Ready Access to Anhydrous Anionic Lanthanide Acetates by Using Imidazolium Acetate Ionic Liquids as the Reaction Medium

Abstract

Access to lanthanide acetate coordination compounds is challenged by the tendency of lanthanides to coordinate water and the plethora of acetate coordination modes. A straightforward, reproducible synthetic procedure by treating lanthanide chloride hydrates with defined ratios of the ionic liquid (IL) 1‐ethyl‐3‐methylimidazolium acetate ([C₂mim][OAc]) has been developed. This reaction pathway leads to two isostructural crystalline anhydrous coordination complexes, the polymeric [C₂mim] _n [{Ln₂(OAc)₇} _n ] and the dimeric [C₂mim]₂[Ln₂(OAc)₈], based on the ion size and the ratio of IL used. A reaction with an IL : Ln‐salt ratio of 5 : 1, where Ln=Nd, Sm, and Gd, led exclusively to the polymer, whilst for the heaviest lanthanides (Dy−Lu) the dimer was observed. Reaction with Eu and Tb resulted in a mixture of both polymeric and dimeric forms. When the amount of IL and/or the size of the cation was increased, the reaction led to only the dimeric compound for all the lanthanide series. Crystallographic analyses of the resulting salts revealed three different types of metal‐acetate coordination modes where η ² μκ ² is the most represented in both structure types.

A straightforward pathway to synthesize anhydrous lanthanide acetate complexes from lanthanide chloride hydrates and the ionic liquid (IL) 1‐ethyl‐3‐methylimidazolium acetate ([C₂mim][OAc]) is described. The combination of the lanthanide size and the amount of IL introduced allows the selective formation of two isostructural crystalline anhydrous compounds the dimeric, [C₂mim]₂[Ln₂(OAc)₈], and the polymeric, [C₂mim] _n [{Ln₂(OAc)₇} _n ].

Introduction

The simple acetate anion ([OAc]⁻) is ubiquitous in Nature and offers an exciting coordination chemistry because of the plethora of coordination modes it can accommodate with metal ions (Figure 1, right), especially in anionic metal complexes. ^[1] With larger metal ions such as alkaline earths (Ae²⁺), lanthanides (Ln³⁺), or uranyl ([UO₂]²⁺), structurally distinct anionic complexes form structures ranging from isolated monomers chelated by [OAc]⁻, to dinuclear oligomers and polymeric complexes with chelating, bridging, and mixed coordination modes. ^[2] Some Ln³⁺ ions, for example, Nd³⁺, have been found with up to three different metal/ligand ratios, pointing towards the large hidden potential of these systems to provide unique acetate complexes. ^[2] This wide range of coordination modes which [OAc]⁻ can adopt with the lanthanides prompted us to explore reproducible routes to homoleptic anionic salts of the series of 4f elements. Only a few structures of homoleptic 4f‐element acetate complexes have been reported so far. ^[3] A particular challenge in the preparation of homoleptic complexes of trivalent lanthanides is their extreme oxophilicity, ubiquitous water frequently enters their coordination environment.

Figure 1.

Top: General structure of dialkylimidazolium acetate ILs. Bottom: Observed coordination modes of the [OAc]⁻ ion with trivalent lanthanide cations.

As a vehicle to provide a high concentration of coordinating ligands promoting the formation of homoleptic complexes, and large stable cations (e. g., dialkylimidazolium cations, Figure 1, left) which can stabilize large and potentially unusual coordination complex anions, ionic liquids (ILs) have proven a valuable synthetic tool, in particular for obtaining homoleptic complexes.[ ² , ⁴ ] Their tendency to promote coordination of large numbers of anions and the exclusion of neutral solvent molecules has led to new coordination numbers, various coordination modes, and polymeric structures, ^[5] in addition to providing routes to new luminescent and magnetic materials. ^[6]

One feature of ILs, which renders them particularly interesting as solvents and reaction media for the conversion of hydrates, particularly for the readily hydrated lanthanide salts, is the stability of ILs to temperatures above the boiling point of water. Certain ILs with coordinating and even weakly coordinating anions ^[7] can be reacted neat with hydrated metal salts on the benchtop and heated, with or without vacuum, to drive off volatiles such as such as H₂O or even HCl. ^[8] Historically, access to anhydrous lanthanide complexes has been limited because of their high affinity for oxygen, ^[9] thus we have been exploring the use of ILs to dehydrate metallic salts leading to ready access to new compounds including [C₂mim][UO₂(OAc)₃], [C₂C₂im][UO₂(OAc)₃], [C₄mim][UO₂(OAc)₃], [C₂C₂im]₂[La(OAc)₅], [C₂mim]₂[Nd(OAc)₅], and [C₂mim] _n [Sr(OAc)₃] _n (where (C₂mim)=1‐ethyl‐3‐methylimidazolium, (C₂C₂im)=1,3‐diethylimidazolium, (C₄mim)=1‐butyl‐3‐methyl‐imidazolium).[ ² , ^4a , ⁸ ] Interestingly, in these reactions the formation of oxo‐ and hydroxy compounds, which frequently occurs when attempting to dehydrate hydrates of strongly Lewis basic metal salts, is suppressed. Moreover, while readily accessible coordination chemistry is thus easily available, systematic studies are still lacking.

In this work, we build upon our observation that the readily available IL [C₂mim][OAc] allows for easy and reliable access to water‐free f‐element compounds and develop a general, reliable synthetic procedure applicable across the entire range of 4f metal ions leading to anhydrous materials. Here, single crystal X‐ray diffraction analysis has allowed us to unveil the complex nature of acetate coordination modes in these anhydrous, homoleptic lanthanide anionic complexes across the entire lanthanide series.

Results and Discussion

Synthesis

The synthetic methodology developed (Scheme 1) is based on reacting readily available starting salts such as LnCl₃ ⋅ xH₂O (Ln=La, Pr, Nd, Sm−Lu, x=6 or 7) with an excess of the IL [C₂mim][OAc] in a fixed molar ratio, driving off the water by heating, and slowly cooling to obtain a crystalline product. The crystalline complexes were characterized by powder and single crystal X‐ray diffraction (PXRD and SCXRD), differential scanning calorimetry (DSC), and infrared spectroscopy (IR).

