Skip to main content
NCBI home page
Acta Crystallographica Section B: Structural Science, Crystal Engineering and Materials logoLink to Acta Crystallographica Section B: Structural Science, Crystal Engineering and Materials
. 2017 Oct 12;73(Pt 6):1032–1042. doi: 10.1107/S2052520617011490

Conformational and structural diversity of iridium di­methyl sulfoxide complexes

Benjamin M Ridgway a, Ana Foi a, Rodrigo S Corrêa b, Damian E Bikiel a, Javier Ellena c, Fabio Doctorovich a, Florencia Di Salvo a,*
PMCID: PMC5744404

Iridium(III) di­methyl sulfoxide complexes, analogues to clinically significant metallodrugs, exhibiting different rotamers in their solid-state structures are investigated. The effect of the amount of the crystallized water and the hydrogen-bonded network in their supramolecular structures on the conformational preferences is studied.

Keywords: solvent polymorphism, crystal engineering, metallodrug, rotamer

Abstract

Transition metal complexes containing dimethyl sulfoxide (DMSO) are important precursors in catalysis and metallodrugs. Understanding the solid-state supramolecular structure is crucial for predicting the properties and biological activity of the material. Several crystalline phases of DMSO-coordinated iridium anions with different cations, potassium (1a) and n-butyl­ammonium (1b), were obtained and their structures determined by X-ray crystallography. Compound (1a) is present in two solvatomorphic forms: α and β; the β form contains disordered solvent water. In addition, the structures exhibit different rotamers of the trans-[IrCl4(DMSO)2]− anion with the trans-DMSO ligands being oriented in anti and gauche conformations. In consideration of these various conformers, the effects of the crystallized solvent and intermolecular interactions on the conformational preferences of the anion are discussed. In addition, density functional theory calculations were used to investigate the energies of the anions in the different conformations. It was found that hydrogen bonds between water and the DMSO complex stabilize the gauche conformation which is the least stable form of the trans-DMSO complex. Consequently, by controlling the number of hydrogen-bond donors and acceptors and the amount of water, it may be possible to obtain different solvatomorphs of clinically significant metallodrugs.

1. Introduction  

Whereas the molecular structure of a drug is undoubtedly the most important feature determining its function and activity, the supramolecular structure of the bulk solid, including its crystallinity or amorphousness, affects its bioavailability, storage properties, solubility and activity. Increasingly, techniques such as continuous crystallization, co-crystallization and large-scale polymorph screening are being used to find new drug forms, improve drug properties and, contentiously, extend patents on existing drugs (Rossum, 1963 ▸; Stromgaard et al., 2009 ▸; Nagy & Braatz, 2012 ▸; Aaltonen et al., 2009 ▸; Ismail et al., 2013 ▸; Shan & Zaworotko, 2008 ▸). Consequently, an understanding of different crystal forms, whether polymorphs or solvatomorphs, is crucial.

Since the discovery of cisplatin, many metal complexes have been investigated for anticancer activity (Köpf-Maier, 1994 ▸; Sava et al., 2011 ▸). Unlike other platinum-group metals, iridium complexes were long thought to be less active against cancer cells (Buckley, 1994 ▸). However, recent work has shown that careful choice of auxiliary ligands can enable the use of Ir complexes as therapeutic agents (Geldmacher et al., 2012 ▸; Leung et al., 2013 ▸).

DMSO (dimethyl sulfoxide) is a common solvent used in transition metal chemistry. In addition, DMSO can coordinate to the metal centre, although it is a very labile ligand. This ease of coordination explains its frequent appearance in the single-crystal X-ray structures of many catalytic complexes. Aside from its unexpected appearances, DMSO metal complexes have applications as synthesis precursors. DMSO–metal complexes have also attracted attention because they have been shown to act in vitro as potent anticancer agents (Sava et al., 1992 ▸; Rademaker-Lakhai et al., 2004 ▸). Concerning its behaviour as a ligand, DMSO can coordinate via oxygen or sulfur. Early work on divalent metal complexes of DMSO suggested that oxygen coordination was more common, the exception being for Pd complexes, for which sulfur coordination was indicated (Selbin et al., 1961 ▸). However, since then, PdII complexes containing DMSO molecules coordinated through both oxygen and sulfur have been identified (Diao et al., 2012 ▸), sometimes in the same crystal (Wayland & Schramm, 1969 ▸; Abbasi et al., 2011 ▸).

The octahedral trans-bis­(di­methyl sulfoxide) complex, [IrCl4(DMSO)2]−, (1), was first reported by Haddad et al. (1974 ▸). They reported a pink–orange salt of (1), whose composition was determined by elemental and IR analyses to comprise (1) and a protonated DMSO dimer. They also investigated several other salts of (1) and found various anion geometries. Notably, yellow crystals containing a cis-coordinated form were found. Later, James et al. determined the structure of the rhodium analogue of (1), which contains protonated DMSO as the countercation or just ion (CSD refcode DMSRHD, P2/n; James et al., 1980 ▸). Cartwright et al. (1988 ▸) determined the X-ray crystal structure of (1); for that salt, as with the rhodium analogue determined by James et al., protonated DMSO was found to be the counterion (CSD refcode TAGREY, P2/c), forming short proton-bridged DMSO dimeric countercations. Then, a ruthenium analogue of (1) was determined by Alessio et al. (1991 ▸), and this structure was also found to contain a proton-bridged DMSO dimer. More recently, Messori et al. (2003 ▸) reported the structure of (1) (CSD refcode TAGREY01, P2/n) which is isostructural with the Rh analogues DMSRHD and TAGREY (although TAGREY was solved in P2/c).

Concerning the medical applications of metal–DMSO complexes, Cebrián-Losantos et al. (2007 ▸) determined the structures of several octahedral osmium tetrachloride complexes with various ligands trans to a coordinated DMSO ligand. These complexes are salts of pyrazine, triazine, and imidazole derivatives where the nitro­gen-containing heterocycle is coordinated trans-axially. In addition, they determined the structure of a trans-bis­(DMSO) complex (CSD refcode VIGKAY, Inline graphic), in which the countercation is also a protonated DMSO dimer, making this osmium complex isostructural with those of the Rh, Ir and Ru complexes (James et al., 1980 ▸; Cartwright et al., 1988 ▸; Alessio et al., 1991 ▸). Furthermore, these osmium complexes show antiproliferative activity against colon and mammary carcinoma cells.

