DOI: 10.1021/cg100315u
Tuning Proton Behavior in a Ternary Molecular Complex Lynne H. Thomas,† Nicholas Blagden,§ Matthias J. Gutmann,# Andras A. Kallay,† Andrew Parkin,† Colin C. Seaton,‡ and Chick C. Wilson*,†
2010, Vol. 10 2770–2774
†
Department of Chemistry and WestCHEM Research School, University of Glasgow, Glasgow G12 8QQ, U.K., ‡The Molecular Materials Centre, School of Chemical Engineering and Analytical Sciences, University of Manchester, P.O. Box 88, Manchester M60 1QD, U.K., §Institute of Pharmaceutical Innovation/School of Pharmacy, University of Bradford, Bradford BD7 1DP, U.K., and #ISIS Facility, STFC Rutherford Appleton Laboratory, Harwell Innovation Campus, Didcot, Oxon OX11 0QX, U.K. Received March 10, 2010; Revised Manuscript Received April 28, 2010
ABSTRACT: The multicomponent ternary complex of 4-dimethylaminobenzoic acid (4-DABA), 3,5-dinitrobenzoic acid (3,5-DNBA), and 4,40 -bipyridine (BIPY) has been studied by variable temperature X-ray and neutron diffraction. Proton disorder is observed within the 4-DABA homodimers present and quantitatively evaluated from neutron data. The effect of the crystal environment and in particular the pyramidalization of the nitrogen atom within the 4-DABA molecule and the consequential effect on the presence of hydrogen atom disorder are discussed with reference to the previously determined pure 4-DABA structure and the binary cocrystal with 3,5-DNBA.
Introduction Intermolecular hydrogen bonding plays an important role in forming anisotropic interactions in condensed systems, and subtle competition between hydrogen bond acceptors/donors can lead to dramatically different solid-state structures. There is significant effort directed at the study of hydrogen bonding in the solid state. Within the field of molecular solids, this includes fundamental studies to determine the geometric parameters in hydrogen bonded systems, to classification of hydrogen-bonding motifs, the field of supramolecular chemistry, and attempts to predict structural motifs based on acceptor/donor properties, through to the use of cocrystallization as a means of altering the physical and chemical properties of one or both of the components.1,2 This type of crystal engineering has been highlighted by Warner as an example of “green chemistry” since it allows the physical properties of a compound to be modified by altering secondary interactions rather than by the more energy-intensive and wasteful process of making and breaking primary bonds.3 Other examples of direct practical application include the cocrystallization of pharmaceutically active compounds to enhance their solubility and hence bioavailability.4 Much of this work has focused on the static structures adopted by molecular hydrogen-bonded systems, but the importance of hydrogen atom transfer through hydrogen bonds between molecules has also been identified. Such processes enable charge and energy transfer in solid chemical and biological systems and have widespread implications for issues as diverse as ferroelectrics, electrochemical processes, reaction intermediates, and enzyme action.5 It is becoming increasingly apparent that the positions of the protons involved in hydrogen bonds are vital in understanding structure and properties in such materials, and are highly susceptible not only to the chemical environment but also to the effects of temperature and pressure.
