block. For each kk, a collection of different cell structures is often used in the search, while here we only list the one used for benchmarking. The candidates column counts those benchmarked, which the CPU time estimate is based on.

For higher values of kk, the search was able to succeed by looking for ternary codes built from a closed chain of smaller “cells” (blocks of 6, 12 or 18 “spins” coupled only to their neighbors)—so that the index is a trace of a product of “transfer matrices” and can be evaluated rapidly (in milliseconds)—with the desired automorphism then being a shift along the chain. This is illustrated in Figure 2 with more details in later sections.

Structure of a ternary code presented as a closed chain of cells

Figure 2. Structure of a ternary code presented as a closed chain of LL cells (with L=6L = 6 in the figure). Each cell carries 12 ternary “spins” and shares rr “boundary spins” with the next cell. The shift along the chain is the symmetry (Z6\mathbb{Z}_6 in the figure) to be gauged. The index of the code is then the trace of the LL-th power of a 3r×3r3^r \times3^r transfer matrix MM, with the periodic (shown in the figure) or an anti-periodic boundary condition, as explained in § 3.2.

Since the index is multiplicative under tensor products, we can directly obtain minimal representatives saturating the TMF bound for k=17=8+9k = 17 = 8 + 9, 19=16+319 = 16 + 3, 20=8+1220 = 8 + 12, 22=16+622 = 16 + 6, 23=8+1523 = 8 + 15, and for every k>24k > 24 by tensoring with the index-1 theory at k=24k = 24. This proves

Theorem 1. For every positive integer kk, there is a holomorphic N=1\mathcal{N} = 1 VOA of central charge c=12kc = 12k with index 24gcd⁡(24,k)\frac{24}{\gcd(24,k)}.

The construction for each kk is summarized in Table 2.

kkccImin⁡\mathcal{I}_{\min}constructiongauged
6724(12,3)3(12,3)^3, a Hadamard–Paley cellZ6\mathbb{Z}_6 shifts
784243×(12,3)+(6,3)3 \times(12,3) + (6,3)—
89634×(12,4)4 \times(12,4), Q8Q_8-equivariant cellsQ8Q_8
91088(18,5)3(18,5)^3, a Paley-type cell of width 18Z3\mathbb{Z}_3 shifts
1012012(12,4)5(12,4)^5, a Hadamard–Paley cellZ2=⟨−1⟩\mathbb{Z}_2 = \langle-1\rangle
11132245×(12,3)+(6,3)5 \times(12,3) + (6,3)—
121442Sym⁡2(T6)\operatorname{Sym}^2(T_6)—
13156246×(12,3)+(6,3)6 \times(12,3) + (6,3)—
14168127×(12,3)7 \times(12,3)Z2=⟨−1⟩\mathbb{Z}_2 = \langle-1\rangle
151808[2×(12,3)+(6,3)]3[2 \times(12,3) + (6,3)]^3Z3\mathbb{Z}_3 shifts
1619238×(12,4)8 \times(12,4), Q8Q_8-equivariant cellsQ8Q_8
1720424T8⊗T9T_8 \otimes T_9—
182164[3×(12,4)]3[3 \times(12,4)]^3Z6\mathbb{Z}_6 shifts
1922824T16⊗T3T_{16} \otimes T_3—
202406T8⊗T12T_8 \otimes T_{12}—
212528[3×(12,4)+(6,4)]3[3 \times(12,4) + (6,4)]^3Z3\mathbb{Z}_3 shifts
2226412T16⊗T6T_{16} \otimes T_6—
2327624T8⊗T15T_8 \otimes T_{15}—
242881Sym⁡2(−Sym⁡2(T6))\operatorname{Sym}^2(-\operatorname{Sym}^2(T_6)), or a twisted D4D_4 orbifold of T6⊗4T_6^{\otimes4}—
>24>24——Tk−24⊗T24T_{k-24} \otimes T_{24}—

Table 2. Construction of holomorphic SCFTs with minimal index ±Imin⁡\pm\mathcal{I}_{\min} for each kk (such a theory is denoted as TkT_k). Cell structures listed can differ from the benchmarked ones in Table 1 in several cases. Explicit generator matrices and verification records for every theory are available in [15]. The focus of this paper is the k=6k = 6 case (blue) and the path to k=12k = 12 (teal) and 24 (gold) via its symmetric orbifolds.

In the process, the AI agents found rich structures (e.g. conceptually simple seed theories) and established various properties (e.g. a proof that the Q8Q_8 action used at k=8,16k = 8,16 is anomaly-free) to guide their search, and the interested reader is invited to check out the summary written by the agents themselves [15]. It also links to a more agent-facing document with all the technical details.

In this paper, we will focus on describing one of the k=6k = 6 theories below and the route to obtain an index-1 theory at k=24k = 24 from it. We start with a brief review of the construction of SCFTs via ternary codes in § 2. We then describe the seed theory and a way to evaluate the index efficiently via a transfer-matrix technique in Section 3. After that, in § 4, we discuss the discrete symmetries used in our approach and their gauging to obtain the k=24k = 24 theory. In the two remaining sections, we remark on how results for other values of kk are obtained (§ 5), and on the implications of our results for the geography of 2d supersymmetric theories, especially in relation to the Segal–Stolz–Teichner conjecture (§ 6).

Ternary codes and SCFTs

A linear ternary code—in the sense of an error-correcting code—of length nn is a subspace C⊂F3nC \subset\mathbb{F}_3^n, where F3={0,+,−}\mathbb{F}_3 = \{0,+,-\} is the field of three elements. Elements in CC are called the “codewords,” and the dimension of CC as a linear space is referred to as the dimension of the code, dim⁡(C)\dim(C). Therefore, there are in total 3dim⁡(C)3^{\dim(C)} codewords.

One can choose mm linearly independent codewords to generate a code of dimension mm. These are then the row vectors of a generator matrix R∈F3m×nR \in\mathbb{F}_3^{m \times n}, which, after row operations (i.e. a change of basis of CC) and possibly a permutation of the coordinates, can be brought to the standard form R=(1m∣A)R = (1_m \mid A). The dual code is C⊥={y∈F3n∣x⋅y=0, ∀x∈C}C^\perp= \{y \in\mathbb{F}_3^n \mid x \cdot y = 0,\ \forall x \in C\} with respect to the standard bilinear form, and CC is self-dual if C=C⊥C = C^\perp. To have a self-dual code, nn has to be even with m=n/2m = n/2, and, in the standard form, self-duality is the identity for the m×mm \times m matrix AA,

AAT=−1m.(10)A A^T = -1_m. \tag*{(10)}

Taking the determinant gives det⁡(A)2=(−1)m\det(A)^2 = (-1)^m, and since −1-1 is not a square in F3\mathbb{F}_3, mm must be even, which requires that n=2mn = 2m is in fact a multiple of 4 for a self-dual code of length nn to exist.

The smallest self-dual ternary code is the tetracode of length 4, generated by (+0++0++−)\begin{pmatrix} + & 0 & + & + \\ 0 & + & + & - \end{pmatrix}, while perhaps the best-known one is the ternary Golay code of length 12, with generator matrix

R=(+00000++++++0+0000+0+−−+00+000++0+−−000+00+−+0+−0000+0+−−+0+00000+++−−+0).(11)R = \begin{pmatrix} + & 0 & 0 & 0 & 0 & 0 & + & + & + & + & + & + \\ 0 & + & 0 & 0 & 0 & 0 & + & 0 & + & - & - & + \\ 0 & 0 & + & 0 & 0 & 0 & + & + & 0 & + & - & - \\ 0 & 0 & 0 & + & 0 & 0 & + & - & + & 0 & + & - \\ 0 & 0 & 0 & 0 & + & 0 & + & - & - & + & 0 & + \\ 0 & 0 & 0 & 0 & 0 & + & + & + & - & - & + & 0 \end{pmatrix}. \tag*{(11)}

One can turn a self-dual ternary code into an odd unimodular lattice [16],

Λ(C)={13x:x∈Zn, x mod 3∈C}⊂Rn,(12)\Lambda(C) = \left\{ \frac{1}{\sqrt{3}}x : x \in\mathbb{Z}^n,\ x \bmod3 \in C \right\} \subset\mathbb{R}^n, \tag*{(12)}

and the lattice (super) vertex operator algebra (VOA) VΛ(C)V_{\Lambda(C)}—related to the usual bosonic lattice VOA except that vectors of odd norm now create fermionic states—is a holomorphic VOA of central charge c=nc = n. See [17, 18, 19, 20] for standard facts about lattice VOAs.

Furthermore, this VOA has an N=1\mathcal{N} = 1 supersymmetry and thus automatically defines an SCFT [4]: the 2n2n vectors ±3ei\pm\sqrt{3}e_i of norm 3 give a “3-frame” 3Zn⊂Λ(C)\sqrt{3}\mathbb{Z}^n \subset\Lambda(C) from which one can construct a weight-3/2 field

G=∑i(Γ3ei+Γ−3ei)(13)G = \sum_i \left(\Gamma_{\sqrt{3}e_i} + \Gamma_{-\sqrt{3}e_i}\right) \tag*{(13)}

generating an N=1\mathcal{N}=1 superconformal algebra. Thus every self-dual ternary code of length n=12kn=12k yields a holomorphic SCFT with c=12kc=12k, and a signed permutation of the coordinates (a monomial automorphism in coding-theory terminology, i.e. a monomial matrix preserving CC) [21] acts on the SCFT preserving GG.

The Ramond ground states of VΛ(C)V_{\Lambda(C)} are labeled by the codewords of full support, c=(c1,…,cn)c=(c_1,\ldots,c_n) with all ci≠0c_i\ne0, and such a ground state is bosonic or fermionic according to the parity of the number of coordinates equal to −-. The Witten index of the SCFT is then given by a combinatorial invariant of the code,

I(VΛ(C))=Index⁡(C):=∑c∈Call ci≠0(−1)#{i:ci=−}.(14)I(V_{\Lambda(C)})=\operatorname{Index}(C):=\sum_{\substack{c\in C\\\text{all }c_i\ne0}}(-1)^{\#\{i:c_i=-\}}. \tag*{(14)}

It was conjectured for all kk and proved for even kk [4] that Index⁡(C)\operatorname{Index}(C) is divisible by 24 for every self-dual code of length 12k12k. We will also prove the case of odd kk and show that index 24 is attained for every kk in § 5.

