1 Introduction

Widespread monoterpenoids form the largest group of natural compounds are extensively applied as platform molecules in food, pharmaceutical and perfumery industries [1]. Complicated structures of monoterpenoids with several functional groups and asymmetric centers allow synthesis of a broad range of the products. On the other hand, development of selective catalytic processes for monoterpenoids transformations is a challenging task because of their high reactivity. Hydrogenation of monoterpene oximes is one of the key steps in the synthesis of valuable compounds, including those with carbonyl and amino groups.

Previously it has been found in our work that Au/TiO2 catalyst promoted an one-pot synthesis of dihydrocarvone comprising sequential transformations of carvone oxime to dihydrocarvone [2], which is a novel approach to obtain a valuable additive in food industry [3]. Application of Au/TiO2 catalyst with a low hydrogenation activity provided carvone oxime deoximation followed by selective hydrogenation of the conjugated C=C bond. At the same time carvylamine, which can be formed as a result of oxime group hydrogenation [4], has not been detected in the reaction mixture in the presence of Au/TiO2 catalyst. Such amines are, however, of interest for the synthesis of biologically active compounds [5, 6] and molecularly imprinted polymers [7].

Reduction of menthone oxime with a simpler structure leads to formation of menthylamine [4], which is of great practical interest due to its wide application in asymmetric catalysis [8,9,10] and synthesis of biologically active compounds [11, 12]. The methods for menthylamine synthesis, both synthetic and catalytic ones provided the desired product yield not exceeding 40–60% [4, 13, 14]. Only platinum black in glacial acetic acid and Raney nickel in methanol were proposed as catalysts for menthone oxime reduction. Generally, Pt catalysts exhibited rather high activity in cyclohexanone oxime hydrogenation to the amine compared to Pd, Ru and Rh catalysts [15]. Recently heterogeneous Pd complexes were shown to be highly active for the hydrogenations of oximes to primary amines under hydrogen atmosphere [16].

In the current study Pt catalysts were selected aiming at the development of direct approaches for monoterpenoid amines synthesis via oximes hydrogenation, being important step in fine organic synthesis. Monocyclic monoterpenoid oximes, carvone and menthone oxime, obtained from corresponding natural terpenoids were used to reveal fundamentals of the hydrogenation of oximes over Pt catalysts.

2 Experimental

2.1 Preparation of Catalysts

Platinum catalysts supporting on commercial powdered metal oxides, including TiO2 (Degussa AG, Aerolyst 7708), ZrO2 (Alfa-Aesar), Al2O3 (Sasol), MgO (Vekton), were prepared by the impregnation methods using an aqueous solution of H2PtCl6 (0.1 M) (Krastsvetmet, Russia). The measured BET surface areas of the supports are presented in Table 1. After impregnation procedure the samples were dried at 110 °C for 12 h. Before the catalytic tests, the samples were reduced in hydrogen flow increasing temperature to 350 °C with a temperature ramp of 2 °C/min during 3 h to remove the excess of chloride.

Table 1 Characterization of 2 wt% Pt catalysts and the corresponding supports
Full size table

2.2 Characterization of Supports and Catalysts

The texture of metal oxides was evaluated by N2 adsorption in a Micromeritics TriStar II-3020 device (Micromeritics Instrument Corp., USA). The samples were degassed in vacuum at 300 °C for 4 h using a Micromeritics VacPrep 061-Sample degas system (Micromeritics Instrument Corp., USA).

The platinum content in the catalysts was determined by X-ray fluorescence spectroscopy (XRF) with the powder pellet method using an ARL PERFORM’X spectrometer equipped with a rhodium anode X-ray tube (Thermo Fisher Scientific, USA).

The average size of platinum nanoparticles was analyzed by TEM using JEM-2010 microscope (JEOL, Japan) with a lattice resolution of 0.14 nm at an accelerating voltage of 200 kV. More than 150 particles were measured to provide reliable metal particles size distribution.

