Ming
Liu
a,
Xiao-Han
Peng
a,
Fu-Sheng
Guo
*a and
Ming-Liang
Tong
*b
aInstitute of Fundamental and Frontier Sciences, University of Electronic Science and Technology of China, Xiyuan Avenue 2006, Chengdu, 611731, China. E-mail: fu-sheng.guo@uestc.edu.cn
bKey Laboratory of Bioinorganic and Synthetic Chemistry of Ministry of Education, School of Chemistry, Sun Yat-Sen University, Guangzhou 510006, China. E-mail: tongml@mail.sysu.edu.cn
First published on 17th May 2023
In recent years, very significant progress has been made in single-molecule magnets (SMMs), and one of the major milestone works is the hysteresis blocking temperature beyond the boiling point of liquid nitrogen. As increasingly abundant experimental and theoretical cases are studied, our understanding of how to construct high-performance rare-earth SMMs is becoming clearer, while few good SMMs have been reported for actinides that possess stronger spin–orbital coupling than rare earths. Recently, attempts to replicate the successful strategy on rare-earth SMMs to actinides have proven to be a failure, or at least frustrating, with the main reason being the stronger covalent contribution in the actinide-ligand bonds. In this review, the progress on actinide SMMs is summarized to look back at how far we have come and try to find possible routes forward.
The requirements for high effective energy barrier (Ueff) and magnetic hysteresis blocking temperature (TB) are two crucial keys to achieving the above attractive applications, which remain the unusually heaviest challenges in this area. TB is closely going hand in hand with the Ueff determined by the uniaxial magnetic anisotropy and relaxing through higher excited states. To improve the value of Ueff, the efforts initially focused on the giant transition metal spin-systems providing massive magnetic centers with strong magnetic coupling, which, however, usually ensued with a severe decline in anisotropy itself, and was not very favourable for SMM behaviours.9
An alternative method to maximize the anisotropy, instead of the extreme pursuit of a larger number of spin ground states, came in the subsequent period. Lanthanide ions with large magnetic moments and high magnetic anisotropy, especially Tb(III), Dy(III), Ho(III), and Er(III), are very excellent spin carriers for the synthesis of SMMs.10–13 After the mononuclear compound, [TbPc2]− (Pc = phthalocyanine dianion), with a high energy barrier of 230 cm−1 was reported by Ishikawa et al.10 a large number of lanthanide-based complexes have been studied, and they are the research mainstay in the field of SMMs at present. Moreover, along with the continuous improvement of theoretical research, the understanding of how to construct lanthanide SMMs with large anisotropy barriers and blocking temperatures is becoming clearer, and several symmetry strategies based on an effective charge model, such as Cn (n ≥ 7), S8/D4d, C5h/D5h and S12/D6d are proposed or have been proved to be helpful.14
Among the lanthanide ions, Dy(III) is the most studied and brightest performer. Dysprosium possessing a 4f9 configuration is a Kramers ion and therefore ensures a bistable ground state regardless of the coordination environment. The oblate electron density of Dy(III) provides guidance to chemists on how to design molecular structures: creating a strong crystal field in the axial direction and minimizing the crystal field in the equatorial plane reduces the electron repulsion between the dysprosium ion and the ligand, allowing the ground state to stabilize at a higher MJ value. In terms of a dysprosium complex, although strict point symmetry is not essential for observing slow magnetic relaxation, the closer the coordination geometry is to ideal point symmetry, the better it behaves as a single-molecule magnet. The synthesis of low-coordinate complexes with perfect point symmetry is very challenging, mainly due to the large ionic radii and odd valence states of lanthanide ions. One of the most striking subsets of the SMM family is dysprosium metallocenes, in which the cyclopentadiene ligand generates a dominant axial crystal field. Recently, some of us reported a heteroligand dysprosium metallocene complex, [CpiPr5DyCp*][B(C6F5)4] (CpiPr5 = pentaisopropylcyclopentadienyl; Cp* = pentamethylcyclopentadienyl), which has a more wider Cp–Dy–Cp angle and shorter Dy–Cp distances.11 The strong axial crystal field results in a large Ueff of 1541 cm−1 and a high blocking temperature of 80 K, making it the first SMM to show open magnetic hysteresis above the boiling point of liquid nitrogen.
Compared to lanthanide complexes, actinides appear as potential candidates for the construction of outstanding SMMs because of their significantly stronger spin–orbit coupling, which can result in greater magnetic anisotropy.15 The relatively more diffuse radial extension of 5f orbitals can lead to a greater degree of metal–ligand covalency, contributing to strong magnetic exchange between spin carriers in polymetallic complexes.16 It can be said that actinides combine the advantages of both lanthanides and transition metals. Furthermore, the accessible oxidation states in actinides ranging from +7 to +1 are potentially promising characteristics to design and synthesize various electronic structures.17 To date, however, the actinides have been far inferior to the lanthanides and transition metals in terms of SMMs, with only a small fraction exhibiting slow magnetic relaxation and mainly concentrated in uranium complexes. Actinide SMMs often exhibit remarkably small reversal barriers and no magnetic hysteresis.18,19 Because of the rapid development and excellent performance of transition metals or lanthanides in high-performance SMMs in recent years, some similar successful strategies were attempted to replicate in actinides, but the results so far have been very frustrating, with actinides exhibiting complexities not encountered with transition metals and lanthanides. The main reason for such phenomena is generally traced to the covalent contribution in the actinide–ligand bonds, leading to the occurrence of rapid relaxation progress. Therefore, understanding the relaxation mechanism and exploring their rational strategies in actinide SMMs are highly critical to further progress, yet they remain a big challenge at the current stage.
