174
G. Smith et al.
(4), 5-nitrosalicylic acid (5) and 3,5-dinitrosalicylic acid (6).
Preparations involved heating under reflux for 10 min at ca. 90°C in 20
cm3 of 95% ethanol, the product being subsequently obtained by the
partial or total evaporation of the solvent at room temperature, yielding
for (1) large yellow prisms (m.p. 88–90°C); for (2), yellow
microcrystals (m.p. 130–131°C); for (3), yellow prisms (m.p. 195–
198°C); for (4), yellow prisms (m.p. 161–162.5°C); for (5) a yellow
powder (m.p. 180–182°C); for (6) orange needles (m.p. 244–250°C).
Elemental analyses indicated 1 : 1 stoichiometries for all adducts except
(3) (1 : 2), while (1) was a monohydrate. With (4), the presence of 1.5
waters was found from analysis, later confirmed in the crystal structure
determination as a 2 : 2 trihydrate. Found for (1): C, 58.2; H, 4.3; N, 8.5.
C16H14N2O6 requires C, 58.2; H, 4.2; N, 8.5%. Found for (2): C, 61.5;
H, 3.8; N, 8.8. C16H12N2O5 requires C, 61.5; H, 3.8; N, 9.0%. Found for
(3): C, 57.7; H, 3.6; N, 8.8. C23H17N3O9 requires C, 57.6; H, 3.6; N,
8.8%. Found for (4): C, 50.0; H, 3.5; N, 10.8. C32H28N6O17 requires C,
50.0; H, 3.7; N, 10.9%. Found for (5): C, 58.5; H, 3.6; N, 8.3.
C16H12N2O6 requires C, 58.5; H, 3.7; N, 8.5%. Found for (6): C, 51.8;
H, 2.9; N, 11.3. C16H11N3O8 requires C, 51.5; H, 3.0; N, 11.3%.
Attempts to prepare the adduct with 1,3,5-trinitrobenzoic acid using
either the general refluxing method in ethanol at elevated temperature
or the room-temperature variant (because of the ease of
decarboxylation of the acid even at moderate temperature) gave good
crystals of the previously described 1 : 1 oxine–1,3,5-trinitrobenzene
adduct.[9] Infrared spectra were recorded for all samples as pressed
disks in KBr on a Perkin–Elmer Spectrum 1000 Fourier-transform
infrared spectrometer.
by full-matrix least-squares (on F2) by using SHELXL-97[21] with
anisotropic thermal parameters for all non-hydrogen atoms. Hydrogen
atoms involved in the hydrogen-bonding interactions were located by
difference methods and both positional and thermal parameters were
refined while others were generally included at calculated positions in
the refinements as riding models. The values of A and B in the
weighting scheme w = [σ2(Fo)2+(AP)2+BP]–1 {where P = [max.
(Fo ,0)+2(Fc)2]/3} were 0.0783, 0.2461 (3), 0.0593, 0.1264 (4), and
2
0.1826, 0.7519 (6). With (3), the high thermal motion of the nitro-
oxygens of molecule 1 [O(4) and O(5)] resulted in these being modelled
over two independent sites with occupancies of 0.80 and 0.20
respectively. Compound (6), which we had previously investigated
unsuccessfully crystallographically,[12] when it was found to have what
appeared to be bad disorder in the space group P21/c with Z 4 in a unit
cell with half the present b axial parameter, was resolved in the lower-
symmetry space group Pc with Z 8. In this space group, pseudo-
symmetry is present with the molecules related by an approximate 21
screw operation and inversion symmetry. This is somewhat analagous
to, but not as severe as, the modulated structure described for the 1 : 1
proton-transfer adduct of 2,4,6-trinitrobenzoic acid with 2,6-
diaminopyridine.[22] In addition, the presence of a partial disordered
labile molecule of oxine contributes to the high refinement residual
which could be reduced from the table value (R 0.080) to ca. R 0.05 by
incorporating electron density equivalent to 90% of a ghost oxine
molecule having an occupancy of 0.08. This effect may also be due to
the presence of an overriding but inadequately modelled stacking fault
or an inversion twin. However, this structural problem could not be
further handled in this study and furthermore, the Pc model adequately
describes the basic hydrogen-bonded structural framework which is not
altered by the presence of the the unasssociated ‘ghost oxine’. A similar
phenomenon was observed for the structure of the 1 :1 proton-transfer
compound of Kemp’s triacid with oxine,[13] in which a partial
disordered oxine molecule of crystallization was lost progressively with
time during room-temperature data collection, without significant loss
in crystallinity.
