Exploring a New Family of (Phosphinochalcogenoyl)Europocenes for Magneto‐Optical Thermometry

Abstract

The development of novel materials based on lanthanide metallocenes remains a valuable endeavour, due to the synthetic accessibility, compositional tunability, and impressive magnetic and luminescent properties of this class of well‐studied compounds. Europium(II) complexes are particularly interesting in this regard, given the relative stability, strong paramagnetism, and characteristic Laporte‐allowed 5d→4f emission of this ion. In this study, we tap into the powerful class of phosphinocyclopentadienyl ligands to develop a novel family of potassium (bis(dimethylamino)phosphinochalcogenoyl)cyclopentadienides K[C₅Me₄(PCh(NMe₂)₂] of sulfur, selenium, and tellurium, and explore their coordination chemistry with Eu^II. These ligands readily form homoleptic europocenes where the pendant phosphine chalcogenide arms also coordinate the europium center, and these are characterized via single‐crystal X‐ray diffraction, absorbance, and photoluminescence spectroscopy. Characterization via magnetic circular dichroism (MCD) spectroscopy reveals that these complexes act as extremely sensitive probes for MCD‐based magneto‐optical thermometry, reaching relative thermal sensitivities up to 170 % K⁻¹ at cryogenic temperatures, with temperature uncertainties as low as 0.003 K. These complexes represent the first examples of organometallic MCD thermometers, and their remarkable properties and interesting structures highlight the promise of this new class of ligands and the utility of the Eu^II ion in lanthanide‐based functional materials research.

A new class of substituted cyclopentadienyl ligand featuring a coordinating phosphinothioyl, ‐selenoyl, or ‐telluroyl pendant arm in addition to the typical η⁵‐mode is presented. The homoleptic divalent europocenes obtained from these are the first examples of organometallic probes for magnetic circular dichroism thermometry at cryogenic temperatures, attaining relative sensitivities of up to 170 % K⁻¹ at 1.7 K with resolution as high as 0.003 K.

Introduction

Molecular materials featuring divalent lanthanide ions have attracted significant attention due to their remarkable optical^{[ 1} , ² , ³ , ⁴ , ^{5 ]} and magnetic^{[ 6} , ⁷ , ⁸ , ⁹ , ^{10 ]} properties. Among the divalent lanthanides, the classical Eu^II is of particular interest for several reasons. It is the most accessible and stable of all divalent lanthanide ions, with the Eu^II‐Eu^III redox couple having a small negative reduction potential of only −0.35 V versus the standard hydrogen electrode.^{[ 11 ]} Consequently, the chemistry of the Eu^II ion is rich,^{[ 12} , ¹³ , ^{14 ]} and a variety of coordination and organometallic constructs containing Eu^II have been reported. In terms of optical properties, the spectra of divalent ions like Eu^II are dominated by Laporte‐allowed d‐f transitions, which are intense and broad.^{[ 2} , ¹⁵ , ^{16 ]} Unlike the Laporte‐forbidden, narrow‐bandwidth, and long‐lived f‐f transitions typical of trivalent lanthanide ions, the d‐f transitions of a divalent ion are highly sensitive to the coordination environment,^{[ 17} , ^{18 ]} given the spatially expanded and diffuse nature of the excited‐state 5d orbitals.^{[ 19} , ^{20 ]} As a result, molecular systems based on Eu^II form a family of luminescent materials which together cover the entire visible spectrum and encroach upon the near‐infrared.^{[ 21} , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ^{27 ]} This sensitivity to environmental stimulus also means that Eu^II‐based luminescent materials can be employed for band‐shift luminescence thermometry,^{[ 1} , ¹⁴ , ²⁰ , ²⁸ , ^{29 ]} complementarily to the ratiometric or lifetime‐based methods commonly demonstrated for trivalent lanthanide complexes.^{[ 30} , ³¹ , ³² , ³³ , ³⁴ , ^{35 ]}

Additionally, the strong paramagnetism of Eu^II makes it interesting from a magnetochemical perspective as well. Eu^II is isoelectronic with Gd^III, thus has maximal single‐ion ground‐state spin owing to its 4f⁷, ⁸S_7/2 ground state electronic configuration. Eu^II‐based materials have been successfully employed in magnetic resonance imaging^{[ 36} , ³⁷ , ^{38 ]} and, despite the formally isotropic electron density distribution of the free ion, have even demonstrated slow magnetic relaxation in a tailored ligand field.^{[ 39 ]} Unlike Gd^III, Eu^II can emit in the visible region,^{[ 40 ]} so multifunctional materials that can simultaneously exhibit interesting optical and magnetic properties become accessible with Eu^II. Thus, developing new (magneto‐)optical materials based on Eu^II is a focal point of our research.