In this review, we aim to present a survey of representative actinide SMMs to look back at how far we have come. Meanwhile, we pay attention to the influence of the ligand field and magnetic exchange of paramagnetic centres on the magnetic dynamics, trying to seek out a feasible strategy for top-performing An-SMMs. The survey consists of two parts, mononuclear actinide SMMs, and multinuclear An-SMMs with a magnetic exchange. Due to the radioactivity of the actinides, only a handful of laboratories are qualified to handle and study them, and most of them are concentrated in uranium. Transuranic elements have more restrictions because of their scarcity and instability, so there are even fewer cases of SMMs, and in the section of mononuclear actinide SMMs, transuranic elements are listed with uranium, classified by the unpaired electron number.
Classification | Complex | Local symmetry | U eff (cm−1) | H dc (Oe) | τ 0 (s) | Ref. | |
---|---|---|---|---|---|---|---|
Cal. | Exp. | ||||||
a The dc field used in ac susceptibility measurements. | |||||||
3 unpaired electrons | U(Ph2BPz2)3 | D 3h | 190 | 20 | 0 | 1.0 × 10−7 | 21 |
U(H2BPz2)3 | D 3h | 230 | 8 | 100 | 1.2 × 10−6 | 22 | |
U(BcMe)3 | C 3h | — | 33 | 750 | 1.0 × 10−7 | 24 | |
[K(18-crown-6)] [U(OSi(OtBu)3)4] | — | — | 18 | 0 | 2.6 × 10−7 | 25 | |
[K(18-crown-6)] [U(N(SiMe3)2)4] | — | — | 16 | 0 | 2.2 × 10−8 | 25 | |
(BDI)U(ODipp)2 | C 2v | — | 15 | 500 | 3.4 × 10−7 | 26 | |
[(Tp*2U)-CCPh] | C 2 | — | 7 | 1000 | 2.1 × 10−6 | 27 | |
[(Tp*2U)(THF)] | C 2 | — | 9 | 500 | 4.3 × 10−6 | 27 | |
[(Tp*2U)(MeCN)2] | C 2 | — | 8 | 1000 | 4.0 × 10−6 | 27 | |
[U({SiMe2NPh}3-tacn)(OPPh3)] | C 3v | 147 | 15 | 0 | 1.1 × 10−7 | 29 | |
[U((SArAd,Me)3mes)] | C 3v | — | 10 | 2000 | 2.8 × 10−6 | 30 | |
[UI3(THF)4] | C 2v | — | 13 | 2000 | 6.0 × 10−6 | 31 | |
[U(N(SiMe3)2)3] | C 3v | — | 22 | 2000 | 1.5 × 10−6 | 31 | |
[U(BIPMTMS)(I)2(THF)] | Cs/C1 | — | 16 | 2000 | 2.8 × 10−6 | 31 | |
[U(N(SiMe2tBu)2)3] | D 3h | — | 15 | 600 | 3.1 × 10−11 | 32 | |
[Li(DME)3] [UIII(COT′′)2] | C 2v | — | 19 | 1000 | 4.6 × 10−6 | 34 | |
[Np(η8-COT)2] | D 8h | 1400 | 29 | 300 | 1.1 × 10−5 | 39 | |
4 unpaired electrons | [(LAr)NpCl] | C 2v | — | — | 1000 | — | 40 |
5 unpaired electrons | [PuTp3] | C 3h | 332 | 18 | 100 | 2.9 × 10−7 | 41 |
[K(2.2.2-cryptand)] [(η5-C5iPr5)2U] | C 2v | — | 14 | 0 | 1.1 × 10−7 | 17 | |
1 unpaired electron | [U(O)-(TrenTIPS)] | C 3v | — | 15 | 1000 | 2.6 × 10−7 | 42 |
Replacing the phenyl group with a hydrogen atom produced a closely related derivative [U(H2BPz2)3] (2) (Fig. 1), which showed the same structure symmetry as that of complex 1 but a subtle elongation was observed along the direction of pseudo-C3 axis.22 Compared with the isostructural complex 1, a much smaller energy barrier of 8 cm−1 (τ0 = 1.2 × 10−6 s) under the applied dc field was quantified for the analogue 2, which is likely attributed to the axial extension. Subsequently, the impact of magnetic dilution on the relaxation behaviour in 2 with different concentrations was further elucidated by Long and co-workers.23 Dilution-induced SMM behaviour in the highest dilution (U:Y = 1:90) was presented, yielding a thermal-activated barrier Ueff = 16 cm−1, two times larger than that of the previously concentrated sample. Intriguingly, computational studies of 1 and 2 revealed that the values of the first excited states are 190 and 230 cm−1, respectively, which significantly exceed the Ueff determined from experiments.18
Fig. 1 The crystal structure of [U(H2BPz2)3]. U atom is drawn in yellow, B in orange, N in blue, and C in grey; H atoms are omitted for clarity except for those bound to B atoms. |
Donor atoms may play crucial roles in determining the magneto-dynamic characteristics of uranium-based complexes because of the discrepancies in the strength of the ligand field effect and uranium-ligand covalency. The long group compared the behaviour of SMM in two isostructural complexes, [U(BpMe)3] (3a) (BpMe = dihydrobis(methylpyrazolyl)borate) and [U(BcMe)3] (3b) (BcMe = dihydrobis(methylimidazolyl)borate), finding that the stronger ligand field of N-heterocyclic carbene in complex 3b gives rise to greater covalency and anisotropy, exhibiting slower relaxation than complex 3a.24 The 12% magnetically dilute version of 3b yielded a large barrier of 33 cm−1 with τ0 = 1 × 10−7 under a field of 750 Oe, dominated by an Orbach relaxation in the high-temperature range.
