Transferability of Multipole Charge Density Parameters for

Transferability of Multipole Charge Density Parameters for Supramolecular Synthons: A New Tool for Quantitative Crystal Engineering ... (G.R.D.) Fax: ...
2 downloads 0 Views 3MB Size
DOI: 10.1021/cg101540y

Transferability of Multipole Charge Density Parameters for Supramolecular Synthons: A New Tool for Quantitative Crystal Engineering

2011, Vol. 11 616–623

Venkatesha R. Hathwar, Tejender S. Thakur, Tayur N. Guru Row,* and Gautam R. Desiraju* Solid State and Structural Chemistry Unit, Indian Institute of Science, Bangalore 560 012, India Received November 20, 2010

ABSTRACT: The modularity of the supramolecular synthon is used to obtain transferability of charge density derived multipolar parameters for structural fragments, thus creating an opportunity to derive charge density maps for new compounds. On the basis of high resolution X-ray diffraction data obtained at 100 K for three compounds methoxybenzoic acid, acetanilide, and 4-methylbenzoic acid, multipole parameters for O-H 3 3 3 O carboxylic acid dimer and N-H 3 3 3 O amide infinite chain synthon fragments have been derived. The robustness associated with these supramolecular synthons has been used to model charge density derived multipolar parameters for 4-(acetylamino)benzoic acid and 4-methylacetanilide. The study provides pointers to the design and fabrication of a synthon library of high resolution X-ray diffraction data sets. It has been demonstrated that the derived charge density features can be exploited in both intra- and intermolecular space for any organic compound based on transferability of multipole parameters. The supramolecular synthon based fragments approach (SBFA) has been compared with experimental charge density data to check the reliability of use of this methodology for transferring charge density derived multipole parameters.

Introduction Experimental charge density studies on molecular crystals using high resolution X-ray diffraction data have provided significant inputs for the evaluation of both the strength and the directional preferences of strong and weak intermolecular interactions.1 Indeed, efforts have been focused in the past few years to evaluate finer features of experimental charge density distribution in nonbonded regions to obtain topological properties.2 These approaches are based on the multipole formalism suggested by Hansen and Coppens3 to derive charge density distributions from high resolution X-ray diffraction studies as well as from conventional ab initio methods. Bader’s “Atoms in Molecules” approach4 provides a convenient platform to define boundaries between atoms through their density based properties. Most of these experimental data sets are available at subatomic resolution, and hence the derived electrostatic properties can be derived with high precision and accuracy. However, force field calculations are still preferred when macromolecules such as proteins, polypeptides, and DNA fragments are involved. In recent years, several groups5-7 have been interested in examining the prospect of transferring the high resolution results obtained from charge density analysis on small molecules to proteins and other biomolecules. In these approaches, atoms recognized as being chemically similar based on connectivity and bonding considerations are extracted from molecular densities obtained for a library of small molecules. In the next step, it is demonstrated that charge density derived multipole parameters for these individual atoms can be transferred with a reasonable degree of accuracy to larger molecules like polypeptides and proteins. The success of these approaches have culminated in three *To whom correspondence should be addressed. (T.N.G.R.) Fax: (þ91) 80-23601310; tel: (þ91) 80-23932796; e-mail: [email protected]. (G.R.D.) Fax: (þ91) 80-23602306; tel: (þ91) 80-22933311; e-mail: [email protected]. pubs.acs.org/crystal

Published on Web 01/06/2011

major databases for ready reference and effective transfer to large macromolecules whose X-ray diffraction data sets can be obtained fairly routinely from both conventional and synchrotron sources to a resolution of nearly 1.5 to 1 A˚. The success of such transferability using charge density databases, for example, LSDB,5 INVARIOM,6 and ELMAM7 has been noteworthy. Even though charge density studies of small molecules are now being pursued actively, the time, effort, and cost of these experiments act as limiting factors to build large libraries in a short time. The other rate determining deterrent is the availability of high quality single crystals for high resolution X-ray diffraction experiments. Any effort that is able to establish some degree of modularity in the building up of electron density maps is therefore deemed to be of importance. The synthon is a modular unit whose relevance and significance goes beyond its utility as a static descriptor of crystal structure.8 It is a supramolecular structural fragment which is of kinetic significance and may even be invoked as an intermediate in crystallization.9 In this article, we examine the possibility of transferability of multipolar features to molecular and supramolecular fragments, more specifically synthons. Experimental and Computational Details Crystallization. The compounds chosen for the study, methoxybenzoic acid (1), acetanilide (2), 4-(acetylamino)benzoic acid (3), 4-methylbenzoic acid (4), and 4-methylacetanilide (5), were purchased from Sigma Aldrich and recrystallized from MeOH (2) and DMSO (1, 3, 4, and 5) by the slow evaporation method. Data Collection and Structure Refinement Details. Good quality single crystals of size ∼0.3 mm were selected under a polarizing microscope and affixed to Hampton Research Cryoloops using Paratone-N oil for data collection. The crystals were cooled to 100 K with a liquid nitrogen stream using an Oxford cryosystems N2 open flow cryostat. High resolution X-ray data sets up to (sin θ/λ)max = 1.08 A˚-1 with redundancy (∼9-13) and completeness ∼ 100% were collected on a Bruker Kappa Apex II CCD diffractometer using Mo-KR radiation at 100 K for compounds 1-4 (Table 1). Data r 2011 American Chemical Society

Article

Crystal Growth & Design, Vol. 11, No. 2, 2011

617

Table 1. Crystallographic and Structure Refinement Details for Compounds 1-5 high resolution charge density data sets 1 CCDC No. molecular formula formula weight crystal system space group a (A˚) b (A˚) c (A˚) R (°) β (°) γ (°) V (A˚3) Z Fcalc (g/cm3) F(000) μ (mm-1) T (K) λ (A˚) (sin θ/λ)max(A˚-1) reflns collected unique reflns completeness (%) redundancy Rint R1 (F2) wR2 (F2) goodness-of-fit reflns used [I > 3σ(I)] unique reflns R1 (F2) wR2 (F2) goodness-of-fit ΔFmin, max (e A˚-3)