In 2012, Abbasi et al. determined the structures of several related potassium ion-mediated coordination polymers (Abbasi et al., 2012 ▸). One of these (FANYEA, P21/c) is a rhodium DMSO complex containing disordered water (0.25 occupation) molecules (Haddad et al., 1974 ▸). Interestingly, FANYEA is the only potassium salt of a trans-IrL 2Cl4 (where L is another ligand) anion listed in the CSD, the other structures being either cis-anions or ammonium salts. In 1987, James et al. determined two polymorphic structures of the di­ethyl­ammonium salt of the rhodium analogue of (1), both of which formed by the acid hydrolysis of an enamine complex (Gamage et al., 1988 ▸).

Here, we report the X-ray crystal structures of several iridium–DMSO complexes, all of which were found to contain sulfur-coordinated DMSO ligands, and we present a different route for the preparation of (1), which is obtained via nucleophilic attack of n-butyl­amine on penta­chloro­nitro­syl iridate(III). Using this method, we have obtained crystal structures of two different potassium salts of the trans-isomer and an n-butyl­ammonium salt, all of which were obtained from the same reaction mixture. Concerning the potassium salts, because these compounds contain the same constituents but exhibit different amounts of crystallized solvent, they are not polymorphs but solvatomorphs, a term that has become popular to describe structures differing only in the number of solvent molecules (Brittain, 2009 ▸). We have used single-crystal and powder X-ray diffraction measurements (PXRD) and density functional theory (DFT) calculations to investigate the effect of solvent on the conformation of the trans-complexes of (1).

2. Experimental  

2.1. Synthesis  

K[IrCl5NO] was purchased from Strem and was purified by recrystallization in dry aceto­nitrile under an atmosphere of nitro­gen. n-Butyl­amine (99.5%) was purchased from Aldrich and used without further purification. DMSO was purchased from Aldrich. K[IrCl5(NO)] (20 mg, 0.044 mmol) was dissolved in DMSO under a nitro­gen atmosphere, forming a dark-brown solution. To this, n-butyl­amine (9.2 µL, 0.092 mmol) was added via a syringe, and the solution immediately became a light-orange colour. The solution was stirred for one day even though 1H NMR experiments indicate that the reaction completes in 4 h. Then, a few drops of water were added to the solution and it was left to evaporate. Yellow and yellow–orange crystals began to form after two weeks. The crystals were isolated from solution and inspected under a polarizing microscope and then analysed by single-crystal X-ray diffraction. Several batches of the reaction were studied and in all cases three different solvatomorphs were identified: trans-K[IrCl4(DMSO)2]·H2O (1aα), trans-K[IrCl4(DMSO)2]·­0.25H2O (1aβ) and trans-BuNH3[IrCl4(DMSO)2]·0.5H2O (1b). Although the colour of the different solvates was very similar, the crystals of (1b) were less cloudy than those of (1a). Furthermore, the morphologies were different, tetragonal prisms for (1a) and hexagonal crystals for (1b) (see a scheme of the procedure in the supporting information).

2.2. Crystallization experiments under relative humidity controlled atmosphere  

Vials containing DMSO solutions of [IrCl4(DMSO)2]− anions were placed in several sealed desiccators with different relative humidities (RH). Then, the samples were left for crystallization. The relative humidity of the closed desiccator was set by using saturated aqueous salt solutions at room temperature, i.e. 32% with CaCl2·6H2O, 45% with K2CO3, 63% with NH4NO3, 86% with KCl, and 95% with Na2HPO4·12H2O (Greenspan, 1977 ▸). After a few weeks, yellow–orange crystals were obtained in all cases, but single-crystal X-ray diffraction screening showed that the crystalline solvates were identical to those obtained by the crystallization methods described in §2.1 (see a scheme of the procedure in the supporting information).

2.3. Heating experiments  

Heating experiments were carried out on the crystals with the aim of exploring possible phase transformations on the loss of water. A crystal of (1aα) was mounted on a glass fibre with ep­oxy resin (Poxipol), and X-ray diffraction measurements were carried out. Then, the crystal, still mounted on the pin, was removed, clamped to a stand and heated with a heat gun. The temperature was monitored using a thermocouple held at the side of the crystal. After five minutes heating at approximately 453 K, the crystal was remounted on the goniometer and a short data collection was carried out. The X-ray diffraction results suggested a single-crystal to single-crystal transformation from the trans-K[IrCl4(DMSO)2]·H2O to a cis-K[IrCl4(DMSO)2] (2) complex. Due to the poor quality of data, hydrogen atoms of the water molecules were not assigned or refined (see CIF file and supporting information for details).

2.4. X-ray single-crystal diffraction  

Structural data were collected and treated as follows. The crystals were isolated from solution and inspected under a polarizing microscope. Suitable crystals were submerged in vegetable oil and cut down to a size suitable for X-ray diffraction experiments. The samples were mounted in a nylon loop and analysed either at room temperature or under a stream of cooled, dried air at 170 K. X-ray single-crystal diffraction experiments using Mo Kα (λ = 0.71073 Å) radiation were carried out using on a Nonius KappaCCD diffractometer and an Oxford Diffraction Gemini E [structure (2) only, see supporting information] diffractometer. A short exploratory data collection was made to assess crystal quality and to determine the unit cell and then, the full data collection was performed. The Oxford Diffraction Gemini E diffractometer [structure (2)] had a generator voltage and current of 50 kV and 40 mA, respectively. The data collection was planned and executed and the data were reduced using CrysAlisPro (Agilent, 2013 ▸). The Nonius Kappa-CCD diffractometer [structures (1a), (1aβ) and (1b)] had generator voltage and current of 55 kV and 33 mA, respectively. Data collections were made using the COLLECT program (Enraf–Nonius, 1997 ▸) and the data were reduced using the DENZO and SCALEPACK programs (Otwinowsky & Minor, 1997 ▸) and corrected Lorentz–polarization corrections were made. The data were scaled and merged using SORTAV (Blessing, 1995 ▸) for (1aα) and (1aβ). A Gaussian method (Coppens et al., 1965 ▸) implemented in WinGX (Farrugia, 1999 ▸) was used for the absorption correction. Structure solution was carried out using direct methods (SHELXS; Sheldrick, 2008 ▸). Refinement was performed using full-matrix least squares as implemented in WinGX (Farrugia, 1999 ▸) and OLEX2 (Dolomanov et al., 2009 ▸). The CIF was prepared using SHELXLE (Hübschle et al., 2011 ▸) with SHELXL2016 (Sheldrick, 2015 ▸).

2.5. Powder diffraction data  

Samples were characterized by PXRD using a graphite-filtered Cu Kα radiation (λ = 1.5406 Å) on a Siemens D5000 diffractometer.

