Comparison of the mixed-fermion-action Effects using different fermion and gauge actions with 2+1 and 2+1+1 flavors
CLQCD Collaboration, Zun-Xian Zhang, Mengchu Cai, Bolun Hu, Xiangyu Jiang, Yi-Bo Yang, Dian-Jun Zhao · 2025
Read Comparison of the mixed-fermion-action Effects using different fermion and gauge actions with 2+1 and 2+1+1 flavors by CLQCD Collaboration, Zun-Xian Zhang, Mengchu Cai, Bolun Hu, Xiangyu Jiang, Yi-Bo Yang, Dian-Jun Zhao (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-2512-19265.txt.
Introduction
Zun-Xian Zhang, Mengchu Cai, Bolun Hu, and Xiangyu Jiang measure a key constant of mixed-action lattice QCD, where the quarks used to generate gauge fields differ from those used to measure. The constant, Delta mix, captures the leading mismatch effect in the low-energy theory. Knowing its size helps correct mixed-action results.
The team computed it on ensembles with up, down, strange, and charm quarks, using HISQ fermions and an improved gauge action at four lattice spacings. Comparing with other actions and a three-flavor case, they find the fermion action matters most. The gauge action has a smaller, real effect, and charm loops are negligible for now.
Catalog record
| CREATOR | CLQCD Collaboration, Zun-Xian Zhang, Mengchu Cai, Bolun Hu, Xiangyu Jiang, Yi-Bo Yang, Dian-Jun Zhao |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-19265
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Comparison of the mixed-fermion-action Effects using different fermion and gauge actions with 2+1 and 2+1+1 flavors CLQCD Collaboration, Zun-Xian Zhang, Mengchu Cai, Bolun Hu, Xiangyu Jiang, Yi-Bo Yang, Dian-Jun Zhao Abstract The leading-order low-energy constant Δ mix \Delta_{\rm mix} in mixed-action chiral perturbation theory is calculated using 2 + 1 + 1 2+1+1 -flavor gauge ensembles with HISQ fermions and a tadpole-improved Symanzik gauge action at four lattice spacings a ∈ [ 0.048 , 0.111 ] a\in[0.048,0.111] fm. By comparing our results to those from different actions and a 2 + 1 2+1 -flavor case, we find that the fermion action has the major impact, but the gauge action and charm quark loops also contribute significantly. I Introduction Lattice QCD provides the primary non-perturbative framework for ab initio calculations of strong interaction physics. A central challenge in the field is the choice of fermion discretization, which must balance computational cost against the preservation of fundamental symmetries, most notably chiral symmetry. In this context, the mixed-action approach has emerged as a powerful pragmatic strategy. By generating gauge ensembles with a computationally efficient “sea” fermion action and calculating observables using a more theoretically clean but expensive “valence” fermion action, one can leverage the statistical advantages of large ensembles while maintaining good control over systematic errors. Mixed-action lattice QCD has been employed in calculations of glue helicity 36 , nucleon charges (e.g., 21 ; 31 ; 35 ), hadron vacuum polarization 14 , and other observables. The success of this approach hinges on a quantitative understanding of the additional discretization artifacts introduced by the action mismatch, known as mixed-action effects . A key calculable quantity for quantifying these effects is the mass m_{\pi,{\rm vs}} of the mixing pion with one valence quark and one sea anti-quark, which leads to the leading-order low-energy constant in the mixed-action partially-quenched chiral perturbation theory (MAPQ \chi PT) 7 ; 6 ; 22 : \Delta_{\text{mix}}\equiv m_{\pi,\text{vs}}^{2}-\frac{m_{\pi,\text{vv}}^{2}+m_{\pi,\text{ss}}^{2}}{2}|_{m_{\pi,\text{vv}}=m_{\pi,\text{ss}}}, (1) where m_{\pi,\text{vv}} ( m_{\pi,\text{ss}} ) is the pion mass with the valence (sea) quark and anti-quark. Recent systematic study 37 has observed that when the sea fermion action preserves chiral symmetry, \Delta_{\text{mix}} scales as \mathcal{O}(a^{4}) —a significant improvement over the expected \mathcal{O}(a^{2}) behavior. This favorable scaling has been observed over a wide range of lattice spacings, a\in[0.04,0.19] fm, for chirally symmetric actions such as Domain wall 18 ; 16 ; 17 and Highly Improved Staggered Quark (HISQ) fermions 28 . In contrast, when a clover sea fermion action with explicit chiral symmetry breaking is used, \Delta_{\text{mix}} remains substantially larger, independent of the valence action. However, the comparison in Ref. 37 presents potential loopholes. The clover fermion ensembles used the tadpole-improved tree-level Symanzik (TITLS) gauge action, while the HISQ and Domain wall ensembles used different gauge actions (one-loop improved Symanzik and Iwasaki, respectively). Consequently, the possibility that the gauge action itself significantly impacts \Delta_{\text{mix}} cannot be excluded. Furthermore, the number of dynamical fermion flavors differs between the clover ( N_{f}=2+1 ) and HISQ ( N_{f}=2+1+1 ) ensembles, introducing another confounding factor. In this work, we perform a controlled study to isolate these effects. We