Catalyst-Solvent System for PASE Approach to Hydroxyquinolinone-Substituted Chromeno[2,3-b]pyridines Its Quantum Chemical Study and Investigation of Reaction Mechanism

The Pot, Atom, and Step Economy (PASE) approach is based on the Pot economy principle and unites it with the Atom and Step Economy strategies; it ensures high efficiency, simplicity and low waste formation. The PASE approach is widely used in multicomponent chemistry. This approach was adopted for the synthesis of previously unknown hydroxyquinolinone substituted chromeno[2,3-b]pyridines via reaction of salicylaldehydes, malononitrile dimer and hydroxyquinolinone. It was shown that an ethanol-pyridine combination is more beneficial than other inorganic or organic catalysts. Quantum chemical studies showed that chromeno[2,3-b]pyridines has potential for corrosion inhibition. Real time 1H NMR monitoring was used for the investigation of reaction mechanism and 2-((2H-chromen-3-yl)methylene)malononitrile was defined as a key intermediate in the reaction.


Introduction
Multicomponent reactions (MCRs) employ three or more reactants to obtain heterocycles containing structures of all starting materials in a one-pot process under fixed reaction conditions [1][2][3][4]. It provides powerful productivity to satisfy modern green chemistry requirements, but new synthetic strategies with robust efficiency are still demanded. In this connection, the PASE approach has recently emerged [5,6], it is based on the Pot economy principle and unites it with the Atom and Step Economy strategies (PASE), thereby ensuring high efficiency, simplicity and low waste formation [7][8][9][10]. Nowadays, it is emerging as a fast-paced research front of organic chemistry [11][12][13][14][15][16].
Corrosion is a gradual destruction of refined metal by means of reactions with environment. Today it causes heavy losses to the economy [17][18][19]. Due to electronic configuration of heterocyclic compounds, they are capable of corrosion inhibition or coating metals. The development of PASE approaches to anti-corrosion heterocycles is beneficial for protection of metals.
Chromeno [2,3-b]pyridines are heterocycles with special and useful electronic configuration. Thus, they showed gastric antisecretory activity [20], inhibit mitogen-activated protein kinase 2 and suppress expression of tumor necrosis factor alpha (TNFα) in U937 cells [21]. Similar compounds were earlier synthesized via microwave irradiation of aldehydes, malononitrile dimer and kojic acid in EtOH [22]. The synthesis was carried out under reflux conditions in the presence of Et 3 N as a catalyst (73-90% yields). Bis-chromeno [2,3-b]pyridine derivatives were obtained via reaction of bis-aldehydes, malononitrile dimer and dimedone under reflux conditions. The reaction was carried out for 5 h in EtOH with a large amount of piperidine [23]. Phenyl substituted derivatives were obtained by reaction of benzaldehydes, malononitrile dimer and naphthols (12%-62% yields). The reaction was carried out under reflux conditions in H 2 O-EtOH with an equivalent amount of Et 2 NH [24] and in solvent-free conditions with guanidine hydrochloride catalysis (100 • C, 2 h) [25]. Solvent-free conditions were also used for the transformation of benzaldehydes, malononitrile dimer and coumarin [26]. Carbazole and indole derivatives were synthesized (67-85%) under reflux conditions and microwave irradiation: by reaction of aldehydes, malononitrile dimer and hydroxycarbazole or hydroxyindole in anhydrous EtOH with EtONa [27].
