The coordination chemistry of p-tert-butylcalix[8]arene with transition and lanthanide metal ions

Abstract

A survey of the literature and the Cambridge Structural Database reveals thirty nine p-tert-butylcalix[8]arene-supported transition (3d, limited to V–Cu) and lanthanide metal complexes ranging in nuclearity from one to eighteen, twenty of which are homometallic and nineteen of which are heterometallic. We provide a review of the coordination chemistry of these complexes, including our own work in the area. We also provide our thoughts and perspectives on the common structural themes observed, identify gaps in knowledge and evaluate how that may inform and direct future synthetic efforts. What is clear is that the coordination chemistry of H₈TBC[8] with paramagnetic metal ions remains hugely underexplored.

---

Abhijnan Chakraborty

Abhijnan Chakraborty is currently a 2nd year PhD student in Professor Euan Brechin's group. Originally from India, he obtained his MChem degree from the University of Edinburgh in 2023. His research focuses on exploring the coordination chemistry of calixarenes with various transition metals and lanthanides.

Lucinda R. B. Wilson

Lucinda R. B. Wilson completed both her MChem (2020) and PhD (2024) degrees at The University of Edinburgh, the latter under the supervision of Prof E. K. Brechin. Her PhD research involved exploring the structure–property relationships of calix[n]arene-supported clusters whilst her current post-doctoral research focuses on the synthesis and characterisation of iron-based molecular minerals.

Scott J. Dalgarno

Scott J. Dalgarno is Chair of Supramolecular Chemistry at Heriot-Watt University. His research interests are in self-/metal-directed assembly, crystal engineering, host–guest chemistry and energetic materials. He has been the recipient of the Royal Society of Chemistry Harrison-Meldola Memorial and Sir Edward Frankland prizes, and the Chemical Communications Emerging Investigator Lectureship.

Euan K. Brechin

Euan K. Brechin is the Crum Brown Chair of Chemistry at The University of Edinburgh. His research interests are in synthetic coordination chemistry and molecular magnetism. He has been the recipient of the Royal Society of Chemistry Mond-Nyholm, Tilden, Corday-Morgan and Sir Edward Frankland prizes.

---

Introduction

Recent years have seen a vast expansion in the coordination chemistry of p-tert-butylcalix[4]arene (H₄TBC[4]), particularly in relation to the synthesis of polynuclear 3d, 4f, and 3d–4f clusters that display a range of fascinating physical properties (e.g. single molecule magnetism and magnetic refrigeration).¹ The calix[4]arene (C[4]) scaffold is an extremely attractive ligand in this regard as it can be synthesised and modified with relative ease, but also because it is predisposed to bind a series of metal ions in the polyphenolic pocket. Furthermore, the conformational properties of C[4]s are well-understood, particularly when the lower-rim comprises only hydroxyl groups, and these molecules can form cone, partial-cone, 1,2-alternate and 1,3-alternate conformers.² These conformations can also be accessed through the introduction of small substituents at the lower-rim (e.g. OMe), whilst larger groups can be used to lock out a particular conformer as desired. In most cases, this step would preclude the ability to form polynuclear clusters, so it is not of particular relevance to this review and will not be discussed further. The p-tert-butylthia-, sulfonyl- and sulfinylcalix[4]arene analogues have attracted similar interest in coordination chemistry, building on the fact that their conformations are equally well understood, but also because a common [M₄(S⁻/SO⁻/SO₂⁻TBC[4])Cl]ⁿ⁺ (n = 2, 3) metalloligand results that can be exploited in directed assembly.³ Notably, this metalloligand has been utilised in the formation of a range of spectacular cage structures that have been exploited in areas such as selective guest/gas uptake and sorption.

Some of the larger homologues of H₄TBC[4] are highly desirable for exploring polynuclear cluster formation, an example being H₉TBC[9] which has been seen to adopt a triple-cone conformation akin to a three leaf clover in the solid state upon complexation to lanthanide ions.⁴ This molecule would be very interesting for synthesising highly symmetric polyhedra of 3d/4f ions but, as is the case for many of the very large homologues, it is (to our knowledge) only accessible in relatively small amounts via laborious synthetic and purification protocols. This therefore hinders the potential utility of the larger H_nTBC[n]s in the synthesis of coordination clusters, with the exception of H₈TBC[8] which can be isolated on a large scale with ease. Indeed, it has often been found to be an unwanted product in calixarene synthesis. One might intuitively expect H₈TBC[8] to behave as two H₄TBC[4]s from a coordination chemistry perspective, but the conformations adopted by the octamer upon coordination, coupled with the methylene bridge orientations, preclude such behaviour. To achieve this effect one must tether two H₄TBC[4]s through a single methylene bridge from each to afford H₈-bis-TBC[4], a ligand that does indeed display augmented coordination chemistry that mimics that of fused C[4]s.⁵ It is well known that moving beyond the general C[6] scaffold (i.e. ≥ C[7]) leads to a large increase in both flexibility and the associated number of possible conformers in the free calixarene. It therefore follows that, although larger H_nTBC[n]s present very exciting opportunities for unearthing new coordination chemistry through favoured coordination modes, there are intrinsic challenges arising from conformational versatility prior to tailoring assembly with such moieties.

A recent paper by Kieliszek and Malinska comprehensively described the various conformations of H₈TBC[8] observed in a range of solvated crystal structures, surveying the Cambridge Structural Database in the process, to provide an excellent overview.⁶ A range of intra- and intermolecular interactions are involved in stabilising common conformations observed in the solvates studied, these being chair-loop, chair and pleated-loop (Fig. 1). An excellent early example of how the versatility of H₈TBC[8] can be exploited in host–guest chemistry is its ability to selectively complex C₆₀ from a range of fullerenes in soot.⁷ In this case, the hosts adopt a double-cone conformation, encapsulating C₆₀ molecules to form a complex that can then be harvested in crystalline form and deconstructed to afford pure buckminsterfullerene.

Fig. 1 Common conformations of H₈TBC[8] found in a range of solvates. Chair-loop (top), chair (middle) and pleated-loop (bottom). Colour code: C = grey, O = red, H = white. H(–C) atoms and solvent of crystallisation omitted for clarity.

In this article we provide a review of the coordination chemistry of p-tert-butylcalix[8]arene with 3d transition (limited to V–Cu) and 4f lanthanide metal ions, including our own work in the area, with an emphasis on prevailing magnetic properties. We also provide our thoughts and perspectives on the common structural themes observed, identify gaps in knowledge and evaluate how that may inform and direct future synthetic efforts. A search of the Cambridge Structural Database (CSD, July 2024) and the literature for p-tert-butylcalix[8]arene-supported transition (V–Cu) and lanthanide metal complexes results in 39 hits (Table S1, ESI†).^8–37 Of these, 11 are homometallic 3d complexes, 3 are heterometallic 3d–4/5d complexes, 7 are heterometallic 3d-s complexes, 2 are heterometallic 3d-p complexes, 9 are homometallic 4f complexes, 4 are heterometallic 3d–4f complexes, 2 are heterometallic 3d–5f complexes, and 1 is a heterometallic 4f-s complex. The smallest contains a single metal ion and the largest a total of eighteen metal ions.

V-based clusters

The biggest contributors to this family are polymetallic clusters of V, principally investigated for their catalytic properties.^8–13 The first reported example was the dimer [RNH₃][V^V₂O₂(H₁TBC[8])] (1), published in 1990.⁸ Complex 1 (Fig. 2) was synthesised from the reaction of [VO(OⁱPr)₃] with H₈TBC[8] in THF. The core of the complex describes a [V₂O₂] dimer with the two metal ions bridged by two μ-O(Ph) atoms from the hepta-deprotonated H₁TBC[8] ligand. The latter is in the ‘saddle-shaped’ conformation, twisting in such a way as to provide five O-atoms to each octahedral V^V centre, the sixth coordination site on each metal ion being occupied by a terminally bonded oxide (O²⁻) ion.

Fig. 2 Structure of the anionic [V₂] dimer in 1. All reported [V₂] dimers are structurally similar, differing only in the presence of either an O²⁻ or NR²⁻ ion on the sixth coordination site on each V centre. Complexes 1, 3–6 all contain the calix[8]arene in the hepta-deprotonated H₁TBC[8] form and the complexes are charge balanced by a single cation. Complex 2 contains H₂TBC[8] and is neutral. Colour code: V = dark green, O = red, C = grey. H atoms and counter ions omitted for clarity.

The second example, [V^V₂(p-tolylN)₂(H₂TBC[8])] (2), reported in 2001,⁹ was made from the reaction of [V(p-tolylN)(O^tBu)₃] with H₈TBC[8] in toluene (Fig. S1, ESI†). Its structure is similar to 1, but with an organoimido (NR²⁻) ligand occupying the sixth coordination site on each metal. In this case the calixarene is hexa-deprotonated and the complex is neutral. Four other [V^V₂] dimers are known (Fig. S2 and S3, ESI†), [PPh₄][V^V₂O₂(H₁TBC[8])] (3), [Li(MeCN)₄][V^V₂O₂(H₁TBC[8])] (4), [Na(MeCN)₅][V^V₂O₂(H₁TBC[8])] (5) and [^tBuNH₃][V^V₂(p-tolylN)₂(H₁TBC[8])] (6).^10,11 All are structurally similar to 1 and 2. In each case the sixth coordination site on each V centre is occupied by a terminally bonded O²⁻/NR²⁻ ion, with the calix[8]arene being hepta-deprotonated and charge balance maintained through the presence of a single cation. Compound 5 is made via the reaction of H₈TBC[8] with [NaVO(O^tBu)₄], the latter being formed in situ from VOCl₃. Increasing the ratio of [NaVO(O^tBu)₄] in the reaction leads to the formation of [Na(MeCN)₆][V^V₈Na₇O₁₆(TBC[8])₂(MeCN)₆] (7).¹¹ The anionic cluster (Fig. 3) sits on an inversion centre, with each half containing four VO₂ moieties bonded to four Na ions, one of which is at half-occupancy. Each of the V ions is four coordinate and in tetrahedral {VO₄} geometry. The fully deprotonated TBC[8] ligand adopts a pleated-loop conformation, bonded to the four V ions in the asymmetric unit (ASU). Two neighbouring O(Ph) groups possess μ-bridging O-atoms that serve to link the two halves of the molecule together, bonding to a single V ion in one half and to two Na ions on the other half. Charge balance is maintained through the presence of two half-occupancy [Na(MeCN)₆]⁺ cations.

