Single-Molecule Magnets Based on Heteroleptic Terbium(III) Trisphthalocyaninate in Solvent-Free and Solvent-Containing Forms

Abstract

Binuclear heteroleptic triple-decker terbium(III) phthalocyaninate (Pc)Tb[(15C5)₄Pc]Tb(Pc), where Pc²⁻ is phthalocyaninate dianion and 15C5 is a 15-crown-5 moiety, has been synthesized as a solvent-free powder (1) and a well-defined crystal solvate with o-dichlorobenzene (Pc)Tb[(15C5)₄Pc]Tb(Pc)⋅6C₆H₄Cl₂ (2). In the crystal structure of 2, the Tb-N(Pc) distances to the nitrogen atoms in the outer and inner decks are 2.350–2.367(4) and 2.583–2.598(4) Å, respectively, and the Tb–Tb distance is 3.4667(3) Å. The twist angle between the outer and the inner decks is 42.6°. The magnetic properties were studied for both 1 and 2. The χ_MT magnitude of 23.3 emu⋅K/mol at 300 K indicates a contribution of two Tb^III centers with the ⁷F₆ ground state. The χ_MT product increases with decreasing temperature to reach 38.5 emu⋅K/mol at 2 K. This is indicative of ferromagnetic coupling between Tb^III spins in accordance with previous data for triple-decker lanthanide phthalocyaninates of a dipolar nature. Both forms show a single-molecule magnet (SMM) behavior manifesting the in-phase (χ′) and out-of-phase (χ″) AC susceptibility signals in an oscillating field of 3 Oe with estimated effective spin-reversal energy barriers (U_eff) of 222(9) and 93(7) cm⁻¹ for 1 and 2, respectively. The compounds show narrow hysteresis loops in the −1 – +1 kOe range, and the splitting between the zero-field-cooling and field-cooling curves is observed below 6 K. Thus, in spite of similar static magnetic characteristics, each form of the Tb(III) complex shows a different dynamic SMM behavior.

1. Introduction

Single-molecule magnets (SMMs) can show a slow relaxation of magnetization below blocking temperatures. Great progress has been made in this field as increasingly higher spin-reversal barriers (U_eff) have been achieved, and several synthetic strategies have been offered to increase the blocking temperatures for these SMMs [1,2,3,4]. SMMs can potentially find application in molecular electronics, for example, for high-density information storage, spintronics, quantum computers and other devices due to blocked magnetic moment and essential quantum effects [5,6,7,8,9,10]. Among different SMMs, lanthanides are promising candidates due to high unquenched orbital moments and significant inherent spin-orbit coupling, which can lead to strong single-ion anisotropy [11,12,13,14].

In spite of the maximal barriers and blocking temperatures found for substituted metallocene Dy(III) derivatives [4,15], terbium(III) phthalocyanine complexes also attract great attention due to the possibility of the chemical modification of their structure and properties [16], as well as their high air stability and the relative ease of fabricating films to be used in different devices [14]. The SMM properties of double-decker phthalocyanine complexes were first reported by Ishikawa for the Tb^IIIPc₂ complexes with a U_eff of 230 and 410 cm⁻¹ [1,17]. After that, the barriers were increased to 642–652 cm^–1 for heteroleptic Tb^IIIPc₂ derivatives with eight electron-rich NBu₂ or p-tBuPhO substituents on the periphery of one of the Pc macrocycles [18,19].

Triple-decker phthalocyanines can also show properties of SMMs [20,21,22,23,24,25,26]. All these systems also have D_4d axial symmetry, giving a possibility for a significant orbital contribution from the Tb^III ions. This can provide a large separation between the states and, as a result, high U_eff values, which, however, are lower than the barriers for the double-decker complexes [27] because of the nonequivalence of the bond lengths between Tb^III and the outer/inner Pc ligands. Nevertheless, triple-decker complexes provide wider possibilities to study the influence of structural and electronic factors on SMM behavior [14], as these scaffolds show a higher conformational flexibility in comparison with double-decker sandwiches [28,29].

