Chiral, Heterometallic Lanthanide–Transition Metal Complexes by Design

: Achieving control over coordination geometries in lanthanide complexes remains a challenge to the coordination chemist. This is particularly the case in the ﬁeld of molecule-based magnetism, where barriers for magnetic relaxation processes as well as tunneling pathways are strongly inﬂuenced by the lanthanide coordination geometry. Addressing the challenge of design of 4f-element coordination environments, the ubiquitous Ln(hfac) 3 moieties have been shown to be applicable as Lewis acids coordinating transition metal acetylacetonates facially leading to simple, chiral lanthanide–transition metal heterodinuclear complexes. The broad scope of this approach is illustrated by the synthesis of a range of such complexes LnM : LnM(hfac) 3 ( µ 2 -acac-O,O,O (cid:48) ) 3 (Ln = La, Pr, Gd; M = Cr, Fe, Ga), with approximate three-fold symmetry. The complexes have been crystallographically characterized and exhibit polymorphism for some combinations of 4f and 3d metal centers. However, an isostructural set of systems spanning several lanthanides which exhibit spontaneous resolution in the orthorhombic Sohncke space group P 2 1 2 1 2 1 is presented here. The electronic structure and ensuing magnetic properties have been studied by EPR spectroscopy and magnetometry. The GdFe , PrFe , and PrCr complexes exhibit ferromagnetic coupling, while GdCr exhibits antiferromagnetic coupling. GdGa exhibits slow relaxation of the magnetization in applied static ﬁelds.