*To whom correspondence should be addressed. E-mail: c.c.wilson@ chem.gla.ac.uk. pubs.acs.org/crystal
Published on Web 05/13/2010
The targeted synthesis of molecular complexes as a crystal engineering activity has evolved approaches to assign selectivity to specific intermolecular interactions within a structure.6,7 This has been achieved with a high degree of success. This design success is based on a static view of the assembly process, with verification of the solid-state molecular assembly being obtained from single-crystal diffraction studies. An important aspect of such solid-state structure, where intermolecular hydrogen bonding plays an important role, or, particularly in cases where hydrogen transfer activity may be present, is the reliable determination of hydrogen atom parameters. Accurate hydrogen atom location through singlecrystal X-ray and neutron diffraction techniques is becoming increasingly routine.8 However, the issue of the dynamics of specific intermolecular interactions within the target architectures is less understood, as it is still not common to apply variable temperature diffraction measurements to such systems. Within the context of rationalizing in greater detail molecular solid design and function which utilizes hydrogen bonding networks, multicondition studies of hydrogen bonds can allow identification of evolving hydrogen atom behavior within hydrogen bonds such as proton migration and disorder effects.9 Typically, systems that exhibit proton disorder, where a second hydrogen site can be occupied on increasing the temperature, tend to be within the class of moderate hydrogen bonds having an associated energy in the 4-15 kcal mol-1 range and a donor-acceptor distance of 2.5-3.2 A˚.10 In this case, the potential surface in which the hydrogen atom sits can be described as an asymmetric double-well potential,11,12 where there is a significant energy difference between the two wells, induced in the solid-state due to the anisotropy of crystal field effects. To this end, imaging the electron density of such hydrogen atoms through Fourier difference maps derived from single-crystal X-ray diffraction has been shown to provide useful qualitative information on such effects.13 However, among diffraction approaches, it is neutron diffraction which remains the definitive technique for determining precise hydrogen atom parameters and in unequivocally identifying the presence of proton disorder or migration.11 r 2010 American Chemical Society
Article
Crystal Growth & Design, Vol. 10, No. 6, 2010
Crucially, it is also possible to refine site occupancies of hydrogen atom positions from these neutron data. Moreover, by combining X-ray and neutron single-crystal diffraction, the behavior of hydrogen can be probed more deeply when utilizing complementary information. Fourier difference maps derived from high quality X-ray data directly probe the bonding density of the hydrogen bond, whereas those derived from neutron data are probing the nuclear density; significantly, these may not be identifying the same effects particularly in the case of short, strong hydrogen bonds.14 It is our aim to design molecular systems with tunable proton transfer or disorder in the solid-state. pKa matching and mismatching has proved to be helpful in some cases9,15 and misleading in others,16,17 making it an unreliable technique for application to such problems. The driver therefore has to remain experimental, with the design of molecules and complexes aimed at perturbing the electronic structure of the system of interest in order to induce such proton transfer effects. While it is difficult to predict with any reliability whether the presence of hydrogen atom disorder will be observed, previously we have shown that the crystal environment perturbing the local symmetry of related molecular building blocks can have a significant influence on the presence of hydrogen atom disorder. In pure 4-dimethylaminobenzoic acid (4-DABA), the crystal structure consists of hydrogen-bonded dimers between the carboxylic acid groups.18 Within these dimers, hydrogen atom disorder has been observed through imaging of the electron density through Fourier difference maps.16 By perturbing the local environment by addition of the coformer 3,5-dinitrobenzoic acid (3,5-DNBA) into the crystal lattice, while the homodimer persists,19 no hydrogen disorder was observed within it.20 Instead, the disorder moved to the 3,5-DNBA homodimers. This could be attributed to an increased pyramidalization of the nitrogen atom in the 4-DABA molecules when these were housed within the cocrystal due to the charge transfer interaction between the 4-DABA and 3,5-DNBA molecules. The work presented here explores the perturbation of the crystal environment of 4-DABA molecular complexes further by introducing a third component into the crystal structure thus moving our study from that of a binary complex to that of a ternary complex. The utility of a ternary system is that the crystal environment of the intermolecular interaction under study may be more significantly modified by variation of the structure and the available bonding motifs of the third species than is possible with a binary system. While the designed creation of ternary complexes has been less studied compared to binary systems,7,20 the design processes used in such materials may be applied in these cases. Complementary intermolecular interactions must be identified between the various components; thus for this system, a third component was selected such that the homodimer between 3,5-DNBA molecules was disrupted while retaining the 4-DABA homodimer. As 3,5-DNBA is a stronger acid and better hydrogen bond donor compared to 4-DABA, the introduction of a basic proton accepting molecule such as 4,40 -bipyridine (BIPY) would be expected to lead to selective binding of this moiety to the 3,5-DNBA and not to 4-DABA (Scheme 1). Experimental Section Equimolar quantities of 3,5-DNBA, 4-DABA, and BIPY were dissolved in methanol at 50 °C and then the solution was cooled to
