betaFeedback

WorksPhysics and Astronomy

Global Green-Schwarz response on quaternionic sphere bundles

Abstract

Local anomaly polynomials describe local variations, but a normalized Green-Schwarz phase also depends on global background data. We explain this dependence for ten-dimensional three-sphere bundles of quaternionic lines over closed spin seven-manifolds. The backgrounds retain two specified Spin(16)\mathrm{Spin}(16) gauge bundles and an integral differential trivialization on the base, including its boundary identifications. The trivialized class is the sum of half the first Pontryagin class of the tangent bundle and the second Chern class of the quaternionic line's underlying complex bundle. For these backgrounds, the calculation combines a generally nonzero local Green-Schwarz integral with a real eta invariant evaluated by a twelve-dimensional Atiyah-Patodi-Singer filling. Together, these terms determine an integral relative characteristic number; the additive response is one third of that number, modulo integers. When the base's third and fourth rational cohomology groups vanish, the response equals eight times the Crowley-Goette invariant, without assuming that the base is simply connected. A calibration using the actual quaternionic line bundle gives additive value one third on Sp(2)\mathrm{Sp}(2), recovering the known third-root phase. Distinct integral String data on one fixed ten-manifold give all three phase values, showing which global background information the response can detect. The physical identifications retain their prescribed fermion data, while the separate application to integral topological modular forms and vertex operator superalgebras requires compatible integral lifts. The response calculation constructs neither a microscopic realization nor an equivalence of complete theories.

Global information in the Green–Schwarz response

Green and Schwarz showed how suitable local interactions can cancel gauge and gravitational anomalies in ten-dimensional supergravity coupled to Yang–Mills for the appropriate gauge groups [1]. Global information requires a further description. In the anomaly-line framework, curvature measures the local anomaly, while a flat line can retain nontrivial holonomy [2 §2.2]. The scalar studied here compares a normalized Green–Schwarz section with a spin reference; a nontrivial comparison does not by itself signal a failure of anomaly cancellation.

Evaluating this normalized response on a closed background requires global data even when the background gauge connections on the ten-manifold are trivial. Tachikawa–Zhang identified a nontrivial third-root response in their comparison of heterotic Green–Schwarz terms [3]. Quaternionic sphere bundles provide a family of ten-manifolds on which the response can be computed and its dependence on the background made explicit. The calculation relates a local Green–Schwarz integral and a closed fermion eta phase to one integral relative characteristic number. Its value can change with the integral String datum while the underlying ten-manifold and the fibre period remain fixed. Related worldsheet developments include the tertiary invariant of two-dimensional minimally supersymmetric field theories [4] and the study of non-supersymmetric heterotic branes and their Green–Schwarz pairings [5]. Wada–Yonekura [6], which appears simultaneously with this paper, evaluate discrete Green–Schwarz couplings by a complementary worldsheet/tertiary-invariant method, including an order-three response on Sp⁡(2)\operatorname{Sp}(2). The present calculation instead uses a spacetime sphere-bundle construction and a twelve-dimensional APS filling to evaluate the normalized response.

The mathematical derivation of the sphere-bundle formula and the examples below are due to Li–Sun, in their unpublished working manuscript Green–Schwarz scalars on quaternionic sphere bundles [7]. The present companion explains the background data, calculation and physical consequences of those results for readers familiar with anomalies but not with differential cohomology or the Crowley–Goette invariant. Its purpose is to make the local integral, the closed spectral contribution and their common relative pairing visible in one calculation, while retaining the different premises of the physical and generalized-cohomology comparisons.

For a closed ten-manifold MM with differential String data and prescribed trivial gauge connections and frames, the observable is an exponentiated scalar O(M)=exp⁡(2πiϕ(M))\mathcal{O}(M)=\exp(2\pi i\phi(M)), with additive value ϕ(M)∈R/Z\phi(M)\in\mathbb{R}/\mathbb{Z}. Its normalization compares the Green–Schwarz section to a spin reference with trivial framed gauge bundles. Let NN be an eleven-dimensional filling carrying the gauge and differential trivialization data, with three-form component H3H_3, and let WrefW_{\mathrm{ref}} be the spin reference. The local eight-form Y8Y_8 is fixed by the full real fermion anomaly polynomial. The additive phase combines the local integral with a closed eta term:

