Skip to content
Publicly Available Published by De Gruyter March 22, 2021

Synthesis of the scandium chloride hydrates ScCl3·3H2O and Sc2Cl4(OH)2·12H2O and their characterisation by X-ray diffraction, 45Sc NMR spectroscopy and DFT calculations

  • Thomas Bräuniger EMAIL logo , Philipp Bielec , Otto E. O. Zeman , Igor L. Moudrakovski , Constantin Hoch and Wolfgang Schnick

Abstract

The compounds ScCl3·3H2O (SCTH) and [{Sc(H2O)5(μ-OH)}2]Cl4·2H2O (SCOH), have been synthesised and characterised by single-crystal XRD, 45Sc NMR spectroscopy and DFT calculations, with the crystal structure of SCTH reported here for the first time. From 45Sc NMR measurements under static and MAS conditions, both chemical shift and quadrupolar coupling parameters have been determined. The quadrupolar coupling constants χ for the octahedrally coordinated scandium sites in SCTH are 2.0 ± 0.1 MHz for Sc(1) and 3.81 ± 0.05 MHz for Sc(2). For SCOH, where the hepta-coordination of the single scandium site constitutes a less symmetric electronic environment, 14.68 ± 0.05 MHz was found. DFT calculations for the static SCTH structure consistently overestimate the quadrupolar coupling constants, indicating the possible presence of crystal water dynamics on the NMR time scale.

1 Introduction

The chemistry of scandium has been explored less than that of other transition metals, possibly because of its relatively high cost, and the fact that in its inorganic compounds, scandium occurs exclusively in the +III oxidation state [1]. Similarly, technological use of scandium has been mostly limited to specialised areas, such as a doping substance of zirconia [2], as a component of aluminium alloys [3], or recently, as an ingredient of matrix materials for LED phosphors [4], [5]. For determination of structure and dynamics of scandium-containing compounds, nuclear magnetic resonance (NMR) spectroscopy is a valuable analytical tool, complementing X-ray and/or neutron diffraction techniques. The nuclide 45Sc (spin I = 7/2) occurs with 100% natural abundance and has a resonance frequency close to that of carbon. In principle, therefore, 45Sc is an excellent nucleus for observation by solid-state NMR spectroscopy. While a number of such studies of solids may be found in the literature, see e.g. [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], the data base for 45Sc NMR parameters is still scarce in comparison to other nuclides such as 27Al, as is obvious from the relatively short compilation tables in the relevant review articles [15], [16], [17]. Over the last years, solid-state NMR studies have been increasingly accompanied by density functional theory (DFT) calculations of NMR parameters, employing periodic plane waves [18], [19], [20]. Such calculations, which have also been done for 45Sc NMR [8], [9], [10], [11], [12], [13], [14], may be useful for validating experimental results, and for correlating them to structural motifs in the investigated compounds.

Evidently, it is worthwhile to expand the data base for solid-state 45Sc NMR spectroscopy, to improve both DFT calculation accuracy and knowledge about the relation between NMR parameters and structural features. In the current work, we describe 45Sc NMR measurements and corresponding DFT calculations for two scandium chloride hydrates, namely ScCl3·3H2O and Sc2Cl4(OH)2·12H2O. The structures of both compounds have also been established by singe-crystal XRD, with the one of ScCl3·3H2O reported here for the first time.

2 Results and discussion

2.1 ScCl3·3H2O (SCTH)

Scandium trichloride trihydrate (henceforward abbreviated to SCTH), was synthesised as described in the Section 4. A single crystal was isolated and analysed by XRD, with crystallographic details given in Table 1, and the atomic coordinates derived from the refinement procedure listed in Table 2. SCTH crystallises in the monoclinic space group P21/c (No. 14), with Figure 1a showing a projection of the crystal structure. The SCTH structure contains two fundamental building units, as depicted in Figure 1b. In each of the units, Sc3+ is octahedrally coordinated by Cl and H2O ligands. Sc(1) is located inside [ScCl2(H2O)4]+ octahedra (light orange in Figure 1b), which are already known from the structure of the hexahydrate, ScCl3·6H2O [6]. The coordination of Sc(2), [ScCl4(H2O)2] (light blue in Figure 1b), is a novel building unit. ScCl3·3H2O may therefore formally be referred to as tetraaquadichloridoscandium(III) diaquatetrachloridoscandate(III). The packing of the two types of octahedra is reminiscent of the cubic closest packing in terms of the following description. First, the [ScCl2(H2O)4]+ octahedra are arranged according to an a-centered packing, as shown in Figure 1a. Second, the [ScCl4(H2O)2] octahedra complement the structure, which can then be described as a distorted cubic closest packing of ScL6 octahedra (L = ligand). Most likely, this distortion is due to the many hydrogen bonds (dashed lines in Figure 1a) between the octahedral building units. The validity of this structure model was confirmed by a powder XRD analysis (see Supplementary Material, available online), and by solid-state NMR spectroscopy, as described in the following.