In the same year, two novel pseudotetrahedral complexes, [K(18-crown-6)][U(N(SiMe3)2)4] (4a) and [K(18-crown-6)][U(OSi(OtBu)3)4] (4b) with the similar crystal structure but distinct ligands surroundings were reported by Mazzanti and co-workers.25 Both demonstrated slow magnetic relaxation already at zero dc fields but exceedingly similar values of anisotropy barriers (16 cm−1 for 4a and 18 cm−1 for 4b). The insignificant differences were attributed to relatively slight variations in ligand field strength. Recently, another tetrahedral uranium(III) complex, (BDI)U(ODipp)2 (5) (BDI = β-diketiminate, ODipp = 2,6-diisopropylphenolate), was reported by Arnold et al.26 possessing a coordination environment between 4a and 4b. The complex showed obvious SMM behaviour and displayed well-defined peaks from 1.8 K to 4 K in the out-of-phase alternating current susceptibility (χ′′) under an applied 500 Oe field, with associated Ueff = 14.5(1) cm−1 (τ0 = 3.39(24) × 10−7 s), and the thermal barrier is similar to that of 4a and 4b.
Similarly, the insensitivity of dynamic magnetic performance to subtle changes in the ligand field is supported by a homologous series of mononuclear species, [Tp*2UCCPh] (6a), [Tp*2U(THF)](BPh4) (6b) and [Tp*2U(MeCN)2](BPh4) (6c), reported by the Shores group in 2019.27 All complexes show field-induced SMM behaviours, with very small and similar energy barriers of 6.8 cm−1 (τ0 = 2.1 × 10−6 s), 9.0 cm−1 (τ0 = 4.3 × 10−6 s) and 8.4 cm−1 (τ0 = 4.0 × 10−6 s), respectively.
Generally, slow relaxation is common for U(III) complexes, however, a rare example, [U({SiMe2NPh}3-tacn)] (7a) (tacn = 1,4,7-triazacyclononane) with an equatorial charge distribution around the metal center, does not exhibit this kind of behaviour down to 1.7 K.28,29 Then, the structure of complex 7a was modified by coordination of an axial OPPh3 ligand, isolating complex [U({SiMe2NPh}3-tacn)(OPPh3)] (7b) (Fig. 2). Upon the modification, the bond lengths of U–N(amido) and U–N(amino) are longer about 0.1 Å than those in 7a.29 The U–O–P angle of 176.94(14)° is almost linear. Slow magnetic relaxation below 5 K was observed for complex 7b, with peaks in χ′ (in-phase alternating current susceptibility) and χ′′ even in zero applied fields. The energy barriers were obtained by fitting the relaxation time in zero and 2 kOe static field to the Arrhenius law, resulting in 15.2(4) cm−1 and 17.4(2) cm−1, respectively. Although the differences between these two complexes were intuitively explained by that the axial OPPh3 could stabilize large MJ states, the authors also suggested the different behaviours cannot simply be attributed to differences in crystal field.
Fig. 2 The crystal structure of [U({SiMe2NPh}3-tacn)(OPPh3)]. Color codes: U, yellow; Si, pale yellow; O, red; C, grey; N, blue; P, purple. H atoms are omitted for clarity. |
Hard-donor ligands such as alkoxides, aryloxides, and amides have been largely introduced into uranium complexes, while examples of soft ligands involved are very rare. Recently, the Meyer group reported one uranium thiolate complex [U((SArAd,Me)3mes)] (8), with the ligands designed and synthesized through extremely tedious steps.30 Complex 8 crystallizes in the space group of P, and the uranium is located on the C3 symmetry axis (Fig. 3). Compared to its aryloxide analogues, [U((OArAd,Me)3mes)] and [U((OartBu,tBu)3mes)], the U–S interaction is weaker, but a larger covalency contribution was found. An applied external field of 2 kOe allowed observation of the peaks of the out-of-phase components at the low-temperature range, resulting in a thermal barrier of Ueff = 9.7(7) cm−1 with τ0 = 2.8(8) × 10−6 s, as extracted from the fit of the relaxation times to the Arrhenius law below 3.6 K. A significant deviation from linearity in the ln(τ) vs. T−1 plot can be observed, which might involve additional relaxation mechanisms, such as QTM, direct relaxation, and Raman processes.
