J. Phys. Chem. A 2010, 114, 10097–10109
10097
Role of the Substituent Effect over the Squarate Oxocarbonic Ring: Spectroscopy, Crystal Structure, and Density Functional Theory Calculations of 1,2-Dianilinosquairane Carlos E. Silva,† He´lio F. Dos Santos,‡ Nivaldo L. Speziali,§ Renata Diniz,† and Luiz Fernando C. de Oliveira*,† Nu´cleo de Espectroscopia e Estrutura Molecular, Departamento de Quı´mica, Instituto de Cieˆncias Exatas, UniVersidade Federal de Juiz de Fora, Juiz de Fora, MG, 36036-900, Brazil, Nu´cleo de Estudos em Quı´mica Computacional, Departamento de Quı´mica, UniVersidade Federal de Juiz de Fora, MG, 36036-900, Brazil, and Laborato´rio de Cristalografia, Departamento de Fı´sica, UniVersidade Federal de Minas Gerais, MG, 31270-901, Brazil ReceiVed: June 10, 2010; ReVised Manuscript ReceiVed: August 5, 2010
This work presents the crystal structure and the investigation under a supramolecular perspective of a squaric acid derivative obtained from the replacement of the hydroxyl groups by anilines. The squaraine obtained (1,2-dianilinesquaraine) crystallizes in the Pbcn space group, in a unit cell with a ) 26.5911(8) Å, b ) 6.1445(10) Å, and c ) 7.5515(5) Å. The bond lengths in the oxocarbon ring, squarate-N and C-O bonds present the character of double bonds. Also the difference between the longer and shorter C-C bond in the four-membered ring (∆CC) is 0.0667 Å, showing a good degree of equalization of these bond lengths. The phenyl rings are slightly distorted in relation to the squarate ring, and the angle measured between the best plane fitted in each ring is 37.2(9)°. Each molecule is connected to the other through a hydrogen bond involving the N-H · · · O moieties, where the donor · · · acceptor distance is 2.826(1) Å, forming ribbons in a unidimensional arrangement C(5)R22(10) along the b axis. These structures are mutually connected by π-stacking interactions extending the supramolecular structure in a two-dimensional fashion. Besides, an interesting crossed structure can be easily identified in the formed sheets that are built through the C-H/π interactions. DFT calculations at the B3LYP/6-311++G(d,p) level of theory show an approximately planar molecular structure for the isolated molecule. However, when a dimer model built from hydrogen bonds is considered, the optimized structure presents considerable torsion between the phenyl and squarate rings, as observed in the experimental data. The electronic spectrum shows a strong absorption band at 341 nm that is red-shifted compared to the squarate maximum absorption (290 nm), indicating a more effective electronic delocalization. The most characteristic vibrational modes of the oxocarbon species were used as spectroscopic probe to understand how the substituent groups affect the oxocarbon moiety and, consequently, the vibrational spectra. The analysis shows that the modes associated with the C-Cox bonds are the most affected. Also the character of the double bond of squarate-N and the single bond for the phenyl-N are easily identified. In a general form, the calculated vibrational modes of the dimer model were in better accordance with the experimental data, mainly when the mode has a contribution from the acceptor molecule in the intermolecular interaction. 1. Introduction Oxocarbons are cyclic planar organic ions formed only by carbon and oxygen atoms, resulting in a high molecular symmetry (Dnh) and electronic delocalization. These characteristics have been the object of research involving such a class of molecules.1-4 They occur as rings composed of three, four, five, and six carbon atoms, which are denoted as deltate, squarate, croconate, and rodizonate ions, respectively.3-5 The squarate [(C4O4)2-] and croconate [(C5O5)2-] ions are by far the most studied compounds, mainly due to their chemical characteristics, which allow the synthesis of a large variety of derivatives. When these derivatives maintain the main characteristics of the oxocarbon compounds, they are called pseudooxocarbons.1 Among these compounds, the amino-squarate * Corresponding author. E-mail:
[email protected]. Telephone (fax): 55 32 32293310. † Nu´cleo de Espectroscopia e Estrutura Molecular, Departamento de Quı´mica, Instituto de Cieˆncias Exatas, Universidade Federal de Juiz de Fora. ‡ Nu´cleo de Estudos em Quı´mica Computacional, Departamento de Quı´mica, Universidade Federal de Juiz de Fora. § Universidade Federal de Minas Gerais.
derivatives are the most widely studied, being called in the literature as squaraines.1 Both oxocarbons and their derivatives have been extensively studied by our research group through molecular spectroscopy and single crystal X-ray diffraction.6-10 These investigations have shown a straight correlation among structural characteristics present in the solid state and the information supplied from spectroscopic data. More specifically, vibrational spectroscopy has been used for this propose where specific vibrational modes of these species have been shown to be very sensitive to some different interactions in the solid phase. In this way, the goal in this research field is the evolution from molecular aspects to the frontiers beyond the molecule. This work deals with the same subject; however, we are introducing a new feature: the analysis of the effects over the molecular structure and in the crystal packing, produced by the symmetrical replacement of the hydroxyl groups of squaric acid by an aromatic amine. This type of squaraine presents sharp and strong electronic absorption bands, whose positions are strongly influenced by the substituent groups, allowing us to obtain compounds that absorb in a wide range of the optical
10.1021/jp105346h 2010 American Chemical Society Published on Web 08/24/2010
10098
J. Phys. Chem. A, Vol. 114, No. 37, 2010
spectra. Generally, they present multiple emission bands with small Stokes shifts and high quantum fields in the fluorescence spectra.11 Therefore, the electronic features of these compounds are the prime motivation to investigate such molecules, producing chemicals with interesting properties for technological applications, as, for instance, substrates for photovoltaic investigations,12 photoconductors and photoreceptors,13 xerographic sensitizers, and mainly nonlinear optics research.11 Despite this, the spectroscopic analysis of such a family is still scarce and most of the works make use of infrared techniques.1 Cano and co-workers have investigated the Raman spectra of squaraines inside zeolites, concluding that the comparison of the vibrational spectra gives a good indication of the purity of organic guests.14 However, in that paper the authors do not present a discussion about the band assignments. Lopes et al. described the resonance Raman profile of two squaraines in organic solutions, showing that the chemical equilibrium in diluted solutions is more important in the analysis of the chromophoric moiety than the aggregation phenomenon, as believed in the literature.15 In the Supporting Information is a scheme containing the canonical structures of ACSQ. Infrared spectra of cis- and transdimethylsquaraines were described by Lunelli and co-workers,16 investigating the water exchange between them and the partial pressure of water over the hydrate compounds; however, once again, no specific vibrational assignments were done by the authors. An asymmetrical squaraine has been characterized by single crystal X-ray diffraction and vibrational spectroscopy, where a dimer formed by hydrogen bonds and a zigzag motif have been identified. That was the first time the Raman spectra and crystal structure of an aniline squaraine were shown and discussed.17 The present work aims to report a crystallographic, spectroscopic, and theoretical study of a product obtained from the reaction made through the replacement of hydroxyl groups in squaric acid (1,2-dihydroxycyclobutene-3,4-dione) by one aniline moiety. The resulting compound is the cis isomer (3,4bis(phenylamino)cyclobut-3-ene-1,2-dione) denoted here as ACSQ. The crystal structure of this compound and the supramolecular aspects related to the design of the solid, the identification of the main intermolecular interactions that contribute for the supramolecular assembly, and also their implications in the vibrational spectra are given. The main vibrational modes of the oxocarbon moiety are discussed as probes to understand the most significant modifications caused by the aniline groups directly attached to the squarate unity. Therefore, a concise assignment of the main vibrational modes is done and compared with the normal-mode analysis for the ACSQ, computed at the density functional theory (DFT) level of calculations. Additionally, an analysis of the electronic spectrum was also carried out through time dependent (TD)DFT calculations. 2. Experimental Section Synthetic Procedure. All reactants and solvents used were analytical grade. Aniline (Sigma-Aldrich) was previously distilled and stocked in a nitrogen saturated flask. The compounds were synthesized according to the proposition of Neuse and Green,18 however with a few modifications. To 5 mL of squaric acid in DMF solution 1 mol · L-1 was added 0.91 mL of aniline, and this mixture was maintained under reflux for 90 min. During this time the yellow solid formed and to this mixture was added 20 mL of frozen water. The solid formed was removed by filtration and washed exhaustively with water at 50 °C, acetonitrile, hot methanol (60 °C), acetone, and dichloromethane