Further, we have examined the multicomponent reaction in alcohol media. The reaction in EtOH without catalyst resulted in assembling of chromeno [2,3-b]pyridine 4a in 19% yield (Entry 6), in the presence of AcONa it was formed in 35% yield (Entry 7) whereas organocatalysis by Et 3 N and morpholine resulted in 63% and 57% yields (Entries 8 and 9).  Pyridine (Py) has been already employed as a solvent and catalyst for similar processes [36]. Refluxing of salicylaldehyde 1a, malononitrile dimer 2 and hydroxyquinone 3 in Py for 2 h resulted in 61% yield (Entry 10). Py is a good solvent for structures of this type, however, a small amount of chromeno [2,3-b]pyridine 4a always remains dissolved in Py.
Then, reaction in a mixture of EtOH and Py was carried out. In the volumetric range of EtOH:Py (2-4:1), the best yield was in EtOH-Py (3:1) and chromeno [2,3-b]pyridine 4a was isolated in excellent 95% yield (Entry 12). Figure 1 shows that this organic catalyst is more effective than inorganic KF or AcONa catalysts and the 'EtOH-Py' system is more beneficial than 'solvent-free' and 'on-water' approaches to chromeno [2,3-b]pyridine 4a. Py in EtOH keeps the intermediates dissolved and ensures optimal basicity [37] supporting the reaction. Scheme 1. Reaction of salicylaldehyde, malononitrile dimer and hydroxyquinolinone.  Figure 1 shows that this organic catalyst is more effective than inorganic KF or AcONa catalysts and the 'EtOH-Py' system is more beneficial than 'solvent-free' and 'on-water' approaches to chromeno [2,3-b]pyridine 4a. Py in EtOH keeps the intermediates dissolved and ensures optimal basicity [37] supporting the reaction.  When the reaction in EtOH:Py mixture was finished, the final compound was directly crystallized in pure form after cooling. Under the optimal conditions (Entry12: refluxing for 2 h in 4 mL of EtOH-Py (3:1) mixture) multicomponent reactions of salicylaldehydes 1a-i, malononitrile dimer 2 Molecules 2020, 25, 2573 4 of 20 and hydroxyquinolinone 3 were carried out. Chromeno [2,3-b]pyridines 4a-i were obtained in 59-98% yields ( Table 2). In general, substitution reduced yields of chromeno [2,3-b]pyridines 4b-i. Presumably, electronic effects were taking place: electron-donating methyl, methoxy and ethoxy groups tended to support higher yields than bromo and nitro groups. The yield of chlorine substituted compound was bigger than in the case of bromo substituted compound due to a mesomeric effect. Table 2. PASE reaction of salicylaldehydes 1a-i, malononitrile dimer 2 and hydroxyquinolinone 3 1 .
Considering the mechanism of this multicomponent reaction, several path-ways for the reaction were possible. Salicylaldehyde 1 may have undergone condensation either with malononitrile dimer 2 or hydroxyquinolinone 3 (Scheme 2). Hence, four main pathways of the reaction could be When the reaction in EtOH:Py mixture was finished, the final compound was directly crystallized in pure form after cooling. Under the optimal conditions (Entry12: refluxing for 2 h in 4 mL of EtOH-Py (3:1) mixture) multicomponent reactions of salicylaldehydes 1a-i, malononitrile dimer 2 and hydroxyquinolinone 3 were carried out. Chromeno [2,3-b]pyridines 4a-i were obtained in 59-98% yields ( Table 2). In general, substitution reduced yields of chromeno [2,3-b]pyridines 4b-i. Presumably, electronic effects were taking place: electron-donating methyl, methoxy and ethoxy groups tended to support higher yields than bromo and nitro groups. The yield of chlorine substituted compound was bigger than in the case of bromo substituted compound due to a mesomeric effect. Table 2. PASE reaction of salicylaldehydes 1a-i, malononitrile dimer 2 and hydroxyquinolinone 3 1 . 1 Reaction conditions: 1a-i (1 mmol), 2 (1 mmol), 3 (1 mmol) were refluxed in 4 mL of EtOH-Py mixture (3:1). Isolated yields.