Table 1:

Crystallographic details on single-crystal data collection, structure solution and refinement for ScCl3·3H2O (SCTH) and Sc2Cl4(OH)2·12H2O (SCOH). Standard deviations are given in parentheses in units of the last digit.

CompositionScCl3O3H6Sc2Cl4O14H26
Crystal systemmonoclinicorthorhombic
Space groupP21/c, (no. 14)Pnnm, (no. 58)
Lattice parameters/Å, deg.a9.666(2)8.6912(4)
b8.1500(19)15.6523(7)
c9.2018(14)7.1596(3)
β90.821(14)90
V3724.8(3)973.97(7)
Z44
Data collection temperature/K110(2)295(2)
Density (X-ray) ρ/g cm−31.881.64
Structure factor F(000)/e408496
Absorption coefficient/mm−12.01.3
Radiation, wavelength/ÅMo-Kα, 0.71073Mo-Kα, 0.71073
Diffraction angle range 2ϑ/deg.3.270–27.4915.21–59.99
Index range−12 < h < +12−12 < h < +12
−10 < k < +10−22 < k < +22
−11 < l < +11−10 < l < +10
DiffractometerBruker D8 Quest
CorrectionsPolarization, Lorentz, absorpt. effects [25], [26]
Structure refinementFull-matrix least-squares on F2 [27]
No. of collected reflections1121434755
No. of independent reflections16631502
No. of independent reflections (I≥2 σ(I))13241446
No. of least-squares parameters9287
Rint0.07250.0941
Rσ0.04550.0269
R1 for I≥2 σ(I)0.02550.0282
wR2 for I≥2 σ(I)0.04500.0891
R1 (all I)0.04100.0304
wR2 (all I)0.04750.0901
GooF0.9521.330
Residual ρ(e) max; min/e Å−3+0.30; −0.34+0.453; −0.643
CCDC deposition numberCSD-2048718CSD-2048365
Table 2:

Standardized fractional atomic coordinates, Wyckoff positions, point symmetries, and displacement parameters (in Å2) for ScCl3·3H2O (SCTH) derived from single-crystal XRD. For the hydrogen position, Uiso is listed, for all other atoms, the equivalent isotropic displacement parameter Uequiv, which is defined as 1/3 of the trace of the anisotropic displacement tensor. (All standard deviations are given in parentheses in units of the last digit.)

AtomWyck.Symm.xYzUequiv/iso
Sc(1)2a10000.00940(13)
Sc(2)2d11/201/20.00960(13)
Cl(1)4e10.89994(5)0.98385(6)0.24085(5)0.01585(13)
Cl(2)4e10.26952(5)0.07198(6)0.40045(5)0.01490(13)
Cl(3)4e10.57676(5)0.28737(6)0.48069(5)0.01661(13)
O(1)4e10.17819(16)0.1220(2)0.07870(18)0.0161(3)
O(2)4e10.57465(19)0.9491(2)0.28820(16)0.0172(3)
O(3)4e10.1006(2)0.7755(2)0.03930(19)0.0216(4)
H(1)4e10.083(3)0.702(4)0.095(3)0.051(10)
H(2)4e10.210(3)0.198(4)0.039(3)0.045(9)
H(3)4e10.536(3)0.912(3)0.226(3)0.032(8)
H(4)4e10.202(3)0.120(3)0.155(3)0.039(9)
H(5)4e10.643(3)0.979(3)0.263(3)0.041(10)
H(6)4e10.168(4)0.765(4)0.021(3)0.057(12)
Figure 1: Crystal structure of ScCl3·3H2O (SCTH) with (a) projection of the unit cell and (b) the fundamental building units [ScCl2(H2O)4]+ and [ScCl4(H2O)2]−. Atomic coordinates are given in Table 2.
Figure 1:

Crystal structure of ScCl3·3H2O (SCTH) with (a) projection of the unit cell and (b) the fundamental building units [ScCl2(H2O)4]+ and [ScCl4(H2O)2]. Atomic coordinates are given in Table 2.

For solid-state NMR spectroscopy of scandium, which occurs solely as the 45Sc isotope with spin I = 7/2, it is important to recognise the fact that it possesses an electrical quadrupole moment of eQ = 220 mb [21]. The interaction of this quadrupole moment with the electronic surroundings of the nucleus constitutes the quadrupolar interaction, which lifts the degeneracy of the 2I = 7 transitions, separating them into central transition (CT), for which m  =  −1/2→ + 1/2, and six symmetric satellite transitions (ST’s). As long as the quadrupolar interaction is small compared to the fundamental Zeeman interaction (perturbation to first order), the CT resonance position remains unaffected, but the satellite transitions are shifted [22]. If the quadrupolar coupling becomes larger, also the CT position is affected, as described by second-order perturbation terms.[1] For polycrystalline (powder) samples, this leads to characteristic shapes in the NMR spectra, which can be evaluated to extract the quadrupolar coupling parameters. Since the CT powder pattern, even when broadened to second order, is still much narrower than the corresponding ST spectrum, solid-state NMR characterisation of quadrupolar nuclei with half-integer spin is often restricted to observing the position and shape of the central transition. This can be done for static samples, where the line shape is affected by both quadrupolar and chemical shift interactions (plus possibly dipolar coupling), or under magic-angle spinning (MAS) conditions, where chemical shift anisotropy and dipolar coupling are reduced or even completely averaged out. The actual position of the CT pattern is determined by both the isotropic chemical shift δiso, and the additional shift contribution of the quadrupolar interaction to second order. Under most circumstances, all parameters can be determined simultaneously by a numerical fit of the line shape [23].

