Melting point suppression in new lanthanoid(III) ionic liquids by trapping of kinetic polymorphs: an in situsynchrotron powder diffraction study

Abstract

The complexes (N₄₄₄₄)₃[Ln(dcnm)₆] (Ln = La–Nd, Sm; N₄₄₄₄ = tetrabutylammonium) display a decrease in the melting point upon fast cooling from a melt, which is shown by in situsynchrotron based X-ray powder diffraction to be due to the formation of a second, less thermodynamically stable, polymorph.

---

The inclusion of lanthanoids in ionic liquids (ILs) offers an ideal route to incorporate their unique luminescent and magnetic properties into a bulk solution.^1,2 However, lanthanoid compounds often exhibit a poor solubility in commonly used ILs because the IL anions are typically very weakly coordinating, prohibiting the simple dissolution of a lanthanoid salt by complexation to any beneficial extent.² One strategy used to achieve high lanthanoid concentrations is to incorporate the lanthanoid cation directly into an anion that will form ILs.³ This is best accomplished by employing a ligand which readily coordinates to a lanthanoid atom and possesses properties, such as a highly delocalised charge, that are conducive to forming ILs.^4,5 This has been demonstrated by the synthesis of lanthanoid containing ILs with the bis(trifluoromethanesulfonyl)amide³ and thiocyanate ligands,⁶ with the latter being included in the ILs (C₆mim)_5−x[Dy(SCN)_8−x(H₂O)_x] (x = 0–2; C₆mim = 1-hexyl-3-methylimidazolium), which exhibit a stronger response to a magnetic field than previously investigated transition metal based ILs.⁷ Even simple halide salts such as (C₄mim)X (C₄mim = 1-butyl-3-methylimidazolium; X = Cl, Br) in combination with lanthanoid(III) halides gave rise to ionic liquids.⁸

In investigations of ligand systems related to pseudohalides the dicyanonitrosomethanide (dcnm) anion (Fig. 1) has been shown to readily coordinate to lanthanoid(III) cations via η²nitroso groups, to form the 12-coordinate homoleptic complexes [Ln(dcnm)₆]³⁻ (Ln = La–Gd).⁹ The melting point (126 °C) of (N₂₂₂₂)₃[La(dcnm)₆] (N₂₂₂₂ = tetraethylammonium) suggested that an analogue with a larger cation would have a lower melting point, leading to new lanthanoid based ILs.

Fig. 1 The [Ce(dcnm)₆]³⁻ anion from the crystal structure of (N₄₄₄₄)₃[Ce(dcnm)₆], and the dicyanonitrosomethanide (dcnm) anion.

Here we report the first ILs with the hexakis(dicyanonitrosomethanido)lanthanoid trianion, namely (N₄₄₄₄)₃[Ln(dcnm)₆] (Ln = La, Ce, Pr, Nd, Sm), and examine the relationship between polymorphism and thermal behavior with differential scanning calorimetry (DSC) and synchrotron based in situX-ray powder diffraction.

The complexes (N₄₄₄₄)₃[Ln(dcnm)₆] were synthesised by a metathesis reaction between the silver salt of dicyanonitrosomethanide, the requisite lanthanoid chloride, and tetrabutylammonium iodide in alcoholic reaction media. After filtration of the silver halide precipitate, and partial evaporation of the filtrate under vacuum, standing of the solution at −50 °C resulted in crystallization of the target complexes. In the case of Ce and Sm single crystals of sufficient quality for single crystal X-ray diffraction analysis were obtained. The compounds were isomorphous and crystallized in the space groupR32 (no. 155) (Fig. 1).

The melting points of the complexes were determined by DSC and all qualify as ILs (Table 1; Table S1, Fig. S1–S5, ESI†) as their melting points are below 100 °C. However, a substantial deviation between the melting and crystallization temperature exists, with supercooling ranges of ca. 50 °C. Furthermore, when the material that had been heated and recrystallized (from the melt) was heated for a second time the DSC measurements indicated that there was a decrease in the melting point of the materials (Table 1). These lower melting points still give supercooling ranges of ca. 40 °C.