Scheme 1.

General scheme of the synthesis of anhydrous salts of anionic 4f/[OAc]⁻ complexes and influence of the amount of IL introduced on the selectivity of the final product. (*: PXRD indicates that these compounds were obtained, albeit along with unidentified side products.)

When using a 5 : 1 IL : Ln‐salt molar ratio, a combination of PXRD and SCXRD revealed the formation of two series of anhydrous, homoleptic lanthanide acetate complexes, dimeric [C₂mim]₂[Ln₂(OAc)₈] (Ln=Eu, Tb−Lu) and polymeric [C₂mim] _n ‐[{Ln₂(OAc)₇} _n ] (Ln=Nd, Sm−Tb). For the largest lanthanides tested, La and Pr, only Ln(OAc)₃(H₂O) could be isolated (Figure S4 in the Supporting Information). A detailed description of their structures can be found below. Whilst this reaction with light lanthanides (Nd, Sm, and Gd) led to the formation of exclusively polymeric [C₂mim] _n [{Ln₂(OAc)₇} _n ] (see Figure S1 for PXRD patterns of the products), with the heaviest lanthanides (Dy−Lu) the formation of dimeric [C₂mim]₂[Ln₂(OAc)₈] was observed (see Figure S2 for PXRD patterns of the products). The reaction with the intermediate lanthanides, Eu and Tb, resulted in a mixture of both polymeric and dimeric forms (PXRD patterns are shown in Figures S3, S5, and S6). The PXRD patterns of representatives for each of the two compound series, [C₂mim] _n [{Ln₂(OAc)₇} _n ] (Ln=Gd) and [C₂mim]₂[Ln₂(OAc)₈]) (Ln=Er) are shown in Figure 2. However, even if in most cases the final products are obtained as phase pure, for some lanthanides the PXRD patterns indicated that some minor unidentified side products were generated during the reaction.

Figure 2.

Top: Experimental (red) vs. theoretical (black) PXRD pattern of [C₂mim] _n [{Gd₂(OAc)₇} _n ]. Bottom: Experimental (blue) vs. theoretical (black) PXRD pattern of [C₂mim]₂[Er₂(OAc)₈].

There are multiple effects which could result in the preferential formation of each of the three compounds like the ion size or lanthanide salt concentration in the IL, that is, IL−Ln‐salt ratio. To shed more light on this, several experiments were conducted.

To investigate whether it is truly the decreasing ion size, and/or the IL : Ln‐salt ratio that leads to the preferential formation of [C₂mim]₂[Ln₂(OAc)₈] (which would correspond to [C₂mim][OAc] ⋅ Ln(OAc)₃) over [C₂mim] _n [{Ln₂(OAc)₇} _n ] (which could be re‐written as 0.5[C₂mim][OAc] ⋅ Ln(OAc)₃) different IL : Ln‐salt ratios, namely 3 : 1 and 5 : 1 molar equivalents (equiv.) were tested (Scheme 1).

When the reactions were conducted with 3 equiv. of IL, the large lanthanides (Pr, Nd, and Sm) yielded only Ln(OAc)₃(H₂O) as a phase pure compound. However, when a 5 : 1 ratio was used, [C₂mim] _n [{Ln₂(OAc)₇} _n ] formed for Nd and Sm. For Pr, the 5 : 1 ratio already led to Pr(OAc)₃(H₂O). Thus, there is a stronger tendency for the earlier lanthanides to form Ln(OAc)₃(H₂O). In order to prevent its formation, a higher IL : Ln‐salt ratio has to be used.

For that reason we tested a 10 : 1 Ln‐salt ratio. Raising the amount of IL to 10 equiv. led to the dimeric compound even for Pr, Nd, and Sm, which otherwise were only obtained for the heavier lanthanides. However, close examination of the PXRD patterns revealed that only [C₂mim]₂[Sm₂(OAc)₈] was phase pure, whilst for the Pr and Nd reactions, mixtures with other, yet, unidentified crystalline phases were observed.

In summary, the larger the lanthanide is, the higher the IL concentration has to be in order to lead to anhydrous complexes.

Stability and chemical reactivity

Neat products which were carefully washed with acetonitrile to remove all excess IL showed no detectable degradation over two months. However, it was noted that degradation occurred whenever excess IL was present (Figures S7 and S8). The dimeric compounds degraded faster than the polymeric ones, even in a sample container sealed under ambient atmosphere. The powder turned partially into a colorless oil, presumably due to the formation of acetic acid generated from the reaction with moisture in the air as the samples developed the characteristic smell of acetic acid.

By way of illustrative example, the polymer [C₂mim] _n [{Gd₂(OAc)₇} _n ] and the dimer [C₂mim]₂[Er₂(OAc)₈] were followed using PXRD. Without washing with acetonitrile, after four weeks under ambient conditions, the PXRD pattern of the remaining powder indicated that no more dimer was present and indicated the presence of the polymeric form, as well as some undetermined crystalline compounds (Figure S7). This was later confirmed by means of a unit cell check via SCXRD. [C₂mim] _n [{Gd₂(OAc)₇} _n ] was also phase pure after reaction, but after four weeks degraded to Gd(OAc)₃(H₂O) with other unknown phases (Figure S8). This last observation confirms that larger lanthanides seems to have a certain tendency to generate the monohydrate acetate either with a lower amount of IL or after degradation.

Structural results

The two isostructural series of anhydrous, polymeric [C₂mim] _n ‐[{Ln₂(OAc)₇} _n ] and dimeric [C₂mim]₂[Ln₂(OAc)₈] with their unit cell parameters are provided in Table 1. We next describe each series of compounds from a structural point of view.

Table 1.

Comparison of cell parameters for the crystal structures determined in this study.

