WorksPhysics and Astronomy
Magnetic moments of the and baryons from external-field QCD sum rules
Abstract
We compute the magnetic moments of the and isotriplets with external-field QCD sum rules. The heavy-quark propagator is kept with its full mass dependence and with its couplings to the photon, to soft gluons and to both simultaneously, and the magnetic sum rule is normalized by the mass sum rule of the same interpolating field, which removes the residue and gives the baryon masses within a few percent. The quark-condensate magnetic susceptibility , which otherwise dominates the uncertainty, is fixed by the proton magnetic moment computed with the same interpolating field; the neutron moment then follows to about . Uncertainties from the QCD inputs are evaluated with a Monte Carlo analysis, and those from the interpolating field, the Borel window, the scale of the heavy-quark mass, the calibration and the normalization are added. With heavy-quark masses we obtain , , , , and in nuclear magnetons; at this order the results depend sizably on the heavy-quark mass scheme. The moments are carried by the light diquark; the heavy-quark contribution, whose separation depends on the interpolating field, is for and for .
Introduction
Magnetic moments of singly heavy baryons probe how the magnetic response is shared between a heavy quark and a light diquark. In the heavy-quark limit the light-diquark spin in decouples from the heavy-quark spin; the light-diquark contribution to the moment is then the same for and , while the heavy-quark contribution is suppressed by . None of these moments has been measured. Measurements of charm-baryon magnetic moments are, however, being prepared. In a fixed-target experiment at the LHC, charm baryons produced in a target are channeled in a bent crystal, whose strong interplanar electric field induces a measurable spin precession [1, 2, 3, 4, 5]. The dedicated experiment ALADDIN, proposed in a Letter of Intent to the LHCC [6], aims to measure the moments of and , whose lifetimes are – s, with projected uncertainties of 0.012 and 0.018 for the baseline crystal configuration after two years of data taking in LHC Run 4. The method requires weakly decaying baryons, so that the strongly decaying states are not accessible directly. In the constituent quark model, however, the moment equals the charm-quark moment [4, 6], which also enters the moments; such measurements will therefore constrain the heavy-quark contribution and test the approaches used for heavy-baryon moments, including their predictions for . Predictions exist from quark models [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17], lattice QCD for [18], chiral perturbation theory [19, 20], the chiral quark-soliton model [21], external-field QCD sum rules [22] and, recently, light-cone QCD sum rules [23]; for they range from 1.3 to 3.1 . Light-cone QCD sum rules have also been applied to the magnetic moments of baryons [24], of spin-3/2 heavy baryons [25] and to the transitions [26].
In the external-field method [27, 28] the baryon two-point correlator is evaluated in a weak static electromagnetic field, which couples to the quark propagators and, through vacuum susceptibilities, to the QCD vacuum. The only external-field calculation for heavy baryons, Ref. [22], treated the triplet with the Ioffe-type current. It kept the perturbative term and a subset of condensate contributions, did not include the gluon condensate or the couplings of the charm quark to gluons, took the residue from independent mass sum rule analyses, included leading-order anomalous-dimension factors, and estimated the uncertainty by varying the susceptibility . Light-cone sum rules [23] describe the long-distance photon emission through photon distribution amplitudes, whereas in the external-field method it enters through the vacuum susceptibilities; the two approaches are complementary, with different nonperturbative inputs and different sources of uncertainty.
In this work we (i) keep the full heavy-quark mass dependence in all terms, including the couplings of the heavy quark to the photon, to soft gluons, to both simultaneously, and to the gluon condensate; (ii) normalize the magnetic sum rule by the mass sum rule of the same current and the same OPE, so that cancels and the mass sum rule serves as a check; (iii) treat both the and the triplets; (iv) fix with the proton magnetic moment computed with the same current, and evaluate the uncertainties with a Monte Carlo analysis [29, 30] together with the dependence on the interpolating field, the Borel window and the heavy-quark mass scheme; and (v) separate the light- and heavy-quark contributions. This is, to our knowledge, the first external-field analysis of the moments, with the light- and heavy-quark contributions separated and with an uncertainty budget that includes the condensates and susceptibilities, the interpolating field, the Borel window and the scale of the heavy-quark mass. The new OPE ingredients (gluon-condensate terms and the simultaneous coupling of the heavy quark to the photon and to gluons) turn out to be numerically small, about 1% of the OPE; the main differences with respect to Ref. [22] come from the normalization, the treatment of and the error analysis. Unlike Ref. [22], we do not include anomalous-dimension factors.
