Tuning Magnetic Anisotropy Through Ligand Substitution in Five

Article pubs.acs.org/IC

Tuning Magnetic Anisotropy Through Ligand Substitution in FiveCoordinate Co(II) Complexes David Schweinfurth,† J. Krzystek,‡ Mihail Atanasov,*,§,& Johannes Klein,† Stephan Hohloch,† Joshua Telser,*,# Serhiy Demeshko,∥ Franc Meyer,∥ Frank Neese,*,§ and Biprajit Sarkar*,† †

Institut für Chemie und Biochemie, Freie Universität Berlin, Fabeckstraße 34-36, D-14195 Berlin, Germany National High Magnetic Field Laboratory, Florida State University, Tallahassee, Florida 32310, United States § Max Planck Institute for Chemical Energy Conversion, Stiftstraße 34-36, D-45470 Mülheim an der Ruhr, Germany # Department of Biological, Chemical, and Physical Sciences, Roosevelt University, Chicago, Illinois 60605, United States & Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria ∥ Universität Göttingen, Institut für Anorganische Chemie, Tammanstraße 4, D-37077 Göttingen, Germany ‡

S Supporting Information *

ABSTRACT: Understanding the origin of magnetic anisotropy and having the ability to tune it are essential needs of the rapidly developing field of molecular magnetism. Such attempts at determining the origin of magnetic anisotropy and its tuning are still relatively infrequent. One candidate for such attempts are mononuclear Co(II) complexes, some of which have recently been shown to possess slow relaxation of their magnetization. In this contribution we present four different five-coordinated Co(II) complexes, 1−4, that contain two different “click” derived tetradentate tripodal ligands and either Cl− or NCS− as an additional, axial ligand. The geometric structures of all four complexes are very similar. Despite this, major differences are observed in their electronic structures and hence in their magnetic properties as well. A combination of temperature dependent susceptibility measurements and high-frequency and -field EPR (HFEPR) spectroscopy was used to accurately determine the magnetic properties of these complexes, expressed through the spin Hamiltonian parameters: g-values and zero-field splitting (ZFS) parameters D and E. A combination of optical d-d absorption spectra together with ligand field theory was used to determine the B and Dq values of the complexes. Additionally, state of the art quantum chemical calculations were applied to obtain bonding parameters and to determine the origin of magnetic anisotropy in 1−4. This combined approach showed that the D values in these complexes are in the range from −9 to +9 cm−1. Correlations have been drawn between the bonding nature of the ligands and the magnitude and sign of D. These results will thus have consequences for generating novel Co(II) complexes with tunable magnetic anisotropy and hence contribute to the field of molecular magnetism.

■

INTRODUCTION The Cu(I) catalyzed “click” reaction, which selectively produces only one regioisomer of the resulting 1,2,3-triazoles,1,2 has become an important route to novel ligands in coordination and organometallic chemistry.3−6 Taking advantage of this modular synthetic route, a battery of novel ligands, “tuned to the occasion”, have been synthesized. Their resulting metal complexes, such as of Ru(II),7−9 have been probed with respect to their function in photo and redox chemistry and in homogeneous catalysis.10−15 Among our interests is the development of metal complexes with tunable electronic structures, with possible applications as single ion magnets.16,17 In particular, we have recently discovered that click-derived tripodal ligands can deliver special features to their metal complexes with respect to magnetic properties. The magnetic anisotropy of Ni(II)18 and Co(II)17 complexes with such ligands has been investigated. More significantly, Fe(II) and © 2017 American Chemical Society

Co(II) complexes specifically with [(1-benzyl-1H-1,2,3-triazol4-yl)methyl]-amine (tbta) ligands display temperature induced spin crossover behavior.19,20 This interesting property was dependent on the peripheral substituents of tbta.21 Furthermore, tbta is a useful coligand for investigating magnetic exchange in di-Co(II) complexes.22 From among the many paramagnetic 3d ions, Co(II) has played a significant role in recent years with respect to molecular magnetism. Specifically, mononuclear four- and fivecoordinate Co(II) complexes, often with trigonal symmetry,23−25 have emerged as a class of molecules that display slow relaxation of their magnetization.16,23−29 Most of these, and other Co(II) complexes,30−32 were characterized through static and/or dynamic magnetic measurements. Additionally, in some Received: February 9, 2017 Published: April 12, 2017 5253

DOI: 10.1021/acs.inorgchem.7b00371 Inorg. Chem. 2017, 56, 5253−5265

Article

Inorganic Chemistry

calculated and experimental C, H, and N values (see Experimental Section). Single crystals suitable for X-ray diffraction studies were obtained for complexes 2−4 (see Experimental Section; crystallographic details are given in Table S1). As was observed for 1,15 the Co(II) center in 2−4 is in a distorted trigonal bipyramidal environment. The N-donors from the triazole rings of the tripodal ligands take up the equatorial coordination sites. Additionally, the amine-N donor of the tripodal ligands and Cl− (for 3) and NCS− (for 2 and 4) occupy the axial coordination sites (Figure 1; larger views of the cations of 2 and 4 are shown in Figures S1 and S2, respectively). The Co−N bond distances in all complexes are in a similar range with the Co−N(triazole) distances being distinctly shorter compared to the Co− N(amine) distance (Table 1). In 2 and 4, the thiocyanato ligand is bound to the Co(II) centers through the N-donor. The Co-ligand bond distances in these complexes clearly point to the existence of a high spin (HS) Co(II) center at the measurement temperatures. The C−N and C−S distances within the thiocyanato ligands are within the range of values usually observed in metal coordinated thiocyanate complexes (Table 1). For comparison, the metrical data for the previously reported azide complex, [Co(tbta)(N3)]+, 5, is also included.17 The thiocyanato ligand preserves its linear geometry upon metal coordination (Tables 1 and S2), as did the azido ligand in 5. The Co−N2−C1 angle at the coordination interface between cobalt and NCS− is 171.3(3) and 166.2(3)° for 2 and 4, respectively. The Co−N12−N10 bond in 5 is much further from linear (145°), which is typical for azido complexes.35 Even though the geometry around the Co(II) center has been described above as trigonal bipyramidal, it should be borne in mind that distortion from this idealized geometry is rather large in the present cases. Thus, the three triazole donors are “pushed down” in the direction of the central amine donor of the tripodal ligands (Figure 1), leading to N−Co−N angles close to 75° (Table S1). Accordingly, the Cl−Co−N angles open up and are in the range of about 105°. In view of the rather long Co−N(amine) bond distance and the distortion as mentioned above, these complexes tend toward a 4 + 1 pseudo tetrahedral kind of coordination. Overall, the geometry around the first coordination sphere of Co(II) is similar in all of these complexes. The coordination geometry of this series of [Co(txta)X]+ (x = b (benzyl), p (phenyl)) complexes can be further probed using a continuous shape analysis, as pioneered by Alemany, Alvarez, Avnir, and co-workers.36−38 In this case, we compare the complexes to two ideal five-coordinate geometries, trigonal bipyramidal (tbp, D3h symmetry) and square pyramidal (spy, C4v symmetry). The results are summarized in Table S3 and show qualitatively that all five complexes are far from either ideal geometry, although much closer to tbp than to spy. Moreover, 1 and 3 (i.e., the two chlorido complexes) are very similar and further from ideal geometry than the other three, despite having a monatomic axial ligand. The complexes with pseudohalido ligands are similar to 2 being the closest to ideal tbp, then 5, and last 4, with 4 being roughly equidistant in the shape continuum from the other two. Magnetic Measurements. The temperature dependence of the magnetic susceptibility was probed for complexes 2−4. All complexes display a pattern (shown in Figure 2 (left) for 3) that is similar to that previously reported for 1.15 The room temperature χMT values for all of the complexes are in the region of 2.2−2.3 cm3 mol−1K which fits well with the existence

cases high-frequency and -field (HFEPR) spectroscopy was also used to determine spin Hamiltonian parameters including zerofield splitting (ZFS).23,28,29,31 Scorpionate complexes of Co(II), TpR,R′CoL (Tp = hydrido(trispyrazolyl)borate), which as well exhibit roughly trigonal symmetry, were also investigated by HFEPR, but not for their magnetic behavior.33 However, an extensive theoretical study of the magnetic properties of a range of 3d ion scorpionate complexes, [TpMCl]0,+, including M = Co(II), has recently appeared.34 We have reported on the complex [Co(tbta)Cl]BF4 (1) and gave preliminary results on its magnetic properties. 15 Considering the utility of “click”-derived tripodal ligands in generating tunable metal complexes, as mentioned above, we have now turned our attention to a series of Co(II) complexes with such ligands. Herein, we report the synthesis of three new Co(II) complexes: [Co(tbta)NCS]BF4 (2), [Co(tpta)Cl]BF4 (3, tpta = [(1-phenyl-1H-1,2,3-triazol-4-yl)methyl]-amine), and [Co(tpta)NCS]BF4 (4) (Scheme 1). In the series 1−4, we have Scheme 1. Neutral Ligands tbta (Left) and tpta (Right)

varied the substituents on the tripodal ligands (benzyl vs phenyl in tbta vs tpta; these ligands will be referred to generically as txta) and the fifth, the axial ligand (Cl− vs NCS−). Data from single crystal X-ray diffraction measurements are presented to discuss the effects of ligand substitution on the geometric structures of the Co(II) complexes. In addition, results from temperature dependent magnetic measurements, HFEPR spectroscopy, ligand field theory and ab initio calculations are discussed to decipher the effects of ligand substitution on the electronic structures of the complexes, in particular, on the magnetic anisotropy and the ligand field parameters B and Dq. Also relevant in this comparison are results from our previous study on [Co(tbta)N3]ClO4·3CH3CN (5),17 which emphasized understanding the ligand-field of azide ion using data also for TpCoN3. In doing so, we provide here correlations between ligand bonding parameters, the ligand field of the ligands and magnetic anisotropy in these Co(II) complexes.