2.6. Computational details  

DFT calculations were made using parallel ORCA (version 3.0.2) (Neese, 2012 ▸) running on an eight-core AMD desktop computer running Ubuntu 14.04 LTS. Dispersion-corrected (Grimme et al., 2011 ▸, 2010 ▸) geometry optimizations were performed using the B3LYP functional and the Alhrich’s def2-SVP basis set (Schäfer et al., 1992 ▸, 1994 ▸) on all atoms except Ir, which was modelled with the def2-TZVP valence basis set and Stuttgart–Dresden ECP (Weigend & Ahlrichs, 2005 ▸; Andrae et al., 1990 ▸). ECP parameters implemented in ORCA for Ir [SD(60,MWB)] were obtained from the pseudopotential library of the Stuttgart/Cologne group (http://www.theochem.uni-stuttgart.de/pseudopotentials/), whereas ECP basis sets for Ir [DEF2-TZVP] were obtained from the Turbomole basis set library (ftp://ftp.chemie.uni-karlsruhe.de/pub/basen/). Tight SCF convergence criteria were used for geometry optimizations. Coordinate and input file data are available in the supporting information.

3. Results  

Fig. 1 ▸ shows the different rotamers found in each crystal structure, and Figs. 2 ▸ and 3 ▸ summarize the different crystalline forms obtained. Throughout the paper, we refer to the anions by the structure name (1a) and (1b), solvatomorph type (α and β), and rotamer type (gauche and anti). In the following text, we define the rotational isomers as anti, gauche and syn, based on the torsion angle made by the S=O bonds of the trans DMSO ligands. This labelling scheme and the descriptive angles used are shown in Figs. 1 ▸ and S2, respectively. The angles defined in Fig. S2 were used to carry out DFT driver calculations to explore the rotational potential energy surface.

Figure 1.

Figure 1

Open in a new tab

Rotational conformations of the trans-anion: (a) anti, (b) gauche and (c) syn. The definition is dependent on the OCCO dihedral angle and independent of rotation of the chloride plane. Colour code: S, yellow; C, black; H, light grey; O, red; Cl, green. Iridium centres are under the sulfur atoms.

Figure 2.

Figure 2

Open in a new tab

Reaction pathway to obtain (1).

Figure 3.

Figure 3

Open in a new tab

Summary of the crystalline forms and solvatomorphism for (1).

3.1. Formation of trans-[IrCl4(DMSO)2 (1)  

The anion (1) was obtained through the reaction of K[IrCl5(NO)] and n-butyl­amine. As illustrated in Fig. 2 ▸, one equivalent of the amine attacks the NO bonded to the IrIII and a second equivalent is protonated by the generated proton. Consequently, the first step in this reaction is the formation of a labile N-nitroso intermediate; this is then transformed into the corresponding free alcohol after a second nucleophilic attack by a hydroxyl anion generated during the reaction. Finally, the alcohol is lost from the complex and replaced by a solvent molecule, which, in this case, is DMSO (Di Salvo et al., 2008 ▸; Doctorovich & Di Salvo, 2007 ▸). The complex now contains one DMSO ligand, and this may exert a higher trans effect than that of the N-nitroso moiety on the trans-chloride, resulting in loss of the corresponding chloride and coordination of a DMSO molecule to the metal centre. The trans-effect of DMSO ligands is well known; for instance, the trans-effect on square-planar platinum complexes containing DMSO has been previously investigated by Kukushkin (1971 ▸). DFT calculations indicate that the trans-Ir—Cl bond lengthens on the coordination of DMSO (Fig. S3). Thus, the trans-DMSO ligand, (1), is formed.

The conversion of the starting complex into [IrCl4(DMSO)2]− in solution was studied using 1H NMR spectroscopy. After 4 h at room temperature, the intermediates disappeared, the alcohol was generated and the ammonium salt formed (Fig. 2 ▸). Consequently, in comparison with the previously reported synthesis (Haddad et al., 1974 ▸), the desired ion is obtained under mild conditions without the need for concentrated acid.

After leaving the reaction mixture at room temperature and open to the air for two weeks, bright-orange crystals were obtained. The analysis of the resultant material by optical microscope showed two different crystalline forms: hexagonal and tetragonal prisms. Both types of crystal were carefully separated, mounted on a suitable pin (or nylon loop), and analysed using single-crystal X-ray diffraction. Crystallographic data are listed in Table 1 ▸. In our initial screening, the tetragonal prisms were found to be potassium salt hydrates of (1) (Fig. 2 ▸), whereas the hexagonal crystals were found to comprise the n-butyl­ammonium salt (1b). Different crystallization experiments were carried but all of them led to the same kind of crystalline material (see Experimental for details). Fig. 1 ▸ shows the different rotamers found in each crystal structure, and Fig. 3 ▸ summarizes the different crystalline forms obtained. Throughout the paper, we refer to the anions by the structure name (1a) and (1b), solvatomorph type (α and β), and rotamer type (gauche and anti).

Table 1. X-ray crystal structure data for structures (1aα), (1aβ) and (1b).

  (1aα) (1aβ) (1b)
Chemical formula C4H12Cl4IrO2S2·K·H2O C4H12Cl4IrO2S2·K·0.25(H2O) C4H12Cl4IrO2S2·C4H12N·0.5(H2O)
M r 547.37 533.86 573.41
Crystal system, space group Monoclinic, P21/c Monoclinic, P21/c Monoclinic, C2/c
Temperature (K) 293 293 293
a, b, c (Å) 12.1963 (2), 7.4652 (2), 16.4793 (4) 12.7530 (4), 7.8356 (2), 15.3277 (5) 19.8210 (4), 10.0830 (2), 19.5720 (5)
β (°) 96.055 (1) 111.581 (2) 108.657 (2)
V (Å3) 1492.03 (6) 1424.29 (7) 3706.01 (14)
Z 4 4 8
Radiation type Mo Kα Mo Kα Mo Kα
μ (mm−1) 10.21 10.69 8.01
Crystal size (mm) 0.38 × 0.20 × 0.12 0.16 × 0.11 × 0.09 0.21 × 0.12 × 0.11
Diffractometer KappaCCD KappaCCD KappaCCD
Absorption correction Gaussian (Coppens et al., 1965 ▸) Gaussian (Coppens et al., 1965 ▸) Gaussian (Coppens et al., 1965 ▸)
T min, T max 0.064, 0.292 0.134, 0.517 0.263, 0.659
No. of measured, independent and observed [I > 2σ(I)] reflections 8722, 2614, 2582 8611, 2490, 1781 18983, 3424, 3050
R int 0.061 0.104 0.099
(sinθ/λ)max (Å−1) 0.595 0.595 0.606
R[F 2 > 2σ(F 2)], wR(F 2), S 0.030, 0.062, 1.13 0.046, 0.119, 0.95 0.051, 0.132, 1.05
No. of reflections 2614 2490 3424
No. of parameters 141 143 176
H-atom treatment H-atom parameters constrained H-atom parameters constrained H-atom parameters constrained
Weighting scheme [w = 1/[σ2(F o 2) + (0.9117P)2 where P =(F o 2 + 2F c 2)/3] w = 1/[σ2(F o 2) + (0.0507P)2 where P = (F o 2 + 2F c 2)/3 w = 1/[σ2(F o 2) + (0.074P)2 + 31.6594P] where P = (F o 2 + 2F c 2)/3
Δρmax, Δρmin (e Å−3) 0.70, −0.70 1.73, −2.90 3.53, −1.96
Open in a new tab