Sum rules
For with or we use
where is the charge-conjugation matrix and are color indices. This is the proton field of Refs. [30, 31] with . For (Ioffe current) it is, after a Fierz transformation, the standard current with the heavy quark outside a spin-one light diquark. The states are obtained from isospin symmetry, ; this avoids the admixture of that a current of the form (1) would have for two different light flavors.
We consider in a static field and the structure , , which is chiral-even in the light-quark sector and was found to be the most reliable one for the nucleon [30]. After Borel transformation the phenomenological sides of the and of the mass sum rules are
where is the magnetic moment in natural magnetons and accounts for transitions to excited states, which are not suppressed by the Borel transformation. Their ratio,
is free of and ; is the moment in nuclear magnetons.
OPE. For the light quarks () we use the propagator in the Fock–Schwinger gauge to first order in and up to order [27, 28, 30],
where the terms with open indices and are contracted with the gluon field emitted by another quark line, and the susceptibilities are defined by , and ; describes the tensor quark condensate induced by the external field.
For the heavy quark we work in momentum space. In the Fock–Schwinger gauge the photon and gluon fields combine into a constant field strength , and the propagator is generated recursively,
with the same coupling convention as in Eq. (5). To the order needed here,
with (contracted with a gluon from a light-quark line). The second term is the first-order term ; for its Fourier transform coincides with the corresponding term of Eq. (5). The third term follows from the second order in averaged over the gluon condensate, , consistent with . is the mixed term of order (second order of Eq. (6)), and the term of order averaged over (third order). We include all terms up to operator dimension eight, counting and as dimension three, as four, and , and as five, in the constant-field approximation (i.e. without $\langle g_s^3G^3\rangle and derivative terms). and anomalous-dimension corrections are not included; the anomalous dimensions of the heavy-quark current and of the operators differ from the light-baryon case and would require a separate study.
Construction. After the Dirac algebra (FeynCalc [32, 33, 34]) each term has the form , where collects the light quarks, and . The Fourier transform
and its derivatives give . The loop integral is done with a Feynman parameter , leading to with , and is followed by the Borel transformation . The light-quark powers are regulated analytically, . The poles appear only in end-point contributions at and cancel against zeros of for light-quark structures that are polynomial in (e.g. ); they leave finite pole terms proportional to . No logarithm of a renormalization scale remains. Every term has its spectral weight at : a continuous density for , and a function or its derivatives for . We remove all weight above , i.e. the integration is restricted to ; the pole terms at are kept. The perturbative spectral densities are, with ,
The perturbative heavy-quark term vanishes for ; the heavy-quark contribution then comes from the mass-dependent condensate terms. The complete OPE for is given in Appendix Appendix A.
Validation. (i) For the pseudoscalar heavy–light correlator the construction reproduces the known perturbative, quark-condensate, mixed-condensate and gluon-condensate contributions to ten digits for and GeV. (ii) For and the free, photon and soft-gluon terms of the heavy line reproduce the corresponding terms of the proton sum rules obtained with the light-quark propagator (5): the mass sum rules at both chiral structures exactly, and the magnetic sum rule term by term, including the susceptibility terms. The heavy-quark terms with a gluon condensate or with simultaneous photon and gluon coupling have no counterpart in the light-quark propagator and are not part of this comparison; in particular, the term of Appendix Appendix A, which survives for , comes entirely from the term. In the all-light case the sum rule has no term: all diagrams contributing to it are polynomial in and vanish after Borel transformation. (iii) The recursion (6) reproduces both the photon vertex and the gluon-condensate term with a single normalization.