X-ray photoelectron spectra of the fresh catalysts were recorded using a SPECS Surface Nano Analysis GmbH (Germany) equipped with PHOIBOS-150 hemispherical electron energy analyzer, FOCUS-500 X-ray monochromator, and XR-50 M X-ray source with double Al/Ag anode. Monochromatic Al Kα radiation (hν = 1486.74 eV) and energy of fixed analyzer pass of 20 eV under ultra-high-vacuum conditions were used to obtain the core-level spectra. The binding energy of peaks was calibrated by setting the Ti2p3/2 peak at 459.0 eV, the Al2p peak at 74.5 eV, the Zr3d5/2 peak at 182.5 eV and the Mg2p peak at 50.1 eV for titania-, alumina-, zirconia- and magnesia-based catalysts, respectively. The Shirley method was applied for the spectra fitting after the background subtraction using CasaXPS software.

2.3 Catalytic Experiments

Methanol (JT Baker, ≥ 99.8%), (−)-carvone (Aldrich, ≥ 97%), L-menthone (SAFC, trans-/cis-isomer = 85/15) were purchased from commercial suppliers and used as received. (−)-Carvone oxime was synthesized starting from (−)-carvone according to the method presented in [2]. (−)-Menthone oxime was obtained according to the method presented below.

In a typical experiment, a mixture of oxime (1 mmol), methanol (10 ml) and the catalyst (< 63 μm, 0.150 g, the active metal to substrate = 1.5 mol%) was intensively stirred (1100 rpm) at 100 °C under H2 atmosphere (7.5 bar) in a batch reactor. Based on our previous estimation of the Weisz-Prater criterion the experiments were performed in the kinetic regime [2]. During the reaction the samples with the volume of ca. 0.1–0.2 ml were withdrawn and analyzed by gas chromatography using HP-5 column (length 30 m, inner diameter 0.25 mm and film thickness 0.25 μm) and flame ionization detector operating at 300 °C, helium carrier gas (flow rate 2 mL/min, split ratio 5:1), temperature range from 120 to 280 °C, heating 20 °C/min. The products were confirmed by gas chromatography-mass spectrometry (Agilent Tech. 7890 A gas chromatograph with an Agilent 5975C quadrupole mass spectrometer, HP-5MS column, length 30 m, inner diameter 0.25 mm and film thickness 0.25 μm, helium carrier gas, flow rate 1 mL/min, split ratio 10:1, temperature range from 50 to 280 °C, heating 15 °C/min). 1H- and 13C-NMR spectra were recorded using a Bruker DRX-500 spectrometer 500.13 MHz (1H) and 125.76 MHz (13C) in CDCl3 and Bruker AV-400 spectrometer 400.13 MHz (1H) and 100.61 MHz (13C) in CDCl3.

TOF was calculated as a number of converted oxime moles per mole of exposed catalytic site per unit time during the first hour of the reaction corresponding to the linear part of the kinetic curves according to the following equation: \({\text{TOF}} = ({{\text{n}}^{0} - {\text{n}}})/{{\text{n}}_{\text{Me}}} \cdot {\text{D}} \cdot t\), where \(n^{0}\) and \(n\) are the initial and after 1 h molar amounts of oxime, nMe (mol) is the active metal amount in the catalyst, D is the metal dispersion assuming their spherical shape and t = 1 h is the reaction time.

2.4 Synthesis of (2S,5R)-2-Isopropyl-5-Methylcyclohexanone Oxime (Menthone Oxime)

To the solution of L-menthone (9.99 g, 64.8 mmol) in ethanol (40 ml), hydroxylammonium chloride (5.99 g, 86.2 mmol), sodium carbonate (8.93 g, 84.3 mmol) and water (40 ml) were added. The obtained mixture was stirred at room temperature for 5 days. Then ethanol was distilled off, the mixture was diluted with water (40 ml) and extracted by hexane (50 ml) three times. The combined organic layer was washed with brine (30 ml), dried over sodium sulfate and evaporated. The synthesized menthone oxime was purified by column chromatography (SiO2 60–230 μ (Macherey–Nagel), the eluent was EtOAc/hexane (0–3%) yielding 6.68 g (39.4 mmol, 61%) of menthone oxime as the single trans-stereoisomer. NMR 1H and 13C spectra are presented in Supplementary Materials.

