Crystal structures of two magnesium citrates from powder diffraction data

The crystal structures of magnesium hydrogen citrate dihydrate and bis(dihydrogencitrato)magnesium have been solved and refined using synchrotron X-ray powder diffraction data, and optimized using density functional theory techniques.


Chemical context
A systematic study of the crystal structures of Group 1 (alkali metal) citrate salts has been reported in Rammohan & Kaduk (2018). This paper represents the extension of the study to Group 2 (alkaline earth) citrates. The only magnesium citrate previously reported is Mg 3 (C 6 H 5 O 7 ) 2 (H 2 O) 10 Johnson, 1965). I now describe the syntheses and crystal structures of magnesium hydrogen citrate dihydrate, Mg(HC 6 H 5 O 7 )(H 2 O) 2 (I) and bis(dihydrogencitrato)magnesium, Mg(H 2 C 6 H 5 O 7 ) 2 (II). Attempts to prepare Be(H 2 C 6 H 5 O 7 ) 2 , BeHC 6 H 5 O 7 , and Be 3 (C 6 H 5 O 7 ) 2 by HClcatalyzed reaction of Be metal with a citric acid solution have so far yielded only amorphous products (see Fig. S1 in the supporting information).

Structural commentary
The crystal structure of (I) was solved and refined using synchrotron X-ray powder diffraction data, and optimized ISSN 2056-9890 using density functional techniques. (Fig. 1) The root-meansquare Cartesian displacement of the non-hydrogen citrate atoms in the Rietveld refined and DFT-optimized structures is 0.062 Å (Fig. 2) The absolute difference in the position of the Mg cation in the unit cell is 0.055 Å . The excellent agreement between the structures is evidence that the experimental structure is correct (van de Streek & Neumann, 2014): the rest of the discussion will emphasize the DFT-optimized structure. All of the citrate bond distances, bond angles, and torsion angles fall within the normal ranges indicated by a Mercury Mogul geometry check (Macrae et al., 2020). The citrate anion occurs in the trans, trans-conformation (about C2-C3 and C3-C4, respectively), which is one of the two low-energy conformations of an isolated citrate anion (Rammohan & Kaduk, 2018). The central carboxylate group and the hydroxyl group exhibit a significant twist [O17-C3-C6-O15 = À15.6 ] from the normal planar arrangement.
The Mg cation in (I) is six-coordinate (octahedral); the ligands are three carboxylate oxygen atoms, the citrate hydroxyl group, and two cis water molecules. The Mulliken overlap populations indicate that the Mg-O bonds have significant covalent character. The Mg bond-valence sum is 2.22. The citrate anion triply chelates to the Mg cation through the terminal carboxylate O14, the central carboxylate O15, and the hydroxyl group O17 oxygen atoms.
The Bravais-Friedel-Donnay-Harker (Bravais, 1866;Friedel, 1907;Donnay & Harker, 1937) method suggests that we might expect platy morphology for magnesium hydrogen citrate dihydrate, with {200} as the major faces. A 4th order spherical harmonic model was included in the refinement. The texture index was 1.000 (0), indicating that preferred orientation was not significant in this rotated capillary specimen.
The crystal structure of (II) was solved and refined in the same way (Fig. 3) The root-mean-square Cartesian displacement of the non-hydrogen citrate atoms in the Rietveld Comparison of the refined and optimized structures of (I). The refined structure is in red, and the DFT-optimized structure is in blue.

Figure 3
The asymmetric unit of (II) with the atom numbering and 50% probability spheroids.

Figure 4
Comparison of the refined and optimized structures of (II). The refined structure is in red, and the DFT-optimized structure is in blue.