2771
Scheme 1. Selective Disruption of 3,5-DNBA Dimer by BIPY
15 °C by decreasing the temperature by 1 °C every 90 min. Deep red blocks were formed on cooling. X-ray diffraction data sets were collected at 100, 200, and 300 K on a Bruker Nonius Kappa CCD diffractometer equipped with an Oxford Cryosystems low temperature Cryostream device. The structure was solved using SHELXS21 and refined using SHELXL22 within the WinGX program suite.23 The complex crystallizes in the space group P1 with unit cell parameters of a=7.6431(3) A˚, b=8.6569(3) A˚, c=15.4854(6) A˚, R=79.479(2)°, β=82.732(3)°, γ=87.277(2)° at 100 K. The refinement used 374 parameters and gave R1=0.0387 for 3794 data with Fo > 2σ(F), Rw2 = 0.1096 for 4524 unique data. Full crystallographic data for all X-ray and neutron refinements are available within the CIF files as Supporting Information. Neutron diffraction data sets were collected at 40, 100, 200, and 300 K on SXD, the single crystal diffractometer at ISIS, U.K. using the time-of-flight Laue diffraction method. The unit cell was refined from the X-ray cell individually for each frame and an average was taken for the refinement of each temperature. Processing of the data was completed using SXD2001.24 The final refinement was done using SHELXL22 within the program WinGX23 with the initial atomic coordinates taken from the X-ray structure. All hydrogen positional and anisotropic thermal parameters have been fully refined. All Fourier difference maps for both the X-ray and neutron data were generated using MAPVIEW within WinGX.23
Results and Discussion The ternary complex forms in a 2:2:1 4-DABA/3,5-DNBA/ BIPY ratio. Within the complex, the 4-DABA molecules form hydrogen bonded homodimers between the carboxylic acid groups related to one another by an inversion center at the midpoint of the R22(8) ring as seen in the previously reported pure material and in the complex with 3,5-DNBA.16 The O 3 3 3 O distance is 2.637(1) A˚ at 100 K (X-ray) and is characteristic of a moderate strength hydrogen bond. The BIPY sits on an inversion center which relates the two pyridine rings and this is sandwiched by two 3,5-DNBA molecules. An O-H 3 3 3 N hydrogen bond of O 3 3 3 N length 2.547(2) A˚ is formed between these molecules at 100 K (X-ray) (Figure 1). This is characteristic of a short, strong O-H 3 3 3 N hydrogen bond. There is no proton transfer within these three molecule
2772
Crystal Growth & Design, Vol. 10, No. 6, 2010
Thomas et al.
Figure 1. The three-molecule hydrogen bonded block consisting of one BIPY and two 3,5-DNBA molecules (100 K, neutron).
Figure 2. The alternate layers of homodimer and three-molecule blocks showing the close contacts between the planes (neutron, 100 K).
blocks though the O-H distance shows a slight elongation (1.086(9) A˚, 100 K neutron) as would be expected for a short, strong hydrogen bond. There is a small rotation of ∼5° of the 3,5-DNBA molecule with respect to that of the plane of the BIPY at 100 K. The homodimers and three-molecule blocks are stacked alternately above one another and are connected by short C-H 3 3 3 O and C-H 3 3 3 π interactions between the methyl hydrogens of the 4-DABA and the 3,5-DNBA nitro group and BIPY, respectively. The H 3 3 3 O and H 3 3 3 π distances are 2.47(1) A˚ (cf. sum of van der Waals radii 2.72 A˚) and ∼2.73 A˚ at 100 K (neutron) (Figure 2). The non-hydrogen atom positions determined through both X-ray and neutron diffraction are consistent with one another. The thermal ellipsoid describing the hydrogen atom contained within the 3,5-DNBA-BIPY-3,5-DNBA three-molecule hydrogen bonded unit shows no elongation along the hydrogen bond at any temperature indicating a well-behaved, welllocalized hydrogen atom contained within this short, strong hydrogen bond. The thermal parameter increases in size in a manner consistent with all other atoms contained within the 3,5-DNBA-BIPY-3,5-DNBA unit for both the X-ray and neutron determined parameters. The anisotropic thermal parameters derived from the neutron data also show a welllocalized, single H atom position. The hydrogen atom remains associated with the carboxylic acid group at all temperatures studied. The hydrogen atom within the 4-DABA homodimer differs in behavior from the above. In this case, the neutron determined anisotropic thermal parameters show a significant elongation along the hydrogen bond at all temperatures except 40 K (Figure 3). This is indicative of either a hydrogen atom showing disorder in a double well potential with a secondary position being occupied as the temperature increases, or of a hydrogen atom in a flat asymmetric single minimum potential. The position of the centroid of this elongated nuclear density is also seen to move toward the center of the hydrogen bond with increasing temperature,
Figure 3. The evolution of the anisotropic displacement parameters of the hydrogen atom in the homodimer showing the welllocalized position at 40 K moving to an elongated central position at 300 K determined by neutron diffraction.