Silva et al. to remove secondary products. To the residual solid was added hot DMF (50 °C), and the mixture was stirred for about 30 min and then filtrated. The addition of 150 mL of water to the DMF filtrate produced a precipitate of the 1,2-isomer, which, after 12 h of standing at 8 °C, was filtered off and dried, yielding 0.201 g (17%) of a yellow powder made of pure ACSQ. Suitable crystals for X-ray diffraction were obtained from a different method where the aniline was added in molar excess to the acetonitrile solution containing ethyl squarate (Sigma-Aldrich). This mixture was maintained under a nitrogen atmosphere and stirred for about 3 h. Afterward, the reactant mixture was left to stand, and crystals formed from the solution after 5 days. The crystals presented Raman and infrared spectra identical with the data for the obtained powder sample. Spectroscopic Measurements. Fourier-transform Raman spectroscopy of powder solid samples was carried out using a Bruker RFS 100 instrument, equipped with a Nd3+/YAG laser operating at 1064 nm and a CCD detector cooled with liquid N2. Good signal-to-noise ratios were obtained from 512 scans accumulated over a period of about 30 min and 30 mW of laser power, using 4 cm-1 as spectral resolution. All spectra were obtained at least twice to show the reproducibility, and no changes in the band positions and intensities were observed. The infrared spectra were obtained in a Bomem MB-102 spectrometer fitted with a CsI beam splitter, using KBr disks and a spectral resolution of 4 cm-1. A good signal-to-noise ratio was obtained from the accumulation of 128 scans. Also, the infrared spectra were obtained from emulsions in mineral oil from 600 to 250 cm-1 and fluorolube from 4000 to 1350 cm-1, which were pressed into KRS-5 plates. UV and optical spectra were obtained from DMSO solution of ACSQ compound in a quartz cuvette using a Shimadzu UV-1601PC spectrometer with 2.0 nm of spectral resolution and scan rate of 200 nm min-1. X-ray Crystallography. A suitable crystal was mounted on an Oxford GEMINI A-Ultra with a CCD diffractometer with Mo KR (λ ) 0.710 73 Å) at room temperature and 150 K. Data collection, reduction, and cell refinement were performed by the CrysAlis RED program.19 The structures were solved and refined using SHELXL-97.20 A multiscan absorption correction was applied.21 The structures were drawn by ORTEP-3 for Windows22 and Mercury23 programs. Hydrogen atoms were located from Fourier difference maps. Anisotropic displacement parameters were assigned to all non-hydrogen atoms. The crystallographic data were deposited at Cambridge Crystallographic Data Center on CCDC 778399 and 778400. Theoretical Calculations. The geometry of the monomer and dimer models were fully optimized at the DFT level using the B3LYP functional (Becke’s gradient-corrected exchange correlation in conjunction with the Lee-Yang-Parr correlation functional with the three hybrid parameters). The standard Pople split-valence basis set with inclusion of polarization and diffuse functions [6-311++G(d,p)] was used. The harmonic frequencies, Raman activities, and infrared intensities have been calculated with the same basis set and level of theory. The vertical electronic transitions were computed with the TD-DFT at the B3LYP/6-311++G(d,p) level. The solvent effect was included by the polarized continuum model (PCM), where the solute cavity was described by UA0 radii method and the solvent (DMSO) was represented by an isotropic medium with dielectric constant set to ε ) 46.7. The program used to perform the calculations was the Gaussian 03 release C.02.24 3. Results and Discussion 3.1. Molecular and Crystal Structure. The crystallographic data of ACSQ were measured at two temperatures, as mentioned
1,2-Dianilinosquaraine
J. Phys. Chem. A, Vol. 114, No. 37, 2010 10099
TABLE 1: Crystallographic Data formula formula weight/g mol-1 crystal system space group a/Å b/Å c/Å V/Å3 Z µ(Mo KR)/Å no. of reflections no. of unique reflections no. of observed reflections, F02 > 2σ(F02) no. of refined parameters R wR S rms peak (e- Å-3)
150 K
298.15 K
C16H12N2O2 264.28 orthorhombic Pbcn 26.5911(53) 6.1445(12) 7.5515(15) 1233.8(4) 4 0.710 73 27 829 1681 1399 95 0.0399 0.1029 1.045 0.049
C16H12N2O2 264.28 orthorhombic Pbcn 26.7659(8) 6.158 20(10) 7.6556(2) 1261.87(5) 4 27 389 1702 1162 95 0.0420 0.0983 1.085 0.037
TABLE 2: Crystal and Fully Optimized DFT Selected Geometrical Parameters and Hydrogen Bonds for the ACSQ Molecule dimer experiment monomer molecule A molecule B N1-C2 N1-C3 O1-C1 C1-C2 C1-C1i C2-C2i
Bond 1.3316(14) 1.4249(14) 1.2260(14) 1.4730(15) 1.4884(16) 1.4207(15)
Length/Å 1.353 1.416 1.206 1.492 1.532 1.403
1.346 1.420 1.213 1.482 1.513 1.412
1.351 1.419 1.208 1.492 1.529 1.406
C2-N1-C3 O1-C1-C2 O1-C1-C1i C2-C1-C1i N1-C2-C2i N1-C2-C1 C2-C2i-C1i
Bond 125.49(9) 137.18(11) 134.18(7) 88.61(6) 133.00(6) 135.70(10) 91.25(8)
Angle/deg 130.73 137.30 135.16 87.54 129.21 138.34 92.46
130.16 137.08 134.15 88.04 129.75 138.27 91.95
126.11 136.48 135.88 88.63 132.91 137.67 92.36
C3-N1-C2-C1 C3-N1-C2-C2 C2-N1-C3-C4 C2-N1-C3-C8
Dihedral -14.5(2) 168.70(11) 153.51(11) -29.97(16)
-8.965 173.57 169.67 -11.76
-10.61 173.10 148.39 -33.14
N-H/Å H · · · O/Å N-H-O/deg N · · · O/Å
Hydrogen Bond 0.885(18) 1.011 2.006(18) 153.6(15) 2.8262(13)
1.012
1.017 2.110 155.20 3.063
Angle/deg -3.96 177.18 180.00 -2.35
in the Experimental Section. The treatment of both sets of X-ray diffraction data led to the same structure, revealing no molecular or crystalline changing upon cooling. The difference observed was the best quality in the measurement data carried out at 150 K, and due to this fact, the structural analysis and results reported here are obtained from this one. Crystal data for both measurements are listed in Table 1, and some geometrical parameters are shown in Table 2. Figure 1 displays the crystal structure of ACSQ molecule, and the crystal packing is shown in Figures 2 and 3. The four-membered ring is formed by two crystallographic independent carbon atoms and two more atoms generated by symmetry (C1-C2-C1i-C2i, symmetry code i ) -x, y, 1/2 - z), which present a small root mean square (rms) deviation (0.0276) for the best fitted plane containing this atoms, showing good planarity. However, it is slightly distorted when compared to the triclinic sodium squarate where this value is 0.007.25 The
Figure 1. ORTEP view of ACSQ measured at 150 K. Symmetry key: i ) -x, y, 1/2 - z.
average C-C bond distance of the squaric moiety is 1.461 Å and the higher CC bond length difference (∆CC) is 0.0667 Å, whereas these values for the sodium squarate (Na2C4O4 · 3H2O) are 1.470 and 0.0247 Å,25 respectively. In spite of the longer average CC bond of the squarate salt, the bond distances in the four-membered ring are more equalized in this molecule than in the ACSQ compound. However, the longest C-C bond in the ACSQ compound [1.4884(16) Å] is shorter than a formal single bond like in the glyoxal (1.525 Å)26 and the shortest CC bond is longer than a formal double bond like in 3,4dimethylenecyclobutene (1.357 Å).27 All these observations indicate that the electronic delocalization is maintained in the oxocarbon moiety after the hydroxyl replacement by the aniline groups. The C-O bonds have a double bond character [1.226(1) Å], and this is slightly shorter than the average in Na2C4O4 · 3H2O (1.259 Å). Also an imine-like behavior is observed for the Cox-N bonds (subscript “ox” stands for the squaric moiety) where the length is 1.332(1) Å, which is fairly close to that observed for an anilinosquaraine reported with 1.339(2) Å.17 On the other hand, the Cph-N bond involving the phenyl ring [1.425(1) Å] is longer than Cox-N and has characteristics of a single bond. Therefore, a torsional degree of freedom can be expected around the Cph-N bond. In effect, compared with the observed structure for (2-dimethylamino-4-anilino)squaraine (ADTCH3),17 the overall structure is not planar with considerable torsion between the four-membered and the phenyl rings of 37.2(9)°. The angle between the two phenyl rings is stressed and measured at 72.3(5)° showing that the molecular shape is quite distorted. In a first view, the removal of hydroxyl groups seems to lead to a reduction in the hydrogen bond possibilities due to the steric hindrance added by the phenyl rings of the aniline groups. However, as aforementioned, the Cph-N bonds involving the phenyl rings [1.425(1) Å] have characteristics of a single bond, allowing a torsional degree of freedom around this bond. Therefore, despite the presence of bulky groups, the possibility of a hydrogen bond involving the amino group should not be discarded. In effect, the nonplanar molecular conformation reduces the steric effect when compared to the effect of the totally planar conformation, which allows each squaraine molecule to attach to another through a hydrogen bond between the amino and carbonyl groups. It can be classified as being medium to weak intensity in accordance with geometric criteria,28 since the donor-acceptor distance is 2.826(1) Å and the angle ∠CHO is 153.7(2)°. These interactions form ribbons in a unidimensional arrangement, for which the graph set29 is C(5)R22(10), along the a crystallographic axis, as can be seen in Figure 2a. These ribbons overlap each other, forming a stacked fashion. In this arrangement, two different structural motifs are observed for the ACSQ molecule, where each one in the adjacent ribbons is displayed in opposite directions. An inspec-
10100
J. Phys. Chem. A, Vol. 114, No. 37, 2010
Silva et al.
