Synthesis, characterization and thermal behaviour of solid phases in the quasi-ternary system Mg(SCN)₂–H₂O–THF

Abstract

Mg(SCN)₂·4H₂O can be converted into previously unknown compounds Mg(SCN)₂·(4 − x) H₂O·xTHF with x = 0, 2 and 4 by multiple recrystallization in tetrahydrofuran (THF). The phases were characterized by infrared spectroscopy (IR), thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC), and their crystal structures were solved from X-ray powder diffraction (XRPD) data. In the crystal structures isolated Mg(NCS)₂(H₂O)_4−x(THF)_x units form layered motifs. The thermal behavior of Mg(SCN)₂·4H₂O and Mg(SCN)₂·4THF was investigated by temperature dependent in situ XRPD, where Mg(SCN)₂·4THF was found to acquire a room temperature (α-form) and high temperature modification (β-form). The phase transformation is associated with an order–disorder transition of the THF molecules and with a reversion of the stacking order of the layered motifs. Further heating eventually leads to the formation of Mg(SCN)₂·2THF. There thiocyanate related sulfur atoms fill the voids in the coordination sphere of magnesium, which leads to the formation of one dimensional electroneutral _∞[Mg(NCS)_2/2(SCN)_2/2(THF)₂] chains. All investigated Mg(SCN)₂·(4 − x) H₂O·xTHF phases exhibit a remarkable anisotropic thermal expansion, and Mg(SCN)₂·4H₂O and Mg(SCN)₂·2THF were found to show both positive and negative thermal expansion coefficients.

---

Introduction

Magnesium halides and pseudo-halides exhibit a great variety of chemical and structural properties and are used in many different applications, which is why they have been extensively studied.^1–5 However, in the row of magnesium pseudo-halides, i.e. cyanides, cyanates, azides etc., thiocyanates are often overseen, and only little chemical and structural knowledge is available for magnesium thiocyanates and their derivatives.

The first known crystalline magnesium thiocyanate phase was reported to be the tetrahydrate Mg(SCN)₂·4H₂O in 1930,⁶ which was later confirmed by the elucidation of its crystal structure.⁷ In addition, some reports have mentioned the existence of related phases, such as the pentahydrate Mg(SCN)₂·5H₂O or anhydrous Mg(SCN)₂;^8–11 however, only little to no evidence was presented to support this. Several studies showed that Mg²⁺ when coordinated to SCN⁻ anions seems to be capable to form a diversity of solid ionic complexes with different organic molecules such as pyridine (Py), nicotinamide (NIA) and their derivatives,^10–13 although their crystal structures remain unknown.

Besides the tetrahydrate Mg(SCN)₂·4H₂O, crystal structure investigations of Mg(SCN)₂-containing compounds have generally been neglected so far and only spectroscopic studies or thermal analyses are present.^10,12 Further structural studies of related Mg²⁺-SCN⁻-phases are only available for mixed cationic systems involving other cations like K⁺ or Cd²⁺.^14,15 In contrast, other MgX₂ salts with similar anions (X = CN, N(CN)₂) are well known in terms of crystal structures and properties.^16,17

Furthermore, unlike many other thiocyanate salts M(SCN)_n (M = Li, Na, K, Ca, Sr, Ba, Ni, Ag, Hg, …, n = 1 or 2),^18–25 it is yet unclear how to obtain the anhydrous form of Mg(SCN)₂ and under what conditions it would be stable. As mentioned before, some studies suggest the formation of anhydrous Mg(SCN)₂ by simple heating of Mg(SCN)₂·x [L] complexes ([L] = H₂O, pyridine, …);¹⁰ however, the phase formation was not analytically confirmed in any case. In fact, a more likely dehydration mechanism might be similar to that of MgCl₂, which forms a wide variety of hydrates.^1,2,26–28 The hexahydrate MgCl₂·6H₂O is a commonly occurring mineral named bischofite and finds a diversity of industrial usage.^1–4,29 Upon heating of MgCl₂ hydrates different dehydration stages occur; however, the anhydrous state cannot simply be reached as the formation of oxychloride species prevails.^2,3,27 Solvent based methods or rather tedious thermochemical procedures are necessary to produce pure, anhydrous MgCl₂.³

Besides the rich hydrate chemistry of MgCl₂, it also forms different organometallic complexes, which are of great importance in organic chemistry (such as Grignard reagents or for Ziegler–Natta catalysis)^30–33 and have more recently even sparked interest in the battery community.³⁴ These complexes are formed by the inclusion of organic molecules such as pyridine (Py), dimethoxyethane (DME), tetrahydrofuran (THF) etc., and a lot of research studies are devoted to their structural analysis, especially to understand the coordination between metals and organic molecules.^30,33–38

Moving on from halides to pseudo-halides, the previously mentioned coordination chemistry between metals and organic molecules is still a rather unexplored field for crystalline, solid thiocyanates. Some examples of solvent complexes with dimethoxyethane (DME), tetrahydrofuran (THF) and acetonitrile (AN) can be found for rare earth metal thiocyanates and Ca(SCN)₂.³⁹ There also exist coordination compounds with acetone (Ac) of transition metals (TM = Cr, Mn and Fe);⁴⁰ yet for Mg(SCN)₂, only little is known about these kinds of complexes and no structural data are available so far.

In this study we report the synthesis, structure, spectroscopic properties and thermal behavior of the solid phases dithiocyanatetetraaquamagnesium(II) [Mg(SCN)₂(H₂O)₄] (also referred to as magnesium dithiocyanate tetrahydrate, Mg(SCN)₂·4H₂O), bis(tetrahydrofuran)diaquadi-thiocyanatemagnesium(II), [Mg(SCN)₂(H₂O)₂(THF)₂], and dithiocyanatetetrakis(tetrahydrofuran)-magnesium(II), [Mg(SCN)₂(THF)₄] (from hereon we will address these compounds as Mg(SCN)₂·(4 − x) H₂O·xTHF with x = 0, 2, 4).⁴¹ The thermal behaviour of the title compounds was analysed in detail by temperature dependent in situ X-ray powder diffraction and a strong anisotropic lattice expansion as well as several phase transitions of Mg(SCN)₂·4THF eventually leading to the formation of Mg(SCN)₂·2THF were observed.