Fig. 3 The structure of the anionic cluster of compound 7. Colour code: V = dark green, Na = purple, O = red, N = blue, C = grey. H atoms and counter ions omitted for clarity.

The same reaction but employing [KVO(O^tBu)₄] results in the formation of the alkali-metal-free complex [V^V₄O₆(TBC[8])] (8; Fig. 4).¹¹ The metal–oxygen core of 8 describes a near planar [(VO)₄O₂] butterfly, each V ion being five coordinate and in square pyramidal {VO₅} geometry. The fully deprotonated TBC[8] ligand adopts a double-cone conformation, lying “above” the [V₄] plane (as drawn in Fig. 4). Two of the eight O(Ph) atoms are μ-bridging across the body-wing edge of the butterfly, while the remaining six are terminally bonded.

Fig. 4 The structure of compound 8. Colour code: V = dark green, O = red, C = grey. H atoms omitted for clarity.

Employing LiO^tBu in a reaction with VOCl₃ and H₈TBC[8] in THF/Et₂O leads to the isolation of the heterometallic complex [V^V₄Li₆O₈(TBC[8])(O^tBu)₂(THF)₂(Et₂O)₂] (9; Fig. S4, ESI†).¹¹ At the centre of the cluster lies a [V₄Li₄(μ₃-O)₄(μ-O)₄] cage comprising two fused [V₂Li₂(μ₃-O)₄] butterflies, the four μ-bridging O²⁻ ions linking the two halves of the cage together. Each V centre is four coordinate and in tetrahedral {VO₄} geometry. Two additional Li ions sit above the cage and are linked to it via μ-bridging O^tBu molecules. The calixarene adopts a pseudo-pleated loop conformation, with each neighbouring pair of O(Ph) atoms chelating to a V ion.

The four remaining V-based complexes, and those published most recently, are the isostructural heterometallic tetramers [V^V₂Nb^V₂O₅(OⁱPr)₂(TBC[8])(MeCN)₂] (10) and [V^V₂Ta^V₂O₅(OⁱPr)₂(TBC[8])(MeCN)₂] (11; Fig. 5), the hexamer [V^V₃V^IV₁Nb^V₂O₉(OⁿPr)₃(TBC[8])(MeCN)₂] (12; Fig. S5, ESI†) and the monomer [HNEt₃]₂[V^VO₂(H₅TBC[8])] (13; Fig. S6, ESI†).^12,13 Complexes 10 and 11 result from the reaction of H₈TBC[8] with [VO(OⁱPr)₃] and [Nb(OEt)₅] or [Ta(OEt)₅] in toluene, followed by crystallisation from MeCN. The central core of each complex contains a (bent) V–O–Nb/Ta–O–Nb/Ta–O–V chain. The fully deprotonated TBC[8] ligand is in the pleated-loop conformation, with each neighbouring pair of O(Ph) atoms chelating to either a V or Nb/Ta ion. The V ions are four coordinate and in {VO₄} geometries, the fourth site being occupied by a terminally bonded O²⁻ ion. The Nb/Ta ions are six coordinate and in octahedral {Nb/TaO₅N₁} geometries, with their fifth and sixth sites being occupied by MeCN and OⁱPr molecules.

Fig. 5 The structure common to both 10 and 11. Colour code: V = dark green, Nb/Ta = maroon, O = red, N = blue, C = grey. H atoms omitted for clarity.

Replacing [VO(OⁱPr)₃] with [VO(OⁿPr)₃] in the same reaction that affords tetrameric complex 10 results in the formation of the hexameric complex 12.¹² The same central (bent) V–O–Nb/Ta–O–Nb/Ta–O–V chain prevails, but with two additional V ions asymmetrically capping both sides of the Nb–Nb vector, linked by one μ₃-O²⁻ ion and two μ-O²⁻ ions, respectively. The fully deprotonated TBC[8] ligand coordinates in the same way as in 10 and 11. The three OⁿPr ligands are of two types, two are μ-bridging between neighbouring Nb–V–Nb ions, with the third terminally bonded to the same V ion. The Nb ions are six coordinate and in octahedral {NbO₆} geometries. The V ions are of three types, two are four coordinate and in tetrahedral {VO₄} geometries, one is five coordinate and in square pyramidal {VO₅} geometry, while one is six coordinate and in an octahedral {VO₄N₂} geometry with its coordination sphere completed by two MeCN molecules. It is this metal ion that is V^IV.

The reaction of VO(SO₄) with H₈TBC[8] and NEt₃ in refluxing MeOH, followed by crystallisation from MeCN, affords complex 13.¹³ The triply deprotonated H₅TBC[8] ligand sits in a pleated-loop conformation, using two of its deprotonated O(Ph) atoms to bond to the single V^V ion, whose tetrahedral geometry is completed by two terminal O²⁻ ions. The latter are H-bonded to the two HNEt₃ cations that sit one above and one below the calix[8]arene plane. The third deprotonated O(Ph) atom lies directly opposite the {VO₄} moiety, H-bonding to it two nearest neighbours.

Summary of V-based clusters

From the above, it is clear that the coordination chemistry of vanadium (thus far) is dominated by the V^V ion. There are no examples of homovalent V^IV or V^III species in the literature, and there is just one mixed valent species. The vast majority of complexes have been synthesised via air- and moisture-free techniques in non-protic solvents. Very few, if any, have been made in the presence of air/moisture using standard benchtop techniques in protic solvents, and none have been made via high temperature/pressure (solvothermal) methods. Most are of small nuclearity, with dimers being particularly prevalent. Larger species tend to be heterometallic, and the majority of these contain s-block metal ions. This perhaps suggests that the increased flexibility in the coordination number and geometry of the s-block metal ions allows them to be more easily accommodated into the framework of the cluster, whose (often asymmetric) topology is dictated by the conformation of the calix[8]arene. There are no examples of heterometallic V-3d, V-4f or V-5f clusters. Clearly there remains a vast amount of chemistry to be explored before the coordination chemistry of vanadium is fully mapped out. There does not appear to be a common/dominant structural theme arising from the metallic skeletons of the V clusters thus far, albeit {V₂} dimers, {V₄} butterflies and {V₄} tetrahedra are the most common. The calix[8]arene ligands in these clusters show a variety of conformations and are either fully deprotonated, TBC[8], or singly/doubly deprotonated H₁TBC[8]/H₂TBC[8], as one might expect given the coordination chemistry is primarily established for the V^V ion. The most common coordination motif of the ligand is as an η²-chelate with two neighbouring O(Ph) atoms bonded to the same V centre.

Co-based clusters

The second biggest family of TBC[8]-supported clusters belongs to homo- and heterometallic complexes of Co^II.^14–19 The first, and only homometallic cluster in the family is the aesthetically pleasing [HNEt₃]₂[Co^II₂(H₅TBC[8])₂] (14) dimer made in 2007 (Fig. 6).¹⁴ Complex 14 was synthesised via the solvothermal reaction of Co(OAc)₂·4H₂O with H₈TBC[8] and NEt₃ in MeOH at T = 166 °C for 72 hours. The centrosymmetric anionic cluster of 14 possesses a metal–oxygen (magnetic) core describing a [Co₂(OR)₂(HOR)₂] dimer in which the Co ions are ‘linked’ by two Co–O–H⋯O–Co H-bonds. The Co ions are four coordinate and in distorted tetrahedral {CoO₄} geometries, the remaining two O atoms being terminally bonded (deprotonated) O(Ph) atoms of different H₅TBC[8] ligands. The other four O(Ph) atoms on the H₅TBC[8] ligands remain protonated and non-bonded, with both ligands adopting the pleated-loop conformation. Charge balance is provided by two [Et₃NH]⁺ cations. Magnetic susceptibility measurements revealed the presence of weak antiferromagnetic exchange through the Co–O–H⋯O–Co H-bonds, with theoretical calculations indicating significant zero-field splitting at the distorted tetrahedral Co sites.¹⁴

Fig. 6 The structure of the anion of compound 14. Colour code: Co = sky blue, O = red, C = grey. H = white. The Co–O–H⋯O–Co H-bonds are shown by the dashed black lines. These H atoms are included, while all others are omitted for clarity.