Figure 2. Crystal packing formed by a) the NHO hydrogen bond and π-stacking interactions in the parallel layers to the (011) plane (A), and b) the interesting wave arrangement observed along the axis a.
Figure 3. C-H/p interactions and the X-like arrangement formed through this interactions.
tion of π-stacking interactions shows that the centroid-centroid and interplanar distances are less than 4.00 Å (3.77 and 3.27 Å, respectively), indicating that this compound presents an effective π-stacking interaction between the four-membered rings, responding to the connection between the ribbons. Therefore, in a supramolecular view, the molecular assembly allows the hydrogen bonds, producing a unidimensional construction, and the π-stacking interactions extend this bidimensionaly. Comparing to similar compounds,10,17,25,30-33 ACSQ presents the interplanar distance close to that of the potassium squarate salt (3.27 Å)31 and is the smallest among the squaraine molecules32,33 (3.32 and 3.35 Å) with the exception of the (2dimethylamino-4-anilino)squaraine (3.24 Å).17 However, its centroid-centroid distance is quite close to that observed for
the 1,2-bis(dimethylamino)squaraine (3.74 Å)16 and the ammonium squarate (3.71 Å),9 which present higher values for this parameter. The horizontal translation of adjacent rings of ACSQ squaraine has a high value (1.82 Å), suggesting that the rings are not so effectively overlapped as in the others. Therefore, despite the undoubted presence of the π-stacking interactions, in the ACSQ molecule this interaction seems to be weak. As can be seen in Figure 2b, an interesting structural arrangement is formed by two molecular motifs along the b crystallographic axis view where a wavy profile is observed parallel to the plane formed by the axis a and c. This same construction can be found in other squaraines as, for instance, the ADTCH3. The angle between the molecules in the wave building was measured by fitting a mean plane in the squarate
1,2-Dianilinosquaraine ring, since in the ACSQ the rings are not coplanar. The ADTCH3 and the ACSQ have the most significant angles, 67° and 57°, respectively, whereas the two aliphatic squaraines studied by Lunelli et al.16 have a softer waving with 17° and 36° for the trans and cis isomers, respectively. Recently, several investigations have shown the importance of the CH-π interactions in the crystal structure, including selective enantiomeric enrichment,34-37 properties of coordination complexes,38-41 and mainly the folding and properties of molecular biological systems.42-45 This weak electrostatic interaction can be seen in the ACSQ structure and also can be contributing to the hydrogen bonds for the torsion between the squarate and phenyl ring. Each phenyl ring is surrounded by another four rings (Figure 3), arranged in space with an angle of 72.34°, which is the same as found between the phenyl rings in the molecular structure (72.34°). The atoms C5 and C8 have the smallest carbon-centroid distance of 3.606 and 3.632 Å, respectively, whereas for the C4 and C7 this distance is longer and close to 3.96 Å. The average H-centroid distances measured were 2.862 Å for H5 and H8 and 3.534 Å for other ones. Takahashi et al. analyzed a large number of chemical systems deposited in the Cambridge Structural Database (CSD) and concluded that C-H/π intermolecular interactions involving aromatic systems present average distances (Datm) of 2.91 ( 0.12 Å.46 The limiting distance suggested in the literature has been fixed as shorter than 3.05 Å [Dmax ) 2.9 (1.2 Å of H atom plus 1.7 Å for sp2 C) × 1.05]46 and using this value as a cutoff, only H4/C3 and H7/C6 distances are smaller than this (2.858 and 2.868 Å, respectively). Furthermore, Datm is quite shorter than the mean value related in the literature, and these results support the arguments that the CH-π interactions are present in the ACSQ crystal structure. Lately, after describing the most relevant interactions and molecular features present in the crystal, we have been able to trace a general profile of the molecular and supramolecular structure of the ACSQ and the relations between them. Regarding the solid design, each molecule is the center of an X-like assembly that is extended by CH-π interactions in parallel layers to the (011) plane, as can be seen in Figure 3. In the same manner, the π-stacking and N-H · · · O hydrogen bond lengthen the structure in a 2D supramolecular arrangement parallel to the plane formed by the c and b crystallographic axes, but here each interaction grows in ribbons, which are displayed orthogonally among themselves. The set of these interactions form overlapped sheets, where each ACSQ molecule along sheets is organized one by one, generating the wavy profile observed along the a axis. This interesting structural design is a result of the molecular features, as for instance the torsion around the dihedral formed by the squarate and phenyl ring planes. The presence of good acceptors and donors in the molecule, such as carbonyl and phenylamino moieties, makes possible the hydrogen bond interaction, and the presence of cyclic groups rich in π-electrons, such as squarate and phenyl rings, allow the π-stacking and CH-π interactions. The calculated molecular geometry parameters of ACSQ are compared with the available experimental data in Table 2. However, as mentioned above, each ACSQ molecule is connected through hydrogen bonds in the solid state; therefore, in an attempt to rationalize and reproduce the effects caused by this interaction in the molecular structure, a dimeric model (Figure 4) has been proposed. The molecules that constitute the dimer were denominated as molecule A, which acts exclusively as a receptor in the hydrogen bonds, and molecule B, which in turn acts as a donor in the intermolecular interaction. The
J. Phys. Chem. A, Vol. 114, No. 37, 2010 10101
Figure 4. Optimized structure of the dimer model formed by hydrogen bonds. The letters (A) and (B) are related to the acceptor and donor species, respectively.
symmetry found after computing the molecular geometry optimization was C2 group for both models, and therefore, the geometric parameters related by the C2 axis symmetry operation were omitted to avoid redundancy. This observation is in good accordance with the molecular structure determined by X-ray diffraction, where half of each molecule is generated through the 21 symmetry operations. The absence of the imaginary frequencies or negative eigenvalues of the second-derivative matrix confirmed that the stationary points correspond to minima on the potential energy hypersurfaces for both the monomer and dimer. The N1-C2, C1-C2, and C1-C1i bonds have the length overestimated relative to the experimental data with the most significant differences found for the last bond. On the other hand, the N1-C3, O1-C2, and C2-C2i bond distances are underestimated in relation to the experimental values. Comparing the bond lengths of the experiment, monomer, and molecules A and B, listed in Table 2, it can be observed that in molecule A these parameters tend toward the measured by X-ray diffraction, whereas in molecule B these converge to the monomer model. These observations show that the squarate moiety is strongly perturbed by the hydrogen bonds. The NH bond is longer for molecule B than in the monomer and molecule A, showing that for this chemical bond the donor is more sensitive to the intermolecular interaction. Such changes in the bond length in comparison to that for the monomer model also suggest changes in the bond order and, therefore, in the dispersion and reorganization of the electronic density over the oxocarbon ring. As a consequence, when the ChelpG fitted atomic charges of the three molecules (monomer, molecule A, and molecule B) are compared, it is observed that in the dimer the absolute values of the atomic charges decrease in the four-membered moiety. The exception is the C2 atom, where this one increases slightly from 0.09 to 0.12 (values in atomic units). As for the bond distance, the changes in the charges’ distribution are more pronounced in molecule A, where the C1 and O1 atoms are the most affected. In the former an increase in the charge density upon dimerization was found (from 0.36 to 0.28 au), whereas in the latter (O1) a decreasing in the electronic density (from -0.50 to -0.41 au) was predicted. The sensitivity of the squarate ring for the dipolar interactions can be understood as a result of the higher molecular polarizability of the ACSQ (35.76 Å3 for the isotropic polarizability), which is higher than that obtained for p-nitroaniline (13.70 Å3).47 This feature also explains the fact that the acceptor molecule in the hydrogen bonds is more perturbed than the donor.
