Solid-State Catalytic Hydrogen/Deuterium Exchange in Mexidol

Abstract

The reaction between spillover hydrogen (SH) and mexidol upon high-temperature solid state catalytic isotope exchange (HSCIE) is studied experimentally and via DFT B3LYP and MP2 calculations. The distribution of isotopes and the activation energy of the reaction are determined via NMR spectro-scopy and mass spectrometry for samples of the deuterium-labeled mexidol obtained via HSCIE at 50–120°C. The DFT calculated activation energy of hydrogen transfer on the charged alumina cluster HH⁺(‒Al–O–)₃(OH)₆ is found to agree with the experimental data. A mechanism is proposed for HSCIE between the adsorbed compound and charged Brønsted acid sites formed on the surface of the support upon exposure to hydrogen spillover.

INTRODUCTION

In heterogeneous catalysis, spillover is the transport of reactive species adsorbed or formed on one phase to another that, under certain conditions, does not sorb or form these species itself. The transport of activated hydrogen from platinum metals to an inorganic support is referred to as hydrogen spillover (HS) [1]. In this work, we studied the mechanism of the reaction between spillover hydrogen (SH) and mexidol (2-ethyl-6-methyl-3-hydroxypyridine succinate), a multipurpose drug with antihypoxic, antioxidant, anxiolytic, and nootropic activities.

Even though reactions with the participation SH have been known for many years, there is still no common opinion on their nature. According to different hypotheses, hydrogen migrates in the form of a solvated proton, a proton-electron pair, or atomic hydrogen. The migration of SH onto a semiconducting inorganic support is accompanied by the protonation of TiO₂ and the transfer of electrons to the conductivity band [2–4]. DFT calculations of hydrogen spillover over the surfaces of nonreducible metal oxides show that atomic hydrogen cannot migrate into a crystal lattice, and that the observed ability of SH to hydrogenate support-adsorbed molecules cannot be explained for hydrogen atoms [5]. It is known that the selectivity of catalytic isomerization of hydrocarbons is as high as 95–99% under the action of SH [6]. This change in selectivity is due to the reaction proceeding with the participation of new Brønsted acid sites (BASes) that form under the action of HS [7, 8].

The transfer of SH onto an inorganic support in the form of positive particles is confirmed by the formation of negative charges on platinum-group metals [9, 10]. It is assumed that protons migrate over a surface containing bound water and hydroxyl groups via an estafette mechanism [11]. Hydroisomerization, hydrocracking, and isotope exchange can occur on the resulting BASes. It has been shown experimentally that high-temperature solid-state catalytic isotope exchange (HSCIE) occurs on catalytic centers of an inorganic support that form upon exposure to SH [12]. In contrast to liquid-phase reactions, solid-state HSCIE is characterized by extremely weak kinetic effects [13]. The HSCIE of hydrogen for deuterium and tritium in organic compounds occurs without racemization of asymmetric carbon atoms, allowing us to use this reaction for the synthesis of highly tritium-labeled physiologically active compounds [14].

Catalytic reactions with the participation of SH have long been used in industry, but the mechanism of reactions that occur on the resulting strongly acid sites remains poorly studied. The aim of this work was to study the catalytic isotope exchange of hydrogen between SH and alumina-supported mexidol by experimental and quantum chemical means.

EXPERIMENTAL

Deuterium-labeled mexidol was synthesized via HSCIE by treatment of a solid mixture of alumina-supported mexidol and finely divided 5% Pd/BaSO₄ (Fluka) with gaseous deuterium. A solid reaction mixture was prepared as described for the synthesis of glycine labeled with hydrogen isotopes in [14]. A solution of mexidol (1.0 mg) in water (0.5 mL) was mixed with alumina (20 mg) and frozen, and water was removed by freeze drying. Alumina with supported mexidol was mixed with 5% Pd/BaSO₄ (10 mg) (Fluka). The solid reaction mixture was placed in a 10-mL ampule, evacuated, cooled to 0°C, and filled with deuterium up to a pressure of 30 kPa. Isotope exchange was conducted under the conditions given in Table 1 at 50–120°C. Each ampule was then cooled, evacuated, and purged with hydrogen. Mexidol was desorbed using 50% ethanol. The labeled mexidol was dissolved twice more in 50% ethanol and evaporated to dryness to remove labile deuterium. The deuterium-labeled mexidol samples were purified via HPLC on a Kromasil 8 × 150 column using aqueous methanol in a concentration gradient of 15–40% in the presence of 0.1% heptafluorobutyric acid (HFBA) with two-beam UV detection at 220 and 254 nm. The chromatographic retention indices and the extinction ratio for the deuterium-labeled and initial unlabeled mexidol coincided perfectly. The solutions of deuterium-labeled mexidol purified via HPLC were evaporated under reduced pressure, dissolved in distilled water, and freeze dried at −50°C.