The k=6k=6 theory with index 4

We can now start describing the theory with k=6k=6 and index 4 from which we construct the k=24k=24 theory with index 1.

The seed code

The starting point (a “seed”) of our construction is a code of length 72 assembled from three copies of a 12-spin “cell.” The cell is constructed with two 12×1212\times12 matrices over F3\mathbb{F}_3, with rows and columns labeled 0,1,…,110,1,\ldots,11,

Q=(++++00000000++−−00000000+−+−00000000+−−+000000000000++++00000000++−−00000000+−+−00000000+−−+000000000000++++00000000++−−00000000+−+−00000000+−−+),S=(0+++++++++++−0+−+++−−−++−−0+−+++−−−+−+−0+−+++−−−−−+−0+−+++−−−+−+−0+−+++−−−+−+−0+−+++−+−++−+0+++−−++−−+−−0+−+−+++−−+−+0+−−−+++−−+−0++−+−+++−−+−0+).(15)Q=\begin{pmatrix} +&+&+&+&0&0&0&0&0&0&0&0\\ +&+&-&-&0&0&0&0&0&0&0&0\\ +&-&+&-&0&0&0&0&0&0&0&0\\ +&-&-&+&0&0&0&0&0&0&0&0\\ 0&0&0&0&+&+&+&+&0&0&0&0\\ 0&0&0&0&+&+&-&-&0&0&0&0\\ 0&0&0&0&+&-&+&-&0&0&0&0\\ 0&0&0&0&+&-&-&+&0&0&0&0\\ 0&0&0&0&0&0&0&0&+&+&+&+\\ 0&0&0&0&0&0&0&0&+&+&-&-\\ 0&0&0&0&0&0&0&0&+&-&+&-\\ 0&0&0&0&0&0&0&0&+&-&-&+ \end{pmatrix}, \qquad S=\begin{pmatrix} 0&+&+&+&+&+&+&+&+&+&+&+\\ -&0&+&-&+&+&+&-&-&-&+&+\\ -&-&0&+&-&+&+&+&-&-&-&+\\ -&+&-&0&+&-&+&+&+&-&-&-\\ -&-&+&-&0&+&-&+&+&+&-&-\\ -&+&-&+&-&0&+&-&+&+&+&-\\ -&-&+&-&+&-&0&+&-&+&+&+\\ -&+&-&+&+&-&+&0&+&+&+&-\\ -&+&+&-&-&+&-&-&0&+&-&+\\ -&+&+&+&-&-&+&-&+&0&+&-\\ -&-&+&+&+&-&-&+&-&0&+&+\\ -&+&-&+&+&+&-&-&+&-&0&+ \end{pmatrix}. \tag*{(15)}

There are three “boundary” spins {3,5,8}\{3,5,8\}, while the other nine are “internal.” The matrix Q=H4⊕H4⊕H4Q=H_4\oplus H_4\oplus H_4 consists of three copies of the 4×44\times4 Hadamard matrix on the diagonal blocks, and it is orthogonal over F3\mathbb{F}_3,

QQT=4⋅1=1.(16)QQ^T=4\cdot\mathbf{1}=\mathbf{1}. \tag*{(16)}

The matrix SS is the skew Paley conference matrix.4 It is anti-orthogonal,

SST=11⋅1=−1.(17)SS^T=11\cdot\mathbf{1}=-\mathbf{1}. \tag*{(17)}

There is one boundary spin in each Hadamard block. The columns of QQ and the rows of SS that belong to the boundary spins are shaded in blue.

From the cell we form the 12×1212\times12 matrices

D=QPinS,E=QPbdS,Pbd=diag⁡(0,0,0,1,0,1,0,0,1,0,0,0),Pin=1−Pbd.(18)D=QP_{\mathrm{in}}S,\quad E=QP_{\mathrm{bd}}S,\quad P_{\mathrm{bd}}=\operatorname{diag}(0,0,0,1,0,1,0,0,1,0,0,0),\quad P_{\mathrm{in}}=\mathbf{1}-P_{\mathrm{bd}}. \tag*{(18)}

so that D+E=QSD+E=QS, and the code CC is given by the generator matrix

R=(136∣A),withA=(DE00DE−E0D).(19)R=(\mathbf{1}_{36}\mid A),\quad\text{with}\quad A=\begin{pmatrix}D&E&0\\0&D&E\\-E&0&D\end{pmatrix}. \tag*{(19)}

The codewords are the row vectors a⋅R=(a∣a⋅A)a\cdot R=(a\mid a\cdot A) with a=(a1,a2,a3)a=(a_1,a_2,a_3), a1,a2,a3∈F312a_1,a_2,a_3\in\mathbb{F}_3^{12}, and

a⋅A=(a1D−a3E, a2D+a1E, a3D+a2E).(20)a\cdot A=(a_1D-a_3E,\ a_2D+a_1E,\ a_3D+a_2E). \tag*{(20)}

In other words, the block aαa_\alpha contributes aαDa_\alpha D to the α\alpha-th block and passes its three boundary spins vα:=aαQPbdv_\alpha:=a_\alpha QP_{\mathrm{bd}} on to the (α+1)(\alpha+1)-th one, through aαE=vα⋅Sa_\alpha E=v_\alpha\cdot S, cyclically but with a sign when wrapping around, forming an “anti-periodic spin chain.” Although we will mostly use statistical mechanics terminology due to our background, such structures are studied extensively in coding theory, where such a system is known as a trellis [22], here of “tail-biting” type [23].

What makes such a code desirable is that AA is “almost block-diagonal,” with the off-diagonal blocks given by the rank-3 matrix E=QPbdSE=QP_{\mathrm{bd}}S. This makes the cells “weakly coupled,” which greatly facilitates the computations as we will explain later.

Before that, we still need to check self-duality (2.1). This is a straightforward computation using QQT=1QQ^T=\mathbf{1}, SST=−1SS^T=-\mathbf{1} and Pin+Pbd=1P_{\mathrm{in}}+P_{\mathrm{bd}}=\mathbf{1}: the diagonal blocks of AATAA^T are

DDT+EET=Q(Pin+Pbd)QT⋅(−1)=−1,(21)DD^T+EE^T=Q(P_{\mathrm{in}}+P_{\mathrm{bd}})Q^T\cdot(-1)=-\mathbf{1}, \tag*{(21)}

while the off-diagonal ones vanish,

±DET=∓QPinPbdQT=0.(22)\pm D E^{T} = \mp Q P_{\mathrm{in}} P_{\mathrm{bd}} Q^{T} = 0. \tag*{(22)}

The sign in the last row of AA does not affect self-duality—both the periodic chain (without the sign) and the anti-periodic chain are self-dual, and both are useful for different values of kk. The k=6k=6 seed just happens to be anti-periodic.

Indices via transfer matrices

The fact that the code (3.5) has index −24-24 can be checked directly by brute force, by going through the 3363^{36} codewords (or the 236≈7×10102^{36} \approx7 \times10^{10} words whose first 36 components are non-zero). However, as we are also interested in much larger values of kk, a shortcut is needed, which is exactly what the block structure of (3.5) provides and what motivates the use of this very construction.

Write a codeword of full support as c=a⋅R=(a1,a2,a3∣b1,b2,b3)c = a \cdot R = (a_1, a_2, a_3 \mid b_1, b_2, b_3) with aα∈{+,−}12a_\alpha\in\{+,-\}^{12} and

bα=aα−1E+aαD=(vα−1+aαQPin)S,(23)b_\alpha= a_{\alpha-1} E + a_\alpha D = (v_{\alpha-1} + a_\alpha QP_{\mathrm{in}})S, \tag*{(23)}

where, as defined before,

vα−1:=aα−1QPbd,(24)v_{\alpha-1} := a_{\alpha-1} QP_{\mathrm{bd}}, \tag*{(24)}

supported on the three boundary coordinates, is the vector of boundary spins on the “interface” with block α−1\alpha-1 (with v0=−v3v_0 = -v_3 because of the anti-periodicity). The block bαb_\alpha depends on aαa_\alpha and the three spins vα−1v_{\alpha-1} only, factorizing the index computation along the chain. Define the 27×2727 \times27 “transfer matrix,” indexed by the 333^3 possible values of the boundary spins,

Mv,v′=∑a∈{+,−}12aQPbd=v′(−1)#{i:ai=−}σ((v+aQPin)S),(25)M_{v,v'} = \sum_{\substack{a \in\{+,-\}^{12} \\ aQP_{\mathrm{bd}}=v'}} (-1)^{\#\{i:a_i=-\}} \sigma\bigl((v+aQP_{\mathrm{in}})S\bigr), \tag*{(25)}

where aa plays the role of aαa_\alpha, vv that of vα−1v_{\alpha-1} and v′v' that of vαv_\alpha, and

σ(b):={(−1)#{i:bi=−}if all bi≠0,0otherwise.(26)\sigma(b) := \begin{cases} (-1)^{\#\{i:b_i=-\}} & \text{if all } b_i \ne0,\\ 0 & \text{otherwise.} \end{cases} \tag*{(26)}

is the “index” of a length-12 block bαb_\alpha, so that the summand is the contribution of the length-24 word (aα,bα)(a_\alpha,b_\alpha) of one cell.5

The transfer matrix Mv,v′M_{v,v'} records, for an incoming state vv and an outgoing state v′v', the signed number of local configurations that keep all 24 coordinates of the block non-zero. Then

Index⁡(C)=Tr⁡(JM3),(27)\operatorname{Index}(C)=\operatorname{Tr}(JM^3), \tag*{(27)}

where Jv,v′=δv′,−vJ_{v,v'}=\delta_{v',-v} implements the anti-periodic boundary condition. Each entry of MM is a sum over the 2122^{12} local words a∈{+,−}12a\in\{+,-\}^{12} with full support, so the computation involves 27⋅212≈10527\cdot2^{12}\approx10^5 sign evaluations instead of 2362^{36}—a rather dramatic speedup.

In coding theory, this is the trellis evaluation of a weight enumerator [24, 25], in its tail-biting form [26]. Here we have adapted it to the signed count of full-support codewords.

