PCCP

Theoretical analysis of the oxocarbons: The role played by the solvent and counter-ions in the electronic spectrum of the deltate ion
Georgia M. A. Junqueira,ab Willian R. Rocha,a Wagner B. De Almeidab and Helio F. Dos Santos*a
a

NEQC: Nucleo de Estudos em Qumica Computacional, Departamento de Qumica, ICE, Universidade Federal de Juiz de Fora, Campus Martelos, CEP 36036-330 Juiz de Fora, MG, Brasil. E-mail: helius@quimica.ufjf.br LQC-MM: Laboratorio de Qumica Computacional e Modelagem Molecular, Departamento de Qumica, ICEx, Universidade Federal de Minas Gerais, Campus Pampulha, CEP 31270-901 Belo Horizonte, MG, Brasil

Received 24th October 2001, Accepted 18th March 2002 First published as an Advance Article on the web 9th May 2002

The structure and spectroscopic properties of the deltate anion are calculated in the gas phase using ab initio quantum chemical methods and in aqueous solution through a sequential Monte Carlo-quantum mechanical procedure. The eects of the solvent and counter-ions on the electronic spectrum are analyzed, showing that both should be included in the calculation in order to reproduce the observed UV spectrum. For the smallest cyclic oxocarbon, the deltate anion, the calculated electronic transitions were 254 and 246 nm considering the [Li2(C3O3)(H2O)20] species. This is in accordance with the expected behavior for the oxocarbon series, predicting absorption bands close to 200 nm for the deltate anion. Despite the extensive literature on oxocarbons,320 theoretical studies are sparse.1520 In a recent study20 we have analyzed the structure and spectroscopic properties of the croconate ion in the gas phase and aqueous solution using quantum mechanical ab initio methods and Monte Carlo simulation. The main conclusion drawn was that the solvent and counter-ion eects are important to reproduce the observed electronic spectra of the croconate anion in solution.20 In the present study we extend our theoretical proposal reported in ref. 20 to the smallest member of the oxocarbon series, the deltate ion. In addition to the overall solvent eect analyzed for the croconate derivative, in this paper we describe a detailed analysis of the hydrogen bond eect on the electronic spectrum.

Introduction
Oxocarbons are molecules with general formula (CnOn)2 rst identied by West as a new class of aromatic compounds.1 The main oxocarbon representatives are rhodizonate (n 6), croconate (n 5), squarate (n 4) and deltate (n 3) (Fig. 1). Their structures are planar with Dnh symmetry, as was proposed from Raman and IR spectroscopic analysis26 and conrmed by X-ray diraction studies.710 Due to the high degree of electronic delocalization, the oxocarbon ions present a strong absorption in the UV and visible regions, except for the deltate derivative which is expected to absorb at wavelengths lower than 200 nm.11 The electronic structure of the oxocarbons has raised some questions about the potential aromaticity of these compounds, and there are several works in the literature dealing with this interesting property.12,13 The general conclusion is that the aromaticity decreases with the ring size.13 For the three-member ring anion, the most aromatic compound of the oxocarbons series, both s and p aromaticity have been considered to be important.13 An interesting aspect concerning the electronic spectra of the oxocarbons in solution is the presence of two absorption bands that have been attributed to the JahnTeller eect on the rst electronic excited state.11

Theoretical methodology
Gas phase The geometry of the deltate ion was fully optimized in the gas phase at the MllerPlesset second order perturbation theory (MP2) level, using standard split-valence basis-sets with inclusion of polarization [6-31G(d)] and diuse [6-31+G(d)] functions on the O and C atoms. The vibrational frequencies, infrared intensities and Raman activities were also obtained for the (C3O3)2 anion at each level of calculation mentioned before. The electronic spectra in the gas phase were calculated using the ZINDO program,21 within the INDO/CIS semiempirical approach.22 Aqueous solution The solvent was included through Monte Carlo (MC) statistical mechanics simulation employing standard procedures23 with the Metropolis sampling technique24 and periodic boundary conditions using the minimum image method in a cubic box. The simulations were performed in the canonical (NVT) Phys. Chem. Chem. Phys., 2002, 4, 25172523 2517

Fig. 1 Cyclic oxocarbon ions. (a) deltate, (b) squarate, (c) croconate and (d) rhodizonate.