Fig. 3 The crystal structure of [U((SArAd,Me)3mes)]. Color codes: U, yellow; S, pink; C, grey. H atoms are omitted for clarity. |
The previous examples show that modulating the coordination environment by replacing substituents or replacing ligands with different crystal field strengths, does affect the SMM properties. However, the energy barriers or blocking temperatures do not change by orders of magnitude, for example, the energy barriers are always around 20 cm−1 (Table 1). It is also noteworthy that the magnetic relaxation is closer to Raman rather than the Orbach process, therefore to some extent, it makes no sense to simply calculate and analyse the so-called energy barriers.
In 2013, Liddle and Slageren et al. reported in detail three completely unrelated U(III)-based complexes, [UI3(THF)4] (9a); [U(N(SiMe3)2)3] (9b), and [U(BIPMTMS)(I)2(THF)] (9c) [BIPMTMS = CH(PPh2NSiMe3)2] with different ligands and symmetries (C2v, C3v, and Cs/C1 respectively).31 All complexes 9a–c are SMMs, with relatively similar and small Ueff of 12.9, 22.1, and 16.2 cm−1, respectively. Their magnetic bahaviours are comparable with other reported uranium(III) SMMs. This study demonstrates that the character of SMMs is intrinsic to uranium(III), but the ligand types or local symmetries have limited roles in the slow relaxation behaviour.
Subsequently, a bulkier amide ligand, N(SiMe2tBu)2− (N**), was utilized by Mills and co-workers to prepare an actinide complex [U(N**)3] (10) with a trigonal planar geometry.32 Compared to complex 9b, [U(N**)3] is closer to the ideal trigonal symmetry. An energy barrier of 14.9 cm−1 was obtained by the Arrhenius fit to the higher temperature ac data. The authors attribute the reason for the lower value compared to 9b to the selection of the temperature range. However, the relaxation time of 2.6 ms at 2 K is longer than that (0.3 ms) of 9b.
Cyclooctatetraenyl (COT) ligands have been very successful in the construction of lanthanide SMMs, with Er(III) analogues exhibiting high energy barriers and magnetic blocking temperatures.33 To explore the influence of the sandwich-type ligand field on the magnetic relaxation of uranium, in 2015, the Murugesu group reported a U(III) sandwich complex, [Li(DME)3][U(COT′′)2] (11) (COT′′ = bis(trimethylsilyl)cyclooctatetraenyl dianion) (Fig. 4),34 in which the U–C bond distances range from 2.726(0) to 2.755(1) A, and the Ring-U-Ring angle deviates from perfect linearity by 7.2°. Ac susceptibility measurements were performed on the uranium metallocene, showing that it is a field-induced SMM with a magnetic reversal barrier of 27 K and τ0 = 4.6 × 10−6 s. Although its energy barrier is comparable to that of other uranium SMMs, it still lags far behind the Er analogues. The authors suggest that the covalent bonding between uranium and the COT′′ ligands could allow these mononuclear motifs to be ideal for the construction of high-performance multinuclear SMMs, even though a cyclooctatetraenyl ligand creating an equatorial crystal field does not suit well with an oblate U(III).
Fig. 4 The crystal structure of [U(COT′′)2]−. Color codes: U, yellow; Si, light yellow; C, grey. H atoms are omitted for clarity. |
In contrast to a 10 π-electron COT ligand possessing a large diameter, another aromatic ligand, cyclopentadiene with a compact size could provide a strong axial ligand field. Such ligands can stabilize the ground state of oblate dysprosium or terbium at higher MJ values, and a series of outstanding metallocene SMMs have been reported,10–14 of which the most representative is [CpiPr5DyCp*]+,11 the first one breaking through the nitrogen ceiling in hysteresis temperature. There are similarities between uranium(III) and dysprosium(III), for example, ionic radii, and their oblate shape of electron densities, so these drive some groups to replicate the successful strategy of dysprosium to uranium. In 2019, some of us reported the first base-free actinide bis-Cp sandwich complex, [(CpiPr5)2U][B(C6F5)4] (12a) (Fig. 5), which was isolated by using the super-electrophile of [(Et3Si)2(μ-H)][B(C6F5)4] to abstract iodide from the precursor [(η5-CpiPr5)2UI] (12b).35 The two structures are very different, mainly in that the cations have shorter U–Cp bond lengths and wider Cp–U–Cp angles, which is due to the absence of coordination on the equatorial plane. The U–Cp distance in 12a is 2.472(3) Å, slightly longer than that (2.296(1) and 2.284(1) Å) in [CpiPr5DyCp*]+, and the Cp–U–Cp angle of 167.82(8)° is more approaching to linearity than that (162.507(1)°) in dysprosium cation. For both complexes, well-defined peaks in the ac susceptibility plot were only observed upon the application of optimized fields. The magnetic behaviours of these two complexes are quite similar, and there is no performance jump of several orders of magnitude like that between [CpiPr5DyCp*]+ and its precursor.11 The absence of slow relaxation in zero applied fields is attributed to the non-negligible metal–ligand covalency and partial quenching of anisotropy resulting from strong ligand field splitting, which can lead to efficient quantum tunnelling of the magnetization. The temperature dependence of the relaxation time (τ) in the field fits the Raman expression well, indicating that Orbach relaxation processes commonly observed in dysprosium metallocene SMMs are not obvious in these uranium analogues. Therefore, the case of the actinides is far more complicated than that of the lanthanides, and the successful strategy of the lanthanides does not apply to the actinides, at least in unsymmetric uranocenium complexes.