Table 1.   High-temperature solid-state catalytic isotope exchanges of hydrogen for deuterium in mexidol

Mass Spectrometry

The deuterium-labeled [²H]mexidol was analyzed on an LCQ Advantage MAX mass spectrometer (Thermo Electron). The calculated molecular weight value for the light monoisotopic mexidol was 137. The analyzed [²H]mexidol sample was diluted with methanol–acetic acid (100 : 0.1) to a concentration of 10 µg/mL. The spectra for peptides were measured in the positive ion mode in the 50 to 200 range of m/z.

NMR Spectroscopy

The positions of isotopic tags and the degrees of hydrogen/deuterium substitution in the C–H bonds of [²H]mexidol were determined from 1D and 2D ¹H NMR spectra acquired at 20°C by comparing the peak intensities (¹H/¹H-TOCSY with a delay time of 60 ms). ¹H NMR spectra were obtained using an AVANCE 700 MHz NMR spectrometer.

Quantum Chemical Calculations

Quantum chemical calculations of the studied systems were performed via DFT B3LYP/6-31G* with full energy optimization and calculation of the normal mode frequencies. The geometrical parameters of compounds, their electronic characteristics, electron density distribution, full energies, energies of transformations, entropies of transformations, and frequencies of normal modes were calculated. The geometrical structure of molecules was optimized and the frequencies of normal modes were calculated using the 6‑31G* atomic basis sets. All calculations with full geometry optimization of molecules and calculations of normal-mode frequencies were made using the GAUSSIAN-09 program [15] under the LINUX operating system. DFT B3LYP is a combination of the Hartree–Fock approach and density functional theory, in which the gradient-corrected Becke functional with three parameters (B3) [16, 17] and Lee-Yang-Parr correlation energy functional (LYP) [18] are used. The geometrical arrangement of atoms was optimized for each molecule using analytical calculation procedures. Calculations of the frequencies of normal modes using second-order derivatives confirmed that stationary points determined upon geometry optimization were energy minima. Fragments of the potential surface of reactions were calculated. The transitional states of reactions were calculated using the QST2 and QST3 synchronous transit approach [19, 20]. The correspondence between the resulting transition states of a presumed reaction was established using the intrinsic reaction coordinate (IRC) approach [21].

Density functional theory (DFT) calculations provided data on the structure of mexidol (Fig. 1). Figure 1 shows bond lengths and the positions of hydrogen atoms considered to participate in solid-state isotope exchange with SH, and for which quantum chemical calculations of the structures of transitions states of the hydrogen exchange reaction and activation energies were calculated.

Fig. 1.

Geometrical structure of mexidol (DFT B3LYP/6-31G*).

The HSCIE reaction with mexidol was studied in the temperature range of 50–120°C. Figure 2 shows the mass spectrum for the deuterium-labeled [D]mexidol obtained via HSCIE in 5 min at 80°C. The monoisotopic ion of mexidol has m/z = 138. According to the mass spectrum data, the average content of deuterium in [D]mexidol was 4.50 deuterium atoms per [D]mexidol molecule. In this sample, the percentage of molecules containing 3 to 6 deuterium atoms was 88%. The content of unlabeled molecules in this [D]mexidol was less than 2%.

Fig. 2.

Mass spectrum for [D]mexidol obtained via HSCIE at 80°C, 5 min; (monoisotopic ions at m/z = 138).

The regioselectivity of HSCIE between mexidol and SH was analyzed via NMR spectroscopy. Figure 3 shows the ¹H NMR spectra for the initial light [H]mexidol and the heavy [D]mexidol obtained via 5 min of HSCIE at 80°C. According to the NMR data, the average content of deuterium in [D]mexidol was 4.45 deuterium atoms per molecule. Data on the distribution of deuterium over the C–H positions of the molecule are given in Table 1.

Fig. 3.

¹H NMR spectra for [D]mexidol: (top) the initial light [H]mexidol and (bottom) the heavy [D]mexidol, (80°C, 5 min). The ¹H resonance assignment for mexidol is shown above peaks.

