Structure-reactivity relationships on Michael additions of secondary cyclic amines with 3-cyanomethylidene-2-oxindoline derivatives

2.1 General methods

All solvents used purchased from Sigma-Aldrich are spectroscopic grade and used without further purifications. Melting points were determined on a Stuart SMP3 melting point apparatus and are uncorrected. NMR spectra were acquired on a Bruker Avance 500 instrument (500 MHz) in CDCl₃ using residual solvent signals as internal standards.

The electronic absorption spectra and kinetic determinations were recorded with a conventional UV–Vis spectrophotometer (Shimadzu model 1800 PC), whose temperature of cell compartments were preserved around 20 ± 0.1 °C. All kinetic-runs were performed no less than three times under conditions of pseudo-first-order with the 3-cyanomethylidene-2-oxindoline derivatives 3a-e concentration of ∼1 × 10⁻⁴ mol dm⁻³ and a concentration of secondary cyclic amines 4a-c (piperidine, morpholine and pyrrolidine) in the range 1.0 × 10⁻³–1.0 × 10⁻¹ mol dm⁻³. In all realized experiment, the rates data were found to be reproducible to 2–3%.

2.2 General procedure for the synthesis of 3-cyanomethylidene-2-oxindolines 3a-e

To a mixture containing isatins (1) (1 mmol) and malononitrile or ethyl cyanoacetate (2) (1 mmol) in water (5 mL), MCM‐SO₃H (20 mg) was added and stirred at room temperature for 5–10 min. After completion of the reaction (monitored by TLC), the solid product was dissolved in hot ethanol and then the solid catalyst was removed by filtration. The filtrate was cooled to room temperature to afford the pure product 3a-e as dense red crystals. The physical and spectroscopic data of 3a-e were matched as previously reported (Demchuk et al., 2011; Lashgari et al., 2012). Compound 3e was obtained as a mixture of E-/Z isomers in ratio of 8:1 (Junek et al., 1989).

3.1 Synthesis of the Michael acceptors 3a-e

First of all, five Michael acceptors 3a-e were smoothly synthesized in excellent yields (91–99%) via simple, mild, and efficient Knoevenagel condensation reaction of isatins 1 and malononitrile CH₂(CN)₂ or ethyl cyanoacetate (NCCH₂COOC₂H₅) 2 in aqueous medium in the presence of sulfonated mesoporous silica (MCM‐SO₃H) (Hussein and Ahmed, 2017) as heterogeneous acid catalyst (Scheme 1). The reaction proceeded smoothly at room temperature, in short reaction time (5–10 min) and no byproducts have been detected. After achievement of the reaction (monitored by TLC), the raw product was dissolved subsequently in hot ethanol solution. The precipitate was collected by filtration in which heterogeneous solid catalyst was removed. The obtained pure crystals of products after cooling of the filtrate were directly collected as dark red crystals.

Close

Scheme 1

Synthesis of the Michael acceptors 3a-e.

Export to PPT

3.2 UV–Vis spectra of the reactions of 3-cyanomethylidene-2-oxindoline derivatives 3a-e with secondary cyclic amines 4a-c

The UV/Vis spectra of compound 3a-e, recorded before time scan measurements, and of Michael adducts of products 8-10a-e resulting from reaction of 3a-e with piperidine (4a), morpholine (4b) and pyrrolidine (4c), respectively (Scheme 2), recorded at the end of time scan measurements, are given in Supporting information Section (Figs. S1-S14 and Tables S1-S8). Unfortunately, we did not succeed the isolation of Michael adducts 8-10a-e, because they are unstable at room temperature. The UV/Vis spectra of the adducts of piperidine, morpholine and pyrrolidine 4a-c and compounds 3a-e lack the absorption maximum at ∼340 and 550 nm, which appears in the UV/Vis spectra of 3a-e, and corresponds to the conjugated double bond chromophore.

Close

Scheme 2

Reactions of 3-cyanomethylidene-2-oxindolines 3a-e with secondary cyclic amines 4a-c.

