Solvatochromic Probes for the Determination of Solvent Properties : b-Carotene and 2-Chloro-7-nitro-9 H-fluorene

Dipolaridade/polarizabilidade do solvente (SDP) tem sido calculada a partir dos espectros de UV-Vis de 2-(N,N-dimetilamino)-7-nitro-9H-fluoreno e 2-fluoro-7-nitro-9H-fluoreno. Com base em cálculos teóricos (23 solventes) e dados experimentais (56 solventes), revela-se que 2-cloro-7-nitro-9H-fluoreno (disponível comercialmente) pode ser perfeitamente utilizado para o cálculo dessa propriedade, em vez de 2-fluoro-7-nitro-9H-fluoreno. A divisão de SDP nos seus componentes (dipolaridade (SD) e polarizabilidade (SP) do solvente) requer o uso de um composto sintético poliênico cuja síntese é trabalhosa, envolvendo 15 etapas. Nosso grupo de pesquisa mostrou recentemente que o corante natural b-caroteno pode ser convenientemente utilizado para a determinação de SP, permitindo o cálculo de SD. Utilizando estas sondas solvatocrômicas, SDP, SP e SD foram calculados para uma série de 1-bromoalcanos. Para várias séries homólogas, a dependência de SDP (SD e SP para uma série homóloga) com o número de átomos de carbono nos grupos 1-alquila ou acila foi calculada e discutida.


Introduction
The effects of solvents on chemical phenomena, e.g., reactivity, equilibrium, spectroscopic transitions, are usually discussed in terms of non-specific and specific interactions between the species of interest (reactants, transition states, etc.) and the solvent.A successful approach to address solvent effects is to use a linear combination of the contributing solvent properties, i.e., a solvation free energy relationship, e.g.: Effect of solvent = RC Acidity (solvent acidity) + RC Basicity (solvent basicity) + RC Dipolarity/Polarizability (solvent dipolarity/polarizability) (1)   where RC stands for regression coefficient.There are several versions of equation 1, e.g., that of Kamlet-Taft-Abboud; equation 2, [1][2][3] where RCs and the solvent descriptors are termed (a; b; s) and (a; b; p*), respectively (dd is a correction term for p*); and of Catalán (equation 3), 4 where RCs and the solvent descriptors are termed (a; b; dp) (SA; SB; SDP), respectively.For consistency, the acronym SDP is used instead of SPP employed by Catalán in order to denote solvent dipolarity/polarizability.
Effect of solvent = a (a) + b (b) + s (p* + dd) Effect of solvent = a (SA) + b (SP) + dp (SDP) The cavity term has been dropped from equations 1 to 3, because excitation of the probes employed in the present work, vide infra, obeys the Franck-Condon principle.Equations 2 and 3 are applied by studying the phenomenon of interest in a series of solvents, followed by correlating the solvent effect with its descriptors (SA, SB, etc.).The relative importance of solvent acidity, basicity, etc. is deduced from the values of the calculated (standardized) regression coefficients.The solvent descriptors on the righthand side of equations 2 and 3 are called solvatochromic parameters because their values are determined by using solvatochromic substances, hereafter designated as probes.These are dyes whose spectra, absorption or emission, are especially sensitive to solvent properties. 5,6Figure 1 shows the molecular structures of the probes of interest to the present work.The pair 2-(N,N-dimethylamino)-7-nitro-9H-fluorene (DMANF) and 2-fluoro-7-nitro-9H-fluorene (FNF) are homomorphs (compounds with closely similar molecular structures) and are employed for the calculation of SDP.As shown elsewhere, the polyene (ttbP9) is sensitive only to solvent polarizability; its use allows splitting of SDP into its components (SD and SP). 4 The use of some these probes is associated with problems of cost and availability from chemical suppliers.FNF has been discontinued (Aldrich), although custom synthesis is offered; it can be synthesized by nitration of 2-fluoro-9Hfluorene. 7 Relative to the commercially available ClNF, whose molecular structure is similar to that of FNF, the first choice (custom synthesis) is 10 times more expensive.In addition to the labor involved in the synthesis of FNF from 2-fluoro-9H-fluorene, the cost of the latter is twice that of ClNF.Therefore, the use of ClNF is an attractive alternative Scheme 1. Steps of the synthesis of DMANF.(less cost, no labor) for the determination of SDP, provided that its solvatochromic behavior is similar to that of FNF.The synthesis of ttbP9 is a major synthetic undertaking, involving 15 steps. 8As recently shown, the natural dye b-carotene can be conveniently employed instead of ttbP9 for the determination of SP. 9 The aim of the present work is to evaluate the use of ClNF instead of FNF for the determination of SDP.Both theoretical calculation (23 solvents) and experimental results (56 solvents) showed excellent linear correlations between the frequencies of maximum absorption of both probes, i.e., ClNF is a convenient alternative probe to FNF.The values of SP, hence SD of a series of 1-bromoalkanes, have been determined by using b-carotene and DMANF.The dependence of SDP (also SD and SP of 1-bromoalkanes) on the number of carbon atoms in the alkyl-or acyl chain of the studied liquids was calculated and discussed.

