betaFeedback

WorksPhysics and Astronomy

Nucleon sigma terms in lattice QCD at physical pion mass and in the continuum

Abstract

We determine the light, strange and charm σσ-terms of the nucleon in lattice QCD using four Nf=2+1+1N_f=2+1+1 twisted-mass gauge ensembles simulated at physical pion masses and with approximately the same spatial volumes. The four lattice spacings ranging from 0.0800.080 fm to 0.0490.049 fm, with an action that has O(a) improvement, enable extrapolating the results to the continuum limit. Both valence and sea quark contributions are analysed accounting for excited-state contamination from low-lying multihadron states. We perform global fits to all ensembles, that yield σπN=53.0(7.4)(1.6)σ_{πN}=53.0(7.4)(1.6) MeV, σs=58.2(17.0)(2.1)σ_s=58.2(17.0)(2.1) MeV and σc=112(28)(25)σ_c=112(28)(25) MeV, with statistical and systematic uncertainties, respectively.

Introduction

Introduction.—The sigma terms of the nucleon, σq=mq⟨N∣qˉq∣N⟩=mq ∂mN/∂mq\sigma_q = m_q\langle N|\bar{q}q|N\rangle= m_q\,\partial m_N/\partial m_q, measure how sensitive the nucleon mass is to the quark mass mqm_q of each quark flavor qq. They enter the nucleon mass decomposition through the trace of the energy-momentum tensor, together with the gluonic trace anomaly [1, 2]. At leading order in the heavy-quark expansion and in αs\alpha_s, the contribution of a heavy quark Q=c,b,tQ = c,b,t is σQ=227(mN−σπN−σs)\sigma_Q = \frac{2}{27}(m_N-\sigma_{\pi N}-\sigma_s), which makes σc\sigma_c a test of heavy-quark decoupling [1, 3]. The phenomenological value of σπN\sigma_{\pi N} is extracted from isoscalar pion–nucleon scattering using the Cheng–Dashen low-energy theorem [4, 5] up to O(mπ4)O(m_\pi^4). The σs\sigma_s determines the strangeness content of the nucleon, y=2mℓσs/(msσπN)y = 2m_\ell\sigma_s/(m_s\sigma_{\pi N}), where mℓm_\ell is the average light-quark mass.

Nucleon sigma terms also connect the scalar quark content of the nucleon to experimental searches for physics beyond the Standard Model. The corresponding mass fractions fTq=σq/mNf_{Tq} = \sigma_q/m_N enter both the Higgs–nucleon coupling and the amplitude for scalar-mediated dark-matter scattering [6, 7]. In particular, they are essential to translate nuclear-recoil limits from LZ, XENONnT, and PandaX-4T into constraints on dark-matter couplings to quarks [8, 9, 10]. The same matrix elements enter predictions for Higgs-mediated μ→e\mu\to e conversion in the nuclei [11].

Lattice QCD determinations of the light-quark sigma term σπN\sigma_{\pi N} disagree among themselves and with phenomenology. The FLAG average of Nf=2+1N_f = 2+1 results, 42.2(2.4) MeV [12], lies 3.3σ\sigma below the phenomenological value of 56.0(3.5) MeV [5, 13, 14], while the Nf=2+1+1N_f = 2+1+1 average, 60.9(6.5) MeV, agrees with it but rests on only two calculations [15, 16]. Previous lattice calculations relied largely on a single lattice spacing [17] or on joint extrapolations in the lattice spacing and the pion mass [16, 18, 19, 20]. An earlier analysis based on ensembles simulated by the Extended Twisted Mass Collaboration (ETMC) at three lattice spacings at physical quark masses agreed with the smaller value [21]. A well-known difficulty in these calculations is the contamination of the nucleon matrix elements by multiparticle excited states involving the nucleon and pion states, which, if not treated properly, can bias the extracted σπN\sigma_{\pi N} [16, 20, 22].

In this work, we determine σπN\sigma_{\pi N}, σs\sigma_s and σc\sigma_c using an updated analysis at four lattice spacings with physical quark masses. We take multiparticle states into account when extracting the nucleon matrix elements, and we perform a global fit across the four ensembles in addition to fitting the data from individual ensembles. This enables a more stringent continuum extrapolation, while testing the impact of multiparticle contributions without imposing constraints.

The sigma terms can be determined directly from the forward matrix elements of the scalar quark densities,

σq=mq⟨N(p⃗,s)∣qˉq∣N(p⃗,s)⟩=mq(∂mN/∂mq),(1)\sigma_q = m_q \langle N(\vec{p},s)|\bar{q}q|N(\vec{p},s)\rangle= m_q(\partial m_N/\partial m_q), \tag*{(1)}
σπN=mℓ⟨N(p⃗,s)∣uˉu+dˉd∣N(p⃗,s)⟩,\sigma_{\pi N} = m_\ell\langle N(\vec{p},s)|\bar{u}u+\bar{d}d|N(\vec{p},s)\rangle,

with mℓ=(mu+md)/2m_\ell= (m_u+m_d)/2, q=s,cq = s,c, and the states normalized as ⟨N(p⃗ ′,s′)∣N(p⃗,s)⟩=(2π)3δ3(p⃗ ′−p⃗)δss′\langle N(\vec{p}^{\,\prime},s^{\prime})|N(\vec{p},s)\rangle= (2\pi)^3\delta^3(\vec{p}^{\,\prime}-\vec{p})\delta_{s s^{\prime}}. They are renormalization-group invariant and Lorentz scalar observables.

The main uncertainty when using exclusively gauge ensembles at physical pion mass comes from the analysis of the contributions by multiparticle states, which carry the quantum numbers of the nucleon. The scalar current excites certain nucleon and pion states strongly while they are barely visible in the determination of the nucleon mass. Thus, analyses that constrain the excited-state energies from the nucleon mass determination underestimate σπN\sigma_{\pi N} [16, 22, 23, 24, 25]. We allow the effective excited-states gap in the three-point function to differ from that of the nucleon two-point function, and determine it across the four ensembles without priors.