Export to PPT

On the other hand, the peak at ∼380 and 440 nm appears in the UV/Vis spectra of adducts, this decrease in wavelength is due to the reduction of the conjugation accompanying the addition of amine. Analogous behavior has been reported for the reactions of secondary amines with quinone methides (Eq. (1)) (Kanzian et al., 2009) and (E)-4-aryl-4-oxo-2-butenoic acid phenylamides (Eq. (2)) (Cvijetić et al., 2014). In our work, the progress of the reaction can monitored at two wavelengths simultaneously, either through the decrease of absorbance at 340 nm as well as 550 nm or increase of absorbance at 380 nm as well as 440 nm (see supporting information section). In this study, according to absorbance intensity, the rate of the reactions were monitored photometrically from the decays of the absorbance of the electrophiles at a wavelength of 340 nm.

(2)

3.3 Kinetics of the reactions of 3-cyanomethylidene-2-oxindoline derivatives 3a-e with secondary cyclic amines 4a-c

The kinetic investigations were performed under pseudo-first order conditions with at least a 10-fold excess of secondary cyclic amines 4a-c over 2-oxindoline derivatives 3a-e substrate concentration. Experiments were performed by directly mixing a solution of 1.0 × 10⁻⁴ mol/dm³ of 2-oxindoline derivatives 3a-e with the solution of amines 4a-c with concentrations in the range of ([4a-c]_o = 1.0 × 10⁻³ to 1.0 × 10⁻¹ mol/dm³). It therefore appears that all the reactions in this work obeyed first-order kinetics; this undertaken that the pseudofirst-order rate constants (k_obs) were determined from the equation ln(A_∞−A_t) =  − k_obsdt + C. As can be seen for example in Fig. 2, the plot of k_obs versus amine concentration relative to reaction of 2-(2-oxindolin-3-ylidene)malononitrile (3a) with piperidine (4a) showed curves upward (Fig. 2(A)). Similar curved plots are demonstrated for the nucleophilic addition of piperidine on 1-(4-nitrophenoxy)-2,4-dinitrobenzene (Um et al., 2014). Such plots indicate a reaction in which a second molecule of amine acts as a general base catalyst (Um et al., 2007; Um et al., 2012; Crampton et al., 2006). Consequently, one might suggest that the Michael Addition reactions of 3a-e keep through a stepwise-type mechanism with one intermediate (MC^±) as shown in Scheme 2, the process that is, a catalytic route to form anionic σ-adduct MC⁻ from zwitterionic intermediate MC^±. For this reason, it seems reasonable to suggest that addition products 8-10a-e can be formed via the mechanism showed in Scheme 2, involving a nucleophilic addition of secondary cyclic amine and formation of zwitterionic intermediate MC^± in the first step. Thus, under the experimental conditions, the observed first-order rate constant (k_obs) can be expressed by assuming that low concentration of the zwitterions MC^± as follows in Eq. (3) (Ben Salah et al., 2015; Ben Salah et al., 2017).

Close

Fig. 2

Plots of k_obs vs [pip] (A) and k_obs[pip] vs [pip]² (B) for the reaction of 2-(2-oxindolin-3-ylidene)malononitrile (3a) with piperidine (4a) in MeCN at 20 °C.

Export to PPT

Obviously, in agreement with Eq. (3), a good straight line plot (R² > 0.9933) with negligible intercepts were found when the values of (k_obs [NuH]) were plotted versus the square concentration of secondary cyclic amines [NuH]² as shown in Fig. 2(B) and Fig. S15-S20 (see Experimental section), so that Eq. (3) reduces to Eq. (4). A careful examination of these results confirm the proposed mechanism suggesting that the reactions between our 2-oxindoline derivatives electrophiles 3a-e and secondary cyclic amines nucleophiles 4a-c proceed via a nucleophilic addition mechanism, in which the rate-limiting step involve the formation of the N—C bond. Hence, it could be concluded from Eq. (4) that the slopes of the plots (Fig. 2(B) and Fig. S15-S20 (see Experimental section)) induced the second-order rate constants, k₁, which are collected in Table 1.

Table 1 ${\mathit{pK}}_a^{C{H}_{3}CN}$ Values relative to conjugated acid of secondary cyclic amines 4a-c and the obtained rate constants (k₁) for coupling reactions of the 3-cyanomethylidene-2-oxindoline derivatives 3a-e with amines 4a-c.

	pKa = 19.58 a	pKa = 18.92 a	pKa = 16.61 a
	k1 (mol s−1 dm−3)b
3a	2.73 × 10−1	3.22 × 10−2	1.41 × 10−3
3b	2.22 × 10−1	2.72 × 10−2	1.02 × 10−3
3c	4.63	7.85 × 10−1	4.79 × 10−2
3d	3.06	5.16 × 10−1	2.91 × 10−2
3e	9.97 × 10−2	1.12 × 10−2	3.79 × 10−4

^bvalues measured in acetonitrile at 20 °C.

a pK_a data in acetonitrile were taken from (Coetzee and Padmanabhan, 1965).