Furthermore, once MM is known, the same cell repeated LL times gives a self-dual code of length 24L24L with index Tr⁡(JML)\operatorname{Tr}(JM^L) (or Tr⁡(ML)\operatorname{Tr}(M^L) for periodic chains) for every LL. This is what overcomes the exponential-time barrier and makes the search for large kk possible.

For our seed theory, MM has 25 non-zero eigenvalues and

Tr⁡(JM)=0,Tr⁡(JM2)=−240,Tr⁡(JM3)=−24.(28)\operatorname{Tr}(JM)=0,\qquad\operatorname{Tr}(JM^2)=-240,\qquad\operatorname{Tr}(JM^3)=-24. \tag*{(28)}

The numbers on the right of these equations are deceptively small and hide dramatic cancellations between bosonic and fermionic states. In fact, dropping the signs in (25) counts all Ramond ground states instead, and there are 83,000 of them, with 41,488 being bosonic and 41,512 fermionic.

The last equation means that the code is of index −24-24 and hence a legitimate candidate to be considered for discrete gauging. On the other hand, the first equality, Tr⁡(JM)=0\operatorname{Tr}(JM)=0, turns out to be a condition that makes it easier to control the index after orbifolding, and is not guaranteed to hold for arbitrary cells. In fact, the computer search screens candidates for exactly these two conditions: first the cheaper Tr⁡(JM)=0\operatorname{Tr}(JM)=0, then Tr⁡(JM3)=±24\operatorname{Tr}(JM^3)=\pm24 (or without the JJ for periodic chains).

Symmetry and orbifolding

The code (19) has an obvious symmetry—the shift along the anti-periodic chain

g:(a1,a2,a3∣b1,b2,b3)⟼(−a3,a1,a2∣−b3,b1,b2),(29)g:(a_1,a_2,a_3\mid b_1,b_2,b_3)\longmapsto(-a_3,a_1,a_2\mid-b_3,b_1,b_2), \tag*{(29)}

which is a signed permutation of the 72 coordinates with g3=−1g^3=-1, hence of order 6. It acts on VΛ(C)V_{\Lambda(C)} preserving the N=1\mathcal{N}=1 structure, and we want to gauge it.

In such a setting, i.e. CC a self-dual code of length 12k12k with index ±24\pm24 and gg a signed permutation of order m=gcd⁡(k,24)m=\gcd(k,24), one can gauge the Zm\mathbb{Z}_m symmetry of VΛ(C)V_{\Lambda(C)} whenever the two conditions are satisfied [4 Thm. 5.6]: (i) if mm is even, gm/2=−1g^{m/2}=-1;

(ii) if 3∣m3\mid m, then gm/3g^{m/3} permutes the coordinates in 4k4k disjoint three-cycles and the codewords of full support fixed by gm/3g^{m/3} are “balanced,” i.e. with equal numbers of invariant bosonic and fermionic states.

Then the Zm\mathbb{Z}_m action is non-anomalous and the orbifold VΛ(C)/ZmV_{\Lambda(C)}/\mathbb{Z}_m is a holomorphic N=1\mathcal{N}=1 SCFT with index ±24/m\pm24/m.

Here k=m=6k=m=6 and both hypotheses can be checked by inspection. For (i), g3=−1g^3=-1 for our shift symmetry by construction. As for the first part of (ii), g2g^2 shifts by two blocks, (a1,a2,a3)↦(−a2,−a3,a1)(a_1,a_2,a_3)\mapsto(-a_2,-a_3,a_1), so it permutes the 72 coordinates in 24 three-cycles.

Codewords fixed by g2g^2 are those with a2=−a1a_2=-a_1, a3=a1a_3=a_1 (and likewise for bb)—the words that are constant along the chain up to an alternating sign. They are determined by a single cell closed on itself, which forms a length-24 self-dual code, with its signed count of full-support words being exactly the trace Tr⁡(JM)=0\operatorname{Tr}(JM)=0 seen previously. For our seed code above, the 80 such words split evenly into 40 bosonic and 40 fermionic ones.

The orbifold

T:=VΛ(C)/Z6(30)T:=V_{\Lambda(C)}/\mathbb{Z}_6 \tag*{(30)}

then has index

I(T)=−246=−4,(31)I(T)=\frac{-24}{6}=-4, \tag*{(31)}

saturating the TMF bound for k=6k=6. It can therefore serve as a seed theory for our construction. Also notice that a theory obtained by gauging such a faithful anomaly-free discrete symmetry in our setting always has a unique NS vacuum—in this example, every twisted sector for the Z6\mathbb{Z}_6 symmetry in fact has ground-state energy at least 8/38/3.

The k=24k=24 theory

We are now ready to construct the k=24k=24 theory with index ±1\pm1 from the seed theory above via symmetric orbifolding.

The double symmetric square

In general, the permutation symmetry SNS_N of NN copies of a theory can have an ’t Hooft anomaly which would then obstruct its gauging. However, in 2d, the anomaly for the S2=Z2S_2=\mathbb{Z}_2 exchange symmetry only depends on the chiral central charge (or more generally on the gravitational anomaly of the 2d physical theory6 and is given by d mod 8=2c mod 8∈Z8d \bmod8 = 2c \bmod8 \in\mathbb{Z}_8 [27, 8], which vanishes when 4∣c4 \mid c—all cases considered in this work satisfy this—making both symmetric squares in (8) well defined.

In general, there are two choices of a “discrete theta angle” (or a fermionic version of the more familiar “discrete torsion” [28] in 2d) when gauging this Z2\mathbb{Z}_2, with the non-trivial one given by first stacking with the invertible Z2\mathbb{Z}_2-SPT phase Arf⁡⊗Arf⁡\operatorname{Arf} \otimes\operatorname{Arf}. Notice that this would be a trivial phase if one forgets the Z2\mathbb{Z}_2 symmetry but a non-trivial one as a Z2\mathbb{Z}_2-symmetric phase (so genuinely an SPT phase). From the boundary point of view, the exchange acts on the two Majorana zero modes ψ1,ψ2\psi_1,\psi_2 of Arf⁡⊗Arf⁡\operatorname{Arf} \otimes\operatorname{Arf} and maps the only possible mass term iψ1ψ2i\psi_1\psi_2 to iψ2ψ1i\psi_2\psi_1. However, as the two fields anti-commute, any Z2\mathbb{Z}_2-invariant mass term would have to vanish as imψ1ψ2=imψ2ψ1=−imψ1ψ2im\psi_1\psi_2 = im\psi_2\psi_1 = -im\psi_1\psi_2.

When gauging the Z2\mathbb{Z}_2 exchange symmetry, this topological phase would flip the sign of the twisted-sector contribution and disallow symmetric states like ∣a⟩⊗∣a⟩|a\rangle\otimes|a\rangle for bosonic ∣a⟩|a\rangle in the untwisted sector, and therefore the index becomes

I⟼I(I+3)2andI⟼I(I−3)2(32)\mathcal{I} \longmapsto\frac{\mathcal{I}(\mathcal{I}+3)}{2} \qquad\text{and}\qquad \mathcal{I} \longmapsto\frac{\mathcal{I}(\mathcal{I}-3)}{2} \tag*{(32)}

for the two choices respectively. The first is denoted by Sym⁡2(T)\operatorname{Sym}^2(T) and the second by Sym⁡2(T⊗Arf⁡)=Sym⁡2(−T)\operatorname{Sym}^2(T \otimes\operatorname{Arf}) = \operatorname{Sym}^2(-T), as we did in the introduction. Starting from our seed theory with index −4-4, to get the chain −4↦2↦−1-4 \mapsto2 \mapsto-1, the second symmetric square is the twisted one. If one starts with an index-4 theory, both steps would need to be the twisted symmetric square.

D4 orbifold

One can also attempt to combine the two steps into a single one, by starting with four copies T⊗4T^{\otimes4} and gauging a subgroup of S4S_4, instead of first gauging the “inner swaps” Z2×Z2\mathbb{Z}_2 \times\mathbb{Z}_2 and then the “outer” Z2\mathbb{Z}_2.

Notice that the two steps in the double symmetric square are now quite different. The inner Z2×Z2\mathbb{Z}_2 \times\mathbb{Z}_2 is an honest subgroup of S4S_4 generated by the transpositions (13)(13) and (24)(24), while the generator of the “outer” Z2\mathbb{Z}_2 has various lifts: it can be (12)(34)(12)(34), (14)(23)(14)(23), (1234)(1234), or (1432)(1432). These are the four lifts of the generator of the quotient in the extension,

1⟶Z2×Z2⟶D4⟶Z2⟶1.(33)1 \longrightarrow\mathbb{Z}_2 \times\mathbb{Z}_2 \longrightarrow D_4 \longrightarrow\mathbb{Z}_2 \longrightarrow1. \tag*{(33)}

The two lifts of order two, (12)(34)(12)(34) and (14)(23)(14)(23), split this extension, while the two rotations have order four. Together with the inner swaps, any of the lifts generates the same D4D_4 subgroup of S4S_4. Alternatively, it can be viewed as the wreath product Z2≀Z2=D4⊂S4\mathbb{Z}_2 \wr\mathbb{Z}_2 = D_4 \subset S_4.

Gauging a normal subgroup and then the quotient is the same as gauging the extension. Therefore, to do it in one step, one should gauge the D4D_4 symmetry on T⊗4T^{\otimes4}. This can be visualized as the symmetry group of a square whose corners carry the four copies of TT, as illustrated in Figure 1. The normal subgroup Z2×Z2\mathbb{Z}_2 \times\mathbb{Z}_2 is generated by the two diagonal reflections (13)(13) and (24)(24), whose product is the inversion (or π\pi-rotation), while the outer swaps are the edge reflections (12)(34)(12)(34) and (14)(23)(14)(23), with the ±π2\pm\frac{\pi}{2}-rotations also in the same coset.

The twist associated with the outer Z2\mathbb{Z}_2 now becomes a twist of the D4D_4 gauging via pullback along the quotient map, given by the character

χ:D4⟶D4/(Z2×Z2)≃Z2,(34)\chi: D_4 \longrightarrow D_4/(\mathbb{Z}_2 \times\mathbb{Z}_2) \simeq\mathbb{Z}_2, \tag*{(34)}

which is −1-1 exactly on the non-trivial coset.

