First, all the reagents and solvents were purchased from Aldrich and used without further purification. Melting points were obtained with PF-300 SEV equipment. UV/Vis spectra were obtained with a Shimadzu UV-3100S spectrophotometer on MeOH, CHCl₃ solutions of ca 8.8 ´ 10⁻⁵ M for 1 and 1.84 ´ 10⁻⁴ M for 2 as hydrochlorides. As free amines, in MeOH, CHCl₃ solutions using ca of 1.16 ´ 10⁻³ M/7.57 ´ 10⁻⁵ M for 1 and 6.85 ´ 10⁻⁵ M/1.60 ´ 10⁻⁵ M for 2, using CH₂Cl₂ and CH₃OH for 3 at ca 10⁻⁵ M. UV-Vis 1-3 data are reported as: wavelength, l_max (nm), molar absorptivity, ε (cm⁻¹∙mol), transition energy gap, ΔE (cm⁻¹). IR spectra were recorded with a Nicolet Magna-IR 750 spectrophotometer between 4000 cm⁻¹ and 400 cm⁻¹ using KBr pellets. The ¹H-NMR and ¹³C-NMR were recorded with a Bruker Avance III 500 MHz spectrometer for 1 and 2, and with a JEOL Eclipse-400 spectrometer for 3, in CDCl₃ solutions and using TMS as reference for all samples. ¹H-NMR spectra were recorded between 0 and 10 ppm with acquisition time of 2 sec, 16 repetitions. Chemical shifts are reported as δ part per million (ppm) values relative to TMS; s = singlet, d = doublet, t = triplet, m = multiplet. The mass spectra were recorded in a JEOL MStation JMS-700 spectrometer using a FAB⁺ technique. A GBC 932 atomic absorption spectrometer was used to quantify the Zn²⁺ concentrations by the flame method. The hollow cathode lamp was operated at 5.0 mA, with λ = 213.9 nm, a slit of 0.5 nm and burner height of 13.5 mm, the airflow and acetylene rates were used as recommended by the manufacturer and a background correction was carried out with a deuterium lamp. Aqueous solutions of Zn(C₂H₃O₂)₂·H₂O at 0.5, 1.0, 1.5 mg/L concentrations were used to obtain a calibration curve.

General synthetic procedures to obtain 1 and 2: A mixture of acetic acid (0.658 mL, 11.49 mmol), Zn⁰ (657.8 mg, 10.06 mmol) and DAB-Am-4 (320.0 mg, 1 mmol) for 1 or PAMAM G0 (516.5 mg, 1 mmol) for 2 in 8 mL of MeOH were stirred under reflux. Then, 2-pyridinecarboxaldehyde (0.38 mL, 6 mmol), in 3 mL of MeOH was added dropwise over 1.5 hours. The mixture was refluxed for 6 hours for 1 and 72 hours for 2, cooled to room temperature and filtered. Then large amounts of gaseous HCl were bubbled until a white precipitate appeared. The solid was decanted and washed three times with MeOH and Et₂O. The free products are obtained by treating the hydrochlorides with 10 mL of a 1 M NaOH solution, extracted with chloroform and dried.

General synthetic procedure to obtain 3: A mixture of DAB-Am4 (1.49 mmol) with 1 mL of a 5.98 mM solution of salicylaldehyde in MeOH was stirred in a flask. The reaction mixture changed immediately from yellow to orange when salicylaldehyde was added. Then, 1 mL of salicylaldehyde aliquots were added at 0.5 hrs, 1.5 hrs and 4 hrs from the starting reaction time. The reaction was followed by silica TLC, IR and ¹H-NMR spectroscopy. After 24 hrs 3 was extracted with CHCl₃ from the reaction mixture, dried and filtrated.

N,N,N’,N’-tetrakis-[3((pyridine-2-methyl)-amine) propyl]-1,4-butanediamine (1). 1 was obtained as yellow oil (yield: 84.1%). UV-Vis/CH₃OH: l_max/ε/ΔE (nm/cm⁻¹mol/cm⁻¹): 261/9815/38314.2; UV-Vis/CHCl₃ l_max/ε/ΔE (nm/cm⁻¹mol/cm⁻¹): 262/573/38167.9; Selected IR data/KBr (cm⁻¹): _O-H 3416 (br), _N-H 3257 (vs), _C-H 2934 (s), _C-H 2817 (m), _C-N 1649 (w), _C=C 1593 (m), _N-H 1567 (m), _C-Har 1470 (m), _C-N 1116 (w), _C-H 758 (m); ¹H-NMR (500 MHz, CDCl₃) δ (ppm): 8.54 (d, 4H), 7.62 (t, 4H), 7.299 (s, 4H), 7.14 (t, 4H), 3.88 (s, 8H), 2.66 (t, 8H), 2.46 (t, 8H), 2.38 (s, 4H), 1.66 (q, 8H), 1.37 (s, 4H), ¹³C-NMR (500 MHz, CDCl₃) δ (ppm): 159.90, 149.23, 136.39, 122.22, 121.84, 55.29, 54.01, 52.22, 48.17, 27.47, 24.92; MS (FAB⁺) for C₄₀H₆₀N₁₀ [M + H]⁺ calculated: 680.99, found: 681.00.

