betaFeedback

WorksPhysics and Astronomy

A proposed solution for the lifetime puzzle of the 229mTh+ isomer

Abstract

With the example of the 229Th nucleus, which is the most likely candidate for the creation frequency standard of a future, the dynamics of the interplay and the relationship of various resonance conversion mechanisms is analyzed. As a result, a solution is proposed for the so-called thorium puzzle, which consisted of a contradiction between the experimental and theoretical lifetimes of Th+ ions. First, the solution demonstrates the dependence of the lifetime of the nuclear isomer on the ambient conditions. Second, it demonstrates the leveling role of the fragmentation of the single-electron levels, which makes the resonance amplification of the electron-nuclear interaction more likely. Both of these trends lead to a probable decrease of the theoretical lifetime towards agreement with experiment.

Introduction

229Th^{229}\mathrm{Th} is a nucleus with a uniquely narrow doublet of the ground and isomeric first excited level, split by only ∼8 eV\sim8\,\mathrm{eV} [1, 2, 3, 4, 5] and Refs. cited therein. For this reason, its observation and measurement was a longstanding problem. Whereas the first suggestions of its existence appeared decades ago [6], it was only in [7] that its conversion decay was directly detected for the first time. The measured lifetime of about 10 μs10\,\mu\mathrm{s} in neutral atoms [8] precisely coincided with the predicted value [9]. The isomer undergoes deexcitation by a radiative M1M1 transition, with its own half-life of about 2 h, which corresponds to a radiative width of ∼10−20 eV\sim10^{-20}\,\mathrm{eV}. Two years after the first observation, the reality of the isomer was confirmed by direct detection of its hyperfine structure [10].

229Th{}^{229}\mathrm{Th} has the lowest isomer energy amongst the known nuclei. It is the most likely candidate for the frequency reference and nuclear-optical clock of the next generation [11]. For a long time, its energy was considered to be even less, about 3.5 eV. At such an energy, the nuclear level is entangled with many electronic levels with the same multipolarity. This means that we never can observe photons exclusively originating from its deexcitation. A variety of cooperative nuclear-atomic processes takes place instead: from internal conversion (IC) in neutral atoms to discrete, or resonance conversion and electronic bridges in singly and doubly charged ions. It is didactic that in Ref. [12] the authors allegedly reported observation of the nuclear radiation from deexcitation of the isomeric state. Of course, that was impossible, even with the isomer energy of 3.5 eV, as it was commonly accepted at that time. At this energy, deexcitation through the electronic bridges brought about by resonance conversion was the absolutely dominating channel. The discrete conversion RR factor which plays the same role as internal conversion coefficient (ICC) in traditional IC, comprises R=1600R = 1600 [13, 14, 15]. Indeed, the effect of [12] was then attributed to the presence of extraneous α\alpha particles [16, 17].

It was already noted in [7] that there was no observation of any IC signal for the singly charged ions Th+\mathrm{Th}^{+}, whereas quite a distinct signal from doubly charged Th++\mathrm{Th}^{++} ions was detected. The absence of the Th+\mathrm{Th}^{+} signal looked unusual. At the time this finding was tentatively attributed to the small amount of collected 229Th+{}^{229}\mathrm{Th}^{+} ions. This fact was studied in more detail in Ref. [8]. In that paper the problem of a short isomeric lifetime in Th+\mathrm{Th}^{+} was raised for the first time. In Refs. [7, 8], recoil ions of 229Th{}^{229}\mathrm{Th} from the α\alpha decay of 233U{}^{233}\mathrm{U} were collected in the stopping cell, and then transported through the nozzle, RPQ and mass separator to the final microchannel plate detector. The measurements were performed with doubly- and triply charged ions. As a result, certain limitations on the lifetimes in the singly- and doubly-charged ions were retrieved. Data [7, 8] comprised a topic for theoretical investigations [18, 19]. The thorium puzzle received a further examination by Seiferle [20], where further experimental details substantiating the problem can be found. Data from [7, 8, 20] became the starting point of appearance of the thorium puzzle, actively discussed since then.

We consider this question in more detail in the next sections. Various ways of possible solutions are pursued, such as the relation between the structural features of the electronic wavefunction in the ground state and the isomer lifetime, a random resonance, and the most intriguing reason — a possible manifestation of the dependence on the ambient conditions. We show that fragmentation of the atomic wavefunctions is of great importance for the realization of each of the mechanisms. Note that it was already in papers [21, 22] that the decisive role of the electronic level fragmentation in the resonance state had been examined and proven soon after its discovery in the case of RC in the 35.49 keV M1M1 transition in 125Te45+{}^{125}\mathrm{Te}^{45+} ions [23]. The resonance conversion was discovered in this transition at the University of Bordeaux, thanks to a collaboration with the authors of the present investigation, by means of a study of the dependence of the nuclear isomer lifetime on stripping the electron shell. The important role of fragmentation was also considered in Ref. [19], assuming that the isomer energy is 7.6 eV. Herein we advance and refine this approach, in the light of the latest values of the isomer energy as indicated above. We will see that the energy refinement is of crucial importance to solving the puzzle.

