1 Introduction

Presently hydrogen is regarded as one of the main energy sources for environment friendly transport. Herewith the solution of the problem of efficient H2 storage would result in a real alternative to liquid fuels (gasoline and diesel). A very effective approach to the hydrogen storage problem is based on the reversibility of hydrogenation–dehydrogenation reactions of aromatics and heterocycles. Condensed polycyclic aromatic hydrocarbons or heteroaromatic moieties containing nitrogen, oxygen or other heteroatoms are considered as substrates fort his method of hydrogen storage. The conjugated substrates can be pyrenes, coronenes, indenes, carbazole derivatives etc [1]. The pair of N-ethylcarbazole and perhydro-N-ethylcarbazole is one of the promising systems, because as much as 5.8 wt% of hydrogen can be taken up and released in the consecutive hydrogenation and dehydrogenation reactions [2]. Moreover, it was found [3] that the perhydro-N-ethylcarbazole dehydrogenation heat is about 50 kJ/mol Н2, that allows to carryout the dehydrogenation reaction at temperatures lower than 200 °С. Pd-containing catalysts supported on activated carbon, SiO2, and TiO2 were studied using this substrate pair [4,5,6]. The objective of this study is the study of the effect of modification of Pt/Al2O3 catalysts by some non-noble (Cr) and noble (Pd) second components on the activity of the perhydro-N-ethylcarbazole dehydrogenation reaction. The choice of the second metal is based on the ability to catalyze dehydrogenation reactions. Since chromium is much less active compared to palladium, the concentration of Cr in the catalyst was much higher (0.5 wt%) compared to that of Pd (0.1 wt%), but still lower than that of the main metal—Pt (1 wt%).

2 Experimental

2.1 Materials

A Pt/Al2O3 catalyst was prepared by incipient wetness impregnation of γ-Al2O3 (A-64 trademark, Ryazan Refinery, 0.1–0.2 mm fraction) using an H2PtCl6 aqueous solution followed by drying in air at 90 °С. Bimetallic Pt–Cr/Al2O3 and Pt–Pd/Al2O3 catalysts were prepared by incipient wetness impregnation of Pt/Al2O3 using (NH4)2Cr2O7 and [Pd(NH3)4]Cl2 aqueous solutions, correspondingly. After drying in air at 90 °С, the catalysts were calcined in air at 500 °С and reduced in an Н2 flow at 350 °С for 2 h. Then the catalysts were placed under Ar into a static reactor and perhydro-N-ethylcarbazole (H12–NEC) was added.

2.2 Dehydrogenation of perhydro-N-ethylcarbazole (H12–NEC)

The scheme of hydrogenation–dehydrogenation reactions of perhydro-N-ethylcarbazole (H12–NEC) is shown below:

figure b

The catalyst (0.25 g) and H12–NEC (1.0 g) were placed in a batch reactor (a glass vessel, 10 ml volume), the reactor was heated at ambient atmosphere up to 195 °C, and the reaction mixture was kept at this temperature with stirring at 600 rpm for 6 h.

2.3 Reaction Products Analysis

The liquid products were diluted by benzene and analyzed by means of gas chromatography (a “Crystallux” chromatograph) using a capillary column SE-30 (25 m) and a flame-ionization detector at a programmed temperature regime (70 °С isotherm for 4 min, then an increase of temperature, 8 °С/min up to 180 °С). The identification of liquid hydrocarbons was done using Thermo Focus GC (Thermo TR-5MS column) connected with Thermo DSQ II mass-spectrometer. Scanning range—35–350 amu, scan rate—5 per second. The gaseous products (generally H2 with С1–С2 admixtures) were collected in a gas burette (which is measured by the total amount of released hydrogen) and analyzed by using the same chromatograph at 30 °С. The dehydrogenation degree was calculated from the amount of H2 released by dehydrogenation divided by the amount corresponding to completed hydrogenation of 1 g of H12–NEC (650 ml of hydrogen).

2.4 Characterization

X-ray absorption spectra (Pt L3 edge at 11,564 eV and Cr K edge at 5989 eV) were measured in a transmission mode at the HASYLAB X1 station (DESY, Hamburg, Germany) using a Si(111) double crystal monochromator (detuned to give 50% of the maximum intensity for higher harmonics suppression). The experiments were performed with an in situ cell. In a typical experiment, a sample was heated in flowing He at 120 °C and diluted hydrogen (5% in He) to the first reduction temperature at 5 °C min−1 and then it was kept at this reduction temperature for 15 min (stationary reduction regime). After cooling to the liquid nitrogen temperature, the spectrum of the absorption coefficient µ was taken twice (recording simultaneously the spectrum of a Pt(Cr) foil between the second and third ionization chambers for energy calibration). Subsequently, the sample was heated to the previous reduction temperature at 10 °C min−1 and the next temperature was established with the rate 5 °C min−1. Data treatment was carried out using the software package VIPER [7]. A Victoreen polynomial was fitted to the pre-edge region for background subtraction. The smooth atomic background µ0 was estimated using a smoothing cubic spline. The Fourier analysis of the k2-weighted experimental function χ = (µ − µ0)/µ0 was performed with a Kaiser window. For the determination of structural parameters, theoretical references calculated by the FEFF8.10 code [8] for Cr-foil, Cr2O3, CrO3, Pt-foil, and PtО2 were used. To minimize the number of free parameters, equal backscatters were fitted with the same Eo shift wherever possible. The fitting was done in the k- and r-spaces. The shell radius r, coordination number N, Debye–Waller factor σ2 and the adjustable “muffin-tin zero” ΔE values were determined as fitting parameters. The errors in determining the fitting parameters were found by decomposition of the statistical χ2 function near its minimum, taking into account the maximal pair correlations.