Table 1. Input parameters and the Gaussian widths used in the Monte Carlo analysis. The widths of and are our choice; is fixed in each Monte Carlo sample by the proton magnetic moment (central value and calibration uncertainty; see text).
Numerical analysis
The inputs are listed in Table I. The heavy-quark masses are taken in the scheme and the baryon masses from Ref. [40].
Calibration of . Estimates of range from about to [36, 37, 38], partly because depends on the renormalization scale (the windows correspond to and GeV for and ); the lattice gives at GeV [39]. Since dominates the uncertainty, we fix it with the proton magnetic moment. The proton sum rules follow from the same current with (massless); we evaluate them with the same OPE truncation and the ratio method and require [40]. Because the sum rule is linear in , this is done in every Monte Carlo sample with that sample’s other inputs, so that the variations common to the nucleon and sum rules largely cancel. For the proton the term is the leading power correction and the dimension criterion used below cannot be met; we require instead that dimension-eight terms contribute at most of the mass OPE and that the pole contribution exceeds , which together with the proton mass gives – and –. We calibrate at , where these criteria hold, and use the same for all values of . This yields at central inputs; over the Monte Carlo samples, where it follows the other inputs, its spread is (this spread is part of the Monte Carlo uncertainty). As a test, the neutron sum rule (the same OPE with ) then gives and , compared with and , without further adjustment. The neutron moment itself depends strongly on (, , and for , , and ), whereas the ratio varies only between and ; the test thus confirms mainly the isospin structure of the sum rule. Since approximations at the 10% level enter the derivation of leading-order baryon sum rules [29], consistent with the 5% deviation of , we assign a 10% theoretical uncertainty to the nucleon sum rule, which corresponds to , and add the variation for proton windows shifted or enlarged by within the criteria (), giving a calibration uncertainty , which we add as a separate systematic uncertainty. The calibrated is an effective parameter of the leading-order sum rules: it is two to three times larger in magnitude than the lattice value and absorbs part of the corrections common to the nucleon and sum rules. For comparison, was used for the octet baryons in Ref. [30]. The windows of the proton and the sum rules lie at different scales, , and GeV. At leading order the term, which carries the dependence, scales as (from and ); between these scales it grows by 9% for and 16% for . Since anomalous dimensions are not included elsewhere, we do not apply this factor but add the corresponding shift to the calibration uncertainty.
Borel window. We apply the same criteria to both flavors: the pole contribution to the mass sum rule, , is at least 50% at the upper end of the window for all considered, and operators of dimension contribute at most 15% of the mass OPE at the lower end. For the Ioffe current this gives –$2.2\ {\rm GeV}^2$ with – for , and – with – for . In these windows the pole contribution to the magnetic sum rule exceeds 80%. The mass extracted from the mass sum rule, , is – GeV for and – GeV for $\Sigma_b$ {xref}(Fig. 1)`, compared with and GeV; it rises across the window without a plateau, which reflects the simple pole-plus-continuum ansatz. The thresholds – GeV and – GeV lie above the negative-parity and , which contribute to the structures with the same sign as the ground state; separating them would require parity-projected sum rules, and this effect is not included in our uncertainties.
OPE structure. At the OPE in the windows consists of the perturbative term (54–78%), dimension-six terms (24–50%, mostly , as for the nucleon [27]), dimension-eight terms (−2 to −8%) and a gluon-condensate contribution of about 1% (ranges over the windows, the values of and the states). At the lower end of the windows, where the condensates are largest, the dimension-six and dimension-eight terms are 40–50% and −4 to −8% of the OPE, so that the expansion converges as well as for the mass sum rule. Our and terms follow the definitions given with Eq. (5) and the values of Table I; the corresponding terms of Ref. [22] have different coefficients, which amounts to a different definition of . Doubling the term changes the moments by , , and for , , and , well within the uncertainties. Through the term, depends approximately linearly on , with the slopes in Table II. Leaving out all gluon-condensate terms changes by about , and leaving out the mixed condensate by about . With , the value used in Ref. [22], our sum rule gives at central inputs, compared to for the calibrated .