2

785063 C8H8O3 152.14 monoclinic P21/n 3.8760(3) 10.9298(9) 16.620(1) 90 93.383(4) 90 702.9(1) 4 1.438 320 0.111 100(2) 0.71073 1.08 68055 7209 99.9 9.45 0.0325

785065 C8H9NO 135.16 orthorhombic Pbca 9.3710(9) 7.7868(8) 19.530(2) 90 90 90 1425.1(2) 8 1.260 576 0.084 100(2) 0.71073 1.08 96281 7524 100 12.80 0.0328 Spherical atom refinement 0.0363 0.0460 0.1139 0.1278 1.035 1.018 Multipole refinement 5483 5439 311 254 0.0215 0.0284 0.0411 0.0423 1.042 1.314 -0.125, 0.156 -0.131, 0.172

collection strategies were generated using the COSMO module of the Bruker software suite.10 The crystal-to-detector distance was fixed at 40 mm and the data collection was performed with a scan width of 0.5° per frame in all the cases. Cell refinement, data integration, and reduction were carried out using the SAINTPLUS program.10 A numerical absorption correction using the analytical method was employed after indexing the crystal faces in each case.10 Sorting, scaling, and merging of the collected data sets were carried out with the help of SORTAV program.11 The structure solution (direct methods) and refinement (spherical atom approximation based on F2) were performed with SHELXS9712 using the WinGX suite.13 Refinements and the analysis of the charge density data were carried out using the XD software.14 The structure factors were derived using the Su, Coppens, and Macchi wave functions.15 The multipolar nonspherical atom refinements were carried out with theP full-matrix least-squares method. The function minimized was w(|F0| K|Fc|)2, for all reflections with I/σ(I) > 3. The initial scale factor refinement of the diffraction data was performed up to full resolution range. The positional and anisotropic thermal displacement parameters of the non-hydrogen atoms were refined for reflections with sin θ/λ > 0.7 A˚-1. The X-H bond lengths were constrained to the corresponding neutron diffraction values (Car-H=1.077, Cmethyl-H =1.059, Namide-H=1.009, and Ocarboxyl-H=0.967 A˚).16 For nonhydrogen atoms, the scale, positional, and thermal displacement parameters, Pval, Plm, κ, and κ0 were allowed to refine in a stepwise manner, until the convergence was reached. The multipole expansion was truncated at the octupole level (l = 3) for non-hydrogen atoms. Appropriate local site symmetry constraints and chemical constraints were imposed on the multipole populations of all the non-hydrogen atoms during the multipolar refinement. Anisotropic thermal parameters for H-atoms were fixed to the values obtained from SHADE217 analysis. Only monopole, bond directed dipole (dz) and quadrupole (q3Z2-1) components were allowed to refine for the H-atoms. The quantitative analysis of the electronic structure

3

routine data sets 4

3

5

785062 C9H9NO3 179.17 triclinic P1 5.0299(2) 6.8310(3) 12.1945(5) 89.227(2) 80.846(2) 79.318(2) 406.43(3) 2 1.464 188 0.111 100(2) 0.71073 1.08 101574 8574 100 11.84 0.0345

785064 C8H8O2 136.14 triclinic P1 7.2832(3) 7.4255(3) 7.8100(3) 96.067(2) 108.353(2) 117.701(2) 338.60(2) 2 1.335 144 0.096 100(2) 0.71073 1.08 92429 7129 99.9 12.96 0.0202

785060 C9H9NO3 179.17 triclinic P1 5.026(1) 6.822(1) 12.196(3) 89.254(7) 80.804(7) 79.282(5) 405.5(2) 2 1.467 188 0.112 100(2) 0.71075 0.65 9042 1850 99.9 4.9 0.028

785061 C9H11NO 149.19 monoclinic P21/c 11.662(1) 9.4694(8) 7.3997(6) 90 106.559(7) 90 783.3(1) 4 1.265 320 0.083 100(2) 0.71075 0.65 8369 1798 99.7 4.7 0.039

0.0353 0.1021 1.096

0.0371 0.1103 1.015

0.0407 0.1177 1.053

0.0401 0.1008 1.080

7058 460 0.0224 0.0421 1.161 -0.127, 0.156

6044 186 0.0276 0.0601 1.9171 -0.142, 0.242

distribution was performed with the XDPROP module of XD software suite. Routine data sets were collected at 100 K for 3 and 5 (Table 1) on a Rigaku Mercury375R/M CCD (XtaLAB mini) diffractometer using graphite monochromated Mo-KR radiation, equipped with a Rigaku low temperature gas spray cooler. In these cases, data were processed with the Rigaku CrystalClear software.18 Structure solution and refinements were performed using SHELX97 of the WinGX suite. Computational Details. Single point periodic quantum mechanical calculations at the B3LYP/6-31G** level19 were carried out using CRYSTAL0620 with the geometry obtained from the experimental charge density refinement as input. The shrinking factors (IS1, IS2, and IS3) along with the reciprocal lattice vectors were set to 4 (30 k-points in irreducible Brillouin zone). The bielectronic Coulomb and exchange series values for the truncation parameter were set as ITOL1=ITOL2=ITOL3=ITOL4=8 and ITOL5=17 respectively for the CRYSTAL06 calculations. The level shifter was set to 0.7 hartree/cycle. The SCF convergence limit was chosen to ∼10-6 hartree. Molecular geometry and the atomic thermal displacement parameters for all atoms were set to zero in the calculation of a static model. Refinements and analysis of the theoretically obtained charge density data were performed using the XD software package using the same methodology as used for the experimental charge density modeling.