Materials and synthesis
The chemicals were purchased from Alfa-Aeser or Aldrich; the solvents were purified as shown elsewhere, 10 and stored over activated type 4 Å molecular sieves.The purity of the solvents was established by comparing the values of their densities (DMA-4500 digital density meter, Anton Paar), and overall empirical polarity, E T (30) in kcal mol -1 , with literature ones. 3,11ClNF (99% pure) was employed after drying; its elemental analysis calculated for C 13 H 8 ClNO 2 was found to be (element, %): C, 63.5; H, 3.2; N, 5.7; found: C, 63.4; H, 3.0; N 5.5.
The probe DMANF was synthesized in two steps, as shown in scheme 1. 12 First, 2,7-dinitro-9H-fluorene was partially reduced into 2-amino-7-nitro-9H-fluorene (Pd/C catalyst; formic acid; triethylamine).The latter compound, after purification by flash column chromatography (CH 2 Cl 2 eluent; yield 45%), was submitted to reductive amination (reaction with paraformaldehyde) to produce DMANF, 91.0% yield.Elemental analysis calculated for C 15 H 14 N 2 O 2 was: C, 70.85; H, 5.55; N, 11.02; found: C, 70.56; H, 5.65; N, 10.90.The 1 H nuclear magnetic resonance (NMR) spectral data (Varian Inova-300 NMR spectrometer, CDCl 3 ) were similar to those published elsewhere. 12-Vis absorption spectra UV-Vis spectroscopic measurements were carried out on a Shimadzu UV-2500 spectrophotometer, equipped with a model 4029 digital thermometer (Control Company), under the following experimental conditions: each spectrum was recorded at least thrice, at 140 nm min -1 ; slit width 0.5 nm; sampling interval of 0.2 nm.The instrument was routinely checked for λ max accuracy against the known peaks of a holmium oxide glass filter (model 666-F1, Hellma Analytics, Müllheim).Maximum wavelength was determined from the first derivative function, calculated by using a commercial software (GRAMS/32 version 5.10, Galactic Industries); the uncertainty in λ max is ± 0.2 nm.All spectra were recorded at 25 ± 0.1 °C using quartz cells of 1 cm path length.

Calculations
Solvent SDP, SP and SD 4,9 The value of SDP is calculated from equation 4: (4)   where D ñmax;probe;solvent is the difference between the wavenumbers of the longest wavelength (i.e., the solvatochromic) transition of the probes ClNF and DMANF in the solvent of interest i.e., ( ñClNF;solvent -ñDMANF;solvent ); D ñmax;probe;gas is the same difference, but in the gas phase; and D ñmax;probe;DMSO is the corresponding value for (the reference solvent) DMSO.
The experimentally unattainable ñmax;probe;gas (all indicators are solid) is calculated by extrapolation to zero a plot of ñmax;probe;n-alkanes vs. the Lorenz-Lorentz refractive index function f(n) 2 = (n 2 − 1)/(n 2 + 2) (where n = refractive index) for a homologous series of linear alkanes.The values of ñmax;probe;gas are listed at the top of Table S1 in the Supplementary Information (SI) section; the corresponding one for b-carotene is 24402 cm -1 . 9anipulation of the data of DMANF and b-carotene permits the calculation of SD, as shown by equation 5: (5) The plot of ñmax;DMANF;solvent vs. SP b-carotene;solvent is shown in Figure 2, in which the solvents can be grouped into two classes.Group A is located on a straight line; the rest (group B) clearly deviates from linearity.Group A includes gas phase plus alkanes, CCl 4 and CS 2 .These are solvents that influence DMANF only through their polarizability.Group B solvents affect ñmax;DMANF through SD and SP, this leads to deviations from the straight line.Equation 6is obtained from the linear regression of the straight line of Figure 2: where the value ño max;DMANF;solvent represents the frequency that DMANF would have if its interaction with the solvents were governed only by SP.Thus, the difference between ño max;DMANF;solvent and ñmax;DMANF;solvent is proportional to the dipolarity component of this solvent, i.e., its SD.
For group B solvents, the SD value can be obtained from the following equation: (7)   where the strongly dipolar DMSO is taken as a reference.
Therefore, the calculation of SD involves the following steps: determination of (λ max;DMANF;solvent ) of the DMANF charge-transfer band in the solvents of interest, and subsequent calculation of the corresponding frequency ( ñmax;DMANF;solvent ); determination of the same spectral parameters for b-carotene and calculation of the corresponding SP b-carotene from equation 5; calculation of ño max;DMANF;solvent from equation 6; and calculation of SD by application of equation 7.