Ratio. The upper panels of Figure 2 show over a wide range of . Inside the windows it is well described by a straight line, but the lever arm in is short ( and ), so that the linear form is not a strong test; at larger , where the continuum dominates, deviates from the line. Correspondingly, the slope and the intercept are strongly anticorrelated, () and (), and the result depends strongly on the choice of the window, which we include as an uncertainty. As a cross-check of the normalization we also use the conventional procedure, with from the mass sum rule with the experimental baryon mass, averaged over the window, and a fit of alone. This gives and for and (ratio method: and ) and and for and (ratio method: and ) at central inputs. The differences reflect the variation of the extracted mass across the windows (Figure 1); they are largest for , whose mass sum rule falls below the experimental mass in most of the window. We include half of them as a normalization uncertainty, and for and and below for . The lower panels show the moments as functions of for three values of , defined as with from the fit in the window and the mass from the same . Inside the windows varies by less than 1.5%, which partly reflects this construction; the dependence on is about 4% for and 1% for across the ranges considered, and outside the windows the curves rise slowly as the continuum contribution grows.
Figure 1. Baryon mass from the mass sum rule () for the lower and upper value of . Shaded: Borel window; dashed: experimental mass.

Figure 2. Upper panels: the ratio of Eq. (4) at central inputs (, and ). Large symbols: points in the Borel window (shaded); dashed: linear fit in the window, whose slope is in natural magnetons. Lower panels: magnetic moments as functions of at central inputs for three values of , , with fitted in the window at the same .
Monte Carlo and uncertainties. We generate 1000 sets of inputs according to Table I, with calibrated on the proton in each set, uniformly distributed in its range and according to its PDG uncertainty. Each set is used for all six states, so that correlations between them are kept; the moments are averaged sample by sample. For each set we fit Eq. (4) on the Borel grid inside the window and quote the median and the half width of the central 68% interval. We add in quadrature (a) the interpolating-field uncertainty, half the difference between the results for and , i.e. for a variation of the mixing parameter around the Ioffe current with the same windows and the same calibrated . At the window criteria are satisfied; at they are marginally violated at the edges (pole contribution 45 and 48% at the upper end of the windows for the lower , dimension fraction 16% for ), which we accept for this estimate; the pole contribution to the magnetic sum rule stays above 80% throughout; (b) the window uncertainty, half the spread of the central values for windows shifted or enlarged by () or () at either end; (c) the scale uncertainty of the mass, the shift of the central value when is replaced by , i.e. and (three-loop running with ), quoted as a symmetric uncertainty although the shift lowers , while the pole mass raises it; (d) the normalization uncertainty, half the difference between the ratio method and the procedure described above; and (e) the calibration un-
certainty, combined with the scale shift of the term described above. With the calibrated the Monte Carlo uncertainty is three to four times smaller than for a fixed range ; the remaining correlation with , which now varies with the other condensates, is –. The PDG uncertainty of the masses contributes , much less than their scale dependence. The total uncertainties are comparable to those obtained for the octet moments in the external-field method with conservative input widths [30] and to the accuracy of leading-order sum rules estimated in Ref. [29]; analyses that vary only the QCD inputs quote smaller uncertainties [22, 23]. The light- and heavy-quark contributions follow from setting or , since the OPE is linear in the charges. The results are collected in Table II.
Heavy-quark mass scheme. At leading order the scheme of is not fixed. With the pole masses and and the same windows and thresholds, the moments change by the amounts given in the last column of Table II, e.g. for . However, with pole masses the pole contribution to the mass sum rule at the upper end of these windows is only 29–47%, and no choice of satisfies the pole criterion and reproduces the baryon mass at the same time (the mass sum rule then gives 2.7 and 6.2 GeV for large enough for 50% pole dominance). The masses lead to a consistent analysis, and we regard the pole-mass shifts as an upper estimate of the scheme dependence rather than as part of the uncertainty; the scale variation (c) of the mass, which moves in the opposite direction, is included.