Compound	SG	a, Å	b, Å	c, Å	β, °	V, Å3	IL : Ln ratio	CCDC
[C2mim]2[Ln2(OAc)8]
[C2mim]2[Pr2(OAc)8][a]	P21/n	9.535(2)	12.315(2)	16.599(3)	93.830(3)	1944.8(7)	10 : 1	1920417
[C2mim]2[Nd2(OAc)8]	P21/n	9.543(1)	12.493(1)	16.727(2)	93.596(3)	1990.4(4)	10 : 1	1920416
[C2mim]2[Sm2(OAc)8][a]	P21/n	9.487(1)	12.260(2)	16.576(2)	93.709(5)	1924.0(5)	10 : 1	1920418
[C2mim]2[Eu2(OAc)8][a]	P21/n	9.489(8)	12.193(2)	15.563(1)	94.220(3)	1795.8(3)	5 : 1	1920419
[C2mim]2[Tb2(OAc)8][a]	P21/n	9.476(1)	12.132(9)	16.557(2)	94.239(9)	1898.2(3)	5 : 1	1920420
[C2mim]2[Dy2(OAc)8][a]	P21/n	9.449(1)	12.181(1)	16.595(2)	94.155(2)	1905.1(3)	5 : 1	1920421
[C2mim]2[Dy2(OAc)8]	P21/n	9.458(5)	12.308(7)	16.690(9)	93.963(2)	1938.2(2)	5 : 1	1920422
[C2mim]2[Ho2(OAc)8]	P21/n	9.466(8)	12.300(9)	16.700(1)	93.965(3)	1939.8(3)	5 : 1	1920423
[C2mim]2[Er2(OAc)8]	P21/n	9.439(2)	12.227(3)	16.698(5)	93.968(9)	1922.5(9)	5 : 1	1920424
[C2mim]2[Tm2(OAc)8]	P21/n	9.426(2)	12.254(2)	16.685(3)	94.065(5)	1922.3(5)	5 : 1	1920425
[C2mim]2[Yb2(OAc)8]	P21/n	9.421(4)	12.255(5)	16.683(8)	94.113(2)	1921.1(2)	5 : 1	1920426
[C2mim]2[Lu2(OAc)8]	P21/n	9.408(6)	12.237(7)	16.652(9)	94.209(2)	1912.0(2)	5 : 1	1920427
[C2C2im]2[Ln2(OAc)8]
[C2C2im]2[La2(OAc)8][a]	P21/n	9.570(3)	12.473(4)	16.853(5)	93.685(2)	2007.4(1)	10 : 1	1920428
[C2mim] n [{Ln2(OAc)7} n ]
[C2mim] n [Nd2(OAc)7] n [a]	Pca21	19.240(1)	7.847(5)	19.118(1)		2886.5(3)	5 : 1	1920432
[C2mim] n [Eu2(OAc)7] n	Pca21	19.303(3)	7.841(1)	19.138(3)		2896.5(9)	5 : 1	1920433
[C2mim] n [Gd2(OAc)7] n	Pca21	19.281(5)	7.833(2)	19.113(5)		2886.8(9)	5 : 1	1920434
[C2mim] n [Tb2(OAc)7] n	Pca21	19.302(1)	7.868(6)	19.113(1)		2902.7(3)	5 : 1	1920435

[a] Measured at 100 K.

[C₂mim]₂[Ln₂(OAc)₈] (Ln=Pr, Nd, Sm, Eu, Tb−Lu), [C₂C₂im]₂[La₂(OAc)₈]

This family of thirteen isomorphous salts containing the discrete, dimeric [Ln₂(OAc)₈]²⁻ ion all crystallized in the monoclinic space group P2₁/n with one half of the dimeric anion (Figure 3) residing around a crystallographic center of inversion and one [C₂mim]⁺ cation in the asymmetric unit. The [C₂mim]⁺ cations are disordered with the ethyl CH₃ group being spread over three positions emulating both cis and trans conformations of a hypothetical [C₂C₂im]⁺ cation, while the [C₂C₂im]⁺ cation in the La structure was observed solely in the cis‐conformation. This degree of freedom explains the isostructurality of both the [C₂C₂im]⁺ and [C₂mim]⁺ structures.

Figure 3.

[Ln₂(OAc)₈]²⁻ dimeric unit in the crystal structures of [C₂mim]₂[Ln₂(OAc)₈] and [C₂C₂im]₂[La₂(OAc)₈]. Coordination modes are indicated according to Figure 1.

The Ln³⁺ cations are 9‐coordinated with the oxygen atoms in the form of a slightly distorted tricapped trigonal prism. The dimeric units (Figure 3) have C_i symmetry, while each Ln coordination polyhedron is asymmetric. The Ln−Odistances range from 2.27(1) to 2.66(1) Å including minor variations with coordination mode, different data collection temperatures, and following the lanthanide contraction. The [OAc]⁻ ions exhibit three coordination modes – η ³ μ ₂ κ ², η ² μκ ², and η ² μ ₂ κ ² (Figure 1).

The η ² μ ₂ κ ² (4) mode in [C₂mim]₂[Nd₂(OAc)₈] represents the shortest Nd−O bond lengths in the structure – 2.41(1)–2.43(1) Å, while the η ² μκ ² (2) distances are longer (2.50(1)–2.53(1) Å) as a result of partial repulsion between two O atoms binding to the same metal cation. This correlates well with the Nd−O distances observed in [C₂mim]₂[Nd(OAc)₅] (2.53(1)–2.64(1) Å) represented solely by the η ² μκ ² mode. Consequently, the η ³ μ ₂ κ ² (5) mode lies in between with Nd−O=2.437(4) and 2.541(4)–2.622(4) Å depending on how many oxygen atoms have been bound to the same metal center.