In Figure 2, the 45Sc NMR central-transition spectra of SCTH are shown. The MAS spectrum in Figure 2a was recorded with a spinning speed of 23 kHz. Under these conditions, the spectrum is practically free of effects of chemical shift anisotropy and dipolar coupling, and the line broadening exhibited by the two well-resolved resonances of the two scandium sites is caused solely by the quadrupolar interaction. A numerical fit (see Section 4 for details) of this spectrum allows the determination of the quadrupolar coupling parameters to good accuracy, with the results listed in Table 3. In contrast, the static spectrum displayed in Figure 2b is dominated by the chemical shift anisotropy (CSA). Using the quadrupolar parameters derived from the MAS spectrum, a good estimate of the chemical shift parameters could be obtained (see Table 3), with the δiso of both 45Sc sites in the range expected for octahedral coordination [12]. The parameters derived from fitting the spectra at B0 = 9.4 T were also applied for simulation of a static spectrum at B0 = 11.7 T with good results (see Figure S3 of the Supplementary Material), further validating the fit results.

Figure 2: 45Sc NMR central-transition spectra of a powder sample of SCTH at B0 = 9.4 T, with the simulated line shapes shown in red (see text for simulation details). (a) MAS spectrum at 23 kHz spinning speed. (b) Static spectrum with the contribution of the subspectra also shown, Sc(1) in blue and Sc(2) in green.
Figure 2:

45Sc NMR central-transition spectra of a powder sample of SCTH at B0 = 9.4 T, with the simulated line shapes shown in red (see text for simulation details). (a) MAS spectrum at 23 kHz spinning speed. (b) Static spectrum with the contribution of the subspectra also shown, Sc(1) in blue and Sc(2) in green.

Table 3:

45Sc NMR interaction parameters of ScCl3·3H2O (SCTH), ScCl3·6H2O (SCHH), and Sc2Cl4(OH)2·12H2O (SCOH), obtained from numerical fits of NMR spectra of polycrystalline samples, and from DFT calculations using the Castep code. Listed are the isotropic chemical shift δiso, the reduced anisotropy Δδ = δ33δiso, and the asymmetry parameter ηCS = (δ22δ11)/Δδ, with all values referenced against an aqueous ScCl3 solution at 0 ppm. Also given are the quadrupolar coupling constant χ = CQ = e2qQ/h (where eq = Vzz is the largest component of the electric field gradient tensor V), and the quadrupolar asymmetry parameter ηQ = (VxxVyy)/Vzz.

Scandium siteSolid-state NMRDFT
δisoΔδηCSχηQχaηQ
(ppm)(ppm)(MHz)(MHz)
SCTH
 Sc(1)127 ± 1−129 ± 20.0 + 0.12.0 ± 0.11–0.2−6.880.71
 Sc(2)188 ± 195 ± 20.1 ± 0.13.81 ± 0.050.4 ± 0.26.490.79
SCHHb
 Sc(1)125.4 ± 0.5−117 ± 10.08 ± 0.053.9 ± 0.20.77 ± 0.094.700.86
SCOH
 Sc(1)21 ± 1cc14.68 ± 0.050.5 ± 0.1−13.980.53
  1. The indicated errors have been estimated from the fit quality variation upon change of input parameters. aThe absolute sign of χ is available only from calculations, not via NMR experiments. bExperimental data taken from reference [6]. cAveraged out in MAS spectra and therefore not accessible.

Whereas the existence of two components in the 45Sc NMR spectra is consistent with the presence of two crystallographically distinct scandium sites in the SCTH structure, the question of attributing the respective subspectra to Sc(1) and Sc(2) remains. In the current case, the assignment problem may be solved by a simple comparison with a previous NMR spectroscopic study of the hexahydrate, ScCl3·6H2O, by Rossini and Schurko [6]. As explained above, the hexahydrate contains the scandium atoms exclusively in [ScCl2(H2O)4]+ octahedra, and the same coordination is found for Sc(1) in the trihydrate. Looking at the spectra in Figure 3 of reference [6], which were recorded at the same magnetic field, it becomes immediately obvious that the blue subspectrum in our Figure 2b must belong to Sc(1): It shows the same characteristic shape caused by a CSA with ηCS ≈ 0, with the intensity maximum on the left-hand side due to a negative anisotropy Δδ with similar magnitude, and exhibits a practically identical isotropic chemical shift, see Table 3. Remarkably, the quadrupolar coupling parameters of Sc(1) in SCTH strongly differ from those reported for the hexahydrate, with the latter having a coupling constant almost twice as large as the one observed for SCTH. It is well known that the quadrupolar interaction is extremely sensitive to distortions of the electronic environment [28], but when comparing the octahedral coordination of 45Sc in both compounds, they are found to be essentially the same in terms of atomic distances and ‘bond’ angles. An indication for the cause of this unexpected difference in quadrupolar coupling strength may be found in the results of DFT calculations, as explained in the following.