Recall that invertible fermionic phases with GG symmetry in 2d are classified by [29, 30]

fSPT1+1(G)≃Ω2Spin(BG)∨≃Z2×H1(G,Z2)×H2(G,U(1)),(35)\mathrm{fSPT}_{1+1}(G) \simeq\Omega^{\mathrm{Spin}}_2(BG)^\vee\simeq\mathbb{Z}_2 \times H^1(G,\mathbb{Z}_2) \times H^2(G,U(1)), \tag*{(35)}

where the second identification holds as sets (under stacking, the three layers can form a non-trivial extension), and χ∈Hom⁡(D4,Z2)≃H1(D4,Z2)\chi\in\operatorname{Hom}(D_4,\mathbb{Z}_2) \simeq H^1(D_4,\mathbb{Z}_2) specifies the middle factor for G=D4G = D_4.

The other generator for the middle factor H1(D4,Z2)≃Z22H^1(D_4,\mathbb{Z}_2) \simeq\mathbb{Z}_2^2 can be taken to be the sign representation of D4⊂S4D_4 \subset S_4. Stacking a copy of Arf on each of the four TT factors gives exactly this character.

The first factor in (35) is the familiar Arf theory—often excluded as an SPT since it does not require the GG-symmetry. It can be used to flip the sign of the index of the resulting theory. We will need it at the end to go from −1-1 to +1+1.

These three copies of Z2\mathbb{Z}_2 correspond to the choices in the double symmetric orbifold—one can stack an Arf theory at each step on TT, T′T' or T′′T''. In other words, these are the three sign choices in

T′′=±Sym⁡2(±Sym⁡2(±T)).(36)T'' = \pm\operatorname{Sym}^2(\pm\operatorname{Sym}^2(\pm T)). \tag*{(36)}

On the other hand, we do have the extra freedom to modify the discrete torsion η∈H2(D4,U(1))≃Z2\eta\in H^2(D_4,U(1)) \simeq\mathbb{Z}_2, corresponding to some of the additional choices when gauging D4D_4 as a whole that the double symmetric orbifold cannot access.7 Concretely, (a representative of) the class η\eta is a 2-cocycle

η:D4×D4→U(1),(37)\eta: D_{4} \times D_{4} \to\mathrm{U}(1), \tag*{(37)}

and on a torus with a flat D4D_{4}-connection, characterized by a pair of commuting holonomies (g,h)(g,h), it contributes a phase η(g,h)/η(h,g)\eta(g,h)/\eta(h,g) to the partition function of the SPT.

Therefore, there are in total eight twists, obtained by first stacking T⊗4T^{\otimes4} with an invertible fermionic phase labeled by elements in H1(D4,Z2)×H2(D4,U(1))≃Z23H^{1}(D_{4},\mathbb{Z}_{2}) \times H^{2}(D_{4},\mathrm{U}(1)) \simeq\mathbb{Z}_{2}^{3} and then gauging.

In general, the index of a permutation orbifold T⊗n/GT^{\otimes n}/G is

I(T⊗n/G)=1∣G∣∑gh=hgε(g,h) I#orbits of ⟨g,h⟩,(38)\mathcal{I}(T^{\otimes n}/G) = \frac{1}{|G|} \sum_{\substack{g h = h g}} \varepsilon(g,h)\,\mathcal{I}^{\#\text{orbits of }\langle g,h\rangle}, \tag*{(38)}

a sum over commuting pairs, i.e. over GG-bundles on T2T^{2} with the non-bounding spin structure, in which the sign ε(g,h)\varepsilon(g,h) is the partition function of the invertible phase on that torus. In our case, we have

ε(g,h)=(−1)q(χ(g,h))νχ(g,h)η(g,h)η(h,g).(39)\varepsilon(g,h) = (-1)^{q(\chi(g,h))}\nu_{\chi}(g,h)\frac{\eta(g,h)}{\eta(h,g)}. \tag*{(39)}

The first factor is given by the quadratic form qq on H1(T2,Z2)H^{1}(T^{2},\mathbb{Z}_{2}) defined by the spin structure, evaluated at the image of χ\chi on (g,h)(g,h). It can also be viewed as the difference between the Arf invariants of the untwisted spin structure and the one twisted by the D4D_{4} holonomy via the Z2\mathbb{Z}_{2}-valued character χ\chi. The phase νχ\nu_{\chi} is a choice of convention for the origin of the space of discrete torsions.8

Labeling a twist by (χ((13)),χ((12)(34)),η)(\chi((13)),\chi((12)(34)),\eta), with ++ for the trivial and −- for the non-trivial value, the eight twists give at I=−4\mathcal{I}=-4

twist(+,+,+)(+,+,+)(+,+,−)(+,+,-)(+,−,+)(+,-,+)(+,−,−)(+,-,-)(−,+,+)(-,+,+)(−,+,−)(-,+,-)(−,−,+)(-,-,+)(−,−,−)(-,-,-)
I\mathcal{I}55−13-13−1-1−31-31119119898977775959

Table 1 (PDF p. 17).

(see also [15]). The four twists with trivial η\eta, (±,±,+)(\pm,\pm,+), are the double symmetric orbifolds Sym⁡2(±Sym⁡2(±T))\operatorname{Sym}^{2}(\pm\operatorname{Sym}^{2}(\pm T)), with indices 55, −1-1, 119119 and 7777. One of them indeed gives ∣I(T′′)∣=1|\mathcal{I}(T^{\prime\prime})|=1,

Z2×Z2\mathbb{Z}_{2}\times\mathbb{Z}_{2} is the familiar discrete torsion giving the phase (−1)⟨a,b⟩(-1)^{\langle a,b\rangle} on a torus with holonomies a,b∈H1(T2,Z2)a,b\in H^{1}(T^{2},\mathbb{Z}_{2}) for the two Z2\mathbb{Z}_{2} factors paired via the intersection form. On a torus with D4D_{4} holonomy (g,h)(g,h), it gives −1-1 exactly when g≠hg\ne h are both non-trivial and generate a Klein four-group, and it is trivial on the rotation subgroup Z4\mathbb{Z}_{4}.

coming from the twist (+,−,+)(+,-,+), for which ε(g,h)=−1\varepsilon(g,h)=-1 exactly when the pair (g,h)(g,h) has a non-trivial image in D4/(Z2×Z2)D_4/(Z_2 \times Z_2).

This is precisely the twisted double symmetric square, and its index is

I(T⊗4/ ⁣/ ⁣εD4)=I4+6I3+3I2−18I8,(40)\mathcal{I}(T^{\otimes4}/\!/\!_{\varepsilon}D_4)=\frac{\mathcal{I}^4+6\mathcal{I}^3+3\mathcal{I}^2-18\mathcal{I}}{8}, \tag*{(40)}

which evaluates to −1-1 at I=−4\mathcal{I}=-4. Grouping the commuting pairs by the conjugacy class of gg gives the five sectors of Figure 1: the untwisted sector, the inner swaps, the double inner swap (π\pi-rotation), the outer swaps, and the π2\frac{\pi}{2}-rotations respectively contribute

3(I+24),I⋅(I+12),(I2),−(I+12),−I.(41)3\binom{\mathcal{I}+2}{4},\qquad\mathcal{I}\cdot\binom{\mathcal{I}+1}{2},\qquad\binom{\mathcal{I}}{2},\qquad-\binom{\mathcal{I}+1}{2},\qquad-\mathcal{I}. \tag*{(41)}

These can be obtained via a short computation, again pretending TT is the sigma model to I\mathcal{I} points. One does need to note that in the sectors with χ(g)=+1\chi(g)=+1, the wavefunction should be antisymmetric under any h∈D4h\in D_4 with χ(h)=−1\chi(h)=-1, while in the sectors with χ(g)=−1\chi(g)=-1 the states have reversed fermionic parity—the very reason for the minus sign in front of the last two cases—but are symmetric under any hh.

States and fields

The theories TT, T′T' and T′′T'' involved in the double symmetric orbifold of a k=6k=6 seed theory are explicit enough for their spectra to be computed exactly [15].

None of them has fields of weight 12\frac{1}{2} or 11: the seed code has minimum distance 9, so Λ(C)\Lambda(C) has no vectors of norm 1 or 2, and the 72 currents ∂Xi\partial X_i are removed by orbifolding, since gg has no eigenvalue 1. Each has 113 fields of weight 32\frac{3}{2} (among them the supercurrent GG), 2,928 of weight 2 and 42,001 of weight 52\frac{5}{2}. Below weight 3 they are all in the untwisted lattice SVOA sector, while the first states from twisted sectors appear at weight 3.

The NS character is a polynomial in the Hauptmodul

t=(ϑ3η)12=q−1/2+24+276q1/2+2048q+⋯(42)t=\left(\frac{\vartheta_3}{\eta}\right)^{12}=q^{-1/2}+24+276q^{1/2}+2048q+\cdots \tag*{(42)}

of Γθ⊂SL(2,Z)\Gamma_\theta\subset\mathrm{SL}(2,\mathbb{Z}), the index-3 subgroup generated by τ↦τ+2\tau\mapsto\tau+2 and τ↦−1/τ\tau\mapsto-1/\tau stabilizing the NS-NS spin structure. The polynomial has degree c/12c/12, and for TT,

ZTNS=t6−144t5+6984t4−129679t3+751404t2−629867t+13940,(43)Z_T^{\mathrm{NS}}=t^6-144t^5+6984t^4-129679t^3+751404t^2-629867t+13940, \tag*{(43)}

with its constant term being the number of ground states in the Ramond sector. These are far from unique, with about 5×10155\times10^{15} of them in the index-1 theory:

theoryccI\mathcal{I}bosonic Ramond ground statesfermionic Ramond ground states
TT72−4-46,9686,972
T′T'144248,587,87048,587,868
T′′T''288−1-12,360,781,062,549,0292,360,781,062,549,030

Such huge ground-state degeneracies, set against an index of ±1\pm1, ±2\pm2 or ±4\pm4, are typical of the theories constructed in our approach. The small index comes from near-perfect cancellations between bosonic and fermionic ground states, which gives a sense of how deep in the ocean of holomorphic VOAs the theories with minimal index lie.