Results and Discussion Use of supramolecular synthons provides a chemically reasonable division of intermolecular interaction space around molecules in crystals. The present study provides a supramolecular synthon based fragments approach (SBFA) to synthesize the electronic distribution for target molecules in crystalline

618

Crystal Growth & Design, Vol. 11, No. 2, 2011

Hathwar et al.

Figure 1. Residual electron density maps for the multipolar refinement of the experimental charge density data of compounds 1, 2, and 4. Blue (solid lines), red (broken lines), and black (dotted lines) colors represent positive negative and zero contours, respectively. Contours are drawn at the intervals of (0.1 e A˚-3.

Scheme 1

solids. Scheme 1 shows two supramolecular synthons that we have used as test cases. We have chosen these specific examples keeping in mind two prominent interactions which occur in proteins, nucleic acids, and many other macromolecules. Suitable representative molecules having an O-H 3 3 3 O carboxylic acid dimer synthon and an N-H 3 3 3 O amide infinite chain synthon in their crystal structures21 were identified for building the synthon library and are given in Scheme 2. The choice was based on the availability of error free normal quality X-ray data and absence of any kind of disorder in the crystal structure. In set A (Scheme 2), compounds 1 and 2 provided experimental charge density for the synthons shown in Scheme 1 from which the charge density for compound 3 was obtained using the transferability criteria and benchmarked with separately measured experimental charge density. Similarly in set B (Scheme 2), the transferability was evaluated for compound 5 utilizing charge density multipolar features from compounds 2 and 4 and further validating the results from theoretical calculations. The detailed description of the procedures followed is given later in this section. High resolution X-ray data sets were collected for the chosen compounds 4-methoxybenzoic acid (1), acetanilide (2), and 4-methylbenzoic acid (4) at 100 K for building the library of supramolecular synthons (Table 1). The experimental charge densities for these compounds were modeled using the Hansen and Coppens multipolar formalism3 as described in the Experimental Section. The quality of multipole refinements was assessed from the Hirshfeld rigid bond test22 employed to all covalent bonds involving non-hydrogen atoms. The values of maximum differences of mean-square displacement amplitudes (DMSDA) are found to be 5(2)  10-4 A˚2 at O(3)-C(8) for 1, 5(2)  10-4 A˚2 at C(5)-C(6) for 2, and 7(2)  10-4 A˚2 at C(4)-C(8) for 4, respectively. The residual electron densities

Scheme 2. Compounds Chosen for the Study, with Atom Numberinga

a

Set A and set B are used for the check on the respective transferability of experimental and theoretical charge density multipole parameters.

calculated [with I > 3σ (I)] over the asymmetric unit are almost featureless (minimum and maximum density of -0.125 to 0.156 e A˚-3 for 1, -0.131 to 0.172 e A˚-3 for 2, and -0.142 to 0.242 e A˚-3 for 4) and are shown in Figure 1. Compound 4 shows relatively high residual densities around the carboxylic acid group due to the disorder in the H-atom positions. Above tests ensure the suitability of the experimental charge density derived multipole parameters for building of the synthon library. The static deformation density and Laplacian maps obtained from multipolar refinements of experimental and theoretical charge densities (obtained from CRYSTAL06 calculations) were also studied for quality checks (Supporting Information, Figures S1-S7). The electron density derived topological

Article

Crystal Growth & Design, Vol. 11, No. 2, 2011

619

Figure 2. Experimental charge density, CRYSTAL06 and SBFA derived (a) deformation (b) Laplacian density maps for the bonding regions of compound 3.

Figure 3. Experimental charge density, CRYSTAL06 and SBFA derived deformation density maps for the (a) O-H 3 3 3 O carboxylic acid dimer and (b) N-H 3 3 3 O amide infinite chain regions of compound 3.

properties at the bond critical points (Rij, Fb,32Fb, and ε) of all covalent bonds for compounds 1, 2, and 4 obtained from both experimental and theoretical calculations are provided in the Supporting Information (Tables S1, S2, and S4). These values were found to be in excellent agreement with one other, and this also ensures the quality of the modeled charge density parameters for all compounds. Supramolecular Synthon Based Fragments Approach (SBFA). The refined multipole parameters (Pval, Plm, κ, and κ0 ) obtained from the experimental charge density data of compounds 1 and 2 were used to build the supramolecular synthon based fragment library of O-H 3 3 3 O carboxylic acid dimer synthon and the N-H 3 3 3 O amide infinite chain synthons, respectively. Target molecules 4-(acetylamino)

benzoic acid (3) and 4-methylacetanilide (5) were identified to validate the transferability of both qualitative and quantitative features from the synthon library. The target molecules are divided into chemically reasonable molecular fragments based on their interaction environments. Special care was taken while dissecting the molecules into suitable molecular fragments such that atoms at the interfacial region are chemically similar both in terms of their intramolecular geometry and the intermolecular environment. The SBFA methodology involves routine data collection at 100 K, structure determination, and refinement of 3 and 5. This was followed by the direct transfer of experimental charge density derived multipole features from the SBFA library, 1 and 2 for compound 3 (set A) and from 2 and 4 for compound 5

620

Crystal Growth & Design, Vol. 11, No. 2, 2011

Hathwar et al.

Table 2. Topological Analysis of Charge Density Obtained from Experimental X-ray Data and CRYSTAL06 Calculations for Compound 3 and Comparisons with the SBFA Modeled Data bond O1-C7 O2-C7 O2-H2O O3-C8 N1-C4 N1-C8 N1-H1N C1-C2 C1-C6 C1-C7 C2-C3 C2-H2 C3-C4 C3-H3 C4-C5 C5-C6 C5-H5 C6-H6 C8-C9 C9-H9A C9-H9B C9-H9C

method

Rij (A˚)

Fb (e A˚-3)

r2Fb (e A˚-5)

d1 (A˚)

d2 (A˚)

λ1

λ2

λ3

ε

experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A experimental theory set A