In this work, we have synthesized triple-decker Tb^III phthalocyaninate with the inner 15-crown-5-substituted deck (Pc)Tb[(15C5)₄Pc]Tb(Pc) and have studied its magnetic properties. The (Pc)Tb[(15C5)₄Pc]Tb(Pc) complex has been isolated as a solvent-free powder (1) and single crystals of solvate with o-dichlorobenzene (Pc)Tb[(15C5)₄Pc]Tb(Pc)⋅6C₆H₄Cl₂ (2). Its crystal structure has been solved for the first time and discussed. Magnetic properties studied for both the powder and crystal solvate have revealed ferromagnetic coupling of the spins in the binuclear units. Both forms show SMM behavior with different effective spin-reversal barriers.

2. Results and Discussion

The choice of the ligand surrounding for the studied complex is based on our previous experience in SMM studies of crown-substituted heteronuclear triple-decker complexes [(15C5)₄Pc]M*[(15C5)₄Pc]M(Pc), where M and M* is Tb or Y [26]. The Tb(III) ion placed in the homoleptic position M* has smaller U_eff (90 cm^–1) in comparison with the heteroleptic position M (117 cm⁻¹). Thus, (Pc)Tb[(15C5)₄Pc]Tb(Pc) studied herein has both metal centers in an asymmetric heteroleptic position, which expectedly can provide better SMM properties. Whereas the synthesis of (Pc)Tb[(15C5)₄Pc]Tb(Pc) has been reported elsewhere [30], we have made an attempt to improve its preparation. Previously, this complex was isolated in 19% yield by the reaction of tetra-15-crown-5-phthalocyanine with lanthanum bisphthalocyaninate La(Pc)₂ and terbium(III) acetylacetonate in refluxing 1-chloronaphthalene. A lanthanum (III) complex was used as a convenient source of Pc²⁻ species via transmetalation reaction. The modest yield of (Pc)Tb[(15C5)₄Pc]Tb(Pc) is explained by the statistical nature of this reaction, while in the present work, we aim at making it more selective towards the target complex. Therefore, we use the “raise-by-one-story” approach which implies the addition of one more deck to the double-decker complex [(15C5)₄Pc]Tb(Pc) to give a triple-decker sandwich (Scheme 1).

Starting [(15C5)₄Pc]Tb(Pc) was synthesized as described earlier [31]. The synthesis of another precursor—La(Pc)₂—was also reported previously. It was isolated in 47% yield after the reaction of H₂(Pc) with La(acac)₃ nH₂O in refluxing 1-chloronaphthalene in the presence of 1,8-diazabicyclo [5.4.0]undec-7-ene (DBU) [32]. Herein, we show that the addition of n-octanol (OctOH) to the reaction mixture allows one to decrease the reaction time from 3 h to 30 min and increase the yield of La(Pc)₂ to 85%.

The reaction between [(15C5)₄Pc]Tb(Pc) and La(Pc)₂ in the presence of Tb(acac)₃ nH₂O proceeded in a refluxing mixture of 1,2,4-trichlorobenzene (TCB) and OctOH. Monitoring the consumption of starting reagents using UV-vis indicated an exceptionally high reaction rate affording the formation of the target complex within 3–5 min; the latter was isolated in 80% yield using column chromatography with a mixture of chloroform and 0.5–1 vol.% methanol. The evaporation of these solvents from the fractions afforded the powdered form of complex 1, and the absence of solvent molecules was confirmed using elemental analysis. According to the X-ray powder diffraction pattern for 1 (Figure S1), its crystal structure is similar to that of the known (Pc)Sm[(15C5)₄Pc]Sm(Pc) 8CHCl₃ complex [30], with the doubling of the unit cell along the c direction and a slightly smaller unit cell volume.

Slow mixing of a solution of the powder in o-dichlorobenzene (prepared under anaerobic conditions) with n-hexane produced the high-quality large single crystals of 2. All crystals had the same color and shape, and the testing of several crystals from the synthesis showed that only one crystal phase was formed with the same unit cell parameters. The content of C₆H₄Cl₂ was supported using elemental analysis. Crystals were additionally characterized with IR and UV-Vis-NIR spectra in KBr pellets (Figures S2 and S3).