Fig. 5 The crystal structure of [(η5-CpiPr5)2U][B(C6F5)4]. Color codes: U, yellow; F, green; B, deep green; C, grey. H atoms omitted for clarity. |
After the uranocenium work, a perfectly linear uranium(II) metallocene, namely [(η5-CpiPr5)2U] (13), was isolated by the reduction of [(η5-CpiPr5)2UI] with potassium graphite.36 In the uranocene, although the uranium(II) with a 5d36f1 electron configuration has four unpaired electrons and should be placed in the next section, we decided to discuss it here together because of its strong association with the above complex. No slow magnetic relaxation properties were observed in any applied fields, indicating that complex 13 is not an SMM at all. The reason for this is very complicated, in addition to the covalency effects arising from interactions of 5f orbitals with Cp, the Coulomb and exchange interaction between the 5f and 6d/7s could lead to additional splitting and mixing within the low-lying multiplets. Hence, it does not exhibit nice SMM properties, and is not even a SMM, although the non-Kramer U(II) is in a strict point symmetry.
A less bulky cyclopentadienyl ligand, CpiPr4 was used by Arnold et al. to isolate a neutral (CpiPr4)2UI (14a), and a base-free metallocene, [(CpiPr4)2U][B(C6F5)4] (14b), extending the uranium metallocene family.37 The major differences in the structural parameters are that the Cp–U–Cp angles in 14a and 14b are both smaller than those in 12a and 12b. One of the interesting points is that the Cp–U–Cp angle in 14a is about 10° smaller than that in 12a, but upon removing iodide, this parameter has a difference of approximately 25°. Similar to 12a and 12b, both complexes exhibit field-induced slow magnetic relaxation, and the process is dominated by a Raman relaxation mechanism. Remarkably, the field-induced relaxation time of 14b is longer than that of 12a, which is attributed to the flexibility of the CpiPr4 ligand, resulting in a smaller quenching of the orbital angular momentum.
Recently, Mills et al. reported one uranium(III) bis-Cp hypersilanide complex [U(Cp′′)2{Si(SiMe3)3}] (15) (Cp′′ = {C5H3(SiMe3)2-1,3}) by the salt elimination reaction between [{M(Cp′′)2(μ-I)}2] with [K{Si(SiMe3)3}].38 The mean U–Cp distance of 2.4726(2) Å, and the U–Si distance of 3.116(2) Å are comparable to those of other uranocenium and U–Si bond-containing complexes. The Cp–U–Cp is bent with an angle of 131.02(4)°, hence, the uranium center could be considered to be in a distorted pseudo-trigonal planar geometry. Ac susceptibility measurements on it show slow magnetic relaxation below 5 K, the authors suggested that the behaviour was due to Raman and quantum tunnelling of magnetisation processes rather than the Orbach process.
Next to uranium in the periodic table of elements is the artificially prepared transuranium element neptunium. Similar to uranium(III), neptunium(IV) possessing a 5f3 electron configuration, is a Kramer ion and is considered a promising candidate for the synthesis of outstanding SMMs. The first mononuclear transuranic complex showing magnetic memory effects was the homoleptic neptunocene, [Np(η8-COT)2] (16) with an approximate D8h symmetry (Fig. 6).39 At 1.8 K, the magnetic hysteresis presents a butterfly shape, and it is open when the field is above 5 T. The magnetic moment is still far from saturation even at the maximum field of 14 T, indicating strong magnetic anisotropy. The temperature dependence of the out-of-phase susceptibility upon an applied field of 5 kOe shows clear peaks in the temperature range of 2–60 K. The Arrhenius fit is essentially linear at higher temperatures, resulting barrier of 28.5 cm−1 with τ0 = 1.1 × 10−5 s. Other processes are likely to contribute to the overall relaxation, including hyperfine interactions with the I = 5/2 nucleus of 235Np.
Fig. 6 The crystal structure of [Np(η8-COT)2]. Colour code: Np, yellow-green; C, grey. H atoms are omitted for clarity. |
Another actinide complex with 5 unpaired electrons was reported recently by Layfield and co-workers, namely, [K(2.2.2-cryptand)][U(η5-CpiPr5)2] (19) (Fig. 7), which could be isolated by reducing [U(η5-CpiPr5)2] or [IU(η5-CpiPr5)2] with potassium graphite in the presence of 2.2.2-cryptand.17 The oxidation state of the uranium is +1, expanding the range of accessible oxidation states for actinides. Compared to the linear uranium(II) metallocene, the Cp–U–Cp angle of the uranium(I) complex is smaller, and U–Cp distances are longer due to the one-electron reduction. The electron configuration of the rare uranium(I) complex was confirmed to be 5f3(7s/6dz2)1(6dx2−y2/6dxy)1 by theoretical studies and magnetic measurements. Remarkably, unlike its U(III) analogues, complexes 12 and 13, which possess field-induced slow magnetic relaxation behaviours, exhibit SMM properties in zero dc field. The temperature-dependent relaxation time could fit the Arrhenius equation well, resulting in an effective energy barrier of 14(1) cm−1 with τ0 = 1.1(3) × 10−7 s. The behaviour of magnetic hysteresis is similar to that of [Np(η8-COT)2], presenting a narrow butterfly shape at 2 K.