1.2423 1.2425 1.2457 1.3068 1.3068 1.3117 0.9670 0.9670 0.9670 1.2331 1.2330 1.2345 1.4127 1.4128 1.4304 1.3610 1.3610 1.3647 1.0089 1.0088 1.0090 1.3981 1.3982 1.3850 1.4001 1.4001 1.3776 1.4823 1.4823 1.4704 1.3903 1.3904 1.3681 1.0772 1.0771 1.0770 1.4015 1.4016 1.3891 1.0769 1.0769 1.0774 1.3995 1.3996 1.3928 1.3926 1.3927 1.3797 1.0770 1.0769 1.0771 1.0768 1.0768 1.0771 1.5066 1.5067 1.5098 1.0597 1.0596 1.0592 1.0588 1.0589 1.0593 1.0592 1.0592 1.0592

2.840 2.748 2.860 2.455 2.301 2.435 1.980 2.280 2.121 2.944 2.780 2.838 1.958 1.926 1.928 2.314 2.205 2.267 2.282 2.235 2.181 2.157 2.068 2.209 2.116 2.052 2.238 1.804 1.841 1.921 2.117 2.117 2.273 1.909 1.894 2.014 2.123 2.095 2.166 1.931 1.881 2.072 2.138 2.095 2.094 2.107 2.095 2.256 1.903 1.927 1.931 1.962 1.931 2.103 1.751 1.700 1.673 1.839 1.936 1.881 1.809 1.927 1.856 1.798 1.909 2.000

-31.852 -31.921 -33.680 -26.123 -24.645 -22.189 -42.944 -34.478 -40.356 -31.276 -31.180 -34.183 -14.201 -13.174 -12.641 -21.007 -21.328 -20.785 -32.850 -29.894 -36.388 -18.219 -17.229 -17.403 -18.071 -16.390 -17.568 -13.291 -14.084 -13.776 -17.063 -17.797 -19.116 -19.732 -18.564 -17.491 -17.695 -18.064 -17.742 -18.737 -18.445 -20.205 -18.329 -17.937 -16.264 -17.626 -17.305 -18.664 -19.563 -20.172 -20.541 -20.262 -19.982 -17.468 -11.458 -11.071 -9.393 -16.547 -19.171 -15.685 -15.389 -19.700 -16.453 -15.004 -18.787 -17.721

0.7944 0.7912 0.7690 0.8136 0.8218 0.7824 0.7967 0.7494 0.7798 0.7856 0.7899 0.7550 0.8166 0.8053 0.8100 0.7854 0.7997 0.7757 0.7545 0.7430 0.7595 0.6914 0.7050 0.7052 0.6896 0.7106 0.6969 0.7179 0.7085 0.7143 0.6940 0.6883 0.6800 0.7087 0.6978 0.6802 0.6841 0.6762 0.6953 0.6972 0.6874 0.6562 0.7212 0.7378 0.7194 0.6904 0.7010 0.6920 0.7138 0.6981 0.6876 0.6987 0.6886 0.6742 0.7802 0.7953 0.7829 0.6915 0.6779 0.6495 0.6973 0.6766 0.6695 0.7056 0.6789 0.6555

0.4480 0.4512 0.4768 0.4932 0.4850 0.5294 0.1703 0.2176 0.1872 0.4475 0.4430 0.4796 0.5961 0.6075 0.6204 0.5756 0.5614 0.5890 0.2544 0.2658 0.2496 0.7067 0.6932 0.6798 0.7104 0.6895 0.6807 0.7644 0.7738 0.7561 0.6963 0.7020 0.6881 0.3685 0.3793 0.3969 0.7175 0.7254 0.6939 0.3797 0.3895 0.4212 0.6783 0.6618 0.6734 0.7022 0.6917 0.6878 0.3632 0.3788 0.3895 0.3781 0.3882 0.4029 0.7264 0.7114 0.7270 0.3682 0.3816 0.4097 0.3616 0.3822 0.3898 0.3536 0.3802 0.4036

-27.03 -24.13 -27.24 -21.69 -18.32 -21.76 -38.18 -37.03 -38.61 -27.47 -24.84 -27.13 -15.56 -14.26 -15.59 -19.80 -18.00 -20.45 -31.35 -30.46 -31.78 -16.70 -14.65 -17.13 -16.52 -14.51 -17.45 -13.94 -13.24 -15.12 -16.24 -15.18 -17.82 -18.78 -17.55 -18.98 -16.61 -15.30 -17.55 -18.54 -17.37 -19.49 -16.95 -15.18 -16.37 -16.31 -14.84 -17.36 -18.90 -18.10 -18.97 -19.09 -17.79 -19.52 -12.46 -11.21 -12.08 -17.47 -17.54 -17.07 -17.01 -17.56 -17.10 -17.05 -17.24 -18.62

-23.82 -22.27 -23.51 -18.87 -17.71 -18.66 -38.12 -36.59 -38.52 -23.26 -23.02 -24.20 -13.15 -12.83 -13.96 -15.64 -14.73 -16.70 -29.44 -28.59 -30.38 -13.65 -12.70 -14.27 -13.47 -12.21 -14.28 -11.29 -11.35 -12.44 -13.07 -12.75 -14.90 -17.81 -16.97 -18.00 -13.18 -12.75 -13.79 -17.51 -16.49 -18.42 -13.31 -12.64 -13.30 -13.26 -12.68 -15.22 -17.72 -17.46 -18.06 -18.00 -17.36 -18.61 -11.35 -10.44 -10.66 -16.40 -17.23 -15.74 -16.21 -17.07 -16.70 -16.30 -17.00 -17.15

18.99 14.47 17.08 14.43 11.39 18.23 33.35 39.14 36.77 19.45 16.68 17.14 14.51 13.91 16.92 14.43 11.40 16.36 27.94 29.16 25.76 12.14 10.12 14.00 11.92 10.33 14.16 11.94 10.51 13.78 12.25 10.14 13.60 16.86 15.96 19.50 12.10 9.99 13.59 17.32 15.41 17.70 11.93 9.89 13.40 11.94 10.21 13.91 17.06 15.38 16.48 16.83 15.16 20.66 12.35 10.57 13.35 17.33 15.59 17.12 17.83 14.92 17.35 18.35 15.45 18.05