The molecular structure of complex 2 is shown in Figure 1. The asymmetric unit contains only a half of the triple-decker phthalocyanine and three solvent o-dichlorobenzene molecules, one of which is disordered over three orientations. Therefore, the composition of 2 is (Pc)Tb[(15C5)₄Pc]Tb(Pc)⋅6C₆H₄Cl₂. Several parts of the 15-crown-5 rings also show the conformational disorder of the -CH₂-CH₂- moieties, but this disorder cannot be described well by two positions for each fragment. The Tb-N(Pc) distances with the nitrogen atoms of the outer and inner decks are different, being 2.350–2.367(4) and 2.583–2.598(4) Å, respectively. The Tb–Tb distance of 3.4667(3) Å is shorter than the sum of the van der Waals radii of the two terbium atoms, 4.74 Å. The twist angle between the outer and inner decks is 42.6°. This angle is close to 45°, which corresponds to the D_4d axial symmetry. The skew angle in 2 is close to those observed for double-decker terbium(III) phthalocyanines showing SMM behavior. In general, the structural characteristics of the triple-decker molecules in 2 are similar to the structure of the (Pc)Sm[(15C5)₄Pc]Sm(Pc)⋅8CHCl₃ solvate [30].

The triple-decker phthalocyanines form chains arranged along the a axis. They are positioned in such a way that two 15C5 rings of one molecule are positioned over two such rings from the neighboring molecules (Figure S4). The hydrogen atoms of one 15C5 ring are directed to the oxygen atoms of the neighboring ring and form several H…O hydrogen bonds 2.57–2.69 Å in length indicating that hydrogen bonding is a driving force for the packing of (Pc)Tb[(15C5)₄Pc]Tb(Pc) molecules into chains. There are two types of chains directed along the a axis. These chains are arranged in such a way that the phthalocyanine planes belonging to two different chains form an angle of 59.6° (see Figure S4).

The (Pc)Tb[(15C5)₄Pc]Tb(Pc) molecules are surrounded by o-dichlorobenzene molecules that occupy cavities in their packing. Ordered molecules are positioned mainly between the 15C5 rings and above and below the 15C5 rings. The solvent molecules are also positioned above and below the outer Pc macrocycles separating the triple-decker phthalocyanine chains (Figure S5). As a result, multiple shortened Cl(C₆H₄Cl₂)…H(Pc) and O(15C5)…H(C₆H₄Cl₂) contacts are formed.

Static and dynamic magnetic susceptibilities were studied for powdered 1 (Figures S6–S11) and crystalline 2 (Figures S12–S15) sealed in quartz tubes on a SQUID magnetometer. Both complexes are EPR silent in the whole studied temperature range (4.2–295 K) for X-band EPR in the 70–700 mT magnetic field.

Complexes 1 and 2 show similar temperature dependencies of molar magnetic susceptibility. The χ_MT value is 23.3 emu K/mol at 300 K (Figure 2a and Figure 3a). The Tb^III ions have the ⁷F₆ ground state [S = 3, L = 3, g = 3/2], whereas the phthalocyanine macrocycles are diamagnetic. The calculated value for two independent spins with such configuration is 23.62 emu K/mol. This value is close to the experimentally observed one. The χ_MT product increases with decreasing temperature in both samples (Figure 2a and Figure 3a), indicating the ferromagnetic coupling of the spins between two terbium atoms within nearly isolated triple-decker phthalocyanine molecules. This is supported by positive Weiss temperatures, which are +1.5 and +1.7 K, respectively, for 1 and 2 (Figures S8 and S14). The highest value of χ_MT of 38.5 emu K/mol is attained at 2 K (Figure 2a and Figure 3a). This value is only slightly lower than 43.8 emu K/mol calculated for the system where spins of both terbium(III) atoms are arranged parallel to each other, and correspondingly, the system has S = 12 spin state. Previously, the magnetic properties of triple-decker phthalocyanines with different metal atoms have been studied. It has been demonstrated that terbium, dysprosium and holmium show ferromagnetic coupling; according to the calculations, this interaction has mainly a dipole nature, whereas the contribution from the exchange between two Tb^III spins is negligible [33]. Clearly, intermolecular coupling gives no essential contribution to the magnetic properties of this compound as the shortest distances between the Tb^III pairs separated by diamagnetic macrocycles are 15.3–16.6 Å. In both samples, the spins of the Tb^III atoms are arranged parallel to each other at 2 K, and clearly, this state coexists with the SMM behavior discussed below and observed in the same temperature range.