Experimental

Mg(SCN)₂·(4 − x) H₂O·xTHF (x = 0, 2, 4) phases were synthesized by recrystallization in H₂O or THF or by ion exchange. The obtained products were characterized by X-ray powder diffraction (XRPD), infrared (IR) spectroscopy, thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), scanning electron microscopy (SEM) and energy dispersive X-ray (EDX) analysis. The handling and measurements of the samples were conducted under dry inert gas or vacuum unless stated otherwise. All details of the synthesis and analysis can be found in the ESI.†

Results

Synthesis and characterization

Syntheses of Mg(SCN)₂·4H₂O have been reported in a number of different publications employing various synthesis routes, such as:^{6–8,10–12}

The third synthesis route is employed in this study. However, it was discovered that after the removal of ethanol, the formation of Mg(SCN)₂·4H₂O crystals was severely inhibited, and precipitation required long vacuum treatment at rather high temperatures (∼150 °C) with the possible formation of decomposition products. A much easier way to produce crystalline Mg(SCN)₂·4H₂O without risking decomposition was, after removing most of the ethanol, to re-dissolve the obtained slurry in water and then use freeze drying to precipitate the crystals of Mg(SCN)₂·4H₂O. Nonetheless, the purity of Mg(SCN)₂·4H₂O for both the synthesis routes 2 and 3 is limited due to the solubility product of BaSO₄ in water (1.3 × 10^–5 mol L⁻¹)⁴² and KCl in ethanol (3.9 × 10⁻⁵ mol L⁻¹).⁴³ This problem does not occur in the synthesis route 1; however, additional heat treatments or neutralization is required to entirely remove NH₃ or CO₂. To avoid the limits created by these methods, we used cation exchange via an ion exchanger, which directly yielded pure Mg(SCN)₂·4H₂O without further treatments being necessary. Low requirements on the metal-base purity of the precursor KSCN are an additional advantage of this method, as impurity cations contained within the commercially available precursor material will be replaced by Mg²⁺ cations as well.

Both the syntheses of Mg(SCN)₂·2H₂O·2THF and Mg(SCN)₂·4THF required an additional re-crystallization step using THF as the solvent. It was difficult to control the ratio of H₂O to THF, especially for Mg(SCN)₂·2H₂O·2THF, and several synthesis attempts produced a mixture of both Mg(SCN)₂·2H₂O·2THF and Mg(SCN)₂·4THF (Fig. 1). In the case of Mg(SCN)₂·4THF, this problem could be overcome by simply refluxing the solution in THF to remove any remaining H₂O, although this might have caused the formation of an amorphous side phase (more details can be found in the ESI†). With the above described methods three phases with x = 0, 2, 4 in the ternary system Mg(SCN)₂·(4 − x) H₂O·xTHF were successfully synthesized, all of which crystallize in a monoclinic lattice (Table 1).

Fig. 1 XRPD patterns of the series Mg(SCN)₂·(4 − x) H₂O·xTHF showing different syntheses products; pure phases (black) and mixed phases (colored).

Table 1 Crystallographic data for all Mg(SCN)₂·(4 − x) H₂O·xTHF phases

	x = 0	x = 2	x = 4
Space group	P21/a	C2/m	P21/n
a/Å	7.4947(1)	10.7474(2)	15.2137(8)
b/Å	9.0296(1)	8.1893(1)	10.2054(4)
c/Å	7.8721(1)	9.6299(2)	7.7220(3)
β/°	113.650(1)	93.153(1)	94.980(2)
V/Å	488.00(1)	846.29(3)	1194.39(9)
Z	2	2	2

---

The exchange of water with THF molecules leads to fundamental changes in the crystal lattice. From x = 0 to x = 2, the volume per formula unit strongly increases by 179.15 ų, illustrated by the strong downshift of the 001 reflection. This increase is smaller than the volume of two free THF molecules (269 ų). The SEM images in Fig. S4a and b† show that the grain morphologies become more platelet like, which indicated a more pronounced layered motif in the crystal structure after the THF incorporation. When the THF content is increased from x = 2 to x = 4, the volume per formula unit increases by 174.05 ų, which is very similar to the first exchange step. The C-centered monoclinic unit cell becomes primitive (space group P2₁/n for x = 4) upon complete exchange.

Furthermore, the grains adopt a rod-shaped morphology with well-defined crystal faces, as can be seen in Fig. S4c.† The EDX results of x = 4 shown in Fig. S5† match the proposed stoichiometry with an average Mg : S ratio of 1 : 2.

IR spectroscopy was employed to confirm the presence of thiocyanate and THF as well as to derive information on the coordination of Mg²⁺. Fig. 2 shows selected excerpts of the IR spectra for all Mg(SCN)₂·(4 − x) H₂O·xTHF phases.

Fig. 2 IR spectra of Mg(SCN)₂·(4 − x) H₂O·xTHF phases with x = 0 (black, measured in Ar), x = 2 (blue, measured in air), x = 4 (orange, measured in Ar).

The spectra show THF and water related bands situated in the high wavenumber region between 2800 and 3600 cm⁻¹. For Mg(SCN)₂·4H₂O and Mg(SCN)₂·4THF the bands of THF and H₂O are respectively absent (Fig. 2, black and orange line), while the spectrum of Mg(SCN)₂·2H₂O·2THF shows a combination of both (Fig. 2, blue line; further details can be found in Table 2). Interestingly, the characteristic bands ν(C≡N) (∼2080 cm⁻¹) and δ(H–O–H) (∼1630 cm⁻¹) of the SCN⁻ anion and H₂O are split for Mg(SCN)₂·4THF and Mg(SCN)₂·2H₂O·2THF, respectively. In the case of the SCN⁻ anion, such a splitting was also observed for the dihydrate Ca(SCN)₂·2H₂O,²¹ the trihydrate K₂Mg₂(SCN)₆·3H₂O,¹⁴ and various rare earth metal thiocyanate complexes with different solvent molecules.³⁹ In the last case, the splitting was attributed to the bridging nature of SCN⁻ to two cations, enabling both symmetric and asymmetric ν(C≡N) modes. In contrastry, here SCN⁻ does not have a bridging function (see below), and it is more likely that the splitting is connected to the complexity of molecules coordinated to the central cation in the vicinity of SCN⁻. Since this splitting does not appear in the bands of Mg(SCN)₂·4H₂O, it can be attributed to the coordination of THF molecules to Mg²⁺, inducing polarization effects either in H₂O or SCN⁻. However, an unambiguous explanation requires a more rigorous treatment of the spectroscopic features of these phases and is beyond the scope of this report.