3 Results and Discussion

3.1 Catalysts Characterization

The Pt content in the synthesized catalysts determined by XRF was close to the theoretical one (2 wt%). TEM micrographs and histograms of the particle size distribution are presented in Fig. 1. According to TEM data Pt nanoparticles were characterized with a uniform distribution of spherical metal nanoparticles. Pt average nanoparticles size seems to depend on the support nature and its textural properties. Thus, titania and magnesia with a lower BET surface area compared to alumina and zirconia were characterized with Pt nanoparticles with the average size of 2.0 and 2.2 nm, respectively (Table 1, Fig. 1). The average particle sizes of Pt nanoparticles were 0.8 and 0.9 nm for Pt/Al2O3 and Pt/ZrO2 catalysts, respectively (Table 1, Fig. 1).

Fig. 1
figure 1

TEM micrographs and histograms of synthesized Pt catalysts: a Pt/MgO, b Pt/Al2O3, c Pt/ZrO2, d Pt/TiO2

Full size image

The chemical state of platinum was studied by the XPS. The Pt4f core-level spectra are shown in Fig. 2. Typically, Pt4f spectra were characterized by Pt4f7/2-Pt4f5/2 doublet, where the spin–orbit splitting is 3.33 eV. The XP spectra of Pt/MgO, Pt/Al2O3 and Pt/ZrO2 catalysts were approximated by one Pt4f7/2–Pt4f7/2 doublet with the binding energy values for Pt4f7/2 at 72.6–72.9 eV, which can be assigned to Pt2+ species [17,18,19,20]. The latter seems to be present on MgO, Al2O3 and ZrO2 surfaces in the form of PtO, being caused by covering of small Pt nanoparticles with the oxide layer [21, 22]. At the same time Pt/TiO2 catalyst was characterized by one asymmetric doublet with the binding energy values for Pt4f7/2 at 70.7 eV. A low value of the binding energy as well as peak asymmetry is a feature of platinum in the metallic state [23, 24]. In the case of titania Pt nanoparticles decoration by the support (TiOx fragments) can stabilize the metallic state [25, 26].

Fig. 2
figure 2

Pt4f core-level XPS spectra of Pt catalysts (symbols—experimental data, curves—fitting)

Full size image

3.2 Catalytic Results

First platinum catalysts supported over the metal oxides (TiO2, ZrO2, Al2O3, MgO) were used for menthone oxime hydrogenation. The main idea was to explore feasibility of Pt catalysts application for monoterpene oxime hydrogenation to amine and to determine fundamental regularities for hydrogenation of monoterpene oxime with a simpler structure in terms of functional groups capable of competitive hydrogenation. During the reaction the desired menthylamine as well as trans- and cis-menthone were detected in the reaction mixture (Fig. 3). The menthylamine has several stereocenters resulting in diastereoisomers formation. Based on NMR analysis it was observed that two of the three stereocenters were involved in catalytic processes resulting in four diastereoisomers—neomenthylamine, menthylamine, isomenthylamine, neoisomenthylamine (Fig. 3). In the current study, the overall selectivity to menthylamine was mainly determined, while a detailed investigation of the reaction kinetics will be reported separately. Nevertheless, stereoselectivity for both menthone and menthylamine seems to be determined by thermodynamics.

Fig. 3
figure 3

Scheme of platinum-catalyzed transformation of menthone oxime under hydrogen atmosphere

Full size image

Note that the metal oxides per se catalyzed only menthone oxime deoximation to trans- and cis-menthone with small to moderate conversion (Table 2). The catalytic activity was noticeably lower compared to the corresponding Pt catalysts and no menthylamine was observed.

Table 2 The catalytic properties of Pt catalysts in menthone oxime hydrogenation after 7 h
Full size table

In the case of supported Pt nanoparticles, the catalytic activity and selectivity depended significantly on the metal oxide type (Table 2). Platinum supported on magnesia promoted menthone oxime hydrogenation with a higher activity and predominant formation of the desired menthylamine (Table 2). At the same time slightly less active alumina supported platinum catalyst provided more selective menthylamine formation due to absence of side transformation. Platinum nanoparticles on TiO2 and ZrO2, having an average particle size (2.0 and 0.9 nm) similar to MgO and Al2O3 (2.2 and 0.8 nm), respectively, demonstrated completely different activity and selectivity, producing menthone in a quite high yield. Pt/ZrO2 catalyzed menthone oxime transformation with a significantly lower activity and produced menthylamine with a slightly higher selectivity compared to menthone. The substrate deoximation to menthone dominated over Pt/TiO2 catalysts. Probably, menthone oxime hydrolysis via hydroxyl species transfer from the support can lead to menthone formation. Interestingly, no deoximation over Pt/TiO2 and TiO2 proceeded without hydrogen.