The thorium puzzle and ways of its probable solution

In Ref. [8, 20], along with neutral thorium atoms, restrictions on the isomer lifetimes in singly- and doubly charged atoms were obtained. The halflife in any charged state of 229Th{}^{229}\mathrm{Th} but singly charged one turned out to be longer than one minute, as this was the maximum survival time of thorium ions under the achievable experimental conditions. On top of that comes the non-observation of an IC signal in singly charged Th+\mathrm{Th}^{+} ions. Under experimental conditions of Refs. [8, 20], the buffer-gas stopping cell was less than 10 ms. So the non-observation of the IC signal was attributed in [8] as pointing out to a potential reduced halflife of Th+\mathrm{Th}^{+} ions. The latter was thus restricted by the experimental extraction time to 0.01 s as the halflife limit T1/2+<0.01T_{1/2}^{+} < 0.01 s given for Th++\mathrm{Th}^{++}; Th+++\mathrm{Th}^{+++} for more than one minute. This gave basis for separate speculations that the isomer energy might be more than 11.9 eV, which opened the IC channel. But the latter measurements mentioned above make this supposition unlikely. This makes the case of singly charged ions most intriguing. Calculation [19] results in a lifetime of not less than ∼1\sim1 s with a resonance conversion factor of R≈5900R \approx5900. A question arises how one can explain the difference of two orders of magnitude. This situation got the name of thorium puzzle ([20]).

This situation has an explanation based on a physical ground. But before proceeding with detailed consideration, we note that the thorium puzzle arose even before the publication of Ref. [19], in attempts to extract the T1/2+T_{1/2}^{+} from Ref. [24]. Moreover, the discrepancy obtained appeared to be even two orders of magnitude greater than that indicated above. We remind that in Ref. [24], the authors calculated the probability of excitation of the nuclear isomer by means of a two-photon absorption through an electron shell in singly charged thorium ions, see Figure 1a. The valence electronic configuration in the ground state was assumed to be 7s (6d)27s\ (6d)^2. The first of the absorbed photons promotes an electron from the 6d6d- to the 7p7p state, transferring the atom as a whole to the excited level (7s 6d 7p)J=5/2(7s\ 6d\ 7p)_{J=5/2} with an energy of 3.084 eV. Then the same electron absorbs the second photon, appearing in the second intermediate state, from where it transfers to the nucleus already the total energy of both photons, itself returning to the ground state 6d6d. The disadvantages of this scheme were considered in [25, 26], and herein we will no longer dwell on them. Note, however, that, firstly, the second intermediate state was not specified in Ref. [24]; let us call it xx. In fact, in a situation close to a resonant one, quantum numbers of the nearest 8s8s or 7d7d levels can be assigned to it. Then it is easy to see that this scheme excludes the most efficient resonant electron transition 8s−7s8s-7s for pumping the nucleus [27]. This underestimates the efficiency of the entire circuit by up to three orders of magnitude: cf. the above work [12], where it was not taken into account that there is a decay through the 7s−8s7s-8s electronic channel, with the discrete conversion factor R≈1600R \approx1600 expected to be valid instead.

Figure 1. Feynman graphs of: left: two-photon scheme of isomer pumping [24]; middle and right: amplitude of the second photon absorption, calculated in [24], and its reversal, which can be allegedly used for a calculation of the isomer lifetime, respectively. The electronic configurations together with a simplified classification of the single-electronic states for pointing out the electronic circuits are indicated.

Secondly, only a part of the transition amplitude was actually calculated, starting with the second photon, namely, that of the 7p−x−6d7p-x-6d circuit, with the atom remaining in the ground state in the end. This is illustrated in Figure 1b. It may be due to this error, that the calculated amplitude can be used in order to calculate the lifetime of the isomer in singly charged ions—in view of the TT-invariance of quantum electrodynamics, as shown in Figure 1c. This erroneous judgment gave rise to a thorium riddle. The mistake is that the electron bridge in Figure 1c is initiated by the resonance conversion, throwing up the 6d6d electron to the excited xx state (8s8s or 7d7d), from where it then transfers to the 7p7p state. This mechanism, however, does not represent the isomer lifetime, for much stronger conversion transitions are presented by the 7s−8s−7p7s-8s-7p electronic circuits [13], and these transitions essentially remain outside the scope. As a result, our RR factor approximately equals the coefficient GG defined in Ref. [24]. Nevertheless, it is the β\beta coefficient of Ref. [24], which is analogue of the RR factor. But the latter is approximately twenty times less:

R≈β≈120G2.(1)R \approx\beta\approx\frac{1}{20}G_{2}. \tag*{(1)}

This leads to the fact that, at an isomer energy of 7.87.8 eV, the divergence of paper [24] with the experiment turns out to be at least 20 times stronger than the one given above. Thus, a consistent calculation as in [19] significantly reduced the severity of the thorium puzzle, though did not completely solve it yet. There was still a gap with the experiment of two orders of magnitude.