10102
J. Phys. Chem. A, Vol. 114, No. 37, 2010
Silva et al. TABLE 3: Main Experimental and Computed (TDDFT-PCM) Vertical Excitation Energies (Ever) from Ground (S0) to excited (Sn) States state S1 S2 S3 S4
b
expa λ/nm
theor λ/nm
fb
341
343 333 273
0.8977 0.1719 0.4254
1
221
0.1764
1
291
excitation (coeff %)
assignments (π-π*) (π-π*) 1 (π-π*) 1
(π-π*)
70 71 71 71 70 75 71 77 73
r r r r r r r r r
69 69 67 68 65 65 66 66 68
(86.17) (86.24) (9.37) (85.6) (4.68) (5.68) (7.23) (6.68) (66.50)
a The trace symbol refeers to the nonexperimentally observed. Oscillator strength.
Figure 5. Experimental (black) and theoretical (blue) UV-vis absorption spectra.
On the contrary, the bond angle difference between the theoretical and experimental results is smaller than those found for the bond lengths; therefore, one can say there is good accordance between them. However, among the geometrical parameters, the torsion angles have the most pronounced deviations from the experimental data. In particular, three dihedral angles are considered, C3-N1-C2-C1, C2-N1C3-C4, and C2-N1-C3-C8, since these can be seen as indicatives of the torsion degree between the rings. As can be seen in Table 2, the monomer has small values for these dihedral angles (-3.96°, +180.00°, and -2.35°) and, therefore, a small distortion between the squarate and phenyl moieties. In the case of the dimer model, these three torsion angles are -8.97°, +169.6°, and -11.76°, for molecule A, respectively; and they were calculated as -10.61°, +148.10°, and -33.14° for molecule B. These values show that the torsion angle is enhanced in the dimer, mainly in molecule B where the atoms that form these dihedrals are directly involved in the hydrogen bonds. This theoretical result supports the experimental assumption that the hydrogen bonds and the torsional degree of freedom are mutually connected. Despite the undeniable influence of the intermolecular interaction in the molecular frame, the donor-acceptor distance in the hydrogen bond for the dimer model (3.063 Å) is overestimated in relation to the experimental data [2.8262(13) Å]. 3.2. Electronic Spectrum and TD-DFT Excitations Energy. Figure 5 displays the UV-vis absorption spectrum of ACSQ in DMSO solution, which shows a strong band absorption centered at 341 nm with a small shoulder at ca. 290 nm. In Table 3 are listed the experimental and theoretical absorptions of ACSQ. The squarate anion absorbs strongly at 269 nm and, similarly to squaraine, it presents a shoulder in the lower wavelength at 249 nm.2 In the case of the oxocarbon, Sato and Nagakura have shown that the shoulder is a result of the JahnTeller effect due to the vibronic coupling between the two first excited electronic states. The displacement of the squaraine absorption band for the red region in comparison to that for the squarate suggests a more effective electronic delocalization for the derivative. These same features can be proposed for the aniline, since this molecule absorbs strongly at 192 nm.48 The region below 250 nm was not collected since it is not clear due to the strong absorption of the solvent. Aiming to obtain a best description of the electronic transitions and absorption process, the computation of the electronic
Figure 6. Frontier orbitals and electronic transitions obtained from the TD-DFT calculations.
structure has been evaluated through the PCM/TD-DFT approach for a single ACSQ molecule (monomer model). For comparison, a Lorentzian fit of the theoretical electronic excitation energies and the respective oscillator strengths are presented in Figure 5. Among the electronic transitions predicted, the S0 r S1 transition (341 nm) has the most pronounceable oscillator strength (f ) 0.897) and it is worth mentioning the good accordance with the maximum of the absorption (λmax) experimentally measured. The TD-DFT excitation coefficients for this transition present a strong contribution (86%) from the 69 f 70 excitation, which are the higher occupied molecular orbital (HOMO) and lower unoccupied molecular orbital (LUMO), respectively. Therefore, it is assigned as a typical π-π* transition, and these orbitals are shown in Figure 6. A less intense but close to the aforementioned transition is theoretically described at 333 nm (S0 r S2) with f ) 0.1719. This transition is overlapped in the broad band centered at 341 nm. Also here, a single excitation has the major contribution to the transition from MO 69 to 71, characterizing it as a HOMO-LUMO+1 type. The 70 and 71 orbitals are quite similar in shape, suggesting the transition to the S1 state can also be described as π-π*. After the S0 r S1, the transition to the S3 excited state (273 nm) is the most intense (f ) 0.425) with the 68 f 71 as the main excitation referring to the HOMO-1 and LUMO+1 MO, respectively. Looking at Figure 6, we see that the HOMO-1 orbital is spread over the phenyl rings, as well as in the HOMO orbital, but with a smaller
1,2-Dianilinosquaraine
Figure 7. Infrared (A) and FT-Raman (B) spectra of the ACSQ compound. The inset presents the far-infrared spectrum.
contribution from the oxocarbonic portion. In turn, contrasting with the 68 MO, the LUMO+1 is distributed over the three rings. Therefore, the transition HOMO-LUMO+1 presents a considerable redistribution of the charges in both squarate and phenyl rings as well in the phenyl N-squarate moieties. This transition has been associated with the shoulder observed in the experimental spectrum at 290 nm. Despite the region below 250 nm not being assessable experimentally, a transition at 221 nm has been predicted, corresponding to the first peak shown in Figure 5. The excitation from the 68 to 73 MOs presents the most important contribution (66%) for this electronic transition. Looking at Figure 6 one can see that the LUMO+2 (73 MO) is a delocalized orbital over the two phenyl rings, which contrasts with the observation for the HOMO-1 (68 MO). This observation is very interesting because it suggests an intramolecular charge transference involving the two well-defined chemical groups, the oxcarbon and the phenyl rings. 3.3. Vibrational Spectra. For the best understanding of the vibrational spectra and their relation with the structure of the title system, the analysis was sectioned into three parts: the squarate modes, the phenyl modes, and the connector modes. The first group is the most interesting for this study, since it comprises the main spectroscopic features for the influence of the substituent groups in the most characteristic vibrational modes of the oxocarbon unity; the second group involves the characterization of the main phenyl ring modes; and in the last group the vibrational modes related to the chemical bond that link the phenyl and squarate rings are addressed. The infrared spectrum (Figure 7A) presents a broad set of bands around 3000 cm-1 as well as in the range between 1735 and 1275 cm-1; in comparison, the Raman spectrum (Figure 7B) shows sharp and defined bands that make this the most reliable spectroscopic technique in the present investigation. In Table 4, the wavenumber of the main bands are presented with their respective assignments. Theoretical vibrational wavenumbers were computed for the two models, monomer and dimer, and the assignments were made by visual inspection of the atomic motions in the Molekel program.49 As mentioned above, the monomer and dimer were optimized in the C2 symmetry point group and, therefore, only two possible symmetry keys are present for each vibrational mode; however, the symmetry key is shown only for the monomer in Table 4. For each calculated vibrational mode of the monomer, with some few exceptions, two modes of the dimer can be correlated. Due to changes in the force constant induced by the influence of the hydrogen bond, two different modes with coincident assignments can
J. Phys. Chem. A, Vol. 114, No. 37, 2010 10103 occur: one with the major contribution of molecule A and a second with the main contribution from molecule B, as shown in Table 4. Squarate Moiety Vibrational Modes. The pioneer work of West and Ito53 in the 1960s has shown oxocarbon ions belong to the Dnh group, in aqueous solution, and the normal coordinate analysis was done only for the in-plane modes; comparison with the spectroscopic data showed that D4h is the most probably symmetry for the squarate ion in aqueous solution. The infrared spectrum presents a strong and broad band centered at 1530 cm-1 assigned to the conjugated mode ν(CC) + ν(CO) (ν12, Eu), and other three bands at 259, 350, and 1090 cm-1 assigned to out-of-plane CO bending (ν4, Agu), in-plane CO bending (ν14, Eu), and CC stretching (ν13, Eu), respectively. The Raman spectrum shows six well-defined bands for the squarate anion at 194, 647, 723, 1123, 1593, and 1794 cm-1 assigned to the in-plane CO bend (ν4, B1g), ring bending (ν10, B2g), out-of-plane CO bending (ν11, Eg), ring breathing (ν2, A1g), and CC stretching (ν5, B1g), and the last two for the CO stretching (ν9, B2g; ν1, A1g), respectively. Since then, Raman spectra have been used as the main tool to characterize oxocarbon species, instead of infrared spectra. Ribeiro and co-workers,50 investigating the molecular dynamics of the squarate anion in acetonitrile, have shown that there is a blue shift of the Raman wavenumbers of the CO stretching and ring-breathing modes, when compared with the aqueous solution results. This same behavior has been observed in the metal coordination complexes, where the CO stretching is shifted for small wavenumbers when compared to that for the free anion.6,7,51,52 In the literature, data can be seen for several squarate salts, showing that this mode is very affected by the chemical environment, depending directly on the number, arrangement, and interactions of both solvent molecule and counterions.8-10 Therefore, the literature points to the symmetrical CO stretching as a very sensitive mode to the chemical environment experienced by the oxocarbon, and so, it is an excellent spectroscopic probe. This mode has been assigned in the infrared and Raman bands at 1795 cm-1 in the vibrational spectra of the ACSQ, and therefore, despite the effective intermolecular interactions involving the molecule in the solid state, it occurs at the same wavenumber observed for the squarate anion spectra in aqueous solutions.53 From the predicted vibrational spectrum of the dimer, two modes can be described as symmetrical νCO, at 1846 and 1841 cm-1, related to molecules A and B, respectively, whereas for the monomer model this mode was computed at 1846 cm-1. The comparison between these values shows that the hydrogen bond does not produce effective changes in the wavenumber of this vibrational mode and, therefore, it is not as sensitive to the chemical environment changes in the squaraine as in the free anion. In another way, the combined mode Q77, described as asymmetric CO stretching, shows a decrease in the wavenumber from monomer (1795 cm-1) to molecule A (1788 cm-1) to molecule B (1766 cm-1) (in the dimer structure). However, this mode has contributions from the CCox stretching oscillator, which can explain this behavior since, as shown by theoretical calculations and X-ray data (Table 2), the hydrogen bond has a strong influence over the CCox bond lengths. This last vibrational mode has been assigned to the strong infrared band at 1671 cm-1. Besides the numerical error intrinsic of the theoretical method, the large difference between the experimental and theoretical wavenumber for this mode is a consequence of the underestimation of the hydrogen bond strength in the theoretical approach.