DOI: 10.1039/b109738e

This journal is # The Owner Societies 2002

Table 1 Intermolecular potential parameters used in the Monte Carlo simulations (qi in e, ei in kcal mol1 and si in A) Site H2O O H
a

qi 0.834 0.417

ei 0.152 0.000

si 3.151 0.000

(C3O3)2 b C O

0.270 0.940

0.105 0.210

3.750 2.960

[Li2(C3O3)] (Form A)b c C(1) 0.202 O(2) 0.720 C 0.325 O 0.940 Li 0.830

0.105 0.210 0.105 0.210 0.018

3.750 2.960 3.750 2.960 2.126

Fig. 2 Calculated auto-correlation function for the deltate anion in water.

a TIP3P potential.26 b OPLS parameters24 with the charges calculated using the ChelpG tting procedure25 at the MP2/6-31+G(d) level of theory. c The optimized structure and numbering scheme are shown in Fig. 4.

(t).30 The correlation step t can be obtained from the autocorrelation function of the energy [C(t)]. The interval between uncorrelated congurations, the so-called correlation step t, is calculated by performing the integration from zero to innity of C(t) (eqn. (2)). t Z1 Ctdt
0

ensemble. The system under investigation consisted of one deltate anion and 499 solvent molecules (water). The volume of the cubic box was determined by the experimental density of water,25 which is 0.9966 g cm3 at T 298.15 K. The intermolecular interactions were described by the Lennard-Jones plus Coulomb potential, shown in eqn. (1), with 3 parameters for each atom i(ei , si and qi). "   6 # X sij 12 sij qi qj eij 1 Uint 4 rij rij rij i;j The set of intermolecular potential parameters employed in the simulations is shown in Table 1. The s and e parameters were taken from the OPLS force eld26 and the atomic charges were obtained from the ChelpG tting procedure27 at the MP2/6-31+G(d) level. The intermolecular interactions were spherically truncated within a center of mass separation smal ler than the cut-o radius rC 12.374 A. In all simulated systems, the molecules were kept with rigid geometries. The water molecules in their C2v structure with rOH 0.9572 A and  28 cHOH 104.52 and the deltate ion in its D3h structure optimized at the MP2/6-31+G(d) level of theory. The initial positions and orientations of the molecules were generated randomly. Congurations for the subsequent analysis were stored after every 200 MC steps, i.e., after all solvent molecules sequentially attempt to translate in all cartesian directions and also attempt to rotate around a randomly chosen axis. The maximum allowed displacement of the molecules was selfadjusted after 5000 MC steps to give an acceptance ratio of new congurations around 50%. The maximum rotation angle was xed during the simulation in dy 15 . The simulations consisted of a thermalization phase of 10 000 MC steps, followed by an averaging stage of 60 000 MC steps. Thus, the total number of congurations generated from the MC simulation is 3 107. To analyze the solvent eect on the electronic spectrum of the solute molecule we employed a sequential Monte Carlo/ quantum mechanics procedure,29 in which the quantum mechanical calculations are performed on supermolecular clusters, generated by the MC simulations, composed of one deltate molecule and all solvent molecules within a specic solvation shell. Since we have thousands of congurations, a procedure must be applied to reduce the number of supermolecular clusters that are calculated quantum mechanically. The congurations are selected according to the correlation step 2518 Phys. Chem. Chem. Phys., 2002, 4, 25172523

For Markovian processes C(t) follows an exponentially decaying function (eqn. (3) and Fig. 2). In general, congurations separated by 2t, or larger, are considered uncorrelated.28 Ct
n X i

ci et=ti

For the deltate anion in water the following equation was tted from the data depicted in Fig. 2: Ct 0:455et=20:80 0:5453et=586:5 4

