A simple perturbation of Vafa-Witten equations and a transversality result
Bo Dai, Ren Guan · 2025
Read A simple perturbation of Vafa-Witten equations and a transversality result by Bo Dai, Ren Guan (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-2505-14702.txt.
Introduction
A set of maths rules for gauge fields, the Vafa-Witten rules, form a space of answers whose clean shape is key to their use. Bo Dai and Ren Guan nudge those rules in a 2025 study with a small added term. They ask if the nudged space stays smooth for most picks of the nudge size, for two gauge groups.
Catalog record
| CREATOR | Bo Dai, Ren Guan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 15 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-14702
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A simple perturbation of Vafa-Witten equations and a transversality result Bo Dai, Ren Guan Abstract . We consider a simple perturbation of the Vafa-Witten equations, and prove that for generic perturbation parameter, the full rank part of the perturbed Vafa-Witten moduli space satisfies transversality condition, when the structure group is S U ( 2 ) SU(2) or S O ( 3 ) SO(3) . 1. Introduction The Vafa-Witten equations came from a topological twist of \mathcal{N}=4 supersymmetric Yang-Mills theory. Vafa and Witten conjectured that the partition function of their theory transforms as modular form under S -duality, an important duality in quantum field theory. Let X be a smooth, oriented, Riemannian 4-manifold, P be a principal bundle over X with structure group G , \mathfrak{g}_{P} be the adjoint bundle of P . The Vafa-Witten equations have the following form: (1.1) \left\{\begin{aligned} &d_{A}^{*}B+d_{A}C=0,\\ &F_{A}^{+}+\frac{1}{8}[B\centerdot B]+\frac{1}{2}[B,C]=0,\\ \end{aligned}\right. where A is a connection of the principal bundle P , B is a self-dual 2-form with values in the adjoint bundle \mathfrak{g}_{P} , C is a section of the adjoint bundle \mathfrak{g}_{P} , and [B\centerdot B] is a combination of the Lie bracket of \mathfrak{g} and the Lie algebra structure of \Lambda^{2,+}\mathbb{R}^{4}\cong so(3) . The Vafa-Witten equations are invariant under the action of gauge group. The Vafa-Witten moduli space is the space of solutions to the Vafa-Witten equations modulo gauge transformation. Mares and Taubes did fundamental work to the analytic aspect of Vafa-Witten theory, [ 8 , 13 ] , but key challenges remain, such as the compactification and transversality of the Vafa-Witten moduli spaces. Tanaka and Thomas made key contribution to the algebraic aspect of Vafa-Witten theory. They defined Vafa-Witten invariants for projective surfaces by the method of derived algebraic geometry and virtual intersection theory, and verified S -duality in some cases, [ 11 , 12 ] . Since then, the algebraic aspect of Vafa-Witten theory developed rapidly. In this short note, we consider a simple perturbed version of the Vafa-Witten equations, and obtain a transversality result of the perturbed Vafa-Witten moduli spaces. The perturbed Vafa-Witten equations on X are (1.2) \left\{\begin{aligned} &d_{A}^{*}B+d_{A}C=0,\\ &F_{A}^{+}+\frac{1}{8}[B\centerdot B]+\frac{1}{2}[B,C]+\tau B=0,\\ \end{aligned}\right. where the triple (A,B,C) is as above, \tau is a section of the bundle \mathfrak{gl}(\Lambda^{2,+}T^{*}X) . We note that \tau acts on \mathfrak{g}_{P} trivially, i.e., as identity. Thus the perturbed Vafa-Witten equations are invariant under gauge group action, and the perturbed moduli space \mathcal{M}_{VW,\tau}(P) is well-defined. To apply analytic tools, we need to work on appropriate Sobolev spaces. Let \displaystyle\mathcal{C}_{k}(P) \displaystyle:=\mathcal{A}_{k}(P)\times L_{k}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{2,+})\times L_{k}^{2}(X,\mathfrak{g}_{P}), \displaystyle\mathcal{C}_{k-1}^{\prime}(P) \displaystyle:=L_{k-1}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{1})\times L_{k-1}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{2,+}), be the configuration spaces, C^{r}(X,\mathfrak{gl}(\Lambda^{2,+})) be the parameter space, where r>k\geq 2 are integers. Let \mathcal{G}_{k+1}(P) be the group of L_{k+1}^{2} gauge transformations. In the remaining part of this note, we take the structure group G to be SU(2) or SO(3) , and follow notations in [ 2 ] . Theorem 1.1 . Let X be a closed, oriented, smooth, Riemannian 4-manifold, and P\to X be a principal SU(2) - or SO(3) -bundle. Then for generic perturbation \tau\in C^{r}(X,\mathfrak{gl}(\Lambda^{2,+})) , where r>k\geq 2 are integers, the full rank part of the moduli space, \mathcal{M}_{VW,\tau}^{(3)}(P) , is a smooth, oriented, zero-dimensional manifold in the quotient space \mathcal{B}_{k}(P)=\mathcal{C}_{k}(P)/\mathcal{G}_{k+1}(P) . Remark 1.2 . Other kind of perturbations of Vafa-Witten equations had been studied in [ 4 , 5 , 10 ] . Remark 1.3 . A PhD student, Mr. Zhengxiong Cao from Humboldt-Berlin, pointed out an error in our earlier paper [ 2 ] . That is, the three equations in [ 2 ] (4.3) do not hold separately. Theorem 1.1 is a replacement of Theorems 1, 4 and 5 in [ 2 ] . 2. Prove of the main theorem Consider the parameterized Vafa-Witten map (2.1) \displaystyle F:C^{r}(X,\mathfrak{gl}(\Lambda^{2,+}))\times\mathcal{C}_{k}(P) \displaystyle\to\mathcal{C}_{k-1}^{\prime}(P), \displaystyle F(\tau,A,B,C) \displaystyle:=\begin{pmatrix}d_{A}^{*}B+d_{A}C\\ F_{A}^{+}+\frac{1}{8}[B\centerdot B]+\frac{1}{2}[B,C]+\tau B\\ \end{pmatrix}. Lemma 2.1 . With notation in ( 2.1 ), at any point (\tau,A,B,C) where F(\tau,A,B,C)=0 and B has rank 3, the differential dF at this point is surjective. Proof. The differential (dF)_{(\tau,A,B,C)} differs from d^{1}_{(A,B,C)} in [ 8 ] page 28 slightly, given by (dF)_{(\tau,A,B,C)}:C^{r}(X,\mathfrak{gl}(\Lambda^{2,+}))\oplus L_{k}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{1})\oplus L_{k}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{2,+})\oplus L_{k}^{2}(X,\mathfrak{g}_{P})\to\mathcal{C}_{k-1}^{\prime}(P), (2.2) (\delta\tau,a,b,c)\mapsto\begin{pmatrix}d_{A}^{*}b+d_{A}c-[B\centerdot a]-[C,a]\\ d_{A}^{+}a+\frac{1}{4}[B\centerdot b]+\frac{1}{2}[b,C]+\frac{1}{2}[B,c]+\tau b+(\delta\tau)B\\ \end{pmatrix}. The linearization of the action of gauge group \mathcal{G}_{k+1}(P) at a point (\tau,A,B,C)\in C^{r}(X,\mathfrak{gl}(\Lambda^{2,+}))\times\mathcal{C}_{k}(P) is given by d_{(\tau,A,B,C)}^{0}:L_{k+1}^{2}(X,\mathfrak{g}_{P})\to C^{r}(X,\mathfrak{gl}(\Lambda^{2,+}))\oplus L_{k}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{1})\oplus L_{k}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{2,+})\oplus L_{k}^{2}(X,\mathfrak{g}_{P}), (2.3) \xi\mapsto(0,d_{A}\xi,[B,\xi],[C,\xi]). The L^{2} -adjoint of d_{(\tau,A,B,C)}^{0} is (2.4) d_{(\tau,A,B,C)}^{0,*}(\delta\tau,a,b,c)=d_{A}^{*}a+[b\cdot B]+[c,C]. Since (dF)_{(\tau,A,B,C)}+d_{(\tau,A,B,C)}^{0,*} is a family of first order elliptic operators depending on \delta\tau linearly, (dF)_{(\tau,A,B,C)} has closed image. Assume F(\tau,A,B,C)=0 and B has rank 3. By the regularity theorem for elliptic equations with C^{r} coefficients (cf. [ 3 ] , Proposition 3.7 and Remark 5.1), we may assume that (A,B,C) is a C^{r} solution of the perturbed Vafa-Witten equations ( 1.2 ). To prove that (dF)_{(\tau,A,B,C)} is