ϕ(M)=−aN−μ∗(Z),aN=∫NH3Y8,Z=N‾∪MWref.(1)\phi(M)=-a_N-\mu_*(Z),\qquad a_N=\int_N H_3Y_8,\qquad Z=\overline{N}\cup_M W_{\mathrm{ref}}. \tag*{(1)}

The orientation-reversed filling and reference form the closed spin/gauge double ZZ. The closed term μ∗\mu_* uses a real normalization of the Atiyah–Patodi–Singer eta invariant of the specified virtual real fermion coefficient, with ξ=(ηAPS+dim⁡Cker⁡)/2\xi=(\eta_{\mathrm{APS}}+\dim_{\mathbb{C}}\ker)/2 and the factor −1/2-1/2 multiplying ξ\xi, as well as the specified finite tangent-connection transgression. The kernel term, reference and the two minus signs are part of the observable. § 2 gives the complete definition in (5).

The scalar describes the specified anomaly response. Its identification with the Spin⁡(32)/Z2\operatorname{Spin}(32)/\mathbb{Z}_2 heterotic Green–Schwarz phase divided by the E8×E8E_8\times E_8 phase uses the published fermion and Green–Schwarz prescription: the same chirality adjoint gauginos, identical gravitational and ghost sectors, the normalized section and reference pairing, and no additional independent invertible sector. Section 9 states this input and compares the actual closed real Dirac systems. Equality of local polynomials would leave precisely that global identification undetermined. The response studied here does not specify a complete partition function or a microscopic realization.

The geometry makes its global inputs concrete. Let BB be a closed connected spin seven-manifold, with specified Spin structure, metric and a quaternionic line VV with connection. Its unit sphere bundle M=S(V)M=S(V) has dimension ten. If UU is the underlying complex rank-two bundle, write c=c2(U)c=c_2(U) and λ=p1/2\lambda=p_1/2. The extra base datum is an integral differential trivialization hˇ\check{h}, with three-form component hh, satisfying

dhˇ=λˇ(TB)+cˇ,dh=λ(TB)+c.(2)\mathrm{d}\check{h}=\check{\lambda}(TB)+\check{c},\qquad\mathrm{d}h=\lambda(TB)+c. \tag*{(2)}

The second equation is the curvature equation; the first also retains the integral nullhomotopy and its specified morphisms. A primitive of the four-form alone would not retain the integer information used by the response. This datum induces the differential String datum sˇM\check{s}_M on S(V)S(V). All gauge bundles used to construct the filling lift to the actual common cover Spin⁡(16)×Spin⁡(16)\operatorname{Spin}(16)\times\operatorname{Spin}(16).

The disk bundle N=D(V)N=D(V) provides the induced boundary data and the local integral. With the base-then-fibre orientation and the outward boundary orientation used throughout, that integral is aN=13∫Bhca_N=\frac{1}{3}\int_B hc and is generally nonzero. Computing only the eta term would therefore miss the dependence on hˇ\check{h}. To evaluate the closed term in (1), extend the actual pair (B,V)(B,V) over a spin eight-manifold WW with a quaternionic line V~\widetilde{V}. Associating the doubled four-dimensional fibre over WW produces an auxiliary twelve-dimensional spin/gauge filling. An ordinary Atiyah–Patodi–Singer calculation [8], including real index parity and the finite connection correction, evaluates the closed eta phase. This auxiliary filling carries the Spin and gauge data required by the index theorem, but its anomaly four-cocycle is not trivialized. That distinction allows the scalar to be nontrivial.

The specified integral boundary nullhomotopy determines a relative class zz, recording the bulk characteristic class and its chosen boundary trivialization. Its cohomology group and the map to absolute cohomology are

z∈H4(W,B;Z),j:H4(W,B;Z)⟶H4(W;Z).z\in H^4(W,B;\mathbb{Z}),\qquad j:H^4(W,B;\mathbb{Z})\longrightarrow H^4(W;\mathbb{Z}).