starting at O-H and H 3 3 3 O distances of 1.009(7) and 1.645(7) A˚, respectively, at 40 K and reaching a central position at 300 K with O-H and H 3 3 3 O distances of 1.29(3) and 1.33(3) A˚, respectively. It should be noted, however, that this is a consequence of an inadequate model for the H atom (see below); indeed this apparent shift, together with the extreme elongation of the hydrogen anisotropic thermal parameters are usually characteristic of the presence of a second site occupied by the hydrogen and hence of proton disorder. The O 3 3 3 O distance remains invariant throughout this temperature range. The X-ray data also shows an isotropic thermal parameter for this hydrogen atom, which is significantly enlarged compared to that of all other hydrogen atoms, including those not involved in hydrogen bonds. The isotropic thermal parameter is of comparable size to the librating H atoms of the methyl groups at 300 K. Such large isotropic thermal parameters of hydrogen atoms as determined by X-ray diffraction are also often seen in such disordered systems. The position of the maximum of the electron density is also seen to move toward a more central position as the temperature is increased (O-H, 1.10(3) A˚ at 100 K, 1.24(3) A˚ at 300 K), but again recall that the true situation is not described by a single hydrogen atom site. Imaging of the electron density from X-ray data or the nuclear density from neutron data through Fourier difference maps often provides more insight into the behavior of the hydrogen atoms involved in hydrogen bonds than the refinement of thermal parameters. In the case of this ternary complex, the X-ray and neutron data combined confirm the hydrogen atom is disordered and show a consistent trend in the behavior of the 4-DABA dimers (Figure 4). The 40 K neutron data set clearly shows a well-localized hydrogen atom within the 4-DABA dimer, with no evidence of a secondary site. This is consistent with the refined anisotropic displacement parameter. However, at 100 K, both the X-ray and neutron data show the emergence of a second peak of electron density or nuclear density closer to the secondary atom in the Fourier difference maps. The separation of these two positions (∼0.68 A˚, neutron)
Article
Crystal Growth & Design, Vol. 10, No. 6, 2010
2773
Figure 5. The homodimers at 100 and 200 K showing the refined split occupancy sites for the disordered proton derived from neutron data. A clear gap can be observed between the two thermal parameters in the 100 K model, but by 200 K this has reduced to a situation where they are touching.
Figure 4. Fourier difference maps showing the electron density (left) and nuclear density (right) of the disordered hydrogen atom contained within the homodimer.
is completely consistent with previous observations in carboxylic acid dimers.11 The size of this peak increases as the temperature is increased until at 300 K it becomes difficult to resolve two positions and the average hydrogen position refined for the maximum density takes a more central position. The elongation of the anisotropic thermal parameter for the hydrogen atom within the 4-DABA dimer derived from neutron data at temperatures above 100 K, combined with the Fourier difference maps from both the X-ray and neutron data, suggests a model where the hydrogen atom is disordered across two positions. It is possible to distinguish these two positions using the neutron data and to refine relative occupancies of the two sites to be 75:25, in a refinement model in which the isotropic displacement parameters are restrained to take the same value. At 200 K the increased thermal motion results in more overlap of the nuclear density corresponding to the two hydrogen atom positions, but it is still possible to refine the disorder to relative occupancies of 60:40 (Figure 5). Such a change in occupancy is as would be expected for a double minimum potential well, where the increased temperature allows an increasing occupancy of the minor position (representing a higher energy configuration of the dimer in the solid-state crystal lattice). The O-H distances take conventional values at both 100 and 200 K (0.99(2) A˚ (75%), 0.97(5) A˚ (25%) at 100 K; 0.94(3) A˚ (60%), 1.03(4) A˚ (40%) at 200 K). The separation of the two hydrogen sites is 0.68(5) A˚ at 100 K and 0.69(4) A˚ at 200 K, but the increased size of the thermal parameter can be seen to allow a slight overlap of the two thermal parameters at 200 K. However, by 300 K the