Dependence of the nuclear lifetime on the ambient conditions as an argument in solving the thorium puzzle

The next point is that nuclei, decaying via resonance conversion, turn out to be vulnerable to the ambient conditions. Indeed, compare the case with traditional IC, when a bound electron, having received energy from the nucleus, leaves the atom. In contrast, in the case of resonance conversion, the conversion electron remains in the atom, sharing the obtained energy of the nuclear transition with the other electrons. In this case, the further electron fate evidently depends on the impact of the environment. First of all, this is collisional broadening of the electron state. Let us examine the question in more detail.

In fact, the first 0.01 s fresh 229mTh+{}^{229\mathrm{m}}\mathrm{Th}^{+} ions occurring from the alpha decay of 233U{}^{233}\mathrm{U} spend in a funnel in a buffer gas atmosphere at a pressure of about 40 mbar. Broadening from the collisions with the buffer gas molecules is proportional to the concentration of the gas, as well as the pressure. Bearing in mind that the collisional broadening achieves two orders of magnitude at normal conditions, one can estimate that at the pressure of 30 to 40 mbar the broadening may be as large as of a factor of 3. Hence, in this way, from the initial four orders of magnitude, the paradox essentially decreases to an order of magnitude. In turn, this remaining discrepancy can be considered as a result of random resonance coincidence of the isomer energy with one of the intermediate electronic states [28]. Moreover, as analysis shows, such an enhancement of the transition probability by an order of magnitude is not an unlikely event at all.

The isomer undergoes deexcitation through a great deal of the electronic bridge transitions like 7s−8s−7p7s-8s-7p, with a number of intermediate 8s8s and final 7p7p states, brought about by the interaction of the electronic configurations. In these circuits, the first step — lifting up the 7s7s electron to the 8s8s position — is nothing but the bound, or resonance conversion (Figure 2a). Because of the defect of resonance Δ\Delta, this state cannot be the final state of the transition. Emitting a photon, the atom makes the final transition to one of the 7p7p states, restoring the energy conservation. At a nuclear transition energy ωn\omega_n, the RR factor is hence comprised by the sum of contributions of hundreds of intermediate states:

Figure 2. Scheme of the resonance electronic bridge deexcitation of the 229mTh{}^{229\mathrm{m}}\mathrm{Th} isomer: left: as anticipated in the absence of fragmentation; right: with allowance for the fragmentation in the intermediate states.

R(τL;ωn)=∑lαd(l)(τL)Γl/2πΔl2+(Γl/2)2≡∑lRl.(2)R(\tau L;\omega_n)=\sum_l\frac{\alpha_d^{(l)}(\tau L)\Gamma_l/2\pi}{\Delta_l^2+(\Gamma_l/2)^2}\equiv\sum_l R_l. \tag*{(2)}

In (2), αd(l)(τL)\alpha_d^{(l)}(\tau L) is the analogue of the ICC, τL\tau L indicate the type and multipole order of the nuclear transition. The αd(l)(τL)\alpha_d^{(l)}(\tau L) values are practically independent of the nuclear model, similar to the ICC. However, they acquire the dimension of an energy in the case of bound (resonance) conversion. Therefore, they lose the sense of ICC, but they regain another meaning of residues in the sum (2) [19]. We return to this issue in § 4. The αd(l)\alpha_d^{(l)} values are given by the squares of the conversion matrix elements between the initial 7s7s and final 8s8s electronic wavefunctions (e.g. [23]). Γl\Gamma_l is the full width of the intermediate state ll, involving the atomic and nuclear widths, and Δl=ωn−ϵl\Delta_l = \omega_n - \epsilon_l is the defect of the resonance between the electronic and nuclear transitions (e.g., [19, 23]). The Γl\Gamma_l values are typically around 10−8−10−710^{-8} - 10^{-7} eV. For this reason, for a randomly chosen ωn\omega_n, as a rule, one meets the condition Δl≫Γl\Delta_l \gg\Gamma_l. The addition of Γl\Gamma_l in the denominator in (2) is important just near resonances ll, where the RR factor differs considerably from its regular value.

As stated above, in the case of resonance conversion the RR factors from Eq. (2) play the same role as ICC — in traditional IC. That is, the probability of the isomer deexcitation per unit of time through the electronic bridge in Figure 2a, induced by resonance conversion, is

Γc(τL)=RΓγ(τL),(3)\Gamma_c(\tau_L) = R\Gamma_\gamma(\tau_L), \tag*{(3)}

where Γγ\Gamma_\gamma is the proper isomeric radiative width.

