WorksPhysics and Astronomy
Comprehensive Assessment of Properties for Nuclear Clock and Fundamental Physics Applications
Abstract
By employing singles, doubles, and triples excitations within the relativistic coupled-cluster framework, we perform comprehensive calculations of a wide range of atomic properties for the Th ion. These properties are essential for advancing nuclear clock technology and probing fundamental physics. Combining our isotope shift parameters with experimental data, we estimate highly accurate values of the differential nuclear charge radii for Th and Th. Additionally, we determine the nuclear magnetic dipole and electric quadrupole moments for both the ground and isomeric states of Th by combining measured hyperfine structure constants with our theoretical calculations. Our precise evaluations of electric dipole polarizabilities and hyperfine-induced quadrupole moments are critical for assessing systematic uncertainties in Th-based nuclear clock. Notably, we observe unexpectedly significant contributions from higher-order relativistic effects and excitations involving orbitals with higher angular momentum, which markedly influence the energies of the ground state and its fine-structure partner. These results highlight the substantial challenges in achieving highly accurate predictions for these properties.
The thorium isotope Th possesses a unique low-energy isomeric state at approximately 8.3 eV [1, 2, 3], making it the only known candidate for an optical nuclear clock based on this transition. The proposed clock scheme exploits the transition between the states in triply ionized Th, involving the nuclear ground and isomeric states with spins and , respectively (see Figure 1). This clock could reach fractional uncertainties as low as [4] and offers exceptional sensitivity to potential temporal variations of fundamental constants, as well as to dark matter interactions [5, 6]. The francium-like electronic configuration of Th enables accurate ab initio calculations, while its atomic energy levels are well-suited for laser cooling and trapping, establishing an ideal platform for high-precision spectroscopy [7, 8, 9]. Experimentally, two primary methods are considered for cooling Th ions for spectroscopic studies [2, 8]: (i) a closed two-level transition , accessible with a 1087 nm laser, and (ii) a closed three-level (-type) transition , requiring lasers at 690 nm and 984 nm wavelengths.

Figure 1. Energy level diagram of the Th ion with the proposed clock and cooling transition lines involving the ground and isomeric manifolds.
From a fundamental physics perspective, precise characterization of both nuclear and electronic properties of thorium isotopes is essential. In particular, accurate determination of isotope shift (IS) parameters plays a crucial role, as these parameters provide vital insights into nuclear structure and electron-nucleus interactions within thorium isotopes. Despite their significance, relatively few studies have focused on calculating the IS constants of Th; most existing research has predominantly addressed field shift (FS) constants, with limited attention paid to mass shift (MS) constants [10, 11, 12]. Notably, there is an approximately 8% discrepancy between reported values of the FS constant () for the electronic ground state [11, 12], and about a 7% difference in the estimated change in the root-mean-square (rms) nuclear charge radius, , between the isomeric and ground states of Th, as reported in Refs. [11] and [13]. This highlights the urgent need for more accurate and consistent evaluations of IS constants in Th.
The leading-order IS in a transition in two isotopes and is given by [14, 15]
where with are masses of isotopes and is the MS constant, which is decomposed into the normal mass shift (NMS) and the specific mass shift (SMS) components: .
