Ferrocene-Supported Compartmental Ligands for the Assembly of 3d/4f Complexes.

Site-specific coordination ligands, also known as compartmental ligands, have been used for the preparation of heterometallic complexes. These ligands, by virtue of possessing specific binding sites, can encapsulate different metal ions in their coordination pockets. Such compartmental ligands have been widely used for the preparation of heterometallic 3d/4f complexes which have applications in molecular magnetism. This Review summarizes our efforts in the use of ferrocene-based compartmental ligands for the preparation of heterometallic 3d/4f complexes, some of which are single-molecule or single-ion magnets.

Introduction

The field of molecular magnetism has received a significant boost with the discovery of the property of single-molecule magnetism (SMM) behavior in [Mn12O12(OAc)16(H2O)4] (1).[1a] It was shown that 1 possessed a total ground state spin, S = 10, with a significant Ising-type magnetic anisotropy (negative zero field splitting, D) characterized by a D value of ∼ −0.6 cm–1. The Ising-type anisotropy[1b] results in the m state with the highest S value (S = 10) as the ground state, which is predominantly populated upon magnetization and is also retained after the field is switched off. The effective energy barrier, Ueff, for the relaxation of magnetization has been estimated to be 60 K. The potential utility of such molecular magnets in various applications, such as high-density data storage,[2] quantum computation,[3] spintronics,[4] and so on, has propelled an intense interdisciplinary research activity in this area. Studies on SMMs since 1993 have focused on obtaining higher blocking temperatures (TB; the maximum temperature up to which an open hysteresis loop is observed in M vs H plots)[5] and higher-energy barriers (Ueff) for the reversal of magnetization.[5] It is now known that for polynuclear transition metal complexes, the effective energy barrier is related with the ground-state spin (S) and magnetic anisotropy (D) and could be formulated as Ueff = S2|D| {for integer spin}; Ueff = (S2 – 1/4)|D| {for half-integer spin}(Figure S1).[1b] Thus, large values of ground-state spin and magnetic anisotropy help in achieving a large energy barrier to the reversal of magnetization. This resulted in an almost exclusive focus on polynuclear transition metal complexes, at least in the initial stages of development of this research field.[6] However, it was very soon realized that increasing S does not scale linearly to an increasing D.[6] Therefore, it was important to focus on systems that would have significant magnetic anisotropy, as a result of strong spin orbit coupling and structural distortion.[6] This led to investigating complexes containing low-coordinate transition metal ion complexes as exemplified by Co(C(SiMe2ONaph)3)2 [C(SiMe2ONaph)3 see Chart S1][7a] (2) and [K(crypt-222)][Fe(C(SiMe3)3)2][7b] (3) as well as lanthanide complexes such as [DyLz2(o-vanilin)2]·NO3 [Lz = 6-pyridin-2-yl-[1,3,5]triazine-2,4-diamine] (4),[7c] [Dy(Cpttt)2][B(C6F5)4] { Cpttt= 1, 3, 4- tritertiarybutylcyclopentadienyl (5),[7d] [L2Ln(H2O)5][I]3·L2*·(H2O) [Ln= Dy (6); L* = (BuPO(NHPr)2)],[7e] (Et3NH)[(H2L)LnIIICl2] [H4L= 2,6-diacetylpyridine bis(salicylhydrazone) and Ln = Tb (7), Dy (8)].[7f] Another important issue that needed to be solved was the fact that irrespective of the height of the energy-barrier mechanisms such as quantum tunneling of magnetization, Raman and direct mechanisms exist that can undercut the barrier readily (Figure S1).[8] One of the ways of minimizing quantum tunneling is to increase the exchange coupling between the metal ions in di- and polynuclear ensembles.[8] However, in polynuclear lanthanide complexes, such interactions are very weak and can only be increased by the use of an open-shell bridging ligand. However, in the case of heterometallic 3d/4f complexes,[9] a reasonable exchange coupling can be anticipated between the 3d and the 4f metal ions.[9] This design enables the combination of different spin carriers within the same molecular entity that not only possess higher intramolecular exchange coupling due to the presence of the 3d metal ion but also an anisotropic center thanks to the lanthanide ions.[9]Ab initio (CASSCF+RASSI-SO+SINGLE_ANISO) calculations by Rajaraman et al. on eight isostructural 3d–4f complexes, [CuII(L1)(C3H6O)LnIII(NO3)3] (LnIII=DyIII (9), TbIII (10), HoIII (11), and ErIII (12) and [VIVO(L1)(C3H6O)LnIII(NO3)3] [LnIII=DyIII (13), TbIII (14), HoIII (15), and ErIII (16)] {H2(L1) = N,N′-bis(3-methoxysalicylidene)-1,3-diamino-2,2-dimethylpropane},[10] confirmed experimentally by Ishida et al.,[11] revealed that as the ferromagnetic exchange coupling increased the extent of quantum tunneling diminished. Even prior to this study, the presence of ferromagnetic interactions between 3d and 4f metal ions in a heterometallic complex was demonstrated both theoretically and experimentally.[9−11] The first example of an SMM based on a 3d–4f complex was a Cu2Tb2 (17) complex as reported by Osa et al. (Figure ).[12] After this discovery, several 3d–4f complexes containing paramagnetic 3d ions (such as CoII, MnII/III/IV, FeII/III, CrIII, NiII, CuII etc.),[13] and 4f ions (such as DyIII, TbIII, GdIII, HoIII and ErIII) were synthesized, and their interesting SMM properties were explored (Table S1 summarizes some selected examples). Subsequently, it has also been shown, that even if the 3d metal ion is diamagnetic, in some instances, the heterometallic 3d–4f complexes possess SMM properties.[13c]