■

RESULTS AND DISCUSSION Synthesis and X-ray Structures. Complex 3 was synthesized by a route analogous to that of 1, which had been reported earlier.15 Reaction of tpta with a mixture of Co(BF4)2·6H2O and CoCl2·6H2O in acetonitrile led to the formation of 3 (see Experimental Section). For 2 and 4, an analogous procedure with additional NBu 4 NCS as a thiocyanate source followed by recrystallization from acetonitrile/diethyl ether led to the formation of pure complexes. The purity of these purple materials was ascertained by elemental analyses that showed excellent agreement between the 5254


Article

Inorganic Chemistry

temperature dependence of the magnetic data were fitted using a standard spin Hamiltonian: ⎛ 2 1 ⎞ 2 2 / = βeB·g ·S ̂ + D⎜Sẑ − S(S + 1)⎟ + E(Sx̂ − Sŷ ) ⎝ ⎠ 3 (1)

These fits employed an isotropic g value and only axial ZFS, yielding excellent fits for 2−4. The fit parameters are summarized in Table 2; those for 3 are also given in the caption to Figure 2 and those for 2 and 4 are given in their respective figures (Figures S3 and S4). As has been well documented,29 the magnitude, and more importantly the sign of D cannot be reliably obtained from such measurements alone. Indeed, although the fits of magnetic data for 2−4 appear definitive in determining the magnitude of D, no indication of its sign was possible, so that only the absolute value of D is given. For complex 3, however, variable temperature-variable field (VTVH) magnetization measurements were also made (Figure 2, right) and a global fit with a negative D value was more successful. We, therefore, turned to high-frequency and -field EPR (HFEPR) spectroscopy to determine the spin Hamiltonian parameters of all of the complexes of interest more accurately. High-Frequency and -Field EPR (HFEPR). Cobalt(II) complexes possess a 3d7 configuration and in the high spin form the three unpaired electrons give rise to a quartet ground state with S = 3/2. The ZFS removes the degeneracy of the quartet state and splits it into two so-called Kramers doublets with MS = ±1/2 and ±3/2. In principle, EPR transitions are possible inside each Kramers doublet and between different Kramers doublets, but only transitions between the two different Kramers doublets yield information about the ZFS, because they are related to the energy difference between the two Kramers sublevels at zero field.39,40 HFEPR experiments were conducted on powder samples of complexes 1−4. Even though the quality of the spectra varied from one compound to another, inter-Kramers transitions were clearly observed in all cases, which allowed for the direct extraction of the ZFS parameters from these experiments. Figure 3 shows selected spectra of 1−4, which were recorded on freshly ground powders (some field-induced torquing effects were observed in certain complexes but overall good powder patterns were observed in the spectra). Together with the experimental spectra, powder-pattern simulations are shown, which were generated using an S = 3/2 spin Hamiltonian (the same as that in eq 1) with both positive and negative signs of D. The sign of E was fixed to that of D and the D and g tensors were held to be collinear so that gx, gy, and gz values result, but these designations do not necessarily correspond to a known molecular direction. The key point is that, in contrast to simulations of the magnetic susceptibility data, the difference in spectral appearance for the HFEPR simulations with opposite signs of D is quite dramatic. A reliable extraction of spin Hamiltonian parameters from powder samples, however, relies on the analysis of field/ frequency 2-D maps, which include all measured resonances of each complex, rather than on the simulation of a single frequency spectrum. This method allows for determination of frequency-independent ZFS parameters even in the presence of well-known problems such as microcrystallite torquing in high magnetic fields that can distort the ideal powder patterns.39 In Figure 4, such 2-D maps for the complexes 1−4 are shown. In each case, the zero-field energy gap between the MS = ±1/2

Figure 1. ORTEP views of 2 (top), 3 (middle), and 4 (bottom). Ellipsoids are drawn at 50% probability. Hydrogen atoms (except for the triazole C−H atoms) and counteranions have been omitted for clarity.

of a HS-Co(II) center (2.25 cm3 mol−1 K or 4.24 μB for S = 3/2 with g = 2.19). On lowering the temperature, the χMT values drop to 1.4−1.5 cm3 mol−1 K, which because of the large separation between nearest neighbor Co(II) centers is an indication of the operation of ZFS in these complexes. The 5255


Article

Inorganic Chemistry

Table 1. Selected Bond Lengths (in Å) and Angles (in deg) in the [Co(txta)X]+ Complexes (x = b (benzyl), p (phenyl); X− = Cl− NCS−, N3−) bond lengths complex; x, X−

1; b, Cl−

2; b, NCS−

3; p, Cl−

4; p, NCS−

5; b, N3−

Co1−N1 Co1−N11 Co1−N21 Co1−N31 Co1−Cl Co1−N2 C1−N2 C1−S1

2.350(3) 2.037(3) 2.026(3) 2.035(3) 2.2638(8)

2.346(3) 2.028(3) 2.028(3) 2.026(3)

2.388(3) 2.053(3) 2.039(3) 2.036(3) 2.260(1)

2.371(3) 2.017(3) 2.024(3) 2.032(3)

2.369(3) 2.032(3) 2.032(3) 2.042(3)

1.980(3) 1.154(4) 1.625(3)

1.964(3)

1.953(3) 1.153(5) 1.620(5) bond angles

complex; x, X−

1; b, Cl−

2; b, NCS−

3; p, Cl−

4; p, NCS−

5; b, N3−

N11−Co1−N21 N21−Co1−N31 N31−Co1−N11 N1−Co1−Cl1 N1−Co1−N2 Co1−N2−C1 Co1−N2−N3 N2−C1−S1 N2−N3−N4

113.2(1) 114.4(1) 113.1(1) 179.0(1)

114.5(1) 110.6(1) 115.7(1)

113.1(1) 116.8(1) 109.1(1) 177.8(1)

110.0(1) 119.4(1) 109.2(1)

113.5(2) 112.5(2)

177.9(1) 171.2(3)

175.2(1) 166.2(3)

173.8(2)

179.5(4)

178.3(3)

141.1(3) 176.2(4)

Figure 2. Left: Plot of χMT versus T for 3 at an external field of 0.5 T. Right: VTVH magnetization measurements given as Mmol versus μBB/kT for 3. Experimental data are given as circles. Solid lines represent the best global fit with the parameters S = 3/2, giso = 2.23, and D = −6.2 cm−1.

Table 2. Spin Hamiltonian Parameters of the Investigated Complexes Derived from Magnetometry and HFEPR Measurements complex

Da (cm−1)

|E|a (cm−1)

|E/D|a

1 HFEPR magnet. 2 HFEPR magnet. 3 HFEPR magnet. 4 HFEPR magnet.

−5.20(5) |4.6(4)|e −4.41(3) |7.0(4)| −9.02(4) −6.2(3) +8.29(1) |7.2(4)|

0.75(6)

0.14

0.13(5)

gyb

gzb

gavgc

Δgd

2.30(4)

2.15(4)

2.41(8)

0.26

0.03

2.3(4)

2.16(2)

2.18(3)

1.64(2)

0.18

2.324(14)

2.42(2)

2.22(2)

0.45(1)

0.05

2.25(1)

2.24(1)

2.138(2)

2.29 2.17(1) 2.21 2.19(1) 2.32 2.23(1) 2.21 2.17(1)

gxb

0.13 0.20 0.11

a

Magnetometry provided only D values and the sign of D was determined only for 3. The sign of D was directly determined by HFEPR; that of E was not but was assumed to be that of D in simulations. bThe g values are given here as gx, gy, gz, as defined in the EPR simulation software: z being defined as the ZFS (D tensor) principal axis, which does not necessarily correspond to a specific molecular reference frame. cMagnetometry provided only isotropic g values, which are listed here as gavg. dΔg is defined as gmax − gmin. eMagnetic susceptibility data for 1 were published previously.15

and MS = ±3/2 Kramers doublets can be clearly recognized. The particulars of HFEPR spectra of the individual complexes are briefly discussed below. In the spectra of 1, very broad resonances were observed. Nevertheless, inter-Kramers transitions were clearly observed and the magnitude of the ZFS can be reliably determined. The

sign of D can be determined by comparison of single-frequency spectra with simulations for both positive and negative D. The main criteria are not the positions of the resonances, which depend only on the magnitudes of the ZFS parameters, but on their relative intensities, as seen in Figure 3. In the case of 1, the experimental spectrum is best reproduced by using a negative D 5256


Article

Inorganic Chemistry

defined resonances were observed and the good agreement between experimental spectrum and simulation clearly points to a negative D value (compare the simulated and experimental transition intensities at ∼13 T; the group of resonances in the 9−11 T region originates from impurities). This is in contrast to 4, for which a positive D value was confidently established. This can be seen by comparison of the relative intensities of the resonances at 1−2 T to those at 7−8 T. For negative D, the former are stronger and the latter are weaker, while for positive D, and in this case in the experimental spectrum, the former are weaker and the latter are stronger. Table 2 summarizes the spin Hamiltonian parameters, which were obtained from the analysis of the multifrequency 2-D maps. The spectra of complexes 1 and 2 can be fitted with D values of −5.20(5) and −4.41(4) cm−1, respectively, which are similar in magnitude. Even though the spectra of 3 were fitted with a negative D value, in contrast to positive D for 4, the D values of +9.02(4) and −8.29(1) cm−1 are of similar magnitude. Hence, the |D| values of complexes 3 and 4 with the ligand tpta are double the magnitude of the |D| values of complexes 1 and 2 with the ligand tbta. In addition to the axial ZFS, D, we can consider the rhombic ZFS, given by the parameter E. Note that this information can be readily determined from the HFEPR data, but is generally unobtainable from magnetometry. Even though similar |D| values were observed for 1 and 2 (i.e., tbta: Cl− vs NCS−), their respective |E| values, 0.75(6) and 0.13(5) cm−1, are quite different. The same is true for 3 and 4 (i.e., tpta: Cl− vs NCS−), which exhibit |E| values of 1.64(2) and 0.45(1) cm−1, respectively. Complexes 1 and 3, which incorporate a chlorido ligand, possess |E| values that are significantly higher than for 2 and 4, which feature a thiocyanato ligand. However, the rhombicity, given by the ratio |E/D|, is very similar for 1 and 3 with values of 0.14 and 0.18. The same is true for complexes 2 and 4 with similar |E/D| values of 0.03 and 0.05, respectively, which are much smaller than for 1 and 3. This grouping is seen as well for the g-anisotropy, Δg, with 1 and 3 being similar and having higher Δg values compared to those for 2 and 4, which are themselves similar. The rhombicity can be related to the deviation from ideal tbp geometry as given by the continuous shape analysis described above. This is shown graphically in Figure S5, which shows a smooth progression of increasing |E/D| with increasing distortion. It follows that although the geometric structures of the complexes 1−4 are very similar, their magnetic properties, and thus also their electronic structures, as revealed by HFEPR, differ significantly. On the one hand, the influence of substituents on the tripodal ligands is reflected by the magnitude of the |D| and |E| values, which are double the amount for complexes with the ligand tpta (3 and 4) in comparison to complexes with tbta (1 and 2). On the other hand, the influence of the ancillary ligand is reflected by the anisotropy, reflected in the parameters |E/D| and Δg, which is much higher in the case of the chlorido complexes (1 and 3) in comparison to the thiocyanato complexes (2 and 4). These results are totally in contrast to the results of an extensive study on structurally related tetrahedral scorpionate complexes TpR,R′Co(X) (X = Cl−, NCS−), which was reported by some of us earlier.33 In the first case, all investigated complexes TpR,R′Co(X) possessed positive D values. In the case of the scorpionate complexes, the magnitude of the D values was independent of the substitution of the scorpionate ligand,