Our endeavours toward novel Eu^II‐based molecular materials led us to the first organoeuropium(II) band‐shift luminescence thermometer, the hemimetallocene dimer [Cp*Eu(BH₄)(THF)₂]₂,^{[ 28 ]} and later to a systematic study of how different ligands, especially different cyclopentadienides, affect the emission properties in complexes of that nature.^{[ 14 ]} Among these, the ligand ^bdpCp^# (1‐(bis(dimethylamino)phosphino)‐2,3,4,5‐tetramethylcyclopentadienyl)^{[ 41 ]} stood out to us because the complex [^bdpCp^#Eu(BH₄)(THF)₂]₂ bearing this esoteric phosphinocyclopentadienide demonstrated the second‐best thermometric performance of the complexes we investigated. Phosphinocyclopentadienides are highly versatile ambident ligands, having already been extensively studied in the coordination and organometallic chemistry of main group, transition metal, and f‐block elements, and combinations thereof;^{[ 42} , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ^{51 ]} one need only look at the ubiquity of substituted phosphinoferrocenes as an example.^{[ 52} , ⁵³ , ⁵⁴ , ⁵⁵ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ^{61 ]} Furthermore, the free phosphine moiety of such ligands is poised for further chemical elaboration, offering a functional handle from which more diverse and complex systems can be constructed and evaluated for their potential favourable properties.

Phosphinocyclopentadienyl ligands can be conveniently derivatized by transforming the formally P^III phosphine component into a tetracoordinate phosphorus analogue. Doing so allows the electronic properties of the ligand to be modulated, and offers a way to diversify the nature of the potential phosphorus‐based coordinating moiety. Such an approach has generated a variety of lanthanide complexes bearing cyclopentadienylideneaminophosphorane ligands^{[ 62} , ^{63 ]} that, by virtue of their nature as “constrained‐geometry complexes” (CGCs),^{[ 64} , ^{65 ]} have demonstrated utility in small‐molecule activation and polymerization catalysis.^{[ 66} , ⁶⁷ , ^{68 ]} However, in this sense we are more interested in the ability of phosphines to be oxidized to formally P^V phosphino(chalcogenoyl) species, e.g. by direct oxidation with elemental chalcogens.^{[ 69 ]} We take inspiration from the binary Eu^II chalcogenides EuCh (Ch  =  O, S, Se, Te),^{[ 70} , ^{71 ]} which are a famous class of magneto‐optical materials that have been thoroughly studied since the 1960s. These compounds exhibit interesting properties such as ferromagnetic semiconductivity, magneto‐optic Kerr effect,^{[ 72 ]} and strong magnetic circular dichroism (MCD). The bis(dimethylamino)phosphino moiety borne by ^bdpCp^# is particularly disposed toward oxidative reactivity with chalcogens because it is highly Lewis basic, allowing it to form a stable phosphine selenide or telluride and addressing the lability typically intrinsic to such species,^{[ 69} , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ^{79 ]} thus enabling a survey of structure‐property relationships across the whole chalcogen group.

Given that Eu^II metallocenes are already well‐known for their luminescence properties,^{[ 14 ]} we envisioned that combining guidance from the field of solid‐state materials with the well‐understood chemistry of lanthanide metallocenes would be a promising avenue toward new magneto‐optical materials. Further, despite its advantageous characteristics, ^bdpCp^# is unexpectedly understudied, and to the best of our knowledge, there exist only two other reports which discuss it or its derivatives,^{[ 41} , ^{80 ]} and phosphine chalcogenide derivatives thereof are heretofore unknown. Thus, we report herein the synthesis and characterization of a new family of potassium 1‐(bis(dimethylamino)phosphinochalcogenoyl)cyclopentadienides. These salts serve as convenient ligand precursors in the synthesis of the corresponding homoleptic Eu^II metallocenes, in which the pendant phosphine chalcogenide moieties are coordinated to the Eu center. This work focuses on the heavier chalcogens (S, Se, Te) because we are most interested in investigating the properties of soft‐donor ligands. These are still less well‐studied in lanthanide chemistry given the preference of hard lanthanide ions for hard donors, and are advantageous for luminescent materials applications due to their lower phonon energy.^{[ 1} , ^{81 ]} Further, Eu^II is, among lanthanides, a relatively large and soft ion, which is more amenable to accommodating a soft‐donor ligand like a heavy chalcogen.^{[ 1 ]} The Eu^II metallocenes are characterized spectroscopically and found to be highly sensitive probes for MCD‐based thermometry in the cryogenic regime, becoming the first examples of organometallic MCD thermometers.

Results and Discussion

Synthesis and Characterization of Ligand Precursors

Beginning from the phosphinocyclopentadiene H^bdpCp^#,^{[ 41 ]} the synthesis of the phosphinochalcogenoyl‐substituted proligands studied here is straightforward and high‐yielding at the gram‐scale (Scheme 1). In a two‐step, one‐pot procedure, a toluene solution of H^bdpCp^# is first oxidized with elemental chalcogen (sulfur, selenium, or, gratifyingly, tellurium) to furnish the phosphine chalcogenide, followed immediately by deprotonation with potassium hexamethyldisilazide (KHMDS) to give the corresponding potassium cyclopentadienide. We designate these compounds KCp^#PCh (Ch  =  S, Se, Te), in analogy to the cyclopentadienylphosphazene “CpPN” ligands of Sundermeyer et al.^{[ 62} , ⁶⁶ , ^{82 ]} It should be noted that the best results are obtained from these protocols when the starting material H^bdpCp^# is very pure (see Supporting Information). Each of the ligand precursors KCp^#PCh is well‐soluble in polar solvents like tetrahydrofuran (THF) and pyridine, but insoluble in aromatic solvents or alkanes.