Lattice setup

Lattice setup.—We employ four ETMC gauge ensembles with a mass degenerate up and down quarks, a strange and a charm quark (Nf=2+1+1N_f = 2+1+1) with their mass fixed to approximately their physical values [27, 28, 29]. This lattice discretization scheme at maximal twist gives automatic O(a)O(a) improvement of physical observables [30, 31], so the leading cutoff effects are O(a2)O(a^2). The spatial extent of the ensembles is L≃5.1L \simeq5.1–5.5 fm, giving mπL≃3.6m_\pi L \simeq3.6–3.9. The parameters of the four ensembles are listed in Table I. Details of the fits are given in a forthcoming paper on the nucleon scalar charges [32]. We use Osterwalder-Seiler [33] strange and charm quarks with their bare masses tuned from the meson sector [26, 28] at the isospin-symmetric point defined by mπ=135.0m_\pi= 135.0 MeV, mK=494.6m_K = 494.6 MeV, mDs=1967m_{D_s} = 1967 MeV and Fπ=130.5F_\pi= 130.5 MeV.

Ensemble(L/a)3×T/a(L/a)^3 \times T/aaa [fm]LL [fm]mNm_NNcfgN_{\mathrm{cfg}}
cB211.072.64 (B64)643×12864^3 \times1280.07975.10940(5)749
cC211.060.80 (C80)803×16080^3 \times1600.06805.44945(3)400
cD211.054.96 (D96)963×19296^3 \times1920.05695.46945(4)493
cE211.044.112 (E112)1123×224112^3 \times2240.04885.47943(4)467

Table 1. Parameters of the ensembles analyzed. The lattice spacings aa are from [26] (see also Refs. [27, 28]). Their error is one on the last digit given. In the last column we give the number of configurations NcfgN_{\mathrm{cfg}} analyzed per ensemble.

Correlation functions.—To extract the nucleon mass, we compute two-point functions C2(ts)C_2(t_s) with the interpolator JN(t)=∑x⃗ϵabc[uaTCγ5db](x⃗,t)uc(x⃗,t)J_N(t) = \sum_{\vec{x}} \epsilon_{abc} [u^{aT} C\gamma_5 d^b](\vec{x},t) u^c(\vec{x},t) at zero momentum, with Gaussian-smeared quark fields [34, 35] and APE-smeared links [36], tuned to a nucleon smearing radius of about 0.5 fm [21]. The connected three-point functions C3(tins,ts)C_3(t_{\mathrm{ins}},t_s) are computed with the fixed-sink method, requiring a separate inversion for each value of the time separation tst_s between the time of creating and annihilating the nucleon states but for all values of the scalar current insertion times tinst_{\mathrm{ins}}. We use a number of source positions per configuration that grows with tst_s, from one at the smallest to O(100)O(100) at the largest, so that the statistical error is roughly uniform across all values of tst_s. The disconnected quark loops are computed using the one-end trick, hierarchical probing, color–spin dilution [21, 37].

We form the ratio

R(tins,ts)=C3conn(tins,ts)C2(ts)+C3disc(tins,ts)C2highstat.(ts)−⟨Sq⟩.(2)R(t_{\mathrm{ins}},t_s)=\frac{C_3^{\mathrm{conn}}(t_{\mathrm{ins}},t_s)}{C_2(t_s)}+\frac{C_3^{\mathrm{disc}}(t_{\mathrm{ins}},t_s)}{C_2^{\mathrm{highstat.}}(t_s)}-\langle S_q\rangle. \tag*{(2)}

The connected term exists only for q=u,dq=u,d, so σs\sigma_s and σc\sigma_c are purely disconnected. Each term is divided by the two-point function computed with the same statistics. The vacuum expectation value of the scalar loop SqS_q is subtracted. At large time separations R→⟨N∣qˉq∣N⟩R\to\langle N|\bar{q}q|N\rangle, and σq=μqR\sigma_q=\mu_q R, with μq\mu_q the bare quark mass [28]. For twisted-mass fermions at maximal twist the product of the bare quark mass and the bare scalar matrix element is renormalization group invariant, so no renormalization is needed [17, 21]. For the light quarks, the scalar current maps onto a flavor nonsinglet current in the twisted basis, so the singlet–nonsinglet mixing that affects σπN\sigma_{\pi N} with Wilson-clover fermions [16] does not arise.

Extraction of the nucleon matrix elements.—With the nucleon at rest and the positive-parity projector, the correlators have the spectral decompositions

C2(ts)=∑i∣Zi∣2e−Eits,(3)C_2(t_s)=\sum_i |Z_i|^2 e^{-E_i t_s}, \tag*{(3)}
C3(tins,ts)=∑i,jZiZj∗⟨i∣Sq∣j⟩e−Ei(ts−tins)−Ejtins,C_3(t_{\mathrm{ins}},t_s)=\sum_{i,j} Z_i Z_j^* \langle i|S_q|j\rangle e^{-E_i(t_s-t_{\mathrm{ins}})-E_j t_{\mathrm{ins}}},

with ZiZ_i the overlap of the interpolator with the state ii of energy EiE_i with i=0i=0 being the nucleon state and SqS_q the scalar current of flavour qq. Keeping the nucleon and one effective excited state, whose gap may differ between the three- and the two-point function, the ratio Rq,jR_{q,j} of Eq. (2) for flavour qq on ensemble jj, multiplied by the bare quark mass μq,j\mu_{q,j}, becomes