Both samples manifest the in-phase (χ′) and out-of-phase (χ″) AC susceptibility signals when measured in a zero DC field and an oscillating field of 3 Oe (Figures S11 and S15). It is seen that both compounds show a frequency-dependent character of χ′ and χ″, revealing their slow magnetic relaxation and SMM behavior without the need for an external DC field. The temperature dependences of the relaxation times for 1 and 2 show the absence of any temperature-independent regimes even at low temperatures (Figure 2b and Figure 3b). This allows one to fit the data of ln(τ) vs. 1/T without a QTM term. Fitting the data was made with Equation (1).

$${\tau }^{-1}={\tau }_0{}^{-1}{e}^{-U_{\mathrm{eff}}/k_{B}T}+C{T}^n$$

(1)

Here, ${\tau }_0$ and U_eff are parameters of the Orbach process, and $C$ and $n$denote the Raman parameters. The fitting gave the energy barrier U_eff = 222(9) cm^–1 (319(13) K) with ${\tau }_0$ = 1.3(2)⋅10⁻⁹ s and the Raman parameters C = 2.1(2) s^–1K^–n and n = 1.61(4) for the powdered sample 1 (R = 0.998) (Figure 2b). For the crystal solvate 2, an adequate fitting to the experimental data was achieved at U_eff = 93(7) cm^–1 (134(10) K) with = 6.7(3)⋅10⁻⁷ s and the Raman parameters C = 0.8(1) s⁻¹K⁻ⁿ and n = 2.18(7) (R = 0.997) (Figure 3b). It has been demonstrated that two relaxation processes are possible for diterbium triple-decker phthalocyanines, and one process can be observed at higher frequencies and temperatures [23,24]. An essential increase in the χ″ values was observed in 2 when the frequency increased above 400 Hz (Figure S15). However, if maxima were present, they would be observed above a frequency of 1500 Hz, which was the maximum for the SQUID used in this work. As a result, no information on the possible maxima in the high-frequency range could be obtained.

A comparison of the zero-field-cooling and field-cooling curves (Figures S7 and S13) demonstrates that the difference between these curves is observed below 6 K, which allows one to evaluate the blocking temperature (T_b) for both samples. It turns out that 1 and 2 have the same blocking temperatures, though the U_eff values differ more than twice. The magnetization curve for 1 is shown in Figure S9. Magnetization is quickly saturated in a magnetic field of 8–10 kOe, and then it is temperature-independent up to 50 kOe. The maximal value attained at high magnetic fields is only 10.8 μ_BN_A. This value is nearly two times lower than the 18.2 μ_BN_A value expected for the isolated pairs with a parallel arrangement of Tb^III spins. Previously, it has been shown that the magnetization in triple-decker phthalocyanines is strongly anisotropic, and the maximal value can be attained when a crystal is oriented in such a way that the Tb-Tb pair is aligned exactly with the magnetic field direction, while for the perpendicular orientation, the magnetization decreases several times [21]. Therefore, the maximal available magnetization cannot be attained for randomly oriented polycrystals of 2.

The compounds show a rather narrow hysteresis loop in the magnetic field of 20 Oe/s, as demonstrated in Figure 4. This loop is observed in the –1 – +1 kOe magnetic field range. The magnetic hysteresis width for crystals is larger than for the powder (Figure 4). It is likely that the wider hysteresis loop for the powder can be observed only at temperatures essentially lower than 1.9 K. For example, a hysteresis loop is not observed for Tb₂[(BuO)₈Pc]₃ above 1.5 K, and a wider hysteresis loop is found only at very low temperatures (T = 0.04 K) [21].