N,N,N’,N’-tetrakis-[N-((2-methylpyridine)ethyl)propanamide]ethylendiamine (2). 2 was obtained as a yellow oil (yield: 60%). UV-Vis/CH₃OH: l_max/ε/ΔE (nm/cm⁻¹mol/cm⁻¹): 205.1/48757/48756.7, 207.9/48100/ 48100, 211.4/47304/47303.7, 218.3/47304/45808.5, 257.3/38865/38865.1, 261.5/38241/38240.9, 268.1/37300/ 37299.5; UV-Vis/CHCl₃ l_max/ε/ΔE (nm/cm⁻¹mol/cm⁻¹): 252.5/39604/39604, 263.4/37965/37965.1, 270.6/36955/ 36954.9, 309.8/32279/32278.9; IR/KBr_max (cm⁻¹): _N-H 3278 (vs), _C-H 2927 (m), _C=O 1647 (vs), _C=C 1593 (s), _N-H 1568 (m), _C-H 761 (m); ¹H-NMR (400 MHz, CDCl₃): δ ppm: 8.52 (s, 4H), 8.65 - 7.16 (m, 12H), 3.87 (d, 8H), 3.34 (m, 8H), 2.77 (m, 8H), 2.71 (m, 8H), 2.62 (s, 4H), 2.29 (m, 8H); ¹³C-NMR (400 MHz, CDCl₃) δ (ppm): 173.26, 159.55, 149.43, 136.87, 122.61, 122.30, 54.8, 52.12, 50.77, 48.68, 39.46, 34.37; MS (FAB⁺) for C₄₆H₆₈N₁₄O₄: [M]⁺ calculated: 881.10, found: 881.00.

N,N,N’,N’-tetrakis-(3((2-hidroxibenziliden)-amine)propyl)-1,4-butanediamine (3). 3 was obtained as a highly viscous yellow liquid (yield: 83.0%). UV-Vis/CH₃OH: l_max/ε/ΔE (nm/cm⁻¹mol/cm⁻¹): 216.9/45426.4/ 46053.6, 252.61/20013.5/39544.8, 269.5/8926.9/37065.1, 310.8/7415.1/32150.1, 393.5/2879.6/25391.5, 515.1/ 503.9/19396.2; UV-Vis/CHCl₃: l_max/ε/ΔE (nm/cm⁻¹mol/cm⁻¹): 248.9/30429.4/40126.3, 260.2/33836.9/38394.2, 271.3/10380.9/36824.6, 317.7/13550.6/31439.7, 406.0/554.7/24605.9; IR/KBr _max (cm⁻¹): _O-H 3060 (vb), _C=N 1633 (s); ¹H-NMR (400 MHz, CDCl₃) δ (ppm): 1.39 (s, 4H), 1.79 (m, 8H), 2.38 (s, 4H), 2.49 (t, 8H), 3.59 (t, 8H), 7.30 - 6.81 (m, 16H), 8.31 (s, 4H), ¹³C-NMR (400 MHz, CDCl₃) δ (ppm): 25.08, 28.40, 51.33, 53.90, 57.35, 116.94, 118.38, 118.72, 131.07, 132.02, 161.25, 164.82; MS (FAB⁺) for C₄₄H₅₆N₆O₄: [M]⁺ calculated: 732.88 found: 733.00.

A previously reported methodology from our research group was used to synthesize compounds 1 and 2 [17] (Figure 1). This methodology was originally proposed for reactions between linear chain polyamines and 2-pyridinecarboxaldehyde and adapted for the functionalization of the DAB-Am4 and PAMAM cores leading to 1 and 2 in good yields. On the other hand, compound 3 showed in Figure 1 was obtained using a reported methodology from the literature to synthesize branched imines [18] .

The pathway reaction to obtain 1 and 2 was proposed based on the Zn redox potential quantification, IR and NMR spectroscopic studies (Scheme 1), whose discusses are given in the section corresponding.

