Pamoic acid determined from powder diffraction data

The title compound [systematic name: 4,4′-methyl­enebis(3-hydr­oxy-2-naphthoic acid)], C₂₃H₁₆O₆, has one half-mol­ecule in the asymmetric unit. The mol­ecular twofold rotational axis about the central C atom is preserved on crystallization. A chain formed by R₂²(8) hydrogen bonds runs along the c axis and an intra­molecular O—H⋯O=C—OH hydrogen bond is also formed. The crystal structure was solved by simulated annealing from laboratory X-ray powder diffraction data, with data collected at room temperature. Rietveld refinement of this model led to a final R_wp value of 0.0391 at 1.39 Å resolution.

Comment

The title compound, (I), is used in the pharmaceutical industry as a counter-ion to obtain long-acting formulations of certain basic drugs (Jorgensen, 1998), and is also of fundamental inter­est because of the possibility of extensive hydrogen bonding in the solid state. Attempts have been made to grow single crystals of pamoic acid for structure determination (Blackburn et al., 1996, Haynes et al., 2005) but these have proved unsuccessful; thus we have employed X-ray powder diffraction to solve and refine the crystal structure, which is reported here for the first time.

The compound crystallizes in the space group C2/c with one half-mol­ecule of pamoic acid in the asymmetric unit (Fig. 1). A twofold rotation axis passes through C1. The –OH and –CO₂H groups lie in the plane of the attached ring.

A corrugated chain of mol­ecules running along the c axis is formed by inter­molecular R₂²(8) hydrogen bonds (Fig. 2a). The hydrogen bonds themselves are oriented at approximately 45° to the chain propagation vector (Fig. 2b). Other than hydrogen bonding, the only short inter­molecular contact (less than the sum of the van der Waals radii) is an aromatic C—H⋯π inter­action (symmetry code: ½ − x, ½ + y,  − z) between H10 and C10 of 2.684 (8) Å, which can be compared with the van der Waals distance of 2.9 Å. Of 48362 short aromatic–aromatic C—H⋯π contacts in the Cambridge Structural Database (CSD; Version 5.26 of November 2005; Allen, 2002), only 624 (1.3%) are as short or shorter; the median is 2.837 Å. The bis-pyridinium salt of pamoic acid (CSD refcode TABMAK; Blackburn et al., 1996) exhibits a similar C—H⋯π distance of 2.646 Å, although between a different pair of atoms. This type of contact may therefore play a role in stabilizing the observed crystal structures in pamoic acid derivatives; however, further work will be necessary to confirm this.

Figure 1 View of (I), with the atom-numbering scheme. The unlabelled atoms are related to the labelled ones by the symmetry operation (1 − x, y, − z).

Figure 2 The crystal packing viewed (a) along the a axis and (b) along the b axis. The O—H⋯O hydrogen bonds are indicated by dashed lines.

Figure 3 Final observed (points), calculated (line), difference [(yobs − ycalc)] and weighted difference [(yobs − ycalc)/σ] profiles for the Rietveld refinement of the title compound. Change of scale at 40° is a factor of 10 and the increment in 2θ is 0.01°.

Experimental

Pamoic acid (97%+) was obtained from Sigma and used without further purification. No impurities were detected by X-ray powder diffraction. The sample was lightly ground and loaded into a 0.7 mm-diameter Lindemann glass capillary. Data were collected in Debye–Scherrer geometry employing Co Kα₁ radiation.

Crystal data

• C₂₃H₁₆O₆

• M_r = 388.37

• Monoclinic, C 2/c

• a = 19.7348 (7) Å

• b = 4.78768 (12) Å

• c = 19.2544 (4) Å

• β = 108.9622 (17)°

• V = 1720.5 (1) ų

• Z = 4

• D_x = 1.499 Mg m⁻³

• Co Kα₁ radiation

• μ = 2.19 mm⁻¹

• T = 298 K

• Specimen shape: cylinder

• 12 × 0.7 × 0.7 mm

• Specimen prepared at 298 K

Data collection

• Stoe linear PSD diffractometer

• Specimen mounting: 0.7 mm Lindemann glass capillary

• Specimen mounted in transmission mode

• Scan method: step

• Absorption correction: none

• 2θ_min = 2.0, 2θ_max = 80.0°

• Increment in 2θ = 0.0°

Refinement

• Refinement on I_net

• R_p = 0.030

• R_wp = 0.039

• R_exp = 0.035

• S = 1.20

• Profile function: pseudo Voigt (Thompson et al., 1987) with the asymmetry correction (Finger et al., 1994)