Introduction
Recent years have seen a revival of lanthanide coordination chemistry, partially fueled by their promising magnetic properties [1][2][3]. Thus, mononuclear lanthanide complexes, polynuclear lanthanide complexes, lanthanides coordinating organic radicals, and mixed 3d-4f metal complexes have all yielded single molecule magnets [2][3][4][5][6]. Among them, only few exhibit chirality [7][8][9][10]. In many regards, lanthanide-based systems outperform molecular magnets based on transition metals, but in the context of magnetic properties, the flexibility of the lanthanide ions concerning coordination numbers and coordination geometries complicates matters in more than one sense. Firstly, as the ligand field splitting in lanthanide complexes is energetically subordinate to the interelectronic repulsion and spin-orbit coupling, the coordination geometries and ensuing ligand fields become determining for the magnetic properties. Hence, tuning of magnetic properties requires a degree of tailoring the coordination environments, which is often difficult to realize for lanthanides. Secondly, the pliant geometries around lanthanide ions hamper the design of highly symmetrical systems, which can act as building blocks for extended structures since any symmetry imposed by multidentate ligands is easily Inorganics 2018, 6, 72 2 of 12 broken. Thus, although crystallographically strictly axial, trigonal, and tetragonal [11][12][13][14][15][16][17][18][19][20] lanthanide complexes have been obtained by the use of structure-directing ligands, these systems typically feature coordinatively saturated lanthanide centers with no obvious avenues towards extended structures which preserve the tailored symmetry. Addressing this challenge, we forward a simple route towards heterobinuclear transition metal-lanthanide complexes with near-axial symmetry.
Inorganics 2018, 6, x FOR PEER REVIEW 2 of 12 symmetrical systems, which can act as building blocks for extended structures since any symmetry imposed by multidentate ligands is easily broken. Thus, although crystallographically strictly axial, trigonal, and tetragonal [11][12][13][14][15][16][17][18][19][20] lanthanide complexes have been obtained by the use of structure-directing ligands, these systems typically feature coordinatively saturated lanthanide centers with no obvious avenues towards extended structures which preserve the tailored symmetry. Addressing this challenge, we forward a simple route towards heterobinuclear transition metal-lanthanide complexes with near-axial symmetry. The hexafluoroacetylacetonate (hfac) complexes of the trivalent lanthanide ions are ubiquitous starting materials. While the homoleptic neutral entities Ln(hfac)3 are unknown in the condensed phase, many neutral heteroleptic complexes Ln(hfac)3Ln with lanthanide coordination numbers ranging from 7 to 10 [21][22][23][24] have been structurally characterized. Among these, eightfold coordination dominates as exemplified by the aquo-complexes Ln(hfac)3(H2O)2 [25] (cf. Figure 1a). Indeed, the Lewis acidic fragment Ln(hfac)3 has been employed as a building block for several structures by ligation of bidentate transition metal complexes creating heterometallic bi-and polynuclear systems in a systematic fashion [26,27] where one simple system is illustrated in Figure 1b [28]. However, to date, only a single example of a triply bridged system based on the Ln(hfac)3 fragment, namely CuLa(μ2-acac-O,O,O′)2(μ-H2O) (hfac)3 (acac = acetylacetonate), entailing nine-coordination of the large La 3+ ion, has been reported (cf. Figure 1c) [29]. Common to the previously studied systems is the geometric incompatibility of the transition metal and the lanthanide fragments, which inevitably lowers the overall symmetry to C1. Notable is also the system [(NiL)Gd(hfac)2(EtOH)] (L = μ2-1,1,1-tris(N-salicylideneaminomethyl)ethane) [30], where the trifold symmetric Schiff-base ligand provides too much steric encumbrance to preserve three hfac ligands on the lanthanide. Inspired by the above-mentioned CuLa(μ-acac)2(μ-H2O)(hfac) 3 and by examples of triple bridging between two Ln(hfac)3 fragments in, e.g., (hfac)3Ln(μ-PyO)3Ln(hfac)3 (Ln = Eu, Dy; cf. Figure 1d) [31], we decided to pursue the possibility of facial coordination of sterically undemanding, threefold symmetric transition metal complexes terminated in hard oxygen donor ligands to Ln(hfac)3 fragments aiming at threefold symmetric dinuclear systems. This approach necessitates the use of non-competing and hence weakly coordinating solvents, imposing some restrictions on the suitable transition metal building blocks. Common to the previously studied systems is the geometric incompatibility of the transition metal and the lanthanide fragments, which inevitably lowers the overall symmetry to C 1 . Notable is also the system [(NiL)Gd(hfac) 2 (EtOH)] (L = µ 2 -1,1,1-tris(N-salicylideneaminomethyl)ethane) [30], where the trifold symmetric Schiff-base ligand provides too much steric encumbrance to preserve three hfac ligands on the lanthanide. Inspired by the above-mentioned CuLa(µ-acac) 2 (µ-H 2 O)(hfac) 3 and by examples of triple bridging between two Ln(hfac) 3 fragments in, e.g., (hfac) 3 Ln(µ-PyO) 3 Ln(hfac) 3 (Ln = Eu, Dy; cf. Figure 1d) [31], we decided to pursue the possibility of facial coordination of sterically undemanding, threefold symmetric transition metal complexes terminated in hard oxygen donor Inorganics 2018, 6, 72 3 of 12 ligands to Ln(hfac) 3 fragments aiming at threefold symmetric dinuclear systems. This approach necessitates the use of non-competing and hence weakly coordinating solvents, imposing some restrictions on the suitable transition metal building blocks. Evidently, uncharged complexes with three mono-anionic, bidentate ligands meet both the symmetry and solubility requirements and pose an obvious starting point. Indeed, this approach proved feasible and focusing on the pervasive tris-acetylacetonates of the trivalent transition metal ions, a subset of such systems which indeed conforms to the designed threefold pseudosymmetry is presented here (Figure 2). Evidently, uncharged complexes with three mono-anionic, bidentate ligands meet both the symmetry and solubility requirements and pose an obvious starting point. Indeed, this approach proved feasible and focusing on the pervasive tris-acetylacetonates of the trivalent transition metal ions, a subset of such systems which indeed conforms to the designed threefold pseudosymmetry is presented here (Figure 2).

Results and Discussion
Synthesis of the face-sharing mixed 3d-4f metal acetylacetonates LnM(hfac)3(μ-acac)3 proceeds in a facile manner by dehydration of the Ln(hfac)3(H2O)2 precursor through azeotropic distillation with, e.g., benzene in the presence of the transition metal acetylacetonate. In this fashion, triply bridged binuclear complexes have been accessed with Ln = La through Tb. The method provides moderate to good yields (30-84%) of well-crystalline products, which, if needed, can be recrystallized from non-coordinating solvents (e.g., chloroform, hexane…). The heterobinuclear complexes exhibit polymorphism and in the present contribution, we have chosen to focus on an orthorhombic phase which has been obtained pure for Ln = La, Pr, Nd, Sm, Eu, Gd, and Tb with M = Cr, Fe, Ga. This phase can be reproducibly obtained by use of the crystallization procedure described in the experimental section. Other phases belonging to the monoclinic and triclinic crystal systems have been observed for some combinations of 3d and 4f metals, but these will not be discussed here. The heterobinuclear complexes LnM(hfac)3(μ-acac)3 were obtained spontaneously resolved in the Sohncke space group P212121. The 65 Sohncke space groups (often called chiral space groups) contain only rotation or screw axes. They are thus the only groups in which an enantiopure compound crystallize, and hence also the only groups allowing spontaneous resolution. The resolution occurs by conglomerate formation as demonstrated by structure determination of both enantiomorphs of PrFe (cf. Supplementary Materials). Illustrations of the molecular structure and packing of Λ,Δ-GdFe are given in Figure 3. Tables S1-S3 report the relevant crystallographic data for all the described complexes.