Fig. 7 The crystal structure of [K(2.2.2-cryptand)][(η5-CpiPr5)2U]. Color codes: U, yellow; K, purple; O, red; C, grey; N, blue. H atoms are omitted for clarity. |
Thereafter, based on the same TrenTIPS ligand, the same group replaced O2− with N3− to study the influence of varying crystal fields at the uranium(V) centre. A series of dinuclear contact ion pair complexes [{U(TrenTIPS)(μ-N)( μ-M)}2] (21a) (M = Li, Na, K, Rb, Cs), terminal nitride separated ion pair complexes [U(TrenTIPS)(N)][M(12C4)2] (21b), and the contact ion pair complexes [U(TrenTIPS)(μ-N){M(15C5)}] (21c) were prepared.43 All complexes, except 21Li, display SMM behaviours with Ueff of 14–28 cm−1, but show no hysteresis.
Classification | Complex | Local symmetry | U eff (cm−1) | H dc (Oe) | τ 0 (s) | Ref. |
---|---|---|---|---|---|---|
a The dc field used in ac susceptibility measurements. | ||||||
5f–5f exchange | {NpVIO2Cl2}{NpVO2Cl(THF)3}2 | D 5h/Oh | 97.3 | — | — | 15 |
[{U(BIPMTMSH)-(I)}2(μ:η6:η6-C6H5CH3)] | C 4v | — | 0 | — | 44 | |
[Cp3UOUO(THF)(Pacman)] | C 3v | 19.0 | 10000 | 3.3 × 10−8 | 45 | |
4f–5f exchange | [{UO2Dy(py)2(Pacman)}2] | D 5h | — | — | — | 46 |
3d–5f exchange | [{UO2(salen)}2Mn(py)3]6 | D 5h | 98.7 | 0 | 3.0 × 10−12 | 47 |
[{[Mn(TPA)I][UO2(Mesaldien)][Mn(TPA)I]}I] | D 5h | 56.3 | 0 | 5.0 × 10−10 | 48 | |
[{Fe(TPA)Cl}{UO2(Mesaldien)}{Fe(TPA)Cl}]I | D 5h | 37.5 | 0 | 3.4 × 10−9 | 49 | |
[UFe2BPPAH] | D 5h | 6.3 | 0 | 7.8 × 10−6 | 49 | |
[UNi2BPPAH] | D 5h | 19.0 | 0 | 2.4 × 10−8 | 49 | |
[{[UCo2BPPAH][UO2-(Mesaldien)][Co(BPPA)(Py)]}]I | D 5h | 21.2 | 1500 | 2.9 × 10−9 | 50 | |
{[UO2-(salen)(py)][Mn(py)4](NO3)}n | D 5h | 93.1 | 500 | 3.1 × 10−11 | 51 | |
{[UO2(Mesaldien)][Mn(NO3)-(Py)2]}n | D 5h | 84.8 | 0 | 6.2 × 10−12 | 52 | |
2p–5f exchange | [U(TpMe2)2(bipy)] | C 2 | 22.6 | 500 | 2.6 × 10−7 | 55 |
[{(SiMe2NPh)3-tacn}UIV(η2-N2Ph2˙)] | C 3v | 12.2 | 500 | 4.7 × 10−8 | 56 | |
[K(2.2.2-cryptand)][{((Me3Si)2N)3U(III)}2(μ-bpym)] | D 3h | 8.9 | 1000 | 4.7 × 10−8 | 57 | |
[{((Me3Si)2N)3U(IV)}2(μ-bpym)][BPh4] | D 3h | 2.0 | 1000 | 6.4 × 10−4 | 57 |
Fig. 8 The crystal structure of (NpVIO2Cl2)[NpVO2Cl(THF)3]2. Color codes: Np, yellow-green; O, red; C, grey; Cl, yellow. H atoms are omitted for clarity. |
An arene-bridged diuranium complex [{U(BIPMTMSH)-(I)}2(μ:η6:η6-C6H5CH3)] (23) (BIPMTMS = C(PPh2NSiMe3)2 showing an inverted-sandwich structure was reported by the Liddle group.44 Apart from the bridged arene, one carbanion and two nitrogen atoms from the BIPMTMS ligand, and one halide complete the coordination geometry of each uranium. The U–Carene bond lengths are in the range of 2.553(7)–2.616(7) Å, and the average C–C bond length in the toluene ring is 1.436(16) Å, slightly longer than that in free toluene. The none zero χT product at the lowest measurement temperature suggests that there are no strong interactions of antiferromagnetic exchange in the complex, and the authors noted that the magnetism cannot be described as a superexchange interaction between localized uranium ions. The magnetic hysteresis shows a butterfly shape at a low temperature of 1.8 K, and the zero coercivity at zero fields suggests the presence of QTM. Ac susceptibility measurements show that no out-of-phase signals were found in the zero applied field, but clear signals could be observed when it is in an applied field of 1000 Oe, indicating that complex 31 is a field-induced SMM. The energy barrier was not determined because too few ac peaks could be observed.