μq,jRq,j(tins,ts)=Cq,j(tins,ts)1+c1,je−ΔE2pt,j(aj)ts,(4)\mu_{q,j}R_{q,j}(t_{\mathrm{ins}},t_s)=\frac{C_{q,j}(t_{\mathrm{ins}},t_s)}{1+c_{1,j}e^{-\Delta E_{2\mathrm{pt},j}(a_j)t_s}}, \tag*{(4)}
Cq,j(tins,ts)=σq,j(aj)+Bje−ΔE3pt,j(aj)ts+Aj(e−ΔE3pt,j(aj)tins+e−ΔE3pt,j(aj)(ts−tins)),C_{q,j}(t_{\mathrm{ins}},t_s)=\sigma_{q,j}(a_j)+B_j e^{-\Delta E_{3\mathrm{pt},j}(a_j)t_s} +A_j\left(e^{-\Delta E_{3\mathrm{pt},j}(a_j)t_{\mathrm{ins}}}+e^{-\Delta E_{3\mathrm{pt},j}(a_j)(t_s-t_{\mathrm{ins}})}\right),

where σq,j(aj)\sigma_{q,j}(a_j) is the sigma term of flavour qq at the lattice spacing aja_j for ensemble jj, the overlap ∣Z0∣2|Z_0|^2 and the factor e−E0tse^{-E_0t_s} of the nucleon cancel in the ratio. The energy gaps ΔE2pt,j(aj)\Delta E_{2\mathrm{pt},j}(a_j) and ΔE3pt,j(aj)\Delta E_{3\mathrm{pt},j}(a_j) are the two- and three-point gaps for ensemble jj. In the denominator, we truncate the spectral decomposition of C2,jC_{2,j} to the first excited state, so that c1,j=∣Z1,j∣2/∣Z0,j∣2c_{1,j}=|Z_{1,j}|^2/|Z_{0,j}|^2 is the ratio of the overlaps for the first excited state and the nucleon computed on that ensemble. We determine the mass of the nucleon mN,j(aj)m_{N,j}(a_j) and the energy gap ΔE2pt,j(aj)\Delta E_{2\mathrm{pt},j}(a_j) by fitting the effective mass

meff,j(t)=E0,j(aj)+log⁡(1+c1,je−ΔE2pt,j(aj)t1+c1,je−ΔE2pt,j(aj)(t+1))(5)m_{\mathrm{eff},j}(t)=E_{0,j}(a_j)+\log\left(\frac{1+c_{1,j}e^{-\Delta E_{2\mathrm{pt},j}(a_j)t}}{1+c_{1,j}e^{-\Delta E_{2\mathrm{pt},j}(a_j)(t+1)}}\right) \tag*{(5)}

with tt and the energies in lattice units, and then give them as input in Eq. (4) as correlated observables. In the three-point function the overlaps are weighted by ⟨i∣Sq∣0⟩\langle i|S_q|0\rangle (Eq. (3)), so that another state can dominate, which we expect to be NπN\pi at rest, as demonstrated by our latest analysis of the B64 ensemble [38] and supported by other studies [22, 25]. We therefore do not set ΔE3pt,j(aj)=ΔE2pt,j(aj)\Delta E_{3\mathrm{pt},j}(a_j)=\Delta E_{2\mathrm{pt},j}(a_j) and fit it together with the other parameters in the numerator of the ratio. A fit window (t_s^{\mathrm{low}},t_s^{\mathrm{cut}}_{\mathrm{ins}}) contains the separations ts≥tslowt_s\ge t_s^{\mathrm{low}} and, at each tst_s, the insertion times tinscut≤tins≤ts−tinscutt_{\mathrm{ins}}^{\mathrm{cut}}\le t_{\mathrm{ins}}\le t_s-t_{\mathrm{ins}}^{\mathrm{cut}}. From Eq. (4), it is evident that there are two distinct methods to extract σq\sigma_q in the continuum, one where the extraction is performed per ensemble jj and then the continuum limit is taken, and one where a global fit across all ensembles is performed with a continuum ansätze for σq(a)\sigma_q(a) and ΔE3pt(a)\Delta E_{3\mathrm{pt}}(a).

Continuum limit.—We take two approaches:

i) Two-step procedure: We first fit the ratio of Eq. (4) on each ensemble jj separately, at a tslowt_s^{\mathrm{low}} chosen on that ensemble where σj\sigma_j is stable and for two to four values of tinscutt_{\mathrm{ins}}^{\mathrm{cut}}. The per-ensemble value σj\sigma_j is the mean over these windows, with the root mean square of the fit errors as statistical error and the standard deviation over the windows as systematic error, as for the global fit below. Since cutoff effects start at O(a2)O(a^2), we take the continuum limit with a term linear in a2a^2 as σ(a)=σ+ba2\sigma(a)=\sigma+ba^2, by a weighted fit to the σj\sigma_j of the four independent ensembles. We perform the continuum extrapolation using the statistical uncertainties of the ensemble results σj\sigma_j, propagating them to the continuum value. To estimate the systematic uncertainty from the fit windows, we repeat the continuum extrapolation for every combination of the windows on the four ensembles and take the spread of the continuum values. The statistical and systematic uncertainties are then added in quadrature. The uncertainty of the lattice spacing is propagated and changes the results by less than 0.1 MeV.