The remaining Co-based clusters are all heterometallic. The first is a small family of 3d–4f clusters published in 2012 of formula [Ln^III₆Co^II₈(TBC[8])₂O₂(OH)₄(CO₃)₂(OAc)₄(HCOO)₂(dmf)_m(H₂O)_n] (Ln = Sm (15a), m = 8, n = 6; Ln = Gd (15b), Dy (15c), m = 10, n = 4).¹⁵ The complexes were synthesised from the solvothermal reaction of Ln(OAc)₃·6H₂O, Co(OAc)₂·4H₂O and H₈TBC[8] in a basic dmf/MeOH solution (dmf = N,N-dimethylformamide). The CO₃²⁻ originates from the fixation of atmospheric CO₂ and the formate from the oxidation of MeOH. The structure of 15 (Fig. 7) is rather unusual. Each half of the molecule consists of a central [Co₂Ln₂(μ₄-O)] square. On one side of this square lies a [CoLn₂(μ₃-OH)] triangle which shares a vertex at the Ln site. On the other side of the square lies a [Co₂Ln(μ₃-OH)] triangle which shares the Co–Ln edge of the square. Both triangles are positioned approximately perpendicular to the square. The two remaining Ln ions link the two halves of the molecule together. Each half of the molecule contains one fully deprotonated TBC[8] ligand in the double-cone conformation, with each O(Ph) atom μ-bridging between neighbouring metal ions. Six of these link Co–Ln ions within each half of the cluster, while two bridge the two Ln ions in the central square to the two Ln ions that join the halves together. The Co^II ions are of three types, two are six-coordinate and in distorted octahedral {CoO₆} geometries, four are five-coordinate and in distorted trigonal bipyramidal {CoO₅} geometries, and two are four coordinate and in distorted tetrahedral {CoO₄} geometries. The Ln^III ions are in two geometries, four being eight coordinate and in square antiprismatic {LnO₈} geometries, and two being seven coordinate and in capped trigonal prismatic {LnO₇} geometries. Magnetic studies of all three complexes showed the presence of competing (weak) ferro- and antiferromagnetic exchange interactions.¹⁵

Fig. 7 The structure of compound 15. Colour code: Co = sky blue, Ln = light green, O = red, N = blue, C = grey. H atoms omitted for clarity.

The isostructural complexes [Ln^III₄Co^II₄(TBC[8])₂O₂(def)_x(H₂O)_y] (Ln = Dy (16a, x = 8, y = 4); Er (16b), x = 8, y = 4; Y (16c), x = 6, y = 6; Fig. 8) were published in 2017.¹⁶ They were synthesised from the solvothermal reaction of H₈TBC[8], Co(OAc)₂·4H₂O and Ln(OAc)₃·6H₂O in a basic MeOH/def solution (def = N,N-diethylformamide) at T = 130 °C for 3 days. The metallic skeleton of 16 describes a {Ln^III₄Co^II₄} square antiprism, with the top and bottom faces being {Ln^III₂Co^II₂} squares with the 3d and 4f metal ions alternating around the square. These are connected by a central, planar μ₄-O²⁻ ion. The upper and lower square faces are connected to each other through the presence of four μ₃-H₂O molecules. The inorganic core is sandwiched between two tail-to-tail, fully deprotonated TBC[8] ligands which adopt the double-cone conformation. In each case four of the O(Ph) atoms are μ-bridging between Co–Ln ions while the remaining four are monodentate, neighbouring pairs chelating to a single Ln ion. The coordination sphere of each metal ion is completed by a solvent molecule. The Ln^III ions have two different geometries, two being eight-coordinate and square antiprismatic {LnO₈}, and two being seven-coordinate with capped octahedral {LnO₇} geometry. Likewise, the Co^II ions possess two different geometries, five coordinate distorted trigonal bipyramidal {CoO₅}, and six coordinate distorted octahedral {CoO₆}. Magnetic studies reveal that the Dy analogue (16a) is a single-molecule magnet (SMM), possessing frequency and temperature dependent out-of-phase (χ′′) signals in ac susceptibility studies.¹⁶ The ac susceptibility data were fitted to a model assuming a Raman/Orbach relaxation mechanism of the form ln τ = −ln[CTⁿ + τ₀⁻¹ exp(−U_eff/k_BT)] to afford U_eff/k_B = 10.38 K, τ₀ = 9.8 × 10⁻⁶ s, C = 1500 s⁻¹ K^−0.4, n = 0.4.

Fig. 8 The structure of compound 16. Colour code: Co = sky blue, Ln = light green, O = red, N = blue, C = grey. H atoms omitted for clarity.

In 2020 a family of complexes of formula [HNEt₃][Ln^III₆Co^II₆(H₁TBC[8])₃(OMe)₆(OAc)₂(CO₃)(dmf)₆(MeOH)] (17a, Ln = Y; 17b, Ln = Eu; 17c, Ln = Dy; Fig. 9) was published. Each member was synthesised via the solvothermal reaction of Co(OAc)₂·4H₂O, Ln(OAc)₃·6H₂O and H₈TBC[8] in a basic MeOH/dmf solution at T = 150 °C for 3 days.¹⁷ At the centre of the cluster lies a μ₆-CO₃²⁻ anion, generated in situ from the fixation of atmospheric CO₂, which templates the formation of a {Co₆} wheel. The outer rim of this wheel is bridged alternately by three μ-OH⁻ ions and three μ₃-OH⁻ ions, the latter also bridging to a Ln ion in the {Ln₂TBC[8]} fragments. These are all positioned ‘above’ the {Co₆} plane forming a bowl-shaped cage. The three, fully deprotonated, TBC[8] ligands adopt the double-cone conformation, each encapsulating a {Ln₂} dimer and using two of their O(Ph) arms to bridge between the two metals, {Ln₂(OR)₃}. The third bridging O atom in the dimer is provided by either an acetate or methanol. Two of the O(Ph) arms are used to connect the {Ln₂} dimer to the {Co₆} wheel across the same edge as the μ₃-OH⁻ ions. The remaining three O(Ph) atoms are monodentate, with one per ligand protonated. The coordination spheres of all the Ln^III ions are completed by a molecule of dmf, and result in either seven coordinate {LnO₇} capped octahedra, seven coordinate capped {LnO₇} trigonal prisms or eight coordinate {LnO₈} biaugmented trigonal prisms. Each Co^II ion is five coordinate and in a distorted trigonal bipyramidal {CoO₅} coordination geometry. Magnetic studies suggest that the Eu analogue exhibits spin glass behaviour, while the Dy version is an SMM.¹⁷

Fig. 9 The structure of the anion of compound 17. Colour code: Co = sky blue, Ln = light green, O = red, N = blue, C = grey. H atoms omitted for clarity.

The core of the complex [Co^II₂Th^IV₄O₂(OH)₂(H₁TBC[8])₂(dmf)₆] (18) describes two [CoTh₃O(OH)] tetrahedra, sharing a {Th₂} edge (Fig. S7, ESI†).¹⁸ The μ₄-O²⁻ ions sit in the centre of each tetrahedra, with the μ-OH⁻ ions capping one {Th₂} edge each. The hepta-deprotonated H₁TBC[8] ligands adopt the double-cone conformation, with four O(Ph) atoms μ-bridging between neighbouring Co–Th (×3), Th–Th (×1) metal ions across edges of the {CoTh₃} tetrahedra. The remaining four O(Ph) atoms are all terminally bonded to Th ions. The latter are eight coordinate and in distorted square antiprismatic {ThO₈} geometries, while the two Co ions are both 5-coordinate {CoO₅} and trigonal bipyramidal. The six dmf molecules sit in the hydrophobic calix[8]arene cavities, one terminally bonded to each of the Co and Th ions.

The most recent examples of TBC[8]-supported Co clusters are a family of heterometallic Co–Li/Na species reported in 2022.¹⁹ The complex [Co^II₂Li₈(OH)₂(TBC[8])Br₂(MeCN)₄(THF)₆] (19) was made from the reaction of CoBr₂, H₈TBC[8], and LiO^tBu in THF (Fig. S8, ESI†). The metallic skeleton of 19 describes three edge sharing [Co_xLi_4−x] squares, connected to one additional Li ion at each end. This gives it a ladder- or rod-like appearance. In the centre lies a [Co₂Li₂] square, two edges being singly bridged by a μ-O(Ph) atom from the TBC[8] ligand, with the other two edges being doubly bridged by two μ₄-OH⁻ and two μ-Br⁻ ions. The μ₄-OH⁻ ions further bridge to two Li ions on each side of the square, thus forming [CoLi₃] squares on each side of the central square. The three remaining edges of these squares are singly bridged by a μ_2/3-O(Ph) atom from the TBC[8] ligand. The μ₃-O(Ph) atom further bridges to a single Li ion at each end of the cluster, forming a [Li₃O] triangle. All ten metal ions are four coordinate and in distorted tetrahedral geometries. The coordination spheres of the Li ions are completed by a combination of THF and MeCN molecules. The TBC[8] ligand is fully deprotonated and adopts the pleated-loop conformation, with six/two O(Ph) atoms being μ_2/3-bridging, respectively.