Figure 1

A tetranuclear Cu2Tb2 complex, (17), that was shown to be a SMM.[22]

The synthesis of 3d–4f complexes can be accomplished effectively by designing ligands that have selective binding properties toward either transition metal ions or lanthanide metal ions. Such ligands, known as compartmental ligands, have proven to be extremely effective in the assembly of 3d/4f complexes. This Review is focused on a family of compartmental ligands that have been built on the ferrocene scaffold.[14−17]

Design of Compartmental Ligands for the Assembly of 3d–4f Complexes

Control of the nuclearity, as well as the geometry around the metal ions in a heterometallic complex, is crucial, and this can be best accomplished by utilizing rationally designed compartmental ligands. The ligands should have specific coordination sites organized in pockets or compartments that can then bind to either a transition or a lanthanide metal ion. Many such ligands have been assembled by condensation between aldehydes and amines resulting in Schiff-bases or related compounds (Figure ). Most commonly, a diamine or triamine (Figure a) is chosen and reacted with various salicylaldehyde derivatives (Figure c), affording the desired ligand.[18]

Figure 2

Schematic representation of synthesis of compartmental ligands (b), choice of amine (a) and aldehyde (c); site-preference binding of compartmental ligands (d).

Such SALEN-type ligands have been found to be useful in the preparation of various coordination complexes that have applications traversing from catalysis[19] to magnetism.[18]Figure d illustrates the design principle. The ligand, in this instance, has two specific coordination pockets, P1 and P2. The inner coordination pocket (P1) possess a N2O2 coordination unit and selectively binds with transition metal ions (3d), whereas the outer compartment possessing a O2O2′ coordination unit prefers 4f ions.

