View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/D3CC03034B (Communication) Chem. Commun., 2023, 59, 13970-13973

Probing the magnetic and magneto-optical properties of a radical-bridged Tb4 single-molecule magnet†

Niki Mavragani a, Alexandros A. Kitos a, Diogo A. Gálico a, Akseli Mansikkamäki b and Muralee Murugesu *a
aDepartment of Chemistry and Biomolecular Sciences, University of Ottawa, Ottawa, ON K1N 6N5, Canada. E-mail: m.murugesu@uottawa.ca
bNMR Research Unit, University of Oulu, P.O. Box 8000, FI-90014, Finland

Received 23rd June 2023 , Accepted 29th October 2023

First published on 30th October 2023

Abstract

Reaction of the 1,2,4,5-tetrazine (tz˙−) radical and {Cp*2Tb}+ has yielded a tetranuclear radical-bridged TbIII single-molecule magnet. The strong Ln–radical coupling and the electronic differences of the TbIII ions in [(Cp*2Tb)4(tz˙−)4]·3C6H6 (1) are probed via magnetic, magneto-optical and computational studies.

In high-performing single-molecule magnets (SMMs), lanthanides (Ln) have always attracted significant scientific interest. Their large inherent anisotropy, stemming from a near-unquenched orbital angular momentum and strong spin–orbit coupling, has allowed some mononuclear Ln-bearing SMMs to exhibit hysteresis of the magnetization at temperatures as high as that of liquid nitrogen.1,2 This has enabled us to consider SMMs as potential candidates in spin-based electronics, high-density information storage and quantum computing.3 Although many lanthanide ions have been explored for their magnetic interest,4,5 most of the high-performing SMMs are based on Kramers ions (with half-integer spins) such as DyIII (6H15/2 ground term). For the non-Kramers ions, such as TbIII, with an integer spin (7F6), the ±mJ degeneracy is not necessarily guaranteed, thus in these cases, the quantum tunnelling of the magnetization (QTM) is more pronounced. A way to surpass this, is by imposing the ideal local symmetry around the TbIII ions6,7 or inducing strong magnetic coupling between the metal centers.8

However, the latter proves to be challenging due to the core-like nature of the 4f orbitals shielded by the outer 5s and 5p orbitals. Due to negligible magnetic interactions in multinuclear LnIII-based complexes, the observed SMM behaviour arises from single ion anisotropy rather than a collective behaviour of metal ions.9 To overcome this, a radical bridging ligand can be used as a linker.10 With that said, it is critical to consider a radical ligand with a diffuse HOMO orbital that can penetrate the electron density of the outer shells of the Ln ions and subsequently enhance the coupling with the 4f magnetic orbitals.11 Thus, strong Ln–rad coupling is anticipated to improve the magnetic performance owing to the collective behaviour of the spin centres. Amongst the many available organic radicals, tetrazines offer significant advantages such as their synthetic accessibility, structural modification and their easily accessible π* LUMO, which facilitates the formation and stabilization of the radical species.12

Based on these principals, we recently demonstrated that the incorporation of the unsubstituted 1,2,4,5-tetrazine (tz) into dinuclear “Ln2” [(Cp*2Ln)2(tz˙−)(THF)]BPh4 (Ln = Gd, Tb or Dy; Cp* = pentamethylcyclopentadienyl; THF = tetrahydrofuran)13 and tetranuclear “Ln4” [(Cp*2Ln)4(tz˙−)4]·3C6H6 (Ln = Dy or Gd) metallocenes, led to hard-magnet-like behavior.14 The diffuse spin density of the tz˙− effectively promoted strong Ln–rad magnetic coupling which led to remarkable magnetic performance for the “Dy4”, displaying open hysteresis loops with Hc = 30 kOe, as well as slow relaxation of the magnetization via an Orbach process. Given the significantly strong magnetic coupling promoted by tz˙− and the considerable single-ion anisotropy of the 4f8 system of the TbIII ion, herein, we wanted to expand this concept and investigate the effects of the strong Ln–rad coupling for the first time in the “Tb4” analogue. As such, we present the synthesis and magneto-optical studies using both SQUID magnetometry and MCD spectroscopy, as well as computational studies, to probe the magnetic performance of this SMM.