The related complex [Co^II₂Li₆(OH)₂(TBC[8])Br₂(MeCN)₄] (20) can be obtained from the reaction of CoBr₂, LiO^tBu and H₈TBC[8] in Et₂O.¹⁹ The structure of 20 (Fig. S9, ESI†) is very similar to 18 but with the peripheral Li ions and terminal THF molecules being absent. Variations in reaction stoichiometry and reagents appear to play an interesting role in product identity for both 19 and 20, particularly with regard to the Co:Li ratio/positional disorder at the metal sites, such that several non-stoichiometric analogues have been isolated.¹⁹

The sole Na analogue in the family is the complex [Co^II₄Na(H₄TBC[8])₂Br(MeCN)₆] (21) made from the reaction of CoBr₂, H₈TBC[8] and NaH in THF.¹⁹ The structure of 21 (Fig. S10, ESI†) is highly unusual. The metallic skeleton is an irregular polyhedron, perhaps best described as a highly distorted {Co₄Na} square based pyramid, with a Co^II ion at the apex and three Co^II ions and one Na ion forming the square base. However, not all of these ions are directly connected to each other. The apical Co^II ion is linked to one Co^II ion in the square base through a μ-bridging Br⁻ ion and one μ-O(Ph) atom. In turn, this Co^II ion is bridged to the other Co^II ions in the square base via two μ-O(Ph) atoms from the same H₄TBC[8] ligand. The remaining five O(Ph) atoms are uncoordinated, sharing four H-bonds, with the ligand adopting a pseudo-chair loop conformation. The second calix[8]arene ligand is also quadruply deprotonated, H₄TBC[8], but does not bridge between any metal centres. Instead, it adopts the pleated-loop conformation in which each neighbouring pair of O(Ph) atoms chelate to a single metal centre. A total of four H-bonds are observed between the O(Ph) atoms. The four Co^II ions are of three types. Two are four coordinate and in distorted tetrahedral {CoO₃N} geometries, the fourth site being occupied by a MeCN molecule. There is an additional long contact to the μ-Br⁻ ion (∼3 Å). One Co ion is five coordinate and in a distorted trigonal planar {CoO₃NBr} geometry, again coordinating a MeCN molecule. The fourth Co ion is six coordinate and in a distorted octahedral {CoO₃N₂Br} geometry, with its coordination sphere completed by two MeCN molecules. The Na ion is four coordinate and in a distorted {NaO₂BrN} coordination geometry, with another MeCN molecule completing its coordination sphere.

Summary of Co-based clusters

The coordination chemistry of cobalt is (thus far) dominated by the Co^II ion, albeit with just one homometallic cluster, a dimer. All other reported compounds are heterometallic, being either Co-f block or Co-s block species. The prevalence of heterometallic over homometallic compounds is reflective of that seen for vanadium above, and remains a common theme in what follows below. The metallic skeletons of Co-based compounds are often based on dimers and tetramers. The latter can be squares, butterflies or tetrahedra, with larger species also being based on these units, e.g. square-based pyramids and square antiprisms. The calix[8]arene ligands in these clusters display a variety of conformations and protonation levels from the doubly deprotonated H₆TBC[8] to the fully deprotonated TBC[8], with the O(Ph) atoms bonding in both terminal (with neighbouring atoms chelating the same metal centre) and μ-bridging modes.

Fe-based clusters

All three Fe-based clusters in the CSD were published between 2007–2010 and are heterometallic, two include Ge and one includes K.^20–22 [Ge^II₄(TBC[8])][Fe⁰₂(CO)₈] (22; Fig. S11, ESI†) was synthesised from the reaction of [Ge₄(TBC[8])] with [Fe₂(CO)₉] in benzene.²⁰ Complex 22 largely retains the same overall structural features of the [Ge^II₄TBC[8]] starting material, in which two {Ge₂O(Ph)₂} rhombs occupy two separate halves of the molecule, related by a C₂ axis. The two {Fe(CO)₄} groups are bonded to the two Ge ions that are positioned outwith the central cavity of the TBC[8] ligand, pointing downwards with respect to the top of the ligand bowl. The TBC[8] ligand adopts a distorted saddle conformation, wrapping around the Ge ions such that four O(Ph) atoms are bonded to each {Ge₂} pair; two terminally bonded and two μ-bridging. The Fe atoms are formally zero valent, with all four Ge ions in the +2 oxidation state. Interestingly, the same reaction but using the para-unsubstituted H₈C[8] ligand results in the formation of [Ge^IV₄(C[8])][Fe^−I₂(CO)₈]₄ (23) (Fig. 10).²¹ In this case the reaction proceeds via (an unexpected) reductive decarbonylation, with extrusion of one bridging CO group in [Fe₂(CO)₉] and the formation of four [GeFe₂] three-membered rings, arranged in a tetrahedral fashion around the C[8] ligand. Formally, the reaction involves oxidation of Ge^II to Ge^IV and reduction of the diiron fragments to [Fe₂(CO)₈]²⁻.

Fig. 10 The structure of compound 23. Colour code: Fe = pale orange, Ge = teal, O = red, C = grey. H atoms omitted for clarity.

The final Fe-based cluster in the database is the complex [Fe^II₄K₄O₂(H₂O)₂(H₂TBC[8])₂(MeCN)₆] (24; Fig. S12, ESI†) which results from the reaction of [K₂Fe₂(O^tBu)₆(THF)₂] with H₈TBC[8] in toluene, followed by crystallisation from MeCN.²² The ASU in 24 contains a near-linear Fe–O–Fe dimer in which the μ₄-O²⁻ ion further bridges to two K ions lying in a perpendicular plane. Both Fe ions are four coordinate and in distorted tetrahedral {FeO₄} geometries, the other three O-atoms deriving from the H₂TBC[8] ligand. Symmetry expansion reveals that the four K ions form a central square rhomb, connected by four and two bridging MeCN and H₂O molecules, respectively. The hexa-deprotonated H₂TBC[8] ligands adopt the double-cone conformation, and present an unusual bonding motif. Four O(Ph) atoms are μ-bridging between the Fe and K ions, two are terminally bonded to Fe ions, and the two protonated O-atoms are non-bonded. In the latter case, the associated Ph group bonds side-on (η⁶) to a K ion.

Summary of Fe-based clusters

Here it is very clear that the coordination chemistry of H₈TBC[8] with Fe is almost completely unknown, with only three heterometallic complexes published, two rooted in Fe carbonyl chemistry, with the other being an Fe-s block complex. There are zero examples of homometallic Fe clusters in any oxidation state and no clusters containing more than four Fe centres. The calix[8]arene ligands are present in either their fully deprotonated, TBC[8]/C[8], or doubly protonated, H₂TBC[8], forms in distorted saddle and double-cone conformations, respectively. The O(Ph) atoms of the ligands adopt two bonding motifs – terminal, with neighbouring atoms chelating the same metal centre, and μ-bridging. In each case the latter directs the formation of {Fe₂} dimers.

Mn-based clusters

There are also only three Mn-based clusters in the CSD, all published between 2011–2013.^23–25 The first of these was the complex [Mn^III₂Na₄(H₅TBC[8])₂(CO₃)₂(dmf)₆] (25; Fig. 11) which was made from the reaction of MnCl₂·4H₂O, H₈TBC[8] and NaO₂CCMe₃ in a basic MeOH/dmf solution.²³ The pivalate anion is not incorporated into the structure but its presence appears vital for cluster formation. We note with interest that the well-known complex [Mn₆O₂(O₂CR)₁₀(solvent)₄] (R = CMe₃, Ph) is often a major/minor product in Mn-H₈TBC[8] reactions with certain carboxylates.²³ The metallic skeleton of 25 contains a central {Na₄} square, flanked on both sides by a Mn^III ion. The metals are linked together by two μ₅-CO₃²⁻ ions, formed from the fixation of atmospheric CO₂, which each bridge to all four Na ions and one Mn^III ion on the same face of the square. The remaining bridging in the {Mn₂Na₄} moiety comes from two μ₃-bridging and two μ-bridging dmf molecules. The former sit above/below the {MnNa₂} triangles in the two halves of the molecule. The latter bridge between two Na ions across an edge of the {Na₄} square. The presence of multiple bridging dmf molecules, and in two different coordination modes, is rather unusual. The calix[8]arene ligands adopt the pleated-loop conformation, sandwiching the inorganic core. They are triply deprotonated with six of the eight O(Ph) atoms being monodentate and each neighbouring pair chelating to a metal centre in the {MnNa₂} triangle. The remaining two (neighbouring) HO(Ph) arms are uncoordinated. The six coordinate Mn^III ions have their coordination spheres completed with two (trans) dmf ligands and are in Jahn–Teller (JT) distorted octahedral {MnO₆} geometries, the JT axis being defined by the O(dmf)–Mn–O(dmf) vector. Examination of the extended structure of 25 reveals the formation of large and aesthetically pleasing infinite stacked tubular assemblies, and that packing of neighbouring tubules results in the formation of solvent-filled channels within the structure.²³ The structure of 25 has some striking similarities to 24 (Fig. S12, ESI†) despite the different d- and s-block metals employed and the entirely different synthetic methodologies, highlighting perhaps the dominant structure directing role of calix[8]arene ligands. The presence of the bridging dmf molecules is similar to that seen in the vanadium chemistry described above.

Fig. 11 The structure of compound 25. Colour code: Mn^III = magenta, Na = purple, O = red, N = blue, C = grey. H atoms omitted for clarity.

The second Mn-based cluster published employed a heteroleptic ligand strategy which exploited the complementary ligand bridging modes of H₈TBC[8] and phenyl salicylaldoxime (Ph-saoH₂) to drive the formation of a heterovalent {Mn^IIIMn^IV} dimer.²⁴ Thus, the reaction of Mn(NO₃)₂·4H₂O, H₈TBC[8] and Ph-saoH₂ in a basic dmf/MeOH solution afforded [Mn^IIIMn^IV(H₄TBC[8])(Ph-sao)(μ-OMe)(dmf)₂] (26; Fig. 12). The magnetic core of 26 is a [Mn^IIIO(Me)O(Ph)(NO)Mn^IV] moiety in which the two Mn ions are triply bridged by a single μ-OMe ion, a single μ-O(Ph) atom from the calix[8]arene, and the –N–O– oxime unit of Ph-sao²⁻. The Mn^III ion has its coordination sphere completed by the terminally bonded alkoxide O-atom of Ph-sao²⁻ and a molecule of dmf, with the O(dmf)–Mn–O(Ph) vector defining the JT axis. There is also a single dmf coordinated to the Mn^IV ion. Both Mn ions are six coordinate and in distorted octahedral geometries. The quadruply-deprotonated H₄TBC[8] ligand adopts a conformation that is best described as being somewhere between a double-cone and a pleated-loop, wherein three phenol moieties adopt a cone orientation and the other five resemble a regular pleated-loop. The four deprotonated O(Ph) atoms adopt different coordination modes, three being monodentate and the fourth μ-bridging. The four protonated HO(Ph) arms are non-coordinated. It would appear therefore that the presence of the bulky phenolic oxime prevents full adoption of either of the more common calix[8]arene conformations, and restricts the nuclearity of the complex formed.