3.2. Potassium trans-coordinated salts (1a)  

Two hydrates of the potassium salt were found, labelled (1aα) and (1aβ). Both crystallize in the same space group, P21/c. The unit-cell parameters and Z for both structures are given in Table 1 ▸. Bond distances and bond angles for (1aα) are given in Table S3 and those for (1aβ) are given in Table S4. The asymmetric unit of the (1aα)-hydrate contains one water molecule per anion and one gauche-DMSO coordinated anion; in contrast, the (1aβ)-hydrate contains disordered water. During refinement, the U-factor of the water molecule increased; consequently, its occupancy was refined to 0.25 with a reasonable ellipsoid. Because there is neither a clear void space nor a channel through which the solvent can move, the water must be statistically distributed over one in four unit cells. The DMSO ligands are oriented in an anti-orientation (Fig. 4 ▸). (1aβ) contains two half anions in the asymmetric unit, labelled (I) and (II), which both sit on an inversion centre and are, thus, trans-anti-coordinated. However, the coordinated DMSO ligands have slightly different tilt angles, i.e. the O—S—Ir angles for (1aβI) and (1aβII) differ slightly: 115.68 and 115.98°, respectively. The α hydrate contains only one crystallographically unique anion; here, the torsion angle is −66.03° and the O—S—Ir tilt angle is 116.02° (Fig. 4 ▸ d).

Figure 4.

Figure 4

Open in a new tab

Ellipsoid plots at 50% probability, showing the different rotamers: (a) (1aα) (gauche), (b) and (c) (1aβ) (both anti), and (d) and (e) (1b) (I, gauche; II, anti).

3.3. n-Butyl­ammonium trans-coordinated salt (1b)  

The n-butyl­ammonium hydrate, (1b), crystallizes in the space group C2/c. The unit-cell parameters and Z. The unit-cell parameters and Z are given in Table 1 ▸. The bond distances and bond angles are given in Table S2. The unit cell contains two crystallographically unique anions. In each anion, the iridium cation is positioned on a special position: the inversion centre. Thus, the unit cell contains two crystallographically unique half anions, and the two anions have different rotamer configurations. In (1bI), the DMSO ligands are gauche, whereas, in (1bII), an anti-conformation is observed. For (1bI) and (1bII), the torsion angles are −62.94 and −180°, respectively, and the OSIr angles are 114.61 and 117.09°, respectively. Fig. 4 ▸ shows the different anions viewed along the S—Ir—S axis. The crystal structures contain two rotamers of the trans-anion: gauche and anti [(I) and (II) respectively]. The syn rotamer was not observed. Furthermore, a search of the CSD for (1) shows that, of trans complexes, only gauche and anti rotamers have been found previously, as shown in Fig. S4, suggesting that the syn rotamer is unfavourable.

3.4. Analysis of the relative stability of the trans-coordinated rotamers  

Relaxed surface scans were carried out at the DFT/B3LYP level of theory using the def2-SVP basis set for the C, H, O and S atoms and the def2-TZVP basis set for the metal. An SDD effective core potential was placed on Ir (see the Experimental , Fig. 5 ▸ and supporting information for details). The calculations reveal three minima: the anti-rotamer being the global minimum energy structure, followed by a slightly higher minimum for the gauche structure, and finally, the syn rotamer being the highest energy local minimum. The minima are separated by low energy barriers that arise due to the steric clash between the equatorial chloride ligand and the methyl groups of DMSO. However, these barriers are low, approximately 12.5 kJ mol−1, an energy commensurate with that of rotation about a single C—C bond (Yirong, 2011 ▸; Smith & Jaffe, 1996 ▸). Therefore, as expected, rotation in solution is probably facile and rapid, and deviations from the gas-phase minimum energy structures can be expected to be easily compensated by packing effects and the formation of energetically favourable contacts, such as hydrogen bonds to water molecules and electrostatic interactions with large solvated cations. The DFT optimized structure agrees well with the X-ray crystal structure of DMSO (Thomas et al., 1966 ▸), and these differ from the structure of DMSO in the anion mainly by having slightly shorter bond lengths [for example, 1.47 versus 1.52 Å in (1β) and the calculated structure, respectively] possibly due to packing effects; indeed, computational studies have shown that significant distortion to bond lengths and bond angles can occur in DMSO because of intermolecular interactions (Mezey & Kapur, 1980 ▸). In addition, the pyramidalization of the DMSO molecule is reduced on coordination, as indicated by the increased C—S—C and C—S—O bond angles.

Figure 5.

Figure 5

Open in a new tab

Relaxed potential energy scan. Minima correspond to the syn, anti and gauche conformations, decreasing in energy in that order. Black dots indicate the calculated data points. The green curves are deconvoluted Gaussian curves fitted to the calculated data points. The red curve is the sum of the deconvoluted green Gaussian curves.

To investigate the importance of the water molecules in the crystal structure and its role with the rotamer stabilization, we attempted to remove water from a crystal by heating. For that experiment we chose a crystal of (1aα). On heating, the crystal became slightly opaque and lost its lustre, the surface of the crystal became roughened and its shape became rounded. In the subsequent data collection, the reflections were not sharp and some powder rings were present. Nevertheless, the data were of sufficient quality for structure refinement and the crystal was found to have crystallized in the monoclinic space group P21/n. To our surprise, the cell parameters were different from those of (1), the crystal structure did not contain water and, more importantly, the iridium complex had changed from trans to cis geometry [named here as (2)]. To confirm this single-crystal to single-crystal transformation, other crystals of (1aα) and (1aβ) were heated using the procedure outlined above. However, subsequent X-ray analysis showed, despite the poor-quality data, that the crystal structure remained that of the starting compound. We were unable to repeat this transformation in subsequent experiments. DSC experiments performed in (1aβ) (the main phase) were not conclusive regarding a single-crystal to single-crystal transformation on the bulk. Bulk PXRD measurements show that (1) is the only phase present. Possibly, a crystal of the cis-phase grew alongside that of the main phase, (1), and the heating of the crystal reduced the crystallinity of the main phase, allowing the cis-phase to be detected (Fig. S5).

