Differentiation Between the like and unlike Isomers of Dimethyl 3,4-di(p-anisyl)adipate Using 1H NMR Spectroscopy

The electrohydrodimerization 1 of olefin derivatives is a versatile reaction, which has attracted considerable interest in recent years. In the course of an investigation 2 on the electrochemical reduction of substituted cinnamic esters, we have prepared and isolated both diastereomers of dimethyl 3,4-di( p anisyl)adipate ( dAA , Figure 1). At that point, however, we still had the problem of discerning the unlike3 (meso) compound from the like3 isomer racemate. As outlined below, we have accomplished this objective by employing well-established procedures but, to our knowledge, this approach has never been applied to the stereochemical analysis of open-chain molecules, such as substituted adipic acid derivatives.


Introduction
The electrohydrodimerization 1 of olefin derivatives is a versatile reaction, which has attracted considerable interest in recent years.
In the course of an investigation 2 on the electrochemical reduction of substituted cinnamic esters, we have prepared and isolated both diastereomers of dimethyl 3,4-di(panisyl)adipate (dAA, Figure 1).At that point, however, we still had the problem of discerning the unlike 3 (meso) compound from the like 3 isomer racemate.As outlined below, we have accomplished this objective by employing well-established procedures but, to our knowledge, this approach has never been applied to the stereochemical analysis of open-chain molecules, such as substituted adipic acid derivatives.

Results and Discussion
After acquisition of the 300 MHz 1 H NMR spectrum from each sample (in CDCl 3 , at room temperature), we noticed that although both presented the expected 4 AA'BB'XX' second-order sub-spectra (Figures 1, 2a and 3), besides the methoxy singlets and the signals due to the aromatic protons, only one of them (henceforward denoted as Sample I) presented this AA'BB'XX' sub-spectrum sufficiently well resolved for further analysis (Figure 2a).
Hence, after having measured J(AX) = J(A'X') = 8.55 Hz, J(BX) = J(B'X') = 6.20 Hz and δ(X) = 3.411, by assuming that 2a), the AA'BB' part of the AA'BB'XX' sub-spectrum is amenable to direct solution 5 , on taking the centers of the doublets (due to coupling to either the X or the X' nuclei) as the four lines of an AB system, this procedure yielding: δ(A) = 2.548, δ(B) = 2.688 and J(AB) = J(A'B') = 15.43Hz.Using the above data, together with the measured positions of the 15 observed lines, as input for the program LAOCOON 3 6 , we have obtained a very good fit (Figure 2b, RMS Error = 0.0012 Hz) for the 208 calculated lines, together with the calculated value for J(XX') = 5.06 Hz.This value was, then, converted to a 46° dihedral angle (Hx-C-C-Hx') for Sample I, by using the Altona Equation 7(a generalized version of the Karplus Equation).Now, to be able to distinguish between the diastereomers of dAA, an acyclic compound, two conditions must be met: each isomer must have only one strongly preferred rotamer, and these two rotamers must present different H x -C-C-H x' dihedral angles.

O Me
Thus, we have performed Molecular Mechanics calculations (MM+, Hyperchem 3) and found that both above conditions are met, each diastereomer strongly favoring the rotamer with distal p-anisyl groups, represented in Figure 4, below.
Hence we concluded that Sample I is the l-dAA.It should be pointed out that this assignment was only possible because these compounds present second-order sub-spectra: were it otherwise, it would be impossible to obtain the values of the coupling constants between the protons on each side of the relevant symmetry element.
It should be added that X-ray Diffraction analysis results 8 on a crystal obtained from Sample II shows it to be the unlike isomer, thus confirming our conclusion regarding Sample I. Furthermore, the X-ray data show a dihedral angle (CH 2 -C-C-CH 2 ) of 177.51° for the u-dAA (in a monocrystal), which is near enough the 171.50° value our MM+ calculations yielded for the same angle (for one molecule alone in the vacuum) and thus confirms both the accuracy of our calculations and the very strong preference presented by dAA isomers for the rotamers with distal p-anisyl groups.