0.13 0.08 0.16 0.15 0.03 0.17 0.00 0.01 0.00 0.18 0.08 0.12 0.18 0.11 0.12 0.27 0.22 0.22 0.06 0.07 0.05 0.22 0.15 0.20 0.23 0.19 0.22 0.23 0.17 0.22 0.24 0.19 0.20 0.05 0.03 0.05 0.26 0.20 0.27 0.06 0.05 0.06 0.27 0.20 0.23 0.23 0.17 0.14 0.07 0.04 0.05 0.06 0.02 0.05 0.10 0.07 0.13 0.06 0.02 0.08 0.05 0.03 0.02 0.05 0.01 0.09

(set B). As a matter of precaution, we ensured that multipole parameters used for modeling the O-H 3 3 3 O carboxylic acid dimer in 3 correspond to an ordered carboxyl group in the synthon library. This is important in the context of synthon modularity and transferability. Therefore, compound 1 was

preferred over compound 4 for modeling the density around the O-H 3 3 3 O carboxylic acid dimer synthon region of 3. Poor quality modeling was obtained for the O-H 3 3 3 O carboxylic acid dimer synthon region when the SBFA library of 2 and 4 was employed for modeling the charge density properties of 3

Article

synthon

Crystal Growth & Design, Vol. 11, No. 2, 2011

621

Table 3. Geometries of the O-H 3 3 3 O and N-H 3 3 3 O Interactions Obtained from Experimental X-ray Studies compound method X-H (A˚) H 3 3 3 A (A˚) X 3 3 3 A (A˚) X-H 3 3 3 A (°) Car-Car-C/N-O (°)

O-H 3 3 3 O

1 3

N-H 3 3 3 O

4 2 3 5

charge density charge density normal data charge density charge density charge density normal data normal data

0.94(1) 0.91(1) 0.91(3) 0.96(2) 0.84(1) 0.86(1) 0.89(2) 0.89(2)

1.69(1) 1.72(1) 1.72(3) 1.68(2) 2.09(1) 2.09(1) 2.07(2) 2.01(2)

2.628(1) 2.618(1) 2.620(2) 2.626(1) 2.913(1) 2.950(4) 2.954(2) 2.897(1)

-1.7(1) -4.9(1) -4.9(2) -3.8(4) -17.6(1) -39.2(1) -38.9(2) -18.0(2)

175(1) 169(1) 173(3) 169(2) 170(1) 176(1) 176(2) 175(1)

Table 4. Comparison of BCP Properties for the O-H 3 3 3 O and N-H 3 3 3 O Interactions Obtained from the SBFA Modeled and Experimental Charge Density Dataa compound

method

Rij (A˚)

Fb (e A˚-3)

3

experimental theoretical modeled ‘Set A’

1.6639 1.6644 1.6783

0.285 0.324 0.312

3

experimental theoretical modeled ‘Set A’ theoretical modeled ‘Set B’

1.9436 1.9436 1.9519 1.8884 1.8884

0.104 0.139 0.108 0.175 0.177

5 a

r2Fb (e A˚-5)

d1 (A˚)

d2 (A˚)

λ1

λ2

λ3

ε

1.1437 1.1058 1.1378

0.5202 0.5586 0.5405

-2.19 -2.39 -2.40

-2.16 -2.37 -2.36

6.86 7.77 7.04

0.01 0.01 0.02

N-H 3 3 3 O Infinite Chain 2.149 1.2704 1.669 1.2414 1.368 1.2781 2.109 1.2031 1.622 1.2199

0.6732 0.7022 0.6737 0.6853 0.6686

-0.59 -0.85 -0.73 -1.10 -1.19

-0.58 -0.84 -0.71 -1.08 -1.18

3.32 3.36 2.80 4.29 3.99

0.03 0.01 0.04 0.02 0.01

O-H 3 3 3 O Dimer 2.517 3.005 2.278

Theoretically obtained values (CRYSTAL06) for 3 and 5 are also given.

(set C, Supporting Information). The multipole parameters used for modeling the N-H 3 3 3 O amide infinite chain synthons for test cases 3 and 5 were taken directly from 2. The positional and thermal parameters of the non-hydrogen atoms were fixed to values obtained from routine quality data set. The H-atoms were fixed to neutron values and the anisotropic thermal parameters of H-atoms were obtained from a SHADE2 analysis. The multipole parameters along with κ and κ0 for all atoms in 3 were carefully chosen based on the molecular fragments of synthons. The scaling of the normal data set was carried out using the XD2006 package. The charge neutralization of transferred multipoles was achieved by releasing monopoles of all atoms in the target molecule. All other multipole parameters including κ and κ0 were kept fixed during refinements. Validation of SBFA Synthesized Electron Density with Experimental Charge Density. For benchmarking the SBFA methodology, a high resolution experimental charge density data set was also collected for 3 at 100 K (Table 1). The residual electron density map for the multipolar refinement of the experimental charge density data for 3 is provided in the Supporting Information. A comparison with theoretically obtained charge density was also carried out. Static deformation density and Laplacian plots obtained from the SBFA modeled density show a near perfect one-to-one correspondence with the experimental and theoretical charge density maps for both inter- and intramolecular regions (Figures 2 and 3). It is noteworthy that even subtle features in the deformation density map such as accumulation of electron density in the bonding regions and the polarization of lone pair density around the carbonyl oxygen atom were found to be modeled with reasonable accuracy (within 0.02-0.1 e A˚-3) with our approach. Reproduction of the bonding features at the interfacial regions of the synthon fragments (C2-C3 and C5-C6) ensures the quality of modeling. A comparative analysis of bond critical point properties of all covalent bonds was carried out in order to evaluate the quality of SBFA modeled charge density with experimental

Figure 4. CRYSTAL06 and SBFA derived deformation maps for the intramolecular bonding and the N-H 3 3 3 O amide infinite chain region of compound 5.

and theoretical values (Table 2). The derived values were closely comparable with values obtained from experimental charge density analysis. The intermolecular space around the target molecule, 3 can be divided into two supramolecular regions, an O-H 3 3 3 O carboxylic acid dimer synthon and an N-H 3 3 3 O amide infinite chain synthon. Related synthon geometries

622

Crystal Growth & Design, Vol. 11, No. 2, 2011

Hathwar et al.