| Method | ||||
| Energy (in cm) | ||||
| DHF | −206642 | −203213 | −211800 | −207575 |
| RCCSD | −227266 | −222883 | −221669 | −216327 |
| RCCSDT | −227578 | −223257 | −221590 | −216335 |
| +Basis | −1201 | −1197 | −362 | −323 |
| +Breit | −714 | −847 | −44 | −122 |
| +VP | 38 | 37 | 16 | 14 |
| Final | −229455 | −225264 | 221980 | 216766 |
| Experiment | −231065 | −226740 | −221872 | −216579 |
| Accuracy | 1% | 1% | 2% | 2% |
| (in GHz/fm) | ||||
| DHF | 51.26 | 49.06 | 21.90 | 19.16 |
| RCCSD | 53.61 | 51.47 | 21.5 | 19.60 |
| RCCSDT | 53.09 | 50.96 | 21.34 | 19.48 |
| +Basis | 0.25 | 0.17 | 0.09 | 0.12 |
| +Breit | ∼ 0.0 | ∼ 0.0 | 0.08 | −0.15 |
| +VP | 0.77 | 0.60 | 0.25 | 0.09 |
| Final | 54.1(15) | 51.7(12) | 21.8(10) | 19.5(10) |
| SD+CI [11] | 55.0(55) | 53.0(53) | 23.3(23) | 20.5(20) |
| CI [12] | 51.4 | |||
| (in GHz-amu) | ||||
| DHF | 3609.8 | 3624.1 | 3500.2 | 3456.4 |
| RCCSD | 3926.4 | 3933.4 | 3637.1 | 3580.4 |
| RCCSDT | 3937.7 | 3945.3 | 3640.5 | 3584.4 |
| +Basis | −6.1 | −4.3 | 4.3 | 13.3 |
| +Breit | 35.0 | 41.5 | 1.4 | 5.6 |
| +VP | −3.8 | −3.6 | −1.8 | −5.2 |
| Final | 3963(84) | 3979(75) | 3644(78) | 3598(70) |
| (in GHz-amu) | ||||
| DHF | −9367.7 | −9074.5 | −3097.2 | −3169.4 |
| RCCSD | −8743.8 | −8549.4 | −2621.4 | −2727.6 |
| RCCSDT | −8852.6 | −8669.4 | −2640.0 | −2739.9 |
| +Basis | 19.1 | 15.2 | −5.1 | −5.2 |
| +Breit | −121.8 | −109.5 | −13.7 | −12.1 |
| +VP | 0.9 | 0.7 | −0.5 | −0.5 |
| Final | −8954(165) | −8763(150) | −2659(100) | −2758(109) |
Table 1. Energies and IS constants of the ground and low-lying excited states of Th at different levels of approximation. The energy values are compared with experimental data [20, 21] to assess their accuracy, while the FS constants are contrasted with existing theoretical calculations.
By incorporating the relativistic forms of the MS operators and the Fermi-charge distribution for the FS operator (see Ref. [15] for the explicit expressions), the IS constants are estimated using the finite-field (FF) approach through an expansion
where is the energy of an atomic state with valence orbital due to the atomic Hamiltonian , is the energy obtained from the total atomic Hamiltonian with is an arbitrary parameter, and represents the corresponding IS operator. In
this work, the atomic Hamiltonian is initially considered to be the Dirac-Coulomb (DC) Hamiltonian and correction terms from the Breit and vacuum polarization (VP) interactions are subsequently incorporated. Calculations are first performed at the singles and doubles (RCCSD) approximation within the relativistic coupled-cluster theory, followed by more comprehensive calculations including singles, doubles, and triples (RCCSDT method) starting from the Dirac-Hartree-Fock (DHF) method. Differences between the RCCSD and RCCSDT methods would highlight the significance of triple excitations in enhancing the accuracy of the results. A detailed description of these methods can be found in Refs. [15, 16, 17, 18]. Due to computational constraints, orbital excitations are usually restricted to orbitals with lower orbital angular momentum () values. However, it is later demonstrated that orbitals with contribute significantly to the energy calculations of Th. To illustrate this, initial calculations are performed using orbitals up to , followed by calculations including orbitals up to . The differences between these results are referred to as the “+Basis” contributions.
Our calculated energies for various states of Th, obtained at different level of approximations, are summarized in Table I. Notably, the energy of the ground state is estimated to be higher than that of the excited states at the DHF method; however, accurate trend is achieved after including correlation effects. As previously indicated, unusually large contributions from +Basis to the energies are observed, consistent with the findings reported in Ref. [19]. Breit interaction contributions to the energies of the ground state and its fine-structure partner are found to be significant, whereas VP contributions are moderate. A comparison between our final results and the experimental values from Refs. [20, 21] shows that our calculated energies agree within 1–2%, which could potentially be improved by incorporating higher-level excitations in our RCC calculations.