The formal average Pt metal particle size was calculated using formula from [9], the calculation procedure was described earlier [10].

3 Results and Discussion

3.1 Dehydrogenation Reaction

According to chromato-mass-spectrometry data, the reaction mixture contained three intermediate products: H10-N-ethylcarbazole, H6-N-ethylcarbazole, and H2-N-ethylcarbazole together with totally dehydrogenated N-ethylcarbazole. For instance, in the case of the Pt–Pd/Al2O3 catalyst (195 °C, 6 h), the mixture contained H10-N-ethylcarbazole—5.8 wt%; H6-N-ethylcarbazole—1.9 wt%, and H2-N-ethylcarbazole—2.6 wt%, along with the desired product N-ethylcarbazole—89.7 wt%. Therefore the average conversion is below 89.7% (75%). This fact shows that during the dehydrogenation reaction, sequential formation of intermediate products takes place following subsequent dehydrogenation. The performances of the catalysts are presented in Table 1.

Table 1 Dehydrogenation of perhydro-N-ethylcarbazole at 195 °C catalyzed by mono- and bimetallic Pt-containing catalysts
Full size table

The introduction of a second component results in an increase of the catalytic activity compared to that of Pt/Al2O3 (Table 1). Both Pt–Cr/Al2O3 and Pt–Pd/Al2O3 catalysts provide after a prolonged time (over 10 h) the complete dehydrogenation of H12-N-ethylcarbazole, while the Pt/Al2O3 catalyst is not capable to overcome the limit of 71% under the chosen conditions (195 °C).

In addition, bimetallic catalysts show a higher selectivity toward the dehydrogenation reaction, with much lower concentrations of by-products (methane and ethane) in evolving hydrogen. This parameter was considered most important in view of the potential application of the bimetallic catalysts in cyclic hydrogen storage by consecutive hydrogenation of NEC and dehydrogenation of H-NEC, since the formation of light products as a result of cracking or hydrocracking/hydrogenolysis, like methane and ethane will reduce the potential hydrogen storage capacity of the substrates, especially if the life time of over 1000 or more cycles of hydrogenation–dehydrogenation is to be taken into account. Therefore the Pt–Cr/Al2O3 catalyst can be a preferred choice, even in comparison with the Pt–Pd/Al2O3 catalyst, because (1) it is characterized by much lower concentrations of light gases formed at the dehydrogenation stage, (2) chromium is a non-noble metal unlike Pd, (3) the overall activities of the Pt–Cr and Pt–Pd catalysts are comparable. Thus, this particular catalyst was studied in detail by XAS spectroscopy.

3.2 X-ray Absorption Spectroscopy

The bimetallic Pt–Cr/Al2O3 and monometallic Pt/Al2O3 catalysts were selected for the comparative study of the effect of the second component on the electronic state and local structure of platinum by X-ray absorption spectroscopy, because the effect of Cr is more pronounced in terms of the dehydrogenation selectivity and suppression of the formation of by-products (methane and ethane). In addition, the monometallic Cr-containing catalyst does not show any activity in this reaction.

The Pt L3 X-ray absorption near-edge spectra (Fig. 1) of the Pt/Al2O3 and Pt–Cr/Al2O3 samples reduced at 350 °С demonstrate that XANES spectra of these catalysts are similar to Pt-foil regarding the energetic position and “white” line intensity.

Fig. 1
figure 1

Pt L3 XANES spectra of Pt/Al2O3, Pt–Cr/Al2O3 and reference compounds

Full size image

Fourier transformation of the oscillating part of the spectrum (EXAFS, Fig. 2) of the Pt/Al2O3 catalyst differs from Pt-foil only in the double peak intensity. This demonstrates that platinum in this sample is completely reduced. At the same time, the EXAFS spectrum of Pt–Cr/Al2O3 shows that the first peak corresponds to a Pt–O atomic pair, whereas the second peak—to a Pt–Pt atomic pair.

Fig. 2
figure 2

Pt L3 EXAFS spectra of Pt/Al2O3, Pt–Cr/Al2O3 and reference compounds

Full size image

For fitting the EXAFS Pt L3 edge spectrum of the monometallic Pt sample, crystallographic data of Pt-metal were used. Fitting was done for one shell (double shell) that contains 12 Pt atoms at the 2.807 Å real distance.

The Pt L3 EXAFS spectrum of the Pt–Cr bimetallic sample was fitted with a two-shell model: Pt–O and Pt–Pt. Crystallographic data of PtO2 and Pt-metal were used.