In 2016, Arnold and co-workers reported a U(IV)–uranyl(V) complex [Cp3UOUO(THF)(H2L)] (24; L = “Pacman” Schiff-based polypyrrolic macrocycle) formed by one electron transfer from Cp3U to uranyl(VI) dication.45 At room temperature, the χT product approaches the expected value for a U(IV)–U(V) pair, confirming that the reduction of the uranyl group occurs. Ac susceptibility measurements were performed to investigate the dynamic magnetic property of complex 24, and well-defined peaks in the out-of-phase component imply a slow relaxation of the magnetization below 4 K. The thermal activation barrier of 19 cm−1 could be obtained by fitting to the relaxation time, and this value is quite similar with other mononuclear U(V) SMMs. Given that U(IV) is often non-magnetic in low symmetry geometries, and that no apparent sign of magnetic coupling between uranium ions, the authors attributed the SMM behaviour to the uranyl(V) group.
Simple models facilitate the analysis of magneto-structural relationships and help one to tune the SMM properties. Based on this, Mazzanti et al. presented two trimeric heterodimetallic complexes [{[M(TPA)I][UO2(Mesaldien)][M(TPA)I]}I] (M = Mn (27a), Cd (27b) (Mesaldien = N,N′-(2-aminomethyl)diethylenebis(salicylidene imine, TPA = tris(2-pyridylmethyl)amine).48 In the structures, the two terminal [M(TPA)I]+ are bridged by the central [UO2(Mesaldien)]− through the two oxo atoms. The mean U–O–Mn angle and the Mn–U–Mn angle are 169.7(1.7)° and 173.77(5)°, respectively, in the linear {Mn–OUO–M} core of complex 27a. Dc magnetic susceptibility measurements were performed on complexes 27a and 27b, and the diamagnetic Cd(II) in complex 27b could help to understand the spin–orbit and ligand field effects contribution from the uranium(V) centre. The fit to dc magnetic susceptibility yields J = +7.5 cm−1, and this positive value indicates ferromagnetic coupling between uranium(V) and manganese(II) ions. Ac susceptibility measurements were performed to investigate the magnetization dynamics of complex 27a in a zero dc field, showing strong frequency-dependent peaks. An energy barrier of 56.3 cm−1 with τ0 = 5.02 × 10−10 s was obtained by fitting to the Arrhenius equation. The SMM behaviour of complex 27a can be attributed to the magnetic coupling between uranium(V) and manganese(II) ions and the large Ising-type anisotropy of the OUO unit. Slow magnetic relaxation could be observed for 27b under an applied dc field due to the single-ion anisotropy of uranium(V).
Soon afterwards, a series of closely related trimer [{Fe(TPA)Cl}{UO2(Mesaldien)}{Fe(TPA)Cl}]I (28a) and [UM2BPPAH] (Fe for 28b and Ni for 28c) (BPPAH = bis(2-picolyl)(2-hydroxybenzyl)amine) were reported,49 showing similar linear (M–OUO–M) cores. The maxima in the χ′′M(v) plot were observed at 2.1–5.7 K for 28a, at 1.8–5 K for 28b, and at 1.8–3.3 K for 28c in a zero applied field, with the anisotropy barriers of 37.5 cm−1 (τ0 = 3.40 × 10−9 s), 6.3 cm−1 (τ0 = 7.82 × 10−6 s) and 19.0 cm−1 (τ0 = 2.40 × 10−8 s), respectively. All the complexes are SMMs without the applied field, meaning that intramolecular magnetic coupling in U(V)-TM(II) (TM = transition metal) has a significant influence on the dynamic magnetic properties.
The same group then extended the UO2-TM(II) family by preparing two cobalt-containing complexes, {Co(TPA)}{UO2(Mesaldien)}I (29a) (Fig. 10 left) and [{[UCo2BPPAH][UO2-(Mesaldien)][Co(BPPA)(Py)]}]I (29b) (Fig. 10 right).50 The isolation of the dinuclear or trinuclear complexes is controlled by the appropriate choice of the ligand binding with the cobalt ion. The trinuclear complex shows SMM behaviour with an energy barrier of 21.2 cm−1 originating from the coupling between uranium(V) and cobalt ions, but it needs an applied field. In contrast, the dinuclear complex with a bent Co–O–U does not show frequency dependency due to no evidence of U(V)–Co(II) magnetic coupling.
Moreover, two U(V)-TM(II) coupling-based complexes [{[UO2-(salen)(py)][Mn(py)4](NO3)}]n (30)51 and {[UO2(Mesaldien)][Mn(NO3)-(Py)2]}n (31),52 showing single chain magnet behaviours were developed. Their energy barriers are 93.1 cm−1 and 84.8 cm−1, respectively. It should be emphasized that, unlike the previous U(V)-TM(II) SMMs, the hysteresis curves of these two complexes have nice rectangular shapes and show large coercive fields with no obvious drop in the zero fields.
In general, the magnetic bistability of the ground state is required for slow magnetic relaxation, so tetravalent uranium with two single electrons is not considered a good spin carrier for the construction of SMMs as this non-Kramers ion exhibits an orbital singlet ground state at low temperatures. In 2015, Pereira and co-workers reported one mononuclear uranium(IV) complex with an zobenzene radical ligand [{(SiMe2NPh)3-tacn}UIV(η2-N2Ph2˙)]56 (34) (N2Ph2 = bipyridine radical) obtained by the addition of azobenzene to the trivalent uranium precursor 7a. Surprisingly, complex 34 showed field-induced slow magnetic relaxation in external dc fields in the range of 1–2.5 kOe, and a relaxation barrier with a value of 12.2 cm−1 at 2.5 kOe could be obtained. The opening of the butterfly-shaped hysteresis loops could be apparently observed at 1.7 K but without a zero-field coercivity. This study provides the first SMM example based on the 5f2 uranium(IV). The unusual magnetic properties for 34 were derived from the magnetic coupling between uranium(IV) center and the radical, which seems to switch the parity from non-Kramers to Kramers.