In Table I, we also present our calculated constants, as well as and values for the investigated states of Th at the same level of approximation used for the energy calculations. Notably, for the FS, the VP contributions exceed that from the Breit interaction and +Basis. We compare our computed values with those reported in Refs. [11, 12]. While Ref. [12] em
| Transition | IS [7] | This work (in fm) | ||
| (in MHz) | ||||
| −9856(10) | 0.316(19) | |||
| −10509(7) | 0.313(16) | 0.316(9) | ||
| −9890(9) | 0.317(16) | |||
| Method | ||||
| DHF | 202.96 | 105.16 | 332.44 | 121.80 |
| RCCSD | 225.44 | 80.96 | 430.12 | −30.80 |
| RCCSDT | 224.11 | 80.89 | 432.70 | −36.22 |
| +Basis | −0.56 | 0.34 | 1.04 | 0.68 |
| +Breit | 3.56 | 2.28 | 4.00 | 1.36 |
| +VP | 0.12 | −0.24 | 0.80 | −1.04 |
| +BW | −0.72 | 0.92 | −4.36 | 5.24 |
| Final | 226.5(4) | 84.2(8) | 434.2(28) | −30.0(67) |
| Ref. [9] | 229.2 | 86.1 | 431.5 | −36.7 |
| Ref. [28] | 230.53 | 87.12 | 431.85 | −23.31 |
| Ref. [19] | 224.8(1) | 83.6(12) | 425.2(48) | −41.6(88) |
| (Expt) [7] | 82.2(6) | 31.4(7) | 155.3(12) | −12.6(7) |
| (Expt) [8] | 82.0(2) | −12.9(3) | ||
| Extracted | 0.3620(11) | 0.3729(90) | 0.3576(36) | 0.430(96) |
| Recommended value in | ||||
| This work | 0.3618(12) | |||
| Others | 0.360(7) [9], 0.359(9) [28], 0.3663(60) [19] | |||
| 0.54 [24], 0.45(4) [26] |
Table 2. The weighted value inferred by combining our IS factors with IS measurements from Ref. [7].
ploys a configuration interaction (CI) method, Ref. [11] utilizes a combined approach of linearized RCCSD (SD) and CI (SD+CI) methods. Although all methods yield reasonably consistent results, our present calculations are more precise. The differential FS constants for the , , and states relative to the ground state are 2.4(19), 32.3(18), and 34.6(18) GHz/fm, respectively, against the reported values of 2(2), 33(8), and 35(8) GHz/fm in Ref. [10] that were obtained by accounting for pair-correlation (PC) effects through Brückner orbitals. It is worth noting that our RCC methodology incorporates both PC and core-polarization (CP) effects to all-orders, alongside correlations among PC, CP, non-PC, and non-CP effects together. In the NMS and SMS constants, the most significant corrections originate from Breit interactions, followed by +Basis and VP effects.
Utilizing the calculated IS factors, the extracted rms values are summarized in Table II, with the weighted average (denoted by bar) from all considered transitions provided therein. When considering only the FS factors derived from our calculations, we obtain fm, which is consistent with the values reported in previous studies: 0.320(15) fm [11] and 0.299(15) fm [13]. The discrepancies are likely attributable to contributions from triple excitations. Including MS effects, our
estimate increases to 0.316(9) fm, representing approximately a 3% enhancement and highlighting the importance of the MS contributions. Additionally, by employing the ratio of isomeric to ISs reported in Ref. [22], we estimate fm.