Thus, in spite of the similar magnetic characteristics of both forms of (Pc)Tb[(15C5)₄Pc]Tb(Pc), they show different dynamic behavior, which is manifested in a more than two times higher U_eff value for powder 1 (222 cm⁻¹) in comparison with crystal solvate 2 (93 cm⁻¹). It is seen that the formation of well-defined crystals and the presence of solvent (C₆H₄Cl₂) molecules, which isolate and surround triple-decker (Pc)Tb[(15C5)₄Pc]Tb(Pc) phthalocyanines, do not increase the barrier compared to the solvent-free powdered phase. Unfortunately, the exact crystal structure of 1 is unknown, which prevents the determination of the difference in the geometry of the Tb^III environment in both forms. We can only suppose that the appearance of the solvent molecules can affect the dihedral angle between the Pc planes or some other parameters.

Previously, it has been elucidated that the twist angle between the Pc macrocycles can affect the D_4d symmetry of Tb^III and, hence, the dynamic magnetic properties of SMMs [14]. Thus, the different packing and arrangement of the surrounding cations and solvent molecules can affect the geometry and properties of double- and triple-decker complexes. For example, the cationic surrounding of the double-decker Tb^III(Pc)₂ anions changes when PPN⁺ (bis(triphenylphosphoranylidene)ammonium) cations are replaced by the {Cryptand(Na⁺)} cations. This leads to a change in the twist angle from 41.2 to 38.8° and, as a result, the first compound is an SMM with a high U_eff = 538 cm⁻¹,whereas the second compound shows no signs of SMM behavior [34]. There are several examples of different SMM behaviour depending on the nature of the solvent molecules in the crystal lattice. In particular, Tb(Pc)₂ can be crystallized from either CH₂Cl₂/MeOH or CHCl₃/MeOH, affording the solvate Tb(Pc)₂ CH₂Cl₂ or solvent-free crystals, respectively [35]. In spite of only a subtle difference in the geometries of the complex itself in both samples, a butterfly-shaped hysteresis loop is observed up to 10 K only in the case of the solvate, while solvent-free crystals show almost no opening even at 2 K. Binuclear monophthalocyaninate (μ-Pc)[Er(dpm)₂]₂ (dpm⁻—dipivaloylmethanate anion) can be solvated either with C₆H₆ or CH₂Cl₂ in the course of “crystal-to-crystal” transformation [36]. Fast relaxation with a U_eff = 2.6 cm^–1 under a 600 Oe dc field is observed for the benzene solvate, while the dichloromethane solvate behaves as a field-induced SMM with a higher U_eff = 34.3 cm⁻¹.

