1,8-Bis(dimethylamino)naphthyl-2-ketimines: Inside vs outside protonation

Introduction

It is well known that the extremely high basicity of 1,8-bis(dimethylamino)naphthalene (1, DMAN; pK_a = 12.1 in H₂O [1,2], 18.62 in MeCN [3], 7.5 in DMSO [4]) and its derivatives originates from the electrostatic repulsion between the unshared electron pairs of nitrogen atoms, which strongly destabilises the neutral base. This repulsion is additionally fortified by the steric inhibition of resonance (both NMe₂ groups cannot be conjugated to the aromatic system at the same time) preventing charge delocalisation. The protonation results in the formation of the N–H···N bonded cation 1H⁺, the removal of the electrostatic and steric strain and thus a considerable free energy gain (Scheme 1) [5]. This is also a reason why the vast majority of DMAN derivatives are protonated to the internitrogen space, even if other centres of basicity are present in the molecule. The only exceptions are compounds 2 and 3 which are protonated to aza- and carbonyl groups, respectively (Scheme 2) [6]. This unusual protonation site originates from the conjugation between the C=N (C=O) and the NMe₂ groups, leading to the formation of stabilised cations.

In our recent work, we observed for the first time that the ortho-ketimino group in 4a can compete with the proton sponge moiety for the proton, which results in the formation of equilibria between the forms 4b (protonated to the proton sponge fragment) and 4b’ (protonated to the imino group) in the DMSO solution (Scheme 3) [7]. The aim of the present work is to investigate the major factors influencing the effective basicity of the imino functionality. As primary objects for this study, we consider a series of imines 4a–7a (Scheme 4), which differ by the number and position of an OMe group in an aryl substituent. We hope that increasing the number of strong electron-donating groups will result in a significant transfer of basicity to the imino function. Selected series of neutral molecules 4a–7a, as well as their mono- (4b–7b) and diprotonated (4c–7c) species were investigated by ¹H NMR spectroscopy in DMSO-d₆, CD₃CN and acetone-d₆ in a wide range of temperatures. Results obtained in solutions were compared to single crystal X-ray structures and calculated equilibrium geometries of isolated molecules in a vacuum.

Results and Discussion

Synthesis of DMAN-ortho-ketimines and their cations

Target compounds 4a–7a were synthesised by techniques previously developed in our laboratory (Scheme 4) [7,8]. Compounds 4a–6a were obtained by the treatment of the ortho-lithium derivative 8 with the corresponding nitrile (path a). Unfortunately, the strong election-donating effect of two and more OMe groups makes this approach not effective for the synthesis of compounds 6a and 7a. To overcome this difficulty, we used a reversed method based on the treatment of ortho-cyanide 9 with the corresponding aryllithiums, which produces the compounds 5a–7a with good to moderate yields (path b). The monocations 4b–7b and dications 4c–7c were prepared by the treatment of the corresponding imines with one or two equivalents of HBF₄ in Et₂O (Scheme 5). The so obtained tetrafluoroborates were used for NMR measurements after recrystallisation from EtOH.

Structural investigations

The structural investigations started with a ¹H NMR analysis of the imines 4a–7a in acetone-d₆, CD₃CN and DMSO-d₆ in a wide range of temperatures. The main attention was paid to the chemical shift and the shape of the C=NH and NMe₂ groups’ signals. Upon temperature decrease, the NH signal splits into two peaks of unequal intensity, which is best seen in the case of the acetone-d₆ solution (Figure 1). For example, for compound 4a, the C=NH proton signals are at δ = 9.9 ppm and 10.7 ppm at −80 °C, with the major signal being in the higher field. In contrast, in the spectra of 5a–7a, bearing OMe groups in ortho-position to the imino group, the major NH signal is located in the lower field (Figures S2–S4 in Supporting Information File 1). We assigned these signals to the NH^a and NH^b protons of Z- and E-isomers, as shown in Scheme 6. We believe that in solution the Z-form is more stable for 4a, while the E-form is more stable for 5a–7a. Indeed, the X-ray analysis shows that compound 6a in the solid state exists in the E-form stabilised by the N−H···O intramolecular hydrogen bond (IHB) with an ortho-OMe group (Figure 2a), while compound 4a, as it was revealed in our previous work [7], exists exclusively as a Z-isomer (Figure 2b). Moreover, gas phase quantum chemical calculations show that the E-form is preferable for 5a (ΔE = E_Z − E_E = 1.66 kcal/mol), 6a (ΔE = 1.19 kcal/mol), and 7a (ΔE = 1.90 kcal/mol), while the Z-form dominates for 4a (ΔE = −2.62 kcal/mol, Figure 3). The contribution of the N−H···O IHB to the stabilisation of the E-forms of 5a–7a appears to be small: in the ¹H NMR spectra, the NH^a signal is shifted to the low field by less than 1 ppm in comparison with Z-4a (Figure 1) [9]; in a DFT calculation, the mutual orientation of C=NH and the OMe groups is not optimal for IHB formation.