Following a similar synthetic procedure to the previously published “Dy4” and “Gd4”,14 the equimolar reaction of [Cp*2Tb][BPh4] with tz and freshly prepared potassium graphite (KC8) in benzene (Fig. 1(A)), afforded a dark-red solution that led to the formation of dark red prismatic crystals of [(Cp*2Tb)4(tz˙−)4]·3(C6H6) (1) (Fig. 1(B) and Fig. S2–S5, ESI†). Single-crystal X-ray diffraction analysis reveals that 1 crystallizes in the orthorhombic Cmca space group similarly to the isostructural “Dy4” and “Gd4” analogues (Fig. S1, ESI†). Crystallographic details, as well as selected bond distances and angles are summarized in Tables S1–S3 (ESI†), along with the analysis of the structural features of 1 (see ESI†).


image file: d3cc03034b-f1.tif
Fig. 1 (A) Synthesis of [(Cp*2Tb)4(tz˙−)4]·3(C6H6) (1), by reaction of tz and [Cp*2Tb][BPh4] in the presence of KC8 in benzene (C6H6). (B) Molecular structure (side and top views).

The magnetic behaviour of 1 was initially probed through direct current (dc) magnetic susceptibility measurements at 1000 Oe and between 1.8–300 K (Fig. S6, ESI†). The room temperature χT product of 48.75 cm3 K mol−1 is in good agreement with the theoretical value of 48.76 cm3 K mol−1 for four TbIII ions (S = 3, L = 3, 7F6, g = 3/2, C = 11.82 cm3 K mol−1) and four radical species (S = 1/2, C = 0.37 cm3 K mol−1). As the temperature decreases, the χT product gradually decreases until ∼170 K, reaching a shallow minimum of 47.55 cm3 K mol−1. Bellow this temperature, a rapid increase of the χT value is observed until it reaches a maximum of 120.96 cm3 K mol−1 at 13 K. As in the case of the previously published tz˙−-bridged “Ln2” and “Ln4” analogues, this rise in χT is attributed to the spin alignment of the TbIII ions caused by the strong antiferromagnetic Ln–rad coupling, which gives rise to a high angular momentum, “giant spin” state. The strength of this Ln–rad coupling was previously quantified in the “Ln4” analogues.14 Due to the presence of two crystallographically independent LnIII ions, we were able to probe two exchange parameters JLn–rad for the “Gd4” (J1 = −24 cm−1; J2 = −15 cm−1) and “Dy4” (J1 = −21 cm−1; J2 = −27 cm−1). Given that the χT vs. T plot of 1 shows similar trends it is expected that the respective exchange parameters JTb–rad are of comparable strength. In the low temperature region, the rapid drop of the χT product suggests the pinning of the orientation of the total molecular spin by the strong magnetic anisotropy. This is further evident by the zero-field-cooled/field-cooled (ZFC/FC) magnetic susceptibility data where a clear divergence of the two data sets below 5.4 K is observed (Fig. S7, ESI†). This, in addition to the s-shape of the magnetization curves at low temperatures (1.9 and 3 K; Fig. S8, ESI†), sparked our interest in measuring the hysteresis of 1. As such, hysteresis measurements were performed between 70 to −70 kOe in the 1.8 to 5.2 K temperature region using an average sweep rate of 25 Oe s−1 (Fig. S9, ESI†). At 1.8 K and H = 0 Oe, the hysteresis loop is closed but opens upon increasing the magnetic field. The butterfly-shaped hysteresis loops remain open until 5.2 K, above which they are no longer open. This trend agrees with the observed trends from the ZFC/FC measurements and is slightly lower than the blocking temperature of the “Dy4” analogue (6 K). However, the non-retention of the magnetic moment when nearing 0 Oe, is indicative of QTM that leads to the absence of coercivity in 1 in contrast to the “Dy4” analogue. Given the structural similarities between the Tb and Dy-congeners, it is reasonable to attribute these differences to the intricate electronic structures of the individual LnIII ions.

