Crystal structures of dimetal terephthalate dihydroxides, M 2(C8H4O4)(OH)2 (M = Co, Ni, Zn) from powder diffraction data and DFT calculations

A new ordered C2/c model is proposed for the dimetal terephthalate dihydroxides M = Co, Ni, and Zn, compared to the previous disordered C2/m model for M = Co.


Chemical context
Dicobalt terephthalate dihydroxide, Co 2 (C 8 H 4 O 4 )(OH) 2 , was first prepared by Sherif (1970). A powder pattern was reported, but no unit cell or crystal structure were determined. The powder pattern from this reference is included in the Powder Diffraction File (Gates-Rector & Blanton, 2019) as entry 00-034-1897. A search of the nine peaks of this entry against the PDF-4 Organics 2022 database yielded no additional terephthalate compounds.
Approximately 20 years ago, one of us (JAK) solved and refined the structure of Zn 2 (C 8 H 4 O 4 )(OH) 2 using synchrotron powder data, first in a C2/m cell with disordered terephthalate anions. It then became apparent that if the c-axis were doubled, the systematic absences corresponded to space group C2/c. This doubled unit cell removed the disorder and yielded a more satisfactory refinement. This structure was deposited in the Cambridge Structural Database (Kaduk, 2016;refcode PUCYAO01), but never otherwise published or discussed. Since that time, another polymorph of Zn 2 (C 8 H 4 O 4 )(OH) 2 (in space group P2 1 /c) has been reported (Carton et al., 2009;PUCYAO). Some of our recent attempts to prepare Co and Ni porous metal-organic frameworks (MOFs) yielded instead cobalt and nickel terephthalate hydroxide. We took advantage of the opportunity to re-refine the structures (as well as that of Zn) in what we believe to be the correct space group, and to optimize the structures using density functional techniques.

Structural commentary
Doubling the c-axis of the previously reported disordered C2/m model for Co results in a chemically-reasonable ordered C2/c structure for these compounds. The X-ray powder diffraction patterns show that the three compounds are isostructural (Fig. 1). The root-mean-square Cartesian displacements of the non-H atoms in the Rietveld-refined and DFT-optimized structures are 0.125, 0.143, and 0.339 Å for Co,Ni,and Zn,. The good agreement provides strong evidence that the structures are correct (van de Streek & Neumann, 2014). This discussion concentrates on the DFT-optimized structures. The asymmetric unit (with atom numbering) is illustrated in Fig. 5. The best view of the crystal structure is down the b-axis (Fig. 6). A view down the caxis is shown in Fig. 7.

Figure 4
Comparison of the Rietveld-refined (red) and VASP-optimized ( The peak profiles are dominated by microstrain broadening. The generalized microstrain model was used for Co and Zn, but the limited Ni data supported refinement of only an isotropic broadening coefficient. The average microstrain is similar for Co and Zn (21042 and 20094 ppm, respectively), while that for Ni is much larger, at 114830 ppm. Perhaps this greater microstrain indicates that some square-planar Ni coordination also occurs. Analysis of the contributions to the total crystal energy of the structure using the Forcite module of Materials Studio (Dassault Systè mes, 2021) suggests that for Co and Ni, the bond and angle distortion terms dominate intramolecular deformation energy, but that torsion terms are also significant. For Zn, the angle distortion terms dominate the intramolecular deformation energy. The intermolecular energy in all three compounds is dominated by electrostatic attractions, which represent the M-O bonds.

Figure 6
The crystal structure of Co 2 (C 8 H 4 O 4 )(OH) 2 , viewed down the b-axis direction.

Figure 7
The crystal structure of Co 2 (C 8 H 4 O 4 )(OH) 2 , viewed down the c-axis direction.

Figure 8
The layers in the crystal structure of Co 2 (C 8 H 4 O 4 )(OH) 2 , viewed down the a-axis direction. that we might expect elongated (with [010] as the long axis) or platy (with {200} as the major faces) morphology for these compounds. A 2nd order spherical harmonic model was included in the refinement. The texture indices were 1.003, 1.417, and 1.016 for Co, Ni, and Zn respectively, showing that preferred orientation was significant only for the flat-plate Ni specimen.