Table 2 Wavenumbers of IR spectra in Fig. 2 for all Mg(SCN)₂·(4 − x) H₂O·xTHF phases with tye respective assignments according to the literature^{18,21,44–46}

x = 0; ν/cm^−1	x = 2; ν/cm^−1	x = 4; ν/cm^−1	Assignment
ν 1: symmetric stretch, ν2: asymmetric stretch; δ: scissoring or bending; ω: wagging; T: twisting; ρ: rocking; i–p: in-plane.
[1] 3511	[1] 3509		ν 2(O–H) [H2O]
[2] 3400	[2] 3402
[3] 3363	[3] 3365
[4] 3314	[4] 3328
[5] 3244	[5] 3247		ν 1(O–H) [H2O]
	[6] 2982	[1] 2977	ν 2(C–H) [THF]
	[7] 2955	[2] 2963
	[8] 2931	[3] 2932
	[9] 2892	[4] 2893	ν 1(C–H) [THF]
[6] 2123	[10] 2117	[5] 2110	ν(C≡N) [SCN]
[7] 2081	[11] 2083
	[12] 2072	[6] 2075
		[7] 2034
[8] 1636	[13] 1637		δ(H–O–H) [H2O]
	[14] 1606
	[15] 1476	[8] 1459	δ(H–C–H) [THF]
	[16] 1458	[9] 1447
	[17] 1372		ω(H–C–H) [THF]
		[10] 1344
	[18] 1318
	[19] 1296	[11] 1296
	[20] 1247	[12] 1261
	[21] 1172	[13] 1177	ν 2(ring) [THF]
		[14] 1092	T(H–C–H) [THF]
	[22] 1028	[15] 1028	ν 2(ring) [THF]
		[16] 981	2δ(S–C≡N) [SCN]
	[23] 962	[17] 958	ρ(H–C–H) [THF]
[9] 940			2δ(S–C≡N) [SCN]
	[24] 926	[18] 921	ν 1(ring) [THF]
	[25] 880	[19] 882	[THF]
	[26] 861
	[27] 847
[10] 789	[28] 790	[20] 797	ν(S–C) [SCN]
	[29] 778
[11] 719	[30] 722		T(H–O–H) [H2O]
		[21] 705
	[31] 678	[22] 678	ρ(H–C–H) [THF]
[12] 610	[32] 616	[23] 581	δ(ring i–p) [THF]
[13] 495	[33] 518
[14] 468		[24] 488	δ(S–C≡N) [SCN]
[15] 458	[34] 458	[25] 477
[16] 432	[35] 432

---

In summary, the combination of XRPD and IR spectroscopy shows that THF molecules can successfully be introduced into the structure of Mg(SCN)₂ producing new crystalline phases with distinct properties.

Thermal behavior

In the family of magnesium salts, MgCl₂ is a close relative to magnesium thiocyanate for which the thermal behaviour has already been studied thoroughly, especially regarding the complexity of obtaining pure, anhydrous MgCl₂.^2,3 In comparison, the hydrates of LiSCN, NaSCN and Ca(SCN)₂ were reported to be readily dehydrated by simple heating of the hydrated material.^18,19,21 Various anhydrous thiocyanate salts have been shown to melt congruently around 280–470 °C followed by decomposition above ∼550 °C,²² and bimetallic thiocyanates first dissociate before decomposing to sulphides and gaseous species.¹⁵

Fig. 3 shows the TGA measurement of Mg(SCN)₂·4H₂O along with the corresponding signals from mass spectroscopy (MS). According to the TGA and MS analyses, the tetrahydrate passes through different hydrates (nominal a dihydrate at 144 °C and nominal a monohydrate at 154 °C), until it finally decomposes starting around 161 °C. However, along with the recorded peaks for H₂O in the MS, another signal with a molar mass of 44 g mol⁻¹ (e.g. CO₂) was observed, which could indicate that the dehydration is accompanied by a partial decomposition process already below 161 °C. The here recorded TGA measurement agrees with that from Mojumdar et al.,¹⁰ even though the temperatures seem to be slightly shifted. Regarding the interpretation, however, some clarifications are necessary: (i) Given the hygroscopy of Mg(SCN)₂·4H₂O, it is possible that the additional H₂O molecule reported in reference 10 was actually just adsorbed surface water. (ii) There is no evidence for the intermediate formation of anhydrous Mg(SCN)₂ and the MS analysis rather suggests a thermal decomposition. (iii) Even though the formation of MgS as a decomposition product is indeed possible and would be in line with other thiocyanate systems,¹⁵ it remains challenging to substantiate this given the limited information obtainable by e.g. XRPD or IR spectroscopy (Fig. S9†).

Fig. 3 TGA measurement of Mg(SCN)₂·4H₂O under N₂: (a) TGA signal; (b) MS signal.

Further TGA and DTA measurements of magnesium thiocyanate complexes with various incorporated organic molecules often show a step-wise mass loss of the organic molecule followed by thermal decomposition.^10,12 A similar mechanism was observed in the TGA measurement of Mg(SCN)₂·4THF (Fig. 4) concerning the loss of THF. According to the combination of TGA and MS, three intermediates were formed with 3.5 (113 °C), 3 (156 °C) and 2THF (210 °C) molecules followed by thermal decomposition.

Fig. 4 TGA measurement of Mg(SCN)₂·4THF under N₂: (a) TGA signal; (b) MS signal.