Fig. 12 The structure of compound 26. Colour code: Mn^III = magenta, Mn^IV = brown, O = red, N = blue, C = grey. H atoms omitted for clarity.

Magnetic susceptibility and magnetisation studies on 26 revealed the presence of a moderate ferromagnetic interaction between the two metal ions (J = +9.81 cm⁻¹, g = 1.99; formalism) resulting in the stabilisation of an S = 7/2 ground state. A fit of the magnetisation data in the context of the giant spin model suggested the ground state to have a significant zero-field splitting, D_S=7/2 = −0.91 cm⁻¹, with ac susceptibility data confirming slow relaxation of the magnetisation through the presence of temperature and frequency dependent out-of-phase (χ′′) signals below T ≈ 4 K. Complex 26 was therefore the first reported TBC[8]-based SMM.

The complex [Mn^II₅(H₄TBC[8])(OH)₂(OMe)_1.5(HCOO)(OAc)_0.5 (O₂CNMe₂)(dmf)₅] (27; Fig. 13) was synthesised via the solvothermal (T = 130 °C, 3 days) reaction of Mn(OAc)₂·4H₂O, H₈TBC[8] and KCl in a basic MeOH/dmf solution.²⁵ Neither K⁺ or Cl⁻ are incorporated into the structure of the product, but it would appear that addition of the salt is required for 27 to form. The dimethylcarbamate and formate anions are generated in situ from the decomposition of dmf/oxidation of MeOH, respectively. The metallic skeleton of 27 describes a {Mn₅} square based prism, with the apex connected to the square base via a combination of μ₃-OMe and μ₃-OH ions. The four Mn ions in the base are further connected by the μ₃-O₂CNMe₂ ligand that sits below the square. All five Mn^II ions are six coordinate and in distorted octahedral {MnO₆} geometries. Each has a single terminally coordinated dmf molecule directed above the square, with the apical Mn ion bonded to an additional carboxylate (OAc/O₂CH) ligand. The calix[8]arene is tetra-deprotonated, H₄TBC[8] and sits below the {Mn₄} square in the pleated-loop conformation, each neighbouring pair of O(Ph) atoms chelating to a single Mn^II ion. Magnetic susceptibility studies revealed the presence of weak antiferromagnetic exchange (J = −2 to 3 cm⁻¹) between neighbouring Mn^II ions.²⁵

Fig. 13 The structure of compound 27. Colour code: Mn^II = pale pink, O = red, N = blue, C = grey. H atoms omitted for clarity.

Summary of Mn-based clusters

Given that the coordination chemistry of H₄TBC[4] with Mn is very well established, and shown to be a hugely fruitful source of numerous polymetallic clusters,^1,26 the isolation of only three Mn-based complexes of H₈TBC[8] is surprising, in turn suggesting that a systematic investigation of this chemistry has not been performed. Drawing conclusions regarding structural trends is rather difficult, though two of the three metallic skeletons are based on squares. It is interesting to see that all three common Mn oxidations states (Mn^II, Mn^III and Mn^IV) exist in the three complexes isolated, and all the calix[8]arenes are only partially deprotonated, H₄TBC[8] or H₅TBC[8]. We also note that in two of the three complexes isolated, molecules not added to the reaction mixtures become incorporated into the structures – CO₃²⁻ from the fixation of CO₂, formate from the oxidation of MeOH and dimethylcarbamate from the breakdown of dmf. This perhaps suggests that the deliberate introduction of these (or similar) molecules would be worth exploring. Indeed, the presence of the CO₃²⁻ ion is a common theme in the Ln^III-based chemistry discussed below (and in Ln coordination chemistry more generally), but it is very rarely/intentionally employed as a bridging ligand.

Ni-based clusters

The sole Ni-based cluster in the CSD is the complex [Ni^II₂Th^IV₅O₄(OH)₂(H₁TBC[8])₂(dmf)₅(MeOH)₂] (28; Fig. 14).¹⁸ The metallic skeleton of 28 describes a highly distorted {Ni₂Th₄} octahedron, capped on one face by the fifth Th^IV ion, and stabilised by the presence of four O²⁻ ions which sit on the four equatorial edges of the {Ni₂Th₂} square. Three are μ₄-bridging (2 × {Th₃Ni}, 1 × {Th₂Ni₂}), and one is μ₃-bridging (1 × {Th₃}). The two Th^IV apices are connected to this square by eight {OR} alkoxides, one on each edge. Six are derived from μ-O(Ph) atoms of the calix[8]arene ligands, with the remaining two, on neighbouring edges, being μ₃-OH ions. The latter further bridge to the remaining Th^IV ion, which caps a {Ni₂Th} face of the octahedron. The hepta-deprotonated H₁TBC[8] ligands lie approximately perpendicular to each other, sandwiching the metal–oxygen core, and adopt the double-cone conformation. Six of the eight O(Ph) atoms are μ-bridging and two (neighbouring) O-atoms are monodentate. A solvent molecule completes the coordination at each metal centre, a dmf on the eight coordinate Th ions and a MeOH on the six coordinate Ni ions.

Fig. 14 The structure of compound 28. Colour code: Ni = green, Th = deep pink, O = red, N = blue, C = grey. H atoms omitted for clarity.

Summary of Ni-based clusters

The lack of any homometallic (or 3d–4f) Ni clusters is surprising, particularly due their potentially appealing magnetic properties. Perhaps it is even more surprising that there are no homometallic/heterometallic clusters of Ni with H₄TBC[4], both highlighting glaring omissions in the literature. A systematic exploration of Ni-based clusters with both H₄TBC[4] and H₈TBC[8] is required across a range of experimental conditions to map out the associated coordination chemistry of these systems.

Cr-based clusters

There is a single family of isostructural Cr-based clusters of formula [Cr^IIILn^III₆O₂(OH)(TBC[8])₂(H₂O)₄(dmf)₅(MeOH)₂] (Ln = Gd (29a), Tb (29b), Dy (29c), Tm (29d); Fig. 15),²⁷ the general metallic skeleton of which describes a capped distorted octahedron. The {Ln₅Cr} octahedron is bridged internally via two μ₄-O²⁻ ions; an alternative description of the core is therefore of two edge-sharing {Ln₄O}/{Ln₃CrO} tetrahedra. The remaining Ln ion caps a {Ln₂Cr} face, being coordinated to it through the presence of three μ₃-bridging O-atoms derived from H₂O molecules. The sole μ-bridging OH⁻ ion spans a {Ln₂} edge of the octahedron. Both fully deprotonated TBC[8] ligands adopt the double-cone conformation sandwiching the inorganic core of the molecule, lying perpendicular with respect to each other. They coordinate in two different ways. The first uses six of its O(Ph) atoms to μ-bridge between neighbouring metal ions, the two remaining O(Ph) atoms, positioned next to each other, being monodentate and chelating a single Ln ion.

Fig. 15 The structure of compound 29. Colour code: Cr = golden, Ln = light green, O = red, N = blue, C = grey. H atoms omitted for clarity.

The second TBC[8] ligand contains five monodentate O(Ph) atoms, all positioned next to each other, and three μ-bridging O(Ph) atoms. The Cr^III ion is six coordinate and in a distorted {CrO₆} octahedral geometry, the sixth site being occupied by a MeOH molecule. The Ln^III ions are of two types: five are eight coordinate and in square antiprismatic {LnO₈} geometries, and one is seven coordinate and in pentagonal bipyramidal {LnO₇} geometry, their coordination completed through the presence of dmf, MeOH and water molecules. Magnetic measurements revealed the Tb^III analogue (29b) to be a SMM. A fit of the ac susceptibility data to the Arrhenius law (τ = τ₀ exp[U_eff/(k_BT)]) revealed an effective energy barrier to magnetisation reversal of U_eff/k_B = 18.13 K with a pre-exponential factor of τ₀ = 7.5 × 10⁻⁸ s.²⁷

Summary of Cr-based clusters

As with the limited coordination chemistry of Ni^II, Fe^II/III, Co^II, Mn^II/III/IV and Co^II described above, the lack of any published homometallic Cr^III compounds of H₈TBC[8] (and H₄TBC[4]) reveals a serious gap in knowledge when it comes to polymetallic clusters of paramagnetic 3d ions. The isolation of 29 clearly shows this (and more 3d–4f chemistry) will be possible, and that it will lead to some exciting new compounds. We also note also that there are zero H₈TBC[8]-supported clusters of Cu. All of this chemistry remains to be explored, but is clearly rife for exploitation.