3.5. Discussion of the supramolecular structures and the role of the hydrogen bond on the rotamer stability  

The α hydrate of (1a) contains one unique anion, and the DMSO ligands are trans to each other. In addition, one molecule of water per unit cell is present. This water molecule forms a hydrogen bond with the oxygen atom of the coordinated DMSO ligand. Its molecular structure presents a distorted gauche conformation of the DMSO molecules coordinated to the octahedral IrIII centre, as shown in Fig. 6(a) ▸. The shortest intermolecular interactions in this structure are the hydrogen bonds between water molecules (O1w) and oxygen atoms from coordinated DMSO, i.e. O2⋯O1w = 2.989 Å and O—H⋯O = 176.50° (Fig. 6 ▸). Hydrogen bonds can be classified as classical and non-classical. Classical hydrogen bonds contain those from donor heteroatoms such as nitro­gen, oxygen and halogens. These hydrogen bonds can be further classified as strong, medium or weak. This classification is often made based on the separation between the atoms and the angle made by the D—H⋯A hydrogen bond. For common hydrogen bonds, this scheme works reasonably well. However, in very well packed structures, such as those discussed in this paper, it is often difficult to make a clear assessment of the hydrogen-bond energetics (Steiner, 2002 ▸). In addition, H1w forms an asymmetrically bifurcated hydrogen bond with the chloride ligands of a neighboring iridium complex. Thus, the hydrogen atoms form hydrogen bonds to both Cl1 and Cl4.

Figure 6.

Figure 6

Open in a new tab

Detail of the (a) coordination environment of potassium (K1) in (1aα), and (b) the centrosymmetric hydrogen-bonded dimer that forms the sheet-like structure shown in Fig. 6 ▸(a). Not shown are hydrogen bonds from O1w (via H1w) to neighbouring chloride ligands (Cl1 and Cl4).

Concerning the coordination environment of the potassium cation, K1 is surrounded by eight ligands: one water molecule, three oxygen atoms from two symmetry-related DMSO molecules and four chloride ligands from two symmetry-related complex anions (Fig. 6 ▸). The symmetry-related potassium cations are separated by 4.3 Å. A centrosymmetric hydrogen-bonded dimer is formed by the water molecule (O1w), which connects the two symmetry-related parts of the dimer through a hydrogen bond to a DMSO oxygen (O1w⋯O2); this DMSO oxygen also forms a contact with the potassium (K1) and the second DMSO oxygen (O1) forms short contacts with potassium only (Fig. 6 ▸ b). Expanding the hydrogen-bonded contacts leads to the formation of a sheet-like structure, as shown in Fig. 7 ▸. Fig. 7 ▸(a) shows the top of this sheet-like structure viewed along the a axis. Fig. 7 ▸(c) shows the side view of the sheet observed along the c axis, clearly showing the sandwich-like structure of the anions on the outside and a ‘filling’ of potassium cations. The exterior surface of the sheet is rich in methyl groups and chloride anions. These sheets stack together, as shown in Fig. 7 ▸(b), sustained by CH⋯Cl short contacts.

Figure 7.

Figure 7

Open in a new tab

View along the a axis of (1aα); (b) stacking of these sheets along the a axis, note the methyl–methyl contacts and methyl–chloride contacts; (c) view along the c axis showing potassium cations sandwiched between anions; and (d) perspective view showing the methyl- and chloride-rich outer layers.

As in the (1aα) crystal structure, (1aβ) contains solvent water, a trans-anion and a potassium cation. The potassium cation, as for (1aα), is octahedrally coordinated (Fig. 8 ▸). However, in contrast with the α-form, it forms three contacts to oxygen (one from water and two from DMSO) and five close contacts to chloride. Also, the water of crystallization is disordered with an overall occupancy of 25% per water site.

Figure 8.

Figure 8

Open in a new tab

Coordination environment around potassium in (1aβ). Only partial anions are shown.

The iridium complex in (1aα) crystallized with only one form of the trans-isomer, gauche; in contrast, the (1aβ) form crystallized with two independent anions in the asymmetric unit. Both anions are the anti-rotamer, and the iridium atoms sit on an inversion centre. Again, as for the α-form, an extended sheet-like structure is observed (similar to Fig. 7 ▸ a) with a core of potassium and chloride anions (similar to Fig. 5 ▸ c). Unlike (1aα), because the supramolecular sheets are composed of iridium complexes, an infinite two-dimensional network of K···Cl—Ir—Cl⋯K contacts is present. In the sheet-like structure, electrostatic interactions between potassium (K1) and the chloride ions of the iridium complex (Cl3, Cl4) form a four-membered cyclic structure. The chain is propagated by the inversion symmetry through Ir2, and the chain continues by contacts through K1 and Cl2 (and the symmetry equivalent of Cl2), forming a second four-membered ring (see Fig. 8 ▸ for a close-up of these short contacts). The K⋯Cl—Ir2—Cl⋯K chains run along [100], thus forming the sheet-like structure in the (002) plane, as shown in Fig. 9 ▸. The distances between K1 and Cl2, Cl2*, Cl3 and Cl4 are 3.229, 3.279, 3.218 and 3.272 Å, respectively. In addition, K1 makes short contacts with O1, O2 and O1w, which have separations of 2.826, 2.853 and 3.061 Å, respectively. These layers stack along the c axis (Fig. 9 ▸) but in (1aα), the hydrogen-bonded sheets are quite separated (Fig. 7 ▸ b) with no hydrogen bonds between the sheets. In contrast, for (1aβ) the DMSO methyl groups form an interdigitated structure (Fig. 9 ▸).

Figure 9.

Figure 9

Open in a new tab

Infinite sheets showing methyl group interdigitation in (1aβ). Note the position of the water molecules indicated by pink circles.