Figure 3: Crystal structure of [{Sc(H2O)5(μ-OH)}2]Cl4·2H2O (SCOH), showing (a) a projection of the unit cell, and (b) the structure of the cationic dinuclear units, [{Sc(H2O)5(μ-OH)}2]4+. The colour coding for the elements is identical to Figure 1, with the atomic coordinates listed in Table 4.
Figure 3:

Crystal structure of [{Sc(H2O)5(μ-OH)}2]Cl4·2H2O (SCOH), showing (a) a projection of the unit cell, and (b) the structure of the cationic dinuclear units, [{Sc(H2O)5(μ-OH)}2]4+. The colour coding for the elements is identical to Figure 1, with the atomic coordinates listed in Table 4.

Calculations of NMR interaction parameters using density functional theory (DFT) with the periodic plane wave approach have become increasingly accurate over recent years [18], [19], [20], and have also shown to produce results very close to experimental findings for 45Sc NMR spectroscopy [8], [9], [10], [11], [12], [13], [14]. It is well known that in most cases, the atomic coordinates taken from XRD analyses have to be energy-optimised by the DFT algorithm in order to reach good agreement between calculation and experiment [20]. We have carried out such computations for both SCTH and SCOH using the Castep code (see Section 4 for details). For SCTH, it became apparent that the DFT algorithm consistently delivers quadrupolar coupling constants that are much larger than the experimental values, with the non-optimised structure (i.e. coordinates directly taken from Table 2) resulting in χ = Cq = −7.5 MHz for Sc(1) and 6.1 MHz for Sc(2), not even reproducing the experimentally established fact that Sc(2) shows a larger coupling constant. For fully optimised structures, with the coordinates of all atoms in the unit cell undergoing changes according to the DFT-calculated energy landscape, the quadrupolar coupling constants became even bigger. The values listed in Table 3 stem from a structure where only the atomic coordinates of the water molecules have been optimised by the algorithm, and they are the smallest values (and hence closest to experiment) that could be obtained. We encountered the same effect when computing the quadrupolar parameters of the hexahydrate [6], with the best values shown in Table 3. While the underlying reason for the discrepancy between calculations and experiment cannot be established with certainty, we think that the most likely cause is the presence of water dynamics in both the tri- and hexahydrate. If these exchange processes take place on the required time scale, they may lead to (incomplete) averaging of the quadrupolar coupling, generating an effective coupling constant, which is generally smaller [29], [30], [31]. We also note that the existence of water dynamics could explain that the experimentally determined coupling constant for Sc(2) in SCTH, which resides in [ScCl4(H2O)2] octahedra, is larger than that of Sc(1) in [ScCl2(H2O)4]+ coordination, since the octahedron of the latter contains more water molecules which potentially could contribute to the averaging process. Otherwise, however, the experimental evidence for water exchange processes taking place in the hydrates is scarce. The displacement parameters derived from the XRD refinement (Table 2) do not show excessive values indicative of dynamics, but it should be kept in mind that from electron density maps, it is very difficult to distinguish between disorder which is either static or dynamic on the NMR time scale. The singular hint for the presence of dynamics, therefore, comes from the DFT calculations, which come closest to the experimental values when only the coordinates of the water molecules are energetically optimised. Detailed experimental investigation of the water dynamics in the tri- and hexahydrate would require extensive temperature-dependent measurements, or alternatively the use of 2H-NMR spectroscopy of isotopically labelled samples, which are outside the scope of the current work.

2.2 Sc2Cl4(OH)2·12H2O (SCOH)

Sc2Cl4(OH)2·12H2O, henceforward abbreviated to SCOH, may be formally described as bis-(μ-hydroxido)decaaquadiscandium(III) tetrachloride dihydrate, and can also be represented by the more detailed sum formula [{Sc(H2O)5(μ-OH)}2]Cl4·2H2O. The compound was synthesised as described in the Section 4. A single crystal was isolated and analysed by XRD, with crystallographic details given in Table 1, and the atomic coordinates derived from the refinement procedure listed in Table 4. The results are in close agreement with those of previous XRD studies [32], [33], [34]. SCOH crystallises in the orthorhombic space group Pnnm (No. 58), with two formula units per unit cell, as depicted in Figure 3a. Each scandium atom, located at Wyckoff position 4g, is coordinated by five H2O molecules and bridged to another scandium atom by two hydroxide anions (see Figure 3b), resulting in hepta-coordination with a capped trigonal prismatic geometry. The two [{Sc(H2O)5(μ-OH)}2]4+ cationic dinuclear units per unit cell are linked by hydrogen bonds via the crystallisation water molecules, forming chains along the crystallographic a axis, as was already noted in an earlier publication [33].