Scheme 1.

Synthesis of (phosphinochalcogenoyl)cyclopentadienides KCp^#PCh and their corresponding Eu^II complexes 1‐Ch.

Analysis of the multinuclear NMR spectra of all compounds in pyridine‐d ₅ reveals the expected pattern of resonances, and the spectral features are largely unremarkable. In the case of KCp^#PSe and KCp^#PTe, an uncommonly‐observed ⁴ J _C‐P coupling is resolved in the ¹³C resonance ascribed to the 3,4‐methyl substituents, due to the “rigid W” effect.^{[ 83 ]} Oxidation of the phosphino substituent engenders a significant upfield shift in the ³¹P resonances, which occur within the typical range for the corresponding (dimethylamino)phosphine chalcogenide^{[ 84 ]} in all cases (KCp^#PS 74.8 ppm, KCp^#PSe 68.2 ppm, KCp^#PTe 28.7 ppm cf. the free phosphine K^bdpCp^# at 115.9 ppm). A ³¹P‐coupled doublet at ‐167 ppm is observed in the ⁷⁷Se spectrum of KCp^#PSe, well downfield of (Me₂N)₃PSe or (Me₂N)₂PhPSe (ca. ‐370 ppm). The magnitude of the coupling constant ¹ J _P‐Se (660 Hz) is rather smaller than those observed in (Me₂N)₃PSe or (Me₂N)₂PhPSe (ca. 800 Hz), and is instead closer to 1,1′‐bis(alkylphosphino)ferrocene diselenides (ca. 700 Hz) or even Me₃PSe (684 Hz). These observations may imply a significant contribution of the P═Se multiple‐bonded resonance form of the phosphinoselenoyl substituent in this compound, compared to the usually predominant dipolar P⁺‐Se⁻ form.^{[ 85 ]} Based on analysis of its NMR spectra, KCp^#PTe seems to exhibit some coordinative fluxionality of the phosphinotelluroyl substituent in solution, wherein the dissolved species might interconvert between metallocene‐like polymers (vide infra), smaller oligomers or clusters, and monomers; this is discussed in more detail in the Supporting Information.

Synthesis of Europocenes

With the ligand precursors in hand, we proceeded to investigate their complexation reactivity with Eu^II. The reaction of [Eu^II(BH4)₂(THF)₂]^{[ 22 ]} with 2 eq. KCp^#PS in THF followed by crystallization from THF/pentane at ‐35 °C yielded yellow blocklike crystals suitable for X‐ray diffraction analysis, which were found to be the metallocene solvate [(Cp^#PS)₂Eu(THF)]·THF (1‐S·THF, Figure S14). The lanthanide borohydrides [Ln^X(BH₄)_x(THF)_2–4] typically exhibit the useful property of allowing stepwise exchange of their tetrahydroborate ligands, thus serving as a useful platform for the synthesis of heteroleptic complexes.^{[ 86 ]} Given this, we were surprised to find that we could not obtain any mono(cyclopentadienyl) products from the equimolar reaction of [Eu^II(BH4)₂(THF)₂] and KCp^#PS. Instead, the metallocene product was always formed, leaving 0.5 eq. of [Eu^II(BH4)₂(THF)₂] unreacted.

We discovered that crystals of 1‐S·THF are extremely easily desolvated, quickly collapsing to a bright orange powder when not protected by their mother liquor or immersed in oil. Indeed, treatment of these crystals with an excess of any non‐coordinating solvent afforded complete deterioration to give said orange powder, which is well‐soluble in aromatic solvents. This behaviour precluded any further characterization of 1‐S·THF, so this complex will not be discussed further. It is necessary to conduct the metathesis reaction in THF to ensure solubility of both precursors. However, removal of THF upon completion of the reaction and redissolution of the residue in fluorobenzene gives, after a facile workup, the pure, solvent‐free homoleptic europocene [(Cp^#PS)₂Eu] (1‐S) as an orange powder in good yield. Accordingly, the corresponding solvent‐free phosphinoselenoyl and phosphinotelluroyl complexes [(Cp^#PSe)₂Eu] (1‐Se) and [(Cp^#PTe)₂Eu] (1‐Te) were synthesized in an analogous manner and were obtained as orange powders in good yield. Complexes 1‐Ch are well‐soluble in aromatic solvents and sparingly soluble in alkanes; their alkane solubility decreases on going down the chalcogen group. These complexes were found to be indefinitely stable in the solid state under inert conditions. Notably, solid 1‐Ch can be handled briefly in air, decomposing over the course of minutes or hours instead of instantaneously. Their resistance towards aerobic decomposition decreases on going down the chalcogen group, as expected.