Considering the mechanism of this multicomponent reaction, several path-ways for the reaction were possible. Salicylaldehyde 1 may have undergone condensation either with malononitrile dimer 2 or hydroxyquinolinone 3 (Scheme 2). Hence, four main pathways of the reaction could be Considering the mechanism of this multicomponent reaction, several path-ways for the reaction were possible. Salicylaldehyde 1 may have undergone condensation either with malononitrile dimer 2 or hydroxyquinolinone 3 (Scheme 2). Hence, four main pathways of the reaction could be discussed. In order to gain insight into the reaction mechanism, we have performed 1 H NMR monitoring, to obtain constituent data on the reaction. To reduce the influence of sample preparation, the transformation of starting materials into chromeno[2,3-b]pyridine 4a was carried out and monitored directly into a spectrometer without catalyst to slow down the reaction.
During the NMR study, six major components were recorded: salicylaldehyde 1a, malononitrile dimer 2 and hydroxyquinolinone 3, In order to gain insight into the reaction mechanism, we have performed 1 H NMR monitoring, to obtain constituent data on the reaction. To reduce the influence of sample preparation, the transformation of starting materials into chromeno[2,3-b]pyridine 4a was carried out and monitored directly into a spectrometer without catalyst to slow down the reaction.
During the NMR study, six major components were recorded: salicylaldehyde 1a, malononitrile dimer 2 and hydroxyquinolinone 3, The starting materials are well known (we estimated the presence of salicylaldehyde 1a by the singlet at 10.25 ppm, malononitrile dimer 2 by the singlet at 3.85 ppm and hydroxyquinolinone by the singlet at 5.78 ppm); compound 4a is described in this manuscript and was estimated by the signal at 5.61 ppm; intermediate 5 was described previously (7.75 ppm) [40]; intermediate 7 was estimated by the signals at 8.04-8.11 ppm [41].  The starting materials are well known (we estimated the presence of salicylaldehyde 1a by the singlet at 10.25 ppm, malononitrile dimer 2 by the singlet at 3.85 ppm and hydroxyquinolinone by the singlet at 5.78 ppm); compound 4a is described in this manuscript and was estimated by the signal at 5.61 ppm; intermediate 5 was described previously (7.75 ppm) [40]; intermediate 7 was estimated by the signals at 8.04-8.11 ppm [41].  Based on these data and taking into consideration earlier published results [42,43], we suggest that the first stage was a rapid formation of intermediate 7 with expulsion of a hydroxide anion [37]. This hydroxide anion instantly catalyzed a rapid cyclization of intermediate 7 into intermediate 5.
To prove this conclusion, the intermediate 5 was isolated in a one-step reaction. Further addition of hydroxyquinolinone 3 to the obtained intermediate 5 in a distinct reactor afforded chromeno[2,3-b]pyridine 4a. The reaction conditions were the same as developed for multicomponent process and the yield of chromeno[2,3-b]pyridine 4a was 95% again. More than that, the two-component reaction of salicylaldehyde 1a and hydroxyquinolinone 3 in a separate reactor did not form intermediate 9. This experimental data and 1 NMR monitoring data were consistent with each other. Thus, the results from both intermediate synthesis and 1  To sum up, we postulated Path 1 for the assembling of chromeno[2,3-b]pyridine 4a in this multicomponent process. The proposed mechanism for the developed reaction is outlined in Scheme 3. Based on these data and taking into consideration earlier published results [42,43], we suggest that the first stage was a rapid formation of intermediate 7 with expulsion of a hydroxide anion [37]. To sum up, we postulated Path 1 for the assembling of chromeno[2,3-b]pyridine 4a in this multicomponent process. The proposed mechanism for the developed reaction is outlined in Scheme 3.

