Selective arrangement of three types of calcium–terbium tetranuclear cores by a thiacalixarene ligand using thermodynamic and kinetic strategies

Abstract

In this study, we report thermodynamic and kinetic strategies for arranging three types of Ca–Tb heterotetranuclear cores in a square configuration sandwiched by thiacalix[4]arene-p-tetrasulfonate (TCAS) ligands in aqueous solutions. In the thermodynamic strategy, the components were mixed under optimum pH conditions to afford a complex with a desired ratio of Ca:Tb:TCAS. In the kinetic strategy, the Ca₁Tb₃TCAS₂ complex was formed via mixing kinetically stable Tb₃TCAS₂ with Ca²⁺. Interestingly, the resulting complexes (Ca_xTb_4−xTCAS₂, x = 1–3) exhibited Tb-centered luminescence upon excitation of the TCAS center with a high quvdantum yield (ϕ = 0.11–0.14) and a long luminescence lifetime (approximately 1.2 ms). The thermodynamic strategy can be applied to Sr²⁺ instead of Ca²⁺, but it is not suitable for first transition metal ions. However, the kinetic strategy is versatile and can be applied to first transition metal ions to afford M₁Tb₃TCAS₂ (M = Sr²⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺).

---

Introduction

The functionality of metal complexes arises from the ligand and the metal center. Introducing different metal centers to a complex to form a heteronuclear complex can result in novel functions imparted by metal–metal interactions and are of significant interest to chemists. Many heteronuclear complexes have been used as luminescent materials, magnetic materials (single molecule magnets), and catalysts.¹

Lanthanide (Ln) complexes exhibit potential as bio-imaging materials, such as luminescence probes and contrast agents for magnetic resonance imaging (MRI).² To enhance the performance of Ln-based probes, the design of a heteronuclear complex featuring different metal centers (d–f or f–f metal centers) can be useful.^1a,3 The d–f and f–f heteronuclear complexes can provide multifunctionality and novel functions through interactions between the metal centers, such as dual luminescence, up-conversion, and multimodalities (e.g., luminescence imaging and MRI).^4–7

Thermodynamic and kinetic strategies (Fig. 1) are known approaches for the selective synthesis of a single heteronuclear complex.^3a In the thermodynamic strategy, two coordination sites of a ligand are designed to provide selectivity towards two metal ions, enabling the binding of distinct metal centers. This thermodynamic strategy is straightforward to implement because of the simple mixing of the ligand and metal ions. Furthermore, the metal centers are in close proximity for their interactions. However, the chemical properties of the metal ions should be significantly different; consequently, the combination of metal ions is limited to d–d or d–f. In the f–f combination, obtaining site-selective coordination is impossible. Moreover, the resulting complex should be kinetically stable, otherwise the Ln and Ln′ ions in the formed complex shuffle between different coordination sites.

Fig. 1 Schematics of the (a) thermodynamic and (b) kinetic strategies for the selective synthesis of a heteronuclear complex.

Two methods are used to implement the kinetic strategy. The first method is the step-wise method, which involves the initial synthesis of a mononuclear complex. The mononuclear complex subsequently reacts with a different metal ion to form a heteronuclear complex. The other method is the covalent joining of two mononuclear complexes. These methods require kinetic stability of the mononuclear intermediates such that they do not release the metal center. Therefore, an endo-type ligand, such as 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) is usually used as a coordination moiety (The definition of an endo-type ligand is one that can encapsulate a metal ion inside the ligand.^10b). However, an endo-type ligand and linker moiety induce the opening of a large distance between the metal centers, which causes weakening of the f–f interaction due to wrapping of the metal centers by the endo-type ligand and separation by the linker moiety.

A coordination bond is dynamic, and its timescale significantly depends on the kinetic property of the metal center. Therefore, the kinetic stability of a heteronuclear complex is crucial in determining the formation of heteronuclear complexes. If a heteronuclear complex is kinetically unstable, the metal centers shuffle between the coordination sites, resulting in a decrease in the intended heteronuclear complex and the formation of unwanted homo- and heteronuclear complexes. Furthermore, the synthesis of a single heteronuclear complex is difficult due to the low selectivity of the coordination moiety of the ligand. This low selectivity of the coordination moiety can also result in the concomitant formation of unwanted homo- and heteronuclear complexes. Therefore, the development of selective synthesis methods and kinetically stable single heteronuclear complexes in aqueous solutions remains a significant challenge in research.

In a previous study, we reported that thiacalix[4]arene-p-tetrasulfonate (TCAS, Fig. 2) can be a ligand for complexes containing multiple Ln and M–Ln cores.⁸ For example, Ln ions self-assemble to form a trinuclear Ln–TCAS complex (Ln₃TCAS₂, Fig. 2).⁹ The TCAS ligand functions as an exo-type ligand, coordinating to the tri-Ln core through O,S,O donor sets located on the surface of TCAS. (The definition of an exo-type ligand is one that binds to a metal center at the surface of the ligand.^10b). The Ln₃TCAS₂ complex exhibits Ln-centered luminescence (Ln = Tb, Yb, and Nd) through an energy-transfer mechanism with good luminescence properties, and possesses high proton relaxivity, and thus functions as a suitable contrast agent for MRI with kinetic stability.^9,10 Furthermore, d–f and f–f heteronuclear complexes, such as Ag–Ln–TCAS, Cd–Ln–TCAS, and Ln–Ln′–TCAS systems, have been successfully synthesized to attain distinctive functions that have not been achieved by mononuclear Ln₁TCAS₁.^7,11

Fig. 2 Two-dimensional representations of TCAS, Ln₁TCAS₁, Ln₃TCAS₂, and heterotetranuclear Ca–Tb–TCAS complexes.

Simply mixing the components at pH 10 according to the thermodynamic strategy afforded Ag₄Tb₁TCAS₂, which showed an exceptionally long-lived Tb-centered luminescent lifetime (4.6 ms) in an aqueous solution by eliminating coordinated water molecules through the O₈-cubic coordination geometry provided by the two TCAS ligands.^11a,b In another example, the Tb–Yb–TCAS system provided a mixture of homo- and heterotrinuclear complexes (Tb_xYb_3−xTCAS₂, x = 0–3), in which heterotrinuclear complexes (Tb_xYb_3−xTCAS₂, x = 1, 2) exhibited a Tb→Yb energy transfer, that is, f–f communication.⁷ In contrast, following the kinetic strategy, the Tb₁Yb₂TCAS₂ complex can be selectively synthesized by isolating an intermediate complex (Ln₁TCAS₁, Fig. 2).⁷ Thus, TCAS is a suitable ligand for the formation of d–f and f–f heteronuclear complexes.