The thermal behavior of Mg(SCN)₂·4H₂O and Mg(SCN)₂·4THF was further investigated by DSC (Fig. 5). The DSC measurements showed that in a semi-closed system (cold welded aluminium pan) the tetrahydrate melts at 144 °C with a pronounced hysteresis (supercooling) effect as crystallization does not occur before 80 °C. Beyond 144 °C the material showed further endothermic peaks starting at 167 °C, indicating thermal decomposition. The decomposition temperature measured by DSC fits well to that observed by TGA in Fig. 3a. In contrast, Mg(SCN)₂·4THF displayed higher thermal stability than the tetrahydrate with the first endothermic peak occurring at 197 °C. This finding is rather counterintuitive, since the exchange of H₂O with organic ligands would be expected to decrease the thermal stability of the material. Thereafter two more endothermic peaks are observed at 219 °C and 279 °C. Since no exothermic peaks appear in the cooling run as well as considering the results from TGA, we can assume that the peak at 279 °C indicates the thermal decomposition of the material. Furthermore, the TGA measurement showed that at around 210 °C the material loses two THF molecules, which is connected with the DSC peaks at 197 °C and 219 °C. Subsequent heating to higher temperatures merely revealed further decomposition of the material.

Fig. 5 DSC measurements of different phases in the series Mg(SCN)₂·(4 − x) H₂O·xTHF shown for (a) x = 0 and (b) x = 4.

In order to corroborate the above findings of TGA and DSC, temperature dependent in situ XRPD measurements were conducted for both Mg(SCN)₂·4H₂O and Mg(SCN)₂·4THF (Fig. 6). For the tetrahydrate, all reflections in the XRPD pattern disappear between 149 °C and 186 °C in agreement with the melting point at 144 °C and the decomposition at 167 °C from DSC. The formation of a new phase is observed at 190 °C, distinctly different from the originally present Mg(SCN)₂·4H₂O in terms of the XRPD pattern. Even though a structure solution of this emergent phase was not possible due to anisotropic peak broadening, it appears to be a yet unknown stacking faulted decomposition product of Mg(SCN)₂·4H₂O.

Fig. 6 Temperature dependent in situ XRPD data of Mg(SCN)₂·4H₂O (a and b) and Mg(SCN)₂·4THF (c and d) in a sealed capillary under Ar showing (a and c) selected XRPD patterns and (b and d) the thermal progression of reflections.

Mg(SCN)₂·4THF undergoes several phase transitions upon heating before it melts or becomes amorphous (Fig. 6). At 138 °C a phase transition occurs from the parent structure α-Mg(SCN)₂·4THF to a new polymorphic phase β-Mg(SCN)₂·4THF, evident from the changes in the XRPD pattern. It should be noted here that most likely this β-phase was not observed in the TGA measurement shown in Fig. 4, as THF first evaporated in the continuous N₂ gas stream of the TGA before the β-phase could form. Yet with the sealed capillary employed in the temperature dependent in situ XRPD measurement, THF is contained long enough in the sample for the β-phase to evolve. It remains unclear, however, why for the α- to β-transition no signal was observed in the DSC. Given the long measurement times in XRPD (hours for each temperature point), the kinetics of the phase transition are most likely slow. Furthermore, within the employed temperature ranges, any vibrational entropic contributions to the overall change in energy is probably small and difficult to resolve by DSC. It is therefore very likely that the phase transition could not be resolved by DSC and was consequently not observed. Upon further heating THF loss is inevitable until at 164 °C (DSC: 197 °C, TGA: 210 °C) a new phase starts to form yielding Mg(SCN)₂·2THF. The formation temperature here is a bit lower than that observed by TGA and DSC, which can be attributed to both different measurement set-ups as well as a different environment from the gas phases (the closed system in XRPD, the semi-closed system in DSC and the open system in TGA). The new phase Mg(SCN)₂·2THF is stable up to 234 °C at which point, it first turns into an unknown, stacking faulted decomposition product and then finally melts or becomes amorphous starting around 251 °C (DSC: 279 °C).

The obvious differences in thermal behaviour between Mg(SCN)₂·4H₂O and Mg(SCN)₂·4THF are rather surprising, especially considering their similar coordination chemistry. Even though it appears counterintuitive for the tetrahydrate to have a lower thermal stability, as the formation of bridging hydrogen bonds should stabilize the structure, without having any knowledge of anhydrous Mg(SCN)₂, a direct comparison between the phases cannot be made analytically.

With both TGA and XRPD the formation of anhydrous Mg(SCN)₂ was not observed and it remains to be clarified under which conditions it could be obtained. The intermediately formed Mg(SCN)₂·2THF phase is to date the only magnesium thiocyanate with only two coordinating molecules. Regarding the progression of reflections in the XRPD patterns (Fig. 6b and d), a significant shifting of selected reflections to lower angles was observed for both phases. This shifting indicates a strong and possibly anisotropic lattice expansion, which we will address later in more detail. Given the delicate synthesis and storage of Mg(SCN)₂·2H₂O·2THF, we decided to only report the results of room temperature measurements.

Structure description

In the series Mg(SCN)₂·(4 − x) H₂O·xTHF with x = 0, x = 2 and x = 4, the Mg²⁺ cations always exhibit a distorted octahedral coordination sphere with the thiocyanate anions located in the trans configuration with the N pointing towards Mg²⁺ (Fig. 7). In Mg(SCN)₂·4H₂O water molecules occupy the equatorial positions (Fig. 7a, green atoms) and the magnesium is situated on the centre of inversion. After replacing two water molecules by THF in Mg(SCN)₂·2H₂O·2THF all ligand pairs remain in the trans configuration. The magnesium cation, the thiocyanate anion and the THF related oxygen atom are situated on a mirror plane, which leads to a positional disorder of the non-planar THF molecule (Fig. 7b, red and blue atoms and bonds). The complete replacement of water by THF yields a propeller-like complex [Mg(NCS)₂(THF)₄] with magnesium situated on the centre of inversion. Two out of four THF molecules in the trans configuration are positionally disordered (Fig. 7c, red and blue atoms and bonds). In all complexes the average bond lengths between the central atom and ligands follow this trend: d[Mg–N] < d[Mg–O(H₂O)] < d[Mg–O(THF)], which is to be expected.

Fig. 7 Octahedral coordination of Mg²⁺ with SCN⁻, H₂O and THF in the series Mg(SCN)₂·(4 − x) H₂O·xTHF for (a) x = 0, (b) x = 2, (c) x = 4 (α-phase). Two THF molecules (red and blue atoms and bonds) with identical oxygen positions were used to model the positional disorder.