To explore the relaxation dynamics of 1, alternating current (ac) susceptibility measurements within the 0.1–1488 Hz frequency range were performed. In the absence of a static dc field (Hdc = 0 Oe), a single peak of the in-phase (χ′) and out-of-phase (χ′′) (Fig. S10, ESI†) susceptibilities was observed between 9 to 1.8 K. More specifically, the χ′′ peaks showed an exponential increase from 6 to 9 K suggesting the presence of a temperature-dependent relaxation mechanism, while bellow 6 K the peaks started to overlap suggesting that QTM is dominating this temperature region. Fitting of the ac susceptibilities with CCFit-215 using a generalized Debye model (Fig. 2(A)), the relaxation times (τ) of the magnetization were extracted (Table S4, ESI†). Subsequently, the τ−1vs. T plot was fitted to a combination of QTM and Orbach relaxation processes, as described by eqn (S1) (ESI†), yielding the following best-fit parameters: τQTM = 0.41 s, τ0 = 1.67 × 10−10 s and Ueff = 87.8 cm−1 (Fig. 2(B)). The observed Ueff value is slightly smaller than the respective Ueff value for the “Dy4” (Ueff1 = 91 cm−1; Ueff2 = 80 cm−1). This is expected given the presence of QTM which acts competitively with the thermally activated mechanism (Orbach). Even though the presence of the radical should lead to the effective suppression of the QTM, it is apparent that this is not the case for 1.


image file: d3cc03034b-f2.tif
Fig. 2 Top: Cole–Cole plots of 1 (Hdc = 0 Oe; (A) and Hdc = 1200 Oe; (C)). Fitting parameters are summarized in Tables S4 and S6 (ESI†), respectively. Bottom: Temperature-dependence of the relaxation times (τ) for 1 (Hdc = 0 Oe; (B) and Hdc = 1200 Oe; (D)), with the respective estimated standard deviations (gray bars). The estimated standard deviations of the τ were estimated from the α-parameters of the generalized Debye fits and the log-normal distribution.15 The solid red lines represent the best-fit based on eqn (S1) (ESI†) for (B) and eqn (S3) (ESI†) for (D), while the dashed orange and magenta lines in (B) represent the individual components of the magnetic relaxation for QTM and Orbach process, respectively.

In an effort to eliminate the presence of QTM in the magnetic relaxation of 1, the effect of the applied static field was probed. Ac susceptibility measurements were undertaken at various static fields (0–3000 Oe) at a constant temperature of 6 K, where a field-dependent signal of the χ′′ of 1 was observed (Fig. S11, ESI†). The extracted τ (Table S5, ESI†) showed a field-dependent increase with the increase of the applied static field between 0 to 1200 Oe (Fig. S12, ESI†). At higher fields, the relaxation times become field-independent indicating that QTM has been successfully suppressed and the thermally activated process (i.e. Orbach) is mediating the relaxation of the magnetization. An excellent fit of the field-dependent τ was achieved for a combination of QTM and Orbach processes using eqn (S2) (ESI†) with the following best-fit parameters: B1 = 0.25 s, B2 = 1.94 × 10−5 Oe−2, τ0 = 1.84 × 10−10 s and Ueff = 88.0 cm−1 (Fig. S12, ESI†).

Hence, the 1200 Oe was selected as the optimal field where ac susceptibility measurements were conducted between 9 and 5.8 K, where frequency-dependent χ′ and χ′′ (Fig. S13, ESI†) signals were observed. Fitting of the ac susceptibility (Fig. 2(C)), yielded longer τ (Table S6, ESI†). As expected, the peaks of the susceptibility showed only an exponential increase with the decrease in temperature, indicative of an Orbach-only process. As such, the extracted τ were fitted, using eqn (S3) (ESI†), yielding the following best-fit parameters: τ0 = 1.27 × 10−10 s and Ueff = 89.4 cm−1 (Fig. 2(D)). The presence of an Orbach-only mechanism for the magnetic relaxation is also evident upon examining the Arrhenius plot of the ln(τ) vs. T−1, where a linear trend of the τ was observed, as expected (Fig. S14, ESI†).