10104
J. Phys. Chem. A, Vol. 114, No. 37, 2010
Silva et al.
TABLE 4: Computed Harmonic (Unscaled) and Observed Fundamental Wavenumbers (cm-1) theoretical/cm-1
experimental/cm-1 assignmentb
monomer
dimer
IR
Q90
3575
3198 mw
B
ν(NH)
Q89
3570
3153 w
A
ν(NH)
Q87
3195
3091 w
B
2
Q86
3191
3080 w
A
20b
Q84 Q82
3180 3171
3053 w
A B
7a 7b
Q79 Q78
3146 1846
B A
13 ν(CO)
Q77
1795
B
ν(CO) + ν(CCox)
Q76
1661
A
ν(CC)ox + δ(CNH) + ν(CoxN)
Q74
1642
1619 sh
B
8b + δ(CNH)
Q73
1637
1604 s
B
8a + δ(CNH)
Q72
1629
1600 s
A
ν(CC)ox + 9a
Q71
1555
1561 vw
B
δ(NH) + ν(CN) + 8b
Q70
1532
1552 vw
A
19a
Q69
1530
B
19a + δ(NH)
Q68
1525
1499 w
A
δ(NH) + 8b + ν(CC)ox
Q67
1473
1489 w
B
δ(NH) + 19b
Q66
1463
1459 s
A
δ(NH) + 19b
Q65
1415
1438 sh
B
ν(CN) + ν(CC)
Q64
1374
Q62
1338
Q61
1329
Q60
1296
Q59
1244
Q58
1218
Q57
1217
Q56
1189
Q55
1184
Q54
1184
Q53
1117
Q52
1112
Q50
1053
Q49
1020
Q47
1012
Q46
1009
3573 3487 3570 3495 3197 3196 3188 3185 3173 3176 3161 3151 1846 1841 1788 1766 1664 1653 1641 1640 1639 1634 1626 1623 1560 1550 1531 1529 1529 1528 1522 1519 1473 1471 1463 1460 1439 1417 1370 1362 1337 1333 1331 1327 1284 1281 1252 1240 1216 1207 1214 1207 1189 1184 1186 1181 1186 1181 1117 1110 1113 1107 1051 1050 1028 1053 1015 1015 1014 1013
3052 w 3038 w 3012 w 1795 m
Raman
sym keya
vibr no.
1797 w
1671 s 1661 s
1564 s
1547 s
1491 s
1340 w
1340 w
B
3
1316 sh
1313 vw
B
14 + δ(CNC)
1288 w
A
14 + δ(CNC)
1251 vs
A
ν(CC)ox + ν(CN) + 18a
1224 w
1222 w
B
ν(CN) + 18a
1181 w
1181 w
A
9a
1169 vw
1169 vw
B
9a
1157 mw
A
ν(CC)ox + δ(CNC)
1151
1150
B
9b
1108 vw
1108 vw
A
18b
1090 w
1088 w
B
18b + ν(CC)ox
A
18b + ν(CC)ox
A
18a; (18a + νCCox)
A
ν(CC)ox + 12
1001 ms
A
12 + ν(CC)ox
1006 sh
B
12 + ν(CC)ox
1081 w 1029 ms 1022 w
1,2-Dianilinosquaraine
J. Phys. Chem. A, Vol. 114, No. 37, 2010 10105
TABLE 4: Continued theoretical/cm-1
a
vibr no.
monomer
dimer
Q44
1007
Q43
978
Q38
857
Q37
840
Q36
840
Q35
798
Q34
766
Q33
765
Q32
753
Q31
708
Q30
703
Q29
703
Q28
661
Q27 Q26
630 624
Q25
606
Q24
586
Q23
571
Q22
570
Q20
515
Q19
505
Q17
415
Q15
365
Q14
358
Q13
294
Q11 Q10 Q8
242 238 158
Q6
103
1004 990 980 974 858 857 836 834 835 833 800 798 768 766 765 764 771 753 723 709 701 699 699 698 668 663 631 626 624 609 602 596 594 577 572 572 566 513 506 502 500 421 421 374 374 369 363 300 296 240 246 155 147 124 108
experimental/cm-1 IR
Raman
962
sym keya
assignmentb
B
5
900 w
898 vw
A
17b
847 vw
849 w
A
1 + δ(CCN)
838 vw
838 vw
B
10a
825 vw
A
10a
760 m
B
δ(CCN)ox + δ(CCO)ox + 6a
780 w
A
11
B
11
760 sh
753 m 787 mw
789 w
A
βCCC + 11
733 mw
729 ms
A
ringox breathing + 1
B
4
689 w
A
4
655
B
δ(CCC) + 6b
611 mw
A B
6b 6b + δ(CCC)
601 m
A
τ(NCCO)ox + β(NH) + 1
B
β(NH) + δ(CNC) + 1
690 m
653 606 w 600 vw 598 vw 576 vw
577 mw
A
τ(CCCC)ox + β(CNC) + 1
563 w
563 w
B
δ(CNC) + δ(CCO)
502 vw
A
16b + β(CNC)
B
β(NH) + 16b
B
16a
B
δ(CO) + δ(CNC) + 1
371 vw
A
δ(CO) + δ(CNC) + 1
311 m
A
δ(CO) + δ(CNC)
266 m 241 w
B A A
δ(CNC) + δ(CCO) β(CNC) + 16b δ(CNC)
221 w
A
τ(NCCN) + τ(OCCO)
499 w 409 mw 368 w 370 w
279 w
The symmetry keys are related to the monomer model. b For most benzene modes, the Wilson description was here adopted.