Utilizing this paradigm several multinuclear 3d–4f complexes have been synthesized [Figure ; (a): {Fe(3-MeOsaldmen)Gd(NO3)2}2(μ-O) (18);[20a] (b): CoIIYIII(μ-L2)(μ-OAc)(NO3)2 (19),[20b] L2 = 2-{(2-hydroxy-phenylimino)-methyl}-6-methoxyphenol; (c): {NiII(μ-L3)LnIII(DBM)3} (20),[20c] H2(L3) = N,N′-dimethyl-N,N′-(2-hydroxy-3-methoxy-5-methylbenzyl)ethylenediamine, DBM– = anion of 1,3-diphenylpropane-1,3-dione, LnIII= GdIII, TbIII, and DyIII; (d): (L4)ZnII(OAc)ZnII(NO3)ZnII(MeOH)ErIII(NO3) (H2O) (21),[20d] H6(L4) = macrocycle, synthesized by condensation of 2,3-dihydroxybenzene-1,4-dicarbaldehyde, (R,R)-1,2-diphenylethylenediamine; (e): {NiII(L5)LnIII(hfac)2(EtOH)} (22),[20e] H3(L5) = 1,1,1-tris[(salicylideneamino)methyl]ethane, LnIII = EuIII, GdIII, TbIII, and DyIII; hfac = hexafluoroacetylacetonate. Some of the complexes shown in Figure possess interesting SMM behavior.[20] It is worth mentioning that higher nuclearity can be achieved by judicial modification of either the bridging ligands or replacing the cluster terminating −OMe group with the cluster propagating −OH group.[17]

Figure 3

Line diagrams of 18 (a),[20a]19 (b),[20b]20 (c),[20c]21 (d),[20d] and 22 (e).[20e]

We have been working on the assembly of molecular magnets by utilizing various types of compartmental ligands. Thus, the trinuclear CoII2GdIII complexes [(L6)2CoII2GdIII][NO3] (23) were synthesized by the use of the ligand SP[N(Me)N = CH-C6H3-2-OH-3-OMe]3 ((Figure a).[21] We have also utilized other compartmental ligands such as [{NC(N(CH3)2)}2{NP{N(CH3)N CH-C6H3-(o–OH)(m-OCH3)}2}(H2L7)][22] (Figure b) and [N1,N3-bis(6-formyl-2-(hydroxymethyl)-4-methylphenol)diethylenetriamine (H5L8)] (Figure c).[23]

Figure 4

Representative examples of 3d–4f complexes afforded by utilization of various compartmental ligands: (a) [(L6)2CoII2GdIII][NO3] (23),[21] (b) [{(L7)CuIIDyIII(NO3)4}·CH3COCH3] (24),[22] and (c) NiII4DyIII4(H3L8)4(μ3–OH)4(μ2–OH)4]4Cl·30.6H2O·2CHCl3 (25).[23]

Synthesis of 3d–4f Complexes Utilizing Ferrocene-Based Compartmental Ligands

Ferrocene, the archetypical organometallic compound, has very rich chemistry because of several reasons.[24] The low rotational energy barrier of the organometallic bond and the ease of derivatization of the cyclopentadienyl motif make ferrocene extensively useful as an organometallic scaffold.[25] This has allowed ferrocene to be functionalized and utilized in several applications including in the design of new ligands, some of which have been used for preparing metal complexes with applications in catalysis.[24,25] Ferrocene-based ligands have been explored for the synthesis of various 3d or 4f metal complexes.[24,25] In spite of this, there were no reports on the utilization of ferrocene to design ligands that can be used for assembling 3d/4f complexes. Recognizing the opportunity, we have designed two new ferrocene-based ligands {H2(L9) and H4(L10)}. These were prepared by the reaction of 1,1′-ferrocene dihydrazone with o-vanillin or 2,3-dihydroxy benzaldehyde (Scheme ). These ligands possess two distinct coordination pockets suitable for selective binding of metal ions (Scheme ). Accordingly, utilizing these ligands, we synthesized four families of heterometallic complexes: [ZnII-LnIII], [NiII-LnIII], [CoII–YIII], and [ZnII4-LnIII4].[14−17] The structure and magnetic properties of these families of complexes are discussed in the following.