Table 4:

Standardized fractional atomic coordinates, Wyckoff positions, point symmetries, and displacement parameters (in Å2) for Sc2Cl4(OH)2·12H2O (SCOH) derived from single-crystal XRD. For the hydrogen position, Uiso is listed, for all other atoms, the equivalent isotropic displacement parameter Uequiv, which is defined as 1/3 of the trace of the anisotropic displacement tensor. (All standard deviations are given in parentheses in units of the last digit.)

AtomWyck.Symm.xYzUequiv/iso
Sc(1)4g..m0.13138(3)0.57994(2)00.01759(16)
Cl(1)4g..m0.33814(6)0.29716(3)00.03292(18)
Cl(2)4g..m0.72785(5)0.11695(3)00.03165(18)
O(1)8h10.45872(14)0.16811(9)0.28874(18)0.0488(4)
O(2)4g..m0.30387(16)0.68393(9)00.0326(3)
O(3)8h10.21336(17)0.04195(8)0.2770(2)0.0504(4)
O(4)4g..m0.10110(14)0.44923(7)00.0241(3)
O(5)4g..m0.0904(2)0.14306(11)00.0389(4)
H(1)4g..m0.602(4)0.320(2)00.043(7)
H(2)8h10.416(3)0.2003(14)0.213(3)0.047(5)
H(3)4g..m0.162(4)0.4184(17)00.028(6)
H(4)4g..m0.015(5)0.145(3)00.056(11)
H(5)8h10.277(3)0.498(2)0.276(4)0.074(8)
H(6)8h10.042(3)0.3439(14)0.271(4)0.053(6)
H(7)4g..m0.709(4)0.269(2)00.043(8)
H(8)8h10.180(3)0.0683(17)0.212(4)0.057(7)
H(9)4g..m0.128(4)0.194(3)00.063(10)

The coordination of 45Sc in the cationic dinuclear units (see Figure 3b) constitutes an electronic environment which is less symmetric than those present in the two octahedral scandium coordinations of the ScCl3·3H2O structure, leading to a quadrupolar coupling constant of χ = 14.68 ± 0.05 MHz. For such large coupling constants, the acquisition of a distortion-free static spectrum of even just the central-transition signal becomes problematic, because of the limited excitation bandwidth of the RF pulses. Therefore, the CT resonances of MAS spectra at two different magnetic field strengths were evaluated, with both experimental and fitted line shape displayed in Figure 4. The width of the second-order broadened CT spectrum of SCOH is much reduced at higher field (see Figure 4b), since this broadening scales inversely with the Larmor frequency. The NMR parameters resulting from fitting the MAS spectra are listed in Table 3. We note that a previous solid-state NMR spectroscopic study of SCOH [34] found a coupling constant of 13.7 MHz, which is not too far from our value. However, for currently unknown reasons, the line shape of the 45Sc NMR MAS spectrum reported in Ref. [34] is distinctly dissimilar from those shown in Figure 4, leading to a strongly different asymmetry parameter of 0.9, which we find neither experimentally nor by DFT calculations.

Figure 4: 45Sc NMR central-transition spectrum of a powder sample of SCOH at (a) 24.5 kHz MAS and B0 = 11.7 T, and (b) 30 kHz MAS and B0 = 21.1 T, with the simulated (see text for details) line shapes shown in red.
Figure 4:

45Sc NMR central-transition spectrum of a powder sample of SCOH at (a) 24.5 kHz MAS and B0 = 11.7 T, and (b) 30 kHz MAS and B0 = 21.1 T, with the simulated (see text for details) line shapes shown in red.

These Castep calculations reproduce the quadrupolar coupling parameters for SCOH within 5% of our experimentally determined values, see Table 3. Such good agreement between experiment and the DFT calculations of a static structure seems to imply that the dynamic processes of the water molecules invoked for SCTH are absent in SCOH, or at least possess different kinetic parameters. Chemical intuition indeed doubts that the two bridging hydroxide anions in the dinuclear unit shown in Figure 3b would undergo fast exchange, but this does not apply to the remaining five water molecules in the coordination sphere. Hence, similar to SCTH, the decision about presence or absence of water dynamics in SCOH must await further experimental investigation.

3 Conclusions

Two scandium chloride hydrates, namely ScCl3·3H2O (SCTH), and Sc2Cl4(OH)2·12H2O (SCOH), have been synthesised and characterised by single-crystal XRD, 45Sc NMR spectroscopy and DFT calculations. The crystal structure of SCTH is reported here for the first time. It can be understood as a distorted cubic closest packing, built up by an a-centered arrangement of [ScCl2(H2O)4]+ octahedra complemented by [ScCl4(H2O)2] octahedra. From 45Sc NMR measurements, the chemical shift and quadrupolar coupling parameters have been determined for both compounds. As expected, the quadrupolar coupling parameters for the octahedrally coordinated 45Sc in SCTH are smaller than those for SCOH, where the hepta-coordination of the scandium atom constitutes a less symmetric electronic environment. However, DFT calculations for the static SCTH structure consistently overestimate the quadrupolar coupling constants, indicating the possible presence of crystal water dynamics on the NMR time scale. In contrast, this overestimate is not observed for SCOH. While unequivocal experimental evidence for the presence or absence of water dynamics in the respective compounds is not yet available, the present study may point to another potential benefit of augmenting solid-state NMR experiments with DFT calculations: The possible uncovering of dynamic processes, which may hide behind incompletely averaged NMR interaction parameters.