Quantum Chemistry Approach
One of the most well-known corrosion inhibitors are heterocycles, such as azoles [45,46], pyridines [47], several bioorganic compounds [48] and chromenes [49]. The mechanism of their action is the adsorption on the metal's surface (whatever metal is used-copper, steel, brass, etc.). Generally, the greater the number of electron-donating groups in molecule, the higher the probability of adsorption and inhibition activity. As it was shown previously, the inhibitory activity of organic compounds can be predicted properly by quantum-chemical methods [50][51][52]. This reduces the time of the study and hazardous reactant wasting by selecting the hit-compounds. They must have appropriate electronic properties, which can be estimated by quantum chemistry calculations, to interact with metal's surface.
To estimate electronic properties, the next method can be used [53]: according to Koopman's theorem, the energy of the highest occupied orbital (EHOMO) is equal to the ionization potential (I) having the opposite charge, the energy of the lowest unoccupied orbital (ELUMO) is related to the electron affinity (A).

Quantum Chemistry Approach
One of the most well-known corrosion inhibitors are heterocycles, such as azoles [45,46], pyridines [47], several bioorganic compounds [48] and chromenes [49]. The mechanism of their action is the adsorption on the metal's surface (whatever metal is used-copper, steel, brass, etc.). Generally, the greater the number of electron-donating groups in molecule, the higher the probability of adsorption and inhibition activity. As it was shown previously, the inhibitory activity of organic compounds can be predicted properly by quantum-chemical methods [50][51][52]. This reduces the time of the study and hazardous reactant wasting by selecting the hit-compounds. They must have appropriate electronic properties, which can be estimated by quantum chemistry calculations, to interact with metal's surface.
To estimate electronic properties, the next method can be used [53]: according to Koopman's theorem, the energy of the highest occupied orbital (E HOMO ) is equal to the ionization potential (I) having the opposite charge, the energy of the lowest unoccupied orbital (E LUMO ) is related to the electron affinity (A).
Considering the quantum chemical parameters of the inhibitor, the higher the energy of the highest occupied orbital (HOMO), the higher the donation of the inhibitor to metal's vacant d-orbital; the lower by the energy of the lowest unoccupied orbital (LUMO), the greater is the electron acception from the metal to inhibitor. The electron release is characterized by energy difference (∆E (L-H) ). The electron release is easier when the ∆E (L-H) is lower, and then the adsorption of inhibitor is stronger [54]. The electronic density distribution is an important parameter as well, and it can be estimated by a frontier orbital localization study and partial atomic charge distribution [55].
In addition to HOMO and LUMO, several significant parameters were calculated as well. Electronegativity shows the ability of the molecule to attract electrons towards itself and therefore a molecule with the lowest value tends to has the highest ability to donate electrons.
The global chemical hardness is related to the resistance of a molecule to charge transfer [56]. The global chemical softness (σ) is inversely proportional from the chemical hardness. The higher softness, the better adsorption [57]. Chemical hardness (η) as well as electronegativity (χ) can be calculated from ionization potential and electron affinity, and then chemical softness can be calculated from chemical hardness. Since the molecule of inhibitor must donate its electron to vacant metal orbitals, its nucleophilicity should be high, opposite to electrophilicity and hence to the global electrophilicity index (ω) [58].
All quantum chemical calculations were performed using the Gaussian09 program package [59]. The structures 4a-i were optimized and the required parameters were calculated using the density functional theory method B3LYP [60] with the 6-311G(d,p) basic set. This method is recommended for calculation of frontier orbitals energies and related values, such as electronegativity, polarizability etc. [61]. Our study was carried out for molecules in gas phase as well as for solvated ones in water (PCM model) [62].
The results of quantum chemical calculations for gas phase, solvated forms and protonated forms are presented in Table 3, frontier orbitals of several studied compounds are shown in Table 4.   Frontier orbitals of these compounds are also presented in Figures S43--S55

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55 of Supplementary Materials.

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55 of Supplementary Materials.

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55 of Supplementary Materials.

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55 of Supplementary Materials.

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and  Frontier orbitals of these compounds are also presented in Figures S43--S55 of Supplementary Materials.

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and   Frontier orbitals of these compounds are also presented in Figures S43-S55 of Supplementary Materials.

Gas Phase Calculations
During quantum chemical calculations, the structures of the studied compounds were optimized as shown in Figure 4 for 4a. The chromeno[2,3-b]pyridine fragment was not planar and hydroxyquinolinone ring was localized nearly perpendicular to chromeno[2,3-b]pyridine ring. Calculated key parameters of the studied compounds 4a-i are shown in Table 3.
Calculated key parameters of the studied compounds 4a-i are shown in Table 3. As mentioned above, corrosion inhibition depends on the energies and distributions of frontier orbitals. For most cases in considered compounds, the HOMOs are localized on the benzene ring atoms of chromeno [2,3-b]pyridines, making them sites for interaction with metal cations that formed by dissolving the metal in acid. The LUMOs are localized in the hydroxyquinolinone ring of the substituent. Such a type of distribution of the frontier orbitals is represented in compounds 4a, 4b, 4d and 4i. The frontier orbitals of several compounds are shown in Table 4.
Introducing of a strong electron-withdrawing substituent, such as a nitro-group, tended to change the orbital distribution. The HOMO became localized on the atoms of the pyridine ring, and the LUMO on the nitro-group of the substituent (4h, Table 4). A similar distribution was observed for the compounds 4e, 4f and 4g: the HOMOs were localized in the pyridine ring, while the LUMOs were in the hydroxyquinolinone ring (Table 4). This was caused by the withdrawing effect of the halogen atom. It is noteworthy that compound 4c had the same distribution of the orbitals, which was probably related to poor overlapping of oxygen orbital of methoxy-group with ones of benzene ring.
The energy difference is another important parameter. The lower the energy difference of the molecule, the higher the reaction ability (due to easier electron release). For the studied compounds these values were comparable, but 4b, 4e and 4i had the lowest ones. The global electronegativity χ values for 4b, 4e and 4i were almost twice as low as values calculated for similar tricyclic cationic inhibitors (<7.2 is accepted, [53]). The global electrophilicity index ω was low as well. It was almost five times lower than the values calculated for the studied inhibitors (<19.3 is accepted, [53]).