Figure 3. Experimental HFEPR spectra (black traces) of complexes 1−4 and their powder-pattern simulations for positive (red traces) and negative (blue traces) D values. The spectrum of 1 was recorded at 217.6 GHz and 10 K; those of 2 and 4 were recorded at 406 GHz and 10 K, and that of 3 was recorded at 324 GHz and 5 K. Simulations used the best-fit values from Table 2 that were each only minimally adjusted to match the specific spectrum presented: Spectrum of 1 was simulated with |D| = 5.21 cm−1, |E| = 0.75 cm−1, g = [2.30, 2.15, 2.41], 2 with |D| = 4.21 cm−1, |E| = 0.12 cm−1, g = [2.16, 2.16, 2.18], 3 with |D| = 9.02 cm−1, |E| = 1.64 cm−1, g = [2.32, 2.42, 2.22], and spectrum of 4 with |D| = 8.3 cm−1, |E| = 0.36 cm−1, g = [2.19, 2.25, 2.13].

Figure 4. 2-D maps of the turning points in the HFEPR spectra of the investigated complexes. Squares are experimental data at 5 or 10 K; lines were drawn using the least-squares best-fitted spin Hamiltonian parameters as in Table 2. Red lines represent B0∥x turning points (i.e., external field aligned with the ZFS tensor x direction); blue lines, B0∥y; and black lines B0∥z resonances.

value, although the agreement is not ideal due to torquing effects (an experiment on a pressed pellet failed). This is best seen by comparison of the highest field resonances. For positive D, these simulated features are relatively weak, while for negative D, and in the experimental spectrum, they are relatively strong. Complex 2 produced only weak resonances, but the magnitude of the ZFS and a negative D value can analogously be unequivocally established. In the case of 3, well5257


Article

Inorganic Chemistry but relied only on the nature of the ancillary ligand. The complexes TpR,R′Co(NCS) with differently substituted scorpionate ligands exhibited D values in the close range from +2.39(3) to +3.34(1) cm−1, whereas the complexes TpR,R′CoCl had much higher D values in the close range from +10.88(1) to +12.72(4) cm−1. Furthermore, the rhombicity factor |E/D| was higher for the thiocyanato complexes TpR,R′Co(NCS) with values in the range 0.056−0.15, compared to the chlorido complexes TpR,R′CoCl with values in the range 0.012−0.061. All those observations are in contrast to the results on the complexes 1−4 presented in this study. Electronic Absorption Spectroscopy and Classical Ligand Field Theory (LFT). The vis−NIR spectra in acetonitrile solution (the same solvent generally used for crystallizations) of the four Co(II) complexes of interest are shown in Figure 5, with the full UV−vis−NIR spectra shown in

Although the actual symmetry is lower (see structures section), idealized tbp geometry (D3h point group symmetry) will be used at the outset to describe the spectra, particularly of the thiocyanate complexes, as these have N5 donor sets. The free-ion ground term for Co(II) is 4F and in D3h symmetry this splits into 4A2′, 4A1″, 4A2″, 4E′, and 4E″ states. The excited state quartet term, 4P, splits into 4A2′ and 4E″ states, which mix with the corresponding 4F states. A qualitative diagram of this splitting, based on that given by Lever,42 is shown in Figure S7. There are also a variety of spin doublet excited states, which will be included in the full d7 basis treatment, but we will disregard these states for the purposes of assigning the electronic transitions, as these will all be assumed to be spinallowed transitions. The ground state in tbp geometry is 4A2′, with the degenerate 4A1″ and 4A2″ states lying slightly higher in energy. Next are the orbital doublet states arising from 4F, 4E″, and then 4E′, followed by the 4A2′ and 4E″ states from 4P. In true D3h symmetry, the allowed transitions are 4A2′ → 4A1″ (with z polarization) and 4A2′ → 4E′ allowed (with x,y polarization). However, in C3v symmetry, the transition 4A2 → 4 A2 (4A2′ in D3h) also becomes allowed (with z polarization) as are 4A2 → 4E (4E′ or 4E″ in D3h), and in C3 symmetry, all transitions are allowed. We assign the highest energy visible band (at 19 000−21 000 cm−1 in the series) to 4A2′ → 4E″(P) (in D3h), as was done by Larrabee and co-workers for [Co(Me6tren)Cl]+.44 The other assignments are not as clear-cut, but if we follow Larrabee, then the band at 16 000−17 000 cm−1 in the series is assigned to 4 A2′ → 4A2′(P).49,33 The next band shows the greatest difference between the chlorido and thiocyanato complexes, appearing respectively at 11 000 cm−1 and 13 000 cm−1. This band is assigned to 4A2′ → 4E′. In single electron d orbital terms, this transition corresponds to an electron promotion to the dz2 orbital: (dxz,yz)4 (dx2−y2,dxy)2 (dz2)1 →(dxz,yz)3 (dx2−y2,dxy)2 (dz2)2; alternatively, using D3h labels, as (e″, e′)6(a1′)1 → (e″)3(e′)2(a1′)2. Thus, it may be more affected by the axial ligand than the other transitions, which do not involve dz2 to as great an extent. Thiocyanato is a stronger field ligand than chlorido, so the dz2 orbital is higher energy in the [Co(txta) (NCS)]+ complexes than in [Co(txta)Cl]+. Lastly, we come to the NIR bands, which appear to have several components. A shoulder at ∼7500 cm−1 may be assigned to 4A2′ → 4E″(F); the main band at 6000−7000 cm−1 and a possible shoulder at ∼5000 cm−1 to 4A2′ → 4A1″,4A2″. With these tentative assignments, we can use the angular overlap model (AOM), as was done by Larrabee and coworkers.44 We initially use an idealized tbp geometry, with θ = 90° and ϕ = 0, 120, 240°, as depicted in Figure S8. There are, however, three types of ligand: the axial amine nitrogen atom, the three equatorial nitrogen atom donors from the triazole arms, which are held to be equivalent, and the ancillary axial ligand, Cl− or N of NCS−. The bonding parameters determined44 for [Co(Me6tren)Cl]+ are a reasonable starting point for a fitting procedure. The axial amine ligand is a σdonor only; the axial (pseudo)halido ligand is a cylindrical πdonor, and for simplicity, here we will disregard π-bonding by the (equivalent) equatorial triazole nitrogen ligands. The difference between the vis-NIR spectrum of the tbta and tpta complex are well within any error of the fitting procedure (and indeed of the band positions), so for both NCS− and Cl− complexes, we fit only the tpta version. The QCT section deals with each tripodal ligand explicitly. The fitting procedure using classical LFT is described in detail in Supporting Information.

Figure 5. Electronic absorption spectra of complexes 1 (red trace), 2 (blue trace), 3 (magenta trace), and 4 (green trace) recorded in acetonitrile solution at room temperature.

Figure S6 and the data tabulated in Table S2. There are two stronger (ε > 200 mol−1 L cm−1) bands in the Vis region, each of which may have at least two components, and two weaker (ε < 100 mol−1 L cm−1) bands in the NIR region. The two complexes with the thiocyanato ancillary ligand exhibit a strong band (ε > 1000 mol−1 L cm−1) beginning at 28 000 cm−1 that is absent in the chlorido complexes. We assign this to a LMCT band involving the NCS − π* MOs (i.e., formally [CoII(NCS −)]+ → [Co I(NCS• )]+ ). All of the studied complexes exhibit very strong bands above 30 000 cm−1 that we attribute to π−π* absorption in the ligands, as expected for the triazole moieties. These bands are red-shifted in the tpta complexes (more easily compared in the chlorido complexes) as expected for the conjugated phenyl ring of this ligand, as opposed to the isolated benzyl group in tbta. The assignment of metal-centered UV−vis−NIR bands of pentacoordinate Co(II) complexes with idealized trigonal bipyramidal geometry has been extensively studied and is best summarized by Lever,42 and also by Banci et al., as part of their broader discussion of Co(II) complexes,43 and by Ciampolini as part of a survey of five-coordinate complexes.41 More recently, Larrabee and co-workers have studied UV−vis− NIR and MCD spectra of 5-coordinate Co(II) in both monoand dinuclear complexes.44−48 5258


Article

Inorganic Chemistry

Table 3. Assignments and Calculated Electronic Absorption Bands (in cm−1) for [Co(tpta)X]+ (X− = Cl−, NCS−) Complexesa assignment in D3h (C3v) symmetryb A2′ ( A2(1)) → A2″ (4A1)

4

complex

4

4

A2′ → E″[ F] (4E(1))

4

4

A2′ → 4A1″ (4A2(2))

4

4

A2′ → 4E′[4F] (4E(2))

4

4

A2′ → 4A2′[4P] (4A2(3))