4 Experimental

4.1 Synthesis

For the synthesis of scandium trichloride trihydrate, 1 g scandium oxide (99.9%, Smart Elements) was dissolved in concentrated hydrochloric acid (37%, Merck). After inspissating the solution in high vacuum at 25 °C, an excess of thionyl chloride (97%, Sigma Aldrich) was added under inert conditions (Ar). The mixture was stirred for 18 h and afterwards thionyl chloride residues were removed in high vacuum. ScCl3·3H2O was obtained as a colourless crystalline solid in almost quantitative yield.

Sc2Cl4(OH)2·12H2O (SCOH) was synthesised according to a previously published procedure [32]. 1.9 g of Sc2O3 was added to a round bottom flask with 34 ml of distilled H2O and 16 ml HCl (37%). The mixture was then refluxed for 2 h and left to cool down to room temperature. The remaining solid was filtered off and most of the solvent was removed in a rotation evaporator, obtaining the target compound in almost quantitative yield as a colourless crystalline solid. A synthesis of SCOH starting from metallic scandium has also been reported in the literature [34].

4.2 X-ray structure analyses

For single crystal diffraction studies, suitable specimens of single crystals of ScCl3·3H2O and Sc2Cl4(OH)2·12H2O were selected under a binocular with polarization filter and, owing to their hygroscopy, sealed in thin-walled glass capillaries (see Figure S1 in the Supplementary Material). The crystals were centered on the goniometer of a Bruker D8 Quest diffractometer system equipped with a Mo-Kα radiation source (IμS microfocus tube), Goebel mirror optics and a CCD detector Photon II. Data of the entire Ewald sphere were collected with combined φ and ω scans and corrected for Lorentz and polarization effects [25]. Absorption corrections were performed on the basis of multiply registered reflections (multi-scan algorithm) [26]. Structure solution for the new compound ScCl3·3H2O was performed with direct methods [27]. For the structure refinement of Sc2Cl4(OH)2·12H2O the literature model [32] was used for starting values. Refinements of the structure models were performed with full-matrix least-squares cycles [27]. All non-hydrogen atoms were refined with anisotropic treatment of the displacement factors. The positions of hydrogen atoms were taken from difference Fourier syntheses and refined with isotropic displacement parameters applying no restraints whatsoever. Details on powder diffractometric investigations and Rietveld refinement of the structure model of ScCl3·3H2O are compiled in the Supplementary Material (available online).

CSD-2048718 (SCTH) and CSD-2048365 (SCOH) contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via http://www.ccdc.cam.ac.uk/data_request/cif.

4.3 Solid-state 45Sc NMR spectroscopy

45Sc NMR MAS spectra were acquired on a Bruker Avance-III 400 spectrometer, at a Larmor frequency of ν0(45Sc) = 97.20 MHz, a Bruker Avance-III 500, at ν0(45Sc) = 121.53 MHz, and a Bruker Avance-II 900, at ν0(45Sc) = 218.55 MHz. For all measurements, a commercial 2.5 mm MAS probe was used, with the actual MAS frequencies given in the figure captions. Recycle delays were 5 s for SCTH and 1 s for SCOH. All spectra were referenced against an aqueous ScCl3 solution at 0 ppm. Deconvolution and numerical fit of the NMR spectra were done with the help of the DMFIT simulation program [23]. The simulated spectra shown in the respective figures were re-calculated from the determined parameters with the Simpson package [35].

4.4 DFT calculations

All calculations were performed using the Castep DFT code [18], [19] using the GIPAW algorithm [36], [37]. The version used in this work is integrated within the Biovia Materials Studio 2018 suite. The computations use the generalized gradient approximation (GCA) and Perdew-Burke-Ernzerhof (PBE) functional [38], with the core-valence interactions described by ultra-soft pseudopotentials [37]. Integrations over the Brillouin zone were done using a Monkhorst-Pack grid with k point spacings generally being less than 0.04 Å−1 [39]. The cut-off energy used was in the range of 630–900 eV. The convergence of calculated NMR parameters was tested for all systems with respect to the size of a Monkhorst-Pack k-grid and a basis set cut-off energy. Since calculated NMR properties are known to be very sensitive to atomic positions, geometry optimisation runs prior to the NMR calculation are highly recommended, as already mentioned above. In a crystal structure containing hydrogen atoms, their positions are the first to be optimised, starting from the experimentally obtained coordinates. The positions of other atom types are optimised whenever strong forces (in excess of 1 eV/Å) acted on these atoms as reported by Castep for the non-optimised structure. Our geometry optimisation calculations were performed using the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm [40], with the same functional, k-grid spacings and cut-off energies as in the single-point energy calculations. Convergence tolerance parameters for geometry optimization were as follows: maximum energy 2.0 × 10−5 eV/atom, maximum force 0.05 eV/Å, maximum stress 0.1 GPa and maximum displacement 0.002 Å. The calculations produce the quadrupolar coupling constant χ and asymmetry parameter η, which can be compared directly to the experimentally measured parameters.