Solvated Form Calculations
Several parameters changed by taking into consideration the solvation processes. Thus, the frontier orbitals in compounds 4a, 4b, 4d, 4f and 4i were located in the same manner as in non-solvated ones, but for 4c, 4e, 4g and 4h shifting of the HOMOs to the carbonyl atoms of hydroxyquinolinone ring is observed, while the LUMOs remained in the same place ( Table 4).
The analysis of the basic parameters (Table 3) shows that assuming solvation led to changing of some values mostly for compound 4h; we can expect the highest anticorrosive effect from 4b, 4c and 4d due to their low values of general electrophilicity ω, which can be predicted because of electron-donating substituents. The values of electronegativity χ for 4b, 4c and 4d remained low, As mentioned above, corrosion inhibition depends on the energies and distributions of frontier orbitals. For most cases in considered compounds, the HOMOs are localized on the benzene ring atoms of chromeno [2,3-b]pyridines, making them sites for interaction with metal cations that formed by dissolving the metal in acid. The LUMOs are localized in the hydroxyquinolinone ring of the substituent. Such a type of distribution of the frontier orbitals is represented in compounds 4a, 4b, 4d and 4i. The frontier orbitals of several compounds are shown in Table 4.
Introducing of a strong electron-withdrawing substituent, such as a nitro-group, tended to change the orbital distribution. The HOMO became localized on the atoms of the pyridine ring, and the LUMO on the nitro-group of the substituent (4h, Table 4). A similar distribution was observed for the compounds 4e, 4f and 4g: the HOMOs were localized in the pyridine ring, while the LUMOs were in the hydroxyquinolinone ring (Table 4). This was caused by the withdrawing effect of the halogen atom. It is noteworthy that compound 4c had the same distribution of the orbitals, which was probably related to poor overlapping of oxygen orbital of methoxy-group with ones of benzene ring.
The energy difference is another important parameter. The lower the energy difference of the molecule, the higher the reaction ability (due to easier electron release). For the studied compounds these values were comparable, but 4b, 4e and 4i had the lowest ones. The global electronegativity χ values for 4b, 4e and 4i were almost twice as low as values calculated for similar tricyclic cationic inhibitors (<7.2 is accepted, [53]). The global electrophilicity index ω was low as well. It was almost five times lower than the values calculated for the studied inhibitors (<19.3 is accepted, [53]).

Solvated Form Calculations
Several parameters changed by taking into consideration the solvation processes. Thus, the frontier orbitals in compounds 4a, 4b, 4d, 4f and 4i were located in the same manner as in non-solvated ones, but for 4c, 4e, 4g and 4h shifting of the HOMOs to the carbonyl atoms of hydroxyquinolinone ring is observed, while the LUMOs remained in the same place ( Table 4).
The analysis of the basic parameters (Table 3) shows that assuming solvation led to changing of some values mostly for compound 4h; we can expect the highest anticorrosive effect from 4b, 4c and 4d due to their low values of general electrophilicity ω, which can be predicted because of electron-donating substituents. The values of electronegativity χ for 4b, 4c and 4d remained low, almost the same as was calculated for the gas phase. It is noteworthy that the value of the global softness σ for these compounds tended to decrease while taking to consideration the solvation effects.

Protonated form Calculations
In the case of a corrosion inhibition study in high acidic media, we must take in consideration the possibility of protonation of our compounds because of the basicity of some functional groups.
As it was shown previously [63], the main center of basicity in aminopyridines is the nitrogen atom in heterocyclic ring. We decided to confirm this statement by calculating the protonation energy for the compound 4a (Scheme 4).
Molecules 2020, 25,2573 13 of 20 almost the same as was calculated for the gas phase. It is noteworthy that the value of the global softness σ for these compounds tended to decrease while taking to consideration the solvation effects.