Dimeric [Ln₂(OAc)₈]²⁻ units are known in the literature and even found in some applications in cryogenics, ^[10] but are extremely rare. The only previously published compound that contains this unit with identical coordination modes is a Gd hexaaza macrocycle complex (2[Gd(C₂₂H₂₆N₆)‐(C₂H₃O₂)₂][Gd₂(C₂H₃O₂)₈] ⋅ C₆H₆ ⋅ 4CHCl₃) where [Gd₂(OAc)₈]²⁻ plays the role of a counterion. ^[11] Hydrates such as [Ln₂(OAc)₈(H₂O)₂]²⁻ have been observed with a few lanthanides (La−Sm), ^[12] but with different acetate coordination modes (three η ² μκ ² and one η ³ μ ₂ κ ² per Ln atom). Other pure, dimeric Ln tetra‐carboxylato complexes include [Eu₂(CH₃BzO)₈]²⁻ and [Ho₂(butenoate)₈]²⁻ with rather large aromatic counterions. ^[13] Lastly, similar tri‐carboxylato dimers [Ln₂(CH₃COO)₆(H₂O)₄]⋅nH₂O[ ¹⁰ , ¹⁴ ] have been observed for the heavier Ln (Eu−Lu) exhibiting the same η ² μκ ² and η ³ μ ₂ κ ² modes as the latter. Though the η ² μ ₂ κ ² coordination mode could still be obtained in the fluorinated Tb₂(CF₃COO)₆(H₂O)₆, all initial η ³ μ ₂ κ ² bridges have been distorted resulting in η ² μ ₂ κ ².

[C₂mim]_n[{Ln₂(OAc)₇}_n] (Ln=Nd, Eu‐Tb)

This second isostructural series is very sensitive to the lanthanide size and the reaction ratio (5 : 1 IL : Ln‐salt). Here the acetate ligand again shows its versatility by allowing a lower number of coordinating acetates by forming anionic polymers. These compounds crystallize in the orthorhombic space group Pca2₁ with two independent Ln positions and one independent cation in the asymmetric unit. The linear polymeric anion consists of repeating [Ln₂(OAc)₇]⁻ where each Ln is coordinated by nine oxygen atoms in the form of a slightly distorted tricapped trigonal prism. Ln1 is coordinated to two chelating and five unidentate acetate ligands, while Ln2 to three chelating and three unidentate ligands. The distribution of unidentate and chelating ligands violates any possible mirror symmetry for a trigonal prism, and as a result the polymer chains are chiral with the metal ion acting as the chiral center. Both positions alternate along the b axis being connected by three acetate bridges. The [OAc]⁻ ions exhibit η ³ μ ₂ κ ² (5), η ² μκ ² (2), and η ² μ ₂ κ ² (4) coordination modes similar to [C₂mim]₂[Ln₂(OAc)₈]; however, the proportion of the terminating η ² μκ ² mode is significantly lower (Figure 4).

Figure 4.

Polymeric chain [{Nd₂(OAc)₇} _n ]⁻ in the crystal structure of [C₂mim] _n ‐[{Nd₂(OAc)₇} _n ]. Coordination modes are indicated according to Figure 1.

To analyze the bonding patterns we will use [C₂mim] _n [{Nd₂(OAc)₇} _n ] as an example. In addition to two symmetry‐independent Nd positions, there are seven independent acetate ions resulting in a wide spread of the Nd−O bonds (2.36(1)–2.77(1) Å). Although the Nd−O bonds in different coordination modes show partial overlap, the monodentate coordinated [OAc]⁻ (η ² μ ₂ κ ² and in part η ³ μ ₂ κ ² similarly to [C₂mim]₂[Nd₂(OAc)₈]) cover the lower range of that spectrum – 2.365(7)–2.497(6) Å, while the bidentate [OAc]⁻ (η ² μκ ² and η ³ μ ₂ κ ²) partially overlapping the upper one, 2.446(7)–2.767(6) Å. These distances are also comparable to those in the two known Nd acetates ^[15] (see below).

Among the published Nd acetates, [C₂mim] _n [{Nd₂(OAc)₇} _n ] is structurally closer to the acetate hemihydrate Nd(OAc)₃ ⋅ 0.5 H₂O ^[15a] or the Pr acetate monohydrate ^[16] than to the anhydrous Nd(OAc)₃. ^[15b] The two former exhibits a linear polymeric chain with similar acetate coordination modes – only η ³ μ ₂ κ ² and η ² μ ₂ κ ². Both crystallographically independent Ln positions in Nd(OAc)₃ ⋅ 0.5 H₂O and Pr(OAc)₃ ⋅ H₂O are nine‐coordinated in the form of distorted tricapped trigonal prisms. Nd(OAc)₃ crystallizes with ten independent Nd positions and in addition to η ³ μ ₂ κ ² introduces two new coordination modes – η ³ μ ₃ κ ² (6) and η ⁴ μ ₃ κ ² (7). The Nd positions are mostly 9‐coordinated with coordination polyhedra that are significantly more irregular capped trigonal prisms. Polymeric chains are also present in the structure, but due to additional Nd(OAc) _n groups, with the new coordination modes η ³ μ ₃ κ ² and η ⁴ μ ₃ κ ², they merge into layers.

Among the related polymeric Ln acetates an even more complex connectivity has been observed in catena‐(μ ⁴‐acetato)‐(μ ³‐acetato)‐bis(μ ²‐acetato)‐aqua‐di‐europium(II) ^[17] introducing another coordination mode η ⁵ μ ₄ κ ² (8) resulting in a denser layered motif. It is also worth noting that the acetate anion may offer plenty of other coordination modes, for example, η ⁴ μ ₄ κ ², η ⁵ μ ₅ κ ² or η ⁶ μ ₅ κ ², though they have so far not been observed in pure lanthanide compounds.

Vibrational spectroscopy

The different coordination modes also show in the infrared spectra of the compounds. [C₂mim]₂[Er₂(OAc)₈], and [C₂mim] _n ‐[{Gd₂(OAc)₇} _n ], were chosen as representatives for each of the two structure series.