Herein we assume that the expected energy of the isomer takes a random value within the range of Δωn=0.4\Delta\omega_n = 0.4 eV around 8.28.2 eV, ωn∈[8.0,8.4]\omega_n \in[8.0, 8.4] eV. This energy interval is in agreement with the measurements of the isomer energy, cited previously. And we are thus interested in the probability P(Rc)P(R_c) such that the RR factor must be at least RcR_c or more, where RcR_c may considerably exceed the regular value on the interval due to a near coincidence with an intermediate electronic level ll. This imposes a condition on the Δl\Delta_l value. Provided that R≥RcR \ge R_c, Rc≫1R_c \gg1, one finds by means of (2) that

P(Rc)=2Δl/Δωn=2αdΓπRc/Δωn.(4)P(R_c) = 2\Delta_l/\Delta\omega_n = \sqrt{\frac{2\alpha_d\Gamma}{\pi R_c}}/\Delta\omega_n . \tag*{(4)}

It immediately follows from Eq. (2), that an increase of the atomic decay width Γl\Gamma_l by a factor of KK causes the same increase of the RR factor and the related shortening of the nuclear lifetime also by the same factor of KK times. This is on one hand. On the other hand, in a situation when the isomer energy is not precisely known, the probability of a random enhancement of the RR factor due to coincidence with an intermediate electronic level increases simultaneously, but more slowly. According to (4), it increases proportionally to the square root: K\sqrt{K} times.

Influence of fragmentation on the RR factor of resonance conversion: qualitative consideration

In our case, we are in a position that the transition energy is unknown on the level of the required accuracy. Let us assess a scatter of the RR values which can be expected in such a situation. Our method will be similar to that used for treating the dynamic enhancement of the effect of spatial parity nonconservation in the compound nuclei due to fragmentation of the nucleon wave functions in Ref. [29]. Consider a schematic model, similar to the picket-fence model [30].

Our calculations were performed within the framework of the multiconfiguration Dirac-Fock (MCDF) method, taking into account the interaction of the electronic configurations, as in Ref. [19]. Concerning the electron configuration of the ground state in Th+\mathrm{Th}^{+}, different information can be found in the literature. One possibility is

I:7s(6d3/2)2J=3/2(5)I:\qquad7s(6d_{3/2})^{2}\qquad J=3/2 \tag*{(5)}

(e.g., Ref. [31]). The other is

II:(7s)26d3/2J=3/2(6)II:\qquad(7s)^{2}6d_{3/2}\qquad J=3/2 \tag*{(6)}

(e.g., Ref. [32]) Our MCDF calculation shows that the ground state is structured according to Eq. (5). Nevertheless, these two single-electron states are practically degenerate, so that there is a strong admixture of the configuration (6) in the ground state. Therefore, in the main approximation, bearing in mind configuration II, one can put down the ground-state wavefunction as follows:

ψ0∼7s(6d)J=3/22.(7)\psi_{0} \sim7s(6d)^{2}_{J=3/2}. \tag*{(7)}

Resonance conversion transition transfers the 7s7s electron to the intermediate 8s8s state. For simplicity, suppose that taking into account of the interelectronic interaction splits the 8s8s state into two levels. Then their wavefunctions may be expressed as follows:

ψin∼12(α∣8s(6d)2⟩+β∣7s8s6d⟩)J=3/2,(8)\psi_{\mathrm{in}} \sim\frac{1}{\sqrt{2}}\left(\alpha\lvert8s(6d)^{2}\rangle+\beta\lvert7s8s6d\rangle\right)_{J=3/2}, \tag*{(8)}

with ∣α∣≈∣β∣≈1\lvert\alpha\rvert\approx\lvert\beta\rvert\approx1. If one uses wavefunction II in the ground state, then the first term in Eq. (8) gives the predominant contribution to the 7s→8s7s \to8s transition. Oppositely, if wavefunction IIII is used in the ground state, then the main contribution will be due to the second term in Eq. (8). Anyway, the αd\alpha_d value, which is proportional to the square of the wavefunction ψin\psi_{\mathrm{in}} in (8), becomes now αd′\alpha'_d, which is twice smaller for each level:

αd′∼αd2,(9)\alpha'_d \sim\frac{\alpha_d}{2}, \tag*{(9)}

as only the first term on the right side of (8) gives a significant contribution. In contrast, the width of the state (8) changes weaker, as both terms contribute. Taking into account that now we have two levels in the interval, the probability (4) becomes approximately

P′(Rc)=22αd′ΓπRc/Δωn=2P(Rc),(10)P'(R_c)=2\sqrt{\frac{2\alpha'_d\Gamma}{\pi R_c}}/\Delta\omega_n=\sqrt{2}P(R_c), \tag*{(10)}

that is by a factor of 2\sqrt{2} higher. Correspondingly, if one assumes splitting into NN levels, an amplification of P′(Rc)P'(R_c) by N\sqrt{N} times will be expected.