Ln-based clusters

The first Ln-based cluster published was [Eu^III₂(H₂TBC[8])(dmf)₅] (30) reported in 1987, made from the reaction of Eu(ClO₄)₃ and H₈TBC[8] in a basic dmf solution (Fig. S13, ESI†).²⁸ The two Eu^III ions are bridged by two μ-O(Ph) atoms from the H₂TBC[8] ligand and one μ-O(dmf) atom, affording a {Eu₂(OR)₃} core. The remaining six O(Ph) atoms on the calix[8]arene are monodentate, with two being protonated as evidenced by the significantly larger Eu–O bond lengths. The ligand adopts a double-cone conformation with near-twofold rotational symmetry. The Eu^III ions are both eight coordinate and in {EuO₈} bicapped trigonal prismatic geometries. The dimeric {Ln₂} unit is extremely common in the coordination chemistry of TBC[8] with lanthanide ions, and indeed there are multiple isostructural [Ln^III₂(H₂TBC[8])(solvent)₅] complexes reported which differ only in the coordinated solvent present (Ln = Eu, La, Pr, Lu, Tm; solvent = dmf, dmso; Fig. S14, ESI†).^29–32

A related monomeric complex, [Eu^III(H₆TBC[8])(NO₃)(dmf)₄] (31; Fig. S15, ESI†) can be made from the reaction of Eu(NO₃)₃, H₈TBC[8] and K₂CO₃ in a basic dmf solution at T = 60 °C.³³ The doubly deprotonated H₆TBC[8] ligand adopts a pleated-loop conformation, bonding to the Eu^III ion in a bidentate coordination mode. The metal is eight coordinate and in a distorted {EuO₈} square antiprismatic geometry, with the remaining coordination sites being filled with dmf molecules and a bidentate nitrate ion. A similar compound, [Gd^III(H₆TBC[8])Cl(dmso)₄] (32; Fig. S16, ESI†), can be made via the reaction of H₈TBC[8] with GdCl₃·6H₂O in a basic MeCN/dmso solution, followed by diffusion with hexane.³⁴ The structure of 32 is essentially the same as that in 31 but with a single, monodentate Cl⁻ ion replacing the bidentate NO₃⁻. The same paper also reported the synthesis of a number of Ln^III complexes varying in nuclearity form Ln₁–Ln₇ from a series of related chemical reactions. Interestingly, in a number of experiments the same general [Ln^III₂(H₂TBC[8])(solvent)₅] complex was obtained and the five new complexes discovered all contain a related {Ln₂} moiety which is in fact a smaller fragment in the construction of frameworks generally based on octahedra. In each case the calix[8]arene adopts either a perfect or near-perfect double-cone conformation, but with significant differences in the degree of deprotonation.

Reaction of H₈TBC[8] with Ce(NO₃)₃·6H₂O in a basic MeCN/dmf solution, followed by diffusion with hexane, affords the complex [Ce^IV₄O₂(H₂TBC[8])₂(dmf)₄] (33; Fig. 16).³⁴ At the centre of the molecule lies a non-planar/bowl-shaped {Ce₄O₂} butterfly. The Ce^IV ions are all seven-coordinate and in distorted {CeO₇} pentagonal bipyramidal geometries. Each H₂TBC[8] ligand exists as the hexa-anion, H₂TBC[8], with two intra-molecular H-bonds between neighbouring O(Ph) atoms per calix[8]arene molecule, and bridges three Ce^IV ions across one side of the butterfly. Four ligated dmf molecules complete the Ce^IV coordination spheres, occupying the cavities generated by each of the H₂TBC[8] molecules.

Fig. 16 The structure of compound 33. Colour code: Ce = lime green, O = red, N = blue, C = grey. H atoms omitted for clarity.

A similar reaction but employing TbCl₃·6H₂O in a basic MeCN/dmso solution affords the complex [Tb^III₅O(OH)₄(H₃TBC[8])Cl(dmso)₈(H₂O)₃]Cl₃ (34; Fig. 17).³⁴ The metallic skeleton of 34 describes a distorted square-based pyramid. The Tb^III ions in the square base are connected to each other via one μ₄-O²⁻ ion and four μ-O(Ph) atoms, and to the apical Tb^III ion through four μ₃-OH⁻ ions, one on each triangular face of the pyramid. The coordination sites of the metal ions are completed by the presence of several terminally bonded solvent molecules (dmso, H₂O), with the dmso molecules occupying the cavities generated by each of the H₃TBC[8] molecules. The Tb^III ions are eight coordinate and in square antiprismatic {TbO₈} or {TbO₇Cl} geometries. The three protonated arms of the H₃TBC[8] ligand are each H-bonding to one Cl⁻ counter anion. The latter are also H-bonded to the terminally bonded H₂O molecules. Perhaps surprisingly, the apical Tb^III ion is not coordinated to the H₃TBC[8] ligand at all, being connected to the rest of the molecule only through the four μ₃-OH⁻ ions. Its “exterior” square face is entirely comprised of four terminally ligated solvent molecules (3 × dmso, 1 × H₂O), perhaps suggesting the possibility of further cluster growth by solvent substitution.

Fig. 17 The structure of the cation of compound 34. Colour code: Tb = cyan, O = red, C = grey, S = dark yellow, Cl = yellow. H atoms omitted for clarity.

Repeating the reaction that produces 33 but using a different solvent combination (MeOH/dmf) and in the presence of NH₄ClO₄ results in the formation of [Ce^IV₆O₄(H₂TBC[8])₂(OMe)₄(dmf)₆] (35; Fig. 18).³⁴ The metallic skeleton of 35 is a slightly ‘flattened’ or ‘squashed’ {Ce₆} octahedron. The two apical sites are connected to the four ions in the square plane (and to each other) via two μ₄-O²⁻ ions, four μ-OMe⁻ ions and four μ-O(Ph) atoms from the H₂TBC[8] ligands. The remaining two, near linear, μ-O²⁻ ions each link neighbouring Ce^IV ions across an edge of the square plane. Each H₂TBC[8] exists as a hexa-anion – there is one intra-ligand H-bond between neighbouring O(Ph) atoms per molecule and one inter-ligand H bond – and bridges three Ce^IV ions across one triangular face of the octahedron. The dmf molecules again occupy the calixarene cavities. All six Ce^IV ions are eight coordinate and in square antiprismatic {CeO₈} geometries. The addition of NH₄ClO₄ to the reaction seems critical to the formation of 35 even though neither cation or anion is incorporated into the cluster. Repeating the reaction that produced 34 but using DyCl₃·6H₂O in a basic dmf/MeOH solvent combination and under solvothermal conditions (T = 130 °C, 24 h) results in the formation of [Dy^III₇O₂(OH)₄(H₁TBC[8])(H₂TBC[8])(dmf)₉] (36; Fig. S17, ESI†).³⁴ The metallic skeleton of 36 describes a distorted, capped octahedron. The central cavity is again occupied by two μ₄-O²⁻ ions, with the capping Dy^III ion connected to the octahedron through a combination of two μ₃-OH⁻ ions and one μ-OH⁻ ion. The remaining OH⁻ ion μ-bridges between two Dy^III ions in the square plane of the octahedron. The two calix[8]arene ligands are mono- and doubly deprotonated, respectively. In the former the single proton is shared between neighbouring O(Ph) atoms, both of which are terminally coordinated. This ligand bridges a total of six Dy^III ions. For the latter, the two protonated arms are bonded terminally with very large Dy⋯O distances. This ligand bridges a total of five Dy^III ions. The Dy^III ions are of two types: eight coordinate and in square antiprismatic {LnO₈} geometries, or seven coordinate and in pentagonal bipyramidal {LnO₇} geometries. The remaining metal coordination sites are filled with terminally bonded dmf molecules. The magnetic properties of 32, 34 and 36 were not reported.

Fig. 18 The structure of compound 35. Colour code: Ce = lime green, O = red, N = blue, C = grey. H atoms omitted for clarity.

The final Dy-based cluster reported is the complex [NaDy^III(H₆TBC[8])(OAc)₂(dmf)₃] (37; Fig. S18, ESI†), made from the reaction of Dy(OAc)₃·6H₂O, H₈TBC[8], 1,1,1-tris(hydroxymethyl)ethane (thmeH₃) and NEt₃ in dmf/ethyl acetate (Fig. S18, ESI†).³⁵ The doubly deprotonated H₆TBC[8] ligand adopts the pleated-loop conformation, with the two neighbouring deprotonated O(Ph) atoms chelating the Dy ion. The two OAc⁻ ligands also chelate to the Dy ion, using one O atom to μ-bridge to the Na ion. The coordination geometry of the Dy is completed with one molecule of dmf. The resulting {DyO₈} environment is best described as distorted triangular dodecahedron. The Na ion is five-coordinate and has a distorted square pyramidal {NaO₅} geometry, the remaining bonds comprising one molecule of dmf and two neighbouring, protonated O(Ph) atoms from the calix[8]arene. Interestingly, the {NaDy} lies outwith the H₆TBC[8] plane, perhaps due to the bridging acetates being located on the same side of the {NaDy} dimer. The role of the thmeH₃ ligand in the reaction is unclear, but its addition appears necessary for the isolation of the product. Complex 37 is a field induced SMM, albeit with a very small U_eff value.³⁵

Replacing DyCl₃·6H₂O with GdCl₃·6H₂O in an otherwise identical reaction to that which produces 36 leads to the isolation of the complex [Gd^III₈(H₁TBC[8])₂(CO₃)₄(HCOO)₂(dmf)₈] (38; Fig. 19).³⁴ The metallic skeleton of 38 conforms to the 26th Johnson solid (J₂₆), the gyrobifastigium. This 8-vertex polyhedron is composed of four squares and four triangles constructed around a central square plane. It is perhaps best envisioned as two triangular prisms sharing a square face, with one prism given a quarter-turn.

Fig. 19 The structure of the cation of compound 38. Colour code: Gd = cyan, O = red, N = blue, C = grey. H atoms omitted for clarity.