ii) Global fit.—Instead of fitting each ensemble separately and then taking the continuum limit, we fit them simultaneously allowing for a linear in a2a^2 dependence in the fit parameters that describe physical quantities, namely the sigma term σ(a)\sigma(a) and the energy gap ΔE3pt(a)\Delta E_{3\mathrm{pt}}(a). In Eq. (4) we include this a2a^2 dependence explicitly by setting σj=σ(aj)\sigma_j=\sigma(a_j) and ΔE3pt,j=ΔE3pt(aj)\Delta E_{3\mathrm{pt},j}=\Delta E_{3\mathrm{pt}}(a_j), where jj labels the ensemble, σ(a)=σ+ba2\sigma(a)=\sigma+ba^2 and ΔE3pt(a)=ΔE3pt+bEa2\Delta E_{3\mathrm{pt}}(a)=\Delta E_{3\mathrm{pt}}+b_Ea^2. The energy gap ΔE2pt(a)\Delta E_{2\mathrm{pt}}(a) and c1,jc_{1,j} are extracted from a correlated two-state fit to the effective masses of all four ensembles for 0.6≤t≤2.20.6\leq t\leq2.2 fm, with E0(a)E_0(a) and E1(a)E_1(a) linear in a2a^2, which gives mN=947.1(4.9)m_N=947.1(4.9) MeV in the continuum and ΔE2pt=540\Delta E_{2\mathrm{pt}}=540–640640 MeV on the four ensembles; they are used as inputs when fitting the ratio. σ\sigma, ΔE3pt\Delta E_{3\mathrm{pt}}, bb and bEb_E are the only parameters common to all ensembles when fitting the ratio. The amplitudes AjA_j and BjB_j are free on each ensemble, no priors are used, and the covariance is block diagonal because the ensembles are independent. The fit thus extrapolates σ(a)\sigma(a) to the continuum in one step. The global fit uses the same RR windows r=(tslow,tinscut)r=(t_s^{\mathrm{low}},t_{\mathrm{ins}}^{\mathrm{cut}}) on every ensemble and gives the continuum value σr\sigma_r directly in each window. We quote the mean over the windows, σˉ=1R∑rσr\bar{\sigma}=\frac{1}{R}\sum_r\sigma_r, with statistical error (1R∑rδσr2)1/2\left(\frac{1}{R}\sum_r\delta\sigma_r^2\right)^{1/2}, an upper bound since the windows are correlated, and systematic error (1R∑r(σr−σˉ)2)1/2\left(\frac{1}{R}\sum_r(\sigma_r-\bar{\sigma})^2\right)^{1/2}.

Fits for σπN\sigma_{\pi N}.—We fit the light-quark ratio of Eq. (2), with the connected and disconnected contributions combined, by minimizing the fully correlated χ2\chi^2, and compute all statistical errors with the Γ\Gamma-method [39] in pyerrors [40]. In both routes the fits start at tslow≈0.8t_s^{\mathrm{low}}\approx0.8 fm, below which two states do not describe the data, and use separations up to ts≈1.3t_s\approx1.3 fm on every ensemble. The fit windows variations in tslowt_s^{\mathrm{low}} and tinscutt_{\mathrm{ins}}^{\mathrm{cut}} are given in Ref. [32]. The fitted gap, ΔE3pt≈360\Delta E_{3\mathrm{pt}}\approx360–450450 MeV on the four ensembles, lies close to the lowest noninteracting levels in our volumes, the NππN\pi\pi threshold at 2mπ≈2802m_\pi\approx280 MeV and N(pˉ)π(−pˉ)N(\bar{p})\pi(-\bar{p}) at 290290–310310 MeV above mNm_N, as also found in Refs. [16, 24, 38], and well below ΔE2pt≈540\Delta E_{2\mathrm{pt}}\approx540–640640 MeV, which is close to the Roper resonance about 500 MeV above the nucleon [41].

Fits for σs\sigma_s and σc\sigma_c.—The purely disconnected strange and charm ratios are symmetrized under tins→ts−tinst_{\mathrm{ins}}\to t_s-t_{\mathrm{ins}}, so that only the independent half tins≤ts/2t_{\mathrm{ins}}\leq t_s/2 enters the fits. For σs\sigma_s we proceed as for σπN\sigma_{\pi N} but set Bj=0B_j=0, since the strange data do not resolve this term. The charm ratio shows no resolvable excited-state curvature, so we fit a constant to the central region of the largest separations; since the four lattice spacings do not resolve an a2a^2 dependence of σc\sigma_c, the linear fit giving 64(89) MeV, we take its continuum limit with a constant and add half the difference to the linear fit as a systematic error to the global result. Adding an excited-state term with a free gap, or extending the fits to ts=2.0t_s=2.0 fm, shifts σc\sigma_c by up to one statistical standard deviation.

Results and discussion.—Figure Figure 1 shows the global fits together with the ratios of the four ensembles, which they describe with χ2/dof≤1.2\chi^2/\mathrm{dof}\leq1.2 in every window. Even at the largest separations the light and strange ratios lie 30–45% below σ(aj)\sigma(a_j), which shows the size of the excited-state contribution removed by the fit, while the charm ratio is flat within its errors. We show the results from the two analysis approaches in Figure 2. We quote the global fit results, which extrapolates the ratios to the continuum in one step and takes the correlation between σ(a)\sigma(a) and the gap into account, and obtain

Figure 1. Results on the global fits to the ratios of Eq. (2), shown as μq,jRq,j\mu_{q,j}R_{q,j} times the denominator of Eq. (4), versus tins−ts/2t_{\mathrm{ins}}-t_s/2. From left to right we show the fits for the B64, C80, D96 and E112 for the ratio of σπN\sigma_{\pi N}: (top), for σs\sigma_s (middle) and σc\sigma_c (bottom). The curves and bands are the fits at tslow=0.8 fmt_s^{\mathrm{low}}=0.8\ \mathrm{fm} and tinscut=0.24 fmt_{\mathrm{ins}}^{\mathrm{cut}}=0.24\ \mathrm{fm} for σπN\sigma_{\pi N} and σs\sigma_s, and at 1.36≤ts≤1.54 fm1.36\leq t_s\leq1.54\ \mathrm{fm} and tinscut=0.50 fmt_{\mathrm{ins}}^{\mathrm{cut}}=0.50\ \mathrm{fm} for σc\sigma_c. The gray band and the number in each panel give σ(aj)=σ+baj2\sigma(a_j)=\sigma+b a_j^2 of the global fit at the lattice spacing of that ensemble, averaged over the fit windows; σc\sigma_c is fitted with a constant, the same on every ensemble.

Figure 2. Continuum extrapolation of σπN\sigma_{\pi N}, σs\sigma_s and σc\sigma_c (top to bottom). Circles: σ(aj)=σ+baj2\sigma(a_j)=\sigma+b a_j^2 of the global fit at the lattice spacing of each ensemble, averaged over the fit windows (constant for σc\sigma_c); squares: the fits of each ensemble separately. Solid (dashed) curves with bands and stars at a=0a=0: the continuum extrapolation of the global (per-ensemble) route, linear in a2a^2 except for σc\sigma_c, with statistical and systematic errors added in quadrature. Diamond: the Roy–Steiner value converted to isospin-symmetric QCD [5, 13].