increased thermal motion combined with the inferior data quality and thus increased noise means that it is no longer possible to easily resolve the two positions. However, the central position of the anisotropic thermal parameter representing the maximum in the nuclear density for this atom strongly suggests an approximate 50:50 occupancy at this temperature, consistent with the observed trend. The observation of proton disorder in this system can again be related to the degree of pyramidalization of the nitrogen in the 4-DABA molecule. In the case of the binary cocrystal in which there is no proton disorder in the 4-DABA dimer, the methyl groups are significantly pushed out of the plane of the molecule;16 however, this effect is considerably smaller in both the pure material and in the ternary complex in both of which hydrogen disorder is found. The deviation of the position of the methyl groups from the mean plane of the molecule (ring-N-Cmethyl angle) can be seen to decrease (∼9° at 40 K; ∼3° at 300 K; they become more planar) as the temperature is increased and this would also favor the observed increase in the proportion of disorder present. This is consistent with there being a significant role played by a more delocalized electron density across the molecule as a whole in the cases in which disorder is present. The structures of the two disordered systems also have a broader similarity, both being layered in nature and with fewer close interactions pulling the methyl groups out of the plane. The binary complex is the only non-layered crystal structure in this group of compounds and this is also the only system where disorder is not observed within the 4-DABA homodimer. In fact, it shows disorder within the 3,5-DNBA homodimer instead.16 Conclusions This work has confirmed the significant effect that the crystal environment can have on the presence of disorder in the homodimers of 4-DABA housed within the ternary complex of this compound with 3,5-DNBA and BIPY. A systematic comparison of this system with the previously reported pure and binary 4-DABA systems suggests that where there are significant short contacts pulling the methyl groups of the 4-DABA out of the plane of the benzene ring, no disorder is observed. This is most likely due to the disruption of the delocalizing effect across the whole dimer as a result of this pyramidalization of the N atom. However, where there are no such interactions, such as in the case of the ternary complex presented here, neutron diffraction has definitively shown the presence of proton disorder with an increased occupancy of the higher energy state as the temperature is increased.
2774
Crystal Growth & Design, Vol. 10, No. 6, 2010
Visualization of the electron and nuclear densities through Fourier difference maps clearly show the emergence of a second site as the temperature is increased in a consistent manner, and it is possible to refine the occupancies of the two positions from the neutron data at 100 and 200 K; this allows the disorder to be quantified to be in the ratios 75:25 and 60:40 at 100 and 200 K, respectively. The increased thermal motion at 300 K and the accompanying reduction in the quality of the data make it impossible to resolve the two positions and hence derive the relative occupancies at this temperature. However, the almost central position is strongly suggestive of an approximate 50:50 occupancy at this temperature which would also be consistent with the trend found in the 100 and 200 K data. This systematic study demonstrates that the presence of disorder in 4-DABA dimers can be predicted from the heavy atom average structure of both the pure material and its complexes through observation of the degree of pyramidalization of the nitrogen atom in the 4-DABA molecule. Acknowledgment. A.A.K. is funded by EPSRC/STFC (EP/F021666). Access to neutron beamtime at ISIS was provided by STFC. Supporting Information Available: Crystallographic information files. This material is available free of charge via the Internet at http://pubs.acs.org.
References (1) Aaker€ oy, C. B.; Beatty, A. M. Comp. Coord. Chem. II 2003, 1, 679–688. (2) Braga, D.; Desiraju, G. R.; Miller, J. S.; Orpen, A. G.; Price, S. L. CrystEngComm 2002, 4, 500–509. (3) Cannon, A. S.; Warner, J. C. Cryst. Growth Des. 2002, 2, 255–257. (4) Walsh, R. D. B.; Bradner, M. W.; Fleischman, S.; Morales, L. A.; Moulton, B.; Rodrı´ guez-Hornedo, N.; Zaworotko, M. J. Chem. Commun. 2003, 186–187. (5) Zundel, G. Adv. Chem. Phys. 2000, 111, 1–217. (6) Etter, M. C. Acc. Chem. Res. 1990, 23, 120. (7) Aaker€ oy, C. B.; Beatty, A. M.; Helfrich, B. A. Angew. Chem., Int. Ed. 2001, 40 (17), 3240–3242. (8) Wilson, C. C. Cryst. Rev. 2007, 13, 143–198.