Table 5. Analysis of Multipole Derived Atomic Charges Obtained from Experimental, Theoretical (CRYSTAL06), SBFA Modeling (sets A and B) for Compounds 3 and 5 3

5

atom

experimental

theoretical

direct

Set A

atom

theoretical

Set B

O(1) O(2) O(3) N(1) C(1) C(2) C(3) C(4) C(5) C(6) C(7) C(8) C(9) H(2O) H(1N) H(2) H(3) H(5) H(6) H(9A) H(9B) H(9C)

-0.4763 -0.4748 -0.4874 -0.2806 0.1256 -0.0514 -0.0099 0.0816 0.0282 -0.0056 0.0990 -0.0291 -0.1775 0.4322 0.2367 0.1558 0.0916 0.1633 0.1123 0.1434 0.1544 0.1682

-0.2065 -0.1780 -0.1632 -0.0279 -0.0275 -0.0550 -0.0629 -0.1102 -0.0453 -0.0371 -0.0732 -0.1026 -0.1292 0.2184 0.1877 0.1190 0.0993 0.1365 0.1096 0.0980 0.1272 0.1228

-0.2679 -0.3623 -0.0753 -0.0723 -0.1545 -0.0872 -0.0482 0.1551 0.0002 -0.1252 0.0864 0.2387 -0.0173 0.3715 0.3769 -0.0556 -0.0552 0.2064 -0.1923 0.0217 0.0860 -0.0297

-0.2617 -0.3290 -0.1125 0.0308 -0.1597 -0.0452 0.0036 0.1422 0.1105 -0.1074 0.0866 0.2457 0.1405 0.3441 0.2813 -0.0751 -0.1447 0.0944 -0.1767 -0.0404 0.0514 -0.0787

O(1) N(1) C(1) C(2) C(3) C(4) C(5) C(6) C(7) C(8) C(9) H(1N) H(2) H(3) H(5) H(6) H(8A) H(8B) H(8C) H(9A) H(9B) H(9C)

-0.3032 0.0235 -0.0577 -0.0604 -0.0481 -0.0963 -0.0365 0.0243 -0.0715 0.0573 -0.0319 0.1254 0.0614 0.0608 0.0344 0.0397 0.0239 0.0495 0.0646 0.0493 0.0326 0.0587

-0.3189 -0.0859 0.0478 0.1238 0.0448 -0.0834 -0.0515 0.1240 0.0286 0.1243 0.1003 0.0509 -0.0424 0.0527 -0.0006 -0.0766 0.0039 0.0279 0.0211 -0.0637 0.0188 -0.0458

for compounds 1-5 are given in Table 3. The charge density features for 3 were transferred from 1 for the O-H 3 3 3 O (carboxylic acid dimer) synthon region, and from 2 for the N-H 3 3 3 O (amide chain) synthon (Scheme 2, set A). Additionally, very good agreement between the topological properties at bond critical points (BCP) derived from the SBFA modeled charge density and from the experimental charge density at O-H 3 3 3 O and N-H 3 3 3 O interaction regions were obtained (Table 4). Notably, the BCP properties (Fb, r2Fb) for the O-H 3 3 3 O carboxylic acid dimer synthon are regenerated, for example, 0.312 and 2.278, based on the SBFA methodology. This may be compared with the experimental values of 0.285 and 2.517. A slight deviation in the r2Fb values for the N-H 3 3 3 O (amide infinite chain) synthon was seen for the SBFA modeled BCP properties. This can be rationalized on the basis of the difference of ∼17° in the CdC-N-C torsion angles between the N-H 3 3 3 O synthon fragments in 2 and 3 (Table 3). The SBFA approach was employed to synthesize the charge density of 4-methylacetanilide (5) in an independent manner using multipole parameters from the refined experimental charge density data of compounds 2 and 4 (Scheme 2, set B). A comparison was made with the theoretically modeled data obtained from the periodic CRYSTAL06 calculations at the B3LYP/6-31G** level of theory. This was done to assess the quality of modeling using our approach. Such comparison of the deformation density maps (Figure 4) and BCP properties assured us about the reliability of our SBFA methodology when compared with the theoretically obtained models from well-known databases such as LSDB5 (Table 4 and Supporting Information). The nearly similar conformations of the amide groups (CdC-N-C torsion angle; ∼1°) in 2 and 5 implies that a better fit for the modeled N-H 3 3 3 O synthon is expected in 5 rather than in 3 (Table 4). A general similarity in crystal packing in the starting set and target compounds is found to be desirable but not completely mandatory. For example, the amide torsions in 2 and 3 are different, but this does not prevent a derivation of a “synthetic” charge density map for 3 from the experimental multipolar parameters of 1 and 2.