The metallic core is held together by a combination of two μ₅-CO₃²⁻, three μ₄-CO₃²⁻ and two syn, syn, μ-HCOO⁻ ligands. The CO₃²⁻ and HCOO⁻ ligands are derived from the fixation of atmospheric CO₂ and the oxidation of MeOH, respectively. Both are commonly observed in Ln cluster chemistry, and especially so under solvothermal (high temperature, high pressure) reaction conditions.³⁶ The two calix[8]arenes sandwich the inorganic core of the complex and are singly protonated, H₁TBC[8], the proton being shared between O(Ph) neighbours which are all are terminally coordinated. Each ligand bonds to a total of five Gd centres, all being seven coordinate and in distorted pentagonal bipyramidal {GdO₇} geometries. It is interesting to note that complex 38 contains no O²⁻/OH⁻ ions, in contrast to 33–36 despite the similarities in synthetic methodologies. It is also clear the that CO₃²⁻ ions generated in situ have acted as their “replacement”. In turn, this suggests that deliberately incorporating carbonate ions into this chemistry (and other Ln chemistry) would be a fruitful source of new complexes.³⁶ The magnetic properties of 38 were not reported.

By far the largest Ln-based cluster reported is [Ln^III₁₈O₃(OH)₁₂(TBC[8])₃(CO₃)₂Cl₆(H₂O)₆(dmf)₁₈Cl]OH (Ln = La (39a), Nd = (39b), Gd = (39c); Fig. 20), made from the reaction of [Ln^III₂(H₂TBC[8])(DMSO)₃(H₂O)] with trimesic acid in a basic CHCl₃/dmf/MeOH solution under solvothermal conditions (120 °C, 48 hrs).³⁷ The metallic skeleton of 39 describes an intercalated trigonal prism and an irregular polyhedron that consists of six equilateral triangles and eight isosceles trapezia. This framework is templated by a single Cl⁻ anion that sits in the centre of the cage. Twelve μ₃-OH⁻ ions, six μ₃-Cl⁻ ions, three μ₄-O²⁻ ions and two μ₃-CO₃²⁻ ions connect the Ln^III ions together in the core of the molecule. The cationic cluster has D_3h (trigonal planar) symmetry with the C₃ axis passing through the two central C-atoms of the CO₃²⁻ ligands.

Fig. 20 The structure of the cation of compound 39. Colour code: Ln = light green, O = red, N = blue, C = grey, Cl = yellow. H atoms omitted for clarity.

The three fully deprotonated TBC[8] ligands sit on the three triangular vertices, fully encapsulating the inorganic core and adopt the double-cone conformation. All eight O(Ph) atoms μ-bridge between neighbouring Ln^III ions. The remaining coordination sites on the Ln^III ions are occupied by terminally bonded H₂O and dmf molecules. The three unique Ln^III ions in the ASU unit adopt distorted square antiprismatic {LnO₈}, distorted square antiprismatic {LnO₆Cl₂}, and distorted capped square antiprismatic {LnO₈Cl} geometries. The OH⁻ counter anion is H-bonded to three coordinated water molecules that sit on one of the triangular {Ln₃} faces of the cluster. The release of Cl⁻/CO₃²⁻ ions during the reaction is clearly essential for the construction of the complex. These originate from the dehalogenation of CHCl₃/fixation of atmospheric CO₂. Although not incorporated into the structure, the addition of trimesic acid also appears to be vital as repetitions of the reaction without it fail to produce 39. The authors suggested that its presence in the reaction mixture may prevent the isolation of smaller TBC[8]-based Ln complexes.³⁷

The magnetic properties of 39b (Nd analogue) and 39c (Gd analogue) are dominated by the presence of very weak antiferromagnetic exchange interactions, with a Curie–Weiss constant, θ = −3.1 K for 39c. This was corroborated by heat capacity (c_p) measurements in which the zero-field magnetic contribution to c_p is nicely simulated by a Schottky model for eighteen S = 7/2 spins and an effective magnetic field B_eff = 0.6 T that mimics the magnetic interactions, gμ_BSB_eff = 2.8 K ≈ |θ|, for g = 2.0.

Summary of Ln-based clusters

An inspection of the metallic skeletons of the Ln-based complexes reveals the prevalence of dimeric {Ln₂} species and clusters based on octahedra or the component parts of octahedra – capped triangles, square-based pyramids, capped octahedra. For example, a simple comparison of Fig. 18–20 highlights some significant similarities. Calix[8]arene ligands have proven to be excellent scaffolds for the oxophilic Ln^III ions, and this is reflected in the number of homo- and heterometallic species isolated. They display a variety of conformations and a range of deprotonation levels. Unsurprisingly, the size/nuclearity of these species appears to be directly related to the deprotonation level of the ligand. Interestingly, however, there remains only six homometallic clusters with a nuclearity of four or more, {Ce^IV₄}, {Tb^III₅}, {Ce^IV₆}, {Dy^III₇}, {Gd^III₈} and {Ln^III₁₈}. The latter was made via a tandem templating strategy and a dimeric {Ln₂} starting material, perhaps pointing towards a method for constructing larger species. Given the general preference of TBC[4]-supported Ln^III ions to form octahedra, the combination of both H₄TBC[4] and H₈TBC[8] in the same reaction may prove to be complementary. Likewise, heterometallic 4f species, e.g. Ce^IV and Gd^III, may lead to some interesting discoveries.

Given that V-based TBC[8]-supported clusters have found utility in catalysis,^8–13 Ln-based species may also be of interest from that perspective. For example, complex 31 is monometallic and the assembly itself is in a chiral conformation.³³ Although it would require resolution of this species, if achieved, this would have potential utility in catalysis, as may other larger polymetallic Ln-based clusters.

Although not covered in this review (given the focus on structural and magnetic properties), it is worth noting that Ln-based calixarene complexes have fascinating luminescent properties which can be tuned/manipulated through synthetic alteration of the C[n] framework. A relatively recent review by Massi and Ogden highlights the antenna effect calixarenes can provide in such complexes, but also covers early work in which the photophysical properties of compound 30 in detail.³⁸ The Ln-based systems covered herein should be revisited to study such properties in detail, and indeed any further examples isolated should be studied to provide greater insight into how these species may be incorporated into materials, for example.

Conclusions

Systematic exploration of the coordination chemistry of H₄TBC[4] previously allowed us to establish a series of empirical metal ion binding rules that also translate very effectively to H₈-bis-TBC[4].^1,3 Coordination chemistry is promoted by the common cone conformation adopted by C[4]s, and this may benefit from synthetic modification at the methylene bridge positions in future as this (in principle) does not interfere with the polyphenolic pocket. The larger, odd-numbered homologues of H₄TBC[4] hold enormous potential for application in the synthesis of coordination clusters. However, greatly improved synthetic protocols are required for H_nTBC[n] where n is 7 or 9 as these would need to be scaled up for exploratory cluster synthesis with a wide range of 3d and 4f ions. In this review we have shown that, to date, a large number of clusters formed using H₈TBC[8] as a ligand have been synthesised. Interestingly, only a relatively small number have been prepared using benchtop or solvothermal conditions. The varying degrees of deprotonation possible in H₈TBC[8], coupled with the ability to form homo/heterometallic assemblies, suggests that a systematic study of the associated coordination chemistry of this ligand will unearth a huge number of new polymetallic clusters. For example, one only has to compare the Mn complexes of H₄TBC[4] with those formed with H₈TBC[8] to identify an area worthy of further investigation. Indeed, a systematic study of the entire 3d series is warranted. Such studies would ideally afford a set of empirical ion binding rules akin to those for H₄TBC[4] that, once in hand, could potentially be used to target hetero-calixarene assemblies, i.e. systems supported by both ligand types that display their own structural preferences. Inspection of the clusters reported herein already delivers some structural trends, e.g. a tendency towards dimers, butterflies and octahedra. Fig. S19–S24 (ESI†) collect all the metallic skeletons of the polymetallic complexes reported herein for comparison. Formation of the very large Ln₁₈ cluster suggests that it may be possible to access even larger assemblies that conform to highly symmetric solids capped by TBC[8]s in the double-cone conformation; this would be a direct extension from the well-documented structure-capping behaviour of TBC[4].

From a synthetic standpoint, a systematic exploration of the coordination chemistry of any chosen H_nTBC[n] should carefully examine the role of solvothermal conditions in the clusters that result from this chemistry. This synthetic approach clearly has the capability to deliver fascinating new species,³⁷ but achieving a far greater understanding of the influence of these conditions over the calixarene (cluster support) and components that are found in the products will be vital to (a) compiling a set of binding rules for any particular H_nTBC[n], and (b) tailoring the formation of desirable assemblies.

A broad review of coordination chemistry of the larger calixarenes by Redshaw in 2003 demonstrated how the field had developed to that point, identifying a wide range of potential routes of investigation.³⁹ Despite the number of species covered in this review, the coordination chemistry of H₈TBC[8] may still be considered embryonic. Work in our labs has begun a systematic study of the 3d and 4f ions, the results from which will be reported in due course. Ready access to other large(r), odd and even-numbered H_nTBC[n]s on a useable scale will also be pursued, as these ligands offer fascinating opportunities depending on ring size, preferred conformations, etc. However, it does not always follow that larger ligands will deliver larger clusters, an example being that H₁₆TBC[16] adopts a pseudo-spherical conformation in the solid state due to folding and ring size (with phenolic groups all pointing inwards).⁴⁰ This feature may hinder large cluster formation, so one must consider many factors when attempting to systematically programme the assembly of such complex systems. Nevertheless, calix[n]arene coordination chemistry remains a fascinating area for scientific exploration, and we hope this review stimulates activity in the area.

Data availability

No primary research results, software or code have been included and no new data were generated or analysed as part of this review.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

EKB and SJD thank the EPSRC for funding grants EP/I031421/1 & EP/I03255X/1. EKB thanks the Leverhulme Trust (RPG-2021-176).