Thomas et al. (9) Wilson, C. C.; Shankland, K; Shankland, N. Z. Kristallogr. 2001, 216, 303–306. Steiner, T.; Wilson, C. C; Majerz, I. Chem. Commun. 2000, 1231–1232. Steiner, T.; Majerz, I; Wilson, C. C. Angew. Chem., Int. Ed. 2001, 40, 2651–2654. (10) Jeffrey, G. A. In An Introduction to Hydrogen Bonding; OUP: New York, 1997; Chapter 2, pp 11-16. (11) Wilson, C. C.; Shankland, N.; Florence, A. J. J. Chem. Soc. 1996, 92, 5051–5057. Wilson, C. C.; Shankland, N.; Florence, A. J. Chem. Phys. Lett. 1996, 253, 103–107. (12) Gilli, P.; Bertolasi, V.; Ferretti, V.; Gilli, G. J. Am. Chem. Soc. 1994, 116, 909–915. Gilli, P.; Bertolasi, V.; Ferretti, V.; Gilli, G. J. Am. Chem. Soc. 2000, 122, 10405–10417. Gilli, P.; Bertolasi, V.; Pretto, L.; Ferretti, V.; Gilli, G. J. Am. Chem. Soc. 2004, 126, 3845–3855. (13) Wilson, C. C.; Goeta, A. E. Agnew. Chem. Int. Ed. 2004, 116, 2147– 2151. Parkin, A.; Harte, S. M.; Goeta, A. E.; Wilson, C. C. New J. Chem. 2004, 28, 718–721. (14) Thomas, L. H.; Florence, A. J.; Wilson, C. C. New J. Chem. 2009, 33, 2486–2490. (15) Aakeroy, C. B.; Desper, J.; Urbina, J. F. Chem. Commun. 2005, 2820–2822. (16) Parkin, A.; Seaton, C. C.; Blagden, N.; Wilson, C. C. Cryst. Growth Des. 2007, 7, 531–534. (17) Ishida, H.; Kashino, S. Z. Naturforsch. 2002, 57a, 829–836. Childs, S. L.; Strahly, G. P.; Park, A. Mol. Pharmaceutics 2007, 4 (3), 323–338. (18) Vyas, M.; Sakore, T. D.; Biswas, A. B. Acta Crystallogr. 1978, B34, 1366–1368. Anulewicz, A.; Hafelinger, G.; Krygowski, T. M.; Regelmann, C.; Ritter, G. Z. Naturforsch. B, Chem. Sci. 1987, B42, 917–927. (19) Sharma, C. V. K.; Panneerselvam, K.; Pilati, T.; Desiraju, G. R. J. Chem. Soc., Perkin Trans. 2 1993, 2209–2216. (20) Aaker€ oy, C. B.; Desper, J.; Urbina, J. F. Chem. Commun. 2005, 2820–2822. Aaker€oy, C. B.; Desper, J.; Urbina, J. F. Cryst. Growth Des. 2005, 5 (3), 1283–1293. Aaker€oy, C. B.; Salmon, D. J. CrystEngComm 2005, 439–448. Aaker€oy, C. B.; Beatty, A. M.; Lorimer, K. Mol. Cryst. Liq. Cryst. 2006, 456, 163–174. Aaker€oy, C. B.; Desper, J.; Smith, M. M. Chem. Commun. 2007, 3936–3938. Bhogala, B. R.; Basavoju, S.; Nangia, A. CrystEngComm 2005, 551–562. Bhogala, B. R.; Basavoju, S.; Nangia, A. Cryst. Growth Des. 2005, 5 (5), 1683– 1686. Bhogala, B. R.; Nangia, A. New J. Chem. 2008, 32 (5), 800–807. Friscic, T.; Trask, A. V.; Jones, W.; Motherwell, W. D. S. Angew. Chem. Int Ed. 2006, 45 (45), 7546–7550. (21) Sheldrick, G. M. SHELXS97; University of G€ottingen: Germany, 1997. (22) Sheldrick, G. M. SHELXS97; University of G€ottingen: Germany, 1997. (23) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837–838. (24) Keen, D. A.; Gutmann, M. J.; Wilson, C. C. J. Appl. Crystallogr. 2006, 39, 714–722.