Figure 1
The expanded asymmetric unit of (I) with the atom numbering and 50% probability spheroids. Symmetry-generated atoms [Mg19(x, y, z À 1) and O13(x, y, z + 1)] are linked by dashed bonds. refined and DFT-optimized structures is 0.043 Å (Fig. 4). The excellent agreement between the structures is evidence that the experimental structure is correct (van de Streek & Neumann, 2014) and this discussion will emphasize the DFToptimized structure. All of the citrate bond distances, bond angles, and torsion angles fall within the normal ranges indicated by a Mercury Mogul geometry check (Macrae et al., 2020). The citrate anion occurs in the trans, gauche-conformation (about C2-C3 and C3-C4, respectively), which is one of the two low-energy conformations of an isolated citrate anion (Rammohan & Kaduk, 2018). The central carboxylate group and the hydroxyl group exhibit a significant twist [O17-C3-C6-O16 = 10.6 ] from the normal planar arrangement.
The magnesium cation in (II) is six-coordinate (octahedral) and resides on a twofold axis; the ligands are two cis hydroxyl groups and 4 central carboxylate groups O16. Ionizing the central carboxylate group of citric acid first is the normal pattern (Rammohan & Kaduk, 2018

Figure 6
The crystal structure of (II), viewed down the c axis.
ionized terminal carboxylic acid groups form centrosymmetric R 2 2 (8) loops, which link the citrate anions into chains along the c-axis direction. The hydroxyl group O17 forms an intermolecular hydrogen bond to the central carboxylate O15. The energies of the O-HÁ Á ÁO hydrogen bonds were calculated using the correlation of Rammohan & Kaduk (2018). Weak C-HÁ Á ÁO hydrogen bonds are also present (

Database survey
Details of the comprehensive literature search for citrate structures are presented in Rammohan & Kaduk (2018). A search of the Cambridge Structural Database (Groom et al., 2016) using a citrate fragment and Mg, C, H, and O only yielded Mg 3 (C 6 H 5 O 7 ) 2 (H 2 O) 10 (MGCITD; Johnson, 1965). Reduced-cell searches using the unit cells of both compounds of this study yielded no citrate structures. A search of the Powder Diffraction File (Gates-Rector & Blanton, 2019) yielded entry 02-063-3628 calculated from MGCITD, as well as the experimental entry 00-001-0186 (Hanawalt et al., 1938) for the same compound.

Synthesis and crystallization
To prepare (I), magnesium hydrogen citrate dihydrate was synthesized by dissolving 2.0798 g (10.0 mmol) of H 3 C 6 H 5 O 7 (H 2 O) in 10 ml of water, and adding 0.8427 g (10.0 mmol) of 'MgCO 3 ' to the clear solution [the magnesium carbonate reagent was actually Mg 5 (CO 3 ) 4 (OH) 2 ]. After slow fizzing, a clear colorless solution was obtained. This solution was dried in a 333 K oven to yield (I) as a white solid.
Compound (II) was obtained from the scale [94.5 (1) wt% magnesian calcite Ca 0.84 Mg 0.16 CO 3 , 5.3 (4) wt% brucite Mg(OH) 2 , and 0.2 (1) wt% vaterite polymorph of CaCO 3 ] in a Megahome water still. The still was cleaned by filling the tank with tap water (from Lake Michigan), adding several tablespoons of citric acid monohydrate, and boiling for $2 h. The pale-yellow solution was decanted into a plastic pail, and allowed to evaporate at ambient conditions. Over five months, several white solids (calcium citrates, which will be discussed in another paper) crystallized, and were isolated. After five months, a clear yellow syrup remained. This was dried at 423 K to yield (II) as a white powder.