σπN=53.0(7.4)(1.6) MeV=53.0(7.6) MeV,σs=58.2(17.0)(2.1) MeV=58.2(17.1) MeV,σc=112(28)(25) MeV=112(38) MeV.(6)\begin{aligned} \sigma_{\pi N} &= 53.0(7.4)(1.6)\ \mathrm{MeV}=53.0(7.6)\ \mathrm{MeV},\\ \sigma_s &= 58.2(17.0)(2.1)\ \mathrm{MeV}=58.2(17.1)\ \mathrm{MeV},\\ \sigma_c &= 112(28)(25)\ \mathrm{MeV}=112(38)\ \mathrm{MeV}. \tag*{(6)} \end{aligned}

where the first error is statistical and the second systematic, and the final uncertainty, given last, is the sum of the two in quadrature. For σc\sigma_c the systematic error includes the 24 MeV from the unresolved a2a^2 dependence. The a2a^2 slopes of the global fits are compatible with zero; for σπN\sigma_{\pi N}, b=−1.7(1.3) GeV fm−2b=-1.7(1.3)\ \mathrm{GeV\,fm}^{-2} amounts to −21(15)%-21(15)\% at the coarsest lattice spacing. Global fits with σ\sigma and ΔE3pt\Delta E_{3\mathrm{pt}} constant in a2a^2 give σπN=46.2(4.3)(0.6)\sigma_{\pi N}=46.2(4.3)(0.6) MeV and σs=46.2(6.6)(2.2)\sigma_s=46.2(6.6)(2.2) MeV, and the linear fit without B64 gives σπN=58.9(11.5)(2.5)\sigma_{\pi N}=58.9(11.5)(2.5) MeV, all within the errors of our results. The finite-volume shift of σπN\sigma_{\pi N} predicted by next-to-next-to-leading-order chiral perturbation theory [16, 42] with the low-energy constants of Ref. [43], −1.7-1.7 MeV for B64 and −1.3-1.3 MeV for the other ensembles, is well within our statistical error and would change the continuum value by about 1 MeV, so we do not correct for it. The energy gap, ΔE3pt=310(72)(21)\Delta E_{3\mathrm{pt}}=310(72)(21) MeV from the light quark sector, agrees within its error with the NπN\pi and NππN\pi\pi levels of the four ensembles and is well below ΔE2pt≈600\Delta E_{2\mathrm{pt}}\approx600 MeV [32]. The per-ensemble route which is the continuum extrapolation of the results of the four ensembles fitted individually, gives σπN=52.0(9.3)(2.5)\sigma_{\pi N}=52.0(9.3)(2.5) MeV, σs=55.1(17.0)(1.8)\sigma_s=55.1(17.0)(1.8) MeV and, with

σc\sigma_c constant, σc=102(25)(5) MeV\sigma_c=102(25)(5)\ \mathrm{MeV}, where the first error is statistical and the second the systematic error from the variation over the fit windows.

Since we use the isospin symmetric QCD, defined using the neutral pion masses, while within the Roy–Steiner analysis the isospin limit is defined with the charged-pion mass, we need to convert the latter value to our convention to compare. The value of σπN=59.1(3.5) MeV\sigma_{\pi N}=59.1(3.5)\ \mathrm{MeV} from Ref. [5] is reduced by 3.1(5) MeV3.1(5)\ \mathrm{MeV} in isospin symmetric convention, giving 56.0(3.5) MeV56.0(3.5)\ \mathrm{MeV} [13]. This is the value shown in Figure 2 and it is compatible with our value. The covariant chiral perturbation theory analysis of pion–nucleon scattering data of Ref. [14] gives σπN=59(7) MeV\sigma_{\pi N}=59(7)\ \mathrm{MeV}, which the same shift reduces to 56(7) MeV56(7)\ \mathrm{MeV}, also compatible with our value.

In Figure 3, we compare our values with other recent lattice QCD determinations. We distinguish three classes of determinations: those that compute the nucleon matrix elements of the scalar operator, as in this work, those that use the Feynman–Hellmann approach, and those that use the hybrid method of MILC for σs\sigma_s [44, 45]. Our values for σπN\sigma_{\pi N} and σs\sigma_s are in agreement with most lattice determinations except the one from JLQCD which obtained a smaller value for σπN\sigma_{\pi N}. Unlike our earlier analysis at three lattice spacings [21], which gave a smaller σπN\sigma_{\pi N}, we keep ΔE3pt\Delta E_{3\mathrm{pt}} a free parameter and perform a global fit.

Figure 3. Our continuum results (stars and bands) for σπN\sigma_{\pi N} (left), σs\sigma_s (middle) and σc\sigma_c (right) compared with lattice determinations from direct matrix elements (squares) [16, 17, 18, 20, 21, 46, 48], the Feynman–Hellmann method (triangles) [15, 19, 47, 49, 50] and the hybrid method (diamonds) [44, 45]. We include the FLAG averages [12] (gray boxes) and the phenomenological determinations from pion–nucleon scattering [5, 14], converted to isospin-symmetric QCD [13]. Filled (open) symbols use Nf=2+1+1N_f=2+1+1 (Nf=2+1N_f=2+1) ensembles. Statistical and systematic errors are added in quadrature.

For σc\sigma_c there are three other lattice values [17, 21, 46], all in agreement with the value extracted from this analysis. The heavy-quark expansion [1] gives at leading order σcHQE=227(mN−σπN−σs)≃61 MeV\sigma_c^{\mathrm{HQE}}=\frac{2}{27}(m_N-\sigma_{\pi N}-\sigma_s)\simeq61\ \mathrm{MeV}, which agrees with our quoted value. However, the present 34% precision on σc\sigma_c does not provide a stringent test of heavy-quark decoupling at the charm mass. The results on σπN\sigma_{\pi N} extracted probing the dependence of the nucleon mass on the light quark mass within the Feynman–Hellmann approach tend to give lower values [19, 47].