Structural Analysis

The ease with which 1‐S·THF is desolvated was surprising to us, given that removal of coordinated THF from highly oxophilic lanthanide ions, where possible at all, typically requires forcing conditions.^{[ 87 ]} Such behaviour is uncommon but not unprecedented, having been observed in, e.g., the bulky thiophenolate Eu^II(SAr*)₂ (Ar* = 2,6‐bis(2,4,6‐triisopropylphenyl)phenyl)),^{[ 88 ]} the samarocenes (COT)Sm^III(C₅R¹ ₄PR² ₂)(THF)_n (COT  =  cyclooctatetraenyl dianion, R¹  =  H or Me, R²  =  Me or Ph, n  =  1 or 4),^{[ 89 ]} and other “CGC‐type” complexes of cyclopentadienylideneaminophosphoranes.^{[ 62} , ^{90 ]} Clearly, the stereoelectronic characteristics of (Cp^#PCh)⁻ reduce the affinity of its metallocene complexes to bind solvent ligands, thereby avoiding the often undesirable “permanent coordination” of, e.g., THF to lanthanide centers. This is a valuable characteristic, demonstrated even for a relatively large ion like Eu^II which would normally demand a high coordination number.

The solvent‐free europocenes 1‐Ch can each be recrystallized from fluorobenzene or fluorobenzene/pentane to furnish orange blocklike crystals suitable for X‐ray diffraction analysis, and their solid‐state structures are shown in Figure 1.^{[ 91 ]} These complexes are isomorphous in terms of molecular geometry but differ in their crystalline arrangement. Each complex demonstrates a typical bent‐metallocene coordination environment courtesy of the two η⁵‐cyclopentadienyl moieties, and the two chalcogen atoms from the pendant phosphinochalcogenoyl substituents complete the coordination sphere of Eu. The complexes are helically chiral as a result of the coordination of the phosphinochalcogenoyl arms,^{[ 90} , ^{92 ]} but crystallize as racemates in non‐Sohncke space groups in all cases. These structural features are reminiscent of the remarkable ytterbocene [(Ph₂P(C₅Me₄)NAd)₂Yb^II] reported by Petrov et al.^{[ 90 ]} 1‐S crystallizes in the orthorhombic space group Pbcn, with one half‐molecule in the asymmetric unit. 1‐Te also crystallizes in an orthorhombic space group (Pna2₁), but with one complete molecule in the asymmetric unit. Meanwhile, 1‐Se instead crystallizes in the monoclinic space group P2/c, with two halves of two different molecules (one of each enantiomer) comprising the asymmetric unit.

Figure 1.

Solid‐state structures of Eu^II complexes 1‐Ch. Labels are given only to crystallographically unique atoms. Thermal ellipsoids are shown at 50% probability; hydrogen atoms are omitted for clarity. Packing diagrams are given in Figures S16–S18.

Although we were unsuccessful in obtaining X‐ray quality crystals of any of the ligand precursors KCp^#PCh, it is important to note that the infrared spectra of the precursors are extremely similar to their corresponding complexes 1‐Ch (Figures S6–S8). Only small shifts in vibrational frequencies are observed, which are ascribable simply to coordination to Eu^II. This implies that the solid‐state structures of KCp^#PCh probably also contain metallocene‐like motifs.^{[ 93 ]} To corroborate this, the vibrational frequencies of KCp^#PCh formulated as “half‐metallocene” fragments resembling the asymmetric unit of 1‐S (Figure S9) were calculated via DFT at the CAM‐B3LYP^{[ 94 ]}/def2‐TZVPD^{[ 95} , ^{96 ]} level. The calculated frequencies were scaled using a single linear scaling equation^{[ 97 ]} (see Supporting Information), ultimately giving calculated infrared spectra that agree well with the experimental spectra (Figures S11–13). From this analysis, we can observe that the vibrational modes most affected by coordination to Eu are those dominated by motions of the ‐(NMe₂)₂PCh substituent. These results uphold that the most likely solid‐state structure of each KCp^#PCh consists of chains of potassium cations sandwiched between (Cp^#PCh)⁻ anions, which also explains the similarity in the infrared spectra between KCp^#PCh and their corresponding Eu^II complexes 1‐Ch.

Selected metrical parameters of all 1‐Ch are shown in Table 1. The Eu‐Cp(cent) distances are, overall, average among related europocenes,^{[ 98 ]} and become progressively slightly elongated on going down the chalcogen group. The Cp(cent)‐Eu‐Cp(cent) angles are also average among bent europocenes^{[ 99 ]} and show no correlation with the identity of the chalcogen. However, 1‐S is distinctly less bent than its heavier congeners, which is probably due to decreased steric demand from the comparatively smaller sulfur atoms. Similarly, our europocenes are considerably less bent than [(Ph₂P(C₅Me₄)NAd)₂Yb^II] (Cp(cent)‐Yb‐Cp(cent) angle 132.6(1)°)^{[ 90 ]} which bears bulkier ligands. With reference to the phosphinochalcogenoyl substituents and torsional angle τ,^{[ 53} , ^{92 ]} complex 1‐S adopts a synclinal eclipsed conformation of the cyclopentadienyl rings, whereas complexes 1‐Se and 1‐Te adopt anticlinal staggered conformations instead (Figure S19), which is likely due again to reasons of steric demand.