3.4 Effect of basicity of secondary cyclic amines 4a–c on k₁: Brönsted relationship

Examination of the rate constants (k₁) relative to the couplings of Michael acceptors (3a-e) with secondary cyclic amines 4a–c (values obtained in acetonitrile solution, see Table 1) reveals that the obtained rates relative to this addition-type reactions have been found remarkably sensitive to the effect of basicity of present secondary cyclic amines. On the basis of these findings, the observed effect is illustrated in Fig. 3, which shows clearly the linear dependence of the log k₁ with ${\mathit{pK}}_a^{C{H}_{3}CN}$ corresponding to the conjugated acid of secondary cyclic amines 4a–c ( ${\mathit{pK}}_a^{C{H}_{3}CN}$ values are measured in acetonitrile solution (Coetzee and Padmanabhan, 1965; Hall-Jr, 1957) and given in Table 1). The following Eqs. . (5)–(9) describes this linear Brönsted-type relationship dependence:

Close

Fig. 3

Effect of the secondary cyclic amine basicity on the rate constants (k₁) of the coupling reactions of 3-cyanomethylidene-2-oxindoline derivatives 3a-e with amines 4a–c in MeCN at 20 °C.

Export to PPT

Our experimental β_nuc values of the reactions of 3-cyanomethylidene-2-oxindoline derivatives 3a-e with secondary cyclic amines 4a-c deduced from the slope of log(k₁) = f(pK_a) plots are in the range of 0.63 and 0.77 (Eqs. (5)–(9)) proves comparable and in accord to those reported in literature for some nucleophilic addition reactions in which a major diagnostic feature of reaction mechanism that the first step is a rate-limiting step that corresponding to nucleophilic attack step (Um et al., 2012; Castro et al., 1997; Bordwell and Hughes, 1986; Dixon and Bruice, 1972; El Guesmi et al., 2010). As mentioned for example for the coupling reactions of secondary cyclic amines with benzoselenadiazolium in MeCN at 20 °C when obtained β_nuc value is 0.55 (Ben Salah et al., 2017), also, and in a same condition, similar value of β_nuc = 0.52 has been signalized for addition reactions of secondary cyclic amines on 1-iodo-2,4-dinitrobenzene (Um et al., 2012), while for the reactions of a bis(4-nitrophenyl)thionocarbonates with various alicyclic amines in water at 25 °C, β_nuc is 0.50 (Castro et al., 1997).

Moreover, it is interesting to note in the present work that the value obtained of β_nuc is remarkably higher than β_nuc value of 0.41 previously reported for the reaction of benzothiadiazolium cation with various substituted anilines in acetonitrile solution at 20 °C (Ben Salah et al., 2015). We note particularly that the β_nuc value of 0.63 < β_nuc < 0.77 measured for the coupling of 3-cyanomethylidene-2-oxindoline derivatives 3a-e with amines 4a-c using in this work can be properly interpreted for in terms that the initial Nitrogen-Carbon coupling is rate-limiting step.

Furthermore, the linearity obtained in Brönsted plots (Fig. 3) for the secondary cyclic amines (4a-c) informs us on the same reaction mechanism for all reactions in this study. More importantly, the Brönsted coefficient of 0.63 < β_nuc < 0.77 obtained in this work which represent the degree of bond formation in transition state indicates that the nitrogen–carbon bond formation for coupling reactions Amines/Michael acceptors is more than half complete.

3.5 Mayr Relationship: Quantification of reactivity electrophile in term of E parameter of Michael acceptors 3-cyanomethylidene-2-oxindoline derivatives 3a-e

In accordance with numerous kinetic investigations, it has been demonstrated that it is possible that the rate constants for the reactions of some electrophiles such as, carbocations and related quinone methides with n-, σ- and π-nucleophiles can be depicted by an Eq. (10) with three parameters, wherein E is an electrophile dependent parameter and s and N are nucleophile dependent parameters (Mayr and Ofial, 2005; Ofial and Mayr, 2004; Mayr and Ofial, 2004; Mayr et al., 1998a; Mayr et al., 1998b).