• 362 reflections

• 98 parameters

• H-atom parameters constrained

• Weighting scheme based on measured s.u.'s w = 1/σ(Y_obs)²

• (Δ/σ)_max = 1.61

• Preferred orientation correction: none

Table 1
Hydrogen-bond geometry (Å, °)

D—H⋯A	D—H	H⋯A	D⋯A	D—H⋯A
O2—H2A⋯O3^i	0.82	1.85	2.643 (4)	166
O1—H1A⋯O3	0.90	1.72	2.529 (4)	147
Symmetry code: (i) -x+1, -y-1, -z+1.

As noted in _pd_proc_ls_special_details, the refinement was characterised by a relatively shallow minimum with respect to the lattice parameters and zero point which were refined for the final cycle to generate s.u.'s..

Two powder diffraction patterns were collected. Initially a pattern was collected using a Philips Xpert diffractometer operating in Bragg–Brentano mode with a flat-plate sample. No monochromation or collimation was employed. A full pattern was collected in under five minutes; this was used to solve the crystal structure. Subsequent attempts to use this pattern in a Rietveld analysis were not successful owing to difficulties in modelling the diffraction peak shapes from the high-intensity operation mode. Thus a second pattern was collected for approximately 24 h using a monochromatic Stoe Stadi-P instrument operating in Debye–Scherrer geometry with the sample contained in a glass capillary. The resulting pattern proved suitable for a full Rietveld analysis.

The initial X-ray powder diffraction pattern was used for structure determination with the program DASH (David et al., 2004). The powder pattern was truncated to 45.3° 2θ (Cu Kα), corresponding to a real-space resolution of 2.0 Å. The background was subtracted with a Bayesian high-pass filter (David & Sivia, 2001). Peak positions for indexing were obtained by fitting with an asymmetry-corrected Voigt function, followed by indexing with the program DICVOL (Boultif & Louer, 1991). Pawley refinement was used to extract integrated intensities and their correlations, from which the space group was determined using Bayesian statistical analysis (Markvardsen et al., 2001). Possible space groups were Cc or C2/c, the latter implying that the mol­ecule sits on a special position; both space groups were tried. Simulated annealing was used to solve the crystal structure from the powder pattern in direct space. The starting mol­ecular geometry was taken from entry TABMAK (Blackburn et al., 1996) from the CSD. When choosing Cc as the space group, the asymmetric unit consists of a full pamoic acid mol­ecule, which then has four independent flexible torsions, one for each of the two carboxylic acid groups and two across the central C atom. All these four torsion angles were left fully flexible during the simulated annealing, which, combined with three translational and three rotational degrees of freedom, gives a total of ten degrees of freedom. When choosing C2/c as the space group, the pamoic acid mol­ecule is constrained by symmetry to sit on a twofold rotation axis through the central C atom, C1. Suitable constraints for atom C1 were therefore included in the simulated annealing runs for that space group, namely fixing its x coordinate at 1/2, fixing its z coordinate at 3/4 and setting its occupancy to 0.5 to account for site multiplicity. Imposing these constraints reduces the number of degrees of freedom from ten to five. The background subtraction, peak fitting, indexing, Pawley refinement, space-group determination and simulated annealing algorithms used are as implemented in the program DASH. With the default settings for the simulated annealing, ten simulated annealing runs for each of both space groups readily yielded ten identical crystal structures. The two space groups Cc and C2/c gave identical crystal structures with comparable figures of merit, indicating that the higher-symmetry space group was the correct one, and Rietveld refinement was carried out in C2/c only. Suitable constraints were imposed on bond lengths, angles and planar groups, including bonds to H atoms. The CH and CH₂ C—H distances were constrained to be 0.93 and 0.97 Å respectively, with C—C and C—O contraints taken from CSD entry TABMAK (Blackburn et al., 1996). The refinement (Fig. 3), using the GSAS software suite (Larson & Von Dreele, 2000), converged readily to yield acceptable figures of merit (χ² = 1.425, R_p = 0.0298 and R_wp = 0.0392) and a chemically reasonable structural model. An overall isotropic displacement parameter was employed to model the entire mol­ecule. Standard deviations are taken from the program employed and represent statistical uncertainties rather than estimates of the absolute error, which are likely to be considerably greater.