5 Supporting information

Additional data about optical microscopy, powder x-ray diraction and solid-state NMR spectroscopy of SCTH are given as supplementary material available online (https://doi.org/10.1515/znb-2021-0009).


Corresponding author: Thomas Bräuniger, Department of Chemistry, University of Munich (LMU), Butenandtstr. 5–13, 81377Munich, Germany, E-mail:

Acknowledgments

The authors would like to thank Victor Terskikh (National Ultrahigh-Field NMR Facility for Solids, http://nmr900.ca) for access to the 21.1 T NMR spectrometer, and for his support in technical and theoretical matters. We are also grateful to Hans-Christian Böttcher (LMU München) for helpful advice on the intricate nomenclature of metal complexes.

  1. Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.

  2. Research funding: None declared.

  3. Conflict of interest statement: The authors declare no conflicts of interest regarding this article.

References

1. Cotton, S. A. Polyhedron 1999, 18, 1691–1715.10.1016/S0277-5387(99)00039-XSearch in Google Scholar

2. Badwal, S. P. S. J. Mater. Sci. 1987, 22, 4125–4132.10.1007/BF01133368Search in Google Scholar

3. Røyset, J., Ryum, N. Int. Mater. Rev. 2005, 50, 19–44.10.1179/174328005X14311Search in Google Scholar

4. George, N. C., Brgoch, J., Pell, A. J., Cozzan, C., Jaffe, A., Dantelle, G., Llobet, A., Pintacuda, G., Seshadri, R., Chmelka, B. F. Chem. Mater. 2017, 29, 3538–3546.10.1021/acs.chemmater.6b05394Search in Google Scholar

5. Liu, Y., Hu, S., Zhang, Y., Wang, Z., Zhou, G., Wang, S. J. Eur. Ceram. Soc. 2020, 40, 840–846.10.1016/j.jeurceramsoc.2019.10.022Search in Google Scholar

6. Rossini, A. J., Schurko, R. W. J. Am. Chem. Soc. 2006, 128, 10391–10402.10.1021/ja060477wSearch in Google Scholar PubMed

7. Kim, N., Hsieh, C.-H., Stebbins, J. F. Chem. Mater. 2006, 18, 3855–3859.10.1021/cm060590lSearch in Google Scholar

8. Zhang, L., Fehse, C., Eckert, H., Vogt, C., Hoffmann, R.-D., Pöttgen, R. Solid State Sci. 2007, 9, 699–705.10.1016/j.solidstatesciences.2007.05.010Search in Google Scholar

9. Chandran, C. V., Cuny, J., Gautier, R., Le Pollès, L., Pickard, C. J., Bräuniger, T. J. Magn. Reson. 2010, 203, 226–235.10.1016/j.jmr.2009.12.021Search in Google Scholar PubMed

10. Sen, S., Edwards, T., Kim, S. K., Kim, S. Chem. Mater. 2014, 26, 1918–1924.10.1021/cm4041852Search in Google Scholar

11. Heying, B., Haverkamp, S., Rodewald, U. C., Eckert, H., Peter, S. C., Pöttgen, R. Solid State Sci. 2015, 39, 15–22.10.1016/j.solidstatesciences.2014.11.001Search in Google Scholar

12. Bräuniger, T., Hofmann, A. J., Moudrakovski, I. L., Hoch, C., Schnick, W. Solid State Sci. 2016, 51, 1–7.10.1016/j.solidstatesciences.2015.11.002Search in Google Scholar

13. Jaworski, A., Charpentier, T., Stevensson, B., Eden, M. J. Chem. Phys. C 2017, 121, 18815–18829.10.1021/acs.jpcc.7b05471Search in Google Scholar

14. Rakhmatullin, A., Polovov, I. B., Maltsev, D., Allix, M., Volkovich, V., Chukin, A. V., Boča, M., Bessada, C. Inorg. Chem. 2018, 57, 1184–1195.10.1021/acs.inorgchem.7b02617Search in Google Scholar PubMed

15. Alba, M. D., Chain, P., Florian, P., Massiot, D. J. Phys. Chem. C 2010, 114, 12125–12132.10.1021/jp1036525Search in Google Scholar