Materials
Deuterochloroform and methyl p-methoxycinnamate were used as received from Aldrich, after being checked for purity.Both isomers of dimethyl 3,4-di(p-anisyl) adipate (dAA) were simultaneously formed by cathodic reduction of methyl p-methoxycinnamate in methanol, using either platinum or mercury cathodes and TEAB as support electrolyte, as described elsewhere 2,9,10 .The isomers of dAA were subsequently resolved and purified by column chromatography on silica-gel 60.The complete assigned 1 H and 13 C NMR dataset for these compounds is presented in Table 1, below.The HETCOR ( 1 H-13 C) experiments were performed in order to assure the internal consistency of the assignments here presented.

Instruments and Methods
1 H and 13 C NMR spectra were recorded on a Bruker Avance DPX 300 instrument.Standard microprograms from Bruker Software Library were employed.All measurements were performed in 5 mm o.d.tubes, using a deuterium lock, at 20 o C. Samples were prepared by dissolving ca. 5% v/v of each compound in 0.5 mL of CDCl 3 , containing 0.01 % v/v of TMS as internal standard. 1H spectra (300.13MHz) were acquired with a sweep width 2250 Hz, corresponding to a final digital resolution 0.137 Hz per data point.32 scans were accumulated, using a pulse duration of ca.15º, with an acquisition time ≥ 7.4 s and a 10.0 s relaxation delay.
Broadband 1 H decoupled 13 C spectra (75.47 MHz), were acquired with a sweep width ≥ 8000 Hz, corresponding to a final digital resolution ≥ 0.488 Hz per data point.64 scans were accumulated, using a pulse duration of ca.36º, with an acquisition time ≥ 1.0 s and no relaxation delay.Raw data were zero-filled and Fourier transformed under matched-filter conditions.
HETCOR ( 1 H-13 C) experiments 11 were performed at 7.05 T, using a low decoupler power in CW mode (1.5 W) and composite pulse decoupling (CPD) with polarization transfer from 1 H to 13 C.The following pulse sequence was employed: where: D1 = 2 s; D3 = 4 ms; D4 = 2 ms; D0 = (3 + 178.6 n) µs and 0 ≤ n ≤ 512.The FIDs for the 2D experiments were acquired with ca.8000 Hz of sweep width in F2; 32 scans were accumulated, with an acquisition time of 1.5 s; 512 such experiments were performed with the evolution time incremented so as to provide an effective sweep width of ca.2800 Hz in F1.The delays D3 and D4 were chosen to show correlations with peaks of all multiplicities by assuming a J(CH) of ca.125 Hz.The final data matrix was 512 x 4 Kbytes.Raw data were zero-filled in F1 and a gaussian window function was applied in both F1 and F2 prior to Fourier transformation.

Figure 1 .
Figure 1.Representation of dAA, highlighting the protons of the AA'BB'XX' sub-system.

Figure 2 .
Figure 2. The experimental (a) and simulated (b) AA'BB'XX' subspectra (at 300.13 MHz) of l-dAA (the signals marked with * are due to ca. 10% of u-dAA present in the sample, while that marked with x is due to an unidentified inpurity).

Table 1 .
Assigned NMR Data from the Isomers of Dimethyl 3,4-Di(p-Anisyl)Adipate a) 300.13 MHz 1 H NMR δ (multiplicity) of dAA Isomers in CDCl 3 (TMS = 0) Positions relative to the MeO group in the anisyl moiety; b Attached to the p-anisyl moiety; c See Figure1; d J = 5.79 Hz; e The coupling constants of this AA'BB'XX' system were not determined due to insufficient resolution; f J = 5.94 Hz; g For the coupling constants of this AA'BB'XX' system see Table2.Positions relative to the MeO group in the anisyl moiety; b Attached to the p-anisyl moiety.
a a

Table 2 .
Values of the 1 H NMR Scalar Coupling Constants for l-dAA.