Data collection: WinXPow (Stoe & Cie, 1999); cell refinement: GSAS (Larson & Von Dreele, 2000); program(s) used to solve structure: DASH (David et al., 2004); program(s) used to refine structure: GSAS; molecular graphics: PLATON (Spek, 2003).

Supporting information

Computing details

Data collection: Stoe Software (Stoe & Cie, 2002); cell refinement: GSAS (Larson & Von Dreele, 2000); program(s) used to solve structure: DASH (David et al., 2004); program(s) used to refine structure: GSAS; molecular graphics: PLATON (Spek, 2003).

4,4'-methylenebis(3-hydroxy-2-naphthoic acid)

Crystal data

C23H16O6	Z = 4
Mr = 388.37	F(000) = 808.0
Monoclinic, C2/c	Dx = 1.499 Mg m^−3
Hall symbol: -C 2yc	Co Kα1 radiation, λ = 1.78892 Å
a = 19.7348 (7) Å	µ = 2.19 mm^−1
b = 4.78768 (12) Å	T = 298 K
c = 19.2544 (4) Å	yellow
β = 108.9622 (17)°	cylinder, 12 × 0.7 mm
V = 1720.5 (1) Å^3	Specimen preparation: Prepared at 298 K

Data collection

Stoe linear PSD diffractometer	Data collection mode: transmission
Radiation source: sealed X-ray tube, Stoe STADI-P	Scan method: step
Primary focussing, Ge 111 monochromator	2θmin = 2.0°, 2θmax = 79.99°, 2θstep = 0.01°
Specimen mounting: 0.7 mm Lindemann glass capillary

Refinement

Refinement on Inet	98 parameters
Least-squares matrix: selected elements only	82 restraints
Rp = 0.030	H-atom parameters constrained
Rwp = 0.039	Weighting scheme based on measured s.u.'s 1/σ(Yobs)^2
Rexp = 0.035	(Δ/σ)max = 1.61
R(F^2) = 0.07174	Background function: Chebyshev polynomial
7199 data points	Preferred orientation correction: none
Profile function: pseudo_Voigt (Thompson et al., 1987) with the asymmetry correction (Finger et al., 1994)

Special details

Refinement. CW Profile function number 3 with 19 terms Pseudovoigt profile coefficients as parameterized in P. Thompson, D·E. Cox & J·B. Hastings (1987). J. Appl. Cryst.,20,79–83. Asymmetry correction of L·W. Finger, D·E. Cox & A. P. Jephcoat (1994). J. Appl. Cryst.,27,892–900. #1(GU) = 77.113 #2(GV) = 62.885 #3(GW) = 0.872 #4(GP) = 0.000 #5(LX) = 4.481 #6(LY) = 0.000 #7(S/L) = 0.0130 #8(H/L) = 0.0430 #9(trns) = 0.00 #10(shft)= 0.0000 #11(stec)= 0.00 #12(ptec)= 0.00 #13(sfec)= 0.00 #14(L11) = 0.000 #15(L22) = 0.000 #16(L33) = 0.000 #17(L12) = 0.000 #18(L13) = 0.000 #19(L23) = 0.000 Peak tails are ignored where the intensity is below 0.0010 times the peak