Conclusions.— A comprehensive analysis is carried out to extract the light, strange and charm sigma terms of the nucleon in the continuum limit, using only ensembles at physical pion mass and taking into account systematic uncertainties. The analysis includes the accommodation of multiparticle states extracting an energy gap close to the NπN\pi threshold, a finding that is consistent with a variational analysis for B64 [38], where adding nucleon–σ\sigma-meson interpolators and solving the generalized eigenvalue problem [51] substantially reduces the time dependence of the light-quark ratio. Our values for σs\sigma_s and σc\sigma_c provide valuable input for phenomenological searches for beyond the Standard Model physics. For dark-matter direct detection, our σπN\sigma_{\pi N} and σs\sigma_s give the Higgs–nucleon coupling at leading order in the heavy-quark matching [1, 6], fN=29+79(σπN+σs)/mN=0.314(16)f_N=\frac{2}{9}+\frac{7}{9}(\sigma_{\pi N}+\sigma_s)/m_N=0.314(16), in agreement with Ref. [7], where the uncertainty is dominated by σs\sigma_s.

Acknowledgments

We would like to thank all members of the Extended Twisted Mass Collaboration for a very enjoyable cooperation. C.A., C.I., G.K., Y.L., S.P. and F.P. acknowledge partial support from the projects 3D-nucleon (EXCELLENCE/0421/0043), IMAGE-N (EXCELLENCE/0524/0459), Muon-HVP (EXCELLENCE/0524/0017), PulseQCD (EXCELLENCE/0524/0269), Baryon8 (EXCELLENCE/0524/0001), RobustSigma (EXCELLENCE/0925/0325), DeNuTra (EXCELLENCE/524/0455), LENA-QCD (POST-DOC/0925-ROandOther/0180) and partonWF (VISION ERC/0525/0010) funded from the European Regional Development Fund and the Republic of Cyprus through the Cyprus Research and Innovation Foundation. C.A. is supported by European Union’s HORIZON MSCA Doctoral Networks programme, under Grant Agreement No. 101072344, project AQTIVATE (Advanced computing, QuanTum algorIthms and data-driVen Approaches for science, Technology and Engineering). The authors gratefully acknowledge the Gauss Centre for Supercomputing e.V. (www.gauss-centre.eu) for funding this project by providing computing time through the John von Neumann Institute for Computing (NIC) on the GCS Supercomputers JUWELS, JUWELS Booster [52] and JUPITER Booster at Jülich Supercomputing Centre (JSC). We also acknowledge computing time granted on Piz Daint and Alps at Centro Svizzero di Calcolo Scientifico (CSCS) via the projects s1174 and lp13§.

References

  • Corresponding author: [email protected] @page 5/7 starts_mid=1 ends_mid=1 agreement with Ref. [7], where the uncertainty is dominated by σs\sigma_s.

Acknowledgments

We would like to thank all members of the Extended Twisted Mass Collaboration for a very enjoyable cooperation. C.A., C.I., G.K., Y.L., S.P. and F.P. acknowledge partial support from the projects 3D-nucleon (EXCELLENCE/0421/0043), IMAGE-N (EXCELLENCE/0524/0459), Muon-HVP (EXCELLENCE/0524/0017), PulseQCD (EXCELLENCE/0524/0269), Baryon8 (EXCELLENCE/0524/0001), RobustSigma (EXCELLENCE/0925/0325), DeNuTra (EXCELLENCE/524/0455), LENA-QCD (POST-DOC/0925-ROandOther/0180) and partonWF (VISION ERC/0525/0010) funded from the European Regional Development Fund and the Republic of Cyprus through the Cyprus Research and Innovation Foundation. C.A. is supported by European Union’s HORIZON MSCA Doctoral Networks programme, under Grant Agreement No. 101072344, project AQTIVATE (Advanced computing, QuanTum algorIthms and data-driVen Approaches for science, Technology and Engineering). The authors gratefully acknowledge the Gauss Centre for Supercomputing e.V. (www.gauss-centre.eu) for funding this project by providing computing time through the John von Neumann Institute for Computing (NIC) on the GCS Supercomputers JUWELS, JUWELS Booster [52] and JUPITER Booster at Jülich Supercomputing Centre (JSC). We also acknowledge computing time granted on Piz Daint and Alps at Centro Svizzero di Calcolo Scientifico (CSCS) via the projects s1174 and lp13§.