In a recent experiment aimed at preparing Tb_xYb_3−xTCAS₂ (x = 1, 2), we accidently found new heteronuclear complexes containing TCAS, Ca²⁺, and Tb³⁺ (Fig. 2) in the reaction mixture. The Ca–Tb–TCAS system modulates the stoichiometry of the Ca–Tb core by manipulating the Ca : Tb ratio and pH condition, which results in energy-transfer luminescence. Calcium ions play significant roles in cell membrane responses, serving as a biomaterial component, with applications for cancer therapy.¹²

In previous research on heteronuclear complexes,^4–6 a precise design and complicated synthesis of the ligands were required for the selective formation of heteronuclear complexes. For example, Piguet et al. reported some d-f heteronuclear complexes with a precisely designed ligand providing different coordination sites of 6 and 9 coordination numbers to d and f ions, respectively.⁴ In contrast, our system only involves mixing the components (Ca, Tb, and TCAS) while controlling the pH condition for the formation of Ca–Tb–TCAS complexes without modification of the TCAS ligand.

In this study, we demonstrate the selective synthesis of three types of heterotetranuclear Ca–Tb–TCAS complexes using thermodynamic and kinetic strategies. Moreover, the study reveals the luminescent properties of the resulting complexes, and extending the strategies to metal ions other than Ca²⁺ is also successfully demonstrated.

Results and discussion

Self-assembly of heterotetranuclear Ca–Tb–TCAS complexes using the thermodynamic strategy

In our previous study, we found an unknown trace in a reaction mixture used for selective synthesis of heterotrinuclear complexes (Tb_xYb_3−xTCAS₂, x = 1, 2).^7b It was confirmed that the unknown trace was a heteronuclear complex containing Tb Tb³⁺, TCAS (H₄tcas⁴⁻), and Ca²⁺. Then, to synthesize Ca–Tb–TCAS complexes, the ratios of Ca:Tb:TCAS and pH conditions were varied. As a result, three types of heterotetranuclear Ca–Tb–TCAS (Ca_xTb_4−xTCAS₂, x = 1−3) were selectively formed under the appropriate conditions, as confirmed by high-performance liquid chromatography (HPLC) peaks (peak 1, peak 2, and peak 3) with different retention times (Fig. 3). The HPLC conditions were 45 wt% CH₃CN in H₂O containing 10 mmol per kg HEPES (apparent pH 7.4) and 30 mmol per kg TBABr (HEPES = 2-[4-(2-hydroxyethyl)-1-pyperadinyl]ethanesulfonic acid, TBABr = tetrabutylammonium bromide).

Fig. 3 Chromatograms of Ca–Tb–TCAS systems at different pH conditions. pH 6.0 condition: [Ca²⁺] = 10 μM, [Tb³⁺] = 30 μM, [TCAS] = 20 μM, [MES] = 20 mM, pH 6.0, t = 24 h, 60 °C; pH 10 condition: [Ca²⁺] = [Tb³⁺] = [TCAS] = 20 μM, [CAPS] = 20 mM, pH 10, t = 24 h, 60 °C; pH 11 condition: [Ca²⁺] = 30 μM, [Tb³⁺] = 10 μM, [TCAS] = 20 μM, pH 11 (adjusted with NH₃), t = 1 h, 60 °C.

Because previously reported Ln₃TCAS₂ complexes showed high kinetic stability under strongly acidic conditions (Ln = Yb: t_1/2 = 1.53 h at pH 1.16 and Ln = Gd: t_1/2 = 2.4 h at pH 2),^10a,b it was expected that the heterotetranuclear Ca–Tb–TCAS complexes would exhibit high kinetic stability, implying that the dissociation reaction during the HPLC measurement is negligibly slow. The stoichiometry of the heterotetranuclear Ca–Tb–TCAS complexes was determined by employing the molar ratio method and electro-spray ionization-mass spectrometry (ESI-MS) analysis. Hereinafter, we describe the self-assembly of the ternary complex at the optimum pH (6.0, 10, and 11). In addition, conditions with different pH values resulted in a mixture of the complexes and/or TCAS ligand (as an example, at pH 7.4, as shown in Fig. S1†).

At pH 6.0, the molar ratio of Tb³⁺/TCAS was fixed at 1.5, and the concentration of Ca²⁺ was varied to investigate the height of peak 1 (Fig. 4). In zero- or low-concentration regions of Ca²⁺, the peak of Tb₃TCAS₂ was observed. As the concentration of Ca²⁺ increased, peak 1 appeared, and its height increased, reaching a maximum at the ratio of Ca²⁺/TCAS = 0.5. In contrast, when the ratio of Ca²⁺/TCAS was fixed at 0.5 and the concentration of Tb³⁺ was varied, the peak of TCAS was observed at zero or low concentrations of Tb³⁺ (Fig. 5). As the concentration of Tb³⁺ increased, the height of peak 1 reached a maximum at Tb³⁺/TCAS = 1.5.

Fig. 4 (a) Chromatograms of the molar ratio method and (b) molar ratio curves between [Ca²⁺] and [TCAS] for the Ca–Tb–TCAS system at pH 6.0. [Ca²⁺] = 0–16 μM, [Tb³⁺] = 30 μM, [TCAS] = 20 μM, [MES] = 20 mM, pH 6.0, λ = 316 nm, t = 24 h, 60 °C.

Fig. 5 (a) Chromatograms of the molar ratio method and (b) molar ratio curves between [Tb³⁺] and [TCAS] for the Ca–Tb–TCAS system at pH 6.0. [Ca²⁺] = 20 μM, [Tb³⁺] = 0–0.10 mM, [TCAS] = 40 μM, [MES] = 20 mM, pH 6.0, λ = 316 nm, t = 24 h, 60 °C.