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Ų)

	x	y	z	Uiso*/Ueq
C1	0.5	0.4911 (9)	0.75	0.1191 (10)*
H1	0.52716	0.6087	0.72753	0.1191 (10)*
C2	0.44887 (7)	0.3128 (8)	0.69098 (14)	0.1191 (10)*
C3	0.47616 (10)	0.1354 (10)	0.6498 (3)	0.1191 (10)*
C4	0.42915 (11)	−0.0317 (10)	0.5937 (2)	0.1191 (10)*
C5	0.45887 (14)	−0.2268 (11)	0.5504 (3)	0.1191 (10)*
C6	0.35639 (12)	0.0016 (10)	0.5744 (2)	0.1191 (10)*
H6	0.32608	−0.1053	0.5367	0.1191 (10)*
C7	0.32741 (10)	0.1893 (9)	0.6134 (2)	0.1191 (10)*
C8	0.25262 (11)	0.2114 (10)	0.5963 (2)	0.1191 (10)*
H8	0.22249	0.1056	0.5582	0.1191 (10)*
C9	0.22445 (14)	0.3951 (11)	0.6325 (3)	0.1191 (10)*
H9	0.17496	0.4089	0.6205	0.1191 (10)*
C10	0.26911 (15)	0.5591 (10)	0.6896 (3)	0.1191 (10)*
H10	0.24890	0.6825	0.7146	0.1191 (10)*
C11	0.34234 (13)	0.5319 (8)	0.7105 (2)	0.1191 (10)*
H11	0.37152	0.6486	0.7466	0.1191 (10)*
C12	0.37302 (8)	0.3591 (8)	0.6683 (2)	0.1191 (10)*
O1	0.54800 (13)	0.1241 (11)	0.6627 (3)	0.1191 (10)*
H1A	0.5579	−0.024	0.6387	0.1191 (10)*
O2	0.41729 (19)	−0.3766 (9)	0.5024 (3)	0.1191 (10)*
H2A	0.4407	−0.479	0.4852	0.1191 (10)*
O3	0.5278 (2)	−0.2538 (10)	0.5663 (3)	0.1191 (10)*

Geometric parameters (Å, º)

C1—H1	0.97	C7—C8	1.4078 (12)
C1—H1^i	0.97	C7—C12	1.4035 (11)
C2—C3	1.3845 (11)	C8—H8	0.93
C2—C12	1.4337 (11)	C8—C9	1.3494 (12)
C3—C4	1.4201 (11)	C9—H9	0.93
C3—O1	1.3587 (12)	C9—C10	1.4046 (12)
C4—C5	1.4930 (12)	C10—H10	0.93
C4—C6	1.3707 (12)	C10—C11	1.3744 (12)
C5—O2	1.244 (4)	C11—H11	0.93
C5—O3	1.299 (4)	C11—C12	1.4253 (12)
C6—H6	0.93	O1—H1A	0.90
C6—C7	1.4063 (11)	O2—H2A	0.82
H1—C1—H1^i	109	C7—C8—H8	120
C3—C2—C12	118.84 (8)	C7—C8—C9	120.28 (12)
C2—C3—C4	120.13 (14)	H8—C8—C9	120
C2—C3—O1	119.80 (11)	C8—C9—H9	120
C4—C3—O1	120.07 (13)	C8—C9—C10	120.66 (14)
C3—C4—C5	119.97 (15)	H9—C9—C10	120
C3—C4—C6	120.54 (9)	C9—C10—H10	120
C5—C4—C6	119.25 (9)	C9—C10—C11	120.60 (11)
C4—C5—O2	119.4 (2)	H10—C10—C11	120
C4—C5—O3	120.09 (12)	C10—C11—H11	120
O2—C5—O3	120.31 (14)	C10—C11—C12	119.26 (10)
C4—C6—H6	120	H11—C11—C12	120
C4—C6—C7	120.12 (9)	C2—C12—C7	119.36 (9)
H6—C6—C7	120	C2—C12—C11	120.75 (9)
C6—C7—C8	119.99 (10)	C7—C12—C11	118.57 (14)
C6—C7—C12	119.98 (13)	C3—O1—H1A	109
C8—C7—C12	120.01 (9)	C5—O2—H2A	109

Symmetry code: (i) −x+1, y, −z+3/2.

Hydrogen-bond geometry (Å, º)

D—H···A	D—H	H···A	D···A	D—H···A
O2—H2A···O3^ii	0.82	1.85	2.643 (4)	166
O1—H1A···O3	0.90	1.72	2.529 (4)	147

Symmetry code: (ii) −x+1, −y−1, −z+1.