The correlation step calculated from eqn. (2) and (4) was t 328 MC steps and congurations separated by approximately 2.5t ($800 MC steps) have been selected, giving only 14% correlation between the selected congurations. As we have 500 molecules in the simulation, the congurations are selected at each 4 105 MC congurations. A total of 75 nearly uncorrelated congurations were selected during the simulation and considered in the quantum mechanical electronic spectrum calculation. The solvation shells were dened from the radial distribution functions (RDF). The electronic spectra in aqueous solution were then calculated using the semi-empirical ZINDO program,21 within the INDO/CIS semiempirical approach.22 As the appropriate Boltzmann weights are included in the Metropolis Monte Carlo sampling technique, the average values of the transition energies and solvatochromic shifts are given as an arithmetic average over 75 uncorrelated congurations. All ab initio calculations reported here were performed using the Gaussian 98 program.31 The Monte Carlo statistical mechanics simulations were performed using the DICE program.32

Results and discussion

Structure and vibrational spectrum of the (C3O3)2 ion The optimized structural parameters obtained for the deltate anion are reported in Table 2. At both levels of theory used the structure was found to be planar with D3h symmetry. The X-ray structure for the deltate derivative has not yet been

Table 2 Structural parameters calculated for the free deltate iona MP2/6-31G(d) C=O CC cCCC cCCO
a

MP2/6-31+G(d) 1.300 1.443 60.0 150.0

1.292 1.455 60.0 150.0

Bond lengths in A and bond angles in degrees.

Raman spectrum. The calculated frequencies and Raman activities are reported in Table 4 and compared with the experimental available data.33 The A20 mode at 777 cm1 assigned to d(CO)/n(CC) is inactive in both the IR and Raman, and was not included in Tables 3 and 4. It can be seen from the values of Table 4 that there is good agreement between the theoretical and experimental frequencies, relative intensities and normal mode assignment. In a systematic way the frequency values are smaller when diuse functions are added to the basis set, the deviation being more pronounced (see Fig. 3 for the normal mode representation). Electronic spectrum in the gas phase The computed gas phase electronic spectrum for the deltate ion showed two degenerate transitions centered at 279 nm with oscillator strength f 0.3, attributed to the transitions from HOMO to LUMO and from HOMO to LUMO + 1. The experimental data for the electronic spectrum of the deltate has not been reported, however, according to the literature two absorption bands at wavelength lower than 200 nm are expected.11 As observed for the other members of the oxocarbons class, the presence of these bands in the experimental spectrum of deltate ion has been attributed to the break of degeneracy due to JahnTeller distortions of the rst electronic excited state.11 In our previous paper,20 it was shown that the inclusion of counter-ions, coordinated to the solute, and solvent molecules splits the absorption band and a blue shift is calculated relative to the gas phase value. In order to include the counter-ion eect on the structure and electronic properties of the deltate ion, two complexed forms with stoichiometry [Li2(C3O3)] were proposed (Fig. 4). In the complexed form A the deltate ion acts as a bis-bidentate ligand, with the lithium atoms sharing a common oxygen atom. In the form B the deltate also acts as bis-bidentate ligand, but with the lithium atoms lying on opposite sides of the ring. Both structures were fully optimized at the MP2/6-31+G(d) level. The energy dierence is only 0.64 Kcal mol1 favoring the form A. The calculated electronic transitions were (in nm/f): 261/0.2, 254/0.2 (form A) and

reported. However, considering the good agreement between theory and experiment obtained for the croconate-like molecule,20 we do believe that the results reported here should be reliable enough. Analyzing the values in Table 2 it can be seen that the inclusion of diuse functions does not aect the structure signicantly. The vibrational infrared and Raman spectra were calculated and the frequencies and absolute intensities are reported in Tables 3 and 4 respectively. Regarding the IR spectrum of the deltate ion, four active absorption bands were assigned (Table 3) relative to the out-of-plane CO deformation and combination of asymmetric CO stretching and in-plane CO bending (Fig. 3). In general the calculated band positions obtained in the gas phase at the MP2/6-31G(d) and MP2/631+G(d) levels are in good agreement with the experiment carried out in the solid phase.33 With regard to intensities, the greatest deviations were found for the low frequency modes. As can be observed in Table 3, the inclusion of diuse functions on the basis-set causes a systematic decreasing of the frequency values relative to those obtained at the MP2/6-31G(d) level. The IR intensities are not very sensitive to the inclusion of diuse functions, except for the normal mode close to 1500 cm1 that show1ed intensity equal to 675 [MP2/6-31G(d)] and 1221 km mol1 [MP2/6-31+G(d)]. The frequency calculated for this mode is also more sensitive to the basis-set. The deviation from the observed value was 55 [MP2/6-31G(d)] and 6 cm1 [MP2/6-31+G(d)]. According to the vibrational representation of the D3h point group, the deltate ion should have six active modes in the