The determined barriers are attributed to the energy gap between the ground and first-excited Stark sublevels. The barrier for powder 1 is higher than that for the powdered heteroleptic triple-decker terbium(III) phthalocyanine with two (15C5)₄Pc decks—159 cm⁻¹. Therefore, when both terbium atoms are bound in an asymmetric heteroleptic environment, this can increase the barrier in comparison with the situation when the terbium atoms are in hetero- and homoleptic positions as in [(15C5)₄Pc]Tb[(15C5)₄Pc]Tb(Pc) [26]. Nevertheless, the homoleptic phthalocyanine with three octa(butoxy)phthalocyanine decks has nearly the same barrier [21]. At the same time, the U_eff barrier determined for the crystals of 2 is one of the lowest among triple–decker terbium(III) phthalocyanines. Previously, it has been demonstrated that the dynamic characteristics of SMMs based on triple-decker complexes depend mainly on the twist angle between Pc planes, and a dipolar term can affect the dynamic properties.¹⁴ The twist angle of 42.6° in 2 is close to 45°, which is favorable for manifestation of high U_eff. The f-f dipole interaction generally acts as exchange bias, suppressing the quantum tunneling. As a result, triple-decker complexes showing the f-f interaction generally manifest higher U_eff barriers in comparison with similar phthalocyanines but containing only one Tb^III atom and another diamagnetic lanthanide. For example, the barrier for triple-decker (Tb, Tb) phthalocyanine with two (15C5)₄Pc decks is 159 cm⁻¹, whereas such barriers for triple-decker phthalocyanines with the (Tb, Y) and (Y, Tb) combinations are only 90 and 117 cm⁻¹, respectively [26]. The dipolar term depends substantially on the distance between two terbium ions proportionally to 1/R³, where R is the distance between two terbium atoms [33]. Information on the effect of these distances on the barrier can be obtained only from the oxidation of Tb₂[(BuO)₈Pc]₃. It shows the SMM behavior with U_eff = 230 cm⁻¹ at a dihedral angle of only 32^o and the Tb-Tb distance of 3.517 Å [21]. As the HOMO of this molecule is antibonding relative to the Tb–N bonds, oxidation depopulates the HOMO, decreasing the Tb–Tb distance in {Tb₂[(BuO)₈Pc]₃}²⁺(SbCl₆⁻)₂ to 3.435 Å [37]. As a result, the U_eff value decreases to 167 cm⁻¹ despite the fact that the dihedral angle in this complex, 44.2°, is close the 45°, which should provide a higher U_eff. It should be noted that both compounds show the f-f interactions with an increase in χ_MT values at 2 K (this increase is higher for the oxidized dicationic sample). Therefore, a decrease in the Tb-Tb distance can enhance the f-f interactions and decrease U_eff. The Tb–Tb distance of 3.4667(3) Å in 2 is one of the shortest distances among neutral triple-decker terbium(III) phthalocyanines. Therefore, the contribution of the dipole term is rather high in 2 and can be one of the reasons for the decrease of U_eff.

Some of the crystallographic and magnetic data for 1 and 2 are summarized in

Table 1. Summary of selected structural and magnetic characteristics of samples 1 and 2.

Compound	1	2
The Tb-Tb distance, Å	-	3.4667 (3)
Twist angle between inner and outer planes, °	-	42.6
Ueff, cm−1	222	93
Blocking temperature, K	~6	~6
Type of coupling between TbIII spins	ferromagnetic	ferromagnetic

.

3. Materials and Methods

3.1. Synthesis

Starting phthalocyanines were synthesized according to the previously reported procedures [31,38]. o-dichlorobenzene (C₆H₄Cl₂) was distilled over CaH₂ under reduced pressure, and n-hexane was distilled over Na/benzophenone. The solvents were degassed and stored in a glove box. Chloroform was distilled over CaH₂. All other solvents were used as received from commercial suppliers.

All manipulations for the isolation of solvate 2 were carried out in an MBraun 150B-G glove box with controlled atmosphere and a content of H₂O and O₂ less than 1 ppm. The crystals were stored in the glove box and sealed in 2 mm quarts tubes under ambient pressure in argon for EPR and SQUID measurements. KBr pellets for the IR- and UV-visible-NIR measurements were prepared in the glove box.

3.1.1. Improved Procedure for the Synthesis of the La(Pc)₂ Complex

Unsubstituted phthalocyanine (257 mg, 0.5 mmol) was suspended in a mixture of 4 mL of 1-chloronaphthalene and 2 mL of octanol, and DBU (152 mg, 1 mmol, 2 equiv.) was added. The reaction mixture was brought to reflux under an argon atmosphere, and lanthanum acetylacetonate (133 mg, 0.3 mmol, 0.6 equiv.) was added. After half an hour, the reaction mixture was cooled down to room temperature, and a mixture of n-hexane and ethyl acetate (4:1 v/v) was added. The precipitate was filtered off, washed with the same solvent mixture and washed off with a mixture of chloroform and methanol, and the solution was evaporated under reduced pressure. The resulting solid was dissolved in chloroform, and 500 mg of manganese dioxide was added; the resulting mixture was sonicated in an ultrasonic bath for 30 min. The resulting suspension was transferred to a chromatographic column packed with neutral alumina in chloroform, and lanthanum bisphthalocyanate was eluted with a mixture of chloroform with 0 → 4 vol.% methanol. The resulting fractions were combined and evaporated under reduced pressure. The resulting black powder was washed with n-hexane, and the complex was dried overnight at 80 °C. The yield was 245 mg (85%). MALDI TOF: m/z calcd. for C₆₄H₃₂LaN₁₆ 1163.2 found 1163.2—[M]⁺. UV-vis (CHCl₃) λ_max, nm (A_norm) 691 (0.87), 614 (0.34), 490 (0.21), 450 (0.17), 326 (1.00), 278 (0.43). The analytical characteristics of the complex are consistent with previously reported data [32].

