Stacking faults in a layered cobalt tellurium phosphate oxochloride

Solid State Sciences 40 (2015) 67e70

Contents lists available at ScienceDirect

Solid State Sciences journal homepage: www.elsevier.com/locate/ssscie

Stacking faults in a layered cobalt tellurium phosphate oxochloride Iwan Zimmermann, Mats Johnsson* Department of Materials and Environmental Chemistry, Arrhenius Laboratory, Stockholm University, SE-106 91 Stockholm, Sweden

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 November 2014 Received in revised form 29 December 2014 Accepted 2 January 2015 Available online 3 January 2015

The new compound Co2Te3(PO4)O6Cl was synthesized by chemical reactions in a sealed and evacuated silica tube. The crystal structure was solved from single crystal diffraction data and is made up by charge neutral layers. Within the layers two types of chains are made up by edge sharing [CoO6] and [CoO5Cl] polyhedra respectively. The chains are separated by tellurium oxide and phosphate building blocks. There are only weak Van der Waals interactions in between the layers and severe diffuse scattering is observed due to faulted stacking of the layers. Structure solutions in a P-1 triclinic cell and a larger monoclinic cell in P21/c are discussed and compared to a computer generated model. The reasons for the stacking faults may be due to that there are two positions available for each layer that results in similar connectivity to the next layer in addition to the relatively wide channels in between the layers that reduce the Van der Waals interactions in between them. © 2015 Elsevier Masson SAS. All rights reserved.

Keywords: Oxo-halide Lone-pair elements Diffuse scattering Layered crystal structure

1. Introduction Incorporating p-element cations having a stereochemically active lone pair such as e.g. As3þ, Se4þ, Sb3þ or Te4þ into transition metal oxohalides has proven to be an efficient synthetic approach for finding compounds having specific structural and physical properties. The presence of lone pairs and halide ions, that both act as terminal ligands, have shown to result in a high probability of forming low-dimensional arrangements of the structural building blocks. Most commonly layered crystal structures are found from which the halide ions and the lone pairs are protruding. Weak Van der Waals interactions are responsible for the coherence in between layers. Transition metal ions within the covalently bonded layers are therefore well separated along the stacking direction, but can interact within the layers to form topologies may can lead to e.g. frustrated magnetism as in Ni5(TeO3)4X2 and FeTe2O5X (X ¼ Cl, Br) [1,2]. To further separate structural building blocks and to extend the chemical system, the inclusion of tetrahedral building blocks is investigated. Phosphate groups are well defined tetrahedral building units, which are known for the formation of open framework structures due to their predominantly corner sharing coordination behavior. So far the recently synthesized compounds Fe7(PO4)3Sb3O6X3 (X ¼ Cl, Br) are two of few examples [3] of

* Corresponding author. E-mail addresses: [email protected] (I. Zimmermann), mats. [email protected] (M. Johnsson). http://dx.doi.org/10.1016/j.solidstatesciences.2015.01.002 1293-2558/© 2015 Elsevier Masson SAS. All rights reserved.

transition metal phosphate oxohalides having also a lone pair element present. Strong one dimensional diffuse scattering is often observed in faulted layered structures. Stacking faults are fairly common among metals or simple binary compounds which are built from close packing of hexagonal layers such as in e.g. Si [4], SiC [5] or GaN [6] where the energy needed for a layer dislocation is low. Different kinds of stacking disorder can occur e.g. in Gd5Si4-xBix [7] by changing the stoichiometry. Among oxide materials a faulted layer stacking can for example occur from intercalating species such as water molecules in MnH2P3O10$2H2O [8] or b-Ni(OH)2 [9]. Stacking faults are also present in many open framework structures or minerals such as zeolites where the faulted layers are covalently linked together [10] or in e.g. layered brownmillerites [11]. Faulted stacking of layers will lead to one-dimensional diffuse scattering affecting the diffraction lines along the stacking direction. Diffuse diffraction lines often make the structure solution and its refinement difficult. In this work we report the structure and synthesis of a new layered cobalt tellurium phosphate oxochloride having irregular stacking of layers.