It is well established that precise measurements of nuclear moments are crucial for understanding the nuclear structures of isotopes. These values serve as benchmarks for refining nuclear models and enable more reliable predictions of isomeric properties [12, 22, 23]. Although several studies have investigated the magnetic dipole (M1) moment () and electric quadrupole (E2) moment () of the Th nucleus, significant discrepancies remain between the predictions of nuclear theory [24, 25] and those derived from combined calculations and measurements of hyperfine structure constants [7, 9, 10, 19, 26, 27, 28]. In this work, we conduct an independent investigation employing our advanced RCC methods within the expectation value evaluation (EVE) framework.
We evaluate the ratios of the M1 hyperfine structure constants, () and , using the relation where is the nuclear spin, is the electronic angular momentum, and is the electronic component of the M1 hyperfine interaction as defined in Ref. [29]. The calculated values for the considered states are obtained using the DHF, RCCSD and RCCSDT methods. These calculations include corrections from +Basis, Breit, VP, and Bohr–Weisskopf (BW) effects, as summarized in Table III. The BW corrections are estimated following a methodology similar to that described in Ref. [30] and were found to be negligible. By combining the final ratios with precise experimental values [7, 8], we derive values from different states, which are listed in the table above. The averaged obtained from these results aligns well with previous studies, but with improved precision. Similarly, we evaluate the ratios of the E2 hyperfine constants () and given by where represents the electronic component of the E2 hyperfine interaction [29]. The computed values for various states from different theoretical approaches are presented in Table IV. Using measured values from Refs. [7, 8], we extract values for each state, which are also listed in the above table. Notably, the value derived from this work differs significantly from most previous atomic studies but shows good agreement with nuclear calculations. Figure 2 (a) and (Figure 2b) illustrate the predicted and values, respectively, highlighting the contrasting magnitudes reported across different works. Additionally, by employing the ratios of nuclear moments between the isomeric and ground states from Refs. [22, 23], we estimate and b for the isomeric state of Th.
| Method | ||||
| DHF | 534.8 | 571.7 | 611.2 | 648.4 |
| RCCSD | 806.1 | 897.1 | 777.1 | 913.5 |
| RCCSDT | 791.9 | 882.1 | 779.3 | 919.2 |
| +Basis | −1.3 | −1.2 | 2.1 | 2.1 |
| +Breit | −0.7 | 0.2 | −3.5 | −2.6 |
| +VP | 0.0 | 0.0 | −0.1 | 0.1 |
| Final | 790(13) | 881(15) | 778(10) | 919(11) |
| Ref. [9] | 725 | 809 | 738 | 873 |
| Ref. [10] | 740 | 860 | 690 | 860 |
| Ref. [28] | 783 | 879 | 758 | 897 |
| Ref. [19] | 729(10) | 822(13) | 728(6) | 869(11) |
| (Expt.) [7] | 2269(2) | 2550(12) | 2265(9) | 2694(7) |
| (Expt.) [8] | 2270.3(18) | 2695.7(19) | ||
| Extracted | 2.87(5) | 2.89(5) | 2.91(4) | 2.93(4) |
| Recommended value in | 2.91(3) | |||
| This work | 2.91(3) | |||
| Others | 3.11(6) [9], 2.80 [25], 3.11(16) [10] | |||
| 2.95(7) [28], 3.11(2) [19], 4.3(9) [26] |
Table IV. The calculated values (in MHz/b) for the ground and excited states of Th, obtained using the same methods as in the previous table, are listed. The weighted average for Th is then derived by combining these with the measured values. This results in values that agree with some studies but differ substantially from others.

Figure 2. Comparison of (a) and (b) values of Th from different works over the years.
Table 5. Our ab initio calculated ground state , and values (in a.u.) of Th along with a comparison to reported values from earlier studies.