16. Eckert, H., Pöttgen, R. Z. Anorg. Allg. Chem. 2010, 636, 2232–2243.10.1002/zaac.201000197Search in Google Scholar

17. Bräuniger, T., Jansen, M. Z. Anorg. Allg. Chem. 2013, 639, 857–879.10.1002/zaac.201300102Search in Google Scholar

18. Segall, M. D., Lindan, P. J. D., Probert, M. J., Pickard, C. J., Hasnip, P. J., Clark, S. J., Payne, M. C. J. Phys. Condens. Matter 2002, 14, 2717–2744.10.1088/0953-8984/14/11/301Search in Google Scholar

19. Clark, S. J., Segall, M. D., Pickard, C. J., Hasnip, P. J., Probert, M. J., Refson, K., Payne, M. C. Z. Kristallogr. 2005, 220, 567–570.10.1524/zkri.220.5.567.65075Search in Google Scholar

20. Charpentier, T. Solid State Nucl. Magn. Reson. 2011, 40, 1–20.10.1016/j.ssnmr.2011.04.006Search in Google Scholar PubMed

21. Pyykkö, P. Mol. Phys. 2008, 106, 1965–1974.10.1080/00268970802018367Search in Google Scholar

22. Cohen, M. H., Reif, F. Solid State Phys. 1957, 5, 321–438.10.1016/S0081-1947(08)60105-8Search in Google Scholar

23. Massiot, D., Fayon, F., Capron, M., King, I., Le Calve, S., Alonso, B., Durand, J., Bujoli, B., Gan, Z., Hoatson, G. Magn. Reson. Chem. 2002, 40, 70–76.10.1002/mrc.984Search in Google Scholar

24. Moudrakovski, I. L. Ann. Rep. NMR Spectrosc. 2013, 79, 129–240.10.1016/B978-0-12-408098-0.00004-5Search in Google Scholar

25. Apex3. Data Reduction and Frame Integration Program for the CCD Area-Detector System (version 2018.1-0); Bruker AXS Inc.: Madison, Wisconsin (USA), 2018.Search in Google Scholar

26. Sheldrick, G. M. Sadabs, Program for Empirical Absorption Correction of Area Detector Data (version 2.03); University of Göttingen: Göttingen (Germany), 2002.Search in Google Scholar

27. Sheldrick, G. M. Acta Crystallogr. 2008, A64, 112–122.10.1107/S0108767307043930Search in Google Scholar

28. Koller, H., Engelhardt, G., Kentgens, A. P. M., Sauer, J. J. Phys. Chem. 1994, 98, 1544–1551.10.1021/j100057a004Search in Google Scholar

29. Baram, A., Luz, Z., Alexander, S. J. Phys. Chem. 1976, 64, 4321–4328.10.1063/1.432118Search in Google Scholar

30. Schurko, R. W., Wi, S., Frydman, L. J. Chem. Phys. A 2002, 106, 51–62.10.1021/jp0130214Search in Google Scholar

31. Beerwerth, J., Storek, M., Greim, D., Lueg, J., Siegel, R., Cetinkaya, B., Hiller, W., Zimmermann, H., Senker, J., Böhmer, R. J. Magn. Reson. 2018, 288, 84–94.10.1016/j.jmr.2018.01.016Search in Google Scholar

32. Ilyukhin, A. B., Petrosyants, S. P. Zh. Neorg. Khim. 1994, 39, 1517–1520.Search in Google Scholar

33. Ripert, V., Hubert-Pfalzgraf, L. G., Vaissermann, J. Polyhedron 1999, 18, 1845–1851.10.1016/S0277-5387(99)00063-7Search in Google Scholar

34. Boyle, T. J., Sears, J. M., Neville, M. L., Alam, T. M., Young, V. G.Jr. Inorg. Chem. 2015, 54, 11831–11841.10.1021/acs.inorgchem.5b02030Search in Google Scholar PubMed

35. Bak, M., Rasmussen, J. T., Nielsen, N. C. J. Magn. Reson. 2000, 147, 296–330.10.1006/jmre.2000.2179Search in Google Scholar PubMed

36. Pickard, C. J., Mauri, F. Phys. Rev. B 2001, 63, 245101.10.1103/PhysRevB.63.245101Search in Google Scholar

37. Yates, J. R., Pickard, C. J., Mauri, F. Phys. Rev. B 2007, 76, 024401.10.1103/PhysRevB.76.024401Search in Google Scholar

38. Perdew, J. P., Burke, K., Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865–3868.10.1103/PhysRevLett.77.3865Search in Google Scholar PubMed

39. Monkhorst, H. J., Pack, J. D. Phys. Rev. B 1976, 13, 5188.10.1103/PhysRevB.13.5188Search in Google Scholar

40. Pfrommer, B. G., Côté, M., Louie, S. G., Cohen, M. L. J. Comput. Phys. 1997, 131, 233–240.10.1006/jcph.1996.5612Search in Google Scholar


Supplementary Material

The online version of this article offers supplementary material (https://doi.org/10.1515/znb-2021-0009).


Received: 2021-01-22
Accepted: 2021-02-03
Published Online: 2021-03-22
Published in Print: 2021-04-27

© 2021 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 29.3.2024 from https://www.degruyter.com/document/doi/10.1515/znb-2021-0009/html
Scroll to top button