The same computation can be carried out for every seed code. For k=6k = 6, ZNSZ^{\mathrm{NS}} is again a polynomial of degree 6 in tt with leading terms t6−144 t5t^6 - 144\,t^5, while the remaining coefficients can be used to distinguish different theories. Among the 8,748 seed codes of Table 1, these polynomials take 4,579 different values, giving rise to 4,579 pairwise inequivalent theories TT, and as many T′T' and T′′T''.9 The polynomial of every seed is listed in [15]. One can attempt to tell them apart further by finer invariants, such as the Lie algebra structure on the weight-1 fields.

Other values of kk

The search for minimal-index theories for other values of kk follows a similar strategy, with the specific constructions listed in Table 2.

The yield decreases as kk becomes larger, mainly because the index-24 condition on the product of the cell transfer matrices becomes harder to satisfy as the distribution of the index becomes wider. As an example, at k=21k = 21 (chains of four cells with four boundary spins, repeated three times, 2×1042 \times10^4 samples), the index of a random candidate has a median absolute value about 108.610^{8.6} and mean absolute value about 1011.610^{11.6}. Notice that these are still small when compared with the ground-state degeneracies above.

There are three main ingredients that, combined with a collection of observed structures, have greatly helped facilitate the search:

Tensor products. This provides a shortcut to reduce the problem of finding minimal-index theories with k=17,19,20,22,23k = 17, 19, 20, 22, 23 and all k>24k > 24 to the rest of the values of kk. It is rather convenient, as the search with these larger kk values can be quite time-consuming. One might wonder whether there are other useful symmetric orbifolds, but unfortunately the two that we used to go from k=6k = 6 to k=12k = 12 and then to k=24k = 24 seem to be the only ones. The AI agents searched all relevant choices of subgroups of SNS_N using GAP [31], with various twists, applied repeatedly together with tensor products, but found no other way to reach the minimal indices for k≤24k \le24.

Chains of cells with varying sizes. For k=7,9,10,11,13,14,15,18,21k = 7, 9, 10, 11, 13, 14, 15, 18, 21, an index-24 code of length $12k$ was found as a closed chain of cells of width 6, 12 or 18 in the sense of (19), with the same or different cells in different positions but a common number of boundary spins, so that the index is a trace of a product of transfer matrices. When the TMF bound requires a symmetry of order m=3m = 3 or 6 to saturate, the chain is built from a block of cells repeated three times and the balance condition is again a trace, while, for m=2m = 2, one simply gauges g=−1g = -1. At k=18k = 18 and k=21k = 21 the hits came from cells with four boundary spins (81-state transfer matrices), after the three-spin families had produced more than 10810^8 candidates without success. The search samples tuples of cells at random from pools of cells, with the traces computed in exact integer arithmetic (a quick screen modulo two primes, then the exact trace for the few survivors). Although hits are rare, at about one in 10210^2 candidates at k=6,7k = 6, 7, one in 10310^3–10510^5 for k=9,…,14k = 9, \ldots, 14, one in 10610^6 at k=15k = 15, and one in 10710^7–10810^8 for $k = 16, 18, 21, each candidate costs only milliseconds to evaluate.

Quaternionic symmetry. In our approach, for k=8k = 8 and k=16k = 16, saturating the bound by a code orbifold requires a group of order 8, and no cyclic group was found: order-8 shifts are excluded by a 2-adic theorem mentioned below, and every code with a signed-permutation symmetry gg of order 8 with g4=−1g^4 = -1 that we examined has index divisible by 48 rather than 24. Instead, we use codes of length 96 and 192 with a quaternion group Q8Q_8 symmetry. Since every pair of commuting elements of Q8Q_8 generates a cyclic group and every non-identity element has −1-1 as a power, the argument of [4 Thm. 5.6] gives I=24/8=3I = 24/8 = 3 as soon as the Q8Q_8 action is non-anomalous. For this construction to work, one does need to check that the action is anomaly-free, which is a non-trivial step, as part of the anomaly is invisible to genus-one data [32]. The proof, available in [15], connects the code at k=8k = 8 to the sum of 24 tetracodes, through a chain of 23 Q8Q_8-equivariant neighboring codes. Each step is an orbifold by a Z32\mathbb{Z}_3^2 that preserves the Q8Q_8 anomaly, and at the end one uses that the anomaly group Hom⁡(Ω3Spin(BQ8),U(1))≃Z8⊕Z42\operatorname{Hom}(\Omega_3^{\mathrm{Spin}}(BQ_8), \mathrm{U}(1)) \simeq\mathbb{Z}_8 \oplus\mathbb{Z}_4^2 has exponent 8 [33], which divides 24—the number of free-fermion blocks. At k=16k = 16, a similar argument applies.

Other structures in ternary codes. Several observations made along the way are documented in [15], including:

  • a 2-adic theorem, 2v2(L)+2∣Tr⁡(JML)2^{v_2(L)+2} \mid\operatorname{Tr}(JM^L) for L≥4L \ge4, which rules out shift symmetries of

order 8 and 24 in these chains, the reason why non-cyclic groups are used at k=8,16,24k = 8, 16, 24;

  • an explicit derivation of the index of the nine-state cells, and a “selection rule” (an empirical criterion to get a seed code) for the positions of the boundary spins.

Finally, two results settle Conjecture 1.4 of [4].

Theorem 2 (divisibility). The index of every self-dual ternary code of positive length is divisible by 24.

Sketch of proof. Since the index of such a CC is divisible by 12 [4 Thm. 1.3] and is 0 unless the length nn is a multiple of 12, it suffices to show that it is also divisible by 8 when 12∣n12 \mid n. If two self-dual codes CC and C′C' meet in a (non-self-dual) subcode DD of codimension one, a computation of the index of D⊥D^\perp via Poisson summation gives (details in [34 Sec. 2])

Index⁡(D⊥)=3⋅Index⁡(D).\operatorname{Index}(D^\perp) = 3 \cdot\operatorname{Index}(D).

On the other hand, the vectors in D⊥D^\perp outside C∪C′C \cup C' have non-zero (squared) norm and cannot contribute to the index, as every vector of full support has norm 0 (mod 3). This gives

Index⁡(D⊥)=Index⁡(C)+Index⁡(C′)−Index⁡(D),\operatorname{Index}(D^\perp) = \operatorname{Index}(C) + \operatorname{Index}(C') - \operatorname{Index}(D),

leading to

Index⁡(C)+Index⁡(C′)=4Index⁡(D).\operatorname{Index}(C) + \operatorname{Index}(C') = 4\operatorname{Index}(D).

The full-support words ±x\pm x of DD pair up with equal signs, so the index of DD is even and

Index⁡(C′)≡−Index⁡(C)(mod8).\operatorname{Index}(C') \equiv-\operatorname{Index}(C) \pmod{8}.

Any two self-dual codes of the same length are joined by a chain of such neighbors [35], and a sum of tetracodes has index 0, which proves the claim. □

Theorem 3 (existence). For every k≥1k \ge1, there is a self-dual ternary code of length 12k12k with index 24.

Sketch of proof. For each kk, one can build a chain out of kk cells each being a copy of the ternary Golay code Go12\mathrm{Go}_{12}. With a suitable choice of gluing data, this code CkC_k turns out to be self-dual with

Index⁡(Ck)=(−1)k−1⋅Index⁡(Go12)=(−1)k⋅24,\operatorname{Index}(C_k) = (-1)^{k-1} \cdot\operatorname{Index}(\mathrm{Go}_{12}) = (-1)^k \cdot24,

and for odd kk negating one coordinate gives +24+24. The explicit construction of CkC_k and a complete proof are in [34 Sec. 3]. □

Notice that these codes directly give the minimal-index theory when gcd⁡(k,24)=1\gcd(k,24)=1, and for gcd⁡(k,24)=2\gcd(k,24)=2 their orbifold by Z2=⟨−1⟩\mathbb{Z}_2=\langle-1\rangle has index 12. But they do not cover the “harder” cases of k=8k=8 and 16 (for which we used a Q8Q_8 symmetry) or 18 and 21 (the two largest values of kk below 24 not reached by tensor products).

The landscape of 2d theories

So far, we have found a collection of nice SCFTs, but they are still only a handful of landmarks in the vast landscape of 2d supersymmetric theories. What is their significance?

Representing TMF classes

A holomorphic SCFT with k=24k=24 and index 1 is a “periodic element” in the sense that, under the Segal–Stolz–Teichner conjecture, it maps to Δ24∈π576(TMF)\Delta^{24}\in\pi_{576}(\mathrm{TMF}), which gives the 576-periodicity of TMF,

TMF[576]→Δ24TMF.(44)\mathrm{TMF}[576]\xrightarrow{\Delta^{24}}\mathrm{TMF}. \tag*{(44)}

This is an equivalence with the inverse given by Δ−24\Delta^{-24}, which can be represented by an anti-holomorphic theory obtained by simply putting the same theory in the left-moving sector without any right-moving degrees of freedom (cf. the approach in [6]).10

Using this trick, we have

Theorem 4. For every integer kk, there is a holomorphically factorizable SCFT with cR−cL=12kc_R-c_L=12k and index 24gcd⁡(24,k)\frac{24}{\gcd(24,k)}.

Here, stating it as “holomorphically factorizable” is a bit roundabout as one can always choose holomorphic representatives for k>0k>0 and anti-holomorphic ones for k<0k<0 by putting those in the left-moving sector, and, for k=0k=0, the trivial theory is in a sense both holomorphic and anti-holomorphic.

Another remark is that our construction via the double symmetric square is natural, as it is compatible with the E∞E_\infty-ring structure of TMF, where one also expects the very chain −4Δ6↦2Δ12↦−Δ24-4\Delta^{6} \mapsto2\Delta^{12} \mapsto-\Delta^{24} via power operations. Details of this are spelled out in [15 Sec. I], with assumptions on the existence and properties of certain equivariant Thom classes in equivariant elliptic cohomology.

Furthermore, as π∗(MString)→π∗(tmf)\pi_{*}(\mathrm{MString}) \to\pi_{*}(\mathrm{tmf}) is surjective [38], and TMF≃tmf[Δ−24]\mathrm{TMF} \simeq\mathrm{tmf}[\Delta^{-24}], we can find a 2d (0,1)(0,1) theory that conjecturally represents any class in π∗(TMF)\pi_{*}(\mathrm{TMF}) by stacking the anti-holomorphic theories above with a 2d (0,1)(0,1) sigma model with a string manifold target. When the target is connected (which is always possible to arrange), it is reasonable to expect that, for a suitable choice of metric, the ground state in the NS sector is unique, and that, under the RG flow, the theory becomes an SCFT also with unique vacua (provided that the class is non-zero). Therefore, every non-zero TMF class is then expected to be represented by an indecomposable SCFT.