Table 1.

Selected metrical parameters of 1‐Ch. Distances are given in Å and angles are given in degrees.

	1‐S	1‐Se a)	1‐Te
Eu‐Cp(cent) distance	2.5337(1)	2.5379(1)	2.5492(4) b)
Cp(cent)‐Eu‐Cp(cent) angle	143.749(1)	137.728(2)	138.83(2)
Eu‐Ch distance	2.951(1)	3.0816(4)	3.2760(5) b)
Ch‐Eu‐Ch angle	111.09(3)	109.91(1)	108.30(1)
Torsional twist angle τ c)	85.3(2)	118.8(1)	116.2(1)
P‐C(Cp) distance	1.755(3)	1.764(2)	1.763(2) b)
P‐Ch distance	1.997(1)	2.1540(6)	2.3967(6) b)
Avg. P‐N distance	1.654(4)	1.656(2)	1.662(2) b)
N‐P‐N angle	109.9(2)	108.14(8)	109.3(1)
Planarity at N d)	0.192(4) 0.044(4)	0.217(2) 0.003(2)	0.060(2) 0.269(2) 0.244(2) 0.020(2)

^a) The two half‐molecules in the asymmetric unit differ slightly in their metrical parameters but are similar (Figure S20), so one molecule is taken as representative. The difference between values is not crystallographically significant (Table S3), so this value is given as the average.

^b) The asymmetric unit of this structure contains one complete molecule with two slightly different distances (or four, in the case of the P‐N distance), so this value is given as the average.

^c) Torsion angle C‐Cp(cent)‐Cp(cent)‐C between the cyclopentadienyl carbon atoms directly bound to the P atoms of the phosphinochalcogenoyl moiety and the corresponding cyclopentadienyl centroids.^{[ 53 ]}

^d) The distance between the nitrogen atom and the plane defined by the phosphorus atom and the two methyl carbon atoms on N.^{[ 57 ]} Listed in order of atom labeling, i.e., P1‐N1, P1‐N2, etc.

The Eu‐Ch distances in 1‐Ch are in all cases shorter than the corresponding sum of covalent radii and are average among corresponding Eu^II─Ch bonds.^{[ 100} , ¹⁰¹ , ^{102 ]} Especially for 1‐Se, it is interesting to note that the P─Se distance is similar to the average Eu‐alkene(cent) distance in the structurally related tethered olefin complex [(C₅Me₄)SiMe₂(CH₂CH = CH₂)]₂Eu^II] (3.09 Å).^{[ 103 ]} Due to coordination to Eu, the P─Ch distances in 1‐Ch are between that of a typical (alkylamino)phosphine chalcogenide that features some P─Ch multiple bond character (S 1.95 Å,^{[ 104} , ^{105 ]} Se 2.11 Å,^{[ 106} , ^{107 ]} Te 2.36 Å^{[ 108 ]}) and that of a typical P─Ch single bond (S 2.08 Å,^{[ 109 ]} Se 2.22 Å,^{[ 110 ]} Te 2.50 Å^{[ 111 ]}). The P─S distance in 1‐S is noticeably longer than the average P─S distance in the dithioimidodiphosphinate [Eu^II{(NSPPh₂)₂}₂(THF)₂] (1.96 Å)^{[ 1 ]} and is more akin to those found in the phosphinodithioate [Eu^II(Ph₂PS₂)₂(DME)₂] (2.00 Å)^{[ 112 ]} or the phosphoranylidene [(Eu^II{CH(iPr₂P─S)(C₉H₆N‐2)}₂)₂] (2.01 Å).^{[ 113 ]} Meanwhile, the P─Se distance in 1‐Se is similar to the average P‐Se distance in the diselenoimidodiphosphinate [Eu^II{(NSePPh₂)₂}₂(THF)₂] (2.16 Å),^{[ 1 ]} and the phosphinodiselenoate [Eu^II(Ph₂PSe₂)₂(DME)₂] (2.14 Å)^{[ 112 ]} as well as to the CGC‐type zirconocene (η⁷‐C₇H₇)Zr^IV(iPr₂PSe(C₅H₄)) (2.15 Å)^{[ 45 ]} which bears a similar coordinated phosphine selenide arm. We were unable to find any other crystallographically‐characterized P‐Te‐Eu fragments with which to compare 1‐Te, but the P─Te distance in this complex is slightly shorter than average for metal‐bound P─Te motifs.^{[ 114 ]} In all cases, each phosphinochalcogenoyl moiety of 1‐Ch contains one planarized and one pyramidal dimethylamino substituent,^{[ 79 ]} as observed previously in (dialkylaminophosphino)ferrocene chalcogenides.^[16f] The P‐N distances are within the expected range for aminophosphines,^{[ 79 ]} and the P─C(Cp) distance is ∼1.76 Å in all cases; both metrics are average for a (diaminophosphinochalcogenoyl)cyclopentadienyl motif.^{[ 115 ]}