These results suggest that the stoichiometry of the Ca–Tb–TCAS complex for peak 1 was Ca : Tb : TCAS = 0.5 : 1.5 : 1.0, implying the formation of the Ca₁Tb₃TCAS₂ complex. To confirm the formation of Ca₁Tb₃TCAS₂, ESI mass spectrometry was also performed at pH 6.0, adjusted with NH₃ solution, and CH₃CN was added to the sample solution to facilitate the ionization process (Fig. 6 and S2†). The mass spectra exhibited intense peaks at m/z 454.461738, which could be assigned to [Ca²⁺ + 3Tb³⁺ + 2tcas⁸⁻ + 2Na⁺ + 2OH⁻ + NH₃ + CH₃CN]⁵⁻ (m/z 454.4581). All assignment is included in Fig. S2,† and other products with different metal ratios of Ca–Tb–TCAS complexes were not detected.

Fig. 6 ESI mass spectra of the Ca–Tb–TCAS system at pH 6.0. (a) Observed isotopic distribution at m/z 454.461738 under the following conditions: [Ca²⁺] = 10 μM, [Tb³⁺] = 30 μM, [TCAS] = 20 μM, pH 6.0 (adjusted with NH₃), t = 24 h, 60 °C. (b) Calculated isotopic distribution for [Ca²⁺ + 3Tb³⁺ + 2tcas⁸⁻ + 2Na⁺ + 2OH⁻ + NH₃ + CH₃CN]⁵⁻ at m/z 454.4581⁸⁻.

In previous studies, the results of ESI mass spectrometry showed many adducts in the metal complexes using TCAS ligand.^7,9–11 In our case, the adducts of CH₃CN and NH₃ presenting as guest molecules in the TCAS ligand of the Ca–Tb–TCAS complexes is suggested,¹³ and there is also the possibility for a coordinated NH₃ molecule with the latter because the Tb center of the Ca–Tb–TCAS complexes contains two coordinated water molecules (the details of this result will be discussed later, and see Table 1). Moreover, for other compounds with O²⁻ and H⁺, we assigned these compounds as consisting of H₂O or OH⁻. Thus, the Ca₁Tb₃TCAS₂ complex can be selectively formed at pH 6.0.

At pH 10, the molar ratio method at a fixed ratio of Tb³⁺/TCAS = 1.0 with varying concentrations of Ca²⁺ showed that the height of peak 2 reached a maximum at Ca²⁺/TCAS = 1.0 (Fig. S3†). However, when the ratio of Ca²⁺/TCAS was fixed at 1.0, and the concentration of Tb³⁺ was varied, the height of peak 2 reached a maximum at Tb³⁺/TCAS = 1.0 (Fig. S4†). These observations indicate that the stoichiometry of the complex is Ca_iTb_iTCAS_i (i: natural number). Additionally, the ESI mass spectra of the sample solution at pH 10, which was adjusted with NH₃ solution and contained CH₃CN, showed intense peaks at m/z 708.782527, corresponding to the components of [2Ca²⁺ + 2Tb³⁺ + 2tcas⁸⁻ + Na⁺ + 2H⁺ + 3H₂O + 2NH₃]³⁻ (m/z 708.7946) and other assignments, as shown in Fig. S5.† These results indicate that the stoichiometry of the Ca–Tb–TCAS system at pH 10 was Ca₂Tb₂TCAS₂.

At pH 11, the molar ratio plot of peak 3 showed a maximum at Ca²⁺/TCAS = 1.5 when the Tb³⁺/TCAS ratio was fixed at 0.5 (Fig. S6†). Furthermore, a maximum of peak 3 appeared at Tb³⁺/TCAS = 0.5 when the Ca²⁺/TCAS ratio was fixed at 1.5 (Fig. S7†). These results suggest that the Ca₃Tb₁TCAS₂ complex is primarily formed at pH 11; however, peaks of unwanted impurities, such as Ca₂Tb₂TCAS₂ and TCAS, were also observed. Therefore, a more suitable condition for the selective synthesis of the Ca₃Tb₁TCAS₂ complex was necessary.

Considering the success in enhancing the selectivity by adding TBABr during the synthesis of the Ln–Ln′–TCAS heterotrinuclear complex,^7b we added tetrabutylammonium (TBA⁺) bromide (TBABr) salt to the mixture of Ca²⁺, Tb³⁺, and TCAS to enhance the selectivity. Upon the addition of ≥0.2 M TBABr, the proportion of Ca₃Tb₁TCAS₂ increased, and that of Ca₂Tb₂TCAS₂ significantly decreased (Fig. S8†). The molar ratio method under the conditions of 0.2 M TBABr and pH 11 resulted in maxima at Ca²⁺/TCAS = 1.5 (Fig. S9†) and at Tb³⁺/TCAS = 0.5 (Fig. S10†). The effect of TBABr salt on the reaction is currently unclear and is still under investigation. We briefly described the effect of the TBABr salt as a hypothesis based on our previous findings.^7b The presence of TBABr or TBA⁺ affects the rate constant of the complexation reaction during the formation of a sandwich structure, leading to a decrease in the rate of the reaction step:

2Ln₁TCAS₁ → Ln₂TCAS₂ (1)

In the Ca–Tb–TCAS system, three reactions are involved in forming the 2 : 2 complexes, as shown below:

2Tb₁TCAS₁ → Tb₂TCAS₂ (1′)

2Ca₁TCAS₁ → Ca₂TCAS₂ (2)

Tb₁TCAS₁ + Ca₁TCAS₁ → Ca₁Tb₁TCAS₂ (3)

We assume that the presence of TBABr or TBA⁺ selectively decreases the reaction rate of eqn (1′). The ESI-MS analysis of the sample solution of the Ca–Tb–TCAS system, which was adjusted to pH 11 with NH₃ solution and contained CH₃CN, exhibited intense peaks at m/z 382.670041 and 382.8715772. This isotopic distribution corresponds to the components of [3Ca²⁺ + Tb³⁺ + 2tcas⁸⁻ + H⁺ + H₂O]⁵⁻ (m/z 382.6698) and [3Ca²⁺ + Tb³⁺ + 2tcas⁸⁻ + 2H⁺ + H₂O]⁵⁻ (m/z 382.8714), confirming the formation of the Ca₃Tb₁TCAS₂ complex at pH 11. Additionally, other signals and their assignments are shown in Fig. S11.†

We attempted to reveal the structure of these Ca–Tb–TCAS complexes. Despite some crystallization experiments, the single crystal suitable for X-ray diffraction analysis has not yet been obtained and is in progress. We preliminarily investigated isolation of a solid sample of the Ca–Tb–TCAS complexes as a salt with hexadecyltrimethylammonium ion (CTMA⁺). We successfully obtained precipitates produced by the mixing of the Ca–Tb–TCAS complexes and CTMA⁺.