4

A2′ → 4E″[4P] (4E(3))

experimental LFT calculated, trigonal geometryd LFT calculated, real geometrye

5300 (sh)c 4460

6230 5950

3, [Co(tpta)Cl]+ 7800 (sh) 7940

11 000 11 070

15 900, 16 340 16 100

18 500 (sh), 19 700 19 690

4380

5870, 6050

7940

10 650, 11 490

16 090

19 520, 19 830

experimental LFT calculated, trigonal geometryf LFT calculated, real geometryg

5400 (sh)c 4660

12 800 11 130

15 750 (sh), 16 560 16 320

19 330, 20 300 20 290

10 490, 11 790

16 320

20 060, 20 560

4430

4, [Co(tpta) (NCS)]+ 6630 7800 (sh) 6160 8540 5950, 6420

8440

LFT-calculated values are rounded to the nearest 10 cm−1. The experimental error is ca. ±200−300 cm−1 for the resolved NIR bands and ca. ±100 cm−1 for the resolved vis bands. bSee Figure 18 in Ciampolini.41 The sequential numbering of the terms in C3v symmetry is as used in the QCT section below. cThis band identification is very tentative and was not used for fitting. dThe calculation uses the average θ angle, 74.38°, and ideal trigonal ϕ angles with the following parameters: B = 710; εσ(Nax=amine) = 4555; εσ(Neq=triazole) = 4009; εσ(Cl) = 3269; επ(Cl) = 1568. These AOM bonding parameters derived from classical LFT are designated by “ε” (epsilon, the classical letter), while the AOM bonding parameters derived from ab initio LFT (AILFT; see Table 5) are designated by “e”, to avoid confusion between the two parameter sets. eThe calculation uses the real θ and ϕ angles. If the bonding parameters are allowed to vary freely, then the fit drives εσ(Cl) to zero (minimum allowed value) and εσ(Nax) to a very large value, which model has no physical meaning. Therefore, trial and error with constraints on the bonding parameters led to the following consensus fit parameters: B = 730; εσ(Nax=amine) = 5200; εσ(Neq=triazole) = 4300; εσ(Cl) = 3000; επ(Cl) = 2150. Note that in the low symmetry of the real geometry, the orbital degeneracy of the 4E states is removed. fThe calculation uses the average θ angle, 74.62°, and ideal trigonal ϕ angles with the following parameters: B = 725; εσ(Nax=amine) = 4691; εσ(Neq=triazole) = 4339; εσ(NCS) = 3480; επ(NCS) = 2170. gThe calculation uses the real θ and ϕ angles. If the bonding parameters are allowed to vary freely, then the fit drives εσ(NCS) to a very small value and εσ(Nax=amine) to its maximum allowed value, which model has no physical meaning. Therefore, trial and error with constraints on the bonding parameters led to the following consensus fit parameters: B = 730; εσ(Nax=amine) = 5200; εσ(Neq=triazole) = 4300; εσ(NCS) = 3000; επ(NCS) = 2150. Note that in the low symmetry of the real geometry, the orbital degeneracy of the 4E states is removed. a

range 11 000−13 000 and 15 700−17 000 cm−1 that depend on the nature of X (NCS− or Cl−) and the triazol-amine ligand (tbta or tpta) are assigned, respectively to 4A2(1) → 4A2(2) and 4 E(2) (arising from the 4T1(F) parent term in Td symmetry). The energies of these transitions are systematically underestimated by CASSCF/NEVPT2 calculations by an energy of about 2500 cm−1. Finally, d-d bands of highest energy in the range 18 000−20 000 cm−1 are assigned respectively to 4A2(1) → 4A2(3) and 4E(3) (arising from 4T1(P) tetrahedral parent term). These CASSCF/NEVPT2 calculations yield an energy order 4A2(3) < 4E(3) and are again underestimated; to a lesser extent in the case of 4A2(3) (2000 cm−1) but to a larger amount for 4E(3) (5000 cm−1). Ab Initio Ligand Field Theory (AILFT) Analysis of the Metal−Ligand Interactions of 1−4. Ab initio ligand field theory (AILFT) yields the 5 × 5 ligand field matrix and parameters quantifying inter electronic repulsion and spin− orbit coupling within the 3d shell of a transition metal ion. Focusing on the 5 × 5 LF matrix, diagonalization yields a 3dMO ligand field orbital diagrams for 1−4 visualized in Figure 6, where for the sake of comparison, orbital energies for the already reported 5 ([Co(tbta)N3]+; i.e., X = N3−)17 complex are also depicted. All considered complexes yield the same orbital pattern with an energy order: e(1)(dxz,dyz) < e(2)(dxy,dx2−y2) < a1(dz2). Ignoring angular distortions and defining the axial ligand X and Namine as the z-axis, and the equatorial triazole Ndonors in a trigonal arrangement within the xy plane (one ligand necessarily on the x-axis), as is shown in Figure S8, the following expressions for the energies of 3d-type MOs given by the AOM are derived:

The results are presented here in Table 3. The electronic absorption bands are reasonably well fitted and subsequent inclusion of spin−orbit coupling (SOC) reproduces the experimental axial ZFS for those complexes with D < 0 (see Table 5). Quantum Chemical Theory (QCT). Electronic Absorption Spectral Calculations by QCT. The electronic absorption spectra of 1−4 are described in detail above. In addition to classical LFT, quantum chemical theory (QCT) was used, namely, the complete active space−self-consistent field (CASSCF) method with n-electron valence state secondorder perturbation theory (NEVPT2). Using a slightly different geometrical approach than in the LFT section, we can first consider the complexes to be four-coordinate with approximate tetrahedral symmetry, rather than tbp as above. This simplified geometrical model is based on counting the three triazole N donors together with the ancillary X ligand, but with the amine N donor as an additional, perturbing ligand. Thus, in Td symmetry the 4F free-ion term splits into 4A2, 4T2, and 4 T1(F), with the 4P excited term becoming 4T1(P). Lowering the symmetry to the more realistic C3v point group gives the following terms, in order of increasing energy: 4A2(1) ground state, and 4A1, 4E(1), 4A2(2), 4E(2), 4A2(3), and 4E(3) excited states, with each 4A1,2, 4E pair arising from a 4T2,1 term. The CAS-SCF/NEVPT2 calculations on all four complexes showed that spin allowed d-d transitions are predicted to take place from the 4A2(1) ground state into each of these excited states. These results are summarized in Table S4. As in the LFT calculations, the band at 6000−7000 cm−1 in all complexes is assigned to 4A2(1) → 4E(1) (from the 4T2 parent term in Td symmetry), whose computed splitting due to low symmetry cannot be resolved. The two absorption bands appearing in the 5259


Article

Inorganic Chemistry

Table 4. AOM Bonding Parameters (in cm−1) from an AILFT Analysis of CASSCF/NEVPT2 Results on the [Co(txta)X]+ (x = b (benzyl), p (phenyl)) Complexes and Tetrahedral [CoX4]2− Model Complexes (X = N3−, NCS−, Cl−)a X

N3− [Co(tbta)X]

eXσ eXπ

NCS−

Cl−

7351 2900

5749 3178

6991 2852

6029 2984

5287 1409 0.27

3226 1103 0.34

+

6843 3254 [Co(tpta)X]+

eXσ eXπ [CoX4]2− eXσ eXπ eXπ /eXσ

Figure 6. AILFT orbital level scheme for complexes 1−4 from a mapping of CASSCF/NEVPT2 multi reference wave functions/energy eigenvalues to LFT in comparison with the ligand field orbital energies of the previously reported [Co(tbta)N3]+ complex, 5.17

4093 1127 0.28

a The AOM bonding parameters derived from classical LFT (Table 3) are designated by “ε”, while the AOM bonding parameters given here, derived from AILFT, are designated by “e”.

e(d z 2) = eσX + eσam e(dxz , yz) = eπX e(d x 2 − y2 , xy) =

9 ta 3 eσ + eπtas 8 2

ZFS and g-Tensor Parameters. The ZFS parameters D, E and the g-tensor components obtained from CASSCF/ NEVPT2 calculations of complexes 1−4 in their nonrelativistic 4 A2 ground states are listed in Table 5. Considering first the effect of the ligand field for an ideal trigonal symmetry, spin− orbit coupling induces splitting of the S = 3/2 spin state into two Kramers doublets, MS = ±3/2 and ±1/2 with the energy difference being twice the zero-field splitting parameter D; MS = ±3/2 being lowest in energy then implies a negative D (easy axis anisotropy), while the MS = ±1/2 lowest means positive D (easy plane anisotropy). In addition to the axial field, orthorhombic distortions (see structural data in Tables 1 and S1) induce a mixing of MS = ±3/2 and MS = ∓1/2 magnetic sublevels leading to significant E values (Table 5). In agreement with HFEPR results, the D values computed by CASSCF/ NEVPT2 are negative for complexes 1−3; however, it is also calculated to be negative for 4, as expected from the structural and optical spectroscopic similarity among this series of complexes, yet experimentally a positive D is found for this complex. This apparent controversy cannot be adequately resolved and remains a mystery. Nevertheless, effective Hamiltonian theory allows one to quantify the contributions to these ZFS parameters from all microstates spanned by the CoII (d7) configuration, the nine spin quartet and 40 doublet states.51 In the previous study on the azido complex 517 it was found that the largest effect on D and E arises from the lowest 4 A1 and 4E(1) excited states which leads to stabilization of the MS = ±3/2 and ±1/2 levels, respectively resulting in a negative total D and E. Smaller, yet important contributions with an opposite sign of D and E (both positive) originate from the lowest 2A1 and second excited state 2E(2). Considering the dominating effect of 4A1 and 4E(1), second order perturbation theory yields the following expressions for the spin Hamiltonian parameters:

(2)

In eq 2, eXσ and eam σ describe σ antibonding respectively with the X− and Namine donors, while eXπ reflects π-interactions between the dxz and dyz orbital set and X (there are no π electrons for N-amine so eπam is zero). The parameter eσta quantifies the Co(II)-triazole σ interaction, and in contrast to the simplified AOM in the LFT section above, here we do consider π-bonding involving the triazole ligands, as given by etaπs. This parameter, etaπs, describes the interaction between the out-of-plane π orbitals of the triazole rings with the correspondingly oriented Co(II) 3d orbitals, namely dxy,dx2−y2, as quantified in the third equation of eqs 2. Using this model, but accounting for the angular geometry of the ligand donor atoms as given by the X-ray geometries of 1− 4, values of the AOM parameters for each of the ancillary ligands NCS−, N3−, and Cl− have been derived and are listed in Table 4. These parameters specify thiocyanate and azide as > stronger σ-donors when compared with chloride, with eNCS σ eNσ 3. Metal−ligand π-interactions of these ligands are quite large, eXπ ≈ 3000 cm−1 and vary moderately across the series. Comparison of results for the same X between [Co(tbta)X]+ and [Co(tpta)X]+ show that the effect of the benzyl and phenyl substituents is small. The eXσ and eXπ values obtained by the sophisticated AILFT approach are much larger than those derived from simple LFT by roughly a factor of two in each case (note that only [Co(tpta)X]+ X = NCS−, Cl− was analyzed > eCl by LFT), but the ordering eNCS σ σ is in agreement. Finally, for comparison with a simpler, homoleptic system, we computed values of eXσ and eXπ in a tetrahedral [CoX4]2− model complex with a slightly distorted Td geometry.50 Values of eXσ and eXπ from such calculations are also listed in Table 4 and show the same trends as for the triazole-amine complexes. However, we should point out that σ(π) Co−X antibonding is considerably enhanced (suppressed) by the trans/cis effect because of the donor atoms of the triazole-amine ligand. This is of great importance for the magnetic anisotropy (see below). 5260