Refinement
Crystal data, data collection and structure refinement details for (I) are summarized in Table 3. A laboratory powder pattern, measured using Cu K radiation, was indexed using DICVOL (Louë r & Boultif, 2007) as incorporated into FOX (Favre-Nicolin & Č erný, 2002) on a primitive orthorhombic cell with a = 26.9042 (24), b = 5.9323 (4), c = 6.1649 (5) Å , V = 985.27 (17) Å 3 , and Z = 4. Attempts to solve the structure with multiple programs using the laboratory data were unsuccessful. The powder pattern measured at 11-BM using a wavelength of 0.413070 Å was indexed on a primitive orthorhombic cell with DICVOL as incorporated into FOX: a = 26.91159 (14), b = 5.92442 (2), c = 6.15170 (2) Å , V = 980.800 (7) Å 3 , and Z = 4. The Space Group Explorer suggested Pna2 1 , which was confirmed by successful solution and refinement of the structure. The structure was solved using Monte Carlo-simulated annealing techniques as implemented in FOX. The scatterers were a citrate anion, a Mg atom, and two O atoms (water molecules). In the best solution, one of the water molecules was too close to a carboxylate oxygen atom, and was discarded. The Mg coordination was 5/6 of an octahedron, so the second water molecule was placed manually using Materials Studio (Dassault Systems, 2019).    Rietveld plot for (I). 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 vertical scale has been multiplied by a factor of 5Â for 2 > 10.0 , and by a factor of 20Â for 2 > 15.0 . The row of blue tick marks indicates the calculated reflection positions, and the red tick marks indicate the peak positions for the citric acid impurity. The red line is the background curve.
The structure of (I) was refined by the Rietveld method using GSAS-II (Toby & Von Dreele, 2013) (Fig. 8). The initial refinement clarified the presence of extra peaks, which were identified as citric acid (02-061-2110; CITRAC10), which was added as a second phase; its concentration refined to 12.2 wt%. A few very weak peaks indicate the presence of an unidentified impurity. Analysis of potential hydrogen bonding using Mercury (Macrae et al., 2020) made it possible to determine approximate positions for the hydroxyl hydrogen atom H18 and the four water molecule hydrogen atoms. The C1-O12 bond was longer than the other carboxylate distances, and the O12Á Á ÁO16 i distance was 2.62 Å , making it clear that H26, the proton of the un-ionized carboxyl group, was located on O12. All heavy-atom bond distances and angles of the citrate anion were restrained: C1-C2 = C4-C5 = 1.51 (3) (3) . The restraints contributed 1.5% to the final 2 . The hydrogen atoms were included in fixed positions, which were re-calculated during the course of the refinement using Materials Studio. The U iso values of C2, C3, and C4 were constrained to be equal, and those of H7, H8, H9, and H10 were constrained to be 1.3Â that of these carbon atoms. The U iso values of C1, C5, C6, and the oxygen atoms were constrained to be equal, and that of H18 was constrained to be 1.3Â this value. The U iso values of the O atoms of the water molecules were constrained to be equal, and the U iso values of their H atoms to be 1.3Â this value. The background was described by a four-term shifted Chebyshev polynomial, with a peak at 10.84 to describe the scattering from the Kapton capillary and any amorphous component.
A density functional geometry optimization for (I) (fixed experimental unit cell) was carried out using CRYSTAL09 (Dovesi et al., 2005). The basis sets for the H, C, N, and O atoms were those of Gatti et al. (1994), and the basis set for Mg was that of McCarthy & Harrison (1994). The calculation used Rietveld plot for (II). 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 vertical scale has been multiplied by a factor of 2Â for 2 > 3.0 , by a factor of 10Â for 2 > 12.0 , and by a factor of 40Â for 2 > 17.0 . The row of blue tick marks indicates the calculated reflection positions. The red line is the background curve.   (2) , V = 1500.241 (8) Å 3 , and Z = 4. The systematic absences unambiguously determined the space group as P2 1 /c The structure was solved by direct methods using EXPO2009 (Altomare et al., 2013), assuming that it was a Ca salt. During the refinement, the electron density at the metal site and the metal-oxygen bond distances made it clear that it was a Mg salt rather than a Ca compound. The structure was refined by the Rietveld method using GSAS-II (Toby & Von Dreele, 2013) (Fig. 9). Analysis of the refined structure using PLATON (Spek, 2020) and the Find Symmetry module of Materials Studio (Dassault Systems, 2019) suggested the presence of extra symmetry, and that the true space group was C2/c (transformation matrix 1 0 1 / 0 1 0 / 0 0 1). The structure was re-refined in this space group, using the strategy described above for (I). The position of the peak in the background was 5.37 .
A density functional geometry optimization for (II) (fixed experimental unit cell) was carried out using CRYSTAL17 (Dovesi et al., 2018). The basis sets for the H, C, N, and O atoms were those of Gatti et al. (1994), and the basis set for Mg was that of Peintinger et al. (2013). The calculation used 8 k-points and the B3LYP functional, and took $15 h on a 3.54 GHz PC.
A density functional geometry optimization (fixed experimental unit cell) of the structure of magnesium citrate decahydrate (MGCITD) was carried out using CRYSTAL09 (Dovesi et al., 2005). The basis sets for the H, C, N, and O atoms were those of Gatti et al. (1994), and the basis set for Mg was that of McCarthy & Harrison (1994). The calculation used 8 k-points and the B3LYP functional, and took 11 days on a 2.4 GHz PC.