Table 3 Vibrational frequencies (in cm1) and IR intensities (in parentheses, km.mol1) calculated for the free deltate ion (D3h) MP2/6-31G(d) n1 n 2b n 3b n 4b
a

MP2/6-31+G(d) 186(12) 275(4) 970(100) 1464(1221)

Expt.a K3C3O3(s)/ Li2C3O3(s) 258m/236m 325m/341m 965s/995,972s 1445vs/1470vs

Assignment A200 doop(CO) E0 c E0 c E0 c

271(13) 292(3) 979(102) 1525(675)

Values taken from ref. 33. The experimental values were obtained in the solid state for K2C3O3 and Li2C3O3 . The abbreviations s, m, w and vs stand for strong, medium, weak and very strong respectively. b Bands doubly degenerate. c These normal modes arise from combination of asymmetric CO stretching, CC stretching and in plane CO bending (see Fig. 3).

Table 4 Vibrational frequencies (in cm1) and Raman activities (in parentheses, A4 u1) calculated for the free deltate ion (D3h) MP2/6-31G(d) n1 n2 n 3b n 4b n 5b n 6b
a

MP2/6-31+G(d) 1763(28) 758(42) 1464(12) 970(30) 275(10) 645(11)

Expt.a K3C3O3(s)/Li2C3O3(aq) 1818w/1835m 786s/803s 1432w/1446w 966m/992m 321m/346m 689w/696w

Assignment A100 n(CO) A10 ring breathing E0 c E0 c E0 c E00 dring

1787(7) 772(41) 1525(0) 980(10) 292(8) 681(1)

Values taken from ref. 33. The experimental values were obtained in the solid state for K2C3O3(s) and aqueous solution for Li2C3O3(aq). The abbreviations s, m, w and vs stand for strong, medium, weak and very strong respectively. b Bands doubly degenerate. c Additional in-plane modes E0 arise from combination of asymmetric CO stretching, CC stretching and in plane CO bending (see Fig. 3).

Phys. Chem. Chem. Phys., 2002, 4, 25172523

2519

Fig. 4 MP2/6-31+G(d) optimized geometries for the A and B forms of the [Li2(C3O3)] complex.

Electronic spectrum in aqueous solution Initially we will discuss separately the hydrogen bonds between the deltate anion and water molecules. Fig. 5(a) shows the RDF for the oxygens of the deltate ion and water [gOO(r)]. As can be seen there is a hydrogen bond peak that starts at 2.35 A and goes up to 3.45 A. Integration of this peak gives

Fig. 3 Normal modes calculated for the (C3O3)2 anion at the MP2/ 6-31+G(d) level.

387/0.01, 351/0.1 (form B) assigned as HOMO ! LUMO and HOMO ! LUMO + 1 for both structures. By comparing with the values for the free anion (279 nm), the HOMO ! LUMO transition undergoes a blue shift of 18 nm and the HOMO ! LUMO + 1 a corresponding shift of 25 nm for the complexed form A. For the structure B a red shift was observed in both electronic transitions. As is expected for the deltate anion, the absorption bands should be close to 200 nm.11 This shows that the Li+ ions, coordinated on form A, play an important role in determining the electronic spectrum of the molecule. 2520 Phys. Chem. Chem. Phys., 2002, 4, 25172523

Fig. 5 Radial distribution functions obtained from the MC simulation for the (C3O3)2 ion in water (a) OO, (b) cmcm.