The infrared spectrum presents a strong band at 1661 cm-1 assigned to the ν(CC)ox (Q76), which is attributed to the broad infrared band centered at 1530 cm-1 of the free anion. This significant shifting in the wavenumber can be explained through the different oscillators that compose this mode. Observing the displacement vectors for this mode, it can be seen that the main bond that takes part of the oscillator is the C2-C2i, which is the shortest bond in the oxocarbon moiety with considerable difference for the respective bond length in the squarate sodium salt (0.04 Å). This can, in part, explain the wavenumber
enlargement observed for this mode. Another possible reason for the hypsochromic wavenumber shift relative to the squarate anion is the difference in the vibrational mode composition. For the free anion, this mode has a contribution from asymmetric CO stretching,54 whereas for the squaraine there is a strong contribution from the N-H bending and C-N stretching that can change the force constant. The first assumption reflects directly the influence of the replacement group on the squarate ring, increasing the double bond character for the CC bond relative to the squarate anion. The second proposal comprises
10106
J. Phys. Chem. A, Vol. 114, No. 37, 2010
two features, the character of the double bond in the CN moiety as shown from the X-ray diffraction data, and the influence of the hydrogen bond in the frequency of the vibrational mode due to the participation of the NH bending. The last effect can be qualitatively measured from the theoretical frequencies calculations. For the dimer this mode is shifted to a higher wavenumber when the NH is bonded in the intermolecular interaction (molecule B, 1664 cm-1) relative to the free NH (molecule A, 1653 cm-1). Also, it can be used to explain the wavenumber underestimation by the theoretical model compared to the experimental value. The same CCox oscillator is part of the mode Q72 assigned to the strong band at 1600 cm-1 in the Raman spectrum; however, in this case the analysis of the vibrational mode shows that it is accomplished by the 9a benzene mode (CC stretching and CH bending) showing how much the NH and CNox bonds contribute for the force constant enhancement. Georgopoulos et al. have shown that the carbon-carbon stretching modes in the range between 1300 and 1000 cm-1 are very sensitive to the symmetry of the oxocarbon ring.10 Using this characteristic, the authors have proposed that the infrared band at 1090 cm-1 and the Raman band at 1123 cm-1 can be used as probes of the equalization degree of the CCox bond distances (∆CC) and, therefore, of the electronic delocalization degree over the oxocarbon ring. They observed that the break of the oxocarbon ring symmetry is expressed as a splitting of these bands; the wavenumber difference between the two new generated bands (∆νCC) increases, whereas the ∆CC is enlarged. Therefore, this parameter in the infrared spectra for squaric acid as well as lithium, sodium, and potassium squarate salts are 116, 60, 40, and 12 cm-1, whereas in the Raman spectra these are 124, 86, 52, and 36 cm-1, respectively. For the rubidium salt a single band assigned to the νCC in this region is observed, which is similar to that observed in the aqueous solution of the squarate anion, indicating a higher symmetry for the squarate moiety in this salt. In the squaraine, these values are 59 (Q51-Q49) and 94 cm-1 (Q60-Q56), which are close to the measured values for the lithium salt. This feature reveals some interesting aspects about the squarate moiety in the ACSQ molecule; first, it indicates that the higher electronic delocalization degree is kept over the oxocarbon moiety after the substitution by the aniline groups, which is in accordance with the electronic spectroscopy and structural results. The computed ∆νCC in the Raman spectrum are 107, 97, and 95 cm-1 and in the infrared spectrum are 92, 63, and 79 cm-1 for monomer, molecule A, and molecule B (in the dimer), respectively. Thereby, in comparison with the experimental data, molecule A in the dimer has the closest results for the ∆νCC parameters. Furthermore, molecule A shows the smallest values when compared with other entities in the theoretical models, which is evidence that the hydrogen bond can contribute to the equalization of the CCox bonds. The last mode commonly used as probe in the vibrational investigation of the squarate ions is the ring breathing mode, assigned to the band at 723 cm-1; this one is accomplished by a band at 647 cm-1, assigned to the in-plane ring bending mode; these two bands are the most intense in this region of the Raman spectrum. As previously mentioned, the literature points to the ring breathing mode as a potential probe for the chemical environment in which the oxocarbon species is inserted. This mode is assigned to the medium-to-strong Raman band at 729 cm-1 for the ACSQ molecule; however, as in the case of the symmetric CO stretching, no expressive change in wavenumber is observed. The in-plane ring deformation mode is assigned to
Silva et al. the weak band at 653 cm-1, which is different from that observed for the free anion and their salts.10,53 Another band with considerable intensity assigned to the oxocarbon ring bending is observed at 760 cm-1 in the ACSQ Raman spectrum. This mode is not observed in both the Raman and infrared spectrum of the squarate in aqueous solution, and no description of these modes is found in the vibrational spectra of the alkaline squarate salts.10 It is theoretically predicted as an inactive mode at 830 cm-1 in the Raman and infrared spectra, since for the D4h symmetry three inactive modes are expected.53,54 This one was calculated at around 800 cm-1 and has contribution of the 6a benzene mode. This feature, added to the decrease of the D4h symmetry imposed by the substituent groups, can explain the activity of this mode for the squaraine. Phenyl Ring Modes Analysis. The assignments of the main aniline vibrational modes were made using the classical work of Evans55 and the theoretical study of the neutral and radical cation aniline by Wojciechowski et al.56 as references. Here, the Wilson description for the benzene modes are used as shown by Varsanyi for substituted phenyl rings.57 For the monomer model, each phenyl ring mode is duplicated since each ring can vibrate in-phase or out-of-phase. Since the principal symmetry axis is usually located in the center of C2-C2i and C1-C1i, the first one is associated with the A symmetry key, whereas the second one is associated with the B symmetry key, as represented in Table 4. Both infrared and Raman spectra of aniline present intense bands above 3000 cm-1 assigned to the CH stretching modes.56,57 However, there are several discrepancies in the vibrational assignments of these modes in the literature.56 For instance, some authors have assigned the phenyl mode 2 to the infrared bands at 3090 or 3084 cm-1;55,58-64 however, Wojciechowski has reassigned this mode to the strong polarized Raman band at 3072 cm-1.55,56 Additionally, the correlation of the CH stretching modes with the Wilson notation is not totally clear, only the 2 (all in-phase CH stretching) and 20a (out-of-phase CH stretching) modes can be easily identified, being assigned to the Raman band at 3080 cm-1 and to the infrared band at 3091 cm-1. Due to these features, we have decided to assign these modes uniquely as CH stretching, avoiding the correlation with the Wilson description. The infrared spectrum of the ACSQ molecule presents some unexpected bands between 3000 and 2800 cm-1, as can be seen in Figure 7, since this region typically refers to the aliphatic CH stretching. Indeed, no vibrational frequency was theoretically predicted in this region. A possible reason for this behavior is the presence of the solvent molecules in the solid state; however, the CHN microanalysis and mainly the crystalline structure reject this hypothesis. The absence of Raman bands in this region in the calculated spectrum excludes the possibility of existence of fundamental vibrational modes. Therefore, these were assigned as overtones or combination bands intensified by Fermi resonance effects. These bands have been tentatively attributed to the first overtone of the Q66 (2920 cm-1) and Q67 (2971 cm-1) modes and to the combination mode Q74+Q60 (2861 cm-1). As opposed to the CH stretching, most of phenyl ring modes in the Wilson-Varsanyi description could be identified through visual inspection of the calculated normal modes. However, some modes show strong coupling with some oxocarbon moiety vibrations. Most of the pure vibrational phenyl ring modes are in good agreement with the described in literature for the aniline spectra; however, some modes present effective changes in position. The 16a mode (torsional vibration), assigned to the