Crystallography
Crystal Data
(3) [(C9H8NO+)(C7H4NO4–)(C7H5NO4)], CCDC 164598, mol. wt
–
479.4, triclinic, space group P1, a 7.052(2), b 7.497(2), c 21.121(6) Å,
α 97.94(2), β 97.35(3), γ 99.45(2)°, V 1077.9(5) Å3, F(000) 496, Z 2,
Dc 1.477 g cm–3, µ(Mo Kα) 1.16 cm–1, temperature 293(2) K. 4089
reflections measured [2θmax 51°: h, 0 to 8; k, –8 to 8; l, –25 to 24], 3756
unique (Rint 0.012). Final R1* 0.048 (F); wR2† 0.123 (F2) [2845
observed with I > 2.0(I)]; S 1.03. Crystal size 0.50 by 0.30 by
0.25 mm; max/min transmission factors, 0.82/0.70.
Conclusions
The series of quinolin-8-ol compounds with the nitro-
substituted benzoic acids studied here shows a greater
variability in composition and form than any of the other
Lewis-base compounds of the same acids previously studied.
The incorporation of water into the cocrystal make-up in two
of the examples [(1) and (4)] obviously influences the
stabilization of the crystal lattice, which in the case of the
trihydrate (4) links together rows of oxine and dnba
molecules to give an very stable and ordered structure. This
is in contrast to the 1 : 1 compound with dnsa (6), where a
probable stacking fault is present in the structure, and may be
similar to the badly twinned monohydrate (1) where no
structural analysis was possible, despite its macro-crystalline
morphology. Infrared spectroscopy has also proved useful, as
previously noted,[23] for the recognition of proton transfer in
compounds of the type found in this series, particularly with
those for which crystal structure determinations are not
possible e.g. (1), (2) and (5). As with the parallel study
involving the same series of acids with the analogous
nitrogen base 8-aminoquinoline,[24] disappearance of the
characteristic strong C=O stretching frequency at ca.1700
cm–1 is the most definitive sign in the composite spectrum of
proton transfer [except in the case of the 1 : 2 adduct (3)
where free acid is present]. The accompanying broad =N–H+
(4) [(C H NO+) (C H N O –) ·3H O], CCDC 164597, mol. wt
9
8
2
7
3
2
6
2
2
768.6, monoclinic, space group P21, a 14.9606(6), b 7.4287(5), c
15.633(1) Å, β 102.170(4)°, V 1698.4(2) Å3, F(000) 796, Z 2, Dc 1.503
g cm–3, µ(Cu Kα) 1.22 cm–1, temperature 293(2) K. 3851 reflections
measured [2θmax 150°: h, –18 to 0; k, 0 to 9; l, –19 to 19], 3705 unique
(Rint 0.022). Final R1 0.029 (F); wR2 0.084 (F2) [3564 observed with
I > 2.0σ(I)]; S 1.04; absolute structure parameter, 0.01(15). Crystal size
0.43 by 0.30 by 0.23 mm.
(6) [(C9H8NO+)(C7H3N2O7 )], CCDC 164599, mol. wt 373.3,
–
monoclinic, space group Pc, a 8.015(2), b 15.010(2), c 26.546(2) Å, β
95.52(2)°, V 3178.8(9) Å3, F(000) 1536, Z 8, Dc 1.560 g cm–3, µ(Cu
Kα) 1.11 cm–1, temperature 293(2) K. 7322 reflections measured
[2θmax 140°: h, –6 to 9; k, –12 to 18; l, –32 to 32], 7103 unique (Rint
0.013). Final R1 0.080 (F); wR2 0.210 (F2) [6101 observed with
I > 2.0σ(I)]; S 1.03; absolute structure parameter, 0.4(3). Crystal size
0.78 by 0.21 by 0.14 mm; max/min transmission factors, 0.87/0.64.
Data Collection, Structure Solution and Refinement
X-Ray diffraction data for all compounds were measured on Enraf–
Nonius CAD-4 diffractometers by using either crystal-
monochromatized Mo Kα X-radiation (λ 0.71073 Å) [compound (3)]
or Cu Kα X-radiation (λ 1.5418 Å) [compounds (4) and (6)]. Negligible
change in the intensities of three standards monitored throughout the
data collection periods for all adducts indicated no significant crystal
decomposition. Data were corrected for Lorentz and polarization
effects, extinction and for absorption [(3) and (6) only (analytical,
Gaussian)]. The structures were solved by direct methods and refined
*R1 = (⌺ | Fo | – |Fc|)/(⌺ |Fo|).
†wR2 = [⌺w(Fo – Fc )2/⌺w(Fo )2]½.
2
2
2