2. Experimental section Purple plate like single crystals of Co2Te3(PO4)O6Cl were prepared by chemical reactions in evacuated and sealed silica tubes in a yield of 20%. For the synthesis a 1:1:1:1 ratio of starting materials CoCl2 (SigmaeAldrich), CoO (ABCR), TeO2 (SigmaeAldrich) and (NH4)H2PO4 (STREM chemicals) was used. The starting powder was

68

I. Zimmermann, M. Johnsson / Solid State Sciences 40 (2015) 67e70

ground in an agate mortar and heated to 250  C in air to decompose the ammonium salt before sealing in a silica tube under vacuum. The reaction took place at 550  C for ~70 h. Crystals of the title compound for single crystal diffraction experiments were isolated manually from blue powder and unreacted starting materials. EDS analysis of the blue powder revealed that it is an oxide phase and the percentage of the heavy elements were 29.6 at% Co, 46.0 at% Te and 24.5 at% P. Single crystal X-ray diffraction experiments on Co2Te3(PO4)O6Cl were carried out on an Oxford Diffraction Xcalibur3 diffractometer equipped with a graphite monochromator. Data collection was carried out at 293 K using MoKa radiation, l ¼ 0.71073 Å. Absorption correction and data reduction were made with the software CrysAlis RED that was also employed for the analytical absorption correction [12]. The structure solution was carried out with SHELXS-97 and the refinement with SHELXL-97 [13] in the WINGX environment [14]. All atomic positions were refined isotropically as the anisotropic refinement did not give positive temperature parameters for all atoms, which is mainly due to the fact that the intensities from the diffuse scattering streaks could not be integrated properly. The diffuse scattering originating from stacking faults was simulated using the program DIFFaX [15] and results

Table 1 Crystal data and structure refinement parameters for Co2Te3(PO4)O6Cl in two different unit cell settings. Unit cell setting

1

2

Empirical formula

Co2Te3(PO4)O6Cl

Formula weight (g/mol) Temperature (K) Wavelength (Å) Crystal system Space group a (Å) b (Å) c (Å) a ( ) b ( ) g ( ) Volume (Å3) Z Densitycalc. (g cm3) Absorption coefficient (mm1) F(000) Crystal color Crystal habit Crystal size (mm3) Theta range for data collection ( ) Index ranges

1454.16 293 0.71073 Triclinic P-1 5.1671(3) 11.0523(8) 19.1695(17) 98.337(7) 92.823(6) 90.029(5) 1081.82(14) 4 4.464 11.43

Co2Te3(PO4) O6Cl 1454.16 293 0.71073 Monoclinic P21/c 5.1783(3) 11.0826(7) 38.088(5) 90 94.875(8) 90 2177.9(3) 8 4.464 11.43

1288 Pink Plate 0.075  0.047  0.0150 4.14e32.25

7  h  7 16  k  15 28  l  18 Reflections collected eunique 10,292e6647 (2557e2452)a Data/restraints/parameters 6647/0/137 Internal R value Rint ¼ 0.099 Rint ¼ 0.054b 2 Goodness-of-fit on F 1.019/0.865a Final R indicesa [I > 2sigma(I)] R1 ¼ 0.1654 R1 ¼ 0.0759a wR2 ¼ 0.3611 wR2 ¼ 0.1456a R indices (all data) R1 ¼ 0.2468 R1 ¼ 0.1584a wR2 ¼ 0.3917 wR2 ¼ 0.1634a Largest diff. peak and hole 32.346 and 4.648 (2.351 (eÅ3) and 1.838)a a

were compared qualitatively. EDS measurements were carried out on a JEOL 7000F scanning electron microscope. Crystal data are reported in Table 1, atomic coordinates and isotropic temperature parameters for all atoms, tables with bond distances and elemental composition from EDS are given in the Supplementary material. The structural drawings are made with the program DIAMOND [16]. 3. Results and discussion 3.1. Modeling the diffuse scattering Inspecting the reconstructed diffraction pattern of the (0kl) layer one can observe that diffraction lines along l give sharp Bragg peaks for k ¼ 2n and diffuse streaks for k ¼ 2n þ 1, which indicates disorder in the crystal structure along [001], see Fig. 1. The crystal structure was initially solved assuming a triclinic unit cell in the space group P1. As expected the structural model fitted the experimental data poorly (R1 ¼ 16.5% for 2855 reflections), due to the diffuse streaks which cannot be properly integrated by the data reduction software. The refinement could be significantly improved by omitting the reflections originating from the diffuse lines to R1 ¼ 7.6% for 1170 reflections. The structure could also be solved in a monoclinic unit cell with the space group P21/c. This choice is consistent with the systematic absences found in the (h0l) layer. Absent lines with h ¼ 2n þ 1 confirm the c-glide. There are two layers in the monoclinic cell, which are related to each other by the c-glide operation, which leads to the conclusion that the diffuse scattering arises due to irregular stacking of the layers. There are two possibilities for the layers to stack: i) by a glide operation, where the next layer is a mirror image (mirror plane perpendicular to [010]) of the initial layer or ii) by a layer shift, where the next layer is the same as the initial layer shifted along [010]. The two stacking possibilities occur randomly in the structure with equal probabilities. To confirm this theory a computer simulation was set up using the program DIFFaX. To match the input criteria in DIFFaX the monoclinic unit cell was divided into half along [001], giving a

2576 pink plate 4.14e32.25 7  h  7 15  k  16 57  l  43 19,868e7095 7095/0/137 Rint ¼ 0.177 1.076 R1 ¼ 0.1599 wR2 ¼ 0.3448 R1 ¼ 0.2807 wR2 ¼ 0.3886 9.428 and 8.797

Refinement without reflections lying within the diffuse scattering streaks with k ¼ 2nþ1.