Magneto‐Optical Properties and MCD Thermometry

Despite becoming interested in the temperature‐dependent emission properties of 1‐Ch upon observing intense, visibly thermochromic photoluminescence in the solvent‐free 1‐S, we ultimately determined that the photoluminescence characteristics of 1‐Ch are not conducive to their application as luminescence thermometers. The absolute sensitivity of 1‐S is low (only 1.5 cm⁻¹ K⁻¹ at 330 K), the emission of 1‐Se occurs only at low temperatures and is not thermochromic, and 1‐Te exhibits no photoluminescence at all (a detailed discussion is presented in the Supporting Information). However, the strong paramagnetism of Eu^II opens another avenue for applications of these complexes based on their optical properties. Namely, such systems are excellent candidates as probes for MCD thermometry, a magneto‐optical thermometry technique recently introduced by our group.^{[ 116 ]} MCD spectroscopy^{[ 117 ]} measures the differential absorption (ΔA) of left and right circularly polarized components of light (lcp and rcp, respectively), induced by a magnetic field oriented parallel (positive H) or antiparallel (negative H) to the direction of the light propagation. The ΔA signal results from the magnetic perturbation of the states involved in the optical transitions due to the Zeeman effect.^{[ 118 ]} Magneto‐optical thermometers have the potential to show greatly improved relative sensitivity (S_r ) compared to typical luminescence thermometers, since the technique relies on the presence of a highly temperature‐dependent C‐term in the MCD spectra of paramagnetic species at low temperature.^{[ 119 ]} This behaviour enables magneto‐optical thermometers to operate at cryogenic temperatures, complementing conventional luminescence thermometers^{[ 33 ]} that tend to be poorly sensitive in this important temperature range for quantum technologies, spintronics, and superconductors. Hence, we aimed to study the variable‐temperature and variable‐field MCD (VTVH‐MCD) spectra of 1‐Ch in the solid state to evaluate their performance in this area.

Firstly, the bulk magnetic properties of 1‐Ch were probed via direct current (DC) SQUID magnetometry. The DC magnetic susceptibility χ of each complex 1‐Ch was measured at 1000 Oe between 300 and 1.9 K (Figure S27), and the room‐temperature χT products (1‐S 7.88, 1‐Se 7.84, 1‐Te 7.83 cm³ K mol⁻¹) agree well with the theoretical value of 7.88 cm³ K mol⁻¹ for an isolated Eu^II ion (⁸S_7/2, S  =  7/2, g_J =  1.997≅2). The χT product of 1‐S remains expectedly nearly constant over the measured temperature range, while the χT products of 1‐Se and 1‐Te decrease gently with decreasing temperature, with the latter decreasing more quickly than the former. The isothermal magnetization shows typical field‐dependence in all cases (Figure S28a–c). Molecular divalent europium systems tend to exhibit 4f⁷, ⁸S_7/2 ground states consisting of four Kramers doublets separated by a ZFS of up to a few wavenumbers.^{[ 28} , ³⁹ , ^{120 ]} There is a slight downturn in the χT products of all 1‐Ch at very low temperatures and a slight nesting of reduced magnetization isotherms (Figure S28d–f), which are consistent with this model. The saturation magnetization for 1‐S (6.84 Nβ) at 1.9 K and 7.0 T is close to the expected value (7 Nβ). However, in conjunction with the behaviour of the χT product, the saturation values for 1‐Se (6.16 Nβ) and 1‐Te (5.81 Nβ) under the same conditions are lower than expected. The polarizability of the chalcogens increases on going down the group, as does the polarization of the P─Ch bonds,^{[ 121 ]} placing progressively greater electron density at the metal‐bound chalcogen atoms of 1‐Ch. Thus, we attribute the somewhat unusual magnetic behaviour observed in 1‐Se and 1‐Te to an increased nephelauxetic effect in these heavier chalcogen analogues.^{[ 39} , ¹²² , ^{123 ]}