3.1.2. Improved Procedure for the Synthesis of (Pc)Tb[(15C5)₄Pc]Tb(Pc)

A mixture of complexes [(15C5)₄Pc]Tb(Pc) (37 mg, 19.0 µmol [31]) and La(Pc)₂ (22 mg, 19.0 µmol) with terbium acetylacetonate hydrate (27 mg, 57.1 µmol) was refluxed in a mixture of 1,2,4—trichlorobenzene (2.15 mL) with n-octanol (0.23 mL) for 7 min under argon. The reaction mixture was poured into n-hexane and filtered. The solid was washed with n-hexane and dissolved in chloroform, and the target complex (Pc)Tb[(15C5)₄Pc]Tb(Pc) was isolated in 71% yield (35.3 mg) using column chromatography on neutral alumina (CHCl₃ + 0.75% vol MeOH), followed by drying overnight at 80 °C, thus affording a powder sample of 1. MALDI TOF: m/z calcd. for C₁₂₈H₁₀₄N₂₄O₂₀Tb₂ 2615.6 found 2615.9—[M]⁺. UV-vis (CHCl₃) λ_max, nm (A_norm) 696 (0.17), 642 (1.00), 332 (0.63), 294 (0.40). Anal. calc. for C₁₂₈H₁₀₄N₂₄O₂₀Tb₂: C 58.76, H 4.01, Tb 12.15 found C 57.76, H 4.16, Tb 11.82. Chlorine was not found in the samples, suggesting complete removal of chloroform upon sample drying. The analytical characteristics of the complex are consistent with reported data [30].

3.1.3. Preparation of Single Crystals of (Pc)Tb[(15C5)₄Pc]Tb(Pc)⋅6C₆H₄Cl₂ (2)

First, 25 mg of (Pc)Tb[(15C5)₄Pc]Tb(Pc) (0.0095 mmol) was dissolved in 17 mL of o-dichlorobenzene during 24 hours at 50 °C. Obtained green solution was cooled down to room temperature and filtered into a glass tube of 46 ml volume, and n-hexane (26 mL) was layered onto the obtained solution. Slow diffusion of n-hexane during 1 month yielded crystals on the walls of the tube suitable for X-ray diffraction analysis. The solvent was decanted from the crystals, and they were washed with n-hexane to give black blocks of 2 with 62% yield. The composition of the crystals determined from the X-ray diffraction was (Pc)Tb[(15C5)₄Pc]Tb(Pc)⋅6C₆H₄Cl₂ (2). All crystals had the same shape and color. Several crystals tested from the synthesis gave the same unit cell parameters indicating that only one crystal phase was formed. Anal. calc. for C₁₆₄H₁₂₈Cl₁₂N₂₄O₂₀Tb₂: Cl 12.18, found Cl 11.94.

IR-spectrum of 2: 492 w, 565 w, 630 w, 719 w, 732 s, 743 s, 756 w, 780 w, 856 w, 883 m, 934 m, 981 w, 1002 w, 1061 m, 1081 s, 1113 s, 1199 w, 1273 s, 1315 w, 1329 m, 1356 w, 1374 w, 1407 m, 1450 m, 1478 s, 1607 w, 2867 w, 2921 w, 3047 w, 3077 w cm^–1 (w: weak; m: medium; s: strong intensity). IR- and optical spectra are shown in Figures S1 and S2, respectively.