surjective, it suffices to show that (\mathrm{Im}(dF)_{(\tau,A,B,C)})^{\perp}=0 . Suppose (\phi,\psi)\in(\mathrm{Im}(dF)_{(\tau,A,B,C)})^{\perp} . Then we have (2.5) \langle d_{A}^{*}b+d_{A}c-[B\centerdot a]-[C,a],\phi\rangle_{L^{2}}+\langle d_{A}^{+}a+\frac{1}{4}[B\centerdot b]+\frac{1}{2}[b,C]+\frac{1}{2}[B,c]+\tau b+(\delta\tau)B,\psi\rangle_{L^{2}}=0 for all (\delta\tau,a,b,c)\in C^{r}(X,\mathfrak{gl}(\Lambda^{2,+}))\oplus L_{k}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{1})\oplus L_{k}^{2}(X,\mathfrak{g}_{P}\otimes\Lambda^{2,+})\oplus L_{k}^{2}(X,\mathfrak{g}_{P}) . Computing the L^{2} -adjoint operator of (dF)_{(\tau,A,B,C)} from equation ( 2.5 ), we deduce that (\phi,\psi) satisfy the following system of equations (2.6) \begin{cases}d_{A}^{*}\psi-[\phi\centerdot B]-[\phi,C]=0,&\\ d_{A}^{+}\phi+\frac{1}{4}[\psi\centerdot B]-\frac{1}{2}[\psi,C]+\tau^{t}\psi=0,&\\ d_{A}^{*}\phi+\frac{1}{2}[\psi\cdot B]=0,&\\ \langle(\delta\tau)B,\psi\rangle_{L^{2}}=0,\ \forall\,\delta\tau\in C^{r}(X,\mathfrak{gl}(\Lambda^{2,+})).&\\ \end{cases} Note that the first three equations in ( 2.6 ) differ from the expression for the operator d^{1,*}_{(B,C,A)} on page 29 of [ 8 ] by one term \tau^{t}\psi , where \tau^{t} is the transpose of \tau with respect to the Riemannian metric on the bundle \Lambda^{2,+} . Since \delta\tau\in C^{r}(X,\mathfrak{gl}(\Lambda^{2,+})) is arbitrary, from the 4th equation of ( 2.6 ), we deduce that \psi=0 on X^{(3)}(B) , where B has pointwise rank 3. The first equation in ( 2.6 ) implies that (2.7) [\phi\centerdot B]+[\phi,C]=0 on X^{(3)}(B) . Lemma 2.3 below implies that \phi=0 on X^{(3)}(B) . The first three equations in ( 2.6 ) is a system of first order linear differential equations of elliptic type for (\phi,\psi) with C^{r} coefficients. By the unique continuation theorem (cf [ 1 , 7 ] ), we deduce that (\phi,\psi)=(0,0) on X . This implies that (dF)_{(\tau,A,B,C)} is surjective. ∎ Corollary 2.2 . Let \mathfrak{M}_{VW}(P)=\bigcup_{\tau}\mathcal{M}_{VW,\tau}(P) be the parameterized Vafa-Witten moduli space, where \tau ranges over C^{r}(X,\mathfrak{gl}(\Lambda^{2,+})) , and \mathfrak{M}_{VW}^{(3)}(P) be the full rank part of the parameterized Vafa-Witten moduli space where the component B has rank 3. Then \mathfrak{M}_{VW}^{(3)}(P) is an (infinite dimensional) smooth manifold. The projection (2.8) \pi:\mathfrak{M}_{VW}^{(3)}(P)\to C^{r}(X,\mathfrak{gl}(\Lambda^{2,+})),\quad[(\tau,A,B,C)]\mapsto\tau, is Fredholm of index 0. Proof. By Lemma 1 in [ 2 ] , when B has rank 3, we have \mathrm{Stab}(A,B,C)=Z(G) . Lemma 2.1 implies that \mathfrak{M}_{VW}^{(3)}(P) is a smooth manifold. For fixed \tau , (dF)_{(\tau,A,B,C)}+d_{(\tau,A,B,C)}^{0,*} is a first order elliptic operator of index 0. Thus the projection \pi is Fredholm of index 0. ∎ Proof of Theorem 1.1 . By Corollary 2.2 and the Sard-Smale theorem (cf. [ 9 ] , Theorem A.5.1), for generic perturbation \tau in the sense of second category, the full rank part of the perturbed Vafa-Witten moduli space \mathcal{M}_{VW,\tau}^{(3)}(P) is a smooth 0-dimensional manifold. The orientation of the moduli space is established in [ 2 ] and [ 6 ] . ∎ Lemma 2.3 . Suppose that B\in\mathfrak{su}(2)\otimes\Lambda^{2,+}\mathbb{R}^{4} has rank 3, and C\in\mathfrak{su}(2) . Then the linear map \displaystyle L_{B,C}:\mathfrak{su}(2)\otimes\Lambda^{1}\mathbb{R}^{4} \displaystyle\to\mathfrak{su}(2)\otimes\Lambda^{1}\mathbb{R}^{4} \displaystyle\phi \displaystyle\mapsto[B\centerdot\phi]+[C,\phi] is an isomorphism. Proof. According to the singular value decomposition for \mathfrak{su}(2)\otimes\Lambda^{2,+}\mathbb{R}^{4} in [ 8 ,