Spontaneous enanti-omorphism in poly-phased alkaline salts of tris-(oxalato)ferrate(III): crystal structure of cubic NaRb5[Fe(C2O4)3]2.

We show here that the phenomenon of spontaneous resolution of enanti-omers occurs during the crystallization of the sodium and rubidium double salts of the transition metal complex tris-(oxalato)ferrate(III), namely sodium penta-rubidium bis-[tris-(oxalato)ferrate(III)], NaRb5[Fe(C2O4)3]2. One enanti-omer of the salt crystallizes in the cubic space group P4332 with Z = 4 and a Flack absolute structure parameter x = -0.01 (1) and its chiral counterpart in the space group P4132 with x = -0.00 (1). All metal ions are at crystallographic special positions: the iron(III) ion is on a threefold axis, coordinated by three oxalate dianions in a propeller-like conformation. One of the two independent rubidium ions is on a twofold axis in an eightfold coordination with neighbouring oxalate oxygen atoms, and the other one on a threefold axis in a sixfold RbO6 coordination. The sodium ion is at a site of D3 point group symmetry in a trigonal-anti-prismatic NaO6 coordination.

Chemical context

Chirality is the structural property by which a mol­ecule or ion cannot be superposed upon its mirror image through translation and proper rotation operations. This concept along with the related ones of chiral crystal structures and space groups is discussed by Flack (2003 ▸). Chirality is at the core (among other research areas) of the not yet understood origin of the biomolecular asymmetry of life (Meierhenrich, 2008 ▸), enanti­oselective chemical reactions (Knowles, 2001 ▸; Noyori, 2001 ▸; Sharpless, 2001 ▸), biological activity of pharmaceuticals (Nguyen et al., 2006 ▸) and in the design of multifunctional solid-state materials endowed with optical activity and long-range magnetic order (Coronado et al., 2003 ▸) and also in the understanding of the physical properties of chiral liquid crystals and their tailoring for applications in opto-electronic devices (Goodby, 1998 ▸; Coles, 1998 ▸).

While attempting to crystallize the rubidium salt of the tris­(oxalato)ferrate(III) transition metal complex, one of the preparations segregated into a poly-phased crystal system. It contained the intended Rb3[Fe(C2O4)3]·3H2O compound (monoclinic P21/c), which turned out to be isotypic to the reported potassium salt (Junk, 2005 ▸; Piro et al., 2016 ▸), and the triclinic (P ) Rb(C2O4H)(C2O4H2)·2H2O salt (Kherfi et al., 2010 ▸), which is isotypic to the ammonium analogue (Jarzembska et al., 2014 ▸). A third phase consisted of large green crystals of a new cubic (P4332) NaRb5[Fe(C2O4)3]2 salt. Inter­estingly, the isotypic counterpart of this salt where rubidium is replaced by potassium has been reported by Wartchow (1997 ▸) to appear in a mixture with crystals of the monoclinic K3[Fe(C2O4)3]·3H2O salt, hence confirming the tendency of potassium and rubidium alkaline ions to form isotypic crystal analogues (Piro et al., 2016 ▸). Curiously, in a previous work, Henneicke & Wartchow (1997 ▸) reported the chiral counterpart of the cubic NaK5[Fe(C2O4)3]2 salt, which crystallizes in the space group P4132. This prompted us to search for the chiral rubidium analogue in the very same batch as the single-crystals that solved in the space group P4332 NaRb5[Fe(C2O4)3]2. By chance, we picked a single crystal and submitted it to X-ray diffraction scrutiny to find that it now belonged to the chiral space group P4132. This strongly suggests that the NaM 5[Fe(C2O4)3]2 (M = K, Rb) crystal samples could be racemic conglomerates generated by spontaneous resolution, a rare event discovered by Louis Pasteur in 1848 (Pasteur, 1848a ▸,b ▸) in a famous experiment in which he hand-sorted the chirally resolved crystals of sodium ammonium tartrate tetra­hydrate on the basis of their observed morphology and then examined their respective solutions with a polarimeter to find opposite rotations of the plane of light polarization (Flack, 2009 ▸). Recently, we found that the phenomenon could also have occurred in isotypic [M(Lap)2] (M = Cd, Mn; HLap = 2-hy­droxy-3-(3-methyl-2-buten­yl)-1,4-naphto­quinone, C15H14O3) complexes whose enanti­omers crystallize in the tetra­gonal and enanti­omorphic space groups P43212 and P41212 (Farfán et al., 2015 ▸).

Structural commentary

Fig. 1 ▸ shows an ORTEP (Farrugia, 2012 ▸) drawing of the P4332 enanti­omer of the title compound. Bond lengths and angles around iron(III) and within the oxalate dianion are listed in Table 1 ▸ and contact distances around the alkali ions are shown in Table 2 ▸. All metal ions are at crystallographic special positions while the oxalate anion is on a general position. The iron(III) ion is on a threefold axis, C 3 point group symmetry (Wyckoft c site), in an octa­hedral environment (FeO6 core). It is coordinated to three, symmetry-related, oxalate anions acting as bidentate ligands through the oxygen atoms of their opposite carb­oxy­lic groups in a propeller-like conformation and along one electron pair lobe on each oxygen ligand. The FeO6 bond geometry and metrics are consistent with the oxalate being a weak-field ligand that gives rise to the high-spin (S = 5/2) electronic ground state exhibited by the complex, as probed by magnetic susceptibility (Delgado et al., 2002 ▸) and ESR spectroscopy (Collison & Powell, 1990 ▸) in synthetic minguzzite, K3[Fe(C2O4)3]·3H2O, by polarized electronic absorption spectroscopy in single crystal NaMg[(Fe, Al)(C2O4)3]·9H2O mixtures (Piper & Carlin, 1961 ▸) and also by Mössbauer spectroscopy in K3[Fe(C2O4)3]·3H2O (Bancroft et al., 1970 ▸; Sato & Tominaga, 1979 ▸; Ladriere, 1992 ▸) and in the alkali (Na, Rb, Cs) family of tris­(oxalato)ferrate(III) salts (Piro et al., 2016 ▸).

Figure 1

View of NaRb5[Fe(C2O4)3]2 showing the atom labels and displacement ellipsoids at the 50% probability level. For clarity, only the minimum number of oxygen ligands around each metal ion has been labelled. The rest of the environmental oxygen atoms are generated through the symmetry operations of the corresponding point groups: C 3 (Fe), C 2 (Rb1), C 3 (Rb2) and D 3 (Na). Iron–oxalate bonds are indicated by double lines and alkali metal–oxygen contacts by single lines. Symmetry codes: (i) −y + 1, z + , −x + ; (ii) y − , −z + , −x + 1; (iii) −y + , −x + , −z + .

Table 1

Bond lengths and angles (Å, °) around iron(III) and within the oxalate dianion in NaRb5[Fe(C2O4)3]2 P4332 enanti­omer

(a) At a crystal site of C 3 point group symmetry.

Iron(III)a		(C2O4)2−
Fe—O11	2.021 (4)	C1–O12	1.211 (7)	O12—C—O11	125.2 (6)
Fe—O21	1.989 (4)	C1–O11	1.286 (7)	O12—C1—C2	121.2 (6)
		C1—C2	1.540 (9)	O11—C1—C2	113.5 (5)
O21—Fe—O11	80.0 (2)	C2—O22	1.211 (7)	O22—C2—O21	125.3 (6)
O21—Fe—O21i	88.4 (2)	C2—O21	1.283 (7)	O22—C2—C1	121.1 (6)
O11—Fe—O11i	88.7 (2)			O21—C2—C1	113.6 (5)
O11—Fe—O21i	106.2 (2)
O11—Fe—O21ii	160.9 (2)

Symmetry codes: (i) −z + , −x + 1, y − ; (ii) −y + 1, z + , −x + .

Table 2

Bond lengths (Å) around the alkali metal ions in NaRb5[Fe(C2O4)3]2 P4332 enanti­omer.

(a) At a site of C 2 point group symmetry; (b) at a C 3 site; (c) at a D 3 site.

Rb1a		Rb2b		Nac
Rb1—O11	3.009 (4)	Rb2—O22	2.808 (4)	Na—O12	2.439 (4)
Rb1—O11i	3.067 (4)	Rb2—O21iii	3.114 (4)
Rb1—O22ii	2.788 (5)
Rb1—O12ii	3.133 (5)

Symmetry codes: (i) −y + 1, z + , −x + ; (ii) y − , −z + , −x + 1; (iii) −y + , −x + , −z + .

The planes of the carb­oxy­lic –COO− groups of the oxalate ligand are slightly tilted from each other, by 12 (1)° around the C—C σ-bond. As expected, the C—O bond lengths involving the coordinated-to-metal oxygen atoms are significantly longer [1.286 (7) and 1.283 (7) Å] than the ones corresponding to the uncoordinated oxalate oxygen atoms [both equal to 1.211 (7) Å].

There are two independent rubidium ions, one (Rb1) lying on a twofold axis, C 2 point group symmetry (d site) in an eightfold coordination with neighbouring oxalate oxygen atoms, the other one (Rb2) on a threefold axis, C 3 point group (c site) in a sixfold coordination. The sodium ion is at a site of D 3 point group symmetry (a site) in a trigonal–anti­prismatic NaO6 coordination with one oxygen atom of six neighbouring, symmetry-related, oxalate ions.

When dealing with octa­hedral Fe(C2O4)3 tris-chelated metal complexes, it is customary to describe its chirality employing Λ- and Δ-descriptors (Meierhenrich, 2008 ▸). It turns out that the enanti­omeric complexes correlate with the corresponding chiral space groups, as indicated in Fig. 2 ▸.

Figure 2

Views of the Λ and Δ enanti­omers of [Fe(C2O4)3]3−.

The possibility of controlling the crystal chirality and therefore obtaining enhanced optical activity of functional materials has been discussed (Gruselle et al., 2006 ▸). To this purpose, two general synthetic routes have been developed to reach optically active coordination compounds, namely either by enanti­oselective synthesis using enanti­opure chiral species, which yields enanti­opure samples (Knof & von Zelewsky, 1999 ▸) or by spontaneous resolution upon crystallization from a racemate, which yields a conglomerate (Pérez-García & Amabilino, 2002 ▸). As explained above, the chiral NaRb5[Fe(C2O4)3]2 crystals were obtained through the phenomena of spontaneous resolution from a racemic solution of [Fe(C2O4)3]3− complex ions into a racemic conglomerate. This is presumably followed by a structural inductive effect by these chiral mol­ecular ions on the alkali metal ions through shared oxalate ligands. In fact, not only is the FeIII ion a ‘stereogenic centre’ in the Fe(C2O4)3 tris-chelated metal complex, but so also are the sodium and one (Rb2) of the rubidium ions. These metal ions are in a distorted octa­hedral environment coordinated to six oxalate anions, acting as monodentate ligand through one of their oxygen atoms and resembling a six-bladed propeller-like conformation. From the structural data, it turns out that the chirality of this local arrangement around the alkaline ions is coincident with the one of the [Fe(C2O4)3]3− inductor and therefore the chiral crystals reported here can be more conveniently described as Λ-NaΛ-Rb2Rb3[Λ-Fe(C2O4)3]2 (P4332) and Δ-NaΔ-Rb2Rb3[Δ-Fe(C2O4)3]2 (P4132). However, no definitive chirality can be unambiguously assigned to the other independent rubidium (Rb1) ion which is in an eightfold polyhedral coordination.

Database survey

The formation of racemic conglomerates of single crystals, adequate for structural X-ray diffraction, generated by spontaneous resolution is an infrequent phenomenon. In fact, a search of the Cambridge Structural Database (Groom et al., 2016 ▸) invoking the term ‘spontaneous resolution’ showed seventeen entries, and another one using as a target ‘chiral crystals’ produced a further four hits. Among them there were reported the chiral to each other (M)- and (P)-catena-{[μ2-2-(imidazo[4,5-f](1,10)phenanthrolin-2-yl)benzoato-N,N′,O]aqua­chloro­zinc(II)} (CSD refcodes EJINOB and EJINUH; Wei et al., 2011 ▸) and catena-[(μ8-benzene-1,3,5-tri­carboxyl­ato)lithiumzinc] (CSD refcodes WAJHUM and WAJJAU; Xie et al., 2010 ▸).