Discussion
Table III and Figure 3 compare our results with other approaches. Within the uncertainties they agree with the chiral quark-soliton model [21] for all six states and, for and , with the earlier external-field sum rules [22]; for the latter give , about 1.8 standard deviations below our value. The lattice values for and are and (linear) or and (quadratic chiral fit) [18]; our results agree with the quadratic fit and lie 1.2 and 1.4 standard deviations above the linear fit in magnitude. Chiral perturbation theory [19, 20] gives smaller magnitudes. The agreement with Ref. [22] for is partly accidental: with their value of our sum rule gives a smaller moment, while the normalization through the mass sum rule acts in the opposite direction.
Figure 3. Our results (circles, total uncertainty) compared with the chiral quark-soliton model [21] (for its moments equal those of the states with the same light quarks), lattice QCD (linear chiral fit) [18], external-field [22] and light-cone [23] sum rules, covariant [20] and heavy-baryon [19] chiral perturbation theory, and quark models (crosses: the six models in Table III).
The recent light-cone sum-rule analysis [23] agrees with our results within about one standard deviation for all six states; the largest differences, 0.35 and 0.38 , occur for and , where our magnitudes are larger. The two approaches are complementary: the light-cone sum rules parametrize the long-distance photon emission through photon distribution amplitudes, whereas in the external-field method it enters through the susceptibilities , and , of which we fix the dominant one, , with the proton moment. Ref. [23] finds nearly equal moments for corresponding and states; our central values for and are about 10–30% larger in magnitude than those for and , but within our uncertainties the two flavors are also compatible with equal moments (see below). The quark models in Table III give – for and – for ; their spread, which reflects different quark masses, wave functions and treatments of relativistic and screening effects, is comparable to our uncertainty, and our central values lie at or slightly above its upper end for and inside it for . All approaches satisfy, exactly or approximately, the isospin relation , which holds by construction in our analysis; since the charged moments largely cancel in this average, the moments test their relative size, and our and values, and , agree well with the light-cone sum rules and most quark models.
The moments are carried by the light diquark. The heavy-quark contribution is in and
+0.013 ± 0.003 in , and varies by about 0.01 between and . At it comes from mass-dependent condensate terms (Eq. (8)), so that its separation from the light-quark part depends on the interpolating field and the Dirac structure. The nonrelativistic quark-model terms are to for charm and about for bottom (for quark masses of 3.1–3.7 and 4.7–5 GeV), of the same sign and size. Since in this model , the planned measurement of the moment with an uncertainty of about 0.01 [6] would fix the charm term in to a few thousandths of . Taken at face value, our heavy-quark contribution corresponds to , at the upper end of the range 0.15–0.40 of the predictions for compiled in Ref. [6]; given the dependence of the separation on the interpolating field noted above, this comparison is qualitative. The lattice calculation finds a charm contribution (linear) or (quadratic fit), and a light-quark contribution for [18], compared with our 2.50 ± 0.19 . Our heavy-quark contribution is thus larger in magnitude than the lattice value by a factor of two to three (2.5 for the linear fit). Part of this difference may come from the separation itself: on the lattice it is defined through the electromagnetic current, here through the charge factors of the OPE of a particular current and Dirac structure at leading order, in which the heavy-quark term is generated