Article

Inorganic Chemistry

Table 5. Theoretical (CASSCF/NEVPT2) versus Experimental Spin Hamiltonian Parameters (g-Tensor Components and ZFS Parameters D and E (in cm−1)) in the S = 3/2 Ground State of 1−5 [Co(tbta)X]+ −

N3−

X = g1, g2, g3 Da Ea |E/D|a g1, g2, g3 D |E| |E/D|

(5)

c

2.165, 2.171, 2.262 −8.34 (−8.06) −0.33 (−0.26) 0.04 (0.03)

−10.7

[Co(tpta)X]+

−

−

NCS (2) 2.162, 2.171, 2.242 −6.18 (−6.48) −0.43 (−0.38) 0.07 (0.06) 2.16(2), 2.18(3), 2.3(4) −4.41(3) 0.13(5) 0.03

−

Cl (1)

NCS (4)

CASSCF/NEVPT2 2.171, 2.176, 2.244 −5.80 (−6.03) −0.23 (−0.21) 0.04 (0.03) expt.b 2.15(4), 2.30(4), 2.41(8) −5.20(5) 0.75(6) 0.14

Cl− (3)

2.158, 2.178, 2.243 −6.16 (−6.44) −0.98 (−0.86) 0.16 (0.13)

2.170, 2.184, 2.254 −6.60 (−6.58) −0.73 (−0.60) 0.11 (0.09)

2.138(2), 2.24(1), 2.25(1) +8.29(1) 0.45(1) 0.05

2.22(2), 2.324(14), 2.42(2) −9.02(4) 1.64(2) 0.18

The first values of D and E (and |E/D|) presented are those calculated directly by CASSCF/NEVPT2; the second values of D and E, in parentheses, are those calculated using eqs 5 with spin−orbit coupling parameters from AILFT for complexes 5 and 1−4 (ζeff = 515, 515, 513, 515, and 513 cm−1, respectively) and the NEVPT2 g-tensor components given directly above in the table. As described in Supporting Information, LFT gives D = −9.0 cm−1 for 3 and D = −8.2 cm−1 for 4 (E = 0 for both since an ideally trigonal AOM is employed). bFor complexes 1−4, the experimental spin Hamiltonian parameters are those determined here by HFEPR; for complex 5, the parameters are those determined by magnetometry.17 For generality, the g values determined by HFEPR are given here only as g1 ≤ g2 ≤ g3, rather than with assignments to gx, gy, gz as was the case in Table 2, which reported the g values with directions as defined by the EPR simulation software. cFor complex 5, the theoretical spin Hamiltonian parameters given here differ slightly from those originally reported.17 This is due to the current values being calculated using an improved NEVPT2 procedure that gives better agreement with experiment. a

gxx = ge −

8 ζeff 3 Δ( 4T2xg )

gyy = ge −

8 ζeff 3 Δ( 4T2yg )

gzz = ge −

8 ζeff 3 Δ( 4T2zg )

D( 4T 2g ) =

⎡ ⎤ ⎞ 4ζeff 2 ⎢ 1 ⎛ 1 1 1 ⎥ ⎜ ⎟− + 9 ⎢⎣ 2 ⎜⎝ Δ(4E x (1)) Δ(4E y(1)) ⎟⎠ Δ(4A1) ⎥⎦

E( 4T 2g ) =

⎤ 2ζeff 2 ⎡ 1 1 ⎥ ⎢ − 9 ⎢⎣ Δ(4E y(1)) Δ(4E x (1)) ⎥⎦

lower the energy of Δ( 4 A 1 ) ((e″) 4 (e′) 2 (a 1 ′) 1 → (e″)3(e′)3(a1′)1), increasing its negative contribution to D. This π-effect certainly dominates the contributions to D for 5 compared to the other congeners. The actual values of D will also be affected by the spin forbidden transitions, but in short, stronger axial σ- and π-donors will contribute to larger magnitude and negative D. Since D and E are affected by both geometrical distortions away from the tbp geometry (Table S3) and by the axial ligand donors (Table 4 and eq 5), this correlation is only partly reflected in our experimental and computational results (Table 5). In the approximations inherent in eqs 3 and 4, one can show43 that the ZFS and g-tensor parameters are not independent but are related to each other according to eqs 5:

(3)

(4)

where ζeff is the effective spin−orbit coupling parameter of CoII and Δ(4A1), Δ(4Ex(1)), Δ(4Ey(1)) are the energies of electronic transitions from the nonrelativistic 4A2 ground state into the excited 4A1 and 4E(1) (4Ex(1), 4Ey(1)) states, respectively (notation follows C3v point group symmetry). In simple terms, focusing on the 3d-orbital energy diagram (Figure 6) and ignoring the half filled spin-up 3d-shell, excitation of one electron with β-spin from the lowest e(1) orbital into the e(2) and a1 orbitals will lead to 4A1(S = 3/2, e(1)⊗e(2) → A1) and 4E(S = 3/2, e(1)⊗a1 → E), excited states, respectively. It then immediately follows (see Figure 6) that the energy ordering of excited states, 4A1 < 4E, is expected to lead to a negative D (see eq 4, wherein the negative term, that involving Δ(4A1), is dominant). On the basis of the stronger σ- and π-donor ability of NCS− compared to Cl− it is expected that the magnitude of D is larger for 2 than for 1, which can also be derived from eq 4. If we assume that the σand π- effects of the txta ligands do not vary greatly among the series of complexes 1−5, then we should expect that a stronger π-donor character of the axial ligand, as for NCS− relative to Cl−, will lead to a destabilization of e(1) (e″ in D3h), which will

D=−

|ζeff | ⎡ ⎤ 1 g z − (g y + g x )⎥ ⎢ ⎦ ⎣ 6 2

E=−

|ζeff | [(gy − gx )] 12

(5)

The correspondence between theoretical g values and ZFS parameters does agree quite well with use of eq 5 (Table 5). But this is not the case with the experimental results. Equation 5 with the experimental ZFS behavior would give gz > gx ≈ gy, as seen in the theoretical g values, but which is the case only for 1. Moreover, for realistic values of ζeff (see Table 5, note a), the magnitude of D and especially E is overstated. This may be in part because eq 5 is strictly valid if spin−orbit coupling mixing of the S = 3/2 ground state with the S = 1/2 excited states is ignored. This is generally not the case, since in contrast to the g-tensor, the ground state D value is affected by spin−orbit coupling with both quartet and doublet spin excited states.52 LFT equations of this type are also problematic for other highspin ions such as Cr(II) and Mn(III) (3d4) in that there is sizable ZFS (typically, |D| ≈ 2−4.5 cm−1) yet almost no g anisotropy.40,53 5261


Article

Inorganic Chemistry

■

the purple product precipitated. It was filtered, washed with ether, and isolated in 88% yields (266 mg). Single crystals suitable for X-ray diffraction were grown by slow diffusion of ether into an acetonitrile solution. Anal. Calcd for C31H30BCoF4N11S: C, 50.70; H, 4.12; N, 20.98. Found: C, 50.85; H, 4.38; N, 21.44. IR (solid): ṽ = 2067 cm−1 (NCS stretch). [Co(tpta)Cl](BF4), 3. CoCl2·6 H2O (48.8 mg, 0.21 mmol), Co(BF4)2·6 H2O (71.6 mg, 0.21 mmol), and tpta (200 mg, 0.41 mmol) were dissolved in CH3CN (10 mL). The solution was stirred at room temperature for 2 h. Then diethyl ether (20 mL) was added and the purple product precipitated. It was filtered, washed with ether and isolated in 68% yields (95.7 mg). Single crystals suitable for X-ray diffraction were grown by slow diffusion of ether into an acetonitrile solution. Anal. Calcd for C27H24BClCoF4N10: C, 48.42; H, 3.61; N, 20.91. Found: C, 48.47; H, 3.81; N, 20.85. [Co(tpta)NCS](BF4), 4. NBu4NCS (123 mg, 0.41 mmol), Co(BF4)2· 6 H2O (140 mg, 0.41 mmol), and tpta (200 mg, 0.41 mmol) were dissolved in CH3CN (10 mL). The solution was refluxed for 1 h. After the mixture was cooled down, diethyl ether (20 mL) was added, and the purple product precipitated. It was filtered, washed with ether and isolated in 95% yields (270 mg). Single crystals suitable for X-ray diffraction were grown by slow diffusion of ether into a methanol solution. Anal. Calcd for C28H24BCoF4N11S: C, 48.57; H, 3.49; N, 22.25. Found: C, 48.94; H, 3.94; N, 22.86. IR (solid): ṽ = 2065 cm−1 (NCS stretch). X-ray Crystallography. Suitable crystals for the X-ray analysis of all compounds were obtained as described above. The intensity data were collected at 100(2) K on a Bruker Kappa Apex2duo (graphite monochromated MoKα radiation, λ = 0.71073 Å) for 2; and at 173(2) K on a four circle Siemens P4 (graphite monochromated MoKα radiation, λ = 0.71073 Å) for 3, and at 133(2) K on a Stoe IPDS II (graphite monochromated MoKα radiation, λ = 0.71073 Å) for 4. Crystallographic and experimental details for the structures are summarized in Table S1. The structures were solved by direct methods (SHELXS-97) and refined by full-matrix least-squares procedures (based on F2, SHELXL-97).56 CCDC 804871, 911114, and 911115 contain the CIFs for this manuscript. All these data can be obtained free of charge from the Cambridge Crystallographic Data Centre via https://www.ccdc.cam.ac.uk/structures/?. Magnetic Measurements and Analysis. Temperature-dependent magnetic susceptibility measurements were carried out with a Quantum-Design MPMS-XL-5 SQUID magnetometer equipped with a 5 T magnet in the range from 2 to 295 K at a magnetic field of 0.5 T. The polycrystalline samples were contained in a gel bucket, in case of 3 covered with a few drops of low viscosity perfluoropolyether based inert oil Fomblin YL VAC 25/6 to fix the crystals, and fixed in a nonmagnetic sample holder. Each raw data file for the measured magnetic moment was corrected for the diamagnetic contribution of the gel bucket and of the inert oil according to Mdia = χg·m·H, with experimentally obtained χg values. The molar susceptibility data were corrected for the diamagnetic contribution according to χMdia(sample) = −0.5·M × 10−6 cm3·mol−1.57 Temperature-independent paramagnetism (TIP) was included according to χcalcd = χ + TIP. Before simulation, the experimental data were corrected for TIP (440 × 10−6 cm3 mol−1 for 2, 270 × 10−6 cm3 mol−1 for 3, and 370 × 10−6 cm3 mol−1 for 4). Experimental χMT versus T data for 2−4 were modeled with the program julX58 using the same spin Hamiltonian as given in eqn 1 above, but with a fitting procedure using an isotropic g tensor (gx = gy = gz) and axial zero-field splitting (E = 0). Simultaneous fitting of χMT versus T measurements with data obtained from variable temperaturevariable field (VTVH) magnetization measurements was done for 3 with the program julX_2s.59 Error bars of the parameters have been determined by single point variation of the optimized parameters and visual inspection of the results. HFEPR Data Acquisition and Analysis. HFEPR spectra were recorded using Millimeter and Submillimeter Wave Facility60 and Electron Magnetic Resonance (EMR) Facility61 at NHMFL. Tunable sub-THz wave radiation sources (backward wave oscillators) and a variety of solid-state generators were used, in conjunction with a 15/17