Crystal data for 2 at 100(1) K: C₁₆₄H₁₂₈Cl₁₂N₂₄O₂₀Tb₂, F.W. 3498.14, black prism, 0.400 × 0.190 × 0.080 mm, monoclinic, space group P2₁/n, a = 15.6657(3), b = 23.5878(6), c = 20.3168(6) Å, β = 96.146(2) °, V = 7464.3(3) ų, Z = 2, d_calcd = 1.556 M gm⁻³, μ = 1.233 mm^–1, F(000) = 3548, 2θ_max = 54.206°; 50,991 reflections collected, 15,290 independent; R₁ = 0.0428 for 11,779 observed data [>2σ(F)] with 698 restraints and 1090 parameters; wR₂ = 0.1050 (all data); final GoF = 1.049. CCDC 2225666.

3.2. General

UV-visible-NIR spectra were recorded as KBr pellets on a Perkin Elmer Lambda 1050 spectrophotometer in the 250–2500 nm range. FT-IR spectra were recorded as KBr pellets on a Perkin-Elmer Spectrum 400 spectrophotometer (400–7800 cm⁻¹). A Quantum Design MPMS-XL SQUID magnetometer was used to measure the static magnetic susceptibility of 1 and 2 under a magnetic field of 100 mT, under cooling and heating conditions in the 300–1.9 K range. The sample-holder contribution and core temperature-independent diamagnetic susceptibility (χ_d) were subtracted from the experimental values. The χ_d values were estimated from the extrapolation of the data in the high-temperature range by fitting the data with the following expression: χ_M = C/(T − Θ) + χ_d, where C is the Curie constant and Θ is the Weiss temperature. The dynamic magnetic properties of 1 and 2 were studied at 3.0 Oe ac field at the 1−1500 Hz frequencies using a Quantum Design MPMS-5S SQUID magnetometer.

3.3. X-ray Crystal Structure Determination

X-ray diffraction data for 2 were collected on an Oxford diffraction “Gemini-R” CCD diffractometer with graphite monochromated MoK_α radiation using an Oxford Instrument Cryojet system. Raw data reduction to F² was carried out using CrysAlisPro, Oxford Diffraction Ltd. (Abingdon, UK). The structures were solved using direct methods and refined using the full-matrix least-squares method on F² with SHELX-2018/3.[39] Non-hydrogen atoms were refined anisotropically. Positions of hydrogen atoms were included into refinement in a riding model. Crystal structure of 2 contains half independent (Pc)Tb[(15C5)₄Pc]Tb(Pc) unit which is ordered and has full occupancy. In addition, there are three positions of solvent C₆H₄Cl₂ molecules. Two molecules are ordered, but the third molecule is disordered over three orientations with the 0.422(2)/0.419(2)/0.159(2) occupancies. To keep anisotropic thermal parameters of the disordered fragments, the displacement components were restrained using SHELXL instructions of ISOR, SIMU and DELU. That resulted in the use of 698 restraints for the refinement of crystal structure of 2.

4. Conclusions

We have obtained solvent-free (1) and solvent-containing (2) forms of triple-decker phthalocyanine (Pc)Tb[(15C5)₄Pc]Tb(Pc) with the inner 15-crown-5-substituted deck and studied the crystal structure of the o-dichlorobenzene solvated form. These forms also have different crystallinity as 2 was obtained as well-defined single crystals in contrast to powdered 1. Both forms have similar static magnetic properties manifesting ferromagnetic coupling between Tb^III spins. Such behavior is typical of triple-decker terbium(III) phthalocyanines, and this coupling has mainly a dipolar nature [33]. In spite of similar static magnetic properties, the solvent-free form shows more than two times higher U_eff (222 cm⁻¹) in comparison with the o-dichlorobenzene crystal solvate (93 cm⁻¹). Most probably, this difference is explained by the effect of the solvent molecules on the skew angle between the Pc planes and/or some other parameters. As the shortening of the Tb–Tb distance can enhance the f-f coupling and decrease the U_eff barrier, we suppose that the Tb–Tb distance in 2 is minimal among the described neutral triple-decker Tb^III complexes, which can be one of the reasons for a low U_eff barrier. Information on the dynamic properties of structurally characterized triple-decker terbium(III) complexes is still rather limited, and only additional studies will be able to shed light onto the exact correlations between the dynamic properties of SMMs and their geometric parameters.