Furthermore, one can compare what is said above with the results obtained within the conventional Dirac—Fock method with no account of the configuration mixing, making use of the ground-state configuration of type II. In this case, there is one intermediate 8s8s state, together with two final states 7p1/27p_{1/2} and 7p3/27p_{3/2}. The calculated αd(M1;ωn)\alpha_d(M1;\omega_n) value with an isomer energy ωn=8.2 eV\omega_n = 8.2\ \mathrm{eV} is αd=5.6×109 eV\alpha_d = 5.6 \times10^9\ \mathrm{eV}, the radiative width of the intermediate 8s8s state Γ=1.32×10−7 eV\Gamma= 1.32 \times10^{-7}\ \mathrm{eV}, the 7s−8s7s - 8s transition energy being of 6.04 eV6.04\ \mathrm{eV}. Thus, the resonance defect would be tremendous in comparison with the width, making the resonance interaction unlikely. The fragmentation spreads the resonance strength far from the primary non-fragmented line to the inter-electronic-shell area. Note that αd\alpha_d have an important physical meaning of the residues of R(ωn)R(\omega_n), considered as an analytical function in the lower half-plane of the physical sheet of the nuclear transition energy:

−2πiRes⁡R(ωn)∣ωn=Ea(l)=αd(l)(ωn).(11)-2\pi i \operatorname{Res} R(\omega_n)\bigg|_{\omega_n=E_a^{(l)}} = \alpha_d^{(l)}(\omega_n). \tag*{(11)}

Fragmentation of the 8s8s level, accounting for a mixture of configurations, leads to an increase in the number of poles and to their spread over the energy region, accompanied with an adequate decrease in the values of the residues, while the total value of the residues approximately holds:

∑lαd(l)(ωn)=αd(ωn).(12)\sum_l \alpha_d^{(l)}(\omega_n) = \alpha_d(\omega_n). \tag*{(12)}

Let us define RmeanR_{\mathrm{mean}} as the weighted average [30]:

Rmean=∫ρ(En−λ)R(λ) dλ.(13)R_{\mathrm{mean}} = \int\rho(E_n-\lambda)R(\lambda)\,d\lambda. \tag*{(13)}

Here ρ(x)\rho(x) is the averaging function. It is peaked around x=0x = 0, and falls off for large ∣x∣\lvert x\rvert. It is convenient to chose it as follows:

ρ(x)=Δ/2πx2+(Δ/2)2,(14)\rho(x) = \frac{\Delta/2\pi}{x^2+(\Delta/2)^2}, \tag*{(14)}

with Δ\Delta characterizing the energy interval around ωn\omega_n over which the averaging is taken, ωn=8.2 eV\omega_n = 8.2\ \mathrm{eV}, and Δ=0.2 eV\Delta= 0.2\ \mathrm{eV} in our case. The weight function is normalized at unity:

∫ρ(x) dx=1.(15)\int\rho(x)\,dx = 1. \tag*{(15)}

Inserting (14) into (13), neglecting in first approximation the dependence of αd(l)(τL;ωn)Γl\alpha_d^{(l)}(\tau L;\omega_n)\Gamma_l on the nuclear energy, and making use of (11), one arrives at the resulting expression:

Rmean=∑lαd(l)(τL;ωn)(Γl+Δ)/2πΔl2+[(Γl+Δ)/2]2=∑lαd(l)(τL;ωn)Δ/2π(ωl−ωn)2+(Δ/2)2.(16)R_{\mathrm{mean}} = \sum_l \frac{\alpha_d^{(l)}(\tau L;\omega_n)(\Gamma_l+\Delta)/2\pi}{\Delta_l^2+[(\Gamma_l+\Delta)/2]^2} = \sum_l \frac{\alpha_d^{(l)}(\tau L;\omega_n)\Delta/2\pi}{(\omega_l-\omega_n)^2+(\Delta/2)^2}. \tag*{(16)}

The last equality is derived taking into account that Δ≫Γl\Delta\gg\Gamma_l.

Results of the calculation

Bearing in mind the proper halflife of the isomer of approximately 2 h, we conclude that either Rn=2.4×105R_n=2.4\times10^5 or Ra=7.2×105R_a=7.2\times10^5 is needed for agreement with the experimental halflife of 0.01 s, depending on whether the influence of the ambient conditions is taken or not taken into account. Passing to the calculation results, in Figure 3a we present the RR values as calculated against the isomer energy ωn\omega_n, based on the ground state II. The RR values vary drastically on the interval, within well ten orders of magnitude in the peaks. This is very different from the interval around 7.6 eV considered in [19]. To a large degree, the behavior of the plot R(ωn)R(\omega_n) is due to the presence of the strong transitions into the intermediate states 8.148 eV, J=3/2J=3/2 (αd(l)=2.39×109\alpha_d^{(l)}=2.39\times10^9, Rl=3.75×105R_l=3.75\times10^5), and 8.211 eV, J=3/2J=3/2 (αd(l)=3.33×107\alpha_d^{(l)}=3.33\times10^7, Rl=3.15×104R_l=3.15\times10^4). The first state consists of the 8s component by 82%. The second has a more miscellaneous composition. For the energy of ωn=8.2\omega_n=8.2 eV the resulting RR value of R=412066R=412066 is obtained. This turns out to be nearly enough: R≳RaR\gtrsim R_a, in order to explain the observed lifetime.