Fig. 1. Comparison of calculated and experimental diffraction intensities of the (0kl), (1kl) and (h1l) planes. Patterns with white background are calculated, the ones with grey background are experimental.

I. Zimmermann, M. Johnsson / Solid State Sciences 40 (2015) 67e70

new unit cell containing only one layer. This new cell was then further transformed into an orthogonal one (a ¼ 5.18 b ¼ 11.08 c ¼ 18.98 Å) to have the c-axis parallel to the stacking direction, see Fig. 2. This unit cell does not reproduce the 3D crystal structure but describes one single layer required for the input file. In DIFFaX this layer was defined as Layer1 and its mirrored version (mirror plane perpendicular to [010]) as Layer2. The two layers were then allowed to randomly stack with equal probabilities. Layer 1-2 and 2-1 transitions result in stacking by the glide operation and layer 1-1 and 2-2 transitions describe stacking by a layer shift and a value of 0.25 b was used in all the calculations. To qualitatively compare the outcome of the simulation the calculated diffraction intensities were displayed as spheres lying on the reciprocal lattice grid. The outcome from the calculations compared to the reconstructed experimental diffraction patterns for the (0kl), (1kl) and (h1l) planes are shown in Fig. 1. 3.2. Structure description The structure description below of Co2Te3(PO4)O6Cl is based on the structure refinement in the monoclinic crystal system with space group P21/c. The crystal structure consists of layers, which stack along [001] and are held together by only weak Van der Waals interactions. Cavities can be observed running along [100] in between the layers, see Fig. 2. Although the stacking between the layers is irregular causing the diffuse scattering described above, the structure of each layer is well defined. The structure features two different linear cobalt oxochloride chains, which are surrounded by tellurium oxide and phosphate building blocks. All six unique tellurium atoms in the structure show one sided asymmetric coordinations due to the presence of the stereochemically active lone pair. Te1 and Te6 have trigonal pyramidal [TeO3] coordination with bonding distances between 1.79(3) Å e 1.99(3) Å. Te2 and Te3 show [TeO4] see-saw coordination, with three short TeeO bonds (1.86(3) Å e 2.08(3) Å) and one longer TeeO bond (2.14(3) Å, 2.20(3) Å). Both Te atoms have a further oxygen atom close by, which with a bonding length of >2.8 Å lies outside the primary bonding sphere, however this is commonly observed for Te4þ [17]. These two types of connections are very common for Te4þ and

69

found in e.g. Co5Te4O11Cl4 [18]. Te4 and Te5 build [TeO5] coordination polyhedral and this coordination is not commonly observed and seems to be present mainly in tellurium phosphate structures e.g. in Te2O(PO4)2 [19]. The geometry is similar to the one in [TeO4], but an additional long (~2.48 Å) TeeO bond is present. Furthermore there are some short TeeCl interactions of around ~3 Å, which are just at the border to be considered as belonging to the primary coordination sphere. The Te3eTe2 polyhedra form dimers and the Te1eTe6 and Te4eTe5 polyhedra polymerize into chains running along [100]. Two types of chains are formed by edge sharing [CoO6] and [CoO5Cl] building blocks to form [CoO4]∞ and [CoO3Cl]∞ zigzag chains respectively along [100], see Fig. 3. CoeO and CoeCl bonding distances are in between 1.99(3) Å e 2.31(3) Å and 2.516(14) Å e 2.584(14) Å respectively. The phosphate groups form regular tetrahedra with PeO bonding distances from 1.50(3) Å to 1.62(3) Å. The phosphate groups are connected via via corner sharing to the cobalt oxide chains, the cobalt oxochloride chains and to the tellurium oxide building blocks to build the layers which extend in the (110) plane. There are two layers present in the unit cell, which are related to each other by a c-glide operation. In between the layers there are weak van der Waals interactions between the tellurium and oxygen atoms, see Fig. 4. The cavities that are present between the layers are remarkably large compared to other layered oxohalides. 4. Conclusions Single crystals of the cobalt tellurium phosphate oxochloride Co2Te3(PO4)O6Cl were obtained by chemical reactions in sealed and evacuated silica tubes at 550  C. The crystal structure is layered along [001]. The layer coherence is given by weak Van der Waals interactions between tellurium and oxygen atoms. The charge neutral layers are made up by [CoO4]∞ and [CoO3Cl]∞ chains that are connected by tellurium oxide and phosphate building blocks. Diffraction experiments revealed strong diffuse scattering affecting the k ¼ 2n þ 1 diffraction lines, which originates from irregular layer stacking. The crystal structure was first solved in the triclinic system, space group P1, but the structure could also be refined in a larger monoclinic unit cell, space group P21/c. To better understand