Paths to periodicity

However, there are often multiple seemingly distinct realizations of the same TMF class, and it is a highly non-trivial question whether they are the same and how to actually deform one into another.

In particular, the periodicity of TMF concerns the equation Δ24⋅Δ−24=1\Delta^{24} \cdot\Delta^{-24} = 1, and a natural question is why this is true on the physics side—why a c=288c = 288 N=1\mathcal{N} = 1 holomorphic SCFT PP with index 11 or its anti-holomorphic cousin P‾\overline{P} should be invertible.

If one takes P‾⊗P\overline{P}\otimes P, the non-chiral double of an index-1 holomorphic N=1\mathcal{N} = 1 VOA we found, one obtains a 2d N=(1,1)\mathcal{N} = (1,1) SCFT, and it seems far from obvious why this theory can be gapped in a way that preserves half of its supersymmetry. Indeed, such holomorphically factorizable SCFTs have no supersymmetric relevant operators, and a supersymmetric marginal operator would force the index to vanish.11 Therefore, the theories that we want to gap are all SUSY-rigid and are already the endpoints of supersymmetric RG flows, and any supersymmetric deformations of them to the trivial theory must be large and non-perturbative.

One approach is to first go “upward” to find a UV theory that can flow back to it and then, from that UV theory, deform to the trivial theory by gapping it (cf. the approach in [39] when studying theories representing the 0 class in TMF). This is in sharp contrast with the situation in 1d, where the 8-fold periodicity of KO-theory is associated with a quartic interaction for eight Majorana fermions, which is still a relevant deformation in one dimension [40, 41].

This observation seems to tell us that, to solve the periodicity problem, the natural first step is to find a suitable UV description for P‾⊗P\overline{P} \otimes P. One very natural hope is to find a 2d (1,1)(1,1) sigma model that flows to P‾⊗P\overline{P} \otimes P. Then, on the target, one can turn on a Morse function as a superpotential to gap the theory.

Alternatively, there may be a better partner P′P' for P‾\overline{P}, such that P‾⊗P′\overline{P} \otimes P' is easier to work with and can be the infrared limit of a 2d (0,1)(0,1) sigma model with a left-moving fermion bundle on the target. These theories can also be deformed by superpotentials, but it is harder to control phase transitions, which lead to wall-crossing in the parameter space (cf. [42, 43]).

However, exploring these directly for any PP with such a large central charge seems rather difficult. For example, one should expect that the target of the sigma model in the second scenario is at least 192-dimensional, since the central charge, equaling 32\frac{3}{2} times the dimension in the UV, decreases along RG flows.

Our construction might provide a shortcut to achieve this by allowing one to work with simpler theories with c=72c = 72, via the factorization

P‾⊗P=(V‾⊗V)⊗4/G,(45)\overline{P} \otimes P = (\overline{V} \otimes V)^{\otimes4}/G, \tag*{(45)}

with VV now a rather simple lattice SVOA but at the expense of having a very large G=GL×GRG = G_L \times G_R where each half is the wreath product Z6≀D4=Z64⋊D4\mathbb{Z}_6 \wr D_4 = \mathbb{Z}_6^4 \rtimes D_4 of order 64⋅8=10,3686^4 \cdot8 = 10,368 (with the twist described in § 4 when orbifolding).

Therefore, the problem becomes whether one can gap (V‾⊗V)⊗4(\overline{V} \otimes V)^{\otimes4} in a GG-equivariant way that preserves supersymmetry, a question closely related to what is known as symmetric mass generation (see [44] for a recent review).

We will only make a baby step toward this goal, by arguing that, without worrying about preserving GG, (V‾⊗V)⊗4(\overline{V} \otimes V)^{\otimes4} can be gapped. In fact, even a single copy can be gapped. In other words, we are showing “(±24Δ6)⋅(±24Δ−6)=576(\pm24\Delta^6) \cdot(\pm24\Delta^{-6}) = 576” with the 24 as coefficients instead of exponents.

UV theory of a “single double”

The lattice VOA VΛ(C)V_{\Lambda(C)} of a self-dual ternary code CC of length nn is a simple-current extension of V3Z⊗nV_{\sqrt{3\mathbb{Z}}}^{\otimes n}. Each copy V3ZV_{\sqrt{3\mathbb{Z}}} is the chiral half of the “U(1)3_3” WZW model, which can be realized as the infrared limit of the N=2\mathcal{N} = 2 Landau–Ginzburg model with superpotential Φ3\Phi^3.

The current J=i3∂XJ = \frac{i}{\sqrt{3}}\partial X and the weight-32\frac{3}{2} supercurrents G±=e±i3XG^\pm= e^{\pm i\sqrt{3}X} generate the N=2\mathcal{N} = 2 algebra with c=1c = 1. It is the ℓ=1\ell= 1 member of the N=2\mathcal{N} = 2 minimal series c=3ℓℓ+2c = \frac{3\ell}{\ell+2}, whose ℓ\ell-th member is the infrared limit of a Landau–Ginzburg model with superpotential W=Φℓ+2W = \Phi^{\ell+2} [45, 46].

There then must be a rather simple relation between nn copies of this minimal model and V‾Λ(C)⊗VΛ(C)\overline{V}_{\Lambda(C)} \otimes V_{\Lambda(C)}. On the VOA side, the extension data is encoded in the ternary code CC, while for the UV Landau–Ginzburg model, CC determines a subgroup of the Z3n\mathbb{Z}_3^n symmetry. Namely, each codeword cc acts on Φi\Phi_i by ωci\omega^{c_i} with ω=e2πi/3\omega= e^{2\pi i/3} a third root of unity. Therefore, the simplest and most natural lift of the doubled SCFT V‾Λ(C)⊗VΛ(C)\overline{V}_{\Lambda(C)} \otimes V_{\Lambda(C)} is the following.

Claim 1. For every self-dual ternary code CC of length nn, the Landau–Ginzburg orbifold (∑i=1nΦi3)/C(\sum_{i=1}^{n} \Phi_i^3)/C, with CC acting as a diagonal phase by Φi↦ωci⋅Φi\Phi_i \mapsto\omega^{c_i} \cdot\Phi_i, flows to V‾Λ(C)⊗VΛ(C)\overline{V}_{\Lambda(C)} \otimes V_{\Lambda(C)} in the infrared.

The physical justification for this claim is the following. Using the well-tested relation between the UV LG model and the N=2\mathcal{N}=2 minimal model in the infrared, as well as the standard assumption that gauging a symmetry that is preserved along an RG flow leads to a pair of theories that are again connected by an RG flow, one can reduce the claim to a computation just in the IR SCFTs. On one side, we have a product of nn copies of the Φ3\Phi^3 minimal model, where the sectors are labeled by λ∈F3n\lambda\in\mathbb{F}_3^n, with the same label on both sides. The orbifold by CC is expected to add twisted sectors in which the left and right labels can differ by an element of CC, and project onto right labels in C⊥=CC^\perp= C. What remains matches exactly the sectors of VΛ(C)⊗V‾Λ(C)V_{\Lambda(C)} \otimes\overline{V}_{\Lambda(C)} on the other side.

Claim 1 is reminiscent of Gepner models [47, 48] and of other Landau–Ginzburg orbifold constructions using minimal models [49, 50], which might suggest that it generalizes to codes that are not necessarily self-dual and even beyond ternary.

First of all, a non-self-dual but self-orthogonal (i.e. C⊂C⊥C \subset C^\perp) ternary code CC leads to a non-holomorphic VOA (with modules labeled by the cosets in C⊥/CC^\perp/C), but it can also be doubled via the diagonal pairing, and we claim that the doubled theory, V‾Λ(C)⊗C⊥/CVΛ(C)\overline{V}_{\Lambda(C)} \otimes_{C^\perp/C} V_{\Lambda(C)}, is still the RG fixed point of a (∑i=1nΦi3)/C(\sum_{i=1}^{n} \Phi_i^3)/C theory. A special case is the zero code C={0}C = \{0\}. Then Λ(C)=(3Z)n\Lambda(C) = (\sqrt{3}\mathbb{Z})^n, and VΛ(C)=V3ZnV_{\Lambda(C)} = V_{\sqrt{3}\mathbb{Z}}^n. Its diagonal double is indeed the nn copies of ℓ=1\ell= 1 minimal models. From this, one can add codewords to CC one by one, as long as they are linearly independent and orthogonal with themselves and each other, and recover the claim by “gauging one Z3\mathbb{Z}_3 at a time.”12 Then, it is also tempting to state a similar conjecture for (ℓ+2)(\ell+ 2)-nary codes (e.g. codes over Zℓ+2\mathbb{Z}_{\ell+2}) with ℓ>1\ell> 1, but there are two important differences. One is that the central charge of the minimal model is now c=3ℓℓ+2>1c = \frac{3\ell}{\ell+2} > 1, and the minimal model cannot be the doubled theory itself. Also, VOAs coming from (ℓ+2)(\ell+ 2)-nary codes generically don’t have supersymmetry (see [51, 52] for related work). Therefore, a natural conjecture seems to be that, for a self-orthogonal (ℓ+2)(\ell+ 2)-nary code CC, the non-chiral double of its corresponding VOA can be obtained from a SUSY-breaking RG flow from the (∑i=1nΦiℓ+2)/C\left(\sum_{i=1}^{n} \Phi_i^{\ell+2}\right)/C theory. This also has a somewhat trivial generalization for ℓ=0\ell= 0 (i.e. binary code), where the infrared limit of LG model—the “ℓ=0\ell= 0 minimal model”—is just a massive vacuum, and the conjecture should instead state that it can then “flow up” to the doubled theory in a non-supersymmetric way—almost a tautology. Therefore, the ℓ=1\ell= 1 case (ternary code) is at the “borderline” when the central charge of the minimal model has increased from the ℓ=0\ell= 0 case to match that of the VOA, and it will increase further for ℓ>1\ell> 1 to live above the VOA.