Geometry optimization of the structures of FNF and ClNF
FNF and ClNF had their geometries, in the gas phase, optimized in two steps, both using a density functional theory (DFT) calculation that employs B3LYP density functional; 6-311G(d,p) and CC-pVDZ basis sets were employed.
For both probes, the transition energies between the fundamental and excited state were calculated by time-dependent density functional theory (TD-DFT), using BOP density functional and DH (Dunning/Hay double zeta) basis set with the application of long range correction.Values of frequencies of maximum absorption ( ñmax ) of the excited states with the largest oscillator strength are shown in Table S2 in the SI section.Solvent effects were simulated by using the polarizable continuum model (PCM) combined with SMD solvation model.The solvation free energy (DG solvation ) was calculated from the difference between the free energy of the probe (ground state) in the solvent and in the gas phase.All calculations were performed by using the version Oct 1 2010 (R3) of Gamess US program package. 13he quality of geometry optimization of both probes can be shown by the agreement between calculated bond lengths and angles with X-ray data of fluorene (to our knowledge, there are no X-ray data for FNF, ClNF or 7-nitro-9H-fluorene), see Table S3 in the SI section.

Results and Discussion
Can ClNF be reliably employed instead of FNF for the determination of SDP?
In order to answer this question, two strategies were employed: (i) theoretically calculated energies of solvation and the ñmax of both ClNF and FNF for a set of representative solvents were compared; (ii) the experimental ñmax values for a large set of solvents were compared.

Theoretical calculations
With regard to (i), the following procedure was employed for each probe: the geometry was optimized in the gas phase; the optimized probes were then solvated in solvents for which parameterized data are available in the employed software package, vide supra.The set chosen covers nonpolar and weakly polar aliphatic and aromatic hydrocarbons, dipolar aprotic solvents, and aliphatic alcohols.DG Solvation and the values of ñmax for both probes were calculated; the results are listed in Table S2 in the SI section.
Equations 8 and 9 (NS = number of solvents investigated; sd = standard deviation; r = correlation coefficient) and Figure 3 clearly show that there are excellent correlations between DG Solvation and ñmax;Theo for both probes in the different solvents, with slopes of practically, or not far from unity.This indicates that the energies of solvation of the two probes are very similar; hence their solvatochromism responses to the same set of solvents correlate linearly.DG Solvation;ClNF;Theo = -0.839+ 1.003 DG Solvation;FNF;Theo (NS = 23; sd = 0.0831; r = 0.9989) ( Comparison of the experimental ñmax;Exp and the calculated SDP values The experimental values of ñmax;FNF;Exp , ñmax;ClNF;Exp and ñmax;DMANF;Exp are listed in Table S1 in the SI section, along with the corresponding SDP.The equations employed for the calculations of these parameters are given in the Calculations item in the Experimental section. The excellent agreement between the solvatochromic responses of ClNF and FNF is evident from equations 10 and 11 and in Figure 4 (for representative solvents).
Both results show that our initial expectation is confirmed, i.e., because of the similar molecular structures, both probes respond similarly to solvent dipolarity/ polarizability, 2-chloro-7-nitro-9H-fluorene (ClNF) is a convenient probe for the determination of SDP.
After establishing that ClNF can be employed instead of FNF, SDP was determined for a series of 1-bromoalkanes that are extensively employed in the synthesis of ionic liquids with increasing hydrophobic character. 14MANF and b-carotene were employed as solvatochromic probes in order to determine SD and SP for these liquids, the results are listed in Table 1.
The dependence of SDP of some homologous series (Table S1 in     The equations employed for the calculations of these parameters are given in the Calculations item of the Experimental section. (v) Ionic liquids: SDP = 0.999 -0.019 Nc (NS = 6, sd = 0.007, r = 0.9946) Except for one case (1-bromoalkanes), the dependence of SDP on Nc is linear; for these (linear) plots, the slopes are negative, i.e., the dipolarity/polarizability decreases as a function of increasing chain length of the compound.The reason for the non-linear dependence of SDP on Nc for 1-bromoalkanes is not readily explainable at present; data for several homologous series should be examined.It is relevant however, that the slopes of the SD and SP vs. Nc plots for this series are negative and positive, respectively.This behavior agrees with chemistry because dipolar interactions are expected to decrease as a function of increasing alkyl-chain length while keeping the dipolar bond (-C d+ -Br d-) constant.On the other hand, the Lorenz-Lorentz refractive index function of this series (a measure of SP) increases as a function of increasing Nc.

Conclusions
The results of theoretical calculations and experimental data clearly lead to the same conclusion: ClNF and FNF respond similarly to solvent dipolarity/polarizability.This similarity, as evidenced by equations 10 and 11, and Figures 3 and 4, argues convincingly that both probes can be employed interchangeably for the calculation of SDP.The commercial availability of ClNF and the green dye b-carotene at reasonable cost simplify the determination of the solvatochromic parameters, a prerequisite for understanding the effects of solvents on chemical phenomena.For homologous series, the dependence of SDP, SD and SP on their molecular structures is mostly linear, showing that the effects of increasing chain length of the alkyl-or acyl group on these solvent properties are additive.

Figure 2 .
Figure 2. Correlation between ñmax;DMANF and SP b-carotene for different media: group A (squares, including gas phase) and group B (triangles).

Figure 3 .
Figure 3. Correlation between theoretically calculated frequencies of maximum absorption, ñmax;probe of FNF and ClNF.

Figure 4 .
Figure 4. Correlation between experimental SDP values of FNF and ClNF.

Table 1 .
Values of SP and SD for homologue series of 1-bromoalkanes a