Subsequently, nucleophilicity parameters for numerous strong nucleophiles has been assigned, such as carbanions (Phan and Mayr, 2006; Bug et al., 2004; Lucius et al., 2002; Lucius and Mayr, 2000), aliphatic and aromatic amines (Minegishi and Mayr, 2003; Brotzel et al., 2007a; Brotzel et al., 2007b; Brotzel and Mayr, 2007), silyl enol ethers (Dilman and Mayr, 2005; Burfeindt et al., 1998; Patz and Mayr, 1993), enamines (Kempf et al., 2003), and ketene acetals compounds (Tokuyasu and Mayr, 2004) through employing reference electrophiles such as, substituted diarylcarbenium ions and structurally related quinone methides. Consequently, this makes it very reasonable to consider Eq. (10) is also suitable to describe the reactions of electron deficient arenes as well as ordinary Michael acceptors (Seeliger et al., 2007; Berger et al., 2007; Lakhdar et al., 2007; Terrier et al., 2005; Lemek and Mayr, 2003; Remennikov et al., 2003).

In this regard, much attention has been given to Michael-type acceptors involving the initial nucleophilic attack step at the electron deficient double bond. For this purpose, Bernasconi reported the studies of the kinetics of the reactions of various nucleophiles such as, carbanions, amines and alkoxides towards some Michael acceptors, e.g., benzylidene Meldrum’s acids and benzylidene indandiones in DMSO/H₂O mixtures (Bernasconi et al., 1998; Bernasconi, 1989; Bernasconi and Panda, 1987; Bernasconi and Murray, 1986; Bernasconi and Leonarduzzi, 1982; Bernasconi and Fornarini, 1980; Bernasconi and Leonarduzzi, 1980). On the other hand, Rappoport have focused mainly on reactions of chloro-substituted benzylidene Meldrum’s acids on which follow nucleophilic vinyl substitutions (Ali et al., 2006; Bernasconi et al., 1998; Rappoport, 1992). In view of recent investigations, other Michael acceptors have been reported (Oh and Lee, 2002; Oh et al., 2003) during the studies of the kinetic of the addition of benzylamines to substituted benzylidene-acetylacetones in acetonitrile.

3.6 Correlation analysis

It thus appears that Eq. (10) is valid for the reactions of the 3-cyanomethylidene-2-oxindoline derivatives 3a-e with secondary cyclic amines 4a-c since the plots of (log k₁)/s vs. N exhibit slopes very close to 1.0. The five parallel correlation lines in Fig. 4 describe this linear tendency (Eqs. (11)–(15)).

Close

Fig. 4

Relationships between (log k₁)/s and nucleophilicity parameters N for the reactions of the 3-cyanomethylidene-2-oxindoline derivatives 3a-e with secondary cyclic amines 4a–c in MeCN at 20 °C.

Export to PPT

Support for this measure of E parameters for 3a–e is provided by calculation through least squares minimization method [Δ² = Σ(log k – s(N + E))²] using parameters k₁, N and s given in Table 2.

Table 2 Second-order rate Constants, k₁, for coupling reactions of Michael acceptors 3-cyanomethylidene-2-oxindoline derivatives 3a-e with used amines 4a-c in MeCN at 20 °C.

	N = 18.64/s = 0.60	N = 17.35/s = 0.68	N = 15.65/s = 0.74
	log k1
	−0.564 (E = −19.58)	−1.492 (E = −19.54)	−2.851 (E = −19.50)
	−0.654 (E = −19.73)	−1.565 (E = −19.65)	−2.991 (E = −19.69)
	0.665 (E = −17.53)	−0.105 (E = −17.50)	−1.320 (E = −17.44)
	0.486 (E = −17.84)	−0.287 (E = −17.78)	−1.536 (E = −17.73)
	−1.001 (E = −20.31)	−1.951 (E = −20.22)	−3.421 (E = −20.27)