The evaluation of atomic charges gives valuable input on the chemical and physical properties of the molecule. An estimation of the net atomic charge q = Nval - Pval (Pval is the valence population parameter and Nval is the number of valence electrons in a free neural atom) was carried out from the modeled charge density data. The values of atomic charges of compounds 3 and 5 are listed in Table 5. The monopole derived charges on oxygen atoms shows overall negative values in 3 and 5, but the magnitude of the charge was found to be underestimated in the modeled data. For example, the charge on carbonyl oxygen O1 is -0.4763 in the experimental charge density, whereas it is found to be -0.2617 in the SBFA methodology for 3. The N1 atom in the SBFA modeled data set for 3 shows overall positive character in contradiction to that obtained from experimental charge density analysis (þ0.0308 and -0.2806, respectively). Similar inconstancies in modeling atomic charges were observed for C- and H-atoms for SBFA modeled data of 3 and 5. However, these values were found relatively better than the values obtained from the theoretical data. Poor modeling of atomic charges at the interfacial regions of the synthon fragments (C2-C3 and C5-C6) was observed in both the test cases. This is in part expected and may be attributed to the charge neutralization methodology employed during transferability of multipole parameters from the corresponding synthon fragments (refinement by releasing monopole parameters only). Similar anomalies have been observed in the atomic charges obtained from LSDB approach (see Supporting Information, Table S9). A direct transfer of multipole parameters obtained from the experimental charge density data of 3 to routine data was also performed in order to estimate the upper limit of accuracy that can be achieved by SBFA method. Monopole charges obtained from the direct transfer shows better correlation between atomic charges both in terms of magnitude and signature in comparison to both theory and modeled data. Conclusions In summary, we have established that the SBFA method for molecular crystals is well suited to synthon based transferability of multipole parameters. Our approach provides a more

Article

Crystal Growth & Design, Vol. 11, No. 2, 2011

convenient means of obtaining essential topological features of electron density for both intra- and intermolecular space in any organic compound. Such an approach can be very helpful in cases where it is impossible to obtain good quality crystals and/ or a charge density quality data set, notwithstanding also the difficulties inherent in terms of the time required for such data collection. However, the SBFA approach shows discrepancies in modeling the monopole derived atomic charges. The building of an SBFA library for well-known synthons is currently in progress. A natural extension to larger molecules is also being investigated. The modularity of supramolecular synthons seems to extend to charge density properties of molecular crystals. Acknowledgment. V.R.H. and T.S.T. thank Indian Institute of Science for fellowships. T.N.G. thanks the DST for financial assistance through project SR/S1/IC-13/2008. G.R.D. thanks the DST for the award of a J. C. Bose fellowship. Supporting Information Available: (1) Tables S1-S4 Topological analysis of charge density obtained from experimental X-ray data and CRYSTAL06 calculations for compounds 1-4. Table S5 Topological analysis of charge density obtained CRYSTAL06 calculations for compound 5 and comparisons with the SBFA modeled data. Table S6 Analysis of multipole derived atomic charges obtained from experimental, theoretical (CRYSTAL06), LSDB, and SBFA modeling (sets A and C) for compound 3 and 5. (2) Figure S1 Residual electron density maps for the multipolar refinement of the (a) modeled CRYSTAL06 data of compounds 1-4 (b) experimental charge density data for 3. Figures S2-S8 Experimental charge density and CRYSTAL06 (a) deformation (b) Laplacian density maps for the bonding in intermolecular regions of compound 1-4. Figure S9 CRYSTAL06 and SBFA derived Laplacian density maps for the intramolecular bonding and N-H 3 3 3 O amide infinite chain region of compound 5. (3) Crystallographic information files (CIF). This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) (a) Gatti, C.; Saunders, V. R.; Roetti, C. J. Chem. Phys. 1994, 101, 10686–10696. (b) Koch, U.; Popelier, P. L. A. J. Phys. Chem. 1995, 99, 9747–9754. (c) Espinosa, E.; Molins, E.; Lecomte, C. Chem. Phys. Lett. 1998, 285, 170–173. (d) Coppens, P.; Abramov, Y.; Carducci, M.; Korjov, B.; Novozhilova, I.; Alhambra, C.; Pressprich, M. R. J. Am. Chem. Soc. 1999, 121, 2585–2593. (e) Spackman, M. A. Chem. Phys. Lett. 1999, 301, 425–429. (f) Flaig, R.; Koritsanszky, T.; Dittrich, B.; Wagner, A.; Luger, P. J. Am. Chem. Soc. 2002, 124, 3407–3417. (g) Munshi, P.; Guru Row, T. N. CrystEngComm 2005, 7, 608–611. (h) Coppens, P. X-ray Charge Densities and Chemical Bonding; Oxford University Press: Oxford, UK, 1997. (i) Koritsanszky, T. S.; Coppens, P. Chem. Rev. 2001, 101, 1583–1621. (2) (a) Mallinson, P. R.; Smith, G. T.; Wilson, C. C.; Grech, E.; Wozniak, K. J. Am. Chem. Soc. 2003, 125, 4259–4270. (b) Bui, T. T. T.; Dahaoui, S.; Lecomte, C.; Desiraju, G. R.; Espinosa, E. Angew. Chem., Int. Ed. 2009, 48, 3838–3841. (c) Munshi, P.; Guru Row, T. N. J. Phys. Chem. A 2005, 109, 659–672. (d) Hathwar, V. R.; Row, T. N. G. J. Phys. Chem. A 2010, 114, 13434-13441. (e) Mallinson, P. R.; Wozniak, K.; Wilson, C. C.; McCormac, K. L.; Yufit, D. S. J. Am. Chem. Soc. 1999, 121, 4640–4646. (f) Ellena, J.; Goeat, A. E.; Howard, J. A. K.; Punte, G. J. Phys. Chem. 2001, A105, 8696–8708 and references therein . (g) Oddershede, J.; Larsen, S. J. Phys. Chem. A 2004, 108, 1057–1063. (h) Spackman, M. A. Chem. Rev. 1992, 92, 1769–1797. (i) Zuo, J. M.; Kim, M.; O'Keeffe, M.; Spence, J. C. H. Nature 1999, 401, 49–52.