by mass-dependent condensate contributions; the two definitions need not coincide. In the total moment the difference resides mainly in the light-quark sector. The light-diquark contributions in and differ by , where the uncertainty combines the correlated Monte Carlo error (0.06) with the interpolating-field (0.22) and window (0.34) systematics; the scale of cancels in the difference, and adding the normalization uncertainties of the two states would increase the error to about 0.49 . The two flavors are therefore compatible with the heavy-quark limit; the windows lie at different Borel scales, and anomalous-dimension effects, which we neglect, act between them, so that a deviation of the expected size cannot be resolved at this order. The main limitations are the absence of and anomalous-dimension corrections, the heavy-quark mass scheme at leading order, the use of an effective calibrated on the proton, the truncation of the OPE, and the description of excited states through a single transition constant and a continuum threshold, including the
| State | MC | window | scale | cal. | norm. | pole mass | ||||||||
| 2.37 ± 0.65 | 0.18 | 0.41 | 0.17 | 0.32 | 0.29 | 0.08 | 2.50 ± 0.19 | −0.134 ± 0.023 | 0.39 ± 0.18 | −0.38 | 1.43 | +0.67 | ||
| 0.50 ± 0.15 | 0.04 | 0.10 | 0.03 | 0.06 | 0.08 | 0.02 | 0.63 ± 0.05 | −0.134 ± 0.023 | 0.19 ± 0.05 | −0.10 | 0.26 | +0.12 | ||
| −1.39 ± 0.37 | 0.11 | 0.21 | 0.11 | 0.21 | 0.15 | 0.04 | −1.25 ± 0.09 | −0.134 ± 0.023 | −0.02 ± 0.10 | 0.19 | −0.92 | −0.43 | ||
| 2.16 ± 0.65 | 0.19 | 0.19 | 0.26 | 0.36 | 0.28 | 0.26 | 2.15 ± 0.19 | 0.013 ± 0.003 | −0.54 ± 0.17 | −0.26 | 0.93 | +0.80 | ||
| 0.55 ± 0.16 | 0.05 | 0.05 | 0.07 | 0.09 | 0.07 | 0.06 | 0.54 ± 0.05 | 0.013 ± 0.003 | −0.14 ± 0.04 | −0.07 | 0.24 | +0.20 | ||
| −1.06 ± 0.32 | 0.09 | 0.10 | 0.13 | 0.17 | 0.14 | 0.14 | −1.08 ± 0.09 | 0.013 ± 0.003 | 0.25 ± 0.08 | 0.13 | −0.45 | −0.39 |
Table 2. Magnetic moments (in ) for with heavy-quark masses: median and total uncertainty, with its Monte Carlo (MC), interpolating-field (), window, mass-scale, calibration and normalization components; light- and heavy-quark contributions , ; transition strength (GeV); sensitivities (in GeV) and (in GeV) at central inputs; shift of the central value for pole masses (see text).
| Approach | ||||||
| This work | 2.37 ± 0.65 | 0.50 ± 0.15 | −1.39 ± 0.37 | 2.16 ± 0.65 | 0.55 ± 0.16 | −1.06 ± 0.32 |
| Light-cone QCDSR [23] | 2.02(18) | 0.50(5) | −1.01(9) | 2.02(19) | 0.53(6) | −1.01(9) |
| External-field QCDSR [22] | 2.1 ± 0.3 | 0.23 ± 0.03 | −1.6 ± 0.2 | – | – | – |
| Lattice QCD [18] | 1.57(25)/2.22(51) | – | −0.85(13)/−1.07(27) | – | – | – |
| Chiral quark-soliton model [21] | 2.15(10) | 0.46(3) | −1.24(5) | 2.15(10) | 0.46(3) | −1.24(5) |
| HB ChPT, [19] | 1.50 | 0.26(7) | −0.97(5) | – | – | – |
| Covariant ChPT [20] | 1.285 | 0.190 | −0.905 | – | – | – |
| Relativistic three-quark model [8] | 1.76 | 0.36 | −1.04 | 2.07 | 0.53 | −1.01 |
| Nonrelativistic quark model | 2.35 | 0.49 | −1.37 | 2.50 | 0.64 | −1.22 |
| Bag model [9] | 1.68 | 0.32 | −1.04 | 1.62 | 0.42 | −0.78 |
| Bag model [14] | 2.28 | 0.49 | −1.31 | 2.25 | 0.60 | −1.15 |
| Eff. mass, screened charge | 2.10 | 0.43 | −1.23 | 2.19 | 0.56 | −1.06 |
| Hypercentral model [17] | 1.95 | 0.39 | −1.17 | 2.12 | 0.55 | −1.03 |
| Quark models, range [7, 8, 9, 10, 11, 12, 13] | 1.68–3.07 | – | −1.78 to −1.02 | – | – | – |
Table 3. Magnetic moments (in ) compared with other approaches. Lattice values: linear/quadratic chiral extrapolation. Quark models, range: range of the values compiled in Ref. [18]. In the chiral quark-soliton model the moments equal those of the states with the same light quarks [21]. Used as input to fix low-energy constants. Ref. [22] gives 0.6 ± 0.1 in units of , i.e. 0.23 ± 0.03 ; the abstract quotes the former number with the unit . Simple nonrelativistic quark model as evaluated in Ref. [9]. Refs. [15] () and [16] ().