Supramolecular features
The octahedral coordination spheres of Co9 share edges, forming chains running parallel to the b-axis direction; the shared edges are parallel the a-axis direction. The octahedral coordination spheres of Co10 share edges, forming chains propagating along the b-axis; in this case, the shared edges lie parallel to the c-axis direction. Co9 and Co10 share corners (via O11 = the hydroxyl group), forming layers lying perpendicular to the a-axis direction (Fig. 8). The hydroxyl group does not participate in hydrogen bonds.

Database survey
The crystal structure of the 'new terephthalate-based cobalt hydroxide Co 2 (OH) 2 (C 8 H 4 O 4 )' was reported by Huang et al. (2000), and its crystal structure determined [Cambridge Structural Database (Groom et al., 2016) refcode QASLIF] by ab initio methods using X-ray powder diffraction data. The and Z = 2. The structure consists of alternating Co-hydroxide and terephthalate layers, and the terephthalate anions are disordered about an inversion center. Antiferromagnetic ordering in this compound was studied using neutron powder diffraction by Feyerherm et al. (2003), using the same unit cell (QASLIF02). The structure was also determined by Kurmoo et al. (2001;QASLIF01) in the same unit cell, as well as the structure of cobalt terephthalate dihydrate. The structures of a series of (Co,Fe) 2 (C 8 H 4 O 4 )(OH) 2 solid solutions were refined in the same unit cell by Mesbah et al. (2010)

Synthesis and crystallization
Cobalt(II) nitrate hexahydrate (0.0364 g, 0.125 mmol) and terephthalic acid (0.0208 g, 0.125 mmol) were added to a flask followed by 0.125 ml of triethylamine and approximately 5 ml of dimethylformamide. The reaction was stirred for 10 min until a homogenous mixture was obtained. The reaction was heated using a CEM Discover microwave with power set to 150 W using a ramp time of 2 min to reach 423 K with a hold time of 30 min and internal stirring switched off. The vial remained in the microwave until it cooled to 323 K, and the reaction mixture was filtered using vacuum filtration, washed  Step Step Step 2  with DMF and deionized water (10 ml each). The remaining solid was dried in an oven at 343 K under vacuum. Nickel(II) nitrate hexahydrate (0.1948 g, 0.67 mmol) and terephthalic acid (0.2492 g, 1.5 mmol) were dissolved in 10 ml of DMF in a microwave vial. The solution was stirred until homogenous. The solution was then heated using a CEM Mars 6 microwave reactor at 750 W for a total of 85 s, in increments of 25 and 60 s. The resulting green solid was isolated using vacuum filtration, washed with water, methanol, and acetone, and allowed to air dry.
Information on the synthesis of Zn 2 (C 8 H 4 O 4 )(OH) 2 from prior to 1997 is no longer available.

Refinement
Crystal data, data collection and structure refinement details are summarized in Table 1.
Rietveld refinements (Figs. 9-11) were carried out using GSAS-II (Toby & Von Dreele, 2013). All non-H bond distances and angles in the terephthalate anions were subjected to restraints, based on a Mercury Mogul Geometry Check (Sykes et al., 2011;Bruno et al., 2004). The Mogul average and standard deviation for each quantity were used as the restraint parameters. The restraints contributed 0-2.3% to the final 2 . The U iso were grouped by chemical similarity. The U iso values for the H atoms were fixed at 1.3 Â the U iso of the heavy atoms to which they are attached. The peak profiles were described using the generalized microstrain model. The background was modeled using a 3-12-term shifted Chebyshev polynomial.
The structures were optimized with density functional techniques using VASP (Kresse & Furthmü ller, 1996)  The Rietveld plot for the refinement of Ni 2 (C 8 H 4 O 4 )(OH) 2 . The blue crosses represent the observed data points, and the green line is the calculated pattern. The cyan curve is the normalized error plot. The row of tick marks indicates the calculated reflection positions.

Figure 11
The Rietveld plot for the refinement of Zn 2 (C 8 H 4 O 4 )(OH) 2 . The blue crosses represent the observed data points, and the green line is the calculated pattern. The cyan curve is the normalized error plot. The row of tick marks indicates the calculated reflection positions. The vertical scale has been multiplied by a factor of 5Â for 2 > 10.0 , and by a factor of 15Â for 2 > 18.0 .