Acetates have been intensively studied by vibrational spectroscopy since the 1950s. ^[18] Conveniently, sodium acetate was used as a standard for the comparison of acetates. It is known to show two significant peaks at 1440 (CH₃ deformation) and 1578 (C−O stretching) cm⁻¹. ^[19] IR data for the IL [C₂mim][OAc] shows the values of the two peaks of the anion be shifted to 1377 (CH₃ deformation) and 1560 (C−O stretching) cm⁻¹ (Figure S13).

IR spectra of the dimeric and polymeric compounds reveal the non‐equivalency of the acetate oxygen atoms. In the dimeric [C₂mim]₂[Er₂(OAc)₈], there are now three peaks for the acetate at 1404 (CH₃ deformation), 1566 (C−O stretching) and 1608 (C−O stretching) cm⁻¹ (Figure S16). The deformation of the methyl group seems to be similar whatever the type of coordination is. However, there are two types of carbon‐oxygen stretching signals, which confirm the presence of, at least, two different types of coordination modes in the compound. ^[20] It could be also related to the crystal structure bond lengths. The ligands only coordinated to one metal center have bond lengths (C−O) around 2.41 Å whereas when the acetate ligand bridged the two metal centers the bond lengths are between 2.31–2.57 Å.

In the polymeric [C₂mim] _n [{Gd₂(OAc)₇} _n ], there are four bands at 1405 (CH₃ deformation), 1433 (CH₃ deformation), 1566 (C−O stretching), and 1605 (C−O stretching) cm⁻¹ (Figure S15). It seems that in this case the methyl group vibrational frequency is impacted by the type of coordination because of the hydrogen bonding between some CH₃ groups and the oxygen from the bridged acetates. Nevertheless, in the polymeric compounds the C−O bonds lengths are very similar (2.30–2.42 Å) except one (2.86 Å) from the η ³ μ ₂ κ ² coordination mode.

In summary, IR can be used to distinguish between the two different complexes via the stretching signals of the C−O and CH₃ functions. The differences between the bands can be correlated to the bond lengths and coordination modes as observed from the SCXRD analyses.

Thermal properties

To elucidate the thermal properties of the two structure series, differential scanning thermograms of the two representatives for the two series, [C₂mim]₂[Er₂(OAc)₈], and [C₂mim] _n ‐[{Gd₂(OAc)₇} _n ], were recorded. Upon heating [C₂mim] _n ‐[{Gd₂(OAc)₇} _n ] (Figure 5, top) two clear endothermic phase transitions close to each other can be observed. The first transition starts at 180.2 °C, followed by a second at 188.5 °C. Polarized optical microscopy (POM) indicated a transition of the crystalline solid to a higher order mesophase (Figure S17) before melting into an isotropic liquid.

Figure 5.

DSC thermograms of top: [C₂mim] _n [Gd₂(OAc)₇] _n and bottom: [C₂mim]₂[Er₂(OAc)₈].

During the cooling, vitrification can be observed around 33 °C. This reversible process occurs around 35 °C during the second heating. The compound recrystallized upon heating at 86.5 °C (a small feature of this recrystallization process can be even detected in the first heating cycle) and yielded to the mesophase at 179.4 °C, becoming an isotropic liquid at 187.8 °C. Noteworthy, the second heating shows a slightly better separation between the phase transitions from solid to liquid. Indicating that repeated heating and cooling cycle help to enhance crystallinity. All subsequent heating and cooling cycles are fully reproducible.

The reversible cold crystallization and melting of coordination polymers upon heating is well known among halometallate ^[21] and other organometallic ILs.[ ^5h , ²² ] The metal coordination environment and complex formation in such systems can vary under different conditions of temperature and pressure even when the composition is held constant. ^[23] Anhydrous imidazolium salts of [Ln(OAc)₅]²⁻ ions had not been reported to crystallize under the conditions explored here. ^[3] The unpredictable nature of polymeric metal containing ILs along with the highly flexible coordination chemistry of the chosen anion ([OAc]⁻) along with some unreported experimental data strongly imply the existence of yet more undiscovered complexes in the [C_nmim][OAc]/Ln³⁺ systems, which we hope to continue to explore.

The DSC trace of the first heating of the dimeric compound [C₂mim]₂[Er₂(OAc)₈] (Figure 5, bottom) shows a broad endothermic event between 60 to 130 °C which is attributed to the loss of residual acetonitrile. A broad signal at 146.4 °C indicated a melting from solid to isotropic liquid. Upon cooling a vitrification can be observed around 0 °C. The devitrification occurred around 3.5 °C during subsequent heating. A cold recrystallization was observed at 43.4 °C followed by a solid‐solid phase transition at 76.6 °C. This seems to indicate that the compound could have another structural conformation. However it has not been possible to determine it. The compound started to melt at 153.4 °C again with the event extending over several degrees. The second cooling was similar to the previous one with again only one vitrification around 0 °C, and the subsequent third heating cycle was similar to the second.

The dimer melts about 50 °C before the polymer. This can be attributed on the one hand to the lower degree of condensation, that is, less polymeric character of the dimer, but also to reduced connectivity. This might also be the reason, why the polymeric compound features a mesophase, while the coordination dimer displays only a second crystalline phase.