In the mixture of the Ca₁Tb₃TCAS₂ complex and the CTMA⁺ system, the CHN elemental analysis indicated that the composition of the salt was CTMA₆Ca₁Tb₃TCAS₂(H₂O)₁₀Cl₁. This result was consistent with the salt for the Tb₃TCAS₂ complex with quaternary ammonium ions in the previous study, for example, CTMA₇Tb₃TCAS₂(H₂O)₁₀.^16b The other two complexes, Ca₂Tb₂TCAS₂ and Ca₃Tb₁TCAS₂, also showed similar compositions: CTMA₁₀Ca₂Tb₂TCAS₂(H₂O)₁₈(NO₃)₃ and CTMA₁₀Ca₃Tb₁TCAS₂(H₂O)₂₀Cl₃. These CTMA–Ca–Tb–TCAS salts contained H₂O molecules as coordinated water (see Table 1) and crystal solvent, and the coprecipitation of CTMACl or CTMANO₃ salts was suggested due to the experimental conditions (addition of a small excess amount of CTMACl and cooling before filtration).

By simply mixing the components (Ca²⁺, Tb³⁺, and TCAS) at an appropriate ratio and appropriate pH (6.0, 10, and 11), selective formation of the Ca₁Tb₃TCAS₂, Ca₂Tb₂TCAS₂, and Ca₃Tb₁TCAS₂ complexes can be achieved using the thermodynamic strategy. The number of Tb³⁺ in the Ca–Tb–TCAS complexes decreased at high pH, which may be attributed to the fact that Tb³⁺ is more susceptible to hydrolysis compared to Ca²⁺. This phenomenon is influenced by the valence of the metal ions.

The results of the present work show that the ratio of Tb³⁺ and Ca²⁺ (Ca₁Tb₃, Ca₂Tb₂, and Ca₃Tb₁) could be completely controlled by pH conditions despite their properties of chemical similarity. These systems enabled selective synthesis of heteronuclear complexes without the need for precise design or modification of the ligand. In addition, this observation is remarkably unique when compared to previous studies in which TCAS or TCA was used as a ligand containing a M–Ln core, and the chemical properties of M (Ag⁺, Cd²⁺, Mn²⁺, Zn²⁺) and Ln (Tb³⁺, Yb³⁺, Nd³⁺, Gd³⁺, Dy³⁺, Ho³⁺) are significantly different.^11b–d,14

We assume that the Ca–Tb–TCAS complexes possess a double-cone structure in which the tetrametal core is sandwiched by two conical TCAS anions, as exemplified by Ln₃TCAS₂ and Ln₄TCA₂.^9,15 First, Ca₁Tb₃TCAS₂ can also be formed by mixing Ca²⁺ and Tb₃TCAS₂ complex in the kinetic strategy (vide infra), suggesting that Ca₁Tb₃TCAS₂ adopts the double-cone structure. Second, the luminescence properties as well as the number of coordinated water molecules of the Ca₂Tb₂TCAS₂ and Ca₃Tb₁TCAS₂ complexes are very similar to those of Ln₃TCAS₂ and Ca₁Tb₃TCAS₂, indicating the structural similarity among these complexes. Third, the chromatographic behavior of those complexes is very similar. The number of coordination bonds connecting the metal core and TCAS in the Ca–Tb–TCAS complexes is larger than that in the Ln₃TCAS₂ complex. Consequently, the kinetic stability of the heterotetranuclear Ca–Tb–TCAS complexes is expected to be greater than that of the Tb₃TCAS₂ complex.

Time dependence of the self-assembly of the heterotetranuclear Ca–Tb–TCAS complexes

The self-assembly of the heterotetranuclear Ca–Tb–TCAS complexes was measured as a function of time under specific ratios of the components (Ca²⁺, Tb³⁺, and TCAS) at optimum pH (Fig. 7). At pH 6.0, peak 1 (Ca₁Tb₃TCAS₂) immediately appeared upon mixing the components, and reached a plateau after 120 min (Fig. 7(a)). The byproducts (Tb₁TCAS₁, Tb₃TCAS₂, and Ca₂Tb₂TCAS₂) decreased and eventually disappeared at 120 min, indicating the formation of a single Ca₁Tb₃TCAS₂ complex. There was an appearance by peak 2 (Ca₂Tb₂TCAS₂), which reached a plateau at 240 min, while the peaks of the byproducts Ca₁Tb₃TCAS₂ and Ca₃Tb₁TCAS₂ disappeared at pH 10 (Fig. 7(b)), indicating the formation of a single Ca₂Tb₂TCAS₂ complex. In the absence of TBABr at pH 11, not only peak 3 (Ca₃Tb₁TCAS₂) but also significant amounts of unwanted peak 2 Ca₂Tb₂TCAS₂ (approximately 25%) and TCAS appeared and persisted (Fig. S12†). In the presence of TBABr at pH 11, peak 2 and TCAS were substantially suppressed, and peak 3 was predominant for 150 min (Fig. 7(c)).

Fig. 7 Time dependence of the height of the HPLC peaks of the Ca–Tb–TCAS complexes during self-assembly. (a) [Ca²⁺] = 0.10 mM, [Tb³⁺] = 0.30 mM, [TCAS] = 0.20 mM, [MES] = 40 mM, pH 6.0, 60 °C. (b) [Ca²⁺] = 2.0 mM, [Tb³⁺] = 2.0 mM, [TCAS] = 2.0 mM, pH 10 (adjusted with NH₃), 60 °C. (c) [Ca²⁺] = 0.15 mM, [Tb³⁺] = 50 μM, [TCAS] = 0.10 mM, [TBABr] = 0.20 M, pH 11 (adjusted with NH₃), 60 °C.