Table 1 Thermal behaviour of complexes (N₄₄₄₄)₃[Ln(dcnm)₆] as measured by DSC^a

Complex	Cycle	Heating		Cooling
		T m/°C	ΔH/kJ mol^−1	T c/°C	ΔH/kJ mol^−1
a Melting point (Tm), temperature of recrystallization (Tc). Each sample was heated from −20 °C to 120 °C and then cooled back to −20 °C (Cycle 1), and then the heating program was repeated (Cycle 2). DSC measurements of Sm contained in ESI.†
La	1	84.3	48.1	30.6	−24.0
	2	66.7	25.4	30.6	−23.3
Ce	1	84.5	61.8	33.1	−24.5
	2	67.8	33.1	26.3	−23.7
Pr	1	81.2	54.3	32.7	−26.7
	2	72.6	34.3	31.7	−25.8
Nd	1	77.9	47.6	26.2	−28.7
	2	70.2	29.9	28.5	−28.6

---

Characterization by infrared and NMR spectroscopy of the complexes recrystallized from the melt showed that the chemical composition of the material remained unchanged during the heating and cooling processes, eliminating the possibility of an adverse chemical reaction being responsible for the change in the melting point of the recrystallized material. The thermal behaviour cannot be attributed to a loss of the lattice solvent as none is present. The enthalpies of crystallization measured in the first and subsequent cooling cycles were unchanged. However, the enthalpy of fusion determined upon heating in the first cycle was about 20 kJ mol⁻¹ higher than in the following cycles. This suggests that a thermodynamically less stable polymorph may be obtained by crystallization in the DSC. Such a phenomenon has previously been observed for the simple IL (C₄mim)Cl, where a difference in melting points of 25 °C for the two polymorphs was observed.¹⁰ This effect has been attributed to the butyl group of the cation changing conformation, causing a change in the packing and hydrogen bonding within the crystal structure that lowers the melting point of the IL.

X-Ray powder diffraction was employed to determine whether the material indeed had crystallized as a different polymorph, leading to a change in the melting point of the material. To study whether different polymorphs form upon crystallization from a solvent or the melt, crystals of (N₄₄₄₄)₃[Pr(dcnm)₆] were melted under vacuum and then allowed to crystallize upon slow cooling to room temperature under ambient conditions. The powder pattern of the sample that had recrystallized from this melt was the same as that of the sample which had crystallized from the alcoholic reaction solution during the initial synthesis. Thus a solvent effect can be excluded and it suggests that it is rather the rate of crystallization which determined which polymorph of the IL is formed.

This assumption was confirmed by temperature dependent in situsynchrotron X-ray powder diffraction measurements on (N₄₄₄₄)₃[Pr(dcnm)₆]. Ground crystals were sealed in a glass capillary and heated and cooled at a rate of 5 K min⁻¹ by a cryostream, to replicate the conditions under which the DSC measurements were performed. The high intensity of the X-ray beam of the synchrotron source facility allowed the data to be collected in situ while the sample was being constantly heated and cooled, with no pause in the temperature ramping. Typical temperature dependent laboratory X-ray powder diffraction measurements would have resulted in equilibration at a given temperature during data collection, prohibiting the trapping of the thermodynamically unstable polymorph.

The temperatures for the onset of the phase transitions recorded during the in situX-ray powder diffraction experiment were in accord with the DSC measurements (Fig. 2). The temperatures recorded for the solid to liquid phase transition were one to three degrees Celsius higher when measured by PXRD than by DSC; this is due to the PXRD measurements being recorded over a 20 second period while the temperature of the sample is still increasing, introducing an unavoidable, albeit minor, degree of error in the temperature measurements. There was a distinct change in the powder diffraction pattern of the recrystallized sample of (N₄₄₄₄)₃[Pr(dcnm)₆], indicating that a new polymorph was formed (Fig. 2). The reflections of the thermodynamically stable phase could be indexed to a hexagonal cell with unit parameters of a = b = 17.8437(1) Å and c = 22.2154(1) Å, which correspond to the unit cell parameters from the single crystal structures of (N₄₄₄₄)₃[Ln(dcnm)₆] (Ln = Ce, Sm) taking the lanthanoid contraction into account (Fig. S6, ESI†). Furthermore, the PXRD pattern generated from the single crystal structure of (N₄₄₄₄)₃[Ce(dcnm)₆] was found to correlate with the experimental PXRD pattern for the thermodynamically stable polymorph of (N₄₄₄₄)₃[Pr(dcnm)₆] (Fig. S7, ESI†), confirming that the two complexes are isomorphous.

Fig. 2 A 2D plot of the synchrotron PXRD patterns of (N₄₄₄₄)₃[Pr(dcnm)₆]; darker areas represent peaks of higher intensity. The sample was heated and cooled at 5 K min⁻¹. Temperatures on the right of the plot mark the onset of a transition. Measurements commenced and concluded at 0 °C. (b) The different phases of (N₄₄₄₄)₃[Pr(dcnm)₆] and the temperatures at which the phase transitions occurred during the heating and cooling of the sample during in situ measurements of the XPRD patterns.