The n-butyl­ammonium salt (1b) also crystallized with a solvent water molecule of crystallization. The structure, as with (1aβ), contains two independent anions. Each of these shows different conformations of DMSO molecules. These are present in the anti-conformation and the gauche-conformation (Fig. 4 ▸). There is a hydrogen-bonded water molecule which has been modelled having a half-occupancy. Unlike the n-butyl­ammonium structure determined by us, James’ structure [di­ethyl­ammonium salt of the rhodium analogue of (1)] does not contain water (CSD refcode GEZBOC Inline graphic; Gamage et al., 1988 ▸). GEZBOC forms zigzag hydrogen-bonded chains of trans-anions linked via di­ethyl­ammonium. The hydrogen bonds are single and formed between the ammonium proton and the DMSO oxygen atom (D Inline graphic). GEZBOC01 (Inline graphic) also forms hydrogen-bonded zigzag chains, but the hydrogen bonds are bifurcated, forming R Inline graphic(5) rings (Etter et al., 1990 ▸).

n-Butyl­ammonium cations can exist in a variety of conformations, but the linear conformation is the global energy minimum. However, rotations about single bonds are low in energy and in (1b), the cation is not linear (τ1 = τ2 = 180°) but has a slightly bent conformation (τ1 = −74.07 and τ2 = −174.31°, almost linear). As expected, the n-butyl­ammonium cation forms several hydrogen bonds, for example, with a water molecule and the equatorial chloride ligands (Figs. 10 ▸ and 11 ▸). These hydrogen bonds are listed in the supporting information. Chloride is a good, although non-selective, hydrogen-bond acceptor, and thus an R Inline graphic(4) graph-set hydrogen-bonded ring is formed (see Fig. S6). Other hydrogen-bonded rings are formed by the water that hydrogen bonds to both the n-butyl­ammonium cation and an equatorial chloride, in each type of anion [(1bI) and (1bII), Fig. S6]. In the anti-anion, the four equatorial chlorides interact in pairs with an n-butyl­ammonium ion giving place to R Inline graphic(4) rings as mentioned, and then an N—H⋯O hydrogen-bond between two cations and a water molecule is observed (Fig. 12 ▸). For the gauche-anion, two of the four equatorial chlorides interact with two n-butyl­ammonium ions and the other two with a molecule of water, which also interacts with the following pair of n-butyl­ammonium cations (Fig. 13 ▸). As a result, hydrogen-bonded rings Inline graphic(10) and Inline graphic(6) are observed. The gauche form of the anion also results in hydrogen-bonded chains perpendicular to the a axis. These again are propagated through the n-butyl­ammonium cation and water, forming the column shown in Figs. 13 ▸ and 14 ▸. These hydrogen bonds propagate to form a two-dimensional hydrogen-bonded sheet in the ac plane, shown in Fig. 11 ▸. n-Butyl­ammonium–water–chloride hydrogen bonds can be seen propagating in the vertical direction in the figure and the chloride–water hydrogen bonds propagate in the horizontal direction. An N—H bond of n-butyl­ammonium also forms a hydrogen bond to O1DMSO. In addition, there are also several CH⋯O close contacts, notably C1—H1A⋯O2 and C1—H1B⋯O2, both methyl CH from adjacent coordinated DMSO molecules. Short contacts between C—H⋯Cl and C—H⋯O are also present, as listed in Tables S4 and S5.

Figure 10.

Figure 10

Open in a new tab

Hydrogen-bonded network in (1b) containing Inline graphic hydrogen-bonded rings between butylammonium NH and the chloride ligands and Inline graphic rings between DMSO oxygen and the butylammonium NH (see Fig. S6 for further details). (a) View along the b axis and (b) the view along [101].

Figure 11.

Figure 11

Open in a new tab

Sheet-like structure in (1b). The sheet is shown viewed perpendicular to the b axis. The unit cell is shown superimposed. The view along the c axis is shown at the top of the image, and that along the a axis is shown to the left of the image. Anions (1bI) and (1bII) are indicated.

Figure 12.

Figure 12

Open in a new tab

The hydrogen-bonded chain formed between anti-anion (1bII), n-butyl­ammonium and water.

Figure 13.

Figure 13

Open in a new tab

A hydrogen-bonded chain containing the gauche-conformer [1bI], n-butyl­ammonium and water.

Figure 14.

Figure 14

Open in a new tab

A sheet of hydrogen-bonded (1bI) anions. Hydrogen bonds are shown as dashed lines. The hydrogen-bonded column rising from the sheet is the same as that shown in Fig. 13 ▸.

3.6. Comparison of the water content of the crystal structures  

(1aα) contains one water molecule, (1aβ) contains 0.25 water molecules. (1b), in which the counterion is n-butyl­ammonium, has a water molecule with a modelled occupancy of 0.5. In (1b), the water molecule forms a common hydrogen-bonding motif, i.e. a bifurcated hydrogen bond shared between two chloride ligands of anion (1). To better understand the differences in the water content of these structures, we calculated the volume of the water molecules in these structures, as listed in Table 1 ▸. Generally, a water of crystallization occupies about 36 Å3. Table 2 ▸ shows the volume of ‘free’ space in the structures; the cavity in (1b) is small; thus, only half the possible spaces are occupied. (1aα) has a larger cavity and has an occupancy of one. However, (1aβ) has a similar cavity volume to (1aα), yet has even lower water occupancy.

Table 2. Cavity volumes occupied by the water of crystallization.

  Complete structure Water deleted          
  Void volume (%) Void volume (Å3) Void volume (%) Void volume (Å3) ‘Cavity’ size (Å) Z Refined occupancy Volume per water (Å3) Volume per asymmetric unit (Å3)
(1b) 2.6 97.98 4.5 166.74 68.74 8 0.5 17.19 8.59
(1aβ) 6.4 90.86 9.8 139.7 48.84 4 0.25 12.21 12.21
(1aα) 12.1 180.85 15.7 234.23 53.38 4 1 13.35 13.35
Open in a new tab

The Kitaigorodski packing index is a measure of the packing density of a crystal structure. Molecular crystals typically have packing indices ranging from 65 to 77% (Kitaigorodsky, 1973 ▸; van der Sluis & Spek, 1990 ▸). However, structures containing voids and disordered solvent may have significantly lower values. The packing indices for these structures are listed in Table S1 and are typical values for molecular crystals. Interestingly, the packing index for the n-butyl­ammonium structure is lower than those of the potassium salts, and the packing index of (1aβ), which contains disordered water, is high. A possible explanation for the presence of disordered water is that the most energetically favourable packing leaves a small void; this is occupied by water in one out of four asymmetric units, reducing void space but not disrupting the closely packed structure. This structure (1aα) may be the kinetic product, where the less energetically favourable gauche geometry is stabilized by the presence of a hydrogen bond; in contrast, (1aβ) is the thermodynamic product.