To validate the observed magnetic behavior, the local electronic structure of the TbIII ions in 1 was calculated using a previously established approach.14 The calculations utilize multireference ab initio calculations at the NEVPT2//SA-CASSCF level,16,17 with spin orbit coupling treated with the quasi-degenerate perturbation theory18 as implemented in the Orca software version 5.0.4 (see ESI† for further details).19 Due to the non-Kramers nature of the TbIII ion, the local electronic structure consists of quasi-doublets and a single singlet (Tables S9 and S10, ESI†). For both Tb1 and Tb2, the ground quasi-doublets are axial with small but non-negligible tunnelling gaps 0.10 cm−1 and 0.05 cm−1, respectively. The axiality decreases in the higher doublets, and in case of both ions at the first and second excited doublets there is a significant tunnelling gap leading to efficient transitions through the local Ueff. The energy difference between the ground quasi-doublets and the first excited doublets are 47 cm−1 and 78 cm−1 for Tb1 and Tb2, respectively (ESI†). The next doublets lie at 162 cm−1 and 199 cm−1. The effective barrier height fitted for the Orbach process is somewhat higher than the local first excitation energies at the local doublets, and lower than the second excitation energies. This means that the most likely relaxation mechanism takes place via an exchange-coupled giant spin state as in the case of “Dy4”.14 It should be noted that the excitation energy of 78 cm−1 is close to the fitted energy barrier; however, if the relaxation would take place via the local excited states of the individual ions, there should also be a relaxation process that corresponds to a barrier of 47 cm−1, and such a process was not observed.

In addition to the magnetic studies, we also probed the magneto-optical properties of 1via magnetic circular dichroism (MCD). Although for “Dy4” we were not able to visually separate the radical and metal contributions on the MCD spectrum,14 interestingly, this is not the case for 1. MCD spectra obtained at 1.9 K and various applied magnetic fields are shown in Fig. 3(A) and Fig. S16 (ESI†).


image file: d3cc03034b-f3.tif
Fig. 3 (A) MCD spectra of 1 obtained at 1.9 K at different applied magnetic fields. Blue shadowed areas indicate the peak position for the magnetization measurements. See Fig. S16(A) (ESI†) for the spectra in wavenumber scale. (B) Experimental MCD spectrum obtained at 1.9 K with a 1 T applied magnetic field and the resultant fit for the deconvoluted peaks (see Fig. S16(B), ESI† for individual peaks). (C) Normalized ΔA and magnetization plots obtained by MCD and SQUID instruments, respectively. For the MCD data, we selected two different absorptions: the first, at 465 nm, attributed only to the radical and the second, at 486 nm, attributed to radical and TbIII contributions. Black vertical lines in (A) and (B) indicate the position of TbIII excited states, according to Carnall (Table S11, ESI†).21

An intriguing feature of the MCD spectrum for 1 is the presence of a broad band related to the organic ligands spanning all over the ultraviolet and visible range of the electromagnetic spectrum, and additionally, the presence of absorption peaks and dips at wavelengths where most of the TbIII excited states are located (see the vertical black lines in Fig. 3(A) and Table S11, ESI†). It should be noted that all the spectral features arise from a combination of the radical and TbIII 7F6 coupled ground levels. Deconvolution of the broad band in four components (Fig. 3(B) and Fig. S16(B), ESI†) reveals the central position of the ligand-related absorptions to be located at approximately 305, 387, 465, and 573 nm. The presence of TbIII-related absorption peaks and dips allows us to investigate the magneto-optical profile in transitions from the coupled ground states to excited states containing only the radical contribution (465 nm) and in a band where both radical and TbIII contributions are present (486 nm, radical and TbIII 5D4 excited state). Interestingly, when monitoring the transition to the radical excited state, the differential absorption (ΔA) as a function of the applied magnetic field drastically disagrees with the SQUID magnetization (Fig. 3(C)). On the other hand, the band at 486 nm, containing contributions from both radical and metal excited states, follows a similar trend to the magnetic behavior for 1 as observed in the data collected via SQUID magnetometry at 1.9 K. As previously reported,20 the field dependence for ΔA is related to the polarization of the electronic transition. Based on our observations, the absorption component connecting the radical and TbIII 7F6 ground states to the radical and TbIII 5D4 excited states can reflect the bulk magnetic behavior, as observed in the SQUID magnetometry.