The time scale of the formation of each Ca–Tb–TCAS complex should be governed by a rate-determining step (RDS), which is different among the ternary complexes. For example, in Fig. 7(a), the increasing rate of peak 1 and decreasing rate of Tb₃TCAS₂ were the same, implying that the RDS is:

Ca²⁺ + Tb₃TCAS₂ → Ca₁Tb₃TCAS₂ (4)

In the case of Fig. 7(b), steps performed to change the composition of the tetrametal core should control the rate of formation of peak 2. Fig. 7(c) shows that the rate of formation of peak 3 and that of a decrease in TCAS are roughly the same, suggesting the presence of a key intermediate that does not significantly accumulate.

Synthesis of a heterotetranuclear Ca–Tb–TCAS complex using the kinetic strategy

Considering eqn (4), we prepared Ca₁Tb₃TCAS₂ by introducing Ca²⁺ into Tb₃TCAS₂. One equivalent of Ca²⁺ was added to a solution containing Tb₃TCAS₂ obtained by a previously reported method,¹⁵ and the self-assembly process of the complexation was monitored. The peak of Tb₃TCAS₂ was dominant immediately after the addition of Ca²⁺ (Fig. 8). Subsequently, the peak of Tb₃TCAS₂ decreased; however, peak 1 (Ca₁Tb₃TCAS₂) emerged and increased. Furthermore, peak 1 reached a plateau at 240 min, and the peak of Tb₃TCAS₂ disappeared. Thus, mere mixing of the Tb₃TCAS₂ complex and Ca²⁺ afforded Ca₁Tb₃TCAS₂, which formed due to the kinetic stability of Tb₃TCAS₂; thus, this approach can be classified as a kinetic strategy. The times for reaching the plateau were not the same (120 min in Fig. 7(a) and 240 min in Fig. 8), which can be attributed to the difference in the pH conditions.

Fig. 8 Time dependence of the height of the HPLC peaks of the Ca₁Tb₃TCAS₂ complex in the mixture of the Tb₃TCAS₂ complex and Ca²⁺. [Ca²⁺] = 0.10 mM, [Tb₃TCAS₂] = 0.10 mM, pH 10 (adjusted with NH₃), 60 °C.

Luminescence properties of the heterotetranuclear Ca–Tb–TCAS complexes

The excitation spectra of the three heterotetranuclear Ca–Tb–TCAS complexes (Ca_xTb_4−xTCAS₂, x = 1–3) exhibited a broad peak, attributed to the π–π* transition of ligand-centered absorption (Fig. 9(a)). The wavelengths of the excitation peaks were disparate, with values of 314 nm for Ca₁Tb₃TCAS₂, 320 nm for Ca₂Tb₂TCAS₂, and 326 nm for Ca₃Tb₁TCAS₂. The wavelength of the peaks depended on the charge of the metal core; that is, the smaller the number of the charges of the tetrametal core (i.e., +11 for Ca₁Tb₃, +10 for Ca₂Tb₂, and +9 for Ca₃Tb₁), the longer the excitation wavelength. The emission spectra of the complexes Ca_xTb_4−xTCAS₂ (x = 1–3) exhibited sharp emission bands, corresponding to Tb-centered luminescence (⁵D₄ → ⁷F_J, J = 6–3). The ligand-centered absorption and the Tb-centered luminescence suggest that all heterotetranuclear Ca–Tb–TCAS complexes (Ca_xTb_4−xTCAS₂, x = 1–3) exhibited luminescence through energy transfer from TCAS to Tb³⁺.

Fig. 9 (a) Excitation and emission spectra and (b and c) the fine-structure of Tb-centered luminescence for the Ca–Tb–TCAS systems. The Ca₁Tb₃TCAS₂ system: [Ca₁Tb₃TCAS₂] = 1.0 μM, [CAPS] = 10 mM, pH 10, λ_ex = 314 nm, λ_em = 550 nm, t = 24 h, 60 °C; the Ca₂Tb₂TCAS₂ system: [Ca₂Tb₂TCAS₂] = 1.0 μM, [CAPS] = 10 mM, pH 10, λ_ex = 320 nm, λ_em = 550 nm, t = 24 h, 60 °C; and the Ca₃Tb₁TCAS₂ system: [Ca₃Tb₁TCAS₂] = 1.0 μM, [TBABr] = 20 mM, pH 10.5 (adjusted with NH₃), λ_ex = 326 nm, λ_em = 550 nm, t = 1 h, 60 °C. (a) The width of the excitation and emission slits was 5 nm. (b and c) The width of the excitation slit was 20 nm, and the width of the emission slit was 1 nm. Normalized at 550 nm.

In the fine structure of Tb-centered luminescence (⁵D₄ → ⁷F_J, J = 5, 4), the shapes of the bands for the Ca₁Tb₃TCAS₂, Ca₂Tb₂TCAS₂, and Ca₃Tb₁TCAS₂ complexes were not identical (Fig. 9(b and c)). Notably, the luminescence spectra of the Ca₃Tb₁TCAS₂ complex exhibited significantly different shapes compared to those of the Ca₁Tb₃TCAS₂ and Ca₂Tb₂TCAS₂ complexes. In general, the fine structure of the Tb-centered luminescence depends on the coordination environment of the Tb³⁺ center. Assuming a sandwich-type structure for all Ca–Tb–TCAS complexes, the difference in the coordination environment of the Tb³⁺ center can be understood by the electron density of μ₂-O, a phenol oxygen atom of TCAS (O_TCAS), bridging two metal centers according to Ca-O_TCAS-Ca, Ca-O_TCAS-Tb, and Tb-O_TCAS-Tb (Fig. 10).

Fig. 10 Arrangements of the Ca–Tb core and three types of phenol oxygen atom of TCAS in the heterotetranuclear complexes.

The electron density of μ₂-O decreases in the following order: Ca-O_TCAS-Ca > Ca-O_TCAS-Tb > Tb-O_TCAS-Tb. Considering this order, the Tb³⁺ center in the Ca₃Tb₁TCAS₂ is coordinated by two μ₂-O (Ca-O_TCAS-Tb) with an intermediate electron density. For Ca₁Tb₃TCAS₂, two of the Tb³⁺ centers are coordinated by μ₂-O of Ca-O_TCAS-Tb, but one is coordinated by the most electron-deficient μ₂-O (Tb-O_TCAS-Tb). Therefore, it is likely that the difference between the fine structures of the Tb-centered luminescence of the Ca₁Tb₃TCAS₂ and Ca₃Tb₁TCAS₂ complexes originated from the presence or the absence of the Tb-O_TCAS-Tb type of μ₂-O.