The redox activity of 2,2,-bipyrimidine (bpym) allows it to be a radical, and the double-pocket chelating features are often used as a bridging ligand in complexes. Recently, Mazzanti et al. reported a series of bpym-bridged diuranium complexes (Fig. 11), [{((Me3Si)2N)3U(IV)}2(μ-bpym)] (35a), [K(2.2.2-cryptand)][{((Me3Si)2N)3U(III)}2(μ-bpym)] (35b), [K(2.2.2-cryptand)]2[{((Me3Si)2N)3U(III)}2(μ-bpym)] (35c), and [{((Me3Si)2N)3U(IV)}2(μ-bpym)][BPh4] (35d),57 showing analogous structures but different metrical parameters. Complexes 35b and 35d present radical-bridged U(III)–bpym˙−–U(III) and U(IV)–bpym˙−–U(IV) cores, respectively. Variable-temperature dc magnetic susceptibility measurements were performed for all four complexes, and non-strong ferromagnetic couplings were observed, but the weak interactions cannot be completely ruled out. Under a dc field of 2000 Oe, the maxima in the out-of-phase susceptibility component for 35b and 35d, were observed at low temperatures. The reversal barrier of 8.92(2) cm−1 for 35b was determined with τ0 = 8.87(23) × 10−7 s, while for 35d was a miniscule barrier of 1.95(3) cm−1 with an unusually large τ0 value of 6.38(10) × 10−4 s. In contrast, complex 35c containing two uranium(III) ions bridged by a diamagnetic bpym ligand, does not show slow magnetic relaxation behaviours. The differences between complex 35b and 35c could be assigned to the bridging radical, which enhances coupling and suppresses QTM.
Fig. 11 Crystal structure of the [{((Me3Si)2N)3U}2(μ-bpym)] core. Color codes: U, yellow; Si, light yellow; C, grey; N, blue. H atoms are omitted for clarity. |
Recently, Singh et al. suggested that the design of actinides SMMs with high anisotropies should consider not only the coordination geometry but also the nature of the ligand field interaction.58 For a 5f3 ion with an oblate electron density, σ donor ligands are better than π donor ligands to increase the anisotropy in linear geometry. Then, extremely bulky N-heterocyclic carbenes might be helpful in the preparation of high-performance SMMs; however, considering the large ionic radii and high oxidation states of the actinides, this poses a great challenge for the synthesis. In 2022, the Rajaraman group studied a series of three-coordinate U(III) and Pu(III) SMMs by employing ab initio calculations.59 The authors suggested that the blocking barriers of the two reported T-shaped complexes [U(NSiiPr2)2(I)] and [U(NHAriPr6)2I] (AriPr6 = 2,6-(2,4,6-iPr3C6H2)2C6H3) could exceed 900 cm−1, and the pyramidal [Pu{N(SiMe3)2}3] could yield an energy barrier of ∼1933 cm−1. These three complexes are relatively easy to synthesize, so they could be quickly verified. Just this year, the Gagliardi group theoretically investigated the electronic structure and SMM behaviours of californium(III) complex Cf(dbm)3(bpy) (bpy = 2,2′-bipyridine; dbm = dibenzoylmethanoate).60 This study shows that the computed energy barrier of the 5f9 complex is higher than that of the Dy(III) analogue due to the stronger spin–orbit coupling and crystal field splitting. Then, a linear L–Cf(III)–L complex might show promising performance, however, the scarcity and instability limit experimental studies.
By comparison, the development of multinuclear actinide SMMs with magnetic exchange is better, and it seems to be a more promising route to prepare high-performance An-SMMs. However, a big challenge in this direction is that few theoretical calculations have been performed to study the magnetic configuration relations due to the increased complexity of multinuclear An-SMMs with magnetic exchange. Considering that some building blocks, such as Dy-D6h,61–63 Ln-TM5,64,65 and UO2(salen) that exhibit high magnetic anisotropy have been developed, the combinations of lanthanides and actinides might be interesting targets to pursue. Scheme 1 shows two unknown species that might have fascinating SMM properties. The simple structures can help to understand the magneto-structural relationships, for example, by replacing magnetically anisotropic lanthanides with diamagnetic yttrium or isotropic gadolinium, facilitating the determination of magnetic exchange strength and the effect of each component on SMM properties.
Scheme 1 Two potentially excellent single-molecule magnets combined by UO2(salen) and Dy-D6h (left), and metallacrown (right, surrounding ligands are omitted for clarity). |
Since the mononuclear [TbPc2]− SMM was reported, the field of lanthanide SMMs has developed rapidly, with the energy barrier increasing from the initial 230 cm−1 to the latest 1631 cm−1 and the hysteresis blocking temperature rising from the liquid helium region to 80 K. In contrast, the development of actinide SMMs has lagged much behind. However, as more examples of actinides are reported, more theoretical models are developed, and both contribute to each other, this direction will also flourish like the lanthanide SMMs, which of course requires a sincere collaboration between chemists and physicists.
This journal is © the Partner Organisations 2023 |