Results and Discussion
Synthesis of the face-sharing mixed 3d-4f metal acetylacetonates LnM(hfac) 3 (µ-acac) 3 proceeds in a facile manner by dehydration of the Ln(hfac) 3 (H 2 O) 2 precursor through azeotropic distillation with, e.g., benzene in the presence of the transition metal acetylacetonate. In this fashion, triply bridged binuclear complexes have been accessed with Ln = La through Tb. The method provides moderate to good yields (30-84%) of well-crystalline products, which, if needed, can be recrystallized from non-coordinating solvents (e.g., chloroform, hexane...). The heterobinuclear complexes exhibit polymorphism and in the present contribution, we have chosen to focus on an orthorhombic phase which has been obtained pure for Ln = La, Pr, Nd, Sm, Eu, Gd, and Tb with M = Cr, Fe, Ga. This phase can be reproducibly obtained by use of the crystallization procedure described in the experimental section. Other phases belonging to the monoclinic and triclinic crystal systems have been observed for some combinations of 3d and 4f metals, but these will not be discussed here.
The heterobinuclear complexes LnM(hfac) 3 (µ-acac) 3 were obtained spontaneously resolved in the Sohncke space group P2 1 2 1 2 1 . The 65 Sohncke space groups (often called chiral space groups) contain only rotation or screw axes. They are thus the only groups in which an enantiopure compound crystallize, and hence also the only groups allowing spontaneous resolution. The resolution occurs by conglomerate formation as demonstrated by structure determination of both enantiomorphs of PrFe (cf. Supplementary Materials). Illustrations of the molecular structure and packing of Λ,∆-GdFe are given in Figure 3. Tables S1-S3 report the relevant crystallographic data for all the described complexes. only rotation or screw axes. They are thus the only groups in which an enantiopure compound crystallize, and hence also the only groups allowing spontaneous resolution. The resolution occurs by conglomerate formation as demonstrated by structure determination of both enantiomorphs of PrFe (cf. Supplementary Materials). Illustrations of the molecular structure and packing of Λ,Δ-GdFe are given in Figure 3. Tables S1-S3 report the relevant crystallographic data for all the described complexes.   Figure S1. For metrics cf. Table 1.
Notably, of the more than 850 known structures containing the Ln(hfac) 3 fragment only 38 crystallize in one of the Sohncke space groups and only two heterobimetallic Ln(hfac) 3 complexes, both of which are Cu(II)-containing chains, have been found to show spontaneous resolution [32,33]. It must be concluded that the weak dispersion forces from the heavily fluorinated Ln(hfac) 3 fragments are not efficient in enforcing chiral packings. In line with this conclusion, the majority of short inter-molecular contacts in the packing of LnM involve the acetylacetonate ligands, alone or in addition to the hexafluoroacetylacetonate ligands (see Supplementary Materials, Figure S2 for details). In all cases, the ligand sphere around the d-or p-block metal is ordered, but in some cases the coordination sphere around the lanthanide is disordered with both helicities represented. In the systems LaFe, PrFe, and GdFe, no disorder is observed and the chirality of the two metal centers is invariably opposite. We hypothesize that the size of the light-metal fragment governs the locking of the relative configurations around the metal centers and the concomitant order in the lanthanide coordination sphere. The metric parameters for select combinations of metal ions are collected in Table 1. In the first column of Table 1, bond length ranges are given for the d-block metals, including Ga. The geometric perturbation of the M(acac) 3  higher geometric pliancy of Ga(III), and especially Fe(III) as compared to Cr(III), which is expected on the basis of the strong preference of Cr(III) for regular octahedral coordination. The second and third columns of Table 1 summarize the coordination geometry around the lanthanide ions. The observed contraction of ca. 6% in Ln-O acac distances upon going from La to Gd is slightly accentuated over the corresponding contraction of the Ln-O hfac distances. Again, the geometric rigidity of Cr(acac) 3 differentiates this metalloligand from the iron and gallium analogs resulting in longer Ln-O acac bonds for the Cr(III) systems.
Although the binuclear complexes do not have crystallographic threefold symmetry, they approximate this situation quite closely. Thus, angles between centroids of the three-oxygen donor faces of the distant hfac-ligator atoms, the bridging acac-ligator atoms, and the terminal acac-ligator atoms range from 176.10 • to 177.59 • , indicating near co-linearity between the approximate threefold axes of the metal centers. The intermetallic distances span the interval 3.321-3.514 Å with distinct dependence on the nature of both metal centers as also reflected in the first coordination sphere bond distances (vide supra). Despite the only partial charge on the bridging donor atoms, the intermetallic distances are relatively short compared to other triply bridged lanthanide-transition metal complexes as illustrated in Figure 4. The relative proximity of the metal ions as well as the possibility of obtaining isostructural members of the series with diamagnetic centers at either position renders these systems well-suited for the investigation of magnetic properties, in particular 3d-4f magnetic interactions.  The magnetic properties of all these compounds were investigated using dc and ac magnetometry. In Figure 5a Figures S3 and S4). This allows a simple qualitative evaluation of the nature of the coupling that instead dominates the shape of the χT curves of the GdM (M = Ga, Fe, or Cr) derivatives: ferromagnetic (FM) in GdFe and antiferromagnetic (AFM) in GdCr. The nature of the coupling is more difficult to extract at a glance for the complexes containing Pr 3+ due to its pronounced magnetic anisotropy. However, the low-temperature value of the χT product of PrGa (Figure 5a) goes to zero when the temperature is decreased, and the magnetization curves ( Figure  S5a) almost coincide at all temperatures, suggesting that the effect of the ligands is to stabilize the diamagnetic mJ = 0 ground state of Pr 3+ . No hysteresis was observed in any of the samples at the lowest accessible temperature (1.8 K).