1,2-Dianilinosquaraine 409 cm-1 Raman band, is in good agreement with that described by Evans for aniline.55 The almost pure ring puckering vibration is assigned to the band at ca. 690 cm-1 in the infrared and Raman spectra, which is also in very good agreement with that observed for aniline. Q33 and Q34 modes refer to the 11 mode (CH out-of-plane vibrations) and are assigned to the bands at 753 and 780 cm-1, respectively. The difference between them is that the phenyl rings in the squaraine vibrate in-phase for the Q34 and out-of-phase for the Q33 mode; the first is infrared active and is close to the observed wavenumber for the aniline (755 cm-1), while the second is Raman active and is shifted to a higher wavenumber. On the other hand, Q36 and Q37 modes are infrared and Raman active, respectively, but with small difference between them (13 cm-1). The two infrared bands at 753 and 690 cm-1 are described in the literature as indicative of monosubstituted benzene derivatives.65,57 The remaining inplane C-H bending modes are the 9b (Q53), 18b (Q54), 9a (Q57), and 19a (Q70) modes, which are assigned to the bands at 1108, 1150, 1169, and 1552 cm-1 in both Raman and infrared spectra. These assignments agree very well with the description for aniline, the exception is mode 19a, which presents a strong shift of ca. 45 cm-1. This mode has been shown to be very sensitive to modifications in the benzene rings; for instance, in the radical-cation this band is not observed.56 The most characteristic vibrational modes of the phenyl ring present a contribution from oxocarbon vibrations. These mixed modes comprise some of the most intense bands in the vibrational spectra of the ACSQ squaraine. Aniline presents a typical Raman band at ca. 1602 cm-1 assigned to the 8a (CdC stretching) mode, whereas for the squaraine this mode occurs at 1604 cm-1, in good accordance with the observed value for the precursor. Another strong band in the Raman spectrum of aniline is observed at 1028 cm-1, attributed to the 18a benzene mode; this mode has been assigned to the band at 1029 cm-1 in the squaraine spectrum. However, in the dimer model, more specifically for molecule B, this mode also has the contribution of the CCox stretching. As described for the aniline compound by Wojciechowski et al.,56 the mode Q47 can be described as the trigonal ring “breathing” vibration derived from mode 12 of benzene, assigned to the band at 1001 cm-1 in the Raman spectrum of the squaraine together with a shoulder at 1006 cm-1 (Q46). This shoulder is assigned to the out-of-phase phenyl ring vibration of the mode 12. These modes also are coupled with the CC stretching of the oxocarbon ring. Vibrational Analysis of the Connector Moieties. Nakanaga et al. have accurately determined the NH2 stretching frequencies through infrared depletion spectroscopy.66-68 The authors observed two bands at 3508 and 3422 cm-1 with approximately analogous intensities, which were attributed to the symmetrical and asymmetrical NH2 stretching vibration, respectively. In that same paper these modes are assigned to the bands at 3488 and 3395 cm-1 for the cation-radical aniline. These decreases in wavenumber with the increase of the CdN bond character are well described in literature;56,65 for instance, for 2-butanimine the symmetrical stretching has been assigned to the weak infrared band at 3263 cm-1,69 for poly(ethylenimine) these modes were assigned to the infrared and Raman bands at 3219 and 3216 cm-1, representing an effective wavenumber decrease.70 Both symmetrical and asymmetrical NH stretching modes were related to the bands in the infrared spectrum of the ACSQ molecule at 3225 and 3198 cm-1, respectively. Since the structural data show a double bond character for Cox-N, as can be seen in Table 2, this fact can be used to explain these significant bathochromic shifts. Also in the case of the ACSQ
J. Phys. Chem. A, Vol. 114, No. 37, 2010 10107 derivative, the influence of the hydrogen bond involving the N-H moiety should also be considered since, in general, this interaction reduces the force constant of the N-H bond. A comparison between the calculated νN-H frequencies in the monomer and dimer can give a qualitative description of this effect, since as earlier mentioned, the hydrogen bond strength is underestimated in the theoretical representation of this interaction in the dimer system. As can be seen in Table 4, the monomer’s computed wavenumbers are 3575 and 3570 cm-1 for the symmetrical and asymmetrical vibrations, respectively. However, those vibrations with major contribution from molecule B are decreased to 3487 and 3495 cm-1, respectively. An inspection of Table 4 shows that several modes in the 1700-1400 cm-1 region of the vibrational spectra present contributions from the CNH bending. Notwithstanding, four bands will be emphasized in this discussion, since they have the major contribution of the CNH bending vibration. The first refers to the strong infrared band at 1564 cm-1, which is assigned to the combined mode δ(CNH) + ν(CN) + 8b. This same mode is assigned to the weak Raman band at 1459 cm-1; the difference between them is that the CNH moieties vibrate out-of-phase and in-phase, respectively. Two modes also assigned to the δ(CNH) have the same behavior, but in this case they are accomplished by the 19b benzene mode. These are assigned to the strong infrared bands at 1491 and 1438 cm-1 for these in-phase and out-of-phase vibrations, respectively. As in the case of the CNH bending vibration, several modes are influenced by the CN stretching. However, the normal-mode analysis shows two modes (Q65 and Q59) contributing mostly to the vibration, which are assigned to the infrared bands at 1438 and 1224 cm-1. The first mode involves the CN stretching, where the carbon atom is part of the squarate moiety (Cox-N), and in the second case it involves the nitrogen atom bonded to the phenyl ring (Cph-N). Aniline presents the νCN mode at 1280 cm-1, showing that in the squaraine there is a shift to a lower wavenumber indicating that the lone pair of the nitrogen atom is not involved in the resonance with the π-electrons of the phenyl ring, as is very well-known for aniline. In another way, the difference in wavenumbers between Q65 and Q59 indicates a higher character of double bond for the oxocarbon to the nitrogen linkage, showing that the nitrogen lone pair can be involved in the resonance of the π-electrons of the oxocabon moiety. The same feature is observed for a similar squaraine previously investigated (2-dimethylamino-4-anilino)squaraine, where these two modes occur at 1447 and 1223 cm-1, respectively. Therefore, the NH stretching, the CNH bending, and mainly the CN stretching reinforce, over an energetic criterion, the imine character of the Cox-N bond showed by the analysis of the geometrical parameters (Table 2). It is interesting to mention the similarity between the Q65 [ν(CoxN) + ν(CCox)] and Q77 [ν(CO) + ν(CCox)] modes; in both cases the Cox-N or Cox-O bonds in the squaraine oscillate out-ofphase and present contributions from the oxocarbon CC stretching. 4. Conclusions Summing up, the crystallographic data show that replacement of the oxygen atoms by the aniline groups produces significant changes in the bond length when compared to the observations for the squarate salts in the solid state. The four-membered ring presents good planarity and the bond lengths are between the formal double and single C-C bonds, showing, despite the changes caused by the presence of the substituent groups, the main features of the pure oxocarbon compounds; for
10108
J. Phys. Chem. A, Vol. 114, No. 37, 2010
instance, the high electronic delocalization over the structure is maintained. In the solid state the phenyl rings are distorted in relation to the squarate moiety. The solid presents an interesting supramolecular design, where a unidimensional arrangement formed by the NH · · · O hydrogen bonds is extended by π-π and C-H/π interactions in a two-dimensional fashion. Two models have been used in the DFT calculations of this chemical system, the isolated molecule and a dimer constructed by hydrogen bonds. The optimized molecule is approximately planar whereas the dimer tends toward the experimental slightly twisted molecular conformation. This theoretical feature supports the experimental assumption from the X-ray data that the hydrogen bond has the major contribution for the torsion of the phenyl rings, showing an interesting foldable molecular conformation that influences supramolecular assembling. Electronic spectroscopy data and the calculations of the electronic structure suggest a strong delocalized π-electron system involving mainly the oxocarbon and the amine moieties. Some of the most characteristic vibrational modes of the oxocarbon are very affected by substitution, mainly those involving the C-Cox bonds. However, the phenyl ring modes are less affected, with the exception of some out-of-plane bendings and of mode 19a. The vibrational analysis of the amine moiety also shows the character of the Cox-N bond, indicating the involvement of the nitrogen lone pair in the π-conjugation of the oxocarbon moiety. Acknowledgment. We thank CNPq, CAPES, and FAPEMIG (Brazilian agencies) for financial support. Supporting Information Available: Several structural possibilities (or canonical forms) for ACSQ. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Seitz, G.; Imming, P. Chem. ReV. 1992, 92, 1227–1260. (2) de Oliveira, L. F. C.; Mutarelli, S. R.; Gonc¸alves, N. S.; Santos, P. S. Quim. NoVa 1992, 15, 1227–1260. (3) Oliveira, V. E. d.; Diniz, R.; Oliveira, L. F. C. Quim. NoVa 2009, 32, 1917–1925. (4) da Silva, C. E.; Garcia, H. C.; Diniz, R.; Spezialli, N. L.; Yoshida, M. I.; Edwards, H. G. M.; de Oliveira, L. F. C. J. Phys. Chem. A 2007, 111, 11990–11995. (5) West, R. Oxocarbons; Academic Press: New York, 1980. (6) Correˆa, C. C.; Diniz, R.; Chagas, L. H.; Rodrigues, B. L.; Yoshida, M. I.; Teles, W. M.; Machado, F. C.; Edwards, H. G.; de Oliveira, L. F. C. Vibr. Spectrosc. 2007, 45, 82–88. (7) Correˆa, C. C.; Diniz, R.; Chagas, L. H.; Rodrigues, B. L.; Yoshida, M. I.; Teles, W. M.; Machado, F. C.; de Oliveira, L. F. C. Polyhedron 2007, 26, 989–995. (8) Georgopoulos, S. L.; Diniz, R.; Rodrigues, B. L.; de Oliveira, L. F. J. Mol. Struct. 2005, 741, 61–66. (9) Georgopoulos, S. L.; Diniz, R.; Rodrigues, B. L.; Yoshida, M. I.; de Oliveira, L. F. C. J. Mol. Struct. 2005, 157–163. (10) Georgopoulos, S. L.; Diniz, R.; Yoshida, M. I.; Speziali, N. L.; Santos, H. F. D.; Junqueira, G. M. A.; de Oliveira, L. F. J. Mol. Struct. 2006, 794, 63–70. (11) Law, K. Y. J. Phys. Chem. 1987, 91, 5184–5193. (12) Ajayaghosh, A. Acc. Chem. Res. 2005, 38, 449–459. (13) Yanagi, K.; Yakoubovskii, K.; Matsui, H.; Matsuzaki, H.; Okamoto, H.; Miyata, Y.; Maniwa, Y.; Kazaoui, S.; Minami, N.; Kataura, H. J. Am. Chem. Soc. 2007, 129, 4992–4997. (14) Cano, M. L.; Cozens, F. L.; Esteves, M. A.; Marquez, F.; Garcia, H. J. Org. Chem. 1997, 62, 7121–7127. (15) Lopes, J. G. S.; Farani, R. A.; de Oliveira, L. F. C.; Santos, P. S. J. Raman Spectrosc. 2006, 37, 142–147. (16) Lunelli, B.; Roversi, P.; Ortoleva, E.; Destrob, R. J. Chem. Soc., Faraday Trans. 1996, 92, 3612–2625. (17) Silva, C. E.; Diniz, R.; Rodrigues, B. L.; de Oliveira, L. F. C. J. Mol. Struct. 2007, 831, 187–194. (18) Neuse, E. W.; Green, B. R. J. Org. Chem. 1974, 39, 3881–3887. (19) CrysAlis RED, Oxford Diffraction Ltd., Version 1.171.32.38 (release 17-11-2008 CrysAlis171.NET).