In the crystal structures of Mg(SCN)₂·(4 − x) H₂O·xTHF isolated [Mg(NCS)₂(H₂O)_4−x(THF)_x] octahedra form layered motifs (Fig. 8). The layers of the tetrahydrate consist of [Mg(NCS)₂(H₂O)₄] octahedra that show an alternating orientation along the a-axis (Fig. 8a, blue and yellow octahedra), a uniform orientation along the b-axis (Fig. 8b) and are stacked in the [001] direction in a distorted ABA′B′ sequence (ESI, Fig. S11†). The substitution of two water molecules by THF leads to an increase of the inter layer distance. In Mg(SCN)₂·2H₂O·2THF the [Mg(NCS)₂(H₂O)₂(THF)₂] octahedra exhibit a uniform orientation within the layer plane and the layers are stacked in a distorted AA-type fashion (Fig. 8c and d, ESI, Fig. S11†) in the [001] direction. Further replacement of water by THF molecules does not lead to an increase of the inter layer distances. This can be explained by the shift of the AA-stacking order in Mg(SCN)₂·2H₂O·2THF to a distorted ABC-type stacking in Mg(SCN)₂·4THF (Fig. 8e) in the [110] direction and by an increase of the distances between [Mg(NCS)₂(THF)₄] octahedra within the layers.

Fig. 8 Packing diagrams in the series Mg(SCN)₂·(4 − x) H₂O·xTHF for (a and b) x = 0, (c and d) x = 2, (e and f) x = 4. Yellow and blue octahedra indicate different orientations, where the thiocyanate ions on top of the blue octahedra point to the front and on the bottom point to the back and the right hand thiocyanates of the yellow octahedra point to the front and the left hand point to the back.

In terms of layer constitution and stacking order, Mg(SCN)₂·4THF (Fig. 8e and f) exhibits strong similarities to Mg(SCN)₂·4H₂O, whereas Mg(SCN)₂·2H₂O·2THF occupies an exceptional position in the row of Mg(SCN)₂·(4 − x) H₂O·xTHF structures.

Heating of Mg(SCN)₂·4THF (α-form) leads to the high temperature modification of β-Mg(SCN)₂·4THF and eventually to the formation of Mg(SCN)₂·2THF by the release of 2THF molecules. In the crystal structures of the Mg(SCN)₂·xTHF phases, magnesium is always coordinated octahedrally with the thiocyanate anions situated in the trans conformation (Fig. 9). The coordination patterns of α- and β-Mg(SCN)₂·4THF are almost identical (Fig. 9a and b). In the high temperature phase, all THF molecules are positionally disordered (Fig. 9b, blue and red atoms and bonds). After the release of two THF molecules at higher temperatures, the positional disorder of the remaining THF molecules increases (Fig. 9c, blue, red, yellow and green atoms and bonds), and the thiocyanate anions become positionally disordered as well (Fig. 9c, magenta and violet atoms and bonds). Due to the release of two ligands during the formation of Mg(SCN)₂·2THF, the emerging voids in the coordination sphere of magnesium are filled by thiocyanate related sulphur atoms, which leads to the formation of thiocyanate bridged _∞[Mg(NCS)_2/2(SCN)_2/2(THF)₂] chains (Niggli formula:

Fig. 9 Octahedral coordination of Mg²⁺ with SCN⁻ and THF for (a) α-Mg(SCN)₂·4THF; (b); β-Mg(SCN)₂·4THF; and (c) Mg(SCN)₂·2THF. Multiple THF molecules (red, blue, yellow and green atoms and bonds) and of SCN⁻ anions (magenta and violet atoms and bonds) were used to approximate the positional disorder in the structures.

The constitution and the packing of the layered motifs are very similar for the room and the high temperature modification of Mg(SCN)₂·4THF (Fig. 10a–d) and the seemingly increased interlayer distance after the phase transformation is attributed to the strong thermal expansion of α-Mg(SCN)₂·4THF (see below). In the β-form the distorted ABC stacking sequence of α-Mg(SCN)₂·4THF is reverted to a disordered ACB stacking sequence (ESI, Fig. S11†). Releasing two THF molecules from Mg(SCN)₂·4THF causes drastic structural changes. In Mg(SCN)₂·2THF the _∞[Mg(NCS)_2/2(SCN)_2/2(THF)₂] chains are arranged in layers as well with a decreased interlayer distance (Fig. 10e). Within the layers the [Mg(NCS)_2/2(SCN)_2/2(THF)₂] octahedra are oriented uniformly (Fig. 10e and f). Like β-Mg(SCN)₂·4THF, Mg(SCN)₂·2THF exhibits a disordered ABC stacking sequence (ESI, Fig. S11†) in the [110] direction.

Fig. 10 Packing diagrams of (a and b) α-Mg(SCN)₂·4THF at 25 °C, (c and d) β-Mg(SCN)₂·4THF at 150 °C and (e and f) Mg(SCN)₂·2THF at 180 °C. In (f) the C atoms from the THF molecules are omitted for clarity. Yellow and blue octahedra indicate different orientations, where the thiocyanate ions on top of the blue octahedra point to the front and on the bottom point to the back and the right hand thiocyanates of the yellow octahedra point to the front and the left hand point to the back.

Lattice expansion

The thermal expansion was investigated in detail for Mg(SCN)₂·4H₂O, α- and β-Mg(SCN)₂·4THF and Mg(SCN)₂·2THF. Plots showing the variation of the thermal expansion coefficient (α) with the principal directions X₁, X₂ and X₃ (Table S10†) for these phases are presented in Fig. 11. All compounds exhibit a pronounced anisotropic thermal expansion and some even both positive (Fig. 11, red lines) and negative (Fig. 11, blue lines) thermal expansion. The thermal expansion properties can be correlated with certain aspects of the crystal structures.

Fig. 11 Plots showing the variation of the thermal expansion coefficient α with the principal directions X₁, X₂ and X₃ (ESI, Table S10 and Fig. S14†) of (a) Mg(SCN)2·4H₂O, (b) α-Mg(SCN)2·4THF, (c) β-Mg(SCN)2·4THF and (d) Mg(SCN)2·2THF. Red colour indicates expansion, and blue colour indicates compression.