References

References

  1. [1]M. A. Shifman, A. I. Vainshtein, and V. I. Zakharov, Remarks on Higgs Boson Interactions with Nucleons, Phys. Lett. B 78, 443 (1978).
  2. [2]X.-D. Ji, A QCD analysis of the mass structure of the nucleon, Phys. Rev. Lett. 74, 1071 (1995), arXiv:hep-ph/9410274.DOI
  3. [3]M. Hofrichter and J. Ruiz de Elvira, Nucleon mass: trace anomaly and σ-terms (2025), arXiv:2506.23902 [hep-ph].
  4. [4]T. P. Cheng and R. F. Dashen, Is SU(2) x SU(2) a better symmetry than SU(3)?, Phys. Rev. Lett. 26, 594 (1971).DOI
  5. [5]M. Hofrichter, J. Ruiz de Elvira, B. Kubis, and U.-G. Meißner, High-Precision Determination of the Pion-Nucleon σ Term from Roy-Steiner Equations, Phys. Rev. Lett. 115, 092301 (2015), arXiv:1506.04142 [hep-ph].
  6. [6]J. R. Ellis, K. A. Olive, and C. Savage, Hadronic Uncertainties in the Elastic Scattering of Supersymmetric Dark Matter, Phys. Rev. D 77, 065026 (2008), arXiv:0801.3656 [hep-ph].DOI
  7. [7]M. Hofrichter, P. Klos, J. Menéndez, and A. Schwenk, Improved limits for Higgs-portal dark matter from LHC searches, Phys. Rev. Lett. 119, 181803 (2017), arXiv:1708.02245 [hep-ph].
  8. [8]J. Aalbers et al. (LZ), Dark Matter Search Results from 4.2 Tonne-Years of Exposure of the LUX-ZEPLIN (LZ) Experiment, Phys. Rev. Lett. 135, 011802 (2025), arXiv:2410.17036 [hep-ex].DOI
  9. [9]E. Aprile et al. (XENON), First Dark Matter Search with Nuclear Recoils from the XENONnT Experiment, Phys. Rev. Lett. 131, 041003 (2023), arXiv:2303.14729 [hep-ex].
  10. [10]Z. Bo et al. (PandaX), Dark Matter Search Results from 1.54 Tonne-Year Exposure of PandaX-4T, Phys. Rev. Lett. 134, 011805 (2025), arXiv:2408.00664 [hep-ex].DOI
  11. [11]A. Crivellin, M. Hoferichter, and M. Procura, Improved predictions for μ → e conversion in nuclei and Higgs-induced lepton flavor violation, Phys. Rev. D 89, 093024 (2014), arXiv:1404.7134 [hep-ph].DOI
  12. [12]Y. Aoki et al. (Flavour Lattice Averaging Group (FLAG)), FLAG Review 2024 (2024), arXiv:2411.04268 [hep-lat].
  13. [13]M. Hoferichter, J. Ruiz de Elvira, B. Kubis, and U.-G. Meißner, On the role of isospin violation in the pion–nucleon σ-term, Phys. Lett. B 843, 138001 (2023), arXiv:2305.07045 [hep-ph].DOI
  14. [14]J. M. Alarcón, J. Martin Camalich, and J. A. Oller, The chiral representation of the πN scattering amplitude and the pion-nucleon sigma term, Phys. Rev. D 85, 051503 (2012), arXiv:1110.3797 [hep-ph].DOI
  15. [15]C. Alexandrou, W. Drach, K. Jansen, C. Kallidonis, and G. Koutsou, Baryon spectrum with Nf = 2 + 1 + 1 twisted mass fermions, Phys. Rev. D 90, 074501 (2014), arXiv:1406.4310 [hep-lat].DOI
  16. [16]R. Gupta, S. Park, M. Hoferichter, E. Mereghetti, B. Yoon, and T. Bhattacharya, Pion–Nucleon Sigma Term from Lattice QCD, Phys. Rev. Lett. 127, 242002 (2021), arXiv:2105.12095 [hep-lat].DOI
  17. [17]C. Alexandrou, S. Bacchio, M. Constantinou, J. Finkenrath, K. Hadjiyiannakou, K. Jansen, G. Koutsou, and A. Vaquero Aviles-Casco, Nucleon axial, tensor, and scalar charges and σ-terms in lattice QCD, Phys. Rev. D 102, 054517 (2020), arXiv:1909.00485 [hep-lat].DOI
  18. [18]Y.-B. Yang, A. Alexandru, T. Draper, J. Liang, and K.-F. Liu (xQCD), πN and strangeness sigma terms at the physical point with chiral fermions, Phys. Rev. D 94, 054503 (2016), arXiv:1511.09089 [hep-lat].
  19. [19]S. Borsanyi, Z. Fodor, C. Hoelbling, L. Lellouch, K. K. Szabo, C. Torrero, and L. Varnhorst, Ab-initio calculation of the proton and the neutron’s scalar couplings for new physics searches, (2020), arXiv:2007.03319 [hep-lat].arxiv.org/abs/2007.03319
  20. [20]A. Agadjanov, D. Djukannovic, G. von Hippel, H. B. Meyer, K. Ottnad, and H. Wittig, Nucleon Sigma Terms with Nf = 2+1 Flavors of O(a)-Improved Wilson Fermions, Phys. Rev. Lett. 131, 261902 (2023), arXiv:2303.08741 [hep-lat].DOI
  21. [21]C. Alexandrou, S. Bacchio, J. Finkenrath, C. Iona, G. Koutsou, Y. Li, and G. Spanoudes, Nucleon charges and σ-terms in lattice QCD, Phys. Rev. D 111, 054505 (2025), arXiv:2412.01535 [hep-lat].arxiv.org/abs/2412.01535
  22. [22]L. Barca, Current-enhanced excited states in lattice QCD three-point functions, Phys. Rev. D 112, L091503 (2025), arXiv:2508.09006 [hep-lat].arxiv.org/abs/2508.09006
  23. [23]R. Gupta et al., Nucleon form factors and the pion-nucleon sigma term, PoS LATTICE2022, 427 (2023), arXiv:2301.07885 [hep-lat].arxiv.org/abs/2301.07885
  24. [24]C. Alexandrou, G. Koutsou, Y. Li, M. Petschlies, and F. Pittler, Investigation of pion-nucleon contributions to nucleon matrix elements, Phys. Rev. D 110, 094514 (2024), arXiv:2408.03893 [hep-lat].DOI
  25. [25]L. Barca, Evidence of current-enhanced excited states in lattice QCD three-point functions, PoS LATTICE2025, 004 (2026), arXiv:2602.17195 [hep-lat].arxiv.org/abs/2602.17195
  26. [26]C. Alexandrou et al. (Extended Twisted Mass), The hadronic-vacuum-polarization contribution to the muon anomalous magnetic moment in isospin-symmetric lattice QCD with twisted-mass fermions, (2026), arXiv:2610.22224 [hep-lat].arxiv.org/abs/2610.22224