To identify the major phase, crystals obtained by slow evaporation of DMSO were screened by selecting each crystal, cutting it to size, mounting the fragment in a nylon loop, and identifying the unit-cell parameters by collecting several frames. The crystals were separated into groups, based on their unit-cell parameters, and we found that most crystals were (1aβ) (one-quarter of a molecule of water per asymmetric unit). The crystals were then ground and the PXRD data collected (Fig. S7) show that the only X-ray visible phase is that of (1aβ), i.e. the other phases must be present, but at such low concentrations that their contribution to diffraction is unobservable.

4. Conclusions  

DMSO is a labile ligand that can ligate vacant sites on many metal complexes, yielding precursor compounds that can be converted or metabolized in vitro to active metallodrugs. The polymorphic and solvatomorphic behaviour of these complexes is thus of great interest, having consequences for the solubility, bioactivity and storage properties of these compounds. We have described a simple route to obtain complex (1) as potassium and ammonium salts containing different amounts of solvent water. In particular, the two potassium salts include one, (1aα), and one-quarter, (1aβ), of a water molecule per asymmetric unit. Published structures of trans-bis­(DMSO) complexes and DFT calculations indicate that the anti-conformation is the global minimum energy rotamer, and the bond lengths and bond angles in these compounds are typical, based on previous crystallographic studies and calculations. This occurrence of a less favourable rotamer suggests that the formation of a hydrogen-bonded network mediated by water in (1aα) stabilizes the less favourable gauche-conformation. However, the most commonly observed structure was that of (1aβ); in that structure, the packing index is very high, indicating the optimum use of space and, consequently, the lack of a significant hydrogen-bonding network and the presence of the low-energy anti-rotamer. In contrast, (1b) has a low packing index, one more typical of wholly organic crystal structures, and contains half a molecule of water per asymmetric unit. Moreover, in (1b), both crystallographically independent anions are anti, suggesting that at medium packing indices and in the presence of hydrogen-bond donors and acceptors, the gauche conformation is stabilized. Furthermore, in the high packing index low water state, the anti-conformation dominates and is the dominant structure, whereas in the presence of many hydrogen-bond donors and acceptors (water and n-butyl­ammonium), both the low-energy anti-rotamer and a gauche rotamer are present, and extensive hydrogen bond formation compensates for the less tightly packed structure. These results suggest that, if the number of hydrogen-bond donors and acceptors and the amount of water could be controlled, different solvatomorphs of clinically significant metallodrugs could be obtained.

Supplementary Material

Crystal structure: contains datablock(s) 1a_alpha, 1a_beta, 1b, 2. DOI: 10.1107/S2052520617011490/lo5014sup1.cif

b-73-01032-sup1.cif (2.7MB, cif)

Structure factors: contains datablock(s) 1a_alpha. DOI: 10.1107/S2052520617011490/lo50141a_alphasup2.hkl

Structure factors: contains datablock(s) 1a_beta. DOI: 10.1107/S2052520617011490/lo50141a_betasup3.hkl

Structure factors: contains datablock(s) 1b. DOI: 10.1107/S2052520617011490/lo50141bsup4.hkl

b-73-01032-1bsup4.hkl (273.7KB, hkl)

Structure factors: contains datablock(s) 2. DOI: 10.1107/S2052520617011490/lo50142sup5.hkl

b-73-01032-2sup5.hkl (206.7KB, hkl)

Supplementary figures and schemes. DOI: 10.1107/S2052520617011490/lo5014sup6.pdf

b-73-01032-sup6.pdf (1.7MB, pdf)

CCDC references: 1566729, 1566730, 1566731, 1566732

Funding Statement

This work was funded by CONICET grants PICT2012-1335, UBACYT 2014-2017 GC, and No. 20020130100642BA. CNPq grant . FAPESP grant . FAPEMIG grant .