Mg(SCN)₂·4H₂O exhibits both a large positive thermal expansion of 165(4) × 10⁻⁶ K⁻¹ in the [100] direction, corresponding to the principal X₃ direction and a negative thermal expansion of −44(2) × 10⁻⁶ K⁻¹ in the [010] direction, corresponding to the principal X₁ direction (Table S10†). Hence, both the positive and negative thermal expansions occur within the layered motifs, whereas the thermal expansion perpendicular to the layers (principal X₂ axes) is only moderate (32(1) × 10⁻⁶ K⁻¹). Fig. S12† shows four [Mg(NCS)₂(H₂O)₄] octahedra, which are arranged in a diamond motif within the layers. Strong hydrogen bonds between water molecules (dark grey, dashed bonds) and weaker bonds between water molecules and SCN⁻ related sulphur (light grey dotted bonds) mediate the interactions within the layers. During the thermal expansion the thiocyanate anions start to move slightly out of the layer plains (Table S10,† principal X₂ axes). Accordingly, the [Mg(NCS)₂(H₂O)₄] octahedra, whose thiocyanate anions face in the direction of the centre of gravity of the diamond motif (Fig. S12,† black globe), start to move towards each other (blue arrow, corresponding to the principal X₁ axis). Due to steric reasons and in order to maintain a favourable O–H⋯S angle close to 180°, the [Mg(NCS)₂(H₂O)₄] octahedra situated along the a-axis veer away from the center of gravity (red arrow, corresponding to the principal X₃ axis). With respect to the centre of gravity, the expansion behaviour is comparable to Nuremberg scissors.

A complete exchange of water by THF molecules leads to an increase of the total volume expansion from 153(2) × 10⁻⁶ K⁻¹ to 268(2) × 10⁻⁶ K⁻¹ (Table S10†), which can be explained by the weaker interactions between [Mg(NCS)₂(THF)₄] octahedra. Even though the macroscopic structural motifs α- and β-Mg(SCN)₂·4THF are identical (Fig. 10a–d), the thermal expansion properties differ (Fig. 11b and c). Hence the growing disorder of the THF molecules that evolves during the transition from the α- to the β-form has a major impact on the thermal expansion properties. The expansion occurs very anisotropically in α-Mg(SCN)₂·4THF (Fig. 11): while there is a strong positive expansion along the principal X₃ axes of 173(1) × 10⁻⁶ K⁻¹ (Table S10†), along the principal X₁ axis, there is hardly any positive thermal expansion (3(1) × 10⁻⁶ K⁻¹). Within the structure, the strong (Fig. S13a and b,† along X₃ red arrow) and medium (along X₂ green arrow) thermal expansion lead to a beginning shift of the ABC stacking order of the layered motifs towards an ACB stacking. During the phase transition from the α- to the β-form the ordered THF molecule becomes positionally disordered, which leads to a discontinuous increase in the unit cell volume of almost 15 ų (Fig. S15†). The ACB stacking order of the β-modification appears to be more favourable with all THF molecules in the structure being disordered. In addition, the complete disorder of the THF molecules leads to a smaller volume expansion of 217(3) × 10⁻⁶ K⁻¹ (Fig. 11c). A strong positive expansion of 128(3) × 10⁻⁶ K⁻¹ occurs along the principal X₃ axis, which corresponds to the [001] direction. Within the layers, rows of equally oriented [Mg(NCS)₂(THF)₄] octahedra (Fig. S13c and d,† red arrows) elongate. The thermal expansion along the principal X₁ and X₂ axis occurs almost isotropically and corresponds to an expansion both within and in between the layers (Fig. S13c and d,† green arrows).

The release of two THF molecules leads to the formation of one dimensional thiocyanate bridged polyhedra forming _∞[Mg(NCS)_2/2(SCN)_2/2(THF)₂] chains in the crystal structure of Mg(SCN)₂·2THF (Fig. 10e and f). This phase shows a negative thermal expansion of −16(1) × 10⁻⁶ K⁻¹ along these chains (principal X₁ axis) and a strong positive thermal expansion of 219(3) × 10⁻⁶ K⁻¹ perpendicular to the layered motifs (principal X₃ axes, Table S10†). The total volume expansion coefficient of 240(1) × 10⁻⁶ K⁻¹ is larger than in β-Mg(SCN)₂·4THF (217(3) × 10⁻⁶ K⁻¹) but smaller than in the α-form (268(2) × 10⁻⁶ K⁻¹).

Discussion

In the group of alkali and earth alkaline metal thiocyanates, only Be(SCN)₂ has not been investigated and magnesium thiocyanate exists exclusively as coordination compounds Mg(SCN)₂·x [L]. A close relative is LiSCN, which does exist in its anhydrous form, yet is extremely prone to form hydrates.^18,47 In both cases, the unfavourable bonding between Aⁿ⁺ (A = Li, Mg; n = 1, 2) and sulphur atoms in SCN⁻ is the fundamental reason for their extensive solvent coordination chemistry. The investigations of Mg(SCN)₂-coordination compounds reveal a complex thermal decomposition behaviour and disprove any previous claims about the existence of anhydrous Mg(SCN)₂.^8–10 To date, Mg(SCN)₂·2THF is the only Mg(SCN)₂ coordination compound with two solvent ligands, and additionally, the first known Mg²⁺ thiocyanate with Mg–S bonds. Further research of this compound might lead to a procedure to promote Mg–S bonding and successfully isolate anhydrous Mg(SCN)₂. The anisotropic thermal expansion properties of the title compounds are interesting as well. Among the thiocyanates, it was already shown that in KSCN the lattice expansion occurs anisotropically with a linear increase until its well-known phase transition.⁴⁸ Other studies about the thermal expansion of thiocyanates are rather rare and typically limited to more complex systems, such as multi-cationic or organometallic compounds, which are of interest for nonlinear optics.^49,50 Both in Mg(SCN)₂·4H₂O and in Mg(SCN)₂·2THF we observed positive and negative thermal expansion, which is not an unusual phenomenon in pseudo halide compounds like Ag₃[Co(CN)₆].^51,52 In addition, there is an order–disorder phase transition in Mg(SCN)₂·4THF, which affects the thermal expansion properties. Disorder in coordination compounds is very common, e.g. one famous example being the rotational disorder of [CH₃NH₃]⁺ (MA) in MAPbI₃ with its tetragonal to cubic phase transition.⁵³ The research of ligand disorder and its impact on the parent material's chemistry or properties is very established in extended systems, such as metal–organic frameworks (MOFs),⁵⁴ but much rarer in crystalline solids with molecular coordination units. Like other coordination polymers,⁵⁵ Mg(SCN)₂·2THF exhibits a negative thermal expansion along the chains and a strong positive thermal expansion perpendicular to the chain axis.