Calculated R factor of resonance conversion against the nuclear transition energy

Figure 3. Calculated RR factor of resonance conversion against the nuclear transition energy from the isomer to the ground state: left: calculation with the electronic ground-state wavefunction II (5); right: the same with the wavefunction IIII (6).

However, for the other possible values of the isomer energy on the interval, the RR factor can be either much more or less, down to Rmin⁡=7.24×104R_{\min}=7.24 \times10^{4} and 2.11×1042.11 \times10^{4} for the isomer energies of ωn=8.02\omega_n=8.02 and 8.48.4 eV, respectively, on the wings of the interval. In the former case, the RR value turns out to be in agreement with the data. This should not be taken as an argument in favor of the fact that the isomer energy is rather 8.28.2 eV or so. As it is said previously, experimental lifetime can be also explained as far as R>RnR>R_n, if a dependence on the ambient conditions is taken into account.

The calculations were also performed with the ground state (6), corresponding to the configuration IIII. The resulting plot of the RR factor against ωn\omega_n is presented in Figure 3b. These values agree worse with experiment on the interval under consideration. Thus, R=67859R=67859 for ωn=8.2\omega_n=8.2 eV, that is six times less. The peak created by the strongest transitions is shifted towards smaller values of the transition energies.

Continuing our analysis, we note, that the mean RmeanR_{\mathrm{mean}} value on the interval as obtained by means of (16) turns out to be Rmean=4.6×108R_{\mathrm{mean}}=4.6 \times10^{8} and 1.6×1091.6 \times10^{9} for the cases II and IIII, respectively. At first glance, such big values may conflict with the other values of R=412066R=412066 or 6785967859, calculated above at ωn=8.2\omega_n=8.2 eV. However, this example teaches once again how the average values may not be indicative, while they are mainly achieved within the very narrow vicinities of extremely sharp resonances. Actual values are greatly scattered around the mean value, between minimum Rmin⁡=10197R_{\min}=10197 and maximal one Rmax⁡=1.59×1014R_{\max}=1.59 \times10^{14} at ωn=8.48\omega_n=8.48 and 8.148.14 eV, respectively. For comparison, the minimal non-fragmented value is much less: Rmin⁡(nf)=21R_{\min}^{(nf)}=21. Correspondingly, the average non-fragmented value is also much less: Rmean(nf)=25R_{\mathrm{mean}}^{(nf)}=25. Hence, there could be no question of agreement between theory and experiment in the absence of fragmentation.

Now one can determine the probability P(Rc)P(R_c) that the RR factor will be R>RcR > R_c under condition of random distribution of the isomer energy over the chosen interval Δωn=0.4\Delta\omega_n = 0.4 eV around 8.28.2 eV as follows:

P(R)∣R>Rc=1Δωn∫(Δωn)H(R(ω)−Rc) dω(17)\left.P(R)\right|_{R>R_c} = \frac{1}{\Delta\omega_n} \int_{(\Delta\omega_n)} H\left(R(\omega)-R_c\right)\,\mathrm{d}\omega \tag*{(17)}

where H(x)H(x) is the Heaviside step function:

H(x)={1for x≥00for x<0(18)H(x) = \begin{cases} 1 & \text{for } x \ge0 \\ 0 & \text{for } x < 0 \end{cases} \tag*{(18)}

Integration in Eq. (17) is performed over the assumed domain of the nuclear energy Δωn\Delta\omega_n. At the minimum values, P(Rmin⁡)=100%P(R_{\min}) = 100\%.

In Figure 4 we present a plot of the calculated P(Rc)P(R_c) values for both initial electronic configurations. In the case of the ground configuration II, the corresponding probabilities comprise 41 or 24 percent for Rc=2.4×105R_c = 2.4 \times10^5 or Rc=7.2×105R_c = 7.2 \times10^5, respectively, if the influence of the ambient conditions is taken into account or is not. Thus, the calculated probability of a random enhancement of the RR factor by an order of magnitude, needed for agreement between theory and experiment, attains 41 percent, that is comparable with unity. If the influence of the ambient conditions were neglected, the probability of the agreement remained great enough: 24 percent, which remains to be quite a significant quantity, comparable with unity. For comparison, in the case of the ground configuration IIII, corresponding probabilities would comprise 25 and 14 percent, respectively.