The activation energy of HSCIE was calculated from the rate of hydrogen/deuterium substitution in the aromatic and aliphatic moieties of the mexidol molecule as a function of temperature. Experimental E_a of HSCIE were 9.1 kcal/mol for the aromatic H(8) atom and 13.0 kcal/mol for the aliphatic H(16, 17) atoms (Table 1). A mechanism of isotope exchange between mexidol and model acid sites, HAl(OH)₄, HH⁺Al(OH)₄, H(–Al–O–)₃(OH)₆, and HH⁺(–Al–O–)₃(OH)₆, was considered on the basis of DFT calculations. The structures of transitional states were determined, and the activation energies of hydrogen exchange between model acid sites and hydrogen atoms in the mexidol molecule were calculated (Table 1).

RESULTS AND DISCUSSION

The determined activation energies of HSCIE for the aliphatic H(16, 17) atoms of mexidol were close to the experimental E_a values for isotope exchange in glycine and aminoisobutyric acid, which were 15.3 and 13.6 kcal/mol, respectively [13]. The E_a of hydrogen/deuterium exchange in the HSCIE reaction was 14.5 kcal/mol for the glycine fragment of cyclopropylglycine [22].

H₃O⁺ and HAlCl₄ acid centers were considered earlier as a model of acid catalytic sites that form under the action of SH upon consideration of HSCIE [13, 22]. The quantum chemical calculations of E_a of hydrogen exchange on these acid sites correspond to the experimental data for the HSCIE reaction. The data of DFT calculations on the acid catalytic HAlCl₄ center obtained for the hydrogen exchange reaction in cyclopropylglycine agree with the experimental data on the stereo- and regioselectivity of hydrogen/deuterium HSCIE on alumina-supported cyclopropylglycine [22]. To further refine the mechanism of HSCIE with SH, in this work we studied the exchange of hydrogen on model catalytic centers represented by H(–Al–O–)₃(OH)₆ and HH⁺(–Al–O–)₃(OH)₆ alumina clusters [23]. The HSCIE reaction in the alumina-supported mexidol was thus studied experimentally and the hydrogen exchange reaction was studied by MP2 and DFT B3LYP/6-31G* calculations with full energy optimization and calculation of normal-mode frequencies.

DFT calculations of the hydrogen exchange reaction between mexidol and the model catalytic centers HAl(OH)₄, HH⁺Al(OH)₄, H(–Al–O–)₃(OH)₆, and HH⁺(–Al–O–)₃(OH)₆ allowed us to determine the structure of transition states and calculate the activation energies (Table 1). It is clear from the data of Table 1 that the calculated E_a value for the uncharged model catalytic centers exceeds the one determined experimentally, and the presence of a positive charge on the model catalytic center resulted in a dramatic change in E_a to the level of experimental values. The DFT B3LYP/6-31G* calculated data on E_a for the hydrogen exchange reaction between the acid sites and the hydrogen atoms of the mexidol molecule are given in Table 1. Figure 4 shows the structure of the model positively charged HH⁺(–Al–O–)₃(OH)₆ catalytic site that formed after adding protons to the H(‒Al–O–)₃(OH)₆ site.

Fig. 4.

Structure of the model of a charged catalytic HH⁺(–Al–O–)₃(OH)₆ center.

Quantum chemical calculations of hydrogen exchange between the organic compound and the catalytic center were performed for the catalytic HH⁺(‒Al–O–)₃(OH)₆ center. Figure 5 shows the structure of the transitional state of hydrogen exchange between the charged HH⁺(–Al–O–)₃(OH)₆ center and mexidol. According to DFT calculations, the activation energy of hydrogen transfer between the aromatic H(8) atom and the charged catalytic center was 10.78 kcal/mol. Similar DFT calculations of the activation energy were performed for aliphatic hydrogen atoms in the H(16) position, yielding a value of 11.53 kcal/mol. The DFT calculated activation energy for the reaction between hydrogen in the H(8) and H(16) positions of mexidol and the model charged catalytic HH⁺(–Al–O–)₃(OH)₆ center agrees with the experimentally determined E_a of hydrogen/deuterium HSCIE (Table 1).

Fig. 5.

Transition state of hydrogen exchange between the H(8) atom of mexidol and alumina cluster HH⁺[–Al–O–]₃[OH]₆ with a positive charge.

CONCLUSIONS

A mechanism of HSCIE was proposed on the basis of experimental data for the solid-state hydrogen/deuterium exchange reaction in mexidol plus DFT calculations for the hydrogen exchange reaction on alumina catalytic centers. The mechanism of HSCIE can be described as a single-site synchronous hydrogen exchange between the adsorbed compound and Brønsted-type acid catalytic centers formed on the surface of the support upon exposure to SH.