In conclusion, the coordination chemistry of Mg(SCN)₂ turned out to be rich and has the potential to go beyond H₂O and THF as ligands. Moreover, magnesium is very similar to nickel in terms of ionic radius and preference for an octahedral coordination,⁵⁶ so exploring SCN-THF complexes appears to be an excellent opportunity to further expand the rich coordination chemistry of nickel(II) thiocyanates.^57–60 Indeed, similar M^II(SCN)₂·4THF complexes with M = Ni, V and a M^II(SCN)₂·2THF phase with M = Co have been found,⁶¹ but the thermal behaviour and the polymorphism have not been investigated yet.

Conclusion

The recrystallization of Mg(SCN)₂·4H₂O in THF yields the hitherto unknown compounds Mg(SCN)₂·2H₂O·2THF and α-Mg(SCN)₂·4THF. While the mixed H₂O-THF phase is very sensitive to heat or moisture, and even slowly decomposes under an Ar atmosphere, the pure THF phase exhibits a higher thermal stability and can be stored under an inert atmosphere for at least several weeks. In the crystal structures isolated Mg(NCS)₂(H₂O)_4−x(THF)_x units form layered motifs and magnesium is always situated on the center of inversion or on a mirror plane. The latter might be the reason why Mg(SCN)₂·(4 − x) H₂O·xTHF compounds with x = 1 or 3 were not discovered. In the structure of α-Mg(SCN)₂·4THF half of the THF molecules are positionally disordered. Heating of the compounds leads to a phase transformation into the β-form between 130 and 135 °C. After the phase transformation, all THF molecules are positionally disordered and the stacking order of the layered motifs is reversed. Further heating eventually leads to the release of two THF molecules at 170 °C. In Mg(SCN)₂·2THF thiocyanate related sulfur atoms fill the voids in the coordination sphere of magnesium, which leads to the formation of one dimensional _∞[Mg(NCS)_2/2(SCN)_2/2(THF)₂] chains. Anhydrous Mg(SCN)₂ cannot be obtained by further heating as the thiocyanate partially decomposes leading to the formation of a strongly disordered solid that cannot be characterized. All investigated Mg(SCN)₂·(4 − x) H₂O·xTHF phases exhibit a remarkable anisotropic thermal expansion, and Mg(SCN)₂·4H₂O as well as Mg(SCN)₂·2THF were found to show both positive and negative thermal expansion. The order–disorder transition of the THF molecules during the transformation from α- to β-Mg(SCN)₂·4THF has a major impact on the thermal expansion properties, i.e. it reduces the overall volume expansion and makes it more isotropic.

Author contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We want to thank and acknowledge all contributors for the making of this report: Helga Hoier and Claus Mühle (XRPD), Viola Duppel, Bernhard Fenk and Ulrike Waizman (SEM-EDX), as well as Marie-Luise Schreiber (IR). Thank you all for your useful contributions. Open Access funding provided by the Max Planck Society.