The presented MCD study allows us to add another piece to the magneto-optical puzzle of radical-bridged Ln complexes. Although we advanced in understanding the different behavior when monitoring a band with only the radical or with both radical and metal excited state contributions, unfortunately for TbIII ions there is no excited state outside of the radical contribution. Nevertheless, this study demonstrates the clear advantage offered by MCD in monitoring absorption bands originating from strongly coupled systems. However, to understand further the metal-radical coupling through MCD spectroscopy, additional studies with a comprehensive family of molecular lanthanide species that contain radical bridges will be required.

Overall, the synthesis and characterization of the first polynuclear radical-bridged TbIII metallocene complex [(Cp*2Tb)4(tz˙−)4]·3C6H6 (1) was achieved. It exhibits similar magnetic behaviour, i.e., strong magnetic Ln–rad coupling and slow relaxation of the magnetization in the absence of a static dc field (Hdc = 0 Oe) similarly to “Dy4”. The presence of QTM prevents the opening of the hysteresis loop which are significantly impacted and exhibit no coercivity, a vast difference from the “Dy4” congener. However, magneto-optical studies via MCD reveal both radical and TbIII excited state contributions in the absorption spectra in contrast to the “Dy4”. These results provide further insights into how the electronic differences of the individual Ln ions can impact the overall magnetic and magneto-optical properties of radical-bridged polynuclear SMMs.

M. M. acknowledges the Canadian Foundation for Innovation (CFI) and the Natural Sciences and Engineering Research Council of Canada (NSERC) for their financial support. N. M. acknowledges the Stavros Niarchos Foundation for financial support through scholarships. A. M. acknowledges funding provided by the Academy of Finland (grant no. 332294) and the University of Oulu (Kvantum Institute). Computational resources were provided by CSC-IT Center for Science in Finland and the Finnish Grid and Cloud Infrastructure (persistent identifier urn:nbn:fi:research-infras-2016072533).