The map jj forgets the boundary trivialization. Put cW=c2(U~)c_W=c_2(\widetilde{U}), where U~\widetilde{U} underlies V~\widetilde{V}, and λW=λ(TW)\lambda_W=\lambda(TW). The bulk integral and retained local term then combine into

ϕ(S(V),sˇM)=K3(modZ),\phi(S(V),\check{s}_M)=\frac{K}{3}\pmod{\mathbb{Z}},
K=⟨cW⌣z,[W,B]⟩∈Z,jz=λW+cW.(3)K=\langle c_W\smile z,[W,B]\rangle\in\mathbb{Z},\qquad jz=\lambda_W+c_W. \tag*{(3)}

Here ⌣\smile is the cohomological cup product and [W,B][W,B] is the relative fundamental class. In forms, the pairing is ∫WcW(λW+cW)−∫Bh∧c\int_W c_W(\lambda_W+c_W)-\int_B h\wedge c; the boundary subtraction is what turns the APS expression into this integral relative product. Changing WW with all boundary data fixed changes KK by a multiple of 24, and KK is even. Thus the normalized response is a third root of unity on this sphere-bundle domain. The relative integer is the output of the characteristic calculation; no charge interpretation is needed to reach this conclusion.

The relative pairing connects the response to an established secondary invariant of the base. If H3(B;Q)=H4(B;Q)=0H^{3}(B;\mathbb{Q})=H^{4}(B;\mathbb{Q})=0, Crowley–Goette’s invariant tB(V)t_B(V) is defined and their extrinsic formula has the same relative pairing with a different normalization [9 §1, Proposition 1.3]:

tB(V)=K24,ϕ(S(V),s~M)=8tB(V) in R/Z.(4)t_B(V)=\frac{K}{24},\qquad\phi(S(V),\tilde{s}_M)=8t_B(V)\ \text{in }\mathbb{R}/\mathbb{Z}. \tag*{(4)}

These two rational-cohomology vanishings are the exact additional domain restriction; simple connectedness is not required. They make the rational relative lift unique, while the integral differential data remain present. The formula for ϕ\phi in (3) applies more generally and does not define an extension of tt outside its domain.

For the original Sp⁡(2)\operatorname{Sp}(2) background, an actual quaternionic-line filling yields K=−2K=-2, hence ϕ=1/3\phi=1/3 and O=exp⁡(2πi/3)\mathcal{O}=\exp(2\pi i/3). Recovering the published value requires the connection and framing transport, in addition to this number. The general sphere construction has outward fibre period −2-2. The right-framed period +1+1 of the original background is obtained by the special inversion and differential transport on Sp⁡(2)\operatorname{Sp}(2); it is not a generic reidentification of the fibre datum.

The larger domain exposes information that neither the ten-manifold nor its fibre period specifies. At fixed metric and gauge connection, changing h~\tilde{h} by a degree-three differential character with integral characteristic class α∈H3(B;Z)\alpha\in H^{3}(B;\mathbb{Z}) changes the scalar by −⟨α⌣c,[B]⟩/3-\langle\alpha\smile c,[B]\rangle/3. Li–Sun’s construction [7] gives one fixed ten-manifold with choices s~n\tilde{s}_n for which O(S(V),s~n)=exp⁡(4πin/3)\mathcal{O}(S(V),\tilde{s}_n)=\exp(4\pi i n/3), giving all three values for n=0,1,2n=0,1,2. These are distinct integral String data, rather than three vacua or a phase transition. The example lies outside the Crowley–Goette domain. It makes the role of the larger-domain formula tangible: the response can distinguish global background choices that share the same underlying manifold and fibre period.

The same relative formula also has a six-dimensional reading. The virtual complex Weyl coefficient 4U−84U-8, where the integers denote bundle multiplicities, has anomaly density c(λ+c)/3c(\lambda+c)/3. The gauge-dependent part of Lee–Tachikawa’s primitive half-doublet density is c(λ+c)/24c(\lambda+c)/24, and their response uses backgrounds with dH=cdH=c [10 §3.3]. The factor of eight relates these normalizations; the different equations dh=λ+cdh=\lambda+c and dH=cdH=c retain different background data. The resulting divisibility condition and the integral lattice factorization are explained within the established Green–Schwarz quantization framework of Monnier–Moore and Monnier–Moore–Park [11, 12], with no claim to construct a complete six-dimensional theory.