Next, we recorded the VH‐MCD spectra of these complexes, which are shown in (Figures 2a,c,e) and S30–32. Following the trend observed from the DC magnetometry data, the MCD intensity of 1‐Ch is strong for 1‐S and weakens on going down the chalcogen group. Six distinguishable components are selected for analysis in all cases, arranged into two positive (lcp‐absorption dominated) and two negative (rcp‐absorption dominated) bands. The components are 310.3, 350.6, 380, 393.4, 439.8, and 498.3 nm for 1‐S, 281.9, 332, 351.5, 394, 447.3, and 490 nm for 1‐Se, and 300, 322, 346.1, 408.5, 475.8, and 497.1 nm for 1‐Te. In all cases, the transitions observed in the MCD spectra of 1‐Ch correspond with their diffuse reflectance spectra and, where applicable, with their photoluminescence excitation spectra (Figure 2b,d,f). As observed in the diffuse reflectance spectra, transitions occurring below ∼315 nm are predominantly ligand‐based, whereas those occurring above this wavelength are primarily f‐d in nature. For 1‐S and 1‐Se, the f‐d transitions giving rise to the greatest emission intensity (especially those around 400 nm) also exhibit the greatest MCD intensity. Transitions at 281.9 nm (1‐Se) and 300 nm (1‐Te), assigned as putative P‐Ch n‒π* transitions (see Supporting Information), are likewise clearly visualized in the MCD spectra. The application of an equivalent antiparallel magnetic field to all complexes results in the expected mirror images.

Figure 2.

Left column: variable‐field magnetic circular dichroism (VH‐MCD) spectra of 1‐S a), 1‐Se c), and 1‐Te e) at 1.9 K, under applied magnetic fields of 0.1 ‒ 7 T. The dotted lines represent the MCD spectrum under an applied magnetic field of ‐1 T. Right column: comparison of VH‐MCD spectra at selected fields with solid‐state diffuse reflectance (coloured lines in lower portion) and photoluminescence excitation spectra (where applicable, filled grey curves in lower portion) for 1‐S b), 1‐Se d), and 1‐Te (f). Dotted blue lines indicate the distinguishable components observed in the MCD spectra, as described in the text.

The magnetic isotropy of the Eu^II ion, combined with the low symmetry of the complexes (C₂ ),^{[ 124 ]} the large number of overlapping states, and the known complexity of the energetic landscape of Eu^II systems,^{[ 71} , ^{125 ]} would render disambiguation of the exact nature of each transition infeasible. Hence, we have not attempted a more detailed assignment of the MCD spectra. Nevertheless, a highly detailed assignment of the MCD spectra is not required for application in MCD thermometry, so we proceeded to analyze the VT‐MCD spectra for 1‐Ch under several applied magnetic fields (0.1, 0.25, 0.5 and 1.0 T) (the spectra for 1‐S under an applied magnetic field of 0.1 T are shown in Figure 3a as a representative example, and the rest are shown in Figures S33–S35). Upon increasing the temperature, positive bands become less positive and negative bands become less negative. Thus, by taking the difference in integrated area between a lcp‐ and a rcp‐dominated transition band, we can implement a highly sensitive thermometric parameter Δ  = I_lcp   − I_rcp . In accordance with the strong MCD intensity of 1‐S, an acceptable signal‐to‐noise ratio was obtained at 0.1 T, while for 1‐Se and 1‐Te with weaker intensity, the minimum usable field was 0.25 T. A sigmoidal logistic function (Equation 1) was used to fit the Δ parameters:^{[ 126 ]}

(1)

Figure 3.

a) Variable‐temperature magnetic circular dichroism (VT‐MCD) spectra of 1‐S under an applied magnetic field of 0.1 T. Regions I_n of the spectra from which MCD thermometric parameters Δ are derived are indicated by yellow boxes. b) Temperature‐dependence of the Δ parameters of 1‐S under an applied magnetic field of 0.1 T. The solid lines represent the sigmoidal logistic function fitted to the experimental data; best fitting parameters are given in Table S5. Relative sensitivities calculated for each Δ parameter of 1‐S under an applied magnetic field of 0.1 T c), 1‐Se d) and 1‐Te e) each under an applied magnetic field of 0.25 T. The blue line labelled “MP” indicates the sensitivity derived from a multiparametric thermal readout method that combines the temperature‐dependence of all Δ for a given set of conditions into one parameter.

And S_r can then be calculated by dividing the derivative of the fitted function by the fitted Δ values (Equation 2):

$$S_r\left(\%\right)=\left[\frac{\left(\frac{\partial \mathrm{\Delta }}{\partial T}\right)}{\mathrm{\Delta }}\right]·100$$ (2)

Various thermometric parameters consisting of combinations of I_lcp and I_rcp were evaluated to find the best performance (Δ parameters for 1‐S at 0.1 T are given in Figure 3b as a representative example, and the rest are shown in Figures S36–38), and the corresponding S_r values were calculated for each combination (the S_r curves for each complex at their lowest operational field are shown in Figure 3c‐e as representative examples, and the rest are shown in Figures S39–41). A summary of the single‐parameter thermometric figures of merit of 1‐Ch is given in Table 2. We therefore find that complexes 1‐Ch can behave as extraordinarily sensitive MCD thermometers; in particular, 1‐S shows an unprecedented single‐parameter maximum S_r of 93.9 % K⁻¹ at T  =  1.7 K and H  =  0.1 T. To further validate the thermometric performance of these complexes, the temperature uncertainty δT was calculated as reported previously.^{[ 116 ]} At the temperature of maximum sensitivity, values of δT in the range 10⁻³ – 10⁻² K are obtained (Figures S42, S46–48). In accordance with our previous findings regarding MCD thermometers, these results demonstrate that increasing the applied magnetic field broadens the operational range at the expense of maximum S_r , when distinguishable signal‐to‐noise ratio is taken as the only criterion of merit. For a more quantitative comparison, we take a δT value of 1.0 K as the “useful” threshold and define the sensing range as the temperature range over which δT ≤ 1.0 K.