negative-parity states. With fixed, the interpolating field and the scale of the heavy-quark mass give the largest remaining uncertainties for , and the scale, the calibration, the window and the normalization for ; next-to-leading-order corrections would reduce the scale dependence. With the slopes in Table II the results can be rescaled to other values of .
Summary
We have computed the magnetic moments of the and isotriplets in external-field QCD sum rules with the full heavy-quark mass dependence, including the couplings of the heavy quark to the photon and to gluons, and with the mass sum rule of the same current as normalization. The construction reproduces known heavy-light and light-baryon sum rules, and the mass sum rules give the baryon masses within a few percent. The results in Table II agree with the chiral quark-soliton model and with the quadratic chiral extrapolation of lattice QCD for , and with recent light-cone sum rules for all six states; the moments are carried by the light diquark, and within the uncertainties the and moments are compatible with the heavy-quark limit. Fixing with the proton magnetic moment, which also reproduces the neutron moment, removes the largest source of uncertainty; the dependence on the heavy-quark mass scheme is the main open issue at this order.
OPE for
For , in the normalization of Eq. (2), with , , , , , the charge of the two light quarks and that of the heavy quark,
The terms with are pole terms at ; the integrals contain the continuum-subtracted gluon-condensate contributions. The perturbative heavy-quark term and all terms linear in are proportional to and vanish here; they are included for .
References
- [1]V. G. Baryshevsky, Phys. Lett. B 757 (2016) 426.
- [2]A. S. Fomin et al., JHEP 08 (2017) 120, arXiv:1705.03382.DOI
- [3]E. Bagli et al., Eur. Phys. J. C 77 (2017) 828; Erratum: Eur. Phys. J. C 80 (2020) 680, arXiv:1708.08483.DOI
- [4]A. S. Fomin, S. Barsuk, A. Yu. Korchin, A. V. Kovalchuk, E. Kou, M. Liul, A. Natochii, E. Niel, P. Robbe, A. Stocchi, Eur. Phys. J. C 80 (2020) 358, arXiv:1909.04654.DOI
- [5]S. Aiola et al., Phys. Rev. D 103 (2021) 072003, arXiv:2010.11902.DOI
- [6]K. Akiba et al., ALADDIN: An LHC Apparatus for Direct Dipole moments Investigation, Letter of Intent, CERN-LHCC-2024-011, LHCC-I-041 (2024), doi:10.17181/CERN.2G4V0YAO.DOI
- [7]B. Julia-Diaz, D. O. Riska, Nucl. Phys. A 739 (2004) 69, arXiv:hep-ph/0401096.
- [8]A. Faessler, T. Gutsche, M. Ivanov, J. G. Körner, V. E. Lyubovitskij, D. Nicmorus, K. Pumsa-ard, Phys. Rev. D 73 (2006) 094013, arXiv:hep-ph/0602193.DOI
- [9]A. Bernotas, V. Simonis, Lith. J. Phys. 53 (2013) 84, doi:10.3952/physics.v53i2.2668, arXiv:1209.2900.arxiv.org/abs/1209.2900
- [10]N. Sharma, H. Dahiya, P. K. Chatley, M. Gupta, Phys. Rev. D 81 (2010) 073001, arXiv:1003.4338.DOI
- [11]N. Barik, M. Das, Phys. Rev. D 28 (1983) 2823.
- [12]S. Kumar, R. Dhir, R. C. Verma, J. Phys. G 31 (2005) 141.
- [13]B. Patel, A. K. Rai, P. C. Vinodkumar, J. Phys. G 35 (2008) 065001, arXiv:0710.3828.arxiv.org/abs/0710.3828
- [14]V. Visinelli, arXiv:1803.01809.