We first specify the normalized response and its background data in § 2, then follow the disk construction and the direct APS calculation to the relative integer. The comparison, calibration and change law explain what that integer measures. Later sections describe the reduced six-dimensional response, the published heterotic prescription and the separately conditional comparison with integral topological modular forms (TMF) and vertex operator superalgebras. The latter requires actual compatible integral lifts and real equivariant characters, and uses the selected filling with zero local counterterm. These additional inputs are stated in § 10; the response calculation does not construct them.

Neither that conditional application nor the heterotic identification is needed to prove the sphere-bundle evaluation.

The normalized scalar and its background data

The Green–Schwarz response studied here is a phase assigned to a ten-dimensional spin manifold with a specified differential String datum. This datum includes an integral trivialization of the spin characteristic class λ=p1/2\lambda= p_1/2, together with its refinement at the chosen connection. The response combines a local eleven-dimensional integral with a closed real fermion anomaly. The two contributions belong to the same normalization: specifying the anomaly polynomial fixes the local density, but does not by itself specify the real anomaly theory, its boundary line, or the reference against which that line is evaluated.

The connections and String trivialization are background inputs. The eleven- and twelve-dimensional manifolds used below are auxiliary geometric fillings for evaluating their response. Their introduction does not supply additional spacetime degrees of freedom. The issue is which global phase remains after a local anomaly density has been trivialized, so the integral and spectral definitions of that phase must be specified together.

For fermions with an anomaly, the partition function is naturally a vector in a one-dimensional anomaly line over the background. A trivialization of the anomaly supplies a unit vector in that line; its Hermitian pairing with the fermion vector is a number. This is the framework used by Tachikawa–Zhang [3 §4.2, equations (67)–(73)]. The scalar below extracts the Green–Schwarz factor by pairing with a spin reference having trivial gauge fields. In the general invertible-field-theory formulation, an anomaly assigns a graded line to a ten-dimensional background and a phase to a closed eleven-manifold [2 §2.3]. The line used here is the ordered graded tensor product of inverse-Pfaffian lines for the real fermion coefficient specified below. Invertibility refers to the anomaly sector; the fermion path integral can itself vanish in the presence of zero modes [2 §2.1]. The scalar is an anomaly response with specified background data, rather than a calculation of the magnitude or other sectors of a full partition function. The real coefficient, spectral convention and reference pairing fixed below determine its mathematical normalization [7 §2]. The physical identification with the published heterotic prescription is addressed separately in § 9.

Let NN be an eleven-dimensional filling of MM carrying the integral and differential anomaly-trivializing data, and let WrefW_{\mathrm{ref}} be a spin filling of the same boundary tangent geometry with trivial framed gauges. For this scalar the boundary gauge connections on MM are trivial in prescribed frames. Both fillings have product collars, and the boundary identifications include their Spin lifts. Their glued closed spin/gauge manifold is Z=N‾∪MWrefZ = \overline{N} \cup_M W_{\mathrm{ref}}, where the bar reverses orientation. With the conventions developed below, define

O(M)=exp⁡(2πiϕ(M)),ϕ(M)=−aN−μ∗(Z),aN=∫NH3Y8,(5)\mathcal{O}(M) = \exp(2\pi i\phi(M)), \qquad\phi(M) = -a_N - \mu_*(Z), \qquad a_N = \int_N H_3Y_8, \tag*{(5)}

with ϕ∈R/Z\phi\in\mathbb{R}/\mathbb{Z}. Here H3H_3 is the three-form part of the trivialization, Y8Y_8 is fixed by the full fermion coefficient, and μ∗\mu_* is the real reduced eta value at the specified connection. The spin reference need not carry an extension of the String datum. Keeping these two sorts of filling distinct will permit the nonzero response computed in § 4.

The common cover and the full real coefficient

Put λ=p1/2\lambda= p_{1}/2, and write TT for the tangent bundle. The gauge backgrounds are actual principal Spin⁡(16)×Spin⁡(16)\operatorname{Spin}(16) \times\operatorname{Spin}(16) bundles, with associated rank-sixteen real vector bundles E,E′E,E'. This is the common cover on which the comparison is made. A background for a quotient gauge group that does not lift to this cover is outside the present domain. In particular, the cover matters when using the bordism calculation of Basile–Debray–Delgado–Montero [13 §3.3.3, Theorem 3.78].