Notes and references

1. L. R. B. Wilson, M. Coletta, M. Evangelisiti, S. Piligkos, S. J. Dalgarno and E. K. Brechin, Dalton Trans., 2022, 51, 4213 , and references therein.
2. I. Thondorf, Chapter 15 in Calixarenes 2001, ed. Z. Asfari, C. Böhmer, J. Harrowfield, J. Vicens, Kluwer Academic Publishers, Dordrecht, 2001.
3. I. V. Khariushin, V. Bulach, S. E. Solovieva, I. S. Antipin, A. S. Ovsyannikov and S. Ferlay, Coord. Chem. Rev., 2024, 513, 215846.
4. S. Fleming, C. D. Gutsche, J. M. Harrowfield, M. I. Ogden, B. W. Skelton, D. G. Stewart and A. H. White, Dalton Trans., 2003, 3319 , and references therein.
5. M. Coletta, E. K. Brechin and S. J. Dalgarno, in Chapter 25 in Calixarenes and Beyond, ed. P. Neri, J. L. Sessler and M.-X. Wang, Springer Nature, Switzerland, 2016.
6. A. Kieliszek and M. Malinska, Cryst. Growth Des., 2021, 21, 6862 , and references therein.
7. J. L. Atwood, G. A. Koutsantonis and C. L. Raston, Nature, 1994, 368, 229; T. Suzuki, K. Nakashima and S. Shinkai, Tet. Lett., 1995, 36, 249; X. Chen, R. A. Boulos, A. D. Slattery, J. L. Atwood and C. L. Raston, Chem. Commun., 2015, 51, 11413.
8. G. E. Hofmeister, E. Alvarado, J. A. Leary, D. I. Yoon and S. F. Pedersen, J. Am. Chem. Soc., 1990, 112, 8843.
9. V. C. Gibson, C. Redshaw and M. R. J. Elsegood, J. Chem. Soc., Dalton Trans., 2001, 767.
10. E. Hoppe, C. Limberg and B. Ziemer, Inorg. Chem., 2006, 45, 8308.
11. C. Redshaw, M. J. Walton, D. S. Lee, C. Jiang, M. R. J. Elsegood and K. Michiue, Chem. – Eur. J., 2015, 21, 5199.
12. T. Xing, M. Derbyshire, M. R. J. Elsegood and C. Redshaw, Chem. Commun., 2022, 58, 7427.
13. A. Ignaszak, N. Patterson, M. Radtke, M. R. J. Elsegood, J. W. A. Frese, J. L. Z. F. Lipman, T. Yamato, S. Sanz, E. K. Brechin, T. J. Prior and C. Redshaw, Dalton Trans., 2018, 47, 15983.
14. S. Petit, G. Pilet, D. Luneau, L. F. Chibotaru and L. Ungur, Dalton Trans., 2007, 4582.
15. S. C. Du, Y. F. Bi, Y. Yu and W. Liao, Sci. Sin.: Chim., 2012, 42, 1356.
16. H. Han, X.-L. Li, X. Zhu, G. Zhang, X. Hang, J. Tang and W. Liao, Eur. J. Inorg. Chem., 2017, 4879.
17. H. Han., Y.-S. Ding, X. Zhu, T. Han, Y.-Z. Zheng and W. Liao, Inorg. Chem. Front., 2020, 7, 4070.
18. S.-C. Du, H.-Q. Tan, Y.-F. Bi, Y. Yu and W.-P. Liao, Chin. J. Inorg. Chem., 2014, 30, 749.
19. A. Ignaszak, N. Patterson, C. O’Brien, A. True, M. R. J. Elsegood, T. J. Prior and C. Redshaw, RSC Adv., 2022, 12, 11672.
20. R. A. Green, A. L. Rheingold and C. S. Weinert, Inorg. Chim. Acta, 2009, 362, 3159.
21. A. E. Wetherby, L. R. Goeller, A. G. DiPasquale, A. L. Rheingold and C. S. Weinert, Inorg. Chem., 2007, 46, 7579.
22. A. Arbaoui, C. Redshaw, M. R. J. Elsegood, V. E. Wright, A. Yoshizawa and T. Yamato, Chem. – Asian J., 2010, 5, 621.
23. R. D. McIntosh, S. M. Taylor, S. Sanz, C. M. Beavers, S. J. Teat, E. K. Brechin and S. J. Dalgarno, Dalton Trans., 2011, 40, 12265; A. R. E. Baikie, A. J. Howes, M. B. Hursthouse, A. B. Quick and P. Thornton, J. Chem. Soc., Chem. Commun., 1986, 1587.
24. R. McLellan, S. M. Taylor, R. D. McIntosh, E. K. Brechin and S. J. Dalgarno, Dalton Trans., 2013, 42, 6697.
25. S. Du, H. Ke, Y. Bi, H. Tan, Y. Yu, J. Tang and W. Liao, Inorg. Chem. Commun., 2013, 29, 85.
26. G. Karotsis, S. J. Teat, W. Wernsdorfer, S. Piligkos, S. J. Dalgarno and E. K. Brechin, Angew. Chem., Int. Ed., 2009, 48, 8285; S. M. Taylor, G. Karotsis, R. D. McIntosh, S. Kennedy, S. J. Teat, C. M. Beavers, W. Wernsdorfer, S. Piligkos, S. J. Dalgarno and E. K. Brechin, Chem. – Eur. J., 2011, 17, 7521; M. A. Palacios, R. McLellan, C. M. Beavers, S. J. Teat, H. Weihe, S. Piligkos, S. J. Dalgarno and E. K. Brechin, Chem. – Eur. J., 2015, 21, 11212; S. M. Taylor, R. D. McIntosh, C. M. Beavers, S. J. Teat, S. Piligkos, S. J. Dalgarno and E. K. Brechin, Chem. Commun., 2011, 47, 1440; S. M. Taylor, J. M. Frost, R. McLellan, R. D. McIntosh, E. K. Brechin and S. J. Dalgarno, CrystEngComm, 2014, 16, 8098; S. M. Taylor, R. D. McIntosh, S. Piligkos, S. J. Dalgarno and E. K. Brechin, Chem. Commun., 2012, 48, 11190; R. McLellan, M. A. Palacios, S. Sanz, E. K. Brechin and S. J. Dalgarno, Inorg. Chem., 2017, 56, 10044; G. Karotsis, S. Kennedy, S. J. Dalgarno and E. K. Brechin, Chem. Commun., 2010, 46, 3884.
27. H. Han, X. Li, X. Zhu, G. Zhang, S. Wang, X. Hang, J. Tang and W. Liao, Eur. J. Inorg. Chem., 2017, 2088.
28. B. M. Furphy, J. M. Harrowfield, D. L. Kepert, B. W. Skelton, A. H. White and F. R. Wilner, Inorg. Chem., 1987, 26, 4233.
29. J. M. Harrowfield, M. I. Ogden and A. H. White and F. R. Wilner, Aust. J. Chem., 1989, 42, 949.
30. J. M. Harrowfield, M. I. Ogden and A. H. White, Aust. J. Chem., 1991, 44, 1237.
31. J. M. Harrowfield, M. I. Ogden and A. H. White, Aust. J. Chem., 1991, 44, 1249.
32. J.-C. G. Bünzli, F. Ihringer, P. Dumy, C. Sager and R. D. Rogers, J. Chem. Soc., Dalton Trans., 1998, 497.
33. J. M. Harrowfield, M. I. Ogden, W. R. Richmond and A. H. White, J. Chem. Soc., Dalton Trans., 1991, 2153.
34. S. M. Taylor, S. Sanz, R. D. McIntosh, C. M. Beavers, S. J. Teat, E. K. Brechin and S. J. Dalgarno, Chem. Eur. J., 2012, 18, 16014.
35. H.-S. Wang, Y. Chen, Z.-B. Hu, Q.-Q. Long, C.-L. Yin, Z.-C. Zhang and Z.-Q. Pan, Inorg. Chem. Commun., 2019, 105, 76.
36. L. Natarajan, J. Pecaut and M. Mazzanti, Dalton Trans., 2006, 1002; P. C. Andrews, T. Beck, C. M. Forsyth, B. H. Fraser, P. C. Junk, W. Massi and P. W. Roesky, Dalton Trans., 2007, 5651; X.-L. Tang, W.-H. Wang, W. Dou, J. Jiang, W.-S. Liu, W.-W. Qin, G.-L. Zhang, K.-B. Yu and L.-M. Zheng, Angew. Chem., Int. Ed., 2009, 48, 3499; T. N. Hooper, R. Inglis, M. A. Palacios, G. S. Nichol, M. B. Pitak, S. J. Coles, G. Lorusso, M. Evangelist and E. K. Brechin, Chem. Commun., 2014, 50, 3498.
37. Y. Jiao, S. Sanz, J. van Leusen, D. Gracia, A. B. Canaj, M. Evangelisti, E. K. Brechin, S. J. Dalgarno and P. Kögerler, Dalton Trans., 2024, 53, 4624.
38. M. Massi and M. Ogden, Materials, 2017, 10, 1369.
39. C. Redshaw, Coord. Chem. Rev., 2003, 244, 45.
40. C. Bavoux, R. Baudry, I. Dumazet-Bonnamour, R. Lamartine and M. Perrin, J. Inclusion Phenom. Macrocyclic Chem., 2001, 40, 221.

---

Footnote

† Electronic supplementary information (ESI) available: Table of compounds in the review and additional figures. See DOI: https://doi.org/10.1039/d4cc05459h

---