Fig. 2. Different unit cell settings. (a) The triclinic unit cell, (bec) the monocl Different unit cell settings. (a) The triclinic unit cell, (bec) the monoclinic unit cell along [100] and [010] respectively. The orthonormal unit cells of layer1 (violet) and its mirrored (m perpendicular to [010]) version layer2 (blue) used in the calculations with DIFFaX are indicated.

70

I. Zimmermann, M. Johnsson / Solid State Sciences 40 (2015) 67e70

Fig. 3. Detailed view of the cobalt oxide and oxochloride zig-zag chains and the different coordination's of tellurium.

similar connectivity to the next layer and the relatively wide channels in between the layers which reduce the strength of the Van der Waals interactions in between them. Acknowledgments This work has been carried out through financial support from the Swedish Research Council (2011-3303). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.solidstatesciences.2015.01.002. References

Fig. 4. Interlayer coherence through long TeeO interactions.

the diffuse scattering, a computer simulation was set up using the program DIFFaX. The layers were allowed to stack randomly by a simple layer shift or glide operation. The calculated diffraction patterns did show a qualitative match with the experimental ones. Although a layer arrangement in transition metal oxohalides having a lone pair element present is very common, irregular stacking of layers has not been reported in the previously found compounds. The reasons for the stacking faults are most likely due to the fact that there are two positions available for each layer that results in

€ rnroos, P. Lemmens, P. Millet, Chem. Mater. 15 (2003) [1] M. Johnsson, K.W. To 68e73. [2] R. Becker, M. Johnsson, R.K. Kremer, H.H. Klauss, P. Lemmens, J. Am. Chem. Soc. 128 (2006) 15469e15475. [3] I. Zimmermann, M. Johnsson, S. Lidin, Eur. J. Inorg. Chem. 25 (2012) 3971e3974. [4] K. Nordlund, U. Beck, T.H. Metzger, J.R. Patel, Appl. Phys. Lett. 76 (2000) 846e848. [5] A. Boulle, D. Chaussende, F. Pecqueux, F. Conchon, L. LatueRomain, O. Masson, Phys. Stat. Sol. (A) 204 (2007) 2528e2534. [6] M. Barchuk, V. Holý, D. Kriegner, J. Stangl, S. Schwaiger, F. Scholz, Phys. Rev. B 84 (2011) 094113. [7] V. Svitlyk, B.J. Campbell, Y. Mozharivskyj, Inorg. Chem. 48 (2009) 10364e10370. [8] L.S. Ivashkevich, A.S. Lyakhov, A.F. Selevich, Z. Krist. 227 (2012) 334e340. [9] H. Dittrich, P. Axmann, M. WohlfahrteMehrens, J. Garche, S. Albrecht, J. MeeseeMarktscheffel, A. Olbrich, G. Gille, Z. Krist. 220 (2005) 306e315. [10] T. Willhammar, X. Zou, Z. Krist. 228 (2013) 11e27. €ber, T.R. Welberry, R.L. Withers, J.D. Fitz Gerald, Acta Cryst. [11] H. Krüger, S. Sto B67 (2011) 476e485. [12] Oxford diffraction, CrysAlisCCD and CrysAlisRED, Oxford Diffraction Ltd, Abingdon, Oxfordshire, England, 2006. [13] G.M. Sheldrick, Acta Cryst. A64 (2008) 112e122. [14] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837e838. [15] M.M.J. Treacy, J.M. Newsam, M.W. Deem, Proc. R. Soc. Lond. Ser. A 433 (1991) 499e520. [16] G. Bergerhoff, DIAMOND, Bonn, Germany, 1996. [17] J. Zemann, Monatsh. Chem. 102 (1971) 1209e1216. [18] R. Becker, M. Johnsson, H. Berger, Z. Anorg. Allg. Chem. 633 (2007) 422e424. [19] M.K. Kim, S.H. Kim, H.Y. Chang, P.S. Halasyamani, K.M. Ok, Inorg. Chem. 49 (2010) 7028e7034.