Table 5 Hydrogen bond analysis and electronic transitions calculated for the (C3O3)2 ion in the gas phase and in water Hydrogen bondsa b Number of hydrogen bonds (NHB)c Gas phase 3 4 5 6 7 8 9 [C3O3(H2O)22]2
a

Wavelength/nm [Oscillator strengths f ] Percentage of selected congurationsd 23% 27% 17% 20% 5% 7% 1% Number of valence electrons (NE)e 56 64 72 80 88 96 104 208 HOMO ! LUMO 279 276 275 275 274 272 269 272 (267 2) [0.3] [0.3] [0.3] [0.3] [0.3] [0.3] [0.3] [0.3] [0.3] HOMO ! LUMO + 1 279 271 270 269 268 268 268 266 (265 2) [0.3] [0.3] [0.3] [0.3] [0.3] [0.3] [0.3] [0.3] [0.3]

The molecules involved in hydrogen bonds were selected according to the geometric criteria of rOO 4.0 A and cHOO 30 . b The average energy of the hydrogen bonds is (16.9 3.9) kcal mol1. c Number of hydrogen bonds presented in the congurations selected from the MC simulation. d The occurrence of hydrogen bonds was calculated relative to the total number of hydrogen bonds (363) on the 75 uncorrelated congurations. e Number of valence electrons explicitly included in the QM calculations.

4.8 water molecules. In the 75 uncorrelated MC congurations we found a total of 363 hydrogen bonds, giving 4.8 hydrogen bonds for each conguration. This is the same average value obtained from the radial distribution function in Fig. 5(a). In Table 5 the data were organized according to the occurrence of congurations with a distinct number of hydrogen bonds (NHB). We found that 87% of the selected congurations present 36 hydrogen bonds, the congurations with 3 and 4 hydrogen bonds being more abundant. As shown in Table 5, a blue shift of both transitions is observed with increase in the number of hydrogen bonds with the maximum value found to be 10 nm (NHB 8) and 13 nm (NHB 9) respectively for HOMO ! LUMO and HOMO ! LUMO + 1 transitions. In the next step we analyzed the overall solvent eects on the electronic spectrum of the free deltate ion. The solvation shells were dened from the RDF considering the center of mass (cm) of the solute and solvent molecules [gcmcm(r)]. The gcmcm(r) calculated for the deltate ion in water solution is depicted in Fig. 5(b), where two solvation shells are easily dis tinguished. The rst solvation shell was taken from 2.95 A to 5.55 A and the second one nished at 7.35 A. The peak from 4.45 to 5.55 was not considered as a unique solvation shell. This is a consequence of the uctuation in the Monte Carlo simulation and the step of sampling used to calculated the RDF. After spherical integration of the gcmcm(r) over the corresponding intervals, 22 and 57 water molecules were found in the rst and second solvation shells respectively. For each solvation shell a total of 75 electronic spectrum calculations, on nearly uncorrelated congurations, were performed and the values of the wavelength and oscillator strength averaged out. The results obtained are reported in Table 5. Our largest calculation, up to the addition of the rst solvation shell, includes one deltate ion and 22 water molecules. According to our previous results for the croconate ion20 when the solvent is explicitly included, the HOMO ! LUMO and HOMO ! LUMO + 1 transitions remains unaltered up to the third solvation shell. So, for deltate compounds we performed the electronic spectrum calculations considering only the addition of the rst solvation shell. The results obtained show that the interaction of the deltate ion with the solvent shifts the absorption bands in the electronic spectrum by 13 nm. A comparison between the data reported in Table 5 shows that the inclusion of few water molecules yielding small supermolecules is not enough to describe the solvent eect on the electronic spectrum of the deltate ion, being necessary the inclusion, at least, of the rst solvation shell. In the last part of this study we included and analyzed the combined eects from the counter-ion and solvent on the

Fig. 6 Radial distribution function between Li and O atoms, obtained from the MC simulation of the [Li2(C3O3)] complex (Form A) in water.