Conclusion

In summary, we have developed a straightforward synthetic route, based on ionic liquids, towards two anhydrous lanthanide acetate series from their corresponding metal salt hydrates, polymeric [C₂mim] _n [{Ln₂(OAc)₇} _n ] and dimeric [C₂mim]₂[Ln₂(OAc)₈]. PXRD data indicated that for IL : Ln‐salt ratios of 5 : 1, for the heavier lanthanides, that is, Ln=Dy–Lu, the formation of [C₂mim]₂[Ln₂(OAc)₈] is preferred, whilst for the lighter lanthanides, Ln=Nd, Sm, Gd, polymeric [C₂mim] _n [{Ln₂(OAc)₇} _n ] is found. The structural preference crosses over at Ln=Eu and Tb, for which both structures are found. Under the same conditions for La and Pr, only Ln(OAc)₃(H₂O) was isolated. From this one could conclude, that the larger the lanthanide ion is, the higher the tendency to form the lanthanide acetate hydrate. When reducing the IL : Ln‐salt ratio to 3 : 1, even for Ln=Pr, Nd, and Sm, Ln(OAc)₃(H₂O) forms. When a higher IL : Ln‐salt ratio of 10 : 1 is employed, even for Ln=Pr, Nd and Sm, the formation of the dimeric structure is observed. Thus, it appears that a large (e. g., tenfold) excess of IL the formation of the dimer [C₂mim]₂[Ln₂(OAc)₈] is favored for all lanthanides, regardless of their size. This appears to be plausible, as the IL : Ln‐salt ratio is higher in [C₂mim]₂[Ln₂(OAc)₈] (which corresponds to 1 : 1 of [C₂mim][OAc] ⋅ Ln(OAc)₃) than in [C₂mim] _n [{Ln₂(OAc)₇} _n ] (which just has a 1 : 2 ratio according to its composition 0.5 [C₂mim][OAc] ⋅ Ln(OAc)₃)). For La, the use of an IL with a bulkier cation, that is, [C₂C₂im]₂[OAc], allowed access the dimer [C₂C₂im]₂[La₂(OAc)₈], even from aqueous solution.

Both the dimeric and polymeric structures reveal the same acetate coordination modes, namely η ³ μ ₂ κ ², η ² μκ ², and η ² μ ₂ κ ², though with slightly different proportion. The dimeric complexes show low lanthanide size selectivity, though the largest in the series, La, could be crystallized isostructurally only with a larger cation. On the other hand, the acetate anion has been found to be very “creative,” offering extended variety of connectivities with the lanthanide cations and, as a result, multiple formations with the same metal enhancing our interest for further investigations. The formation of the different complexes with their different coordination modes can be easily identified through IR spectroscopy according to the different stretching modes of the C−O and CH₃ groups. The thermal properties of both compound series can be related well to their structural features. As expected, due to the higher degree of condensation, that is, connectivity, the compounds with polymeric structure melt at higher temperatures (ca. 50 °C) than the dimeric ones. For both compounds, before melting the formation of another form is observed. In case of the dimeric material, this is a second solid crystalline phase, whilst in case of the polymeric structure polarizing optical microscopy reveals the formation of a mesophase.

The straightforward methodology established in this work has allowed the formation of extremely rare dimeric [Ln₂(OAc)₈]²⁻ units with a large number of lanthanides which could lead to potential applications depending on the Ln used (magnetism, optical properties, etc.). Moreover some reactions generated either undiscovered phases from the thermal behavior or some undetermined compounds as minority phases observed in PXRD. Taking into account the simplicity of the ligand combined with a huge variety of coordination modes it can deliver, the IL/Ln‐acetate systems still lack deeper understanding and represent a considerable interest in terms of designed synthesis of specific anhydrous lanthanide complexes. There is obviously much room left to explore.

Experimental Section

Syntheses: 1‐Methylimidazole (98 %, Alfa Aesar, Karlsruhe, Germany) and 1‐ethylimidazole (>98 %, Tokyo Chemical Industry, Tokyo, Japan) were distilled over KOH before use. Lanthanide chloride hydrates were obtained from the reaction between Ln₂O₃ (HEFA Rare Earth Canada Co. Ltd, Richmond, BC, Canada) and HCl (37 %, VWR Chemicals, Fontenay‐sous‐Bois, France). Ethyl bromide (98 %, Alfa Aesar, Karlsruhe, Germany) and sodium acetate (99 %, Alfa Aesar) were used as received. All solvents were ‘solvent grade’ and used as received without additional purification.

Synthesis of [C₂C_nim][Br]: In a flask under argon were introduced 1‐alkylimidazole (50 mmol, 1.0 equiv.), ethyl bromide (55 mmol, 1.1 equiv.) and acetonitrile (5 mL). The mixture was heated under reflux for three days. After cooling to room temperature the solution was washed with ethyl acetate and put in a cold bath (−78 °C) to allow the crystallization of the product. The white precipitate was washed several times and then dried under vacuum at room temperature.

[C₂mim][Br] FTIR (υ (cm⁻¹)): 3422, 3136, 3065, 3027, 2975, 2866, 2822, 1770, 1731, 1670, 1571, 1519, 1467, 1446, 1423, 1386, 1366, 1340, 1302, 1251, 1171, 1110, 1093, 1032, 957, 909, 865, 789, 704, 665, 649, 621, 599, 418.

[C₂C₂im][Br] FTIR (υ (cm⁻¹)): 3129, 3056, 2976, 2938, 2796, 2730, 2608, 1563, 1544, 1447, 1408, 1384, 1347, 1295, 1247, 1165, 1122, 1087, 1028, 956, 821, 802, 755, 688, 625, 592.

Synthesis of [C₂C_nim][OAc]: [C₂C_nim][Br] (50 mmol, 1.0 equiv.) and sodium acetate (4.10 g, 50 mmol, 1.0 equiv.) were dissolved in acetone (50 mL) in a round bottom flask. The mixture was stirred overnight then filtrated. The filtrate was evaporated under reduced pressure to yield to a colorless oil.

[C₂mim][OAc] FTIR (υ (cm⁻¹)): 3139, 2975, 3030, 2866, 2748, 1560, 1451, 1427, 1377, 1328, 1256, 1174, 1090, 1002, 1034, 958, 899, 804, 702, 599, 454.

[C₂C₂im][OAc] FTIR (υ (cm⁻¹)): 3422, 3136, 3064, 2978, 2933, 1562, 1410, 1165, 1123, 1090, 1045, 1009, 956, 921, 830, 802, 760, 688, 642, 622, 527, 461.

General procedure for the synthesis of anhydrous lanthanide compounds: In a round‐bottom flask were introduced LnCl₃ ⋅ xH₂O (x=6 or 7; 1.0 mmol, 1.0 equiv.) and [C₂mim][OAc] (850 mg, 5.0 mmol, 5.0 equiv.). The mixture was heated at 100 °C under stirring overnight. The flask was then transferred to a pre‐heated oven at 100 °C. A program was run to slowly decrease the temperature (1 °C h⁻¹) to 25 °C to allow a better crystallization of the final products. The final complexes have been obtained with yields 80 % and higher.