  27. [27]C. Alexandrou et al. (Extended Twisted Mass), Lattice calculation of the short and intermediate time-distance hadronic vacuum polarization contributions to the muon magnetic moment using twisted-mass fermions, Phys. Rev. D 107, 074506 (2023), arXiv:2206.15084 [hep-lat].DOI
  28. [28]C. Alexandrou et al. (Extended Twisted Mass), Strange and charm quark contributions to the muon anomalous magnetic moment in lattice QCD with twisted-mass fermions, Phys. Rev. D 111, 054502 (2025), arXiv:2411.08852 [hep-lat].arxiv.org/abs/2411.08852
  29. [29]C. Alexandrou et al., Simulating twisted mass fermions at physical light, strange and charm quark masses, Phys. Rev. D 98, 054518 (2018), arXiv:1807.00495 [hep-lat].DOI
  30. [30]R. Frezzotti, P. A. Grassi, S. Sint, and P. W. Weisz (Alpha), Lattice QCD with a chirally twisted mass term, JHEP 08, 058, arXiv:hep-lat/0101001.arxiv.org/abs/hep-lat/0101001
  31. [31]R. Frezzotti and G. C. Rossi, Chirally improving Wilson fermions. 1. O(a) improvement, JHEP 08, 007, arXiv:hep-lat/0306014.
  32. [32]C. Alexandrou, C. Iona, G. Koutsou, Y. Li, S. Paul, F. Pittler, and G. Spanoudes, Nucleon scalar charges from lattice QCD at the physical point (2026), in preparation.
  33. [33]K. Osterwalder and E. Seiler, Gauge Field Theories on the Lattice, Annals Phys. 110, 440 (1978).
  34. [34]S. Gusken, U. Low, K. H. Mutter, R. Sommer, A. Patel, and K. Schilling, Nonsinglet Axial Vector Couplings of the Baryon Octet in Lattice QCD, Phys. Lett. B 227, 266 (1989).DOI
  35. [35]C. Alexandrou, S. Gusken, F. Jegerlehner, K. Schilling, and R. Sommer, The Static approximation of heavy - light quark systems: A Systematic lattice study, Nucl. Phys. B 414, 815 (1994), arXiv:hep-lat/9211042.DOI
  36. [36]M. Albanese et al. (APE), Glueball Masses and String Tension in Lattice QCD, Phys. Lett. B 192, 163 (1987).DOI
  37. [37]C. Alexandrou, S. Bacchio, M. Di Carlo, A. Evangelista, C. Iona, F. Pittler, and G. Spanoudes, Flavor decomposition of the nucleon axial and tensor charges from lattice QCD, (2026), arXiv:2611.10816 [hep-lat].arxiv.org/abs/2611.10816
  38. [38]C. Alexandrou, G. Koutsou, Y. Li, S. Paul, and F. Pittler, Excited-State Effects in Nucleon Sigma Terms from a Variational Analysis (2026), arXiv:2610.20362 [hep-lat].arxiv.org/abs/2610.20362
  39. [39]U. Wolff (ALPHA), Monte Carlo errors with errors, Comput. Phys. Commun. 156, 143 (2004), arXiv:hep-lat/0306017.arxiv.org/abs/hep-lat/0306017
  40. [40]F. Joswig, S. Kuberski, J. T. Kuhlmann, and J. Neuendorf, pyerrors: A python framework for error analysis of Monte Carlo data, Comput. Phys. Commun. 288, 108750 (2023), arXiv:2209.14371 [hep-lat].DOI
  41. [41]F. Takahashi et al. (Particle Data Group), Review of Particle Physics, Int. J. Mod. Phys. A 41, 2630011 (2026).
  42. [42]S. R. Beane, Nucleon masses and magnetic moments in a finite volume, Phys. Rev. D 70, 034507 (2004), arXiv:hep-lat/0403015.DOI
  43. [43]M. Hoferichter, J. Ruiz de Elvira, B. Kubis, and U.-G. Meißner, Roy–Steiner-equation analysis of pion–nucleon scattering, Phys. Rept. 625, 1 (2016), arXiv:1510.06039 [hep-ph].
  44. [44]D. Toussaint and W. Freeman (MILC), The Strange quark condensate in the nucleon in 2+1 flavor QCD, Phys. Rev. Lett. 103, 122002 (2009), arXiv:0905.2432 [hep-lat].DOI
  45. [45]W. Freeman and D. Toussaint (MILC), Intrinsic strangeness and charm of the nucleon using improved staggered fermions, Phys. Rev. D 88, 054503 (2013), arXiv:1204.3866 [hep-lat].DOI
  46. [46]M. Gong et al. (XQCD), Strangeness and charmness content of the nucleon from overlap fermions on 2+1-flavor domain-wall fermion configurations, Phys. Rev. D 88, 014503 (2013), arXiv:1304.1194 [hep-ph].DOI
  47. [47]S. Durr et al., Lattice computation of the nucleon scalar quark contents at the physical point, Phys. Rev. Lett. 116, 172001 (2016), arXiv:1510.08013 [hep-lat].DOI
  48. [48]N. Yamanaka, S. Hashimoto, T. Kaneko, and H. Ohki (JLQCD), Nucleon charges with dynamical overlap fermions, Phys. Rev. D 98, 054516 (2018), arXiv:1805.10507 [hep-lat].DOI
  49. [49]G. S. Bali, S. Collins, P. Georg, D. Jenkins, P. Korcyl, A. Schäfer, E. E. Scholz, J. Simeth, W. Söldner, and S. Weißhaupt (RQCD), Scale setting and the light baryon spectrum in N_f = 2 + 1 QCD with Wilson fermions, JHEP 05, 035, arXiv:2211.03744 [hep-lat].DOI
  50. [50]P. Junnarkar and A. Walker-Loud, Scalar strange content of the nucleon from lattice QCD, Phys. Rev. D 87, 114510 (2013), arXiv:1301.1114 [hep-lat].DOI
  51. [51]B. Blossier, M. Della Morte, G. von Hippel, T. Mendes, and R. Sommer, On the generalized eigenvalue method for energies and matrix elements in lattice field theory, JHEP 04, 094, arXiv:0902.1265 [hep-lat].DOI
  52. [52]Jülich Supercomputing Centre, JUWELS Cluster and Booster: Exascale Pathfinder with Modular Supercomputing Architecture at Jülich Supercomputing Centre, Journal of large-scale research facilities 7, 10.17815/jlsrf-7-183 (2021).

Paper details

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

views · 0

downloads · 0

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

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents