CuSeO4 and Cu(SeO3OH)2·6H2O, two novel copper–selenium(VI) oxysalts

The two new copper(II) salts Cu(SeO4) and Cu(SeO3OH)2·6H2O were synthesized at low-temperature hydrothermal conditions (220 °C), and room temperature, respectively. Their atomic arrangements were studied based on single-crystal X-ray investigations [P21/n, a = 4.823(1), b = 8.957(2), c = 6.953(1) Å, β = 94.82(1)°, Z = 4; P1¯\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$P{\overline 1}$$\end{document}, a = 6.133(1), b = 6.303(1), c = 8.648(2) Å, α = 70.45(1), β = 84.60(1), γ = 73.44(1)°, Z = 1]. Cu(SeO4) adopts the MnAsO4 structure type. It exhibits structural as well as topological relations with two formerly known isochemical compounds. They crystallize in the structure type ZnSO4 (mineral name zincosite, Pnma) respectively NiSO4 (Cmcm). The two minerals dravertite, CuMg(SO4)2, and hermannjahnite, CuZn(SO4)2, are isotypic with CuSeO4-P21/n; interestingly, also α-NaCu(PO4) belongs to this structure type: some rotation of the XO4 group allows a supplementary position for the Na atom. — Cu(SeO3OH)2·6H2O represents a new structure type. The protonated selenate group shows an extended Se—Oh bond distance (1.695 Å) as compared to the other Se—O bonds (1.614 to 1.626 Å). One OH dipole of the three independent H2O molecules represents a rather free hydrogen bond. For the other H atoms, the O—H···O lengths vary from 2.585 to 2.799 Å. Interestingly, the distance Ow7···Ow7 of only 2.791 Å does not represent an edge in a coordination polyhedron and it is not preliminary involved in the hydrogen bond scheme. All Cu2+ ions in the two title compounds are in a pronounced [4 + 2] coordination. The Cu2+[4+2] atoms in Cu(SeO4) are linked to chains along [100]; in Cu(SeO3OH)2·6H2O they are not connected among each other.


Introduction
This paper is part of an ongoing series of investigations of zirconium sulfates(VI) and selenates(IV,VI), in part with additional cations Wildner 2018, 2023;Wildner and Giester 2019;Giester et al. 2019;Wildner et al. 2022). As by-products two hitherto unknown copper(II) salts were obtained, that is a new modification of Cu(SeO 4 ) and Cu(SeO 3 OH) 2 ·6H 2 O crystallizing in a new structure type. Already in 1964, Snyman and Pistorius described two modifications of CuSeO 4 : Samples prepared at ambient conditions were obtained by dehydration of the hydrates. They crystallise in the structure type ZnSO 4 (mineral name zincosite, space group Pnma, a = 9.390(6), b = 7.005(5), c = 5.099(4) Å). However, Snyman and Pistorius (1964) described the phase in the non-standard space-group setting Pbnm; the standard setting (Pnma) is obtained by the transformation (010 / 001 / 100). The high-pressure form of CuSeO 4 was prepared in a HP-HT apparatus as described by Griggs and Kennedy (1956). After an exposure to 40 kbar and 400 °C for 30 min the reaction product was quenched to ambient conditions, and finally analysed by powder X-ray diffraction (PXRD). It crystallises in the NiSO 4 -structure type (space group Cmcm, a = 5.42(1), b = 8.55(1), c = 6.61(1) Å). For both the Pnma and Cmcm modifications, Snyman and Pistorius (1964) recorded the intensities of the Bragg reflections but not the full powder patterns as Rietveld refinements were not applicable at that time. These authors calculated the cell metrics and reported the structure types but did not refine any atomic coordinates. However, atomic coordinates can be found in the Inorganic Crystal Structure Database (ICSD) provided by the FIZ Karlsruhe -Leibniz-Institut für Informationsinfrastruktur (collection codes 109069 109073; see e.g., Belsky et al. 2002), but the lack of the full PXRD diagram results in very inaccurate interatomic distances, hampering the discussion of any crystalchemical details.
The present new P2 1 /n modification is a monoclinic distorted form of the Pnma compound (Snyman and Pistorius 1964). For a detailed comparison of the Pnma type and the P2 1 /n type structure of MnAsO 4 see Aranda et al. (1993); for crystal-structure investigations at low temperatures (1.6 to 15 K) and a study of magnetic properties cf. Aranda et al. (1998). For the differentiation, the space-group symmetry is added throughout the paper: Cu(SeO 4 )-Pnma, Cu(SeO 4 )-Cmcm, and Cu(SeO 4 )-P2 1 /n, respectively. Cu(SeO 3 OH) 2 ·6H 2 O represents a new structure type.

Syntheses
The investigated Cu selenates were obtained as side-products in the course of hydrothermal syntheses experiments, aiming at the preparation of ZrCu(SeO 4 ) 3 (Giester and Wildner 2023

Single-crystal X-ray diffraction (SCXRD) investigations
The quality of the obtained crystals was checked by SCXRD experiments. As the cell metrics proved new phases, tiny chips were selected for collection of the Bragg-peak intensities with the following equipment: Bruker APEXII diffractometer (crystaldetector distance 35 mm), CCD (charge-coupled device) area detector, Incoatec Microfocus Source IµS (30 W, multilayer mirror, Mo-Kα), and an Oxford Cryosystems (Cryostream 800 Plus LT) device. Data were collected at 200 K. Several sets of phiand omega-scans with 2° scan-width were combined to achieve respective full-sphere data up to 80° 2θ. For data handling including integration as well as Lorentz, polarization, and absorption correction (multi-scan method), the Bruker Apex3 suite was used (Bruker 2000). Unit-cell parameters were obtained by least-squares refinements of the 2θ values measured for Bragg reflections. Structure solution and refinements were performed by full-matrix least-squares techniques on F 2 considering secondary extinction (SHELX programme suite, Sheldrick 2008Sheldrick , 2015. Complex scattering functions for neutral atoms (Wilson 1992) were applied. For CuSeO 4 -P2 1 /n, the cell metric was found to be close to that of MnAsO 4 (Aranda et al. 1993;1998); therefore, a tentative isotypy was expected. A trial to start the structure refinement based on the atomic coordinates given by Aranda et al. (1993) was successful; their atomic coordinates were transformed according to (001 / 0 1 0 / 100) to obtain the standard setting with a < c. After a few cycles of full-matrix least-squares refinements, the calculations converged satisfactorily. Finally, anisotropic displacement parameters for all atoms were considered.
For the crystal structure of Cu(SeO 3 OH) 2 ·6H 2 O centrosymmetry was suggested by an analysis of the intensity statistics. Assuming space group P1 , the atomic coordinates of the Se and Cu atoms were found by direct methods. In successive difference Fourier summations the positions of the O atoms and later on those of the H atoms were located. During the final refinement cycles, anisotropic displacement parameters of the Se, Cu, and O atoms as well as isotropic ones for the H atoms were refined. Only for the atom O w 7 the displacement parameters of the two H ligands were constrained to equal values during the structure refinement. All other parameters were refined without any constrains. The final difference Fourier summations of both compounds exhibit electron densities of roughly 0.5 eÅ −3 between the Se respective Cu atoms and their oxygen atom ligands indicating the pronounced covalent bonding.
Crystal parameters as well as a summary on the data collections and structure refinements are given in Table 1, final structure parameters and relevant interatomic bond distances and bond angles in Tables 2 and 3. Further details of the crystal structure investigations may be obtained from the joint CCDC/FIZ Karlsruhe online deposition service: https:// www. ccdc. cam. ac. uk/ struc tures/? by quoting the deposition numbers CSD-2231881 (Cu[SeO 3 (OH)] 2 ·6H 2 O) and CSD-2231888 (Cu(SeO 4 )-P2 1 /n).

Discussion
The crystal structure of CuSeO 4 -P2 1 /n A large number of structure types with the general formula M [6] [X [4] O 4 ] is known; often there are many isotypic representatives with distinct M and X atoms. However, Cu 2+ ions have an electron configuration d 9 and, therefore, the octahedral surrounding is tetragonally distorted. Predominantly an elongated tetragonal dipyramid exhibiting a pronounced [4 + 2] instead of a [6] coordination is observed (Jahn-Teller effect; Jahn and Teller 1937). It is to be mentioned that not all ions with a theoretically expected Jahn-Teller effect show such a large distortion of the octahedral environment as Cu 2+ ions. An example is the Co 2+ ion with an uneven electron distribution in the t 2g group of the d-orbitals, where Wildner (1992) could not proof a significant deviation from an octahedral arrangement of the ligands. On the other hand, Zn 2+ atoms may adopt strongly distorted Zn [6] O 6 coordination polyhedra even though its electron configuration d 10 predicts an absence of any distortion caused by the Jahn-Teller effect. Zn 2+ ions may substitute for Cu 2+ atoms if a moderate adoption of the coordination polyhedra are possible in the respective structure type. Consequently, only a limited number of M [6] cations may substitute for the Cu 2+ ions and vice versa. Exceptions are loose and flexible atomic arrangements that can adopt cations with these different environments.
Mn 3+ ions (d 4 configuration) also exhibit a pronounced [4 + 2] coordination caused by the Jahn-Teller effect. Well known is the tetragonal I4 1 /amd symmetry of hausmannite, Mn 3+ 2 Mn 2+ O 4 , instead of the Fd3m symmetry of most other spinels. Therefore, Cu 2+ ions are predestined for a substitution for Mn 3+ ions despite of the different valence.

Crystal chemistry of the MnAsO 4 structure type
Caused by the distinct distortion of the Mn 3+ O 6 polyhedron due to the Jahn-Teller effect, the Mn 3+ As 5+ O 4 -structure type (Aranda et al. 1993(Aranda et al. , 1998) has a limited number of representatives only. As mentioned above, the structure is characterized by two crystallographically distinct Mn sites with a differently strong distortion of the coordination polyhedra; it is evidenced also in CuSeO 4 -P2 1 /n. The adaptability of the structure type enables an occupation of the two sites with distinct kinds of atoms. Due to the pronounced distortion of the Mn1O 6 polyhedron, this M1 site is predestined for an occupation with an ion exhibiting a strong Fig. 1 The crystal structures of (a) Cu(SeO 4 )-P2 1 /n (structure type MnAsO 4 ), (b) Cu(SeO 4 )-Pnma (structure type ZnSO 4 ; Snyman and Pistorius 1964), (c) Cu(SeO 4 )-Cmcm (structure type NiSO 4 ; Snyman and Pistorius 1964), and (d) NaCuPO 4 -P2 1 /n (Kawahara et al. 1993; a transformation to the standard setting according to 001 / 0 1 0 / 001 was applied). In (a), (b), and (d) the "short" and "long" Cu-O bond lengths are indicated in dark blue respectively turquoise colour; in (c) the accuracy of structure refinements does not allow an allocation about distinct Cu-O bond lengths. Drawing with programme ATOMS (Dowty 2013) distortion in the octahedral field. The more regular Mn2O 6 polyhedron and the flexibility of the crystal structure enable the substitution at the M2 site by a variety of atoms. It is verified by the mineral dravertite, CuMg(SO 4 ) 2 (Pekov et al. 2017) and by the recently described mineral hermannjahnite, CuZn(SO 4 ) 2 (Siidra et al. 2018 The flexibility of the structure type is verified by the synthetic compound β-NaCu(PO 4 ) (Kawahara et al. 1993;Fig. 1d). A slight rotation of the XO 4 tetrahedron and of the Cu [4+2] O 6 polyhedra give space for the incorporation of the Na atom at an additional site. α-NaCu(PO 4 ) (Quarton and Kolsi 1983) crystallizes in space group P2 1 2 1 2 1 ; it differently forms zig-zag chains of corner-connected Cu [4+1] O 5 polyhedra in a rugged layer-like arrangement with the Na atoms and phosphate tetrahedra in between.
The title compound and isopointal analogues as well as β-NaCu(PO 4 ) (Kawahara et al. 1993) were geometrically analysed (programme COMPSTRU, de la Flor et al. 2016;Bergerhoff et al. 1999). It allows a comparison between the configurations of crystal-structure pairs. It is outlined by the degree of lattice distortion (S), the maximal displacement between the atomic positions of the paired atoms (d max ), the arithmetic mean of the displacement between the atomic positions (d av ), and the similarity as a function of the differences in atomic positions and the ratios of the corresponding lattice parameters (Δ). The results are compiled in Table 4. In addition, the displacements d X of X atom pairs (X = As, P, S, Se) as well as the arithmetic mean of the displacement between the O atom pairs (d O ) are compiled. Despite the various compositions, S varies by less than 3%; the largest values for Δ were observed for compound pairs where β-NaCu(PO 4 ) is involved. As the M atoms (M = Mg, Mn, Cu, Zn) occupy special positions, the displacement of atom pairs concerns the XO 4 tetrahedron. The largest displacements are observed for β-NaCu(PO 4 ) caused by the space requirement of the additional position for the Na atom. It is supported by the smaller values of d X as compared to d O indicating a rotation of Table 4 Crystal chemical comparison of the compounds crystallizing in the MnAsO 4 structure type (Aranda et al. 1993(Aranda et al. , 1998 Ferraris and Ivaldi (1984).

Hydrogen-bond scheme in Cu(SeO 3 OH) 2 ·6H 2 O
Bond-valence calculations (Tables 2, 3) clearly indicate that the atoms O1, O2, and O3 are acceptor atoms of the hydrogen bonds; O h 4 is the donor atom of a hydroxyl group, whereas the atoms O w 5, O w 6, and O w 7 each belong to an H 2 O molecule. In accordance with the experimentally verified H atom positions, the atoms O1 and O3 are acceptors of two whereas the atom O2 is an acceptor of one H bond (the bond valence sum neglecting the contribution of the H atoms amount: 1.56 and 1.55 respectively 1.74 v.u.). Also the valence sum of the O h 4 atom (1.29 v.u.), those of the atoms O w 5 and O w 6 (0.50 and 0.49 v.u.) agree very well with the assumption of a hydroxyl group respectively H 2 O molecules. The atom O w 7 is neither connected to the Cu nor to the Se atom; H 2 O w 7 represents a rather free H 2 O molecule. The O h -H and all the O w -H bond lengths compare quite well with the expectations for a structure refinement based on X-radiation.
The O h 4-H···O w 7 length is relatively short (2.585 Å), those of O w 3-H···O, O w 4-H···O, and O w 5-H···O are rather uniformly 2.683 to 2.716 Å. All these hydrogen bridges are close to linear. Some problems arise for the O w 7 atom (Fig. 3). Despite the two H atoms H7a and H7b could be located experimentally showing the usual geometry for an H 2 O molecule, only the O w 7-H7a···O2 (angle 175°) bridge is clear-cut. At a first glance, O w 7-H7b···O h 4 is a reasonable assumption for the second H atom. However, O h 4-H4···O w 7 (angle 171°) obviously is a hydrogen bridge from the hydroxyl group. The distance H7b···H4 is already short (1.63 Å) even though the X-ray refinement mirrors the average positions of the electrons but not of the proton of the H atoms. Moreover, the angle O w 7-H7b···O h 4 = 106° is small indicating an at least bi-or multi-furcated hydrogen bond. In this context it should be mentioned that the O w 7···O w 7 distance of 2.793(3) Å is unusually short as it does not represent an edge in Fig. 2 The crystal structure of Cu[SeO 3 (OH)] 2 ·6H 2 O in a projection parallel to [001] on (110). "Short" and "long" Cu-O bond lengths are indicated in dark blue respectively turquoise colour. Drawing with programme ATOMS (Dowty 2013) Fig. 3 The surrounding of the H 2 O w 7 molecule in Cu[SeO 3 (OH)] 2 ·6H 2 O. "Short" and "long" Cu-O bond lengths are indicated in dark blue respectively turquoise colour. Drawing with programme ATOMS (Dowty 2013) any coordination polyhedron. Such short interpolyhedral contacts are rare (cf . Wildner 1990;Zemann 1986). These two O w 7 atoms are symmetry related by an inversion centre; therefore, they cannot represent the acceptor for the hydrogen bond from the H7b atom but it may be involved into a bi-or multi-furcated bond. The atoms O1, O2, O3, O w 5, and O w 6 are in an approximate planar [3] coordination. The O4 atom is bent [2] coordinated.