Notes and references

1. P. A. Agron and W. R. Busing, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1985, 41, 8–10.
2. K. Sugimoto, R. E. Dinnebier and J. C. Hanson, Acta Crystallogr., Sect. B: Struct. Sci., 2007, 63, 235–242.
3. R. K. Motkuri, R. S. Vemuri, D. Barpaga, H. Schaef, J. S. Loring, P. F. Martin, D. B. Lao, S. K. Nune and B. P. McGrail, ACS Sustainable Chem. Eng., 2018, 6, 1048–1054.
4. M. Pannach, S. Bette and D. Freyer, J. Chem. Eng. Data, 2017, 62, 1384–1396.
5. T. S. Daloee and F. K. Behbahani, Mol. Divers., 2020, 24, 463–476.
6. A. Unmack, D. M. Murray-Rust and H. Hartley, Proc. R. Soc. London, Ser. A, 1930, 127, 228–240.
7. K. Mereiter and A. Preisinger, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1982, 38, 1263–1265.
8. D. R. Schultz and R. W. Parry, J. Am. Chem. Soc., 1958, 80, 4–8.
9. K. Yoshihara, H. Ohshima, N. Momozawa, H. Sakai, K. Nishiyama and M. Abe, Colloid Polym. Sci., 1996, 274, 685–695.
10. S. C. Mojumdar, M. Melník and E. Jóna, J. Anal. Appl. Pyrolysis, 1998, 46, 147–156.
11. S. C. Mojumdar, M. Melník, E. Jóna and D. Hudecová, Chem. Pap., 1999, 53, 265–271.
12. D. V. Soldatov, P. A. Trushin, V. A. Logvinenko and E. V. Grachev, J. Struct. Chem., 1993, 34, 232–238.
13. K. Touaj and M. Chabanel, J. Chem. Soc., Faraday Trans., 1995, 91, 4395–4401.
14. C. Wickleder and P. Larsen, Chem. Mater., 2004, 16, 4016–4021.
15. S. Gunasekaran and S. Ponnusamy, Cryst. Res. Technol., 2006, 41, 130–137.
16. U. Berger and W. Schnick, J. Alloys Compd., 1994, 206, 179–184.
17. B. Jürgens, E. Irran and W. Schnick, J. Solid State Chem., 2001, 157, 241–249.
18. O. Reckeweg, A. Schulz, B. Blaschkowski, Th. Schleid and F. J. DiSalvo, Z. Naturforsch., B: J. Chem. Sci., 2014, 69, 17–24.
19. K. Mereiter and A. Preisinger, Z. Kristallogr., 1984, 169, 95–107.
20. S. Yamamoto, M. Sakuno and Y. Shinnaka, J. Phys. Soc. Jpn., 1987, 56, 4393–4399.
21. C. Wickleder and P. Larsen, Z. Naturforsch., B: J. Chem. Sci., 2002, 57, 1419–1426.
22. C. Wickleder, Z. Anorg. Allg. Chem., 2001, 627, 1693.
23. E. Dubler, A. Reller and H. R. Oswald, Z. Kristallogr., 1982, 161, 265–277.
24. I. Lindqvist, Acta Crystallogr., 1957, 10, 29–32.
25. A. L. Beauchamp and D. Goutier, Can. J. Chem., 1972, 50, 977–981.
26. K. Sasvári and G. A. Jeffrey, Acta Crystallogr., 1966, 20, 875–881.
27. H. Schmidt, E. Hennings and W. Voigt, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2012, 68, 4–9.
28. E. Hennings, H. Schmidt and W. Voigt, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 2013, 69, 1292–1300.
29. M. G. Aranda, I. Artioli, T. Bier, À.G De la Torre, D. Freyer, R. Kaden, B. Lothenbach, C. Pritzel, B. Raab, S. Stöber, J. Södje, P. Stemmermann, T. Westphal, F. Winnefeld and Q. L. Yu, Cementitious Materials, De Gruyter, Berlin, Boston, 2017.
30. T. Rüffer, C. Bruhn and D. Steinborn, Z. Anorg. Allg. Chem., 2001, 627, 2408–2412.
31. K. Czaja, M. Białek and A. Utrata, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 2512–2519.
32. W. OchÈ©dzan-Siodłak and M. Nowakowska, Eur. Polym. J., 2005, 41, 941–947.
33. S. Pirinen, K. Jayaratne, P. Denifl and T. T. Pakkanen, J. Mol. Catal. A: Chem., 2014, 395, 434–439.
34. M. Salama, I. Shterenberg, L. J. W. Shimon, K. Keinan-Adamsky, M. Afri, Y. Gofer and D. Aurbach, J. Phys. Chem. C, 2017, 121, 24909–24918.
35. S. Halut-Desportes and C. Bois, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1979, 35, 2205–2207.
36. P. Sobota, T. Pluzinski and T. Lis, Z. Anorg. Allg. Chem., 1986, 533, 215–224.
37. P. Sobota, J. Ejfler, T. Płuziński and T. Lis, Polyhedron, 1991, 10, 2457–2461.
38. B. Neumüller, G. Stieglitz and K. Dehnicke, Z. Naturforsch., B: J. Chem. Sci., 1993, 48, 1151–1154.
39. J. M. Bakker, G. B. Deacon, C. M. Forsyth, P. C. Junk and M. Wiecko, Eur. J. Inorg. Chem., 2010, 6, 2813–2825.
40. E. Shurdha, S. H. Lapidus, P. W. Stephens, C. E. Moore, A. L. Rheingold and J. S. Miller, Inorg. Chem., 2012, 51, 9655–9665.
41. IUPAC, Nomenclature of Inorganic Chemistry, Butterworths, London, 1970, p. 2.
42. D. R. Lide, Physical Constants of Organic Compounds, in CRC Handbook of Chemistry and Physics, 90th Edition (CD-ROM Version 2010), CRC Press/Taylor and Francis, Boca Raton, FL, 2010.
43. Solubility KCl, https://periodic-table-of-elements.org/SOLUBILITY/potassium_chloride, (accessed 5 September 2020).
44. M. Chaplin, Water Absorption Spectrum, http://www1.lsbu.ac.uk/water/water_vibrational_spectrum.html, (accessed 5 September 2020).
45. A. Palm and E. R. Bissell, Spectrochim. Acta, 1960, 16, 459–466.
46. L. H. Jones, J. Chem. Phys., 1956, 25, 1069–1072.
47. D. A. Lee, Inorg. Chem., 1964, 3, 289–290.
48. W. Schranz, in Phase Transitions, 1994, vol. 51, pp. 1–66.
49. X. Wang, D. Xu, Y. T. Chow, W. Liu, S. Li, J. Huang, G. Zhang, M. Lü, D. Yuan, H. P. Chan, Q. Ren and P. L. Chu, Mater. Res. Bull., 2004, 39, 1407–1416.
50. X. Q. Wang, X. F. Cheng, S. J. Zhang, D. Xu, G. H. Zhang, Z. H. Sun, F. P. Yu, X. J. Liu, W. L. Liu and C. L. Chen, Physica B: Condens. Matter, 2010, 405, 1071–1080.
51. A. L. Goodwin, M. Calleja, M. J. Conterio, M. T. Dove, J. S. O. Evans, D. A. Keen, L. Peters and M. G. Tucker, Science, 2008, 319, 794–797.
52. A. L. Goodwin, D. A. Keen and M. G. Tucker, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 18708–18713.
53. A. Poglitsch and D. Weber, J. Chem. Phys., 1987, 87, 6373–6378.
54. R. E. Morris and L. Brammer, Chem. Soc. Rev., 2017, 46, 5444–5462.
55. K. Yadava, G. Gallo, S. Bette, C. E. Mulijanto, D. P. Karothu, I. H. Park, R. Medishetty, P. Naumov, R. E. Dinnebier and J. J. Vittal, IUCrJ, 2020, 7, 83–89.
56. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
57. C. Wellm, T. Neumann, M. Ceglarska, G. Gallo, M. Rams, R. E. Dinnebier and C. Näther, CrystEngComm, 2020, 22, 2350–2360.
58. C. Wellm and C. Näther, Acta Crystallogr., Sect. E: Crystallogr. Commun., 2019, 75, 1685–1688.
59. A. Jochim, M. Ceglarska, M. Rams and C. Näther, Z. Anorg. Allg. Chem., 2018, 644, 1760–1770.
60. A. Jochim, M. Rams, T. Neumann, C. Wellm, H. Reinsch, G. M. Wójtowicz and C. Näther, Eur. J. Inorg. Chem., 2018, 2018, 4779–4789.
61. E. Shurdha, C. E. Moore, A. L. Rheingold, S. H. Lapidus, P. W. Stephens, A. M. Arif and J. S. Miller, Inorg. Chem., 2013, 52(18), 10583–10594.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 2055308–2055312. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/d1dt00469g

---