The results presented in Table 2 for the Pt/Al2O3 and Pt–Cr/Al2O3 catalysts show the average coordination numbers (CN) and the real distance (r) in atomic pairs. The estimation of the average particle size based on the average coordination number depends on the particle morphology. It should be borne in mind that the estimate of the average size is very approximate.

Table 2 Best fit of Pt L3 EXAFS spectra for Pt/Al2O3 and Pt–Cr/Al2O3
Full size table

Using these parameters and with the assumption of a spherical shape, the formal average size (diameter) of Pt particle for Pt/Al2O3 catalyst 10–19 Å was obtained. It should be noted that the presence of the long-distance peaks (> 3.5 Ǻ uncorrected distance) (Fig. 2) in the EXAFS spectrum of this sample indicates that the sample contains large particles as well.

Fitting the spectrum of the bimetallic sample shows that the local structure of the central absorbing atom consists of 0.6 ± 0.1 O atoms at the 2.089 ± 0.003 Å real distance and 4.6 ± 0.6 Pt atoms at the 2.702 ± 0.010 Å real distance. Using parameters for the Pt–Pt shell, the formal average size (diameter) of Pt particles in the range of 6.2 ± 0.6 Å was obtained.

In addition, the difference in the energy position of the white lines in the Pt L3 XANES spectra of the foil and two catalysts (Fig. 1) confirms that the real size of the metallic particles follows the order: Pt-foil > Pt/Al2O3 > Pt–Cr/Al2O3.

It should also be noted that the appearance in the nearest environment of the central Pt atom of oxygen as the nearest neighbor nevertheless does not indicate the oxidation of platinum. The Pt L3 EXAFS spectrum for the sample dehydrated in-situ in a He flow at 120 °C demonstrates the presence of 4–5 oxygen atoms at a distance of 2.043–2.059 Å in the nearest environment of oxidized platinum. At the same time, the distance in the Pt–O atomic pair in the Pt–Cr/Al2O3 spectrum of the catalyst reduced in an H2 flow at 350 °C (Table 2) is 2.086–2.092 Å with CN = 0.5–0.7. We suppose that it is caused by the presence of O atoms that are connected with Cr3+ cations.

Cr K X-ray absorption near-edge spectra (Fig. 3) of the Pt–Cr/Al2O3 catalyst demonstrate that the XANES spectrum of the dehydrated sample differs from both Cr foil and Cr2O3 and is similar to the spectrum of CrO3. This finding confirms the existence of Cr6+ in a tetrahedral lattice. Reduction of the sample at 350 °C leads to the Cr6+ transformation into Cr3+ with the following change of the coordination to octahedral.

Fig. 3
figure 3

Cr K XANES spectra of Pt–Cr/Al2O3 and reference compounds

Full size image

Fourier transformation of the oscillating part of the spectrum (Fig. 4) that belongs to a dehydrated sample differs from that of both Cr2O3 and Cr foil. The EXAFS spectrum of the reduced sample differs from the spectra of all reference compounds and dehydrated sample.

Fig. 4
figure 4

Cr K EXAFS spectra of Pt–Cr/Al2O3 and reference compounds

Full size image

The results presented in Table 3 for the Pt–Cr/Al2O3 catalyst present the average coordination numbers (CN) and the real distance (r) in the first and second coordination spheres obtained by EXAFS spectra fitting using the scattering amplitude and phase of Cr2O3 and CrO3. Best-fit results for the dehydrated catalyst are comparable with the characteristics of CrO3. The crystallographic data of CrO3 consists of 1-st shell (Cr–O atomic para with CN = 1 at 1.575 Å), 2-nd shell (Cr–O with CN = 1 at 1.579 Å), 3-d shell (Cr–O with CN = 2 at 1.748 Å). For fitting we used 1-st shell only.

Table 3 Best fit of Cr K EXAFS spectra for Pt–Cr/Al2O3
Full size table

On the contrary, the coordination number obtained for the reduced sample is lower and the real distance is larger compared with the characteristics of Cr2O3. This fact argues the existence of isolated Cr3+ ions, i.e. a Cr2O3 phase does not form during the reduction. As to the distribution of Cr relative to Pt, we suppose that the presence of Cr results in Pt dilution. At this stage of research, we do not have any evidence of the Pt–Cr interaction that is required for the alloy formation. Thus, chromium ions introduced in the Pt/Al2O3 catalyst provide conditions for stabilization of smaller Pt nanoparticles that are more active in the dehydrogenation of perhydro-N-ethylcarbazole.

4 Conclusions

It has been found that a decrease of the Pt0 particle size in the Pt–Cr/Al2O3 catalyst compared with Pt/Al2O3 promotes an increase of the catalytic activity in the perhydro-N-ethylcarbazole dehydrogenation reaction. Thus, the use of bimetallic catalysts in reversible hydrogenation–dehydrogenation of heteroaromatic substrates can boost the performance of hydrogen storage. The further perspectives in designing efficient catalytic systems for the application in hydrogen storage based on reversible hydrogenation–dehydrogenation of aromatic substrates may be related to the use of hybrid nanomaterials, including core–shell nanoparticles supported onto mesoporous supports [11, 12].