623

(3) Hansen, N. K.; Coppens, P. Acta Crystallogr., Sect. A 1978, 34, 909–921. (4) (a) Bader., R. F. W. Atoms in Molecules-A Quantum Theory; Clarendon: Oxford, 1990. (b) Bader, R. F. W. J. Phys. Chem. A 1998, 102, 7314–7323. (5) (a) Volkov, A.; Messerschmidt, M.; Coppens, P. Acta Crystallogr. Sect. D 2007, 63, 160–170. (b) Volkov, A.; Li, X.; Koritsanszky, T.; Coppens, P. J. Phys. Chem. A 2004, 108, 4283–4300. (c) Dominiak, P. M.; Volkov, A.; Dominiak, A. P.; Jarzembska, K. N.; Coppens, P. Acta Crystallogr. Sect. D 2009, 65, 485–499. (d) Dominiak, P. M.; Volkov, A.; Li, X.; Messerschmidt, M.; Coppens, P. J. Chem. Theory Comput. 2007, 3, 232–247. (6) (a) Dittrich, B.; Hubschle, C. B.; Luger, P.; Spackman, M. A. Acta Crystallogr. Sect. D 2006, 62, 1325–1335. (b) Hubschle, C. B.; Luger, P.; Dittrich, B. J. Appl. Crystallogr. 2007, 40, 623–627. (c) Dittrich, B.; Strumpel, M.; Spackman, M. A.; Koritsanszky, T. Acta Crystallogr., Sect. A 2006, 62, 217–223. (d) Dittrich, B.; Munshi, P.; Spackman, M. A. Crystallogr., Sect. B 2007, 63, 505–509. (e) Dittrich, B.; Hubschle, C. B.; Holstein, J. J.; Fabbiani, F. P. A. J. Appl. Crystallogr. 2009, 42, 110–1121. (7) (a) Zarychta, B.; Pichon-Pesme, V.; Guillot, B.; Lecomte, C.; Jelsch, C. Acta Crystallogr., Sect. A 2007, 63, 108–125. (b) Pichon-Pesme, V.; Lecomte, C.; Lachekar, H. J. Phys. Chem. 1995, 99, 6242–6250. (c) Lecomte, C.; Jelsch, C.; Guillot, B.; Lagoutte, A. J. Synchrotron Rad. 2008, 15, 202–203. (d) Jelsch, C.; Pichon-Pesme, V.; Lecomte, C.; Aubry, A. Acta Crystallogr., Sect. D 1998, 54, 1306–1318. (e) Guillot, B.; Jelsch, C.; Podjarny, A.; Lecomte, C. Acta Crystallogr., Sect. D 2008, 64, 567–588. (8) (a) Desiraju, G. R. Angew. Chem., Int. Ed. Engl. 1995, 34, 2311– 2327. (b) Desiraju, G. R. Angew. Chem., Int. Ed. Engl. 2007, 46, 8342–8356. (c) Desiraju, G. R. In Stimulating Concepts in Chemistry; V€ogtle, F.; Stoddart, J. F.; Shibasaki, M., Eds.; Wiley-VCH: Weinheim, 2000; Chapter 19, pp 293-306. (9) (a) Parveen, S.; Davey, R. J.; Dent, G.; Pritchard, R. G. Chem. Commun. 2005, 1531–1533. (b) Davey, R. J.; Dent, G.; Mughal, R. K.; Parveen, S. Cryst. Growth Des. 2006, 6, 1788–1796. (c) Mondal, R.; Howard, J. A. K.; Banerjee, R.; Desiraju, G. R. Cryst. Growth Des. 2006, 6, 2507–2516. (10) Bruker APEX2 (Version 1.0.22), BIS (Version 1.2.08), COSMO (Version 1.48) and SAINT (Version 7.06A), Bruker AXS Inc.: Madison, Wisconsin, USA, 2006. (11) Blessing, R. H. J. Appl. Crystallogr. 1997, 30, 421–426. (12) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112–122. (13) Farrugia, L. J. WinGX (Version 1.80.03). J. Appl. Crystallogr. 1999, 32, 837-838. (14) Volkov, A.; Macchi, P.; Farrugia, L. J.; Gatti, C.; Mallinson, P. R.; Richter, T.; Koritsanszky, T. S. XD2006, Rev. 5.34. University at Buffalo, State University of New York, NY, USA, 2006. (15) (a) Su, Z.; Coppens, P. Acta Crystallogr., Sect. A 1998, 54, 646. (b) Macchi, P.; Coppens, P. Acta Crystallogr., Sect. A 2001, 57, 656–662. (16) Allen, F. H. Acta Crystallogr. Sect. B 1986, 42, 515–522. (17) (a) Madsen, A. Ø. J. Appl. Crystallogr. 2006, 39, 757–758. (b) Munshi, P.; Madsen, A. Ø.; Spackman, M. A.; Larsen, S.; Destro, R. Acta Crystallogr., Sect. A 2008, 64, 465–475. (18) (a) CrystalClear 2.0; Rigaku Corporation: Tokyo, Japan. (b) Pflugrath, J. W. Acta Crystallogr. Sect. D 1999, 55, 1718–1725. (19) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (b) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789. (20) Dovesi, R.; Saunders, V. R.; Roetti, C.; Orlando, R.; ZicovichWilson, C. M.; Pascale, F.; Civalleri, B.; Doll, K.; Harrison, N. M.; Bush, I. J.; D’Arco, Ph.; Llunell, M. CRYSTAL06 User's Manual; University of Torino: Torino, 2006. (21) (a) Faustoa, R.; Matos-Bejab, A.; Paixaob, J. A. J. Mol. Struct. 1997, 435, 207–218. (b) Johnson, S. W.; Eckert, J.; Barthes, M.; McMullan, R. K.; Muller, M. J. Phys. Chem. 1995, 99, 16253–16260. (c) Kashino, S.; Matsushita, T.; Iwamoto, T.; Yamaguchi, K.; Haisa, M. Acta Crystalogr., Sect. C. 1986, 42, 457–462. (d) Takwale, M. G.; Pant, L. M. Acta Crystallogr., Sect. B. 1971, 27, 1152–1158. (e) Maeda, H.; Kamijo, N.; Fukui, K. Cryst. Struct. Commun. 1976, 5, 129–132. (22) Hirshfeld, F. L. Acta Crystallogr., Sect. A 1976, 32, 239–244.