We corroborated that Zn oxidation potential to produce H^- is feasible at the synthesis reaction conditions. For that purpose a mixture of Zn⁰, acetic acid and methanol were stirred during 7.5 hours. Subsequently, the reaction mixture was filtered, and subsequently diluted with deionized water. The initial and final concentration of Zn²⁺ were quantified by atomic absorption spectrometry. The obtained values were [Zn²⁺]_i = 0.0 mg/L and the [Zn²⁺]_f = 38,125 mg/L, for the initial and final concentrations, respectively. At 343 K (70˚C) and at pH = 2.78 Zn²⁺/Zn the oxidation-reduction potential was quantified using equation (1) [19] .

where R = ideal gas constant, T = 343 K, ν = 2 transferred electrons, F = Faraday constant and Q = [Red_A]^a![Ox_B]^b!/[Ox_A]^a[Red_B]^b. The redox potential value was of E = −0.85 V and kcal∙mol⁻¹, it implies that the Zn oxidation reaction is favored. These results showed that with these reaction conditions, the Zn⁰ oxidation in acid media allowed for hydride formation, H^-. The H^- ions are able to reduce fully the pyridinecarboxaldehyde carbonyl. With this in mind, the proposed reaction mechanism is shown in the Scheme 1(a).

In the first step there is a nucleophilic attack of the amino group contained in the DAB-Am4 or PAMAM-G0 molecules due the carbonyl π-electrons polarization. In the second step a concerted mechanism involving the H^- ions is carried out, which avoids the imine formation. The H^- ion is highly reactive and does not allow the formation of double bonds and therefore the formation of imines.

On the other hand, the current tendency in the conventional condensation reactions to obtain imine [18] is the optimization and reduction of the solvents, energy and secondary reagents used for the synthesis. This is known as “green chemistry” [20] . The obtaining of 3 without using solvents was not possible since the mixture reaction showed a high viscosity and a null reactants mixture results; however the use of solvents in small amounts lead to obtain 3 in good yields. The classic reaction mechanism for 3 is given in the Scheme 1(b).

It is noteworthy that although the salicylaldehyde has an electron-donor hydroxo group in an ortho position with respect to the aldehyde group fast protonation of the oxygen atom was carried out. This shows that when H^- ions are not in the reaction medium, imines compounds will always be obtained [18] . Therefore we conclude that when a strong nucleophile (as the hydride anion) is found in the media reaction the amine formation must be favored.

IR spectra of 1-3 showed the vibrations typical of the materials containing DAB-Am4, PAMAM-G0, 2-pyridinecarboxaldehyde and salicylaldehyde (Supporting Information Figure S1) [21] - [26] . The optimal reaction times for obtaining 1-3 were set by taking samples at different reaction times, observing the change in the spectra. Figure 2 shows the spectra of (a) the starting materials and (b) to (e) the reactions evolution for 1-3 at different reaction times in the zone between 2200 cm⁻¹ and 1100 cm⁻¹.

The band around 1715 cm⁻¹ is assigned to the characteristic vibration _C=O st of the 2-pyridinecarboxalde- hyde, and at 1664 cm⁻¹ is the characteristic vibration _C=O from the salicylaldehyde carbonyl group [22] [25] - [27] . Figure 2(i) and Figure 2(iii) show the IR spectra evolution, at different times of reaction to obtain 1 and 3. For compound 2 the monitoring through the infrared spectra (Figure 2(ii)) was complicated due to the presence of the carbonyl in the amide group, which is very close to the carbonyl signal from the starting material. However, the carbonyl bands in the final spectrum were fitted with Gaussians showing that the carbonyl signal corresponding to 2-pyridinecarboxaldehyde disappeared completely.

The optimal time for the reaction for three compounds was also corroborated by TLC. Finally, it is noteworthy that the energy position of the _OH band of 3 implies two characteristics: 1) the hydroxyl group is conjugated with amino group through the aromatic ring; 2) the equilibrium phenol-imine and keto-amine tautomeric forms are shifted to phenol-imine form (Supporting Information in Figure S2) [27] - [29] . Table 1 summarizes the most important vibration frequencies found in the infrared spectra assigned according the literature [27] - [29] .

The UV-Vis spectra in MeOH show transitions with λ_max = 258.5 nm for 1, λ_max = 205.1 nm, 207.9 nm and 257.3 nm for 2 and λ_max = 216 nm, 252.6 nm, 269.5 nm, 310.7 nm, 393.4 nm and 515 nm for 3 [22] [26] [27] [29] [30] .