The rate and equilibrium constants of this uncatalysed isomerisation [10] depend on the solvent and temperature. By comparing the results obtained in different solvents at equal temperatures, one can see that the H^a H^b exchange rate is noticeably lower in DMSO-d₆ (Figures S9–S12 in Supporting Information File 1) than in acetone-d₆ and CD₃CN (Figures S5–S8 in Supporting Information File 1), while the chemical shift differences between the H^a and H^b signals remain similar in all cases (ca. 1 ppm). The increased energy barrier for the isomerisation could be attributed to the stabilisation of the Z-form through formation of a stronger hydrogen bond between the DMSO molecule and the NH^b proton [11], which is sterically unhindered and accessible for the solvent. Indeed, X-ray data and quantum chemical calculations show that phenyl and naphthalene rings are almost perpendicular (Figure 2 and Figure 3).

After the first protonation of imines 4a–7a with HBF₄, the ¹H NMR spectra display two sets of signals (Figure 4). One of the sets corresponds to the monocations 4b–7b protonated at the proton sponge site: the NHN bridging proton resonates at ≈19 ppm and the positions of the CH signals of the dimethylaminonaphthalene fragment are typical and characteristic for protonated proton sponges [5]. In the other set of signals, the chemical shifts of the NMe₂ groups are practically the same as for the neutral bases 4a–7a. This means that protonation occurs at another basic site, which can only be imine nitrogen, i.e., the structure corresponds to monocations 4b’–7b’ (Scheme 7). Indeed, around 10 ppm, a new signal appears which corresponds to two NH protons. To our knowledge, this is the first example of an imine to be basic enough to successfully compete for a proton with the proton sponge moiety: for example, the basicity of benzophenone imine (pK_a = 6.82) and even that of cyclohexanone imine (pK_a = 9.15) [12] is much lower than that of DMAN (pK_a = 12.1) [1,2]. Moreover, as it was already mentioned in the introduction, this is the third known example of a DMAN derivative with a substituent that is more basic than the peri-dimethylaminonaphthalene core. Such an imine basicity boost can originate from the fact that upon protonation, NMe₂ groups lose conjugation with the ring whereas the C=N moiety does not. As a result, the imino nitrogen saves the electron density supply from both OMe and NMe₂ groups, which increases the s-character of nitrogen hybridisation thus fortifying the basicity [13].

The ratio of two monocations, estimated from ¹H NMR spectra at room temperature, depends on the position and number of the OMe groups and the solvent (Table 1; the values are practically temperature independent). One could expect that strong electron-donating OMe groups being placed in positions 2, 4 or 6 of a phenylimino substituent should increase the basicity of imine nitrogen. Thus, in CD₃CN, both 4b and 5b, containing only one OMe group, show no proton transfer towards the imino function (Figure S13 in Supporting Information File 1). The addition of a second OMe group in the case of 6 leads to the formation of 25% 6b’. Upon the insertion of a third OMe group in 7, the amount of 7b’ expectedly increases to 33%. Switching of the solvent to acetone-d₆ facilitates the proton transfer away from the proton sponge moiety (Figure S14, Supporting Information File 1). For example, the relative amount of 7b’ increases from 33% to 66%. Using DMSO-d₆ leads to the dramatic shift of the equilibrium, resulting in the formation of 90% 7b’ (Figure 4). We believe that the forms protonated at the imino function are additionally stabilised by hydrogen bonds with solvent molecules and thus the amount of b’ forms correlates with the proton accepting ability of the medium [11]. The gas phase calculations show that without any additional interaction with the medium, the forms protonated at the imino group are lowest in energy for the imines 4, 6, 7 (Figure 5). In the solid state, no proton transfer to the imino group was observed: compounds 4a·HClO₄ and 6b·EtOH crystallise in forms protonated at the proton sponge moiety (Figure 6).