electronic spectrum. As was said before, in the gas phase, the coordination of the Li+ cations in form A splits the deltate absorption band (279 nm) into two blue shifted absorptions, centered on 261 and 254 nm. This structure was selected for the MC simulation. The simulation protocol was the same as used for the free (C3O3)2 anion described in the methodology section. The s and e parameters for the Li+ ion were taken from the OPLS force eld and the atomic charges calculated using the ChelpG procedure at the MP2/6-31+G(d) level of theory. The complete set of parameters used for structure A is included in Table 1. In Fig. 6 the RDF between the Li+ ion and the oxygen from water, gLiO(r), is depicted. The narrow peak close to 2 A in the gLiO(r) function indicates that the water molecules are strongly bounded to the Li+ ion. Spherical integration of this peak gives two water molecules per lithium atom. The electronic spectrum calculated for the [Li2(C3O3)(H2O)n] is described in Table 6. As discussed for the free deltate, the hydrogen bonds eects were analyzed separately. From the gOO(r) RDF represented in Fig. 7(a) it can be seen that there is a hydrogen bond peak that starts at 2.35 A and . Integration of this peak gives 2.4 water goes up to 3.15 A molecules. In the 75 MC congurations a total of 154 hydrogen bonds were found, giving 2.1 hydrogen bonds in each conguration. This is close to the average value obtained from the radial distribution function gOO(r) (Fig. 7(a)). As can be observed in Table 6, the maximum number of hydrogen bonds observed was 4, the congurations with 3 hydrogen bonds Phys. Chem. Chem. Phys., 2002, 4, 25172523 2521

Table 6

Hydrogen bond analysis and electronic transitions calculated for the [Li2(C3O3)] complex (form A) in gas phasea and water Wavelength/nm [Oscillator strength f ] Percentage of selected congurationse 16% 27% 35% 8% Number of valence electrons (NE) f 40 48 56 64 192 HOMO ! LUMO 261 259 256 255 255 254 [0.2] [0.2] [0.2] [0.3] [0.3] [0.3] HOMO ! LUMO + 1 254 251 249 249 247 246 [0.2] [0.2] [0.2] [0.2] [0.3] [0.2]

Hydrogen bonds (HB)b c Number of hydrogen bonds (NHB)d Gas phase 1 2 3 4 Li2C3O3(H2O)20

a The optimized structure of the complex [Li2(C3O3)] (form A) is depicted in Fig. 4(a). b The molecules involved in hydrogen bonds were selected according to the geometric criteria of rOO 4.0 A and cHOO 30 . c The average energy of the hydrogen bonds is (6.0 2.3) kcal mol1. d Number of hydrogen bonds presented in the congurations selected from the MC simulation. e The occurrence of hydrogen bonds was calculated relative to the total number of hydrogen bonds (154) on the 75 uncorrelated congurations, 15% do not make hydrogen bonds. Number of valence electrons explicitly included in the QM calculations.

vent eects. Finally, it is important to point out that we are not neglecting the role played by the JahnTeller eect. We are stating that there may be other eects that can also be used to justify the experimental nding.

Conclusions
In the present work the structure and spectroscopic properties of the deltate ion were analyzed in the gas phase and aqueous solution using theoretical quantum mechanical ab initio methods and a sequential Monte Carlo/quantum mechanical approach. The vibrational spectra (infrared and Raman) are well reproduced by the calculations in the gas phase at the MP2/6-31G(d) and MP2/6-31+G(d) levels of theory. A sequential Monte Carlo/quantum mechanical study of the inuence of the hydration shells on the P ! P* transition of the deltate ion was carried out. The solvatochromic shift was analyzed as a function of the solvation shells, starting from the hydrogen bonds and extending to the rst solvation shell. The calculated electronic spectra for the free deltate anion in the gas phase using the MP2/6-31+G(d) optimized geometry showed two degenerate transitions at 279 nm. The inclusion of the solvent splits the absorption into two distinct transitions centered at 267 and 265 nm, attributed as HOMO ! LUMO and HOMO ! LUMO + 1, when the rst solvation shell (22 water molecules) is considered. The inclusion of solvent molecules beyond the rst solvation shell did not aect signicantly the electronic spectrum of the free deltate anion. Recently we have analyzed a new proposal in order to understand at the molecular level the observed electronic spectrum of oxocarbons. It is based on the interaction between the oxocarbon anion and its counter-ion, even in solution. For the deltate ion two distinct complexed structures were optimized and the electronic spectrum calculated. For the most stable form A of the complex [Li2(C3O3)], the UV transitions were calculated in the gas phase at 261 and 254 nm. The inclusion of the solvent up to the rst solvation shell [Li2(C3O3)(H2O)20] shifted the transition to 254 and 246 nm respectively. The experimental electronic spectrum for deltate has not been reported, however, following the behavior observed for the oxocarbon series, absorption bands close to 200 nm are expected for the (C3O3)2 ion, which is in agreement with the values found in this work considering the solvent and the counter-ions eects. Our results suggest that the deltate ion should interact with the counter-ion in solution with the form A being predominant. The general conclusion is that in order to reproduce the experimental electronic spectrum of the oxo-