The virtual real coefficient and its characteristic forms are

ΔR=E⊗RE′−SR+(E)−SR+(E′),(6)\Delta_{\mathbb{R}} = E \otimes_{\mathbb{R}} E' - S_{\mathbb{R}}^{+}(E) - S_{\mathbb{R}}^{+}(E'), \tag*{(6)}
n=λ(E),n′=λ(E′),n = \lambda(E), \qquad n' = \lambda(E'),
X4=λ(T)−n−n′,Y8=n2+nn′+n′2−p2(E)−p2(E′)6.(7)X_{4} = \lambda(T) - n - n', \qquad Y_{8} = \frac{n^{2} + nn' + n'^{2} - p_{2}(E) - p_{2}(E')}{6}. \tag*{(7)}

Here SR+S_{\mathbb{R}}^{+} denotes the real positive half-spin representation, and ΔC\Delta_{\mathbb{C}} is the complexification of the virtual real coefficient. The two subtracted half-spin bundles each have real rank 128; subtraction denotes their inverse anomaly factors. Characteristic classes below also denote their Chern–Weil forms when connections are fixed. The real structure fixes the half factor in the anomaly polynomial. With A^(T)\widehat{A}(T) the Dirac index characteristic form and [⋅]12[\cdot]_{12} the form-degree twelve component, it reads

I12=−12[A^(T)ch⁡(ΔC)]12=X4Y8.(8)I_{12} = -\frac{1}{2}\left[\widehat{A}(T)\operatorname{ch}(\Delta_{\mathbb{C}})\right]_{12} = X_{4}Y_{8}. \tag*{(8)}

We normalize the curvature FF by absorbing 1/(2π)1/(2\pi) into it, so that ch⁡=tr⁡eiF\operatorname{ch} = \operatorname{tr} e^{iF} and c2=tr⁡F2/2c_{2} = \operatorname{tr} F^{2}/2 for the symplectic complex representation [3 §2, §4.3].