Silva et al. (20) Sheldrick, G. M. S. SHELXL-97 - A Program for Crystal Structure Refinement; University of Goettingen: Gottingen, Germany, 1997. (21) Blessing, R. Acta Crystallogr. 1995, A51, 33–36. (22) Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565. (23) Macrae, C. F.; Edgington, P. R.; McCabe, P.; Pidcock, E.; Shields, G. P.; Taylor, R.; Towler, M.; van de Streek, J. J. Appl. Crystallogr. 2006, 39, 453–457. (24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Laham, A. M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian03, Revision C.02; Gaussian, Inc.: Wallingford, CT, 2003. (25) Ranganathan, A.; Kulkarni, G. U. J. Phys. Chem. A 2002, 106, 78113–7819. (26) Durig, J. R.; Hannum, S. E. J. Chem. Crystallogr. 1971, 1, 131–137. (27) Skancke, A. Acta Chem. Scand. 1968, 22, 3239–3244. (28) Diniz, R.; de Abreu, H. A.; de Almeida, W. B.; Sanviero, M. T. C.; Fernandes, N. G. Eur. J. Inorg. Chem. 2002, 2002, 1115–1123. (29) Etter, M. C.; MacDonald, J. C.; Bernstein, J. Acta Crystallogr. B 1990, 46, 256–260. (30) V. Busseti, F. M. Z. Kristallogr. 1997, 212, 302–304. (31) W. M. Macintyre, M. S. W. J. Chem. Phys. 1964, 42, 3563–3568. (32) B. Lunelli, R.; Soares, R. D. Phys. Chem. Chem. Phys. 1999, 1, 1469–1477. (33) Lunelli, B.; Roversi, P.E. O.; Destro, R. J. Chem. Soc., Faraday Trans. 1996, 92, 3611–3623. (34) Horiguchi, M.; Okuhara, S.; Shimano, E.; Fujimoto, D.; Takahashi, H.; Tsue, H.; Tamura, R. Cryst. Growth Des. 2007, 7, 1643–1652. (35) Horiguchi, M.; Yabunaka, S.; Iwama, S.; Shimano, E.; Lepp, Z.; Takahashi, H.; Tsue, H.; Tamura, R. Eur. J. Org. Chem. 2008, 3496–3505. (36) Fujimoto, D.; Takahashi, H.; Ariga, T.; Tamura, R. Chirality 2006, 18, 188–195. (37) Anandamanoharan, P. R.; Cains, P. W.; Jones, A. G. Tetrahedron: Asymmetry 2006, 17, 1867–1874. (38) Gray, J. C.; Pagelot, A.; Collins, A.; Fabbiani, F. P. A.; Parsons, S.; Sadler, P. J. Eur. J. Inorg. Chem. 2009, 2673–2677. (39) Lippold, I.; Vlay, K.; Goerls, H.; Plass, W. J. Inorg. Biochem. 2009, 103, 480–486. (40) Keypour, H.; Shayesteh, M.; Sharifi-Rad, A.; Salehzadeh, S.; Khavasi, H.; Valencia, L. J. Organomet. Chem. 2008, 693, 3179–3187. (41) Miodragovic, D. U.; Bogdanovic, G. A.; Miodragovic, Z. M.; Radulovic, M. D.; Novakovic, S. B.; Kaluderovic, G. N.; Kozlowski, H. J. Inorg. Biochem. 2006, 100, 1568–1574. (42) Shimazaki, Y.; Takani, M.; Yamauchi, O. Dalton Trans. 2009, 7854–7869. (43) Lu, Y.; Wang, Y.; Xu, Z.; Yan, X.; Luo, X.; Jiang, H.; Zhu, W. J. Phys. Chem. B 2009, 113, 12615–12621. (44) Maresca, M.; Derghal, A.; Carravagna, C.; Dudin, S.; Fantini, J. Phys. Chem. Chem. Phys. 2008, 10, 2792–2800. (45) Umezawa, Y.; Tsuboyama, S.; Honda, K.; Uzawa, J.; Nishio, M. Bull. Chem. Soc. Jpn. 1998, 71, 1207–1213. (46) Takahashi, O.; Kohno, Y.; Iwasaki, S.; Saito, K.; Iwaoka, M.; Tomoda, S.; Umezawa, Y.; Tsuboyama, S.; Nishio, M. Bull. Chem. Soc. Jpn. 2001, 74, 2421–2430. (47) Miller, K. J. J. Am. Chem. Soc. 1990, 112, 8533–8542. (48) Sato, H.; Hirota, K.; Nagakura, S. Bull. Chem. Soc. Jpn. 1965, 38, 962–964. (49) Varetto, U. Molekel 5.4S; Swiss National Supercomputing Centre: Manno, Switzerland, 2009. (50) Ribeiro, A. O.; Urahata, S. M.; Ribeiro, M. C. C. Phys. Chem. Chem. Phys. 2005, 6, 2956–2961. (51) Santos, P. S.; Sala, O.; Noda, L. K.; Gonc¸alves, N. S. Spectrochim. Acta A 2000, 56, 1553–1562. (52) Santos, P. S.; Amaral, J. H.; Oliveira, L. F. C. J. Mol. Struct. 1991, 243, 223–232. (53) Ito, M.; West, R. J. Am. Chem. Soc. 1963, 85, 2580–2584. (54) Junqueira, G. M. A.; Rocha, W. R.; De Almeida, W. B.; Dos Santos, H. F. Phys. Chem. Chem. Phys. 2003, 5, 437–435. (55) Evans, J. C. Spectrochim. Acta 1960, 16, 428–442. (56) Wojciechowski, P. M.; Zierkiewicz, W.; Michalska, D. J. Phys. Chem. 2003, 118, 10900–10911.
1,2-Dianilinosquaraine (57) Varsanyi, G. Vibrational spectra of benzene deriVatiVes; Academic Press: New York, 1969. (58) Niu, Z.; Dunn, K. M.; Boggs, J. E. Mol. Phys. 1985, 55, 421–425. (59) Seeger, D. M.; Korzeniewski, C.; Kowalchyk, W. J. Phys. Chem. 1991, 95, 6871–6882. (60) Castella-Ventura, M.; Kassab, E. Spectrochim. Acta 1994, Part A 50, 69–86. (61) Rauhut, G.; Pulay, P. J. Phys. Chem. 1995, 99, 3093. (62) Borisenko, V. E.; Baturin, A. V.; Przeslawska, M.; Koll, A. J. Mol. Struct. 1997, 428, 53–62. (63) Tzeng, W. B.; Narayanan, K. K. C. S.; Tung, C. C. J. Mol. Struct.: THEOCHEM 1999, 489, 47–54.
J. Phys. Chem. A, Vol. 114, No. 37, 2010 10109 (64) Palafox, M. A.; Nun˜ez, J. L.; Gil, M. J. Mol. Struct.: THEOCHEM 2002, 593, 101–131. (65) Colthup, N. B.; Wiberly, S. E.; Daly, L. H. Introduction to infrared and Raman spectroscopy; Academic Press: New York, 1990. (66) Nakanaga, T.; Piracha, N. K.; Ito, F. J. Phys. Chem. 2001, 105, 4211–4215. (67) Nakanaga, T.; Ito, F. Chem. Phys. Lett. 2001, 348, 270–276. (68) Nakanaga, T.; Ito, F. J. Phys. Chem. A 1999, 103, 5440–5445. (69) Zhou, Z.; Zhou, X.; Fu, H.; Fu, A.; Du, D. Spectrochim. Acta Part A 2003, 59, 2593–2601. (70) Hashida, T.; Tashiro, K. Polymer 2007, 48, 7614–7622.
JP105346H