Simulation and parametrization of the magnetic data was performed using a Hamiltonian with the general expression: The magnetic properties of all these compounds were investigated using dc and ac magnetometry. In Figure 5a Figures  S3 and S4). This allows a simple qualitative evaluation of the nature of the coupling that instead dominates the shape of the χT curves of the GdM (M = Ga, Fe, or Cr) derivatives: ferromagnetic (FM) in GdFe and antiferromagnetic (AFM) in GdCr. The nature of the coupling is more difficult to extract at a glance for the complexes containing Pr 3+ due to its pronounced magnetic anisotropy. However, the low-temperature value of the χT product of PrGa (Figure 5a) goes to zero when the temperature is decreased, and the magnetization curves ( Figure S5a) almost coincide at all temperatures, suggesting that the effect of the ligands is to stabilize the diamagnetic m J = 0 ground state of Pr 3+ . No hysteresis was observed in any of the samples at the lowest accessible temperature (1.8 K).
Simulation and parametrization of the magnetic data was performed using a Hamiltonian with the general expression: where the first summation runs over all the paramagnetic centers and contains the Zeeman interaction and Zero Field Splitting (ZFS). The second term, only used for complexes containing two paramagnetic ions, is an isotropic ferromagnetic (j < 0) or antiferromagnetic (j > 0) coupling [34]. For Pr 3+ , the S quantum number in Equation (1) was replaced with the total angular momentum of the ground multiplet J = 4. The parameters employed in the simulations are reported in Table 2. The errors in the parameter values were estimated to be on the last digit reported in Table 2. The evaluation was obtained varying the relevant parameters in the simulations and monitoring the resultant change in the reproduction of the data.  For LaCr and LaFe, an axial ZFS parameter (D) and an isotropic g factor were sufficient to reproduce the shape of both the χT and magnetization curves. The simulations of LaFe, LaCr, and GdGa did not allow for a unique determination of the sign of the ZFS. However, previous studies on Cr(acac) 3 and Fe(acac) 3 suggest a negative ZFS [35,36]; thus, we adopted that sign. As expected, the half-filled Fe 3+ ion has a lower magnetic anisotropy and a g factor closer to 2 compared to the Cr 3+ ion. The small anisotropy expected in GdGa required the use of a more sensitive technique to be accurately unraveled. We thus performed X-band EPR spectroscopy, obtaining the spectrum reported  Figure 6. The simulation provided an estimation of the magnitude of the ZFS but not its sign (the magnetic measurements on GdCr and GdFe suggest a positive ZFS, vide infra).
For LaCr and LaFe, an axial ZFS parameter (D) and an isotropic g factor were sufficient to reproduce the shape of both the χT and magnetization curves. The simulations of LaFe, LaCr, and GdGa did not allow for a unique determination of the sign of the ZFS. However, previous studies on Cr(acac)3 and Fe(acac)3 suggest a negative ZFS [35,36]; thus, we adopted that sign. As expected, the half-filled Fe 3+ ion has a lower magnetic anisotropy and a g factor closer to 2 compared to the Cr 3+ ion. The small anisotropy expected in GdGa required the use of a more sensitive technique to be accurately unraveled. We thus performed X-band EPR spectroscopy, obtaining the spectrum reported in Figure 6. The simulation provided an estimation of the magnitude of the ZFS but not its sign (the magnetic measurements on GdCr and GdFe suggest a positive ZFS, vide infra). The ZFS parameters and g factors obtained for the mono-paramagnetic complexes were fixed for the simulation of the poly-paramagnetic species, allowing for the extraction of the coupling constants reported in Table 2. For the GdCr and GdFe derivatives, a positive ZFS of Gd 3+ was found The ZFS parameters and g factors obtained for the mono-paramagnetic complexes were fixed for the simulation of the poly-paramagnetic species, allowing for the extraction of the coupling constants reported in Table 2. For the GdCr and GdFe derivatives, a positive ZFS of Gd 3+ was found to yield an improved goodness of the simulation compared to the negative case. The coupling in LnFe is significantly weaker than the one in LnCr for both Ln = Gd and Pr. The characteristic feature of the AFM coupling in GdCr (the only derivative that exhibits AFM coupling) is the step-like shape of the magnetization curve in Figure 5: indeed, a field of ca. 26.2 kOe is required to force the magnetic moments to be aligned with the field. A similar behavior with a spin rearrangement corresponding to a ferri-to-ferromagnetic transition at a comparable field was observed both by susceptometry and XMCD for the fluoride-bridged [Dy(hfac) 3 (H 2 O)-CrF 2 (py) 4 -Dy(hfac) 3 (NO 3 )] [37]. The magnitude of the coupling constants is well-comparable with other lanthanide-transition metal complexes connected by monoatomic bridges [38,39].
The magnetic anisotropy of Pr 3+ could not be fully modelled with the available data; thus, we tentatively simulated it using a large positive D parameter (80 cm −1 ) to mimic the diamagnetic ground state and an isotropic Landé factor (g J = 4/5). The introduction of a coupling term allowed for satisfactory reproduction of the magnetic behavior of both PrCr and PrFe, revealing FM coupling in both compounds.
The ac measurements conducted on all the samples evidenced slow relaxation only in GdGa, reported in Figure S6 (B = 0-5000 Oe). The field scan at 1.8 K revealed the presence of a double peak in the imaginary part of the magnetic susceptibility (χ"), corresponding to two active relaxation processes ( Figure S6b), as previously observed in a Gd-EDTA complex. [40] The available experimental frequency window allowed us only to monitor the slow process. A temperature scan with an optimum field ( Figure S7) of 4000 Oe reveals that the complex exhibits a single relaxation time up to 2.8 K, conversely to Gd-EDTA. [40] The origin of the slow relaxation can be attributed either to quantum tunneling (QT) or to a phonon bottleneck since they have a similar ac response. [41] The field and temperature dependence of the slow process was consequently parametrized using the theoretical field dependence predicted for a single QT process ( Figures S8 and S9) [41,42]. The extracted widths of the relaxation time (0.31 < α < 0.62, Figure S10) provide further evidence that the QT regime persists up to 2.8 K since this process is particularly sensitive to strains and disorder [43].