generate new HISQ fermion ensembles using the same tadpole-improved Symanzik gauge action as the clover studies, with N_{f}=2+1+1 flavors at four lattice spacings, and three N_{f}=2+1 flavor ensembles at a\sim 0.11 fm with different gauge actions. By calculating \Delta_{\text{mix}} on these new ensembles, we enable a fair comparison to precisely determine the impact of the gauge action and the dependence on the number of dynamical fermion flavors. II Setup and basic tests of the generated gauge ensembles The fermion action used for the gauge ensembles generated in this work is the HISQ action 28 , which effectively mitigates taste symmetry breaking and discretization errors while retaining acceptable chiral symmetry. It is formulated in the following form: S_{\text{HISQ}}=\sum_{x}\bar{\psi}(x)\left(\gamma\cdot D^{\text{HISQ}}+m\right)\psi(x), (2) with the {\cal O}(a^{2}) improved covariant derivative operator: D^{\text{HISQ}}_{\mu}\equiv\Delta_{\mu}(W)-\frac{a^{2}}{6}(1+\epsilon)\Delta_{\mu}^{3}(X), (3) where the smeared link field W_{\mu}(x) is constructed via a two-stage Fat7 smearing process with intermediate SU(3) projection, and X_{\mu}(x) denotes links smeared with a single application of the ASQTAD smearing operator. The parameter provides improved dispersion relation for heavy quarks with bare mass m by canceling tree-level \mathcal{O}((am)^{4}) errors. We consider a general gauge action of the form: \displaystyle S_{g}=\hat{\beta}\bigg[ \displaystyle\sum_{P}\left(1-\frac{1}{3}\mathrm{Re}\mathrm{Tr}(P)\right) (4) \displaystyle+C_{R}\sum_{R}\left(1-\frac{1}{3}\mathrm{Re}\mathrm{Tr}(R)\right)\bigg], (5) where the sums run over all 1\times 1 plaquettes ( P ) and 2\times 1 rectangles ( R ). Specific gauge actions are characterized by the value of the parameter C_{R} : • C_{R}=0 for the Wilson action (W); • C_{R}=-1/20 for the tree-level Symanzik action ( S^{(0)} ); • C_{R}=-1/(20u_{0}^{2}) for the tadpole-improved Symanzik action, with u_{0}=\left(\langle\mathrm{Re}\mathrm{Tr}(P)\rangle/3\right)^{1/4} . This was employed in previous CLQCD ensembles with clover fermions 30 ; 27 ; • C_{R}=-0.331/(1+8\cdot 0.331)=-0.0907346 for the Iwasaki gauge action. Table 1: Input parameters \hat{\beta} , u_{0} , \tilde{L}^{3}\times\tilde{T} , and \tilde{m}^{\rm b}_{l,s,c} of the 2+1+1 flavor gauge ensembles with HISQ fermion and S^{\rm tad} gauge actions. Ensemble \hat{\beta} \tilde{L}^{3}\times\tilde{T} u_{0} \tilde{m}^{\rm b}_{l} \tilde{m}^{\rm b}_{s} \tilde{m}^{\rm b}_{c} c24P31s 7.29 24^{3}\times 48 0.879440(3) 0.00944 0.055 0.5555 c24P31 7.29 24^{3}\times 48 0.879452(3) 0.00944 0.04721 0.5555 c32P31 7.29 32^{3}\times 48 0.879451(2) 0.00944 0.04721 0.5555 c24P22 7.29 24^{3}\times 48 0.879469(3) 0.00472 0.04721 0.5555 c32P22 7.29 32^{3}\times 48 0.879468(2) 0.00472 0.04721 0.5555 c48P13 7.29 48^{3}\times 48 0.879472(1) 0.00174 0.04721 0.5555 e32P31 7.54 32^{3}\times 64 0.886360(2) 0.007434 0.03715 0.4371 g32P31 7.75 32^{3}\times 64 0.891434(1) 0.00579 0.02895 0.34 g48P31 7.75 48^{3}\times 64 0.891432(1) 0.00579 0.02895 0.34 h48P31 8.20 48^{3}\times 96 0.900600(1) 0.003526 0.01763 0.207 In this work, we generated a set of 2+1+1-flavor gauge ensembles using the aforementioned fermion and gauge actions. The ensemble parameters are detailed in Table 1 (with the symbol \tilde{X}\equiv Xa^{n} for any quantity X with mass dimension n ), including the gauge coupling \hat{\beta} , the tadpole factor u_{0} , lattice volume \tilde{L}^{3}\times\tilde{T} , and the bare dimensionless quark masses \tilde{m}^{\rm b}_{l,s,c} . For all ensembles at \hat{\beta}=7.29 and 7.54 (except c24P31s to be discussed later), we fixed the charm-to-strange mass ratio to \tilde{m}^{\rm b}_{c}/\tilde{m}^{\rm b}_{s}=11.766 , corresponding to the 2+1+1 flavor FLAG average 4 ; 19 ; 20 ; 12 ; 1 . For the ensembles at \beta=7.75 and 8.20, we used \tilde{m}^{\rm b}_{c}/\tilde{m}^{\rm b}_{s}=11.74 , a value consistent with the FLAG average of 11.766(30) within its uncertainty. The strange-to-light mass ratio, \tilde{m}^{\rm b}_{s}/\tilde{m}^{\rm b}_{l} , was set to 27.18 11 for the physical point ensemble (c48P13), to 10 for c24P22 and c32P22, and to 5 for the remaining ensembles (excluding c24P31s). The bare strange quark mass \tilde{m}^{\rm b}_{s} was then tuned to ensure the resulting m_{\eta_{s}} mass was close to the value of 689.63(18) MeV 15 which corresponds to the physical strange quark mass, and the tadpole factor u_{0} was tuned to ensure consistency within 0.001% between its input value and the value from thermalized configurations. Finally, to investigate the impact of the strange sea quark mass, we generated the c24P31s ensemble using the same \hat{\beta} and \tilde{m}^{\rm b}_{l,c} as the c24P31 ensemble but with a \tilde{m}^{\rm b}_{s} enlarged by 17%. The values of u_{0} in Table 1 indicate a clear hierarchy in the parameter dependencies. The gauge coupling \hat{\beta} is the dominant factor. Contributions from the light and strange quark masses, though much weaker ( {\cal O}(10^{-5}) ), are still relevant at