CONCLUSIONS In summary, we have presented here the synthesis of four roughly trigonal−bipyramidal cobalt(II) complexes with clickderived tripodal ligands. The variation among these is either due to different substituents on the backbone, or the “fifth” (pseudo)halide axial ligand on the cobalt(II) center being different. The series of complexes was studied experimentally by magnetometry, HFEPR spectroscopy, and electronic absorption spectroscopy, and analyzed theoretically by firstprinciples ab initio studies in combination with classical LFT (AILFT). In the calculations, it is crucial that core-orbitals along with relativistic corrections are included in the correlation treatment. This combined experimental and theoretical approach showed that the magnitude and the sign of the zero-field splitting parameter D does depend on both the tripodal ligand or the (pseudo)halide. The first principle calculations, supported by an AILFT and the AOM, allowed us to extract metal−ligand bonding parameters that have been correlated with zero-field splitting parameters D and E from HFEPR. The results show that Co(II)-X σ antibonding (quantified by the parameter eσ) increases in the order NCS− (7350 cm−1), N3− (6850 cm−1), Cl− (5750 cm−1), while πantibonding (quantified by the parameter eπ) is large but does not vary much along the series (X = NCS− (2900 cm−1), N3− (3250 cm−1), Cl− (3200 cm−1)). From a comparison with calculations for the much simpler, homoleptic, but still analogous CoX42− model complexes we can conclude that σ(π) Co-X antibonding is considerably enhanced (suppressed) by the trans/cis effect due to the donor atoms of the tripodal triazole ligand. In agreement with previous predictions,17 the NCS− and N3− ligands induce a larger magnetic anisotropy (larger magnitude, negative D) by virtue of their being stronger σ-donors than Cl−. The absence of slow relaxation of the magnetization in complex 554 has been attributed to off-axial (orthorhombic) distortions inducing transversal anisotropy (non-negligible E) that facilitates fast magnetization decay. The results presented here show that incorporation of rigid substituents on the ligand backbone might be beneficial for generating larger magnitude D values, as seen in the comparison between the more rigid tpta and the more flexible tbta ligand. Additionally, decreasing the coordination number by using other related di- and tripodal ligands is also likely to deliver larger D values. Finally, relatively strongly donating pseudohalides such as N3− and NCS− appear to be better axial ligands than Cl− for generating larger axial anisotropy.

■

EXPERIMENTAL SECTION

Caution! Compounds containing azides are potentially explosive. Although we have never experienced any problems during synthesis or analysis, all compounds should be synthesized only in small quantities and handled with great care! Synthesis. All reactions to synthesize the metal complexes were carried out under air. The ligands tbta, tpta, and complex 1 were synthesized according to procedures published in the literature.15,55 For growing single crystals, 10−20 mg of the complexes were dissolved in 2 mL of the polar solvent (acetonitrile for 2 and 3; methanol for 4) in a vial. This solution was layered with 2 mL of the less polar solvent (diethyl ether for all) leading to the isolation of single crystals. The same procedure (also with acetonitrile and ether) was used previously for crystallization of 115 and 5.17 [Co(tbta)NCS](BF4), 2. NBu4NCS (123 mg, 0.41 mmol), Co(BF4)2· 6 H2O (140 mg, 0.41 mmol), and tbta (218 mg, 0.41 mmol) were dissolved in CH3CN (5 mL). The solution was refluxed for 1 h. After the mixture was cooled down, diethyl ether (10 mL) was added and 5262


Article

Inorganic Chemistry T superconducting or 25 T resistive (“Keck”) magnet. Ground powders (typically, 50−100 mg) were placed in Teflon containers. Detection was provided in each case with an InSb hot-electron bolometer (QMC Ltd., Cardiff, UK). Modulation for detection purposes was provided alternatively by chopping the sub-THz wave beam (optical modulation) or by modulating the magnetic field. The relative merits of both types of modulation were discussed in a previous paper by some of us.39 A standard lock-in amplifier (Stanford Research Systems SR830) converted the modulated signal to DC voltage. To extract frequency-independent spin Hamiltonian parameters, and refine them, computer fits (software available freely from Dr. A. Ozarowski, NHMFL) were made to two-dimensional field vs frequency data sets of resonances using a spin Hamiltonian comprised of Zeeman and zero-field splitting (ZFS) terms62 as given in eq 1 above. Further details of tunable-frequency HFEPR methodology are given elsewhere.39 Single-frequency powder HFEPR spectra were simulated using the same software, which made it possible to determine the sign of D in addition to its magnitude. The sign of E was assumed to be the same as that of D, by convention. Continuous Shape Analysis. The program SHAPE, version 2.1,63 was applied to the crystallographic data for the inner coordination sphere of each complex. Ligand Field Analysis. Ligand-field calculations were made using full d7 basis set employed by the programs Ligfield (J. Bendix, U. of Copenhagen, Denmark)64 and DDN (available from J. Telser). A version of DDN, DDNFIT, allows fitting of electronic absorption bands to ligand-field parameters, which can be crystal-field,65,66 AOM,66,67 or single-electron d-orbital energies. Literature sources were used for the free-ion Co(II) Racah (B, C)68 and spin−orbit coupling (ζ)69 parameters. Quantum Chemical Theory (QCT) Computations: General Aspects. All QCT calculations were carried out using the ORCA program suite,70 version 3.0.71 Computations in the present work follow the protocol applied previously.17 An essential new aspect is that rather than using frozen orbitals in the core region, the explicit account of these orbitals has led to values of D and E that compare better with experiment. The analysis of the ab initio results follows the ab initio ligand field theory (AILFT) developed by Atanasov and coworkers, as detailed elsewhere.72−74 Quantum Chemical Theory (QCT) Computations: Details on the CASSCF/NEVPT2 Calculations of Complexes 1−5. All calculations were done using the def2-TZV basis sets of Weigend and Ahlrichs75 along with the resolution of the identity (RI) approximation and the corresponding auxiliary basis functions. In a first step, a DFT calculation was carried out to provide proper starting orbitals for the subsequent correlated CASSCF/NEVPT2 calculations. This DFT calculation yields quasi-restricted orbitals (QROs) for reading into the CASSCF/NEVPT2 calculation. The QROs are then inspected in order to identify the MOs with predominant metal 3d character. In the input file for the CASSCF/NEVPT2 calculation, the MOs are rearranged in such a way that all five predominant 3d MOs appear in the active space. Correlated nonrelativistic calculations are then carried out distributing all seven electrons of Co(II) over these five 3d orbitals, that is, CAS(7,5). Nonrelativistic calculations were done using an electron density resulting from a state averaged Hamiltonian, averaging over all 10 S = 3/2 and 40 S = 1/2 states spanned by the 3d7 electron configuration of Co(II). Sample input files documenting this process for complex 5 are included in Supporting Information.

■

■

Crystallographic information file for [Co(tbta)(SCN)]+, 2 (CIF) Crystallographic information file for [Co(tpta)Cl]+, 3 (CIF) Crystallographic information file for [Co(tpta)(SCN)]+, 4 (CIF)

AUTHOR INFORMATION

Corresponding Authors

*E-mail: *E-mail: *E-mail: *E-mail:

[email protected]. [email protected]. [email protected]. [email protected].

ORCID

Joshua Telser: 0000-0003-3307-2556 Franc Meyer: 0000-0002-8613-7862 Biprajit Sarkar: 0000-0003-4887-7277 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The Fond der Chemischen Industrie (FCI, ChemiefondsStipendium for D.S.) and the Deutsche Forschungsgemeinschaft are kindly acknowledged for financial support of this work. HFEPR studies were supported by the National High Magnetic Field Laboratory, which is funded by the NSF through a Cooperative Agreement DMR 1157490, the State of Florida and the US Department of Energy. We are grateful to Dr. Andrew Ozarowski (NHMFL) for the EPR simulation and fitting software SPIN. We thank Prof. Stefan Stoll (U. of Washington) for helpful discussions.