Understanding the E2 moments () of the hyperfine levels, with , of the clock states of Th is essential for accurately estimating E2 shifts in the clock transitions depicted in Figure 1. Following Ref. [31], the E2 moment () of the level can be expressed as a perturbative expansion: , where the superscripts denote the order with respect to the hyperfine interaction. The relates directly to the atomic E2 moment , with the
E2 operator , given by
The part is expressed as
where s denotes atomic energies, is summed over the M1 () and E2 () components, respectively, and . In Table V, we present values (in atomic units, a.u.) for the state obtained through various approximations, along with comparisons to previously reported values [32, 33], highlighting notable differences. The value for the level of the ground state and the level of the isomeric state were found to be identical, both being 0.614(9) a.u.. However, the values for the level of the ground state are a.u. for and a.u. for . Similarly, for the isomeric level involved in the clock transition, the values are a.u. and a.u. for and , respectively.
Accurate determination of the electric dipole (E1) polarizability, , is essential for precise estimation of Stark and black-body radiation shifts in the clock transition. Previously, both the scalar () and tensor () components of for the ground state of Th were evaluated using the sum-over-states (SOS) approach. However, the SOS method inherently faces limitations in uniformly accounting for contributions from all intermediate states, often requiring a combination of approaches—particularly to incorporate effects from doubly excited and higher-lying configurations. To treat all intermediate states on an equal footing, we determine and through an ab initio methodology based on the expressions , where are the effective E1 operators as defined in Refs. [16, 17], and is the first-order perturbed wave function obtained by solving the inhomogeneous equation
with the E1 operator .
The ab initio values of and computed at various levels of approximation are summarized in Table V. These results are compared with previously reported SOS semi-empirical values. As evident from the table, our findings are in good agreement with existing data, but they are obtained with considerably higher precision. Although the differences between the RCCSD and RCCSDT results for both quantities are relatively small, we emphasize the importance of including contributions from orbitals with to achieve the desired accuracy. For example, the RCCSD result for , when incorporating orbitals up to -symmetry, is approximately 20.0 a.u., which is about 30% higher than the value reported in the table obtained using the RCCSD method.
BKS is supported by ANRF with grant no. CRG/2023/002558 and Department of Space, Government of India. All calculations were performed on the ParamVikram-1000 HPC cluster at Physical Research Laboratory (PRL), Ahmedabad, India.
References
References
- [1]S. Kraemer, J. Moens, M. Athanasakis-Kaklamanakis, S. Bara, K. Beeks, P. Chethri, K. Chrysalisidis, A. Claessens, T. E. Coccolios, J. G. M. Correia et al., Nature (London) 617, 706 (2023).
- [2]E. Peik and C. Tamm, Europhys. Lett. 61, 181 (2003).
- [3]E. Peik, T. Schumm, M. S. Safronova, A. Pálffy, J. Weitenberg, and P. G. Thirolf, Quantum Sci. Technol. 6, 024002 (2021).
- [4]C. J. Campbell, A. G. Radnaev, A. Kuzmich, V. A. Dzuba, V. V. Flambaum, and A. Derevianko, Phys. Rev. Lett. 108, 120802 (2012).
- [5]V. V. Flambaum, Phys. Rev. Lett. 97, 092502 (2006).
- [6]M. S. Safronova, D. Budker, D. DeMille, D. F. J. Kimball, A. Derevianko, and C. W. Clark, Rev. Mod. Phys. 90, 025008 (2018).
- [7]C. J. Campbell, A. G. Radnaev, and A. Kuzmich, Phys. Rev. Lett. 106, 223001 (2011).
- [8]G. Zitzer, J. Tiedau, Ch. E. Düllmann, M. V. Okhapkin, and E. Peik, Phys. Rev. A 111, L050802 (2025).
- [9]M. S. Safronova, U. I. Safronova, A. G. Radnaev, C. J. Campbell, and A. Kuzmich, Phys. Rev. A 88, 060501(R) (2013).
- [10]J. C. Berengut, V. A. Dzuba, V. V. Flambaum, and S. G. Porsev, Phys. Rev. Lett. 102, 210801 (2009).