Finally, in Claim 1, the orbifold is without discrete torsion, and it would be interesting to explore, if one allows it, what the corresponding theories are on the code-VOA side.

A gapping test

The Landau–Ginzburg description in the UV makes it possible to gap the double via an N=(1,1)\mathcal{N} = (1, 1) superpotential—extra credit for preserving more supersymmetries than we need.

The UV theory itself has the real superpotential h0=Re⁡∑iΦi3h_0 = \operatorname{Re}\sum_i \Phi_i^3, and we deform it by a CC-invariant term,

h=h0+ϵ∑i∣Φi∣2.(46)h = h_0 + \epsilon\sum_i |\Phi_i|^2. \tag*{(46)}

For ϵ≠0\epsilon\ne0, hh is a Morse function, resolving, for each Φi\Phi_i, the degenerate critical point at 00 into four non-degenerate ones at

{0,−2ϵ3,−2ϵ3ω,−2ϵ3ω2}.\left\{0,-\frac{2\epsilon}{3},-\frac{2\epsilon}{3}\omega,-\frac{2\epsilon}{3}\omega^2\right\}.

For ϵ>0\epsilon> 0 the first is a minimum of hh (bosonic vacuum), fixed by Z3\mathbb{Z}_3, and the other three are index-1 saddle points (fermionic), permuted by Z3\mathbb{Z}_3. Turning on ϵ\epsilon therefore deforms the theory into the CC-orbifold of nn copies of a gapped theory with 1+31 + 3 massive vacua.

The elliptic genus is unchanged by this deformation and can be computed from the Witten index of the gapped theory. In our case, it is given by the average over pairs (c,c′)(c,c') of the

ensures that two lines have trivial braiding phase—both are required for condensation. When CC is already self-dual, the bulk is invertible and the theory becomes holomorphically factorizable. signed number of vacua fixed by both [53, 54, 49, 55]. Here a field with ci=ci′=0c_i=c'_i=0 contributes −2-2, while any other field contributes 1, from its fixed critical point Φi=0\Phi_i=0 alone.13 This gives the elliptic genus of the LG orbifold model,

I((∑i=1nΦi3)/C)=1∣C∣∑c,c′∈C(−2)#{i:ci=ci′=0}.(47)\mathcal{I}\left(\left(\sum_{i=1}^{n}\Phi_i^3\right)/C\right)=\frac{1}{\lvert C\rvert}\sum_{c,c'\in C}(-2)^{\#\{i:c_i=c'_i=0\}}. \tag*{(47)}

This turns out to equal the index squared of CC for every self-dual code,

Index⁡(C)2=1∣C∣∑c,c′∈C(−2)#{i:ci=ci′=0},(48)\operatorname{Index}(C)^2=\frac{1}{\lvert C\rvert}\sum_{c,c'\in C}(-2)^{\#\{i:c_i=c'_i=0\}}, \tag*{(48)}

which is exactly the elliptic genus of V‾Λ(C)⊗VΛ(C)\overline{V}_{\Lambda(C)}\otimes V_{\Lambda(C)}. Instead of proving the above identity here, we will actually work out the version refined by the left-moving R-symmetry U(1)LU(1)_L and then also match the refined count in Appendix § 8.

Now that we know how to gap the V‾Λ(C)⊗VΛ(C)\overline{V}_{\Lambda(C)}\otimes V_{\Lambda(C)} theory, by first flowing upward to the Landau–Ginzburg orbifold model and then deforming the superpotential, one might wonder whether a similar technique works for T‾⊗T\overline{T}\otimes T or P‾⊗P\overline{P}\otimes P. Unfortunately, this method does not directly extend to these cases, since the symmetries gauged in the construction of TT, T′T' and P=T′′P=T'' act on one chirality only and are invisible in the Landau–Ginzburg description.

Acknowledgment

We thank Theo Johnson-Freyd for discussion, and Ying-Hsuan Lin and Muyang Liu for collaborations on related research projects that in large part inspired the present approach. The work of the author is supported by research grant 42125 from VILLUM FONDEN and the Simons Collaboration on “New Structures in Low-dimensional Topology.” This material is based upon work supported by the National Science Foundation under Grant No. DMS-2424139, while the author was in residence at the Simons Laufer Mathematical Sciences Institute (SLMath) during the Fall 2026 semester. We would also like to thank the staff at SLMath for great hospitality and for helping set up the computational resources on which our computer-aided search was run.

Refined elliptic genera

The LG model has an N=(2,2)\mathcal{N}=(2,2) supersymmetry, with a U(1)L×U(1)RU(1)_L\times U(1)_R R-symmetry. As U(1)LU(1)_L commutes with the action of CC and fixes the right-moving supercharges, the elliptic genus of the orbifold can be refined by the left-moving R-charge J0J_0, whose fugacity we denote as y:=e2πizy := e^{2\pi i z}.

For a single field with W=Φ3W = \Phi^3 (of R-charge 13\frac{1}{3}), twisted by ωa\omega^a in space and by ωb\omega^b in time, the refined elliptic genus is [56, 57]

Fa,b(τ,z)=y−a/3θ1(τ,−2z3+aτ+b3)θ1(τ,z3+aτ+b3).(49)F_{a,b}(\tau,z)=y^{-a/3}\frac{\theta_1\left(\tau,-\frac{2z}{3}+\frac{a\tau+b}{3}\right)}{\theta_1\left(\tau,\frac{z}{3}+\frac{a\tau+b}{3}\right)}. \tag*{(49)}

At z=0z=0 it reduces to the signed counts in (47), with F0,0=−2F_{0,0}=-2 and Fa,b=1F_{a,b}=1 otherwise. The refined elliptic genus of the LG orbifold model is then

ZLG/C(τ,z)=1∣C∣∑c,c′∈C∏i=1nFci,ci′(τ,z).(50)Z_{\mathrm{LG}/C}(\tau,z)=\frac{1}{\lvert C\rvert}\sum_{c,c'\in C}\prod_{i=1}^{n}F_{c_i,c'_i}(\tau,z). \tag*{(50)}

On the other side, the elliptic genus of V‾Λ(C)⊗VΛ(C)\overline{V}_{\Lambda(C)}\otimes V_{\Lambda(C)} refined by J0J_0, with J=i3∑j∂XjJ=\frac{i}{\sqrt{3}}\sum_j\partial X_j acting on V‾Λ(C)\overline{V}_{\Lambda(C)}, is the index of CC (from the right-moving side) times the Ramond-sector supertrace χR(τ,z)=Tr⁡R(−1)FqL0−n/24yJ0\chi^{\mathrm{R}}(\tau,z)=\operatorname{Tr}_{\mathrm{R}}(-1)^Fq^{L_0-n/24}y^{J_0} of V‾Λ(C)\overline{V}_{\Lambda(C)} from the left-moving side.

It turns out that the two agree for every self-dual ternary code CC,

ZLG/C(τ,z)=Index⁡(C)⋅χR(τ,z).(51)Z_{\mathrm{LG}/C}(\tau,z)=\operatorname{Index}(C)\cdot\chi^{\mathrm{R}}(\tau,z). \tag*{(51)}

To see this, let ψu\psi_u, u∈F3u\in\mathbb{F}_3, be the same supertrace over the V3ZV_{\sqrt{3}\mathbb{Z}}-module of label uu, we have

ψu(τ,z)=1η(τ)∑m∈Zm≡u(mod3)(−1)my(m+3/2)/3q(m+3/2)2/6,(52)\psi_u(\tau,z)=\frac{1}{\eta(\tau)}\sum_{\substack{m\in\mathbb{Z}\\m\equiv u\pmod{3}}}(-1)^m y^{(m+3/2)/3}q^{(m+3/2)^2/6}, \tag*{(52)}

with u∈{0,1,2}u\in\{0,1,2\} standing for {0,+,−}\{0,+,-\}. Let ιu=ψu(τ,0)\iota_u=\psi_u(\tau,0) be its Witten index with ι0=0\iota_0=0 and ι±=±1\iota_\pm=\pm1.

The quintuple product identity (see [58] for a survey) gives, for each field,

13∑b∈Z3ω−bkFa,b=−ψ2a+κιa+κ.(53)\frac{1}{3}\sum_{b\in\mathbb{Z}_3}\omega^{-bk}F_{a,b}=-\psi_{2a+\kappa}\iota_{a+\kappa}. \tag*{(53)}

In ZLG/CZ_{\mathrm{LG}/C}, the sum over c′c' projects onto κ∈C⊥=C\kappa\in C^\perp=C, and the change of variables (c,κ)↦(2c+κ,c+κ)(c,\kappa)\mapsto(2c+\kappa,c+\kappa), which maps C×CC\times C onto itself, factorizes the sum into

(∑u∈C∏iψui)(∑w∈C∏iιwi)=χR⋅Index⁡(C).(54)\left(\sum_{u\in C}\prod_i\psi_{u_i}\right)\left(\sum_{w\in C}\prod_i\iota_{w_i}\right)=\chi^{\mathrm{R}}\cdot\operatorname{Index}(C). \tag*{(54)}

At y=1y=1, (A.3) gives the unrefined equality between the Witten indices found in § 6.4. One can similarly show that—now working with the IR minimal model instead of the UV Landau–Ginzberg description—for the other three spin structures, the partition functions of the two theories are also equal. More details are given in [15].