■

REFERENCES

(1) Tornøe, C. W.; Christensen, C.; Meldal, M. Peptidotriazoles on Solid Phase: [1,2,3]-Triazoles by Regiospecific Copper(I)-Catalyzed 1,3-Dipolar Cycloadditions of Terminal Alkynes to Azides. J. Org. Chem. 2002, 67, 3057−3064. (2) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A Stepwise Huisgen Cycloaddition Process: Copper(I)-Catalyzed Regioselective “Ligation” of Azides and Terminal Alkynes. Angew. Chem., Int. Ed. 2002, 41, 2596−2599. (3) Schweinfurth, D.; Deibel, N.; Weisser, F.; Sarkar, B. Mit Klick zu neuen Liganden. Nachr. Chem. 2011, 59, 937−941. (4) Schulze, B.; Schubert, U. S. Beyond click chemistry supramolecular interactions of 1,2,3-triazoles. Chem. Soc. Rev. 2014, 43, 2522−2571. (5) Struthers, H.; Mindt, T. L.; Schibli, R. Metal chelating systems synthesized using the copper(I) catalyzed azide-alkyne cycloaddition. Dalton Trans. 2010, 39, 675−696. (6) Schweinfurth, D.; Hettmanczyk, L.; Suntrup, L.; Sarkar, B. Metal Complexes of Click-Derived Triazoles and Mesoionic Carbenes: Electron Transfer, Photochemistry, Magnetic Bistability, and Catalysis. Z. Anorg. Allg. Chem. 2017, DOI: 10.1002/zaac.201700030. (7) Weisser, F.; Hohloch, S.; Plebst, S.; Schweinfurth, D.; Sarkar, B. Ruthenium Complexes of Tripodal Ligands with Pyridine and Triazole Arms: Subtle Tuning of Thermal, Electrochemical, and Photochemical Reactivity. Chem. - Eur. J. 2014, 20, 781−793. (8) Weisser, F.; Plebst, S.; Hohloch, S.; van der Meer, M.; Manck, S.; Führer, F.; Radtke, V.; Leichnitz, D.; Sarkar, B. Tuning Ligand Effects and Probing the Inner-Workings of Bond Activation Steps: Generation of Ruthenium Complexes with Tailor-Made Properties. Inorg. Chem. 2015, 54, 4621−4635. (9) Weisser, F.; Stevens, H.; Klein, J.; van der Meer, M.; Hohloch, S.; Sarkar, B. Tailoring RuII Pyridine/Triazole Oxygenation Catalysts and

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b00371. Structural information used for LFT, discussion of LFT fitting procedure, and additional magnetic and optical spectroscopic data (PDF) 5263


Article

Inorganic Chemistry Using Photoreactivity to Probe their Electronic Properties. Chem. Eur. J. 2015, 21, 8926−8938. (10) Chan, T. R.; Hilgraf, R.; Sharpless, K. B.; Fokin, V. V. Polytriazoles as Copper(I)-Stabilizing Ligands in Catalysis. Org. Lett. 2004, 6, 2853−2855. (11) Connell, T. U.; Schieber, C.; Silvestri, I. P.; White, J. M.; Williams, S. J.; Donnelly, P. S. Copper and Silver Complexes of Tris(triazole)amine and Tris(benzimidazole)amine Ligands: Evidence that Catalysis of an Azide−Alkyne Cycloaddition (“Click”) Reaction by a Silver Tris(triazole)amine Complex Arises from Copper Impurities. Inorg. Chem. 2014, 53, 6503−6511. (12) Donnelly, P. S.; Zanatta, S. D.; Zammit, S. C.; White, J. M.; Williams, S. J. ’Click’ cycloaddition catalysts: copper(I) and copper(II) tris(triazolylmethyl)amine complexes. Chem. Commun. 2008, 2459− 2461. (13) Presolski, S. I.; Hong, V.; Cho, S.-H.; Finn, M. G. Tailored Ligand Acceleration of the Cu-Catalyzed Azide−Alkyne Cycloaddition Reaction: Practical and Mechanistic Implications. J. Am. Chem. Soc. 2010, 132, 14570−14576. (14) Sangtrirutnugul, P.; Wised, K.; Maisopa, P.; Trongsiriwat, N.; Tangboriboonrat, P.; Reutrakul, V. Trimethylsilyl-substituted triazolebased ligand for copper-mediated single-electron transfer living radical polymerization of methyl methacrylate. Polym. Int. 2014, 63, 1869− 1874. (15) Schweinfurth, D.; Demeshko, S.; Khusniyarov, M. M.; Dechert, S.; Gurram, V.; Buchmeiser, M. R.; Meyer, F.; Sarkar, B. CappedTetrahedrally Coordinated Fe(II) and Co(II) Complexes Using a “Click”-Derived Tripodal Ligand: Geometric and Electronic Structures. Inorg. Chem. 2012, 51, 7592−7597. (16) Rechkemmer, Y.; Breitgoff, F. D.; van der Meer, M.; Atanasov, M.; Hakl, M.; Orlita, M.; Neugebauer, P.; Neese, F.; Sarkar, B.; van Slageren, J. A four-coordinate cobalt(II) single-ion magnet with coercivity and a very high energy barrier. Nat. Commun. 2016, 7, 10467. (17) Schweinfurth, D.; Sommer, M. G.; Atanasov, M.; Demeshko, S.; Hohloch, S.; Meyer, F.; Neese, F.; Sarkar, B. The Ligand Field of the Azido Ligand: Insights into Bonding Parameters and Magnetic Anisotropy in a Co(II)−Azido Complex. J. Am. Chem. Soc. 2015, 137, 1993−2005. (18) Schweinfurth, D.; Krzystek, J.; Schapiro, I.; Demeshko, S.; Klein, J.; Telser, J.; Ozarowski, A.; Su, C.-Y.; Meyer, F.; Atanasov, M.; Neese, F.; Sarkar, B. Electronic Structures of Octahedral Ni(II) Complexes with “Click” Derived Triazole Ligands: A Combined Structural, Magnetometric, Spectroscopic, and Theoretical Study. Inorg. Chem. 2013, 52, 6880−6892. (19) Schweinfurth, D.; Demeshko, S.; Hohloch, S.; Steinmetz, M.; Brandenburg, J. G.; Dechert, S.; Meyer, F.; Grimme, S.; Sarkar, B. Spin Crossover in Fe(II) and Co(II) Complexes with the Same ClickDerived Tripodal Ligand. Inorg. Chem. 2014, 53, 8203−8212. (20) Schweinfurth, D.; Weisser, F.; Bubrin, D.; Bogani, L.; Sarkar, B. Cobalt Complexes with “Click”-Derived Functional Tripodal Ligands: Spin Crossover and Coordination Ambivalence. Inorg. Chem. 2011, 50, 6114−6121. (21) Schweinfurth, D.; Demeshko, S.; Sommer, M. G.; Dechert, S.; Meyer, F.; Sarkar, B. FeII and CoII Complexes with Click-Derived Tripodal Ligands: Influence of the Peripheral Substituents on Geometric Structures and Magnetic Properties. Eur. J. Inorg. Chem. 2016, 2016, 2581−2585. (22) Schweinfurth, D.; Klein, J.; Hohloch, S.; Dechert, S.; Demeshko, S.; Meyer, F.; Sarkar, B. Influencing the coordination mode of tbta (tbta = tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine) in dicobalt complexes through changes in metal oxidation states. Dalton Trans. 2013, 42, 6944−6952. (23) Zadrozny, J. M.; Liu, J.; Piro, N. A.; Chang, C. J.; Hill, S.; Long, J. R. Slow magnetic relaxation in a pseudotetrahedral cobalt(II) complex with easy-plane anisotropy. Chem. Commun. 2012, 48, 3927− 3929. (24) Gomez-Coca, S.; Cremades, E.; Aliaga-Alcalde, N.; Ruiz, E. Mononuclear Single-Molecule Magnets: Tailoring the Magnetic

Anisotropy of First-Row Transition-Metal Complexes. J. Am. Chem. Soc. 2013, 135, 7010−7018. (25) Shao, F.; Cahier, B.; Rivière, E.; Guillot, R.; Guihéry, N.; Campbell, V. E.; Mallah, T. Structural Dependence of the Ising-type Magnetic Anisotropy and of the Relaxation Time in Mononuclear Trigonal Bipyramidal Co(II) Single Molecule Magnets. Inorg. Chem. 2017, 56, 1104−1111. (26) Zadrozny, J. M.; Long, J. R. Slow Magnetic Relaxation at Zero Field in the Tetrahedral Complex [Co(SPh)4]2‑. J. Am. Chem. Soc. 2011, 133, 20732−20734. (27) Zadrozny, J. M.; Telser, J.; Long, J. R. Slow magnetic relaxation in the tetrahedral cobalt(II) complexes [Co(EPh)4]2− (E = O, S, Se). Polyhedron 2013, 64, 209−217. (28) Ruamps, R.; Batchelor, L. J.; Guillot, R.; Zakhia, G.; Barra, A.-L.; Wernsdorfer, W.; Guihery, N.; Mallah, T. Ising-type magnetic anisotropy and single molecule magnet behaviour in mononuclear trigonal bipyramidal Co(II) complexes. Chem. Sci. 2014, 5, 3418− 3424. (29) Bruno, R.; Vallejo, J.; Marino, N.; De Munno, G.; Krzystek, J.; Cano, J.; Pardo, E.; Armentano, D. Cytosine Nucleobase Ligand: A Suitable Choice for Modulating Magnetic Anisotropy in Tetrahedrally Coordinated Mononuclear CoII Compounds. Inorg. Chem. 2017, 56, 1857−1864. (30) Idešicová, M.; Dlháň, L.; Moncol, J.; Titiš, J.; Boča, R. Zero-field splitting in tetracoordinate Co(II) complexes. Polyhedron 2012, 36, 79−84. (31) Idešicová, M.; Titiš, J.; Krzystek, J.; Boča, R. Zero-Field Splitting in Pseudotetrahedral Co(II) Complexes: a Magnetic, High-Frequency and -Field EPR, and Computational Study. Inorg. Chem. 2013, 52, 9409−9417. (32) Korchagin, D. V.; Shilov, G. V.; Aldoshin, S. M.; Morgunov, R. B.; Talantsev, A. D.; Yureva, E. A. Halogen atom effect on the magnetic anisotropy of pseudotetrahedral Co(II) complexes with a quinoline ligand. Polyhedron 2015, 102, 147−151. (33) Krzystek, J.; Swenson, D. C.; Zvyagin, S. A.; Smirnov, D.; Ozarowski, A.; Telser, J. Cobalt(II) “Scorpionate” Complexes as Models for Cobalt-Substituted Zinc Enzymes: Electronic Structure Investigation by High-Frequency and -Field Electron Paramagnetic Resonance Spectroscopy. J. Am. Chem. Soc. 2010, 132, 5241−5253. (34) Perić, M.; García-Fuente, A.; Zlatar, M.; Daul, C.; Stepanović, S.; García-Fernández, P.; Gruden-Pavlović, M. Magnetic Anisotropy in “Scorpionate” First-Row Transition-Metal Complexes: A Theoretical Investigation. Chem. - Eur. J. 2015, 21, 3716−3726. (35) Seok, W.-K.; Klapotke, T. M. Inorganic and Transition Metal Azides. Bull. Korean Chem. Soc. 2010, 31, 781−788. (36) Cirera, J.; Ruiz, E.; Alvarez, S. Continuous Shape Measures as a Stereochemical Tool in Organometallic Chemistry. Organometallics 2005, 24, 1556−1562. (37) Alvarez, S.; Alemany, P.; Casanova, D.; Cirera, J.; Llunell, M.; Avnir, D. Shape maps and polyhedral interconversion paths in transition metal chemistry. Coord. Chem. Rev. 2005, 249, 1693−1708. (38) Alvarez, S. Distortion Pathways of Transition Metal Coordination Polyhedra Induced by Chelating Topology. Chem. Rev. 2015, 115, 13447−13483. (39) Krzystek, J.; Zvyagin, S. A.; Ozarowski, A.; Trofimenko, S.; Telser, J. Tunable-frequency high-field electron paramagnetic resonance. J. Magn. Reson. 2006, 178, 174−183. (40) Telser, J.; Ozarowski, A.; Krzystek, J., High-frequency and -field electron paramagnetic resonance of transition metal ion (d block) coordination complexes. In Electron Paramagnetic Resonance; The Royal Society of Chemistry, 2013; Vol. 23, pp 209−263. (41) Ciampolini, M. Spectra of 3d five-coordinate complexes. Struct. Bonding (Berlin) 1969, 6, 52−93. (42) Lever, A. B. P. Inorganic Electronic Spectroscopy, 2nd ed.; Elsevier: Amsterdam, 1984. (43) Banci, L.; Bencini, A.; Benelli, C.; Gatteschi, D.; Zanchini, C. Spectral-Structural Correlations in High-Spin Cobalt(II) Complexes. Struct. Bonding (Berlin) 1982, 52, 37−86. 5264