Fig. 7 Radial distribution functions obtained from the MC simulation of the [Li2(C3O3)] complex (form A) in water: (a) OO, (b) cmcm.

being more abundant (35%). Considering only the water molecules that are involved in hydrogen bonds with the solute, a blue shift was observed from 261 to 255 nm (HOMO ! LUMO) and 254 to 247 nm (HOMO ! LUMO + 1). In order to analyze the eect of additional solvent molecules on the electronic spectrum of deltate, we included water molecules up to the rst solvation shell dened from the RDF gcmcm(r) depicted in Fig. 7(b). The result obtained for the cluster [Li2(C3O3)(H2O)20] were 254 and 246 nm (Table 6). These are close to the values found considering only congurations with 4 hydrogen bonds. Dierent from the free deltate anion, for the lithium coordination compound the small clusters involving only molecules hydrogen bonded to the solute can be used to describe the overall eect of the solvent on the electronic spectrum. Since electronic transitions close to 200 nm are expected for the deltate anion, our results suggest that the experimental electronic spectrum of that molecule in solution can be reproduced by including the counter-ions and sol2522 Phys. Chem. Chem. Phys., 2002, 4, 25172523

carbons in solution, we must take into account the combined eect of the solvent and the counter-ions.

Acknowledgement
The authors would like to thank the CNPq (Conselho Nacio nal de Desenvolvimento Cientco e Tecnologico) and FAPEMIG (Fundacao de Amparo a Pesquisa do Estado de Minas Gerais) for nancial support. The authors also thank the CENAPAD-MG/CO-NAR-UFJF for providing the computational facilities.

References
1 R. West, H. Y. Niu, D. L. Powell and M. V. Evans, J. Am. Chem. Soc., 1960, 82, 6204. 2 D. Eggerding and R. West, J. Am. Chem. Soc., 1975, 97, 207. 3 M. Ito and R. West, J. Am. Chem. Soc., 1963, 85, 2580. 4 R. West and H. Y. Niu, J. Am. Chem. Soc., 1963, 85, 2586. 5 R. West, J. W. Downing, S. Inagaki and J. Michl, J. Am. Chem. Soc., 1981, 103, 5073. 6 (a) P. S. Santos, J. H. Amaral and L. F. C. de Oliveira, J. Mol. Struct, 1991, 243, 223; (b) L. F. C. de Oliveira, J. G. da Silva Lopes, P. M. V. B. Barone, M. C. C. Ribeiro and P. S. Santos, J. Mol. Struct (THEOCHEM), 1999, 510, 97; (c) J. G. S. Lopes, L. F. C. de Oliveira and P. S. Santos, Spectrochim. Acta, Part A, 2001, 57, 399. 7 M. D. Glick, G. L. Downs and L. F. Dahl, Inorg. Chem, 1964, 3, 1712. 8 M. D. Glick and L. F. Dahl, Inorg. Chem., 1966, 5, 289. 9 F. Dumestre, B. Soula, A. M. Galibert, P. L. Fabre, G. Bernardinelli, B. Donnadieu and P. Castan, J. Chem. Soc., Dalton. Trans., 1998, 24, 4131. 10 N. S. Goncalves, P. S. Santos and I. Vencato, Acta Crystallogr. Sect.C., 1996, 52, 622. 11 M. Takahashi, K. Kaya and M. Ito, Chem. Phys., 1978, 35, 293. 12 J. Aihara, J. Am. Chem. Soc., 1981, 103, 1633. 13 P. von R. Schleyer, K. Najaan, B. Kiran and H. Jiao, J. Org. Chem., 2000, 65, 426. 14 (a) G. Seitz and P. Imming, Chem. Rev., 1992, 92, 1227; (b) B. Zhao and M. H. Back, Can. J. Chem., 1992, 70, 135; (c) M. Dory, J. M. Andre, J. Delhalle and J. O. Morley, J. Chem. Soc., Faraday. Trans., 1994, 16, 2319; (d ) P. L. Fabre, F. Dumestre, B. Soula and A. M. Galibert, Electrochim. Acta, 2000, 45, 2697. 15 R. West and D. L. Powell, J. Am. Chem. Soc., 1963, 85, 2577. 16 L. Farnell, L. Radom and M. A. Vincent, J. Mol. Struct. (THEOCHEM), 1981, 76, 1.