We expected that the proton transfer to the C=N group will force the nearest proton sponge moiety to occupy the in,out-conformation [14] with the 1-NMe₂ group participating in NHN intramolecular hydrogen bonding with the C=NH₂⁺ group (Scheme 8). Indeed, in the case of 7b’, which contains three OMe groups, we have found direct evidence of this process: the 1-NMe₂ group signal shifts from 2.8 to 3.3 ppm – the characteristic region for protonated dimethylamino groups in proton sponges [5] – while the 8-NMe₂ group signals stay at 2.8 ppm. Thus, only the 1-NMe₂ group is protonated, which is only possible if it is out-inverted. Similar chemical shifts were observed for the cation 10 (Scheme 8, right), which is out-protonated due to steric reasons [15]. In refs. [15-17] it was shown that hydrogen bonding to an out-inverted NMe₂ group without the proton transfer does not noticeably change the chemical shift of methyl protons. Thus, in the case of 7b’, the chemical shift at 3.3 ppm proves the proton transfer to the amino nitrogen, as shown in Scheme 8 for (in,out)-7b’. In other words, in the case of 7b’, steric pressure of the trimethoxyphenyl substituent facilitates the 1-NMe₂ group’s out-inversion which is stabilised by an IHB, forming N(sp³)/N(sp²) type of proton sponges, which were recently reported in our paper [18]. It was previously shown that the in,out-conformation can be stabilised by placing the following functionalities in position 2(7): tertiary alcoholic groups which are able to form the O–H···N IHB [16,17,19], bulky substituents rendering a steric pressure onto the 1-NMe₂ group [7,20], or a metal atom coordinating the NMe₂ group as a ligand causing its inversion [21] (Scheme 9). A possibility of out-protonation of 1,8-bis(dimethylamino)naphthalene was discussed in our recent paper [15], while it is experimentally demonstrated here for the first time.

We also prepared dications 4c–7c (Scheme 5) and investigated their structure in solution in CD₃CN, DMSO-d₆ and acetone-d₆. It was found that the exchange rate between protons in the C=NH₂⁺ group strongly depends on the position and number of OMe groups and the temperature. In the case of 4c, signals are coalescent even at −80 °C (Figure 7). The presence of an OMe group in the ortho-position to C=N groups in 5c slows down the exchange and results in NH₂⁺ signals splitting at −40 °C. This can be attributed to the formation of the weak NHO intramolecular hydrogen bond, which becomes stronger upon inserting an additional OMe group. Such a NHO bond is clearly visible in the X-ray structure of 6c (Figure 8). As a result, in solutions of 6c and 7c, NH₂⁺ proton signals are observed separately up to 40 °C (for spectra in CD₃CN see also Figures S20–S23 in Supporting Information File 1).

Conclusion

In summary, we succeeded in preparing a new series of 1,8-bis(dimethylamino)naphthalene-ortho-ketimines with aromatic substituents containing strong electron donating OMe groups. Neutral compounds as well as their mono- and diprotonated species were studied. The obtained compounds demonstrated increased basicity of the imino nitrogen, which was able to effectively compete for the proton with the proton sponge core. Increasing the number of OMe groups facilitated the proton transfer to the imino function by up to 90%. As a result, studied compounds can be considered as superbasic imines. In one case, we were able to experimentally observe the simultaneous out-inversion and protonation of the 1-NMe₂ group for the first time.