Figure 4. Probability P(Rc)P(R_c), in percent, that the RR factor will exceed RcR_c under random choice of the isomer energy ωn\omega_n within the interval 8.0 – 8.4 eV. Full (black) curve — calculation with the ground-state wavefunction II (5), broken (red) curve — with the wavefunction IIII (6).

At the same time, one should not forget about consequences of fragmentation in the ground state. If it is fragmented into NgN_g states, the weight of the 7s7s state is also disseminated over them. Therefore, the interaction force is also diminished by NgN_g times. Partly, it is this effect which is responsible for diminution of the mean and minimal values of the RR factor.

Conclusion

The above analysis shows that in fact the experimental upper bound of the isomeric lifetime in the singly charged ions is at least two orders of magnitude smaller than regular theoretical predictions [19]. This gave rise to the thorium puzzle. A possible solution that the isomer energy might be more than 11.9 eV is not supported by the latest measurements.

Another possible solution assumes a random coincidence of the isomer and one of the electronic levels. Calculations [19] and the above analysis demonstrate that such a coincidence can be expected with the probability of 24 to 41 percent. This is not a small probability. One can say that it may reconcile theory with the experiment. A significant role in the formation of such a probability belongs to the fragmentation of the electronic levels, caused by the configuration mixing. First of all, the fragmentation leads to spreading of the resonance strength over the energy domains between the electronic shells. This allows the resonance condition to be met with higher probability. Second, the fragmentation increases the probability of fluctuations of the RR-factor, as the qualitative consideration in § 4 shows. The picture is completely similar to the case of resonance conversion in the 35.49-keV 125Te45+{}^{125}\mathrm{Te}^{45+} ions [21]. In that case, fragmentation led to the splitting of each of the electronic levels for the conversion electron essentially into two ones, which induced resonance conversion in full scale. For higher ions, the electronic level density decreased, rapidly ruling out the resonance conversion. A similar picture is observed in Th ions: ionization potential in Th++\mathrm{Th}^{++} is 18.3 eV — 6.2 eV higher than that in Th+\mathrm{Th}^{+} [32]. This decreases the electronic level density, making the lifetime longer and giving rise to the thorium puzzle.

Furthermore, an important role belongs to the collisional broadening of the intermediate levels formed in resonance conversion. It is remarkable that it brings about the dependence of the nuclear lifetime on the ambient conditions. This dependence manifests itself also in two respects. First, it may increase the observed value of the RR factor, up to two orders of magnitude at normal conditions, specifically, by several times — under the conditions of experiment [7]. Second, it further increases the probability of random fluctuations of the RR factor. As a result, allowance for the collisional broadening increases the probability of a random coincidence, needed for agreement with experiment, to 41 percent.

The dependence on the ambient conditions stands in the same row with the other effects, such as laser assistance of the isomer decay [33], or mixing of the nuclear levels with different spins [34], yielding acceleration of the isomer decay by factors of few hundred. Each of these effects offer ways of manipulating the nuclear processes, affecting the electron shell.

It is worthy of dwelling on paper [34], where an extreme case of hydrogen-like ions has been considered. Broadening of the isomer state by a factor of seven hundred was predicted due to the interaction between the components of hyperfine structure, formed by vector coupling of the electronic angular momentum j=s=1/2j=s=1/2 to the nuclear spins in the ground and isomeric states. Meanwhile, such an electronic configuration: 1S1/2{}^{1}S_{1/2} in triply charged ions is considered as the most suitable for creation of the nuclear-optical clock in the pioneering work [11]. Estimates fairly similar to [34] show that one should beware of broadening the isomeric line in this configuration by about 100 times. At the same time, a method of aligned states in triply charged thorium ions, based on the valence 5f5f electrons [35], is less susceptible to the broadening. This is because the M1M1 ICC for the 5f5f electrons are by three orders of magnitude smaller than those for the electrons of the 7s7s shell. For this reason, the construction of a nuclear-optical frequency standard on triply charged ions looks more promising. The dependence of the nuclear lifetime on the external conditions demonstrated above should not be overlooked.

The authors would like to express their gratitude to L. von der Wense for inducing discussions. They are also grateful to M. Pfützner and E. Peik for helpful remarks.