For Ca₂Tb₂TCAS₂, there are two possible configurations: a diagonal arrangement (Ca–Tb–Ca–Tb disposition) or an arrangement adjacent to an identical metal center (Ca–Ca–Tb–Tb disposition) (Fig. 10). These configurations represent distinct types of μ₂-O: the Ca–Tb–Ca–Tb disposition has only one type of μ₂-O, that is, Ca–O_TCAS-Tb; however, the Ca–Ca–Tb–Tb disposition has two types of μ₂-O, that is, Ca–O_TCAS–Tb and Tb–O_TCAS–Tb. Because the shape of the luminescence bands of Ca₂Tb₂TCAS₂ is similar to that of Ca₁Tb₃TCAS₂ with Tb–O_TCAS–Tb, the arrangement of the tetrametal core in Ca₂Tb₂TCAS₂ is possibly Ca–Ca–Tb–Tb.

This hypothesis is reasonable by considering the coordination bond length of M–O_TCAS or M–S, which should be different among M = Ca²⁺ and Tb³⁺. The Ca–Ca–Tb–Tb disposition is more tolerant than the Ca–Tb–Ca–Tb disposition toward sandwiching the metal core while maintaining the optimal M–O_TCAS and M–S length. Consequently, two conical TCAS ligands were not in parallel alignment, but rather, were aligned in a slightly tilted manner. In fact, examples of such M–M–Ln–Ln arrangements in M–Ln–thiacalixarene systems have been previously reported;^11b–d,14a however, no arrangements of M–Ln–M–Ln have been reported.

Table 1 summarizes the photophysical properties of the heterotetranuclear Ca–Tb–TCAS complexes (Ca_xTb_4−xTCAS₂, x = 1–3) and the previously reported Tb₃TCAS₂ complex. The Tb-centered luminescence lifetime for the three heterotetranuclear complexes was long, with equal values of approximately 1.2 ms. The number of coordinated water molecules was estimated to be two for each heterotetranuclear complex by Horrocks' equation¹⁷ and the luminescence lifetimes in H₂O and D₂O. The quantum yield of the three heterotetranuclear complexes was 0.11–0.14. The luminescence properties of the heterotetranuclear complexes were as high as those of the previously reported homotrinuclear Tb₃TCAS₂ complex,¹⁶ indicating that they possess a similar sandwich-type structure and octa-coordination environment of the Tb³⁺ center by two O, S, O donor sets from two TCAS ligands and two oxygen atoms from water.⁹

Table 1 Quantum yield (ϕ)^a, luminescence lifetime (τ)^b, and the number of coordinating water molecules per Tb³⁺ center (q)^c for the Ca–Tb–TCAS and Tb₃TCAS₂ complexes

	ϕ/—	τH/ms	τD/ms	q
a [complex] = 0.25 μM, [CAPS] = 10 mM, pH 10, λex = 314 nm for Ca1Tb3TCAS2, 320 nm for Ca2Tb2TCAS2, 326 nm for Ca3Tb1TCAS2, t = 24 h for Ca1Tb3TCAS2 and Ca2Tb2TCAS2, 1 h for Ca3Tb1TCAS2.b For the decay curves, see Fig. S13–15 in ESI. Subscripts H and D for τ denote H2O and D2O solvents, respectively.c Estimated by Horrocks' equation. q = 4.19(τH^−1 – τD^−1).d See ref. 16.
Ca3Tb1TCAS2	0.11	1.19	3.50	2.3
Ca2Tb2TCAS2	0.13	1.19	2.81	2.1
Ca1Tb3TCAS2	0.14	1.22	2.90	2.0
Tb3TCAS2^d	0.14	1.12	3.22	2.4

---

Thermodynamic and kinetic strategies for other metal ions to form heterotetranuclear M–Tb–TCAS complexes

Inspired by the success of the selective preparation of three types of Ca–Tb–TCAS heterotetranuclear complexes using the thermodynamic strategy, and the preparation of Ca₁Tb₃TCAS₂ using the kinetic strategy, we extended the strategy to other alkaline earth and transition metal ions (M) (Fig. S16†). For example, the chromatograms of the reaction mixture in the thermodynamic strategy (M : Tb : TCAS = 1 : 3 : 2, pH 6.0 and M : Tb : TCAS = 2 : 2 : 2, pH 10) showed a single peak in the presence of TCAS and Tb³⁺ with M = Sr²⁺ at the retention time (t_R = 4.53 min for Sr : Tb : TCAS = 1 : 3 : 2 at pH 6.0, 6.56 min for M:Tb:TCAS = 2 : 2 : 2 at pH 10), which was similar to that for one of the Ca–Tb–TCAS complexes (t_R = 4.67 min for Ca₁Tb₃TCAS₂, 6.86 min for Ca₂Tb₂TCAS₂). In contrast, the chromatograms of the transition metal systems (M = Ti³⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺) showed several peaks corresponding to various M–Tb–TCAS and Tb–TCAS complexes. This may be caused by many factors, including the affinity for the TCAS ligand, complexation kinetics, and the tendency towards hydrolysis of transition metal ions other than Ca²⁺. Hence, the optimum conditions for the formation of heterotetranuclear M–Tb–TCAS complexes should be different from those for the formation of Ca–Tb–TCAS.

In contrast, the kinetic strategy using the reaction between Tb₃TCAS₂ and metal ions (M) resulted in a single peak in all systems, and M₁Tb₃TCAS₂ complexes (M = Sr²⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺) were successfully formed. The peak of the Sr–Tb₃TCAS₂ system appeared at t_R = 4.35 min, which is similar to that of the Ca₁Tb₃TCAS₂ complex (t_R = 4.67 min). For the transition metal-Tb₃TCAS₂ systems, a single peak appeared at t_R = 5.22–5.35 min, which is different than that of the Ca₁Tb₃TCAS₂ complex. This may have been caused by the difference in the charge of the M and/or in the deprotonation of the coordinating water to M. Thus, the kinetic strategy can be applied not only to Ca²⁺, but also to various metal ions.