Scheme 1

Synthesis of H2(L9) and H4(L10)

Synthesis of Dinuclear ZnII-LnIII and NiII-LnIII Complexes[14,15]

The sequential reaction of H2(L9) with Zn(OAc)2·2H2O followed by Ln(NO3)3·6H2O (LnIII = DyIII (26), TbIII (27), and HoIII (28)][14] in the presence of triethylamine afforded, as planned, dinuclear complexes, [(L9)ZnII(μ-OAc)LnIII(NO3)2] (26–28) (Figure ).[14]

Figure 5

(a) Synthesis of heterometallic dinuclear complexes [{(L9)ZnII(μ-OAc)LnIII(NO3)2} LnIII = DyIII (26), TbIII (27) and HoIII (28)] and [{(L9)Ni(H2O)(μ-OAc)Ln(NO3)2}·MeCN, LnIII = DyIII (29), TbIII (30), GdIII (31), HoIII (32), and ErIII (33)]; (b) Molecular structures of 26 and (c) molecular structures of 29.[14,15] Reprinted from refs (14, 15). Copyright 2013, 2014 Royal Society of Chemistry.

The molecular structure of the dinuclear complex 26 is shown in Figure (b), which reveals that the ZnII ion occupies the inner coordination pocket and the lanthanide ion is found in the outer coordination pocket. ZnII has a coordination number of 5 (2N, 3O) in a distorted square pyramidal geometry, whereas the lanthanide ion has a coordination number of 9 in a tricapped trigonal prismatic geometry. The ZnII and the LnIII are connected to each other by two phenolate oxygen centers generating a {ZnLnO2} four-membered ring. This is bolstered by a bridging coordination action of the acetate anion. Studies on the magnetic properties of 26–28 revealed the absence of any single-ion magnet behavior.

We wished to examine if Zn(II) can be replaced by paramagnetic ions such as Ni(II) while retaining the overall structural features. An identical synthetic procedure as described above afforded the dinuclear complexes [{(L9)Ni(H2O)(μ-OAc)Ln(NO3)2}·MeCN, LnIII = DyIII (29), TbIII (30), GdIII (31), HoIII (32), and ErIII (33)].[15] The molecular structures of 29–33 reveal that unlike Zn(II) in the previously described examples, NiII has a coordination number of 6 (2N,4O), due to the presence of a water of coordination, and possesses a distorted octahedral geometry. The lanthanide ions have a coordination number of nine (9O) in a tricapped trigonal prismatic geometry. Magnetic properties of 29–33 were studied, which revealed that these possess ferromagnetic interactions at low temperatures. The exchange interaction between NiII ion and GdIII ion through the bridging oxygen reveal that a ferromagnetic interaction is present (JNi–Gd = 0.77 cm–1). However, the intermolecular antiferromagnetic interaction (JNi–Ni′ = −0.17 cm–1), due to presence of intermolecular H-bonding, leads to a sharp decrease of χT upon cooling below 5 K. Dynamic magnetic susceptibility measurements showed that while slow relaxation of magnetization does not occur at zero dc field, in the presence of an applied dc field (1000 Oe), indications of the onset of slow relaxation of magnetization are seen (Figure S2).[15]

CoII–YIII Complexes[16]

Single ion magnets (SIMs) based on CoII complexes are attracting interest in view of the strong spin–orbit coupling present in CoII which can be accentuated by appropriate coordination geometry.[6b] In view of this it was of interest to assemble a CoII–YIII complex where the only paramagnetic ion would be the CoII ion. The beneficial role of a diamagnetic metal ion in such an architecture has been examined by us and others previously.[13c] In view of the foregoing discussion on the capability of the ferrocene-supported ligand to form a dinuclear complex, the use of this ligand for preparing the targeted CoII–YIII complex seemed appropriate. Accordingly, we utilized H2(L9) to synthesize a family of CoII–YIII complexes [Co(μ-L9)(μ-CCl3COO)Y(NO3)2]·2CHCl3·CH3CN·2H2O (34), [Co(μ-L9)(μ-CH3COO)Y(NO3)2]·CH3CN (35), [Co(μ-L9)(μ-PhCOO)Y(NO3)2]·3CH3CN·2H2O (36) and [Co(μ-L9)(μ-tBuCOO)Y(NO3)2]·CHCl3·2H2O (37)]. The only difference in the complexes 34–37 is the change of the bridging carboxylate ligand (Figure e).[16]

Figure 6

Synthesis of 34–37 (e); Molecular structures of 34 (d), 35 (a) 36 (b), and 37 (c).[16] Reprinted from ref (16). Copyright 2019 American Chemical Society.