References

  1. [1]B. R. Beck et al., Phys. Rev. Lett. 98, 142501 (2007).
  2. [2]B. R. Beck et al. LLNL–PROC–415170. Berkeley: 2009.
  3. [3]B. Seiferle, L. von der Wense, P. V. Bilous et al., Nature 573, 243 (2019).
  4. [4]Yamaguchi et al., Phys. Rev. Lett. 123, 222501 (2019).
  5. [5]T. Sikorsky, J. Geist et al., Phys. Rev. Lett. 125, 142503 (2020).
  6. [6]Kroger, L.A., Reich, C.W., Nucl. Phys. A 259, 29 (1976).
  7. [7]L. von der Wense, B. Seiferle, M. Laatiaoui et al., Nature 47, 533 (2016);
  8. [8]B. Seiferle, L. von der Wense, and P. G. Thirolf, Phys. Rev. Lett. 118, 042501 (2017).
  9. [9]F. F. Karpeshin, M. B. Trzhaskovskaya, Phys. Rev. C 76, 054313 (2007).
  10. [10]J. Thielking, M. V. Okhapkin, P. Glowacki, D. M. Meier, L. von der Wense, B. Seiferle, C. E. Duellmann, P. G. Thirolf and E. Peik, Nature 556, 321 (2018), https://doi.org/10.1038/s41586-018-0011-8 .
  11. [11]E. Peik and Chr. Tamm, Europhys. Lett. 61, 181 (2003).
  12. [12]G. M. Irwin and K. H. Kim, Phys. Rev. Lett. 79, 990 (1997).
  13. [13]F.F.Karpeshin, I.M.Band and M.B.Trzhaskovskaya, in: Nuclear Shapes and Nuclear Structure at Low Excitation Energies. Antibes (France) 20 – 25 June, 1994. Abstracts of the Contributed Papers, p. 50; Proceedings of the International Conference, P. 181.
  14. [14]F. F. Karpeshin, I. M. Band, M. B. Trzhaskovskaya, and M. A. Listengarten, Phys. Lett. B 372, 1 (1996).
  15. [15]F. F. Karpeshin, I. M. Band, M. B. Trzhaskovskaya, and A. Pastor, Phys. Rev. Lett. 83, 1072 (1999); Izv. RAN, Ser. Fiz. 63, 38 (1999).
  16. [16]R. W. Shaw et al., Phys. Rev. Lett. 82, 1109 (1999).
  17. [17]S. B. Utter et al., Phys. Rev. Lett. 82, 505 (1999).
  18. [18]R.A. Müller, A.V. Volotka, S. Fritzsche, A. Surzhykov. Nucl. Instr. Meth. in Physics Research, B, (2017).
  19. [19]F. F. Karpeshin, M.B.Trzhaskovskaya, L.F.Vitushkin, Arxive: 1701.05340, 2017; F. F. Karpeshin, M. B. Trzhaskovskaya, Nucl. Phys. A 969, 173 (2018).
  20. [20]Benedict Seiferle. Characterization of the 229Th nuclear clock transition. PHD thesis, Ludwig Maximilians Universität, München, 2019.DOI
  21. [21]F.F.Karpeshin, I.M.Band and M.B. Trzhaskovskaya, Zh. Eksp. Teor. Fiz., 116, 1565 (1999) Engl. transl. JETP, 89, 845 (1999).
  22. [22]M.R.Harston, T.Carreyre, J.F.Chemin, F.F.Karpeshin, and M.B.Trzhaskovskaya, Nucl. Phys. A676, 143 (2000).
  23. [23]F. F. Karpeshin, M. R. Harston, F. Attallah, J. F. Chemin, J. N. Scheurer, I. M. Band, M. B. Trzhaskovskaya, Phys. Rev. C 53, 1640 (1996).
  24. [24]S. G. Porsev et al., Phys. Rev. Let. 105, 182501 (2010).
  25. [25]F. F. Karpeshin, M. B. Trzhaskovskaya, Yad. Fiz. 78, 765 (2015).
  26. [26]F. F. Karpeshin, M. B. Trzhaskovskaya, Phys. Rev., C 95, 034310 (2017).
  27. [27]F. F. Karpeshin, I. M. Band, M. B. Trzhaskovskaya, Nucl. Phys. A 654, 579 (1999).
  28. [28]F. F. Karpeshin, M. B. Trzhaskovskaya, Yad. Fiz. 69, 571 (2006).
  29. [29]O. P. Sushkov, V. V. Flambaum, Uspehi Fiz. Nauk 136, 3 (1982).
  30. [30]O. Bohr, B. Mottelson. Nuclear Structure (World Scientific: Singapore, 1998), Vols. 1, 2.
  31. [31]J. E. Sansonetti and W. C. Martin, Handbook of Basic Atomic Spectroscopic Data. J. Phys. Chem. Data, 34, 1559 (2005).DOI
  32. [32]Kramida, A., Ralchenko, Yu., Reader, J., and NIST ASD Team (2015). NIST Atomic Spectra Database (ver. 5.3), [Online]. Available: http://physics.nist.gov/asd [2016, December 3]. National Institute of Standards and Technology, Gaithersburg, MD.
  33. [33]B.A.Zon, F.F.Karpeshin, Zh. Eksp. i Teor. Fiz., 97, 401, 1990 [Sov. Phys. — JETP, 70, 224 (1990)].
  34. [34]F. F. Karpeshin, S. Wycech, I. M. Band, M. B. Trzhaskovskaya, M. Pfützner and J. Zylicz, Phys. Rev. C 57, 3085 (1998).
  35. [35]C. J. Campbell et al., Phys. Rev. Lett. 108, 120802 (2012).

Paper details

vlnc:W_qqeuux3m
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