Table 2.

Summary of thermometric figures of merit obtained for 1‐Ch applied as probes for MCD‐based thermometry.

		0.1 T		0.25 T		0.5 T		1.0 T
		Single‐Parameter	Multi‐parametric	Single‐Parameter	Multi‐parametric	Single‐Parameter	Multi‐parametric	Single‐Parameter	Multi‐parametric
1‐S	Max. Sr (% K−1) a)	93.9	170	60.3	114	53.8	89.9	46.7	61.0
	δT (K) b)	0.005	0.003	0.02	0.009	0.02	0.01	0.02	0.01
	Sensing Range (K) c)	1.7 ‒ 3.5		1.7 ‒ 10		1.7 ‒ 12.4	1.7 ‒ 18.7	1.7 ‒ 11.7	1.7 ‒ 17.4
1‐Se	Max. Sr (% K−1) a)	n.d.		61.5	112	52.6	90.7	34.3	61.2
	δT (K)	n.d.		0.05	0.03	0.02	0.01	0.07	0.04
	Sensing Range (K) c)	n.d.		1.7 ‒ 8		1.7 ‒ 15		1.7 ‒ 10.9
1‐Te	Max. Sr (% K−1) a)	n.d.		85.9	130	50.8	94.4	35.2	59.4
	δT (K)	n.d.		0.002	0.001	0.02	0.01	0.02	0.01
	Sensing Range (K) c)	n.d.		1.7 ‒ 6		1.7 ‒ 12.5		1.7 ‒ 13.7

^a) Maximum S_r is attained at 1.7 K under all conditions.

^b) Temperature uncertainty at the temperature of maximum S_r .

^c) Temperature range for which δT is ≤ 1.0 K. n.d.  =  not determined.

The thermal sensing performance of these complexes can be further enhanced by considering all combinations of I_lcp and I_rcp simultaneously, via a multiparametric thermal readout method.^{[ 127 ]} This method applies a multiple non‐linear regression to calculate one new S_r from the sum of individual parameters, according to (Equation 3):

$$S_{r,mp}\left(\%\right)=\sqrt{{\sum }_{i=1}^n{\left[\frac{\left(\frac{\partial \mathrm{\Delta }}{\partial T}\right)}{\mathrm{\Delta }}\right]}^2}·100$$ (3)

And correspondingly one new δT calculated based on the new S_r .^{[ 116 ]} Using this approach, dramatic improvements in maximum S_r and δT are achieved in all cases, with 1‐S attaining a maximum S_r of 170 % K⁻¹ with δT  =  0.003 at T  =  1.7 K and H  =  0.1 T. Overall, improvements in sensitivity and temperature uncertainty of 50–60% are observed using the multiparametric approach, except for 1‐S at 1.0 T and 1‐Te at 0.25 T, where the improvements are somewhat less at ca. 25 and 41%, respectively. A summary of the thermometric figures of merit obtained from the multiparametric readout approach is also given in Table 2. Complexes 1‐Ch are thus the first examples of organometallic complexes employed for MCD thermometry, and highlight the huge potential of Eu^II complexes as magneto‐optical thermometers. We predict that further engineering of Eu^II complexes will yield magneto‐optical materials with higher stability and potentially further improvements in thermometric performance.

Conclusion

Three novel potassium cyclopentadienides featuring bis(dimethylamino)phosphinochalcogenoyl substituents containing sulfur, selenium, and tellurium, as well as their homoleptic complexes with Eu^II, were prepared and characterized. The chalcogen atom of these ligands coordinates to the lanthanide left and, in doing so, disfavours the coordination of solvent ligands. Photoluminescence studies show that the identity of the chalcogen atom has an important effect on the luminescence properties of the Eu^II complexes, as the emission is progressively quenched on going down the chalcogen group. These complexes show highly temperature‐dependent MCD and represent the first examples of organometallic complexes as probes for magneto‐optical thermometry via MCD spectroscopy. In this application, values of relative sensitivity up to 170 % K⁻¹ with temperature uncertainty as low as 0.003 K at cryogenic temperatures are obtained, far surpassing what is attainable with conventional luminescence thermometry. Further investigation of lanthanide complexes of these useful and interesting ligands is a continuing area of research in our group.

Supporting Information

The authors have cited additional references within the Supporting Information.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Supporting Information

Diaz‐Rodriguez R. M., Gálico D. A., Murugesu M., Angew. Chem. Int. Ed. 2025, 64, e202517812. 10.1002/anie.202517812

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.