Synthesis and crystallization

As stated in the Chemical context, in one of the preparations generated during the synthesis of the rubidium salt of [Fe(C2O4)3]3−, by reaction of freshly precipitated Fe(OH)3 (obtained by dropwise addition of a small excess of 20% NaOH to an FeIII solution) with rubidium oxalate: Fe(OH)3 + 3Rb(HC2O4) + 3H2O → Rb3[Fe(C2O4)3]·3H2O + 3H2O) (Piro et al., 2016 ▸), we found a relatively complex reaction giving rise to a poly-phased crystal mixture, from which the NaRb5[Fe(C2O4)3]2 chiral pair could be isolated.

Refinement details

Crystal data, data collection procedure and structure refinement results are summarized in Table 3 ▸. The structure was solved by intrinsic phasing with SHELXT (Sheldrick, 2015a ▸). The stereoisomers were determined through refinement of the Flack absolute structure parameter. This is the fractional contribution to the diffraction pattern due to the mol­ecule racemic twin and for the correct enanti­omeric crystal it should be zero to within experimental error.

Table 3

Experimental details

	Cubic, P4332	Cubic, P4132
Crystal data
Chemical formula	NaRb5[Fe(C2O4)3]2	NaRb5[Fe(C2O4)3]2
M r	1090.16	1090.16
Temperature (K)	297	293
a (Å)	13.8058 (4)	13.7995 (3)
V (Å3)	2631.4 (2)	2627.79 (17)
Z	4	4
Radiation type	Mo Kα	Mo Kα
μ (mm−1)	10.42	10.43
Crystal size (mm)	0.48 × 0.42 × 0.38	0.48 × 0.35 × 0.25
Data collection
Diffractometer	Agilent Xcalibur Eos Gemini	Rigaku Oxford Diffraction Xcalibur, Eos, Gemini
Absorption correction	Multi-scan (CrysAlis PRO; Agilent, 2014 ▸)	Multi-scan (CrysAlis PRO; Rigaku OD, 2015 ▸)
T min, T max	0.690, 1.000	0.786, 1.000
No. of measured, independent and observed [I > 2σ(I)] reflections	2960, 959, 767	4284, 961, 814
R int	0.043	0.038
(sin θ/λ)max (Å−1)	0.638	0.638
Refinement
R[F 2 > 2σ(F 2)], wR(F 2), S	0.035, 0.064, 1.00	0.032, 0.068, 1.02
No. of reflections	959	961
No. of parameters	68	68
Δρmax, Δρmin (e Å−3)	0.84, −0.85	1.02, −0.95
Absolute structure	Flack x determined using 225 quotients [(I +)−(I −)]/[(I +)+(I −)] (Parsons et al., 2013 ▸)	Flack x determined using 251 quotients [(I +)−(I −)]/[(I +)+(I −)] (Parsons et al., 2013 ▸).
Absolute structure parameter	−0.013 (12)	−0.003 (10)

Computer programs: CrysAlis PRO (Agilent, 2014 ▸; Rigaku OD, 2015 ▸), SHELXT (Sheldrick, 2015a ▸), SHELXL2014 (Sheldrick, 2015b ▸) and ORTEP-3 for Windows (Farrugia, 2012 ▸).

Crystal structure: contains datablock(s) P4332, P4132, global. DOI: 10.1107/S2056989018008022/hb7735sup1.cif

Structure factors: contains datablock(s) P4332. DOI: 10.1107/S2056989018008022/hb7735P4332sup2.hkl

Structure factors: contains datablock(s) P4132. DOI: 10.1107/S2056989018008022/hb7735P4132sup3.hkl

CCDC references: 1563459, 1563458

Additional supporting information: crystallographic information; 3D view; checkCIF report