In our previous study on one-pot catalytic transformation of carvone oxime to dihydrocarvone via formation of carvone as an intermediate, it was demonstrated that deoximation of the oxime requires both the presence of the gold catalyst and molecular hydrogen [2]. At the same time water addition to the reaction mixture (1 vol%) under hydrogen atmosphere increased selectivity to the corresponding ketone, concomitant with a decrease in the reaction rate and selectivity to dihydrocarvone. Based on the obtained results and the recent publication [16] it can be tentatively proposed, that the oxime hydrogenation proceeds via imine formation followed by its further hydrogenation or hydrolysis depending on the nature of the catalyst.

According to the literature, no structure sensitivity was observed for hydrogenation of a nitro group over Pt [27, 28]. Hydrogenation of nitrocompounds to amines was shown to proceed via an intermediate hydroxylamine derivative formation, while Corma et al. obtained oxime from α,β-unsaturated nitrocompounds with H2 using a gold catalyst [29, 30]. Thus, there could be some common features in the reduction of nitrocompounds and oximes, particularly, in the substrate activation on the catalyst surface. At the same time, in several studies hydrogenation of acetonitrile and imines over Pt catalysts [31] was demonstrated to be structure sensitive. In this work completely different catalytic activity was observed for Pt/TiO2 and Pt/MgO as well as Pt/Al2O3 and Pt/ZrO2 catalysts with almost the same average particle size in hydrogenation of the C=N bond in menthone oxime. In the case of competitive hydrogenation, it was demonstrated by Corma et al. that chemoselectivity over Pt catalysts can be regulated by controlling the metal particles morphology [27, 28], while the support also affected significantly the catalytic activity. The latter seems to be associated with the support influence on Pt rather than the support acidity/basicity.

On the other hand, the metal-support effects for Pt catalyzed hydrogenation of crotonaldehyde [32] and amides [33] were discussed in terms of the functional groups activation on the support surface. Thus, the strength of the C=O bond activation on the support surface affected both the catalyst activity and selectivity. Efficient activation of the functional group resulted in a direct hydrogen transfer from Pt nanoparticles. Among different Pt catalysts on metal oxide, including alumina and zirconia, those which are supported on titania were recognized to provide quite strong interactions with the carbonyl oxygen atom of an amide or an aldehyde. In this work, Pt/TiO2 transformed menthone oxime selectively to menthone under hydrogen atmosphere contrary to other catalysts that also gave menthylamine. Probably stronger interactions of Pt/TiO2 catalyst with the oxime group can change the reaction pathway. In addition, as proposed above, menthone formation can be a result of the oxime hydrolysis via hydroxyl species transfer from the support. Pt catalysts based on alumina with a more acidic character provided higher selectivity to menthylamine, whereas Pt nanoparticles on MgO, ZrO2 and TiO2 enhanced menthone formation.

As a next step synthesized Pt catalysts were used for carvone oxime hydrogenation aiming at synthesis of carvylamine. Contrary to menthone oxime, carvone oxime has a double bond conjugated with the oxime group and an isolated C=C bond, which does not have a steric hindrance from the oxime group. In the presence of platinum catalysts supported on Al2O3 and ZrO2 non-selective hydrogenation of carvone oxime was observed resulting in predominant 5-isopropyl-2-methylcyclohexanamine (amine 1) formation (Fig. 4). The reaction proceeded through first hydrogenation of the C=C bond in isopropyl substituent of carvone oxime (Fig. 5). Subsequent hydrogenation and deoximation of carvotanoacetone oxime gave amine 1 as well as carvotanoacetone and carvomenthones, respectively, with the former being the main product. Note that formation of 5-isopropyl-2-methylcyclohexanamines (four diastereoisomers) was observed by NMR [32]. A higher catalytic activity and overall selectivity to amines 1 was observed in the presence of Pt/Al2O3 catalyst (Table 3). Similar to menthone oxime platinum supported on MgO, ZrO2 and TiO2 enhanced deoximation. Pt/TiO2 and Pt/MgO catalysts initially promoted hydrogenation of carvone oxime to carvotanoacetone oxime, followed by mainly deoximation (Table 3). As mentioned above, deoximation most likely proceeds via transfer of the hydroxyl species from the support. Contrary to other Pt catalysts the one with magnesia as a support was characterized by noticeable side transformations resulting in a complex mixture of ketones and dimers with a minor contribution of amines 1 (Table 3). In the case of carvone oxime transformations over Pt/TiO2 only different ketones were formed, being in a good agreement with the results on menthone oxime hydrogenation,. No amines 1 were detected during the reaction.