Chemicals
All chemicals and solvents were purchased from commercial sources and used without further purification. All the syntheses were performed in open air. [Ln(hfac) 3 ·x(H 2 O) 2 ] (Ln = La, Pr, Gd) (x = 2, 3) were synthesized starting from the respective lanthanide oxides by modified literature methods [44]. [M(acac) 3 ] (M = Cr, Fe) were synthesized from transition metal salts according to the known procedures [45,46] while an adjusted procedure was used for M = Ga based on the synthesis  3 ]. The bulk samples used for the magnetic characterizations were verified by powder X-ray diffraction to be phase pure with diffractograms matching the theoretical ones for the P2 1 2 1 2 1 structures (Supplementary Materials, Figures S11-S18).

Synthesis
Synthesis of [La(hfac) 3 Cr(acac) 3 ] (LaCr). La(hfac) 3 ·3H 2 O (203.5 mg, 0.25 mmol) was boiled close to dryness while stirring in 20 mL of toluene in an Erlenmeyer flask. A new portion of 10 mL of toluene was added, and the solution was again boiled close to dryness. This was repeated two times yielding a pale yellow-brown solution. After the last portion of toluene had been boiled almost dry, the heat was turned off and a solution of [Cr(acac) 3 ] (69.9 mg, 0.20 mmol) in 5 mL of dichloromethane was slowly added to the hot (ca. 60 • C) reaction mixture causing an immediate boil off of dichloromethane. The resulting purple solution was allowed to cool down to room temperature (RT) without stirring, whereupon 10 mL of petroleum spirit was added, and the solution was filtered through a glass frit to remove any unreacted compounds. The solution was transferred to a crystallization bowl with a lid and stored in a refrigerator overnight to yield dark purple crystals of LaCr. To enhance the quality of the crystals, the crude product was recrystallized by boiling in 20 mL of n-heptane, filtering the solution, and then storing it in a refrigerator overnight.