Importantly, Fig. 5 can be employed as a tool for comparing the electrophilicities of the 3-cyanomethylidene-2-oxindoline derivatives 3a-e with some reference electrophiles. It thus appears that the electrophilic reactivity of our five Michael acceptors 3a-e cover a range of almost three orders of magnitude from –17.5 > E > –20.3 (Fig. 5), this seems in good agreement with measured empirical electrophilicity parameters E for series of monoacceptor-substituted ethylenes (H₂C = CH-Acc) and styrenes (PhCH = CH-Acc) recently reported by Mayr group (Allgäuer et al., 2017). It is noteworthy that the electrophile 3e is roughly two orders of magnitude of E parameter less reactive comparing than their analogues 3c and 3d. Thus, this represents a nice reflection of a great effect on the electrophilic reactivity of these Michael-type acceptors when one ester groups was fixed on 3d. Additional, Fig. 5 that also summarize substrates previously studied by Mayr and co-workers, unfolds that the most reactive Michael acceptor studied in this work, i.e. (3c) (E = −17.50) exhibits an electrophilic reactivity that compares well with those of diethyl 4-nitrobenzylidenemalonate (E = −17.67), of (E)-3-(4-cyanophenyl)-1-phenyl-prop-2-en-1-one (E = −17.64) and of (E)-3-(4-nitrophenyl)-1-phenyl-prop-2-en-1-one (E = −17.33) indicating a normal electrophilic ranking of 3c in the Mayr’s scale. It is apparent nevertheless that this value of E remains lower than that of p-substituted-benzylidenemalononitril (−9.32 < E < −13.30), of 5-benzylidene-1,3-dimethylbarbituric and -thiobarbituric acids (−9.15 < E < −13.97) and of 2-(p-substituted-benzylidene)-indan-1,3-dione (−10.11 < E < −13.56). On the other hand it’s clear that E values of our neutral Michael acceptor 3a-e remains relatively lower in term of electrophilic reactivity than that of 4,6-dinitrotetrazolopyridine (E = −4.67) which represent one of the neutral compounds with the most electrophilic character known to date. In undertaking the works of Mayr; it is shown recently that 2,3-dichloro-5,6-dicyano-p-benzoquinone (DCQ) represent at the moment, the neutral compound the most electrophile with an value of E of −3.66. A significant feature of 3a-e that the values obtained of our Michael-type E are considerably higher and thus more electrophile than diethyl 4-substituted-benzylidene malonate (−20.55 < E < −23.80). To be noted is that comparison of individual pairs of electron-deficient Michael acceptors substrates emphasizes that our E values quite similar to those measured for (E)-3-(4-substituted-phenyl)-1-phenyl-prop-2-en-1-one (−17.33 < E < −19.39).

Close

Fig 5

Comparison of the electrophilicity parameters E of 3-cyanomethylidene-2-oxindoline derivatives 3a-e with those of electron-deficient neutral electrophilic substrates.

Export to PPT

Comparison of the data obtained in this study makes it reasonable to consider that the substitution of one cyano group (—CN) in the electrophile 3d by a (—CO₂Et) group to provide 3e is accompanied by a decrease of ∼3 E unit in the electrophilic reactivity. It follows such an E value remains markedly lower than that caused by the replacement of a cyano group (—CN) by a second CO₂Et group into benzylidenemalononitriles to form benzylidenemalonates in which E values are in the range of 9.06–10.48. As an illustration of this behavior can be presumed mainly as reported by Mayr and co-workers (Kanzian et al., 2009) in terms of the reflection of an augmentation of stabilization of the ground state through resonance of benzylidenemalonates that possessing two carbonyl groups (—CO₂Et) (resonance structure (I)) compared to their analogous Michael-type acceptors possessing only one carbonyl group (resonance structures (II) and (III)).

Close

Export to PPT

3.7 The electrophilicity ω index Parr

From a theoretical point of view, a comprehensive of the accurate nature of a chemical reactivity descriptor described in terms of the electrophilicity concept has seduced the attention of many quantum chemists (Pérez et al., 2003; Pérez et al., 2002; Parr et al., 1999; Maynard et al., 1998; Roy et al., 1998). These authors have clearly emphasized that the theoretical scales of electrophilicity are based essentially on descriptors of the electronic structure of molecules in its ground state. It therefore appears that these indexes are considered as static reactivity indexes and consequently forming reactivity absolute scales, without however depending on the nucleophile partners.