Fig. 4
figure 4

Catalytic activity of Pt/Al2O3 (a) and Pt/ZrO2 (b) catalysts in carvone oxime hydrogenation. The reaction conditions: T = 100 °C, p (H2) = 7.5 bar, carvone oxime 1 mmol, methanol 10 ml, catalyst 150 mg (Me/menthone oxime = 1.5 mol%). Amine 1—5-Isopropyl-2-methylcyclohexanamines

Full size image
Fig. 5
figure 5

Scheme of platinum-catalyzed transformation of carvone oxime under hydrogen atmosphere

Full size image
Table 3 The catalytic properties of Pt catalysts in carvone oxime hydrogenation after 7 h
Full size table

Compared to Au/TiO2 catalyst there was both hydrogenation of all reducible functional groups as well as deoximation in the presence of Pt catalysts, with activation of isopropyl fragment being preferred. The current results revealed that Pt catalysts are suitable for only 5-isopropyl-2-methylcyclohexanamine synthesis from carvone oxime via catalytic hydrogenation by molecular hydrogen. Gold nanoparticles supported on the same TiO2 were less active but provided selective carvone oxime deoximation to carvone followed by conjugated C=C bond hydrogenation leading to dihydrocarvone, whereas no carvylamine was also obtained during the transformation [2]. Worth to note that gold was almost not active in activation of C=C bond in isopropyl fragment. Obviously, carvone oxime is activated in a different way over the platinum catalysts compared to the gold ones. Particularly, earlier Corma et al. also recognized that the catalytic behavior of Pt is different compared to Au catalysts in hydrogenation of nitroaromatics [27, 28, 34, 35]. As mentioned above, hydrogenation of nitrocompounds was proposed to proceed via an intermediate hydroxylamine derivative formation. Corma et al. obtained oxime from α,β-unsaturated nitrocompounds as a final product [27, 28]. Typically, Pt interacts very strongly with different functional groups making the catalysts on the one hand highly active and on the other hand not chemoselective in hydrogenation of compounds with several unsaturations.

Thus, in the current work carvone oxime, containing several reducible functional groups and a conjugated oxime group, seems to be activated over Pt catalysts in the way leading to non-selective hydrogenation. Pt catalysts based on metal oxides were successfully applied for menthone oxime hydrogenation into valuable menthylamine. In particular, Pt/Al2O3 provided higher activity and selectivity to the desired amine.

4 Conclusions

In the current work Pt catalysts over metal oxides such as magnesia, titania, zirconia and alumina were used for monoterpene oximes hydrogenation to amines by molecular hydrogen. Monocyclic carvone and menthone oximes synthesized from natural monoterpenoids were utilized to explore the role of the substrate structure on the reaction regularities. Both hydrogenation of all reducible functional groups in carvone oxime and deoximation were observed over Pt catalysts on zirconia and alumina. Pt catalysts were shown to be suitable for only 5-isopropyl-2-methylcyclohexanamine synthesis from carvone oxime. Higher yield of 5-isopropyl-2-methylcyclohexanamine was achieved in the presence of Pt/Al2O3 catalyst. Platinum catalysts on ZrO2, MgO and TiO2 enhanced deoximation. Hydrogenation of menthone oxime over Pt catalysts led to the desired menthylamine formation. The catalytic behavior was noticeably influenced by the support nature. Pt/Al2O3 and Pt/MgO catalysts with the average particle size of 0.8 and 2.2 nm, respectively, exhibited higher activity in menthone oxime hydrogenation to menthylamine, with the former being more selective to the desired amine. Pt/ZrO2 and Pt/TiO2 catalysts with the average particle size of 0.9 and 2.0 nm, respectively, showed significantly lower selectivity to menthylamine at the expense of deoximation resulting in menthone formation.