Single Crystal X-ray Diffraction Measurements
Single-crystal X-ray diffraction data were collected on a D8 VENTURE diffractometer (Bruker AXS, Karlsruhe, Germany) equipped with Mo Kα high brilliance IµS radiation (λ = 0.71073 Å), a multilayer X-ray mirror, a PHOTON 100 CMOS detector (Bruker AXS, Karlsruhe, Germany), and a low-temperature controller (Oxford Cryosystems, Oxford, UK). Crystals were mounted on loops made of Kapton. The diffractometer was controlled using the SAINT program as implemented in the APEX2 software package. Intensity data were corrected for absorption using the multi-scan method implemented in SADABS. The structures were solved using Olex2 by means of the olex2.structure solution program using a quasi-E charge flipping algorithm refined with the olex2.refine refinement package using Gauss-Newton minimization. Racemic twin refinements were carried out on all structures with a Flack parameter deviating more than 3 esds. from zero; in all cases, the crystals were determined not to be twinned with final Flack parameters equal to zero. For the structures with disordered hfac ligands, Shelx ISOR and DFIX restraints were applied to ensure chemically sensible metrics of the hfac ligands.

Powder X-ray Diffraction Measurements
Powder X-ray diffraction data were collected at room temperature on a D8 ADVANCE powder diffractometer (Bruker AXS, Karlsruhe, Germany) operating in a 2θ-θ configuration using Cu Kα radiation (λ = 1.5418 Å).

Magnetic Measurements
EPR spectra of GdGa were recorded on a microcrystalline powder sample using a Bruker Elexsys E500 equipped with a Bruker ER 4116 DM dual-mode cavity, an EIP 538B frequency counter and an ER035M NMR Gauss meter (Bruker Biospin, Rheinstetten, Germany). Effects related to crystallites' preferential orientation were dimmed mediating 90 spectra recorded at different orientations. Ac and dc magnetometry measurements were performed using a Quantum Design MPMS-XL7 SQUID magnetometer (Quantum Design, Darmstadt, Germany). The samples were measured as pressed pellets wrapped into teflon. The data were corrected for diamagnetism using Pascal's constants [47].

Simulations
All the simulations of the dc magnetic data were obtained using the Easy Spin package [48] of MATLAB. The ac susceptibility was fitted using a home-written program in MATLAB.

Conclusions and Outlook
A systematic building-block approach towards chiral and nearly axial heterobinuclear 3d-4f complexes has been forwarded. By employing sterically undemanding, threefold symmetric transition metal complexes it was found to be possible to extend the chemistry based on the Ln(hfac) 3 fragment to encompass triply bridged systems with nine-coordinate lanthanide centers. The resulting systems provide for relatively short 3d-4f distances (3.320-3.514 Å) and concomitant easily quantifiable magnetic interactions. The class of systems introduced here are amenable to extension towards other transition metal building blocks. Such investigations are currently being undertaken, and it has already been verified that other motifs based on, e.g., Ru(III), Re(V) and Co(III) complexes are applicable. It is noteworthy that the crystallization as conglomerates and ensuing spontaneous resolution provides a rare avenue towards chiral Ln(hfac) 3 fragments with potential for complete resolution based on adducts with robust chiral transition metal complexes.