17 C. Puebla and T. K. Ha, J. Mol. Struct. (THEOCHEM), 1986, 137, 171. 18 T. K. Ha and C. Puebla, J. Mol. Struct. (THEOCHEM), 1986, 137, 183. 19 H. Torii and M. Tasumi, J. Mol. Struct. (THEOCHEM), 1995, 334, 15. 20 G. M. A. Junqueira, W. R. Rocha, W. B. De Almeida and H. F. Dos Santos, Phys. Chem. Chem. Phys., 2001, 3, 3499. 21 M. C. Zerner, ZINDO: A Semi-empirical Program Package, University of Florida, Gainesville, FL 32611. 22 (a) J. Ridley and M. C. Zerner, Theor. Chim. Acta., 1973, 32, 111; (b) M. C. Zerner, R. Loew, R. F. Kirchner and V. T. MuellerWesterholf, J. Am. Chem. Soc., 1980, 102, 589; (c) D. Head and M. C. Zerner, Chem. Phys. Lett., 1986, 131, 359; (d ) W. D. Edwalds and M. C. Zerner, Theor. Chim. Acta., 1987, 72, 347. 23 M. P. Allen and D. J. Tildesley, Computer Simulation of Liquids, Oxford University Press, Oxford, 1987. 24 M. Metropolis, M. N. Rosenbluth, A. H. Rosenbluth, A. H. Teller and E. Teller, J. Chem. Phys, 1953, 21, 1087. 25 Handbook of Chemistry and Physics, CRC Press, Boca Raton, 73rd edn., 19921993. 26 W. L. Jorgensen, BOSS version 3.5, Biochemical and Organic Simulation System. 27 C. M. Breneman and K. B. Wiberg, J. Comput. Chem., 1990, 11, 361. 28 W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey and M. L. Klein, J. Chem. Phys., 1983, 79, 926. 29 K. Coutinho and S. Canuto, Adv. Quantum. Chem., 1997, 28, 89. 30 (a) W. R. Rocha, K. Coutinho, W. B. De Almeida and S. Canuto, Chem. Phys. Lett., 2001, 127, 335; (b) K. J. De Almeida, K. Coutinho, W. B. De Almeida, W. R. Rocha and S. Canuto, Phys. Chem. Chem. Phys., 2001, 3, 1583; (c) K. Coutinho and S. Canuto, J. Chem. Phys., 2000, 113, 9132; (d ) S. Canuto and K. Coutinho, Int. J. Quantum. Chem., 2000, 77, 192. 31 Gaussian 98, Revision A.6, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Cliord, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, and J. A. Pople, Gaussian, Pittsburgh, PA, 1998. 32 K. Coutinho and S. Canuto, DICE: A Monte Carlo Program for Liquid Simulation, University of Sao Paulo. 33 R. West, D. Eggerding, J. Perkins, D. Handy and E. C. Tuazon, J. Am. Chem. Soc., 1979, 101, 1710.

Phys. Chem. Chem. Phys., 2002, 4, 25172523

2523