La(OAc)₃(H₂O). White solid. FTIR (υ (cm⁻¹)): 3262, 1601, 1523, 1402, 1335, 1052, 1015, 942, 663, 614, 465, 419.

Pr(OAc)₃(H₂O). Light green solid. FTIR (υ (cm⁻¹)): 3273, 1682, 1531, 1395, 1341, 1170, 1052, 1024, 1007, 956, 943, 748, 670, 642, 610, 502, 477, 421.

[C₂mim] _n [{Nd₂(OAc)₇} _n ]. White solid. FTIR (υ (cm⁻¹)): 3148, 3119, 3090, 3079, 2987, 2929, 1634, 1569, 1391, 1335, 1254, 1167, 1121, 1090, 1048, 1015, 940, 867, 808, 774, 646, 620, 500. ESI TOF m/z (negative mode) 700.4110 (calcd m/z 700.9145).

[C₂mim] _n [{Sm₂(OAc)₇} _n ]. White solid. FTIR (υ (cm⁻¹)): 3142, 3117, 3098, 3079, 2975, 2931, 1609, 1554, 1402, 1336, 1247, 1166, 1118, 1086, 1047, 1013, 942, 869, 808, 775, 641, 666, 611, 489. ESI TOF m/z (negative mode) 718.6951 (calcd m/z 718.9356).

[C₂mim] _n [{Eu₂(OAc)₇} _n ] and [C₂mim]₂[Eu₂(OAc)₈] (mixture). White solid. FTIR (υ (cm⁻¹)): 3380, 3143, 3093, 2999, 2984, 2943, 1605, 1565, 1432, 1404, 1335, 1300, 1249, 1160, 1095, 1045, 1015, 953, 932, 868, 787, 680, 666, 642, 626, 613, 504, 469. [C₂mim] _n [{Eu₂(OAc)₇} _n ] ESI TOF m/z (negative mode) 716.6241 (calcd m/z 716.9348). [C₂mim]₂[Eu₂(OAc)₈] ESI TOF m/z (negative mode) 387.6951 (calcd m/z 387.9743).

[C₂mim] _n [{Gd₂(OAc)₇} _n ]. White solid. FTIR (υ (cm⁻¹)): 3143, 3093, 2984, 1606, 1566, 1434, 1405, 1336, 1300, 1161, 1095, 1045, 1016, 954, 933, 868, 787, 667, 651, 643, 626, 613, 505, 469. ESI TOF m/z (negative mode) 730.6341 (calcd m/z 730.9448).

[C₂mim] _n [{Tb₂(OAc)₇} _n ] and [C₂mim]₂[Tb₂(OAc)₈] (mixture). Brown solid. FTIR (υ (cm⁻¹)): 3550, 3526, 3360, 3150, 3109, 2994, 2932, 1643, 1590, 1543, 1407, 1351, 1338, 1162, 1099, 1054, 1018, 949, 939, 861, 787, 742, 691, 668, 643, 615, 506, 484, 468, 453, 425. [C₂mim] _n [{Tb₂(OAc)₇} _n ] ESI TOF m/z (negative mode) 730.6166 (calcd m/z 730.9444). [C₂mim]₂[Tb₂(OAc)₈] ESI TOF m/z (negative mode) 394.6952 (calcd m/z 394.9791).

[C₂mim]₂[Dy₂(OAc)₈]. White solid. FTIR (υ (cm⁻¹)): 3546, 3372, 3238, 3150, 3109, 3002, 2933, 1592, 1544, 1408, 1351, 1162, 1099, 1053, 1018, 950, 940, 900, 861, 788, 692, 669, 644, 616, 508, 470, 441, 428, 417. ESI TOF m/z (negative mode) 398.6946 (calcd m/z 398.9817).

[C₂mim]₂[Ho₂(OAc)₈]. Light pink solid. FTIR (υ (cm⁻¹)): 3142, 3092, 3002, 2985, 2945, 1609, 1568, 1435, 1406, 1347, 1335, 1301, 1248, 1161, 1095, 1045, 1015, 957, 934, 871, 808, 788, 751, 682, 668, 644, 607, 508, 466. ESI TOF m/z (negative mode) 400.7019 (calcd m/z 400.9841).

[C₂mim]₂[Er₂(OAc)₈]. Light pink solid. FTIR (υ (cm⁻¹)): 3385, 3143, 3092, 2985, 1608, 1566, 1404, 1335, 1301, 1249, 1161, 1094, 1045, 1016, 958, 934, 871, 789, 682, 668, 645, 614, 509, 469, 426. ESI TOF m/z (negative mode) 402.6986 (calcd m/z 402.9851).

[C₂mim]₂[Tm₂(OAc)₈]. White solid. FTIR (υ (cm⁻¹)): 3396, 3154, 3118, 2993, 1560, 1400, 1347, 1169, 1048, 1016, 981, 963, 945, 905, 867, 845, 670, 652, 617, 516, 496, 471. ESI TOF m/z (negative mode) 404.6964 (calcd m/z 404.9880).

[C₂mim]₂[Yb₂(OAc)₈]. White solid. FTIR (υ (cm⁻¹)): 3388, 3153, 3090, 2995, 1618, 1570, 1404, 1345, 1299, 1249, 1169, 1097, 1047, 1018, 959, 930, 878, 793, 687, 671, 643, 618, 512, 474, 425. ESI TOF m/z (negative mode) 409.0997 (calcd m/z 409.4923).

[C₂mim]₂[Lu₂(OAc)₈]. White solid. FTIR (υ (cm⁻¹)): 3384, 1638, 1540, 1453, 1412, 1169, 1049, 1022, 957, 839, 759, 688, 642, 613, 509. ESI TOF m/z (negative mode) 410.6987 (calcd m/z 410.9945).