- [11]V. A. Dzuba and V. V. Flambaum, Phys. Rev. Lett. 131, 263002 (2023).
- [12]Ran Si, Chaoan Shi, Nan Xue, Xiangjin Kong, Chongyang Chen, Bingsheng Tu, Yu-Gang Ma, Sci. China Phys. Mech. Astron. 68, 272011 (2025).
- [13]M.S. Safronova, S. G. Porsev, M.G. Kozlov, J. Thielking, M.V. Okhapkin, P. Glowacki, D.M. Meier, and E. Peik, Phys. Rev. Lett. 121, 213001 (2018).
- [14]W. H. King, Isotope shifts in atomic spectra, Springer Science & Business Media, (2013).
- [15]B. K. Sahoo, S. Blundell, A. V. Olyeinychenko, R. F. G. Ruiz, L. V. Skripnikov, and B. Ohayon, J. Phys. B 58, 042001 (2025).
- [16]A. Chakraborty and B. K. Sahoo, Phys. Rev. A 111, 042807 (2025).
- [17]A. Chakraborty and B. K. Sahoo, Phys. Rev. A 111, 062812 (2025).
- [18]Vaibhav Katyal, A. Chakraborty, B. K. Sahoo, Ben Ohayon, Chien-Yeah Seng, Mihail Gortcheitch, and John Behr, Phys. Rev. A 111, 042813 (2025).
- [19]S. G. Porsev, M. S. Safronova, and M. G. Kozlov, Phys. Rev. Lett. 127, 253001 (2021).
- [20]A. Kramida, Yu. Ralchenko, J. Reader, and NIST ASD Team, NIST Atomic Spectra Database, version 5.8, https://physics.nist.gov/asd (2020).
- [21]J. Blaise and J. Wyart, Selected constants, energy levels, and atomic spectra of actinides, https://www.lac.universite-paris-saclay.fr/Data/Database/.
- [22]J. Thielking, M. V. Okhapkin, P. Glowacki, D. M. Meier, L. von der Wense, B. Seiferle, C. E. Düllmann, P. G. Thirolf, and E. Peik, Nature 556, 321 (2018).
- [23]C. Zhang, Tian Ooi, Jacob S. Higgins, Jack F. Doyle, Lars von der Wense, Kjeld Beeks, Adrian Leitner, Georgy A. Kazakov, Peng Li, Peter G. Thirolf, Thorsten Schumm, and Jun Ye, Nature 633, 63 (2024).
- [24]R. R. Chashman, I. Ahmad, A. M. Friedman, and J. R. Erskine, Rev. Mod. Phys. 49, 833 (1977).
- [25]N. Minkov and A. Pálffy, Phys. Rev. Lett. 122, 162502 (2019).
- [26]S. Gerstenkorn, P. Luc, J. Verges, D. W. Engelekmeier, J. E. Gindler, and F. S. Tomkins, J. Phys. (Paris) 35, 483 (1974).
- [27]C. E. Bemis, F. K. McGowan, J. L. C. Ford, Jr., W. T. Milner, R. L. Robinson, P. H. Stelsen, G. A. Leander, and C. W. Reich, Phys. Scr. 38, 657 (1988).
- [28]F.-C. Li, H.-X. Qiao, Y.-B. Tang, and T.-Y. Shi, Phys. Rev. A 104, 062808 (2021).
- [29]C. Schwartz, Phys. Rev. 97, 380 (1955).
- [30]B. K. Sahoo, Phys. Rev. A 111, L060801 (2025).
- [31]A. Derevianko, Phys. Rev. A 93, 012503 (2016).
- [32]Shi-Cheng Yu, Wen-Ting Gan, Xia Hua, Xin Tong and Cheng-Bin Li, Phys. Rev. A 109, 063115 (2024).
- [33]M. S. Safronova and U. I. Safronova, Phys. Rev. A 87, 062509 (2013).