References

  1. [1]G. Segal, What is an elliptic object?, in Elliptic Cohomology: Geometry, Applications, and Higher Chromatic Analogues (H. R. Miller and D. C. Ravenel, eds.), no. 342 in London Mathematical Society Lecture Note Series, pp. 306–317. Cambridge University Press, 2007.DOI
  2. [2]S. Stolz and P. Teichner, What is an elliptic object?, pp. 247–343. London Mathematical Society Lecture Note Series. Cambridge University Press, 2004.
  3. [3]S. Stolz and P. Teichner, Supersymmetric field theories and generalized cohomology, Proc. Symp. Pure Math. 83 (2011) 279–340, [arXiv:1108.0189].arxiv.org/abs/1108.0189
  4. [4]D. Gaiotto and T. Johnson-Freyd, Holomorphic scfts with small index, Canadian Journal of Mathematics 74 (2022), no. 2 573–601, [arXiv:1811.00589].
  5. [5]S. Gukov, D. Pei, P. Putrov, and C. Vafa, 4-manifolds and topological modular forms, JHEP 05 (2021) 084, [arXiv:1811.07884].arxiv.org/abs/1811.07884
  6. [6]Y.-H. Lin and D. Pei, Holomorphic CFTs and Topological Modular Forms, Commun. Math. Phys. 401 (2023), no. 1 325–332, [arXiv:2112.10724].DOI
  7. [7]Y.-H. Lin, Topological modularity of Monstrous Moonshine, Ann. Henri Poincaré 25 (2024), no. 4 2427–2452, [arXiv:2207.14076].arxiv.org/abs/2207.14076
  8. [8]J. Albert, J. Kaidi, and Y.-H. Lin, Topological modularity of Supermoonshine, PTEP 2023 (2023), no. 3 033B06, [arXiv:2210.14923].DOI
  9. [9]D. Pei and D. H. Wu, E-Strings and Four-Manifolds, arXiv:2511.11807.arxiv.org/abs/2511.11807
  10. [10]M. J. Hopkins, Algebraic topology and modular forms, Proceedings of the ICM, Beijing 1 (2002) 283–309.DOI
  11. [11]C. L. Douglas, J. Francis, A. G. Henriques, and M. A. Hill, eds., Topological Modular Forms, vol. 201 of Mathematical Surveys and Monographs. American Mathematical Society, 2014.DOI
  12. [12]J. F. Duncan, Super-moonshine for Conway’s largest sporadic group, Duke Math. J. 139 (2007), no. 2 255–315, [math/0502267].DOI
  13. [13]R. Dijkgraaf, G. W. Moore, E. P. Verlinde, and H. L. Verlinde, Elliptic genera of symmetric products and second quantized strings, Commun. Math. Phys. 185 (1997), no. 1 197–209, [hep-th/9608096].DOI
  14. [14]C. R. Harris et al., Array programming with NumPy, Nature 585 (2020) 357–362.
  15. [15]Fable 5.1, GPT-6 Astra, and Opus 5.5, “Holomorphic SCFTs with minimal index: research notes, certificates and code.” https://du-pei-physics.github.io/holomorphic-svoa/, 2026. AI-generated research note, supervised by D. Pei.
  16. [16]J. H. Conway and N. J. A. Sloane, Sphere Packings, Lattices and Groups, vol. 290 of Grundlehren der mathematischen Wissenschaften. Springer, 3rd ed., 1999.
  17. [17]R. E. Borcherds, Vertex algebras, Kac-Moody algebras, and the Monster, Proc. Nat. Acad. Sci. USA 83 (1986), no. 10 3068–3071.pubmed.ncbi.nlm.nih.gov/16593694
  18. [18]I. Frenkel, J. Lepowsky, and A. Meurman, Vertex operator algebras and the Monster, vol. 134 of Pure and Applied Mathematics. Academic Press, Boston, 1988.
  19. [19]C. Dong, Vertex algebras associated with even lattices, J. Algebra 161 (1993), no. 1 245–265.DOI
  20. [20]V. Kac, Vertex Algebras for Beginners, vol. 10 of University Lecture Series. American Mathematical Society, 2nd ed., 1998.
  21. [21]W. C. Huffman and V. Pless, Fundamentals of Error-Correcting Codes. Cambridge University Press, 2003.DOI
  22. [22]L. Bahl, J. Cocke, F. Jelinek, and J. Raviv, Optimal decoding of linear codes for minimizing symbol error rate, IEEE Trans. Inform. Theory 20 (1974), no. 2 284–287.
  23. [23]A. R. Calderbank, G. D. Forney, Jr., and A. Vardy, Minimal tail-biting trellises: the Golay code and more, IEEE Trans. Inform. Theory 45 (1999), no. 5 1435–1455.
  24. [24]A. J. Viterbi, Convolutional Codes and Their Performance in Communication Systems, IEEE Trans. Commun. Technol. 19 (1971), no. 5 751–772.
  25. [25]R. J. McEliece, On the BCJR trellis for linear block codes, IEEE Trans. Inform. Theory 42 (1996), no. 4 1072–1092.DOI
  26. [26]G. D. Forney, Jr., Codes on graphs: duality and MacWilliams identities, IEEE Trans. Inform. Theory 57 (2011), no. 3 1382–1397, [arXiv:0911.5508].arxiv.org/abs/0911.5508
  27. [27]T. Johnson-Freyd, The Moonshine Anomaly, Commun. Math. Phys. 365 (2019), no. 3 943–970, [arXiv:1707.08388].
  28. [28]C. Vafa, Modular invariance and discrete torsion on orbifolds, Nucl. Phys. B 273 (1986) 592–606.DOI
  29. [29]L. Fidkowski and A. Kitaev, Topological phases of fermions in one dimension, Phys. Rev. B 83 (2011), no. 7 075103, [arXiv:1008.4138].DOI
  30. [30]A. Kapustin, R. Thorngren, A. Turzillo, and Z. Wang, Fermionic Symmetry Protected Topological Phases and Cobordisms, JHEP 12 (2015) 052, [arXiv:1406.7329].DOI
  31. [31]The GAP Group, “GAP – Groups, Algorithms, and Programming, Version 4.16.1.” https://www.gap-system.org, 2026.
  32. [32]M. Yu, Genus-one data and anomaly detection, Phys. Rev. D 105 (2022), no. 10 106007, [arXiv:2105.02231].DOI
  33. [33]J. Davighi, B. Gripaios, and N. Lohitsiri, Anomalies of non-Abelian finite groups via cobordism, JHEP 09 (2022) 147, [arXiv:2207.10700].
  34. [34]Fable 5.1, GPT-6 Astra, and Opus 5.5, “The index of self-dual ternary codes: a proof of a conjecture of Gaiotto and Johnson-Freyd.” https://du-pei-physics.github.io/gj-conjecture/, 2026. AI-generated research note, supervised by D. Pei.
  35. [35]H. S. Chakraborty, W. Chiari, T. Miezaki, and M. Oura, Neighbors, neighbor graphs and invariant rings in coding theory, arXiv:2411.04647.arxiv.org/abs/2411.04647
  36. [36]Y. Tachikawa and M. Yamashita, Anderson duality of topological modular forms and its differential-geometric manifestations, arXiv:2305.06196.arxiv.org/abs/2305.06196
  37. [37]T. Johnson-Freyd and M. Yamashita, On the 576-fold periodicity of the spectrum SQFT: The proof of the lower bound via the Anderson duality pairing, arXiv:2404.06333.
  38. [38]S. K. Devalapurkar, The Ando-Hopkins-Rezk orientation is surjective, arXiv:1911.10534.
  39. [39]D. Gaiotto, T. Johnson-Freyd, and E. Witten, A note on some minimally supersymmetric models in two dimensions, Proc. Symp. Pure Math. 103.2 (2021) 203–221, [arXiv:1902.10249].arxiv.org/abs/1902.10249
  40. [40]A. Kitaev, Periodic table for topological insulators and superconductors, AIP Conf. Proc. 1134 (2009) 22–30, [arXiv:0901.2686].
  41. [41]L. Fidkowski and A. Kitaev, Effects of interactions on the topological classification of free fermion systems, Phys. Rev. B 81 (2010) 134509, [arXiv:0904.2197].DOI
  42. [42]S. Gukov, D. Pei, and P. Putrov, Tralities of minimally supersymmetric 2d gauge theories, JHEP 04 (2020) 079, [arXiv:1910.13455].DOI
  43. [43]J. Bao, M. Yamazaki, and D. Zhou, Elliptic genera of 2d N = (0, 1) gauge theories, JHEP 04 (2026) 080, [arXiv:2508.06865].DOI
  44. [44]J. Wang and Y.-Z. You, Symmetric Mass Generation, Symmetry 14 (2022), no. 7 1475, [arXiv:2204.14271].DOI
  45. [45]C. Vafa and N. P. Warner, Catastrophes and the Classification of Conformal Theories, Phys. Lett. B 218 (1989) 51.DOI
  46. [46]E. J. Martinec, Algebraic Geometry and Effective Lagrangians, Phys. Lett. B 217 (1989) 431.DOI
  47. [47]D. Gepner, Exactly Solvable String Compactifications on Manifolds of SU(N) Holonomy, Phys. Lett. B 199 (1987) 380–388.
  48. [48]D. Gepner, Space-Time Supersymmetry in Compactified String Theory and Superconformal Models, Nucl. Phys. B 296 (1988) 757.
  49. [49]C. Vafa, String Vacua and Orbifoldized L-G Models, Mod. Phys. Lett. A 4 (1989) 1169.DOI
  50. [50]K. A. Intriligator and C. Vafa, Landau-Ginzburg orbifolds, Nucl. Phys. B 339 (1990) 95.DOI
  51. [51]K. Kawabata and S. Yahagi, Fermionic CFTs from classical codes over finite fields, JHEP 05 (2023) 096, [arXiv:2303.11613].DOI
  52. [52]K. Kawabata and S. Yahagi, Elliptic genera from classical error-correcting codes, JHEP 01 (2024) 130, [arXiv:2308.12592].arxiv.org/abs/2308.12592
  53. [53]L. J. Dixon, J. A. Harvey, C. Vafa, and E. Witten, Strings on Orbifolds, Nucl. Phys. B 261 (1985) 678–686.
  54. [54]L. J. Dixon, J. A. Harvey, C. Vafa, and E. Witten, Strings on Orbifolds. 2., Nucl. Phys. B 274 (1986) 285–314.
  55. [55]F. Hirzebruch and T. Höfer, On the Euler number of an orbifold, Math. Ann. 286 (1990) 255–260.DOI
  56. [56]E. Witten, On the Landau-Ginzburg description of N=2 minimal models, Int. J. Mod. Phys. A 9 (1994) 4783–4800, [hep-th/9304026].
  57. [57]T. Kawai, Y. Yamada, and S.-K. Yang, Elliptic genera and N=2 superconformal field theory, Nucl. Phys. B 414 (1994) 191–212, [hep-th/9306096].DOI
  58. [58]S. Cooper, The quintuple product identity, Int. J. Number Theory 2 (2006), no. 1 115–161.

Paper details

vlnc:W_qt7rfc6b
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 3

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

AI coding agents

View statement
Looking for related papers…
Contents