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. C. A. P. Goodwin, F. Ortu, D. Reta, N. F. Chilton and D. P. Mills, Nature, 2017, 548, 439–442 CrossRef CAS PubMed .
  2. F.-S. Guo, B. M. Day, Y.-C. Chen, M.-L. Tong, A. Mansikkamäki and R. A. Layfield, Science, 2018, 362, 1400–1403 CrossRef CAS PubMed .
  3. E. Coronado, Nat. Rev. Mater., 2020, 5, 87–104 CrossRef .
  4. M. Brzozowska, G. Handzlik, K. Kurpiewska, M. Zychowicz and D. Pinkowicz, Inorg. Chem. Front., 2021, 8, 2817–2828 RSC .
  5. C. D. Buch, S. H. Hansen, C. M. Tram, D. Mitcov and S. Piligkos, Inorg. Chem., 2020, 59, 16328–16340 CrossRef CAS PubMed .
  6. N. Ishikawa, M. Sugita, T. Ishikawa, S. Koshihara and Y. Kaizu, J. Am. Chem. Soc., 2003, 125, 8694–8695 CrossRef CAS PubMed .
  7. K. Kizaki, H. Ozawa, T. Kobayashi, R. Matsuoka, Y. Sakaguchi, A. Fuyuhiro, T. Fukuda and N. Ishikawa, Chem. Commun., 2017, 53, 6168–6171 RSC .
  8. C. A. Gould, K. R. McClain, D. Reta, J. G. C. Kragskow, D. A. Marchiori, E. Lachman, E.-S. Choi, J. G. Analytis, R. D. Britt, N. F. Chilton, B. G. Harvey and J. R. Long, Science, 2022, 375, 198–202 CrossRef CAS PubMed .
  9. M. He, F.-S. Guo, J. Tang, A. Mansikkamäki and R. A. Layfield, Chem. Commun., 2021, 57, 6396–6399 RSC .
  10. S. Calancea, L. Carrella, T. Mocanu, V. Sadohin, M. Raduca, I. Gutu, J. C. da Rocha, M. G. F. Vaz, E. Rentschler and M. Andruh, Eur. J. Inorg. Chem., 2021, 567–577 CrossRef CAS .
  11. G. T. Nguyen and L. Ungur, Phys. Chem. Chem. Phys., 2021, 23, 10303–10310 RSC .
  12. O. Stetsiuk, A. Abhervé and N. Avarvari, Dalton Trans., 2020, 49, 5759–5777 RSC .
  13. N. Mavragani, A. A. Kitos, A. Mansikkamäki and M. Murugesu, Inorg. Chem. Front., 2022, 10, 259–266 RSC .
  14. N. Mavragani, D. Errulat, D. A. Gálico, A. A. Kitos, A. Mansikkamäki and M. Murugesu, Angew. Chem., Int. Ed., 2021, 60, 24206–24213 CrossRef CAS PubMed .
  15. D. Reta and N. F. Chilton, Phys. Chem. Chem. Phys., 2019, 21, 23567–23575 RSC .
  16. (a) B. O. Roos in Advances in Chemical Physics, Ab Initio Methods in Quantum Chemistry II, ed. K. P. Lawley, Wiley, New York, 1987, vol. 69, pp. 399–455 Search PubMed ; (b) P. Siegbahn, A. Heiberg, B. Roos and B. Levy, Phys. Scr., 1980, 21, 323–327 CrossRef CAS ; (c) B. O. Roos, P. R. Taylor and P. E. M. Siegbahn, Chem. Phys., 1980, 48, 157–173 CrossRef CAS ; (d) P. E. M. Siegbahn, J. Almlöf, A. Heiberg and B. Roos, J. Chem. Phys., 1981, 74, 2384–2396 CrossRef ; (e) B. O. Roos, R. Lindh, P. Å. Malmqvist, V. Veryazov and P.-O. Widmark, Multiconfigurational Quantum Chemistry, Wiley, Hoboken, NJ, 2016 CrossRef .
  17. (a) C. Angeli, R. Cimiraglia, S. Evangelisti, T. Leininger and J.-P. Malrieu, J. Chem. Phys., 2001, 114, 10252–10264 CrossRef CAS ; (b) C. Angeli, R. Cimiraglia and J.-P. Malrieu, Chem. Phys. Lett., 2001, 350, 297–305 CrossRef CAS ; (c) C. Angeli, R. Cimiraglia and J.-P. Malrieu, J. Chem. Phys., 2002, 117, 9138–9153 CrossRef CAS .
  18. (a) F. Neese, T. Petrenko, D. Ganyushin and G. Olbrich, Coord. Chem. Rev., 2007, 251, 288–327 CrossRef CAS ; (b) M. Atanasov, D. Aravena, E. Suturina, E. Bill, D. Maganas and F. Neese, Coord. Chem. Rev., 2015, 289–290, 177–214 CrossRef CAS .
  19. (a) F. Neese, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2017, 8, e1327 Search PubMed ; (b) F. Neese, F. Wennmohs, U. Becker and C. Riplinger, J. Chem. Phys., 2020, 152, 224108 CrossRef CAS PubMed ; (c) F. Neese, J. Comput. Chem., 2022, 44, 381–396 CrossRef PubMed .
  20. D. Errulat, K. L. M. Harriman, D. A. Gálico, J. S. Ovens, A. Mansikkamäki and M. Murugesu, Inorg. Chem. Front., 2021, 8, 5076–5085 RSC .
  21. W. T. Carnall, H. Crosswhite and H. M. Crosswhite, Energy level structure and transition probabilities in the spectra of the trivalent lanthanides in LaF3, U.S. Department of Energy-Office of Scientific and Technical Information, United states, N. p., 1978 Search PubMed .

Footnote

† Electronic supplementary information (ESI) available: Synthetic procedures, single-crystal X-ray diffraction data, computational details, additional magnetic and MCD data. CCDC 2271181. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3cc03034b
This journal is © The Royal Society of Chemistry 2023
Click here to see how this site uses Cookies. View our privacy policy here.