Compounds obtained with a 3 : 1 (IL/Ln‐ratio): Nd(OAc)₃(H₂O). Light purple solid. FTIR (υ (cm⁻¹)): 3276, 1682, 1598, 1528, 1395, 1341, 1052, 1024, 1006, 956, 942, 863, 744, 670, 642, 609, 571, 504, 477.

Sm(OAc)₃(H₂O). White solid. FTIR (υ (cm⁻¹)): 3336, 1705, 1531, 1415, 1263, 1051, 1018, 951, 674, 643, 612, 501, 464.

Compounds obtained with a 10 : 1 (IL/Ln‐ratio): [C₂mim]₂[Pr₂(OAc)₈]. Light green solid. ESI TOF m/z (negative mode) 376.6848 (calcd m/z 376.9609).

[C₂mim]₂[Nd₂(OAc)₈]. Light purple solid. ESI TOF m/z (negative mode) 378.6852 (calcd m/z 378.9627).

[C₂mim]₂[Sm₂(OAc)₈]. White solid. FTIR (υ (cm⁻¹)): 3144, 3094, 2999, 2984, 2931, 1602, 1564, 1430, 1404, 1336, 1300, 1249, 1160, 1095, 1045, 1015, 951, 930, 866, 786, 679, 665, 641, 626, 613, 502, 469, 429. ESI TOF m/z (negative mode) 388.6932 (calcd m/z 388.9747).

General procedure for the synthesis of [C₂C₂im][La₂(OAc)₈]: In a round‐bottom flask were introduced LaCl₃ ⋅ 7H₂O (371 mg, 1.0 mmol, 1 equiv.), [C₂C₂im][OAc] (1.840 g, 10 mmol, 10 equiv.), and water (5 mL). The mixture was heated at 100 °C under stirring overnight. The flask was then transferred to a pre‐heated oven at 100 °C. A program was run to slowly decrease the temperature (1 °C h⁻¹) to allow a better crystallization of the final product.

Instrumentation: PXRD data were recorded at ambient temperature on a PANalytical X'pert PRO diffractometer (Malvern Panalytical, Malvern, UK), operating at 45 kV and 40 mA and using Cu_Kα1 radiation. The data were recorded in reflection mode from 5° to 70° with a step size of 0.01° for 60 min.

The data for single crystal analysis at Stockholm University were collected on a Bruker Venture diffractometer (Bruker AXS, Karlsruhe, Germany) using Cu_Kα radiation (λ=1.54178 Å at 297 K) at room temperature or at low temperature (100 K). The crystal structures were solved and refined using SHELXT and SHELXL subroutines within the APEX3 software package. ^[24] Absorption corrections were carried out with the program SADABS. Hydrogen atoms for all structures were assigned geometrically using a riding atom model. To illustrate the crystal structures, the program Diamond 3 (CrystalImpact, Bonn, Germany) was used. ^[25]

The crystal structural measurements at The University of Alabama were carried out on a Bruker SMART APEX II CCD diffractometer (Mo_Kα radiation; λ=0.71073 Å) at 100 K. The intensity data of the reflections and scaling were integrated using SAINT within the APEX3 software package. ^[26] SADABS ^[27] was used for the absorption corrections. The crystal structure solution was performed by direct methods using SHELXT. ^[28] SHELXL‐97 ^[26] and SHELXL‐2013 ^[24c] were used for subsequent difference Fourier analyses and least‐squares refinement. All non‐hydrogen atoms were refined anisotropically, whereas hydrogen atoms were assigned geometrically using a riding atom model.

Differential scanning calorimetry (DSC) was performed with a computer‐controlled PhoenixDSC 204 F1 thermal analyzer (Netzsch, Selb, Germany). Measurements were carried out at a heating rate of 5 °C min⁻¹ from −60 to 210 °C in sealed aluminum crucible with a nitrogen flow rate of 40 mL min⁻¹. The samples were placed in aluminum pans which were cold‐sealed and punctured. Before recording the DSC thermogram the samples were cooled to −60 °C. Given temperatures correspond to the onset.

A SYNAPT G2‐S HDMS Q‐ToF Mass Spectrometer (Waters, Manchester, United Kingdom) with an electrospray ionization (ESI) operated in the positive and negative ion mode, was used in this study. The ion source was set up as follows: capillary voltage: 2500 V, extractor: 1.0 V, RF lens: 0.5 V, ion source temperature: 120 °C and desolvation temperature 250 °C. Nitrogen was used as both the cone and desolvation gas at a flow of 70 L h⁻¹ and 500 L h⁻¹, respectively. Argon was used as a collision gas at a pressure of 2.95 ⋅ 10⁻⁴ mbar.

The infrared spectroscopy (IR) was conducted with a Bruker Alpha‐P ATR‐spectrometer (Bruker AXS) in attenuated total reflection configuration in the range of 4000–400 cm⁻¹. The data evaluation was carried out with the program OPUS (Bruker, Ettlingen, Germany).

Optical analyses were made by heated‐stage polarized optical microscopy (POM) with an Axio Imager A1 microscope (Carl Zeiss Micro Imaging GmbH, Göttingen, Germany) equipped with a hot stage, THMS600 (Linkam Scientific Instruments Ltd, Surrey, UK), and Linkam TMS 94 temperature controller (Linkam Scientific Instruments Ltd). Images were recorded at a magnification of 100× as a video with a digital camera. During heating and cooling, the sample was placed between two cover slips. Heating and cooling rates were 10 °C min⁻¹.

Deposition numbers 1920416, 1920417, 1920418, 1920419, 1920420, 1920421, 1920422, 1920423, 1920424, 1920425, 1920426, 1920427, 1920428, 1920432, 1920433, 1920434, and 1920435 contain the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service www.ccdc.cam.ac.uk/structures.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

G. Bousrez, O. Renier, S. P. Kelley, B. Adranno, E. Tahavori, H. M. Titi, V. Smetana, S.-F. Tang, A.-V. Mudring, R. D. Rogers, Chem. Eur. J. 2021, 27, 13181.