We further investigated the ESI mass spectrometry for the mixture of M and the Tb₃TCAS₂ solution to confirm the formation of the M₁Tb₃TCAS₂ complex (M = Sr²⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺) (Fig. S17–23†). For example, the mixture of Sr²⁺ and Tb₃TCAS₂ solution showed isotopic distribution at m/z 443.24429, which could be assigned to [Sr²⁺ + 3Tb³⁺ + 2tcas⁸⁻ + 2H₂O]⁵⁻ (m/z 443.2449) (Fig. S17†). This result indicates the formation of the Sr₁Tb₃TCAS₂ complex. The others also showed mass spectra corresponding to [M²⁺ + 3Tb³⁺ + 2tcas⁸⁻ + 2H₂O]⁵⁻ (M = Sr²⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Zn²⁺) and [Cu²⁺ + 3Tb³⁺ + 2tcas⁸⁻ + H₂O]⁵⁻. These ESI-MS measurement results are consistent with the above HPLC measurements, implying that the formation of the M₁Tb₃TCAS₂ complex could be possible by the mixture of M and Tb₃TCAS₂ solution at pH 6.0. Moreover, the mixture of M and Tb₃TCAS₂ emitted Tb-centered luminescence due to energy transfer from TCAS ligand, but the intensity at 550 nm was different depending on the type of metal ion (M) (Fig. S24†).

In the case of M = Sr²⁺ or Zn²⁺, the intensity at 550 nm was the same as that of the Tb₃TCAS₂ complex. In contrast, in the case of M = Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, or Cu²⁺, the Tb-centered luminescence was diminished, indicating paramagnetic quenching. According to the above findings, the mixture of M and Tb₃TCAS₂ produces M₁Tb₃TCAS₂ complexes (M = Sr²⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺).

Conclusions

Herein, we presented a facile method to selectively prepare three luminescent heterotetranuclear complexes (Ca_xTb_4−xTCAS₂, x = 1–3), in which heterometal cores are arranged in a square form. In the thermodynamic strategy, three complexes (Ca_xTb_4−xTCAS₂, x = 1–3) were formed by simply mixing Ca, Tb, and TCAS at an appropriate proportion with an optimized pH. The selectivity of the formation of heterotetranuclear complexes seemingly depends upon the fact that the tendency of Tb³⁺ to hydrolyze is greater than that of Ca²⁺ (Fig. 11). This result suggests that the difference in the susceptibility to hydrolysis between heterometal ions is a useful factor to control the heterometal core.

Fig. 11 Schematic of the formation of the heterotetranuclear Ca–Tb–TCAS complexes.

Moreover, this method can be applied not only to the TCAS ligand but also to many other ligands. In the kinetic strategy, Ca₁Tb₃TCAS₂ was afforded by simple addition of Ca²⁺ to Tb₃TCAS₂ in solution (Fig. 11). All heterotetranuclear complexes (Ca_xTb_4−xTCAS₂, x = 1–3) exhibited Tb-centered luminescence by an energy-transfer mechanism with nearly the same lifetime and quantum yield. However, the heterotetranuclear complex is kinetically more stable than the Ln₃TCAS₂ complex because of the multi-coordination bond by the tetranuclear core. Finally, we extended the strategies to the preparation of M–Tb–TCAS complexes (M = Sr²⁺, Mn²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Zn²⁺). Because thiacalixarene ligands can form a metal complex with all Ln³⁺ species⁹ and various transition metal ions,^11d the strategy may be applicable to other metal–lanthanide–TCAS complexes. Thus, the strategies mentioned in this study are promising, and will pave the way for the application of heterometal complexes in multimodal probes and multifunctional materials.

Data availability

The data supporting this article have been included as part of the ESI.†

Author contributions

Ryunosuke Karashimada: conceptualization, data curation, funding acquisition, investigation, project administration, resources, supervision, visualization, writing–original draft, writing–review and editing; Hironori Matsuoka: data curation, investigation, validation, visualization; Nobuhiko Iki: funding acquisition, project administration, resources, supervision, writing–review and editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

ESI-MS measurements were supported by Mr Hiroyuki Momma and Shunsuke Kayamori. This work was partly supported by JSPS KAKENHI (Grant Numbers 18K14248, 20K15308, and 23K04801).