Indeed, based on the work of Maynard (Maynard et al., 1998) and from the knowledge of Density functional theory (Parr and Yang, 1989), a theoretical definition of electrophilicity ω has been quite successful introduced by Parr and co-workers (Parr et al., 1999), which can be expressed as a measure of the energy stabilization of an electrophilic molecule upon receiving an extra amount of electronic charge from the environment. Consequently, the global electrophilicity ω index may be recast and given into the simple expression:

In Eq. (16), μ ≈ (ε_H + ε_L)/2 and η ≈ (ε_L − ε_H) reactivity descriptors are the electronic chemical potential and the chemical hardness, respectively. Note that, the calculation of reactivity descriptors can be approached by the frontier HOMO and LUMO energies.

Calculated values μ, η, and ω of relevance for this work are given in Table 3. So, 3-cyanomethylidene-2-oxindoline derivatives 3a-e acts as an electrophilic molecule owing to the relatively great value of its ω (3.742 < ω < 4.692). So, starting from the reference compound 2-(2-oxindolin-3-ylidene)malononitrile (3a) (ω = 4.230 eV), the N-alkylation with ethyl group (–CH₂CH₃) result in an electrophilic deactivation in compounds 2-(1-ethyl-2-oxindolin-3-ylidene)malononitrile (3b) (ω = 4.164 eV). At the same time, the highest activation effect is produced by introduction of electron withdrawing (–Cl) in compound 2-(5-chloro-2-oxindolin-3-ylidene)malononitrile (3c) (ω = 4.692 eV). In this series, the highest electrophilic deactivation was caused by the substitution of (—CN) group by (—CO₂CH₂CH₃) groups in compound ethyl 2-(5-chloro-1-methyl-2-oxindolin-3-ylidene)-2-cyanoacetate (3e) (ω = 3.742 eV).

Fig. 6 summarizes the comparison between the global electrophilicity index ω calculated using Parr model and the electrophilicity E estimated experimentally using Mayr model. The resulting regression equation for the compounds 3a-d presenting 2 cyano group is:

Close

Fig 6

Correlation between experimental E and theoretical ω electrophilicity parameters and gas phase lowest unoccupied molecular orbital energies (ε_LUMO).

Export to PPT

Excellent linear relationship is obtained for 3a-d between both variables (regression coefficient R = 0.997), thereby suggesting that the electrophilicity index ω may be used to properly evaluate the chemical substitution effect on the electrophilic potential of compounds, and consequently it can be further considered as a trustworthy descriptor of reactivity within this class of electrophiles.

On the other hand and as can be seen in Fig. 6, significant deviation is observed for compound 3e from experimental value, this can be explained as previously reported by Bernasconi (Bernasconi, 1989; Bernasconi and Murray, 1986; Bernasconi and Stronach, 1991; Bernasconi, 1987; Bernasconi and Kanavarioti, 1986) and Oh (Oh et al., 2000; Oh and Lee, 2002; Oh et al., 2003) that the transition states of amine additions to 3e are stabilized by hydrogen bridging from NH to the carbonyl group (six membered transition state TS₆). This results are in good agreement with those reported by Mayr and coworkers for similar system, the additions of secondary amines to 2-benzylidene-indan-1,3-diones 11 (Berger et al., 2007), benzylidene-Meldrum's acid 12 (Kaumanns and Mayr, 2008) and benzylidenebarbituric and -thiobarbituric acids 13 (Seeliger et al., 2007). In line with this observations, Fig. 6 in this work shows also that correlation obtained when using LUMO energies present the same deviation (see Scheme 3).

Close

Scheme 3

Addition of a secondary amine to ethyl 2-(5-chloro-1-methyl-2-oxindolin-3-ylidene)-2-cyanoacetate (3e) (TS₆: six membered transition state, T^±: zwitterionic intermediate).

Export to PPT

Comparisons with literature data shows that the experimental rate constants are generally 20–120 times greater than the calculated values. Though deviations from experimental values by two orders of magnitude are still within the confidence limit of Eq. (10) (Mayr et al., 2001; Mayr et al., 2003). Our study has shown that additions of secondary amines to 3e are only six times faster than predicted by Eq. (10). It has, therefore, been concluded that H-bridging as indicated in TS₆ cannot have a large effect on the transition states of the additions of secondary amines to 3e.

Close

Export to PPT