Protonated form Calculations
In the case of a corrosion inhibition study in high acidic media, we must take in consideration the possibility of protonation of our compounds because of the basicity of some functional groups.
As it was shown previously [63], the main center of basicity in aminopyridines is the nitrogen atom in heterocyclic ring. We decided to confirm this statement by calculating the protonation energy for the compound 4a (Scheme 4). Proton affinity (PA) is Equation: where Ei-the total energy of the corresponding compound We found that for the compound 4a protonation took place on the pyridine ring nitrogen atom, while it was preferable by 15.1 kcal/mol in comparison to the α-amino-group protonation and by 14.2 kcal/mol to the γ-amino-group.
The calculation results confirmed that the protonation took place first on the pyridine ring, so we have not carried out such calculations for the other compounds in order to reduce computation time, and optimized their geometry assuming protonation of the first pyridine position.
The calculation for the protonated molecules shows the great changing of electron density distribution in them. As the nitrogen atom in chromeno[2,3-b]pyridine cycle became positively charged and revealed high electron-withdrawing effect, the electron density in the pyridine ring decreased. This process was also strengthened by the influence of the cyano-group in this ring. In all nine compounds the LUMOs were localized on the atoms of pyridine ring, while the HOMOs were localized either on the benzene ring atoms (compounds 4a, 4b, 4d, 4i, Table 4) or on the carbonyl group of the hydroxyquinolinone ring (compounds 4c, 4e, 4f, 4g, 4h, Table 4).
The key parameters remained low for compounds 4a-i even in the protonated state. Global electrophilicity ω was extremely low (<60 is acceptable, [53]), the softness σ and electronegativity χ were almost twice as low as the values for known cationic dies (<1 and <10 are acceptable, respectively [53]).
The analysis of the key parameters for the protonated molecules of 4a-i shows that the highest anticorrosion effect was possessed by the compounds 4b, 4c and 4d, the lowest by 4f and 4h. However, despite of the presence of an electron-withdrawing substituent and the high value of global electrophilicity ω, compound 4h had the highest value of chemical softness σ, which indicates that this compound can readily react with forming Fe 2+ ions (soft acid with soft base). Proton affinity (PA) is Equation: where E i -the total energy of the corresponding compound We found that for the compound 4a protonation took place on the pyridine ring nitrogen atom, while it was preferable by 15.1 kcal/mol in comparison to the α-amino-group protonation and by 14.2 kcal/mol to the γ-amino-group.
The calculation results confirmed that the protonation took place first on the pyridine ring, so we have not carried out such calculations for the other compounds in order to reduce computation time, and optimized their geometry assuming protonation of the first pyridine position.
The calculation for the protonated molecules shows the great changing of electron density distribution in them. As the nitrogen atom in chromeno[2,3-b]pyridine cycle became positively charged and revealed high electron-withdrawing effect, the electron density in the pyridine ring decreased. This process was also strengthened by the influence of the cyano-group in this ring. In all nine compounds the LUMOs were localized on the atoms of pyridine ring, while the HOMOs were localized either on the benzene ring atoms (compounds 4a, 4b, 4d, 4i, Table 4) or on the carbonyl group of the hydroxyquinolinone ring (compounds 4c, 4e, 4f, 4g, 4h, Table 4).
The key parameters remained low for compounds 4a-i even in the protonated state. Global electrophilicity ω was extremely low (<60 is acceptable, [53]), the softness σ and electronegativity χ were almost twice as low as the values for known cationic dies (<1 and <10 are acceptable, respectively [53]).
The analysis of the key parameters for the protonated molecules of 4a-i shows that the highest anticorrosion effect was possessed by the compounds 4b, 4c and 4d, the lowest by 4f and 4h. However, despite of the presence of an electron-withdrawing substituent and the high value of global electrophilicity ω, compound 4h had the highest value of chemical softness σ, which indicates that this compound can readily react with forming Fe 2+ ions (soft acid with soft base).
The developed approach is facile, it is easy to isolate final compounds directly from reaction mixture and the yields of final compounds are 59%-98%.