- [15]B. Mohan, T. M. S., A. Hazra, R. Dhir, Phys. Rev. D 106 (2022) 113007, arXiv:2211.16418.DOI
- [16]R. Dhir, C. S. Kim, R. C. Verma, Phys. Rev. D 88 (2013) 094002, arXiv:1309.4057.DOI
- [17]A. Majethiya, B. Patel, P. C. Vinodkumar, Eur. Phys. J. A 38 (2008) 307, arXiv:0805.3439.
- [18]K. U. Can, G. Erkol, B. Isildak, M. Oka, T. T. Takahashi, JHEP 05 (2014) 125, arXiv:1310.5915.DOI
- [19]G.-J. Wang, L. Meng, H.-S. Li, Z.-W. Liu, S.-L. Zhu, Phys. Rev. D 98 (2018) 054026, arXiv:1803.00229.DOI
- [20]R.-X. Shi, Y. Xiao, L.-S. Geng, Phys. Rev. D 100 (2019) 054019, arXiv:1812.07833.DOI
- [21]G.-S. Yang, H.-C. Kim, Phys. Lett. B 781 (2018) 601, arXiv:1802.05416.DOI
- [22]S.-L. Zhu, W.-Y. P. Hwang, Z.-S. Yang, Phys. Rev. D 56 (1997) 7273, arXiv:hep-ph/9708411.DOI
- [23]U. Özdem, Eur. Phys. J. A 61 (2025) 62, arXiv:2411.09405.arxiv.org/abs/2411.09405
- [24]T. M. Aliev, K. Azizi, A. Ozpineci, Phys. Rev. D 77 (2008) 114006, arXiv:0803.4420.
- [25]T. M. Aliev, K. Azizi, A. Ozpineci, Nucl. Phys. B 808 (2009) 137, arXiv:0807.3481.DOI
- [26]T. M. Aliev, A. Ozpineci, M. Savci, Phys. Rev. D 65 (2002) 096004, arXiv:hep-ph/0110074.DOI
- [27]B. L. Ioffe, A. V. Smilga, Nucl. Phys. B 232 (1984) 109.
- [28]I. I. Balitsky, A. V. Yung, Phys. Lett. B 129 (1983) 328.
- [29]D. B. Leinweber, Ann. Phys. 254 (1997) 328, arXiv:nucl-th/9510051.DOI
- [30]L. Wang, F. X. Lee, Phys. Rev. D 78 (2008) 013003, arXiv:0804.1779.DOI
- [31]B. L. Ioffe, Nucl. Phys. B 188 (1981) 317.
- [32]R. Mertig, M. Böhm, A. Denner, Comput. Phys. Commun. 64 (1991) 345.
- [33]V. Shtabovenko, R. Mertig, F. Orellana, Comput. Phys. Commun. 256 (2020) 107478.
- [34]V. Shtabovenko, R. Mertig, F. Orellana, Comput. Phys. Commun. 306 (2025) 109357, arXiv:2312.14089.DOI
- [35]M. A. Shifman, A. I. Vainshtein, V. I. Zakharov, Nucl. Phys. B 147 (1979) 385.
- [36]V. M. Belyaev, Ya. I. Kogan, Yad. Fiz. 40 (1984) 1035 [Sov. J. Nucl. Phys. 40 (1984) 659].
- [37]P. Ball, V. M. Braun, N. Kivel, Nucl. Phys. B 649 (2003) 263, arXiv:hep-ph/0207307.DOI
- [38]A. Wainshtein, Phys. Lett. B 569 (2003) 187, arXiv:hep-ph/0212231.
- [39]G. S. Bali, F. Bruckmann, M. Constantinou, M. Costa, G. Endrődi, S. D. Katz, H. Panagopoulos, A. Schäfer, Phys. Rev. D 86 (2012) 094512, arXiv:1209.6015.DOI
- [40]S. Navas, et al. (Particle Data Group), Phys. Rev. D 110 (2024) 030001.