References

  1. [1]M.B. Green and J.H. Schwarz, Anomaly cancellations in supersymmetric D = 10 gauge theory and superstring theory, Phys. Lett. B 149 (1984) 117.
  2. [2]D.S. Freed, “Anomalies and invertible field theories.” arXiv:1404.7224v2, 2014.arxiv.org/abs/1404.7224
  3. [3]Y. Tachikawa and H.Y. Zhang, On a Z₃-valued discrete topological term in 10d heterotic string theories, SciPost Phys. 17 (2024) 077 published 11 September 2024. All citations to this work refer to the published SciPost version.DOI
  4. [4]Y. Tachikawa and K. Yonekura, “On invariants of two-dimensional minimally supersymmetric field theories.” arXiv:2508.04916, 2025.arxiv.org/abs/2508.04916
  5. [5]J. Kaidi, Y. Tachikawa and K. Yonekura, “On non-supersymmetric heterotic branes.” arXiv:2411.04344, 2024.arxiv.org/abs/2411.04344
  6. [6]H. Wada and K. Yonekura, “The tertiary invariant and discrete Green–Schwarz couplings in heterotic string theory.” To appear, 2026.arxiv.org/abs/2611.11731
  7. [7]Y. Li and H.-Y. Sun, “Green–Schwarz scalars on quaternionic sphere bundles.” Unpublished manuscript, 2026.arxiv.org/abs/2611.12810
  8. [8]M.F. Atiyah, V.K. Patodi and I.M. Singer, Spectral asymmetry and Riemannian geometry. I, Math. Proc. Cambridge Philos. Soc. 77 (1975) 43.
  9. [9]D. Crowley and S. Goette, Kreck–Stolz invariants for quaternionic line bundles, Trans. Amer. Math. Soc. 365 (2013) 3193 arXiv:1012.5237v2. Version of 28 October 2011; numbering cited here is from this version.arxiv.org/abs/1012.5237
  10. [10]Y. Lee and Y. Tachikawa, “Some comments on 6d global gauge anomalies.” arXiv:2012.11622v2. Version of 1 February 2021, 2021.arxiv.org/abs/2012.11622
  11. [11]S. Monnier and G.W. Moore, “Remarks on the Green–Schwarz terms of six-dimensional supergravity theories.” arXiv:1808.01334v2. Version of 25 November 2018, 2018.DOI
  12. [12]S. Monnier, G.W. Moore and D.S. Park, “Quantization of anomaly coefficients in 6D 𝒩 = (1, 0) supergravity.” arXiv:1711.04777v3. Version of 26 January 2018, 2018.DOI
  13. [13]I. Basile, A. Debray, M. Delgado and M. Montero, Global anomalies & bordism of non-supersymmetric strings, J. High Energy Phys. 02 (2024) 092 Theorem numbering here is from arXiv:2310.06895v3.arxiv.org/abs/2310.06895
  14. [15]C. Redden, “Trivializations of differential cocycles.” arXiv:1201.2919v2, 2 February, 2012.arxiv.org/abs/1201.2919
  15. [16]D.S. Freed and G.W. Moore, “Setting the quantum integrand of M-theory.” arXiv:hep-th/0409135v3. Numbering cited here is from version 3. Published in Commun. Math. Phys. 263 (2006) 89–132, doi:10.1007/s00220-005-1482-7, 22 December, 2005.DOI
  16. [17]E. Witten and K. Yonekura, “Anomaly inflow and the η-invariant.” arXiv:1909.08775v3, 14 September, 2020.arxiv.org/abs/1909.08775
  17. [18]J. Davigli and N. Lohitsiri, “Omega vs. pi, and 6d anomaly cancellation.” arXiv:2012.11693v2. Version of 11 November 2021, 2021.
  18. [19]N. Seiberg and W. Taylor, Charge lattices and consistency of 6D supergravity, J. High Energy Phys. 06 (2011) 001 [1103.0019v3].DOI
  19. [20]X. Dai and D.S. Freed, Erratum: η-invariants and determinant lines, J. Math. Phys. 42 (2001) 2343.
  20. [21]Y. Tachikawa and M. Yamashita, “Anderson duality of topological modular forms and its differential-geometric manifestations.” arXiv:2305.06196v3. All numbering cited here is from version 3, 30 June, 2025.arxiv.org/abs/2305.06196
  21. [22]S. Stolz and P. Teichner, “Supersymmetric field theories and generalized cohomology.” arXiv:1108.0189v1. Version of 31 July 2011, 2011.arxiv.org/abs/1108.0189
  22. [23]D. Gaiotto and T. Johnson-Freyd, “Mock modularity and a secondary elliptic genus.” arXiv:1904.05788v1. Version of 11 April 2019, 2019.arxiv.org/abs/1904.05788
  23. [24]S.-S. Chern and J. Simons, Characteristic forms and geometric invariants, Ann. of Math. (2) 99 (1974) 48.
  24. [25]X. Dai and D.S. Freed, “η-invariants and determinant lines.” arXiv:hep-th/9405012v4. Numbering cited here is from version 4. Published in J. Math. Phys. 35 (1994) 5155–5194, doi:10.1063/1.530747, 8 June, 1994.DOI
  25. [26]M.F. Atiyah and R. Bott, A Lefschetz fixed point formula for elliptic complexes: II. Applications, Ann. of Math. (2) 88 (1968) 451.
  26. [27]M.F. Atiyah and I.M. Singer, The index of elliptic operators: III, Ann. of Math. (2) 87 (1968) 546.
  27. [28]D.S. Freed and J. Lott, An index theorem in differential K-theory, Geom. Topol. 14 (2010) 903.DOI
  28. [29]X. Dai, Adiabatic limits, nonmultiplicativity of signature, and Leray spectral sequence, J. Amer. Math. Soc. 4 (1991) 265.DOI
  29. [30]R.B. Melrose and P. Piazza, Families of Dirac operators, boundaries and the b-calculus, J. Differential Geom. 46 (1997) 99 The proposition numbering used here is from the authors’ preprint, author-hosted PostScript.DOI
  30. [31]C.C. Carvalho, Cobordism invariance of the family index, Math. Nachr. 285 (2012) 808 arXiv:0805.4412v2.

Paper details

vlnc:W_9kwc6rsb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

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

  • Physics and Astronomy

GPT-5.6, GPT-6, GPT-6.1

View statement
Looking for related papers…
Contents