References

  1. Aaltonen, J., Allesø, M., Mirza, S., Koradia, V., Gordon, K. C. & Rantanen, J. (2009). Eur. J. Pharm. Biopharm. 71, 23–37. [DOI] [PubMed]
  2. Abbasi, A., Geranmayeh, S., Skripkin, M. Y. & Eriksson, L. (2012). Dalton Trans. 41, 850–859. [DOI] [PubMed]
  3. Abbasi, A., Skripkin, M. Y., Eriksson, L. & Torapava, N. (2011). Dalton Trans. 40, 1111–1118. [DOI] [PubMed]
  4. Agilent (2013). CrysAlisPRO. Agilent Technologies Ltd, Yarnton, Oxfordshire, England.
  5. Alessio, E., Balducci, G., Calligaris, M., Costa, G., Attia, W. M. & Mestroni, G. (1991). Inorg. Chem. 30, 609–618.
  6. Andrae, D., Häussermann, U., Dolg, M., Stoll, H. & Preuss, H. (1990). Theor. Chim. Acta, 77, 123–141.
  7. Blessing, R. H. (1995). Acta Cryst. A51, 33–38. [DOI] [PubMed]
  8. Brittain, H. G. (2009). Polymorphism in Pharmaceutical Solids, edited by H. G. Brittain. 2nd ed. Boca Raton, FL: CRC Press.
  9. Buckley, R. G. (1994). Rhodium, Iridium and Palladium Compounds as Experimental Anticancer Drugs In Metal Compounds in Cancer Therapy, edited by S. P. Fricker, pp. 92–108. Netherlands: Springer.
  10. Cartwright, P. S., Gillard, R. D., Sillanpaa, E. R. J. & Valkonen, J. (1988). Polyhedron, 7, 2143–2148.
  11. Cebrián-Losantos, B. K., Krokhin, A. A., Stepanenko, I. N., Eichinger, R., Jakupec, M. A., Arion, V. B. & Keppler, B. K. (2007). Inorg. Chem. 46, 5023–5033. [DOI] [PubMed]
  12. Coppens, P., Leiserowitz, L. & Rabinovich, D. (1965). Acta Cryst. 18, 1035–1038.
  13. Diao, T., White, P., Guzei, I. & Stahl, S. S. (2012). Inorg. Chem. 51, 11898–11909. [DOI] [PMC free article] [PubMed]
  14. Di Salvo, F., Estrin, D. A., Leitus, G. & Doctorovich, F. (2008). Organometallics, 27, 1985–1995.
  15. Doctorovich, F. & Di Salvo, F. (2007). Acc. Chem. Res. 40, 985–993. [DOI] [PubMed]
  16. Dolomanov, O. V., Bourhis, L. J., Gildea, R. J., Howard, J. A. K. & Puschmann, H. (2009). J. Appl. Cryst. 42, 339–341.
  17. Enraf–Nonius (1997). COLLECT. Nonius BV, Delft, The Netherlands.
  18. Etter, M. C., MacDonald, J. C. & Bernstein, J. (1990). Acta Cryst. B46, 256–262. [DOI] [PubMed]
  19. Farrugia, L. J. (1999). J. Appl. Cryst. 32, 837–838.
  20. Gamage, S. N., James, B. R., Rettig, S. J. & Trotter, J. (1988). Can. J. Chem. 66, 1123–1128.
  21. Geldmacher, Y., Oleszak, M. & Sheldrick, W. S. (2012). Inorg. Chim. Acta, 393, 84–102.
  22. Greenspan, L. J. (1977). J. Res. Natl. Bur. Stand. Sect. A, 81A, 89–96.
  23. Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. (2010). J. Chem. Phys. 132, 154104–154119. [DOI] [PubMed]
  24. Grimme, S., Ehrlich, S. & Goerigk, L. (2011). J. Comput. Chem. 32, 1456–1465. [DOI] [PubMed]
  25. Haddad, Y., Henbest, H. & Trocha-Grimshaw, J. (1974). J. Chem. Soc. Perkin Trans. 1, pp. 592.
  26. Hübschle, C. B., Sheldrick, G. M. & Dittrich, B. (2011). J. Appl. Cryst. 44, 1281–1284. [DOI] [PMC free article] [PubMed]
  27. Ismail, S. Z., Anderton, C. L., Copley, R., Price, L. S. & Price, L. S. (2013). Cryst. Growth Des. 13, 2396–2406.
  28. James, B. R., Morris, R. H., Einsten, F. W. B. & Willis, A. (1980). J. Chem. Soc. Chem. Commun. pp. 31.
  29. Kitaigorodsky, A. (2012). Moecular Crystals and Molecules. New York: Academic Press.
  30. Köpf-Maier, P. (1994). Eur. J. Clin. Pharmacol. 47, 1–16. [DOI] [PubMed]
  31. Kukushkin, Y. N. (1971). Chem. Pap. 25, 380–384.
  32. Leung, C. H., Zhong, H. J., Chan, D. S. & Ma, D. L. (2013). Coord. Chem. Rev. 257, 1764–1776.
  33. Messori, L., Marcon, G., Orioli, P., Fontani, M., Zanello, P., Bergamo, A., Sava, G. & Mura, P. (2003). J. Inorg. Biochem. 95, 37–46. [DOI] [PubMed]
  34. Mezey, P. G. & Kapur, A. (1980). Can. J. Chem. 58, 559–566.
  35. Nagy, Z. K. & Braatz, R. D. (2012). Annu. Rev. Chem. Biomol. Eng. 3, 55–75. [DOI] [PubMed]
  36. Neese, F. (2012). Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2, 73–78.
  37. Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307–326. [DOI] [PubMed]
  38. Rademaker-Lakhai, J. M., van den Bongard, D., Pluim, D., Beijnen, J. H. & Schellens, J. H. (2004). Clin. Cancer Res. 10, 3717–3727. [DOI] [PubMed]
  39. Rossum, J. van (1963). J. Pharm. Pharmacol. 15, 285–316. [DOI] [PubMed]
  40. Sava, G., Bergamo, A. & Dyson, P. J. (2011). Dalton Trans. 40, 9069–9075. [DOI] [PubMed]
  41. Sava, G., Pacor, S., Mestroni, G. & Alessio, E. (1992). Clin. Exp. Metastasis, 10, 273–280. [DOI] [PubMed]
  42. Schäfer, A., Horn, H. & Ahlrichs, R. (1992). J. Chem. Phys. 97, 2571–2577.
  43. Schäfer, A., Huber, C. & Ahlrichs, R. (1994). J. Chem. Phys. 100, 5829–5835.
  44. Selbin, J. W., Bull, W. E. & Holmes, L. H. Jr (1961). J. Inorg. Nucl. Chem. 16, 219–224.
  45. Shan, N. & Zaworotko, M. J. (2008). Drug Discov. Today, 13, 440–446. [DOI] [PubMed]
  46. Sheldrick, G. M. (2008). Acta Cryst. A64, 112–122. [DOI] [PubMed]
  47. Sheldrick, G. M. (2015). Acta Cryst. C71, 3–8.
  48. Sluis, P. van der & Spek, A. L. (1990). Acta Cryst. A46, 194–201.
  49. Smith, G. D. & Jaffe, R. L. (1996). J. Phys. Chem. 100, 18718–18724.
  50. Steiner, T. (2002). Angew. Chem. Int. Ed. 41, 48–76.
  51. Stromgaard, K., Krogsgaard-Larsen, P. & Madsen, U. (2009). Editors. Textbook of Drug Design and Discovery. p. 49, Boca Raton: CRC Press.
  52. Thomas, R., Shoemaker, C. B. & Eriks, K. (1966). Acta Cryst. 21, 12–20.
  53. Wayland, B. B. & Schramm, R. F. (1969). Inorg. Chem. 8, 971–976.
  54. Weigend, F. & Ahlrichs, R. (2005). Phys. Chem. Chem. Phys. 7, 3297–3305. [DOI] [PubMed]
  55. Yirong, M. (2011). Comput. Mol. Sci. 1, 164–171.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Crystal structure: contains datablock(s) 1a_alpha, 1a_beta, 1b, 2. DOI: 10.1107/S2052520617011490/lo5014sup1.cif

b-73-01032-sup1.cif (2.7MB, cif)

Structure factors: contains datablock(s) 1a_alpha. DOI: 10.1107/S2052520617011490/lo50141a_alphasup2.hkl

b-73-01032-1a_alphasup2.hkl (209.3KB, hkl)

Structure factors: contains datablock(s) 1a_beta. DOI: 10.1107/S2052520617011490/lo50141a_betasup3.hkl

b-73-01032-1a_betasup3.hkl (199.5KB, hkl)

Structure factors: contains datablock(s) 1b. DOI: 10.1107/S2052520617011490/lo50141bsup4.hkl

b-73-01032-1bsup4.hkl (273.7KB, hkl)

Structure factors: contains datablock(s) 2. DOI: 10.1107/S2052520617011490/lo50142sup5.hkl

b-73-01032-2sup5.hkl (206.7KB, hkl)

Supplementary figures and schemes. DOI: 10.1107/S2052520617011490/lo5014sup6.pdf

b-73-01032-sup6.pdf (1.7MB, pdf)

CCDC references: 1566729, 1566730, 1566731, 1566732

Articles from Acta Crystallographica Section B, Structural Science, Crystal Engineering and Materials are provided here courtesy of International Union of Crystallography

RESOURCES