Through a fitting of the spectra with Gaussians it is possible to know the exact number of electronic

ar = aromatic, al = aliphatic.

transitions contained on each band. For 1, the fitting shows four Gaussians with λ_max = 241.6 nm, 258.8 nm, 261.7 nm and 267.9 nm (Supporting Information in Figure S3), which is in agreement with the number and energy of the π-π^* transitions reported for the pyridine chromophore [29] . For 2 (Figure 3(a)), the diverse functional groups lead to a more complex spectrum presenting two broad bands containing seven Gaussians assigned to four n-π^* electronic transitions with λ_max = 205.1 nm, 207.9 nm, 211.4 nm and 218.3 nm and three n-π^* transitions with λ_max = 257.3 nm, 261.5 nm and 268.1 nm. The spectrum for 3 shows four broad bands containing six Gaussians assigned to two n-π^* electronic transitions with λ_max = 216.9 nm and 252.6 nm and four n-π^* transitions with λ_max = 269.5 nm, 310.7 nm, 393.4 nm and 515 nm in agreement with the structure of 3 (Supporting Information in Figure S4).

A change of the solvent used in the UV-Vis experiments, from MeOH to CHCl₃, had practically no effect on the number and wavelength of the electronic transitions for compound 1. This spectrum showed a band with λ_max = 262 nm, which fitted with four Gaussians with λ_max = 252.5 nm, 263.4 nm, 270.6 nm and 309.8 nm (Supporting Information in Figure S3). However, compounds 2 and 3 UV-Vis spectra showed a significant broadening and shifting of the bands. For 2 the bands with λ_max = 257.3 nm, λ_max = 261.5 nm and λ_max = 268.1 nm in the spectrum of MeOH are shifted to λ_max = 263.4 nm, λ_max = 270.6 nm and λ_max = 309.7 nm in the spectrum of CHCl₃. Figure 3(a) and Figure 3(b) show the UV-Vis spectra of 2 in MeOH and CHCl₃ highligting the bathochromic shifting. We attribute this change in the spectra to the stabilization of the n orbitals due to H-bonding formation between the solvent and the amide groups in compound 2 when MeOH is used as the solvent [30] . For 3, the transitions were recorded in CH₃OH and CH₂Cl₂ (Supporting Information in Figure S4). The OM diagram clearly shows (Supporting Information in Figure S4), the stabilization of 3 in CH₂Cl₂ due the intra-hydrogen- bond between the N of the imine group and the hydroxyl proton group adopting the phenol-imine form (Supporting Information in Figure S2); while the spectrum in MeOH solution shows that the methanol-imine intermolecular hydrogen bridges are present and therefore 3 is destabilized. This corroborates the information obtained through the IR spectra for 3 and must be valid for 2 [29] [31] .

All three compounds are obtained as yellow oils very soluble in MeOH, H₂O and CHCl₃. The obtaining of the desired compounds is corroborated by the number, chemical shifts and integral peak intensities in the ¹H-, ¹³C-NMR spectra, which are in good agreement with the chemical formulas. The peaks are correctly assigned carrying out HSQC experiments, which correlate with the heteronuclear C and H atoms directly bonded.

The pyridine ring signals in the ¹H-NMR spectra are found between 7.14 ppm - 8.54 ppm for 1 (Supporting Information in Figure S5) and between 7.16 ppm - 8.53 ppm for 2, while for 3 the aromatic protons are found between 6.81 ppm - 7.30 ppm (Figure 4(a), Figure 4(b)). The peaks between 3.87 ppm - 1.26 ppm for 1, 3.87 ppm - 2.29 ppm for 2 and 3.59 ppm - 1.39 ppm for 3 are assigned to the aliphatic zone. The aromatic signals are correlated in the ¹³C-NMR spectra with the peaks between 159.90 ppm - 121.84 ppm for 1, between 154.55 ppm - 122.30 ppm for 2 and between 164.82 ppm - 116.94 ppm for 3 and in the aliphatic zone with the peaks between 55.29 ppm - 24.92 ppm for 1, between 54.74 ppm - 29.71 ppm for 2 and between 57.35 ppm - 25.08 ppm for 3 [32] .

To corroborate the reaction mechanism proposed for the synthesis of 1 and 2, the ¹H-NMR spectra of 1 were obtained at 1, 3 and 5 min from the starting reaction time showing that there was no double bond formation (Supporting Information in Figure S6); while the characteristic proton and carbon signals in the HSQC spectrum of 2, 3 were very clear (Figure 4(a), Figure 4(b)).

Thus, NMR studies corroborated the information obtained by IR about the reaction mechanism for obtaining 1 and 2, which does not imply the imine formation due the presence of a strong nucleophile agent (H⁻) which acts fast on the carbonyl carbocation.

Because the molecules contain two perpendicular symmetry axes, only the H⁺ and C atoms of one of the branches in the molecules are shown. The spectra exhibit broad signals because of the flexible nature of these molecules, which is more evident in the spectrum of 2, which contains a PAMAM core and therefore more functional groups.