References

1. (a) F. Artizzu, F. Quochi, A. Serpe, E. Sessini and P. Deplano, Inorg. Chem. Front., 2015, 2, 213–222; (b) J.-R. Jiménez, B. Doistau, M. Poncet and C. Piguet, Coord. Chem. Rev., 2021, 434, 513750–513769; (c) X.-Z. Li, C.-B. Tian and Q.-F. Sun, Chem. Rev., 2022, 122, 6374–6458; (d) Z. Zhu and J. Tang, Chem. Soc. Rev., 2022, 51, 9469–9481; (e) Y. Peng, H. Kaemmerer and A. K. Powell, Chem. Eur J., 2021, 27, 15043–15065; (f) Q. Wang, S. H. Brooks, T. Liu and N. C. Tomson, Chem. Commun., 2021, 57, 2839–2853; (g) Z.-H. Pan, Z.-Z. Weng, X.-J. Kong, L.-S. Long and L.-S. Zheng, Coord. Chem. Rev., 2022, 457, 214419–214440.
2. (a) J.-C. G. Bünzli, Chem. Rev., 2010, 110, 2729–2755; (b) J.-C. G. Bünzli and S. V. Eliseeva, J. Rare Earths, 2010, 28, 824–842; (c) S. V. Eliseeva and J.-C. G. Bünzli, Chem. Soc. Rev., 2010, 39, 189–227; (d) M. Sy, A. Nonat, N. Hildebrandt and L. J. Charbonnière, Chem. Commun., 2016, 52, 5080–5095.
3. (a) N. Iki, Supramol. Chem., 2011, 23, 160–168; (b) A. M. Nonat and L. J. Charbonnière, Coord. Chem. Rev., 2020, 409, 213192–213208.
4. (a) L. Aboshyan-Sorgho, H. Nozary, A. Aebischer, J.-C. G. Bünzli, P.-Y. Morgantini, K. R. Kittilstved, A. Hauser, S. V. Eliseeva, S. Petoud and C. Piguet, J. Am. Chem. Soc., 2012, 134, 12675–12684; (b) D. Zare, Y. Suffren, L. Guénée, S. V. Eliseeva, H. Nozary, L. Aboshyan-Sorgho, S. Petoud, A. Hauser and C. Piguet, Dalton Trans., 2015, 44, 2529–2540; (c) D. Zare, Y. Suffren, H. Nozary, A. Hauser and C. Piguet, Angew. Chem., Int. Ed., 2017, 56, 14612–14617.
5. (a) F. Artizzu, F. Quochi, L. Marchiò, E. Sessini, M. Saba, A. Serpe, A. Mura, M. L. Mercuri, G. Bongiovanni and P. Deplano, J. Phys. Chem. Lett., 2013, 4, 3062–3066; (b) F. Artizzu, F. Quochi, L. Marchiò, R. F. Correia, M. Saba, A. Serpe, A. Mura, M. L. Mercuri, G. Bongiovanni and P. Deplano, Chem. Eur J., 2014, 21, 3882–3885; (c) F. Artizzu, F. Quochi, L. Marchiò, C. Figus, D. Loche, M. Atzori, V. Sarritzu, A. M. Kaczmarek, R. V. Deun, M. Saba, A. Serpe, A. Mura, M. L. Mercuri, G. Bongiovanni and P. Deplano, Chem. Mater., 2015, 27, 4082–4092; (d) F. Artizzu, A. Serpe, L. Marchiò, M. Saba, A. Mura, M. L. Mercuri, G. Bongiovanni, P. Deplano and F. Quochi, J. Mater. Chem. C, 2015, 3, 11524–11530.
6. (a) N. Souri, P. Tian, C. Platas-Iglesias, K.-L. Wong, A. Nonat and L. J. Charbonnière, J. Am. Chem. Soc., 2017, 139, 1456–1459; (b) J. Salaam, L. Tabti, S. Bahamyirou, A. Lecointre, O. H. Alba, O. Jeannin, F. Camerel, S. Cianférani, E. Bentouhami, A. M. Nonat and L. J. Charbonnière, Inorg. Chem., 2018, 57, 6095–6106; (c) A. Nonat, S. Bahamyirou, A. Lecointre, F. Przybilla, Y. Mély, C. Platas-Iglesias, F. Camerel, O. Jeannin and L. J. Charbonnière, J. Am. Chem. Soc., 2019, 141, 1568–1576; (d) R. C. Knighton, L. K. Soro, T. Troadec, V. Mazan, A. M. Nonat, M. Elhabiri, N. Saffon-Merceron, S. Djenad, R. Tripier and L. J. Charbonnière, Inorg. Chem., 2020, 59, 10311–10327; (e) R. C. Knighton, L. K. Soro, A. Lecointre, G. Pilet, A. Fateeva, L. Pontille, L. Francés-Soriano, N. Hildebrandt and L. J. Charbonnière, Chem. Commun., 2021, 57, 53–56; (f) L. K. Soro, C. Charpentier, F. Przybilla, Y. Mély, A. M. Nonat and L. J. Charbonnière, Chem. Eur J., 2021, 3, 1037–1046; (g) R. C. Knighton, L. K. Soro, L. Francés-Soriano, A. Rodríguez-Rodríguez, G. Pilet, M. Lenertz, C. Platas-Iglesias, N. Hilderbrandt and L. J. Charbonnière, Angew. Chem., Int. Ed., 2022, 61, e202113114.
7. (a) R. Karashimada and N. Iki, Chem. Commun., 2016, 52, 3139–3142; (b) R. Karashimada, K. Musha and N. Iki, Bunseki Kagaku, 2022, 71, 145–151.
8. N. Morohashi and N. Iki, Handbook on the Physics and Chemistry of Rare Earths, Elsevier, 2022, vol. 326, pp. 1–280.
9. N. Iki, T. Tanaka, S. Hiro-oka and K. Shinoda, Eur. J. Inorg. Chem., 2016, 5020–5027.
10. (a) N. Iki, E. Boros, M. Nakamura, R. Baba and P. Caravan, Inorg. Chem., 2016, 55, 4000–4005; (b) N. Iki, S. Hiro-oka, M. Nakamura, T. Tanaka and H. Hoshino, Eur. J. Inorg. Chem., 2012, 3541–3545; (c) N. Iki, S. Hiro-oka, T. Tanaka, C. Kabuto and H. Hoshino, Inorg. Chem., 2012, 51, 1648–1656.
11. (a) T. Tanaka, N. Iki, T. Kajiwara, M. Yamashita and H. Hoshino, J. Inclusion Phenom. Macrocyclic Chem., 2009, 64, 379–383; (b) N. Iki, M. Ohta, T. Horiuchi and H. Hoshino, Chem. Asian J., 2008, 3, 849–853; (c) N. Iki, M. Ohta, T. Tanaka, T. Horiuchi and H. Hoshino, New J. Chem., 2009, 33, 23–25; (d) N. Iki, T. Tanaka and H. Hoshino, Inorg. Chim. Acta, 2013, 397, 42–47.
12. Y. Kang, L. Xu, J. Dong, Y. Huang, X. Yuan, R. Li, L. Chen, Z. Wang and X. Ji, Coord. Chem. Rev., 2023, 481, 215050–215068.
13. N. Iki, T. Fujimoto and S. Miyano, Chem. Lett., 1998, 27, 625–626.
14. (a) Y. Bi, Y. Li, W. Liao, H. Zhang and D. Li, Inorg. Chem., 2008, 47, 9733–9735; (b) K. Su, F. Jiang, J. Qian, M. Wu, K. Xiong, Y. Gai and M. Hong, Inorg. Chem., 2013, 52, 3780–3786.
15. A. Bilyk, J. W. Dunlop, R. O. Fuller, A. K. Hall, J. M. Harrowfield, M. W. Hosseini, G. A. Koutsantonis, I. W. Murray, B. W. Skelton, A. N. Sobolev, R. L. Stamps and A. H. White, Eur. J. Inorg. Chem., 2010, 2127–2152.
16. (a) N. Shiraishi, R. Karashimada and N. Iki, Bull. Chem. Soc. Jpn., 2019, 92, 1847–1852; (b) N. Shiraishi, D. Iikura, R. Karashimada and N. Iki, J. Lumin., 2024, 269, 120521–120529.
17. W. D. H. Jr and D. R. Sundnick, Acc. Chem. Res., 1981, 14, 384–392.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ra08259a

---