The unit cell parameters indexed from the PXRD patterns recorded at various temperature intervals in the first heating cycle demonstrated a lengthening of the a and bcell axes and a slight shortening of the ccell axis, resulting in an overall increase in the unit cell volume, corresponding to a thermal expansion of the material (Table S2, Fig. S10–S12, ESI†). Once molten the compound crystallized as a different polymorph of lower symmetry, as indicated by the more complex PXRD patterns (Fig. 2; Fig. S8, ESI†). Single crystals suitable for SXRD could not be obtained. Given that a change in the conformation of the cation in (C₄mim)Cl induced an adjustment in the packing in the crystal structure of the IL, it is plausible that a similar phenomenon occurs in the system reported here. The possible conformational change of the butyl groups of the (N₄₄₄₄)⁺ cation is a source of structural variability, which may give rise to a new polymorph for the complex, consequently affecting the melting point of the material.

In summary the new complexes, (N₄₄₄₄)₃[Ln(dcnm)₆] (Ln = La–Nd, Sm), were shown by DSC to be ILs. When cooled at 5 K min⁻¹, the complexes recrystallized from the melt as thermodynamically less stable polymorphs, as demonstrated by in situsynchrotron based X-ray powder diffraction. These kinetically trapped polymorphs possessed lower melting points than the more thermodynamically stable polymorphs.

We acknowledge financial support by the Australian Research Council (S.R.B. and G.B.D.), the Deutscher Akademischer Austausch Dienst (DAAD) for a Research Grant for Doctoral Candidates and Young Academics and Scientists and the Australian Institute of Nuclear Science and Engineering Inc. (AINSE) for a Post-graduate Research Award (A.S.R.C.). A.-V.M. thanks the Fonds der Chemischen Industrie for a Dozentenstipendium and the DFG for support in the priority program 1166 “Lanthanide specific functionalities”. This research was undertaken on the powder diffraction beamline (10BM1) at the Australian Synchrotron, Victoria, Australia, under beamtime award AS101/PD2296.

Notes and references

1. Y. Yoshida and G. Saito, Phys. Chem. Chem. Phys., 2010, 12, 1675.
2. A.-V. Mudring and S.-F. Tang, Eur. J. Inorg. Chem., 2010, 2569.
3. (a) S.-F. Tang, A. Babai and A.-V. Mudring, Angew. Chem., Int. Ed., 2008, 47, 7631; (b) A. Babai and A.-V. Mudring, Z. Anorg. Allg. Chem., 2008, 634, 938; (c) A. Babai and A.-V. Mudring, J. Alloys Compd., 2006, 418, 122; (d) A. Babai and A.-V. Mudring, Dalton Trans., 2006, 1828; (e) A. Babai and A.-V. Mudring, Chem. Mater., 2005, 17, 6230; (f) A.-V. Mudring, A. Babai, S. Arenz and R. Giernoth, Angew. Chem., Int. Ed., 2005, 44, 5485.
4. (a) S.-F. Tang, J. Cybinska and A.-V. Mudring, Helv. Chim. Acta, 2009, 92, 2375; (b) S.-F. Tang and A.-V. Mudring, Eur. J. Inorg. Chem., 2009, 2769.
5. A.-V. Mudring, Proceed. ISEC, 2008, 2, 1271.
6. P. Nockemann, B. Thijs, N. Postelmans, K. V. Hecke, L. V. Meervelt and K. Binnemans, J. Am. Chem. Soc., 2006, 128, 13658.
7. B. Mallick, B. Balke, C. Felser and A.-V. Mudring, Angew. Chem., Int. Ed., 2008, 47, 7635.
8. (a) R. E. Del Sesto, T. M. McCleskey, A. K. Burrell, G. A. Baker, J. D. Thompson, B. L. Scott, J. S. Wilkes and P. Williams, Chem. Commun., 2008, 447; (b) C. C. Hines, D. B. Cordes, S. T. Griffin, S. I. Watts, V. A. Cocalia and R. D. Rogers, New J. Chem., 2008, 32, 872; (c) B. Mallick, Dissertation, Universität zu Köln, 2009; (d) A. Babai and A.-V. Mudring, Inorg. Chem., 2006, 45, 4874; (e) A. Babai and A.-V. Mudring, Inorg. Chem., 2005, 44, 8168.
9. (a) A. S. R. Chesman, D. R. Turner, E. I. Izgorodina, S. R. Batten and G. B. Deacon, Dalton Trans., 2007, 1371; (b) A. S. R. Chesman, D. R. Turner, G. B. Deacon and S. R. Batten, Eur. J. Inorg. Chem., 2010, 18, 2798.
10. J. D. Holbrey, W. M. Reichert, M. Nieuwenhuyzen, S. Johnson, K. R. Seddon and R. D. Rogers, Chem. Commun., 2003, 1636.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis, characterisation and DSC measurements of complexes Ln. CCDC 790887 and 813884. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c1cc14744g

---