The solid-state structures of these complexes reveal that, as anticipated, CoII ions are held in the inner coordination pocket while YIII is bound in the outer pocket (Figure ). The coordination geometry around the CoII ion is seen to be distorted octahedral, and the magnitude of distortion increased gradually from 34–37. YIII has a coordination number of 9 and is present in a muffin shaped or distorted tricapped trigonal prismatic geometry.[16]

All the complexes showed out of phase susceptibility peak maxima in ac susceptibility measurements in the presence of the corresponding optimum applied magnetic fields (Figure ). Due to the variation in the coordination environment around the complexes caused by the alteration of substituents on the bridging carboxylate, the dynamics of the slow relaxation of magnetization in four complexes appeared to be varied (Figure S3 and Table S3).

Figure 7

Frequency dependence of χM′′ between 2 and 10 K for complexes 34–37 (a to d, respectively) under a 1200 Oe applied magnetic field.[16] Reprinted from ref (16). Copyright 2019 American Chemical Society.

Ab initio calculations revealed that all the complexes have a positive zero-field splitting parameter value (D) along with high transverse anisotropy (E). The values and sign of the parameters were consistent with the experimental values. The electronic state calculation on the complexes showed a multiconfigurational nature of the states (Figure S4) that explains the origin of high transverse anisotropy in the complexes. This high transverse anisotropy promotes fast relaxation of magnetization which also explains why all these complexes were not zero-field SIMs.

Octanuclear Zn4-Ln4 Complexes.[17]

We have seen that the −OMe group present in the H2(L9) is a monodentate ligand and does not participate in bridging coordination and hence the propagation of the nuclearity of the complexes is restricted to two. It seemed reasonable that the replacement of the −OMe group with −OH would result in a situation more favorable to cluster propagation. Such subtle changes in ligand design leading to large changes in the resultant complexes are quite useful and instructive. In this regard, we have synthesized H4(L10) by condensation of 1,1′ ferrocene dihydrazone and 2,3-dihydroxy benzaldehyde (Scheme ).[14−17] Treatment of H4(L10) with Zn(OAc)2 and Ln(NO3)3·6H2O in the presence of triethylamine (to effect complete deprotonation), afforded an iso-structural Zn4Ln4 family [Zn4–Dy4 (38), Zn4–Tb4 (39) and Zn4–Ho4 (40)]. In all of these complexes, the dinuclear core, [(L10)Zn(μ-OAc)Ln], with an acetate bridge remains intact. The newly added phenoxy group functions as a bridging ligand to bind another [(L10)Zn(μ-OAc)Ln] motif to generate the octanuclear complex. Despite such an interesting structure, none of these complexes were SMMs as revealed by detailed magnetic studies (Figure ).

Figure 8

(a) Synthesis of octanuclear Zn4-Ln4 complexes [Zn4–Dy4 (38), Zn4–Tb4 (39), and Zn4–Ho4 (40)]; (b) Molecular structure of 38.[17] Reprinted from ref (17). Copyright 2016 Royal Society of Chemistry.

Summary

The assembly of 3d/4f complexes, where the 3d and the 4f metal ions are connected to each other by a single atom bridge is best accomplished using compartmental ligands that have specific coordination pockets as well as appropriate coordinating motifs. We have explored the use of the ferrocene scaffold for the design of a new family of compartmental ligands which enabled the ready assembly of several dinuclear 3d/4f complexes. Some of these were shown to be a new family of single-molecule/single-ion magnets. An interesting aspect of our study included the modulation of the ligand by replacement of the −OMe group by −OH. Even such a small change in the ligand leads to a large change in the final nuclearity of the complex. In this instance, the nuclearity expands from 2 to 8. Ligand design, thus, remains at the heart of coordination assemblies and hence extracting new properties from them.