Article

Inorganic Chemistry

(62) Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of Transition Ions, Oxford Classic Texts in the Physical Sciences; Oxford University Press: Oxford, UK, 2012. (63) Alvarez, S. SHAPE, 2.1; Barcelona, Spain, 2013. (64) Bendix, J., Ligfield In Comprehensive Coordination Chemistry II, Vol. 2: Fundamentals: Physical Methods, Theoretical Analysis, and Case Studies; Lever, A. B. P., Ed.; Elsevier: Amsterdam, 2003; Vol. 2, pp 673−676. (65) Ballhausen, C. J. In Introduction to Ligand Field Theory; McGraw-Hill: New York, 1962. (66) Figgis, B. N.; Hitchman, M. A. Ligand Field Theory and Its Applications; Wiley-VCH: New York, 2000. (67) Schäffer, C. E. A Perturbation Representation of Weak Covalent Bonding. Struct. Bonding (Berlin) 1968, 5, 68−95. (68) Brorson, M.; Schäffer, C. E. Orthonormal interelectronic repulsion operators in the parametrical dq model. Application of the model to gaseous ions. Inorg. Chem. 1988, 27, 2522−2530. (69) Bendix, J.; Brorson, M.; Schäffer, C. E. Accurate empirical spin orbit coupling parameters znd for gaseous ndq transition metal ions. The parametrical multiplet term model. Inorg. Chem. 1993, 32, 2838− 2849. (70) Neese, F. The ORCA program system. Wiley Interdisciplinary Reviews: Computational Molecular Science 2012, 2, 73−78. (71) ORCA 3.0 − (current version), 2015; Directorship: Frank Neese with contributions from: U.Becker (Parallelization), D. Bykov (SCF Hessian), D. Ganyushin (Spin-Orbit, Spin-Spin, Magnetic field MRCI), A. Hansen (Spin unrestricted coupled pair/coupled cluster methods), D. Liakos (Extrapolation schemes; parallel MDCI), R. Izsak (Overlap fitted RIJCOSX, COSX-SCS-MP3), C. Kollmar (KDIIS, OOCD, Brückner-CCSD(T), CCSD density), S. Kossmann (Meta GGA functionals, TD-DFT gradient, OOMP2, MP2 Hessian), T.Petrenko (DFT Hessian, TD-DFT gradient, ASA, ECA, R-Raman, ABS, FL, XAS/XES, NRVS), C. Reimann (Effective Core Potentials), M. Römelt (Restricted open shell CIS), C. Riplinger (Improved optimizer, TS searches, QM/MM, DLPNO-CCSD), B. Sandhöfer (DKH picture change effects), I. Schapiro (Molecular dynamics), K. Sivalingam (CASSCF convergence, NEVPT2), B. Wezisla (Elementary point symmetry handling), F. Wennmohs (Technical directorship). (72) Atanasov, M.; Ganyushin, D.; Pantazis, D. A.; Sivalingam, K.; Neese, F. Detailed Ab Initio First-Principles Study of the Magnetic Anisotropy in a Family of Trigonal Pyramidal Iron(II) Pyrrolide Complexes. Inorg. Chem. 2011, 50, 7460−7477. (73) Atanasov, M.; Ganyushin, D.; Sivalingam, K.; Neese, F. A Modern First-Principles View on Ligand Field Theory Through the Eyes of Correlated Multireference Wavefunctions. In Molecular Electronic Structures of Transition Metal Complexes II; Mingos, D. M. P., Day, P., Dahl, J. P., Eds.; Springer: Berlin, 2012; pp 149−220. (74) Atanasov, M.; Aravena, D.; Suturina, E.; Bill, E.; Maganas, D.; Neese, F. First principles approach to the electronic structure, magnetic anisotropy and spin relaxation in mononuclear 3d-transition metal single molecule magnets. Coord. Chem. Rev. 2015, 289−290, 177−214. (75) Weigend, F.; Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: Design and assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297−3305.

(44) Larrabee, J. A.; Alessi, C. M.; Asiedu, E. T.; Cook, J. O.; Hoerning, K. R.; Klingler, L. J.; Okin, G. S.; Santee, S. G.; Volkert, T. L. Magnetic Circular Dichroism Spectroscopy as a Probe of Geometric and Electronic Structure of Cobalt(II)-Substituted Proteins: GroundState Zero-Field Splitting as a Coordination Number Indicator. J. Am. Chem. Soc. 1997, 119, 4182−4196. (45) Larrabee, J. A.; Leung, C. H.; Moore, R. L.; Thamrongnawasawat, T.; Wessler, B. S. H. Magnetic Circular Dichroism and Cobalt(II) Binding Equilibrium Studies of Escherichia coli Methionyl Aminopeptidase. J. Am. Chem. Soc. 2004, 126, 12316−12324. (46) Johansson, F. B.; Bond, A. D.; Nielsen, U. G.; Moubaraki, B.; Murray, K. S.; Berry, K. J.; Larrabee, J. A.; McKenzie, C. J. Dicobalt II− II, II−III, and III−III Complexes as Spectroscopic Models for Dicobalt Enzyme Active Sites. Inorg. Chem. 2008, 47, 5079−5092. (47) Larrabee, J. A.; Chyun, S.-A.; Volwiler, A. S. Magnetic Circular Dichroism Study of a Dicobalt(II) Methionine Aminopeptidase/ Fumagillin Complex and Dicobalt II−II and II−III Model Complexes. Inorg. Chem. 2008, 47, 10499−10508. (48) Larrabee, J. A.; Johnson, W. R.; Volwiler, A. S. Magnetic Circular Dichroism Study of a Dicobalt(II) Complex with Mixed 5- and 6Coordination: A Spectroscopic Model for Dicobalt(II) Hydrolases. Inorg. Chem. 2009, 48, 8822−8829. (49) This band has two components, but we attribute this to resolved splitting from spin−orbit coupling, as is the case in other Co(II) complexes. (50) In a truly Td complex, eσX and eπX are correlated by 4 10Dq = 9 (3eσX − 4eπX ), so that no independent eXσ and eXπ values can be determined for such a complex. (51) The 4A2 ground state is not included for this purpose, so there are the nine quartet excited states (each is 3-fold orbitally degenerate; their spin multiplicity is not counted or indicated): T2(F), T1(F), T1(P), and there are the forty doublet excited states (spin multiplicity and free-ion parentage not indicated): T2 × 5 (3 × 5 = 15), T1 × 5 (15), E × 4 (2 × 4 = 8), A1 (1), A2 (1). (52) Neese, F.; Solomon, E. I. Calculation of Zero-Field Splittings, gValues, and the Relativistic Nephelauxetic Effect in Transition Metal Complexes. Application to High-Spin Ferric Complexes. Inorg. Chem. 1998, 37, 6568−6582. (53) Krzystek, J.; Telser, J.; Knapp, M. J.; Hendrickson, D. N.; Aromí, G.; Christou, G.; Angerhofer, A.; Brunel, L.-C. High-frequency and -field Electron Paramagnetic Resonance of high-spin manganese(III) in axially symmetric coordination complexes. Appl. Magn. Reson. 2001, 21, 571−585. (54) Such measurements on 1−4 were inconclusive, but the calculations suggest that none these complexes shows slow magnetic relaxation. (55) Hein, J. E.; Krasnova, L. B.; Iwasaki, M.; Fokin, V. V. CuCatalyzed Azide-Alkyne Cycloaddition: Preparation of Tris((1-Benzyl1H-1,2,3-Triazolyl)Methyl)Amine. Org. Synth. 2011, 88, 238−246. (56) Sheldrick, G. http://shelx.uni-ac.gwdg.de/SHELX/. (57) Bain, G. A.; Berry, J. F. Diamagnetic Corrections and Pascal’s Constants. J. Chem. Educ. 2008, 85, 532−536. (58) Bill, E. julX; Max-Planck Institute for Chemical Energy Conversion: Mülheim/Ruhr, Germany, 2008. (59) Bill, E. LAPACK, NUMERICAL RECIPIES; Max-Planck Institute for Chemical Energy Conversion, Mülheim/Ruhr, Germany, 2014. Matrix diagonalization is done with the routine ZHEEV from the LAPACK numerical package. Parameter optimization is performed with the simplex routine AMOEBA from NUMERICAL RECIPIES. (60) Zvyagin, S. A.; Krzystek, J.; van Loosdrecht, P. H. M.; Dhalenne, G.; Revcolevschi, A. Field-induced structural evolution in the spinPeierls compound CuGeO3: high-field ESR study. Phys. Rev. B 2003, 67, 212403. (61) Hassan, A. K.; Pardi, L. A.; Krzystek, J.; Sienkiewicz, A.; Goy, P.; Rohrer, M.; Brunel, L.-C. Ultrawide band multifrequency high-field EMR technique: a methodology for increasing spectroscopic information. J. Magn. Reson. 2000, 142, 300−312. 5265


Tuning Magnetic Anisotropy Through Ligand Substitution in Five

Recommend Documents