1 Introduction

Biomass has been recently used as a cheap and abundant raw material for production of fuels and chemicals, for example via pyrolysing woody biomass and extraction of lignin to obtain among other products large amounts of phenolic compounds. Such compounds are not suitable directly as fuels due to their high oxygen content thus requiring deoxygenation. Bio-oil obtained by wood pyrolysis has a low pH, is unstable and contains some catalyst poisons. Subsequently catalytic deoxygenation is far from being straightforward. Upgrading of bio-oil and lignin extracts via hydrodeoxygenation (HDO) is currently under intensive research [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26] with both real bio-oil as well as model compounds used as a feedstock.

HDO of guaiacol in particular has been intensively investigated recently using, for example, Fe/SiO2 [20], Mo2N [3, 4], Mo2C/CNF [2, 13], MoC/AC [16], Mo2C/AC [6], W2C/CNF [13], Ni/ZrO2 [23], NiP/SiO2 [25], ReOx [5], Au/TiO2 and Au–Rh/TiO2 [17], Pt/C [8], Pt/alumina silicate [10], PtPd–Al–HMS [26], PtPd–ZrO2–SiO2, [26] Ru/C [1], Ru–TiO2–ZrO2 [15], Rh/ZrO2 combined with alumina silicate [12] as catalysts. HDO is a complex process involving several parallel and consecutive reaction steps, e.g. demethoxylation (1, 11, 16), dehydroxylation (2, 8, 13, 14), alkylation (4) and hydrogenation (3, 7, 12, 15) (Scheme 1; Table 1). Scheme 1 contains different reactions, which can be present during HDO depending on the reaction conditions and catalysts. For example, formation of cresol from phenol can occur through alternative pathways (methanation, reaction 4) and transalkylation with guaiacol giving cresol and catechol (reaction 17). The latter reaction proceeds through a surface methoxide intermediate.

Scheme 1
scheme 1

Reaction scheme for guaiacol transformation based on [4, 12,13,14]

Full size image
Table 1 Different reactions in Scheme 1
Full size table

The benefits of using carbides and nitrides as catalysts are their lower prices, moreover nitrides and carbides exhibit also metal-like properties [27]. However, synthesis of carbides and nitrides requires high temperatures, in the range of 700–1000 °C, giving typically materials with low specific surface area. For this reason, also supported carbides have been recently prepared and tested as catalysts [3, 13].

Guaiacol HDO is typically performed in the temperature range of 250–50 °C under hydrogen pressure varying from 20 to 55 bar (Table 2) [1, 4, 5, 10, 12,13,14, 17, 24].

Table 2 Literature data on guaiacol hydrodeoxygenation
Full size table

The results show that the main product is phenol when molybdenum and tungsten carbides are used as catalysts [4, 13], whereas supported Ni and ReOx catalysts gave cyclohexane as the main product [5, 24]. Several metal supported catalysts, such as TiO2–Pd–SiO2, Rh/ZrO2 and Au–Rh/TiO2 afforded also the deoxygenated product, cyclohexane [12, 14, 17], whereas Pt/C catalyzed only phenyl ring hydrogenation giving 2-methoxycyclohexanol as the main product [10].

Because catalyst deactivation can be a serious issue in HDO of lignin derived substituted phenols, a special emphasis in the current work was put on the mass balance closure. The latter is typically represented through the sum of the masses of reactant and products in the liquid phase determined by GC. Another issue related to the mass balance closure is analysis of the spent catalyst giving, in particular, information about adsorption of organic compounds on the catalyst surface and coking. Coke analysis has been scarcely reported in guaiacol HDO [8, 19, 25]. One example is temperature programmed oxidation of spent Ni2P/SiO2 catalyst showing formation of CO and CO2 mostly at ca. 460 °C [25]. Thermogravimetric analysis performed for the spent catalyst used in the gas-phase guaiacol HDO can be also mentioned showing ca. 20% weight loss in the case of Pt/C catalyst [8]. Coke was also estimated in gas phase HDO of guaiacol using Ni- and Fe-based commercial catalysts [19]. In the current work size exclusion chromatography (SEC) was used to qualitatively confirm the presence of oligomers. Even if oligomer analysis has been applied very rarely in the literature, one example related to HDO of bio-oil is worth mentioning [31] when gel permeation chromatography was utilized to analyze macromolecules.

Kinetics of guaiacol transformation in batch reactors has been very scarcely reported including reactions over W2C/CNF, Mo2C/CNF [13], Mo2N [3] and ReOx [5]. From the discussion above it can be concluded that while molybdenum nitride and carbide have been used in guaiacol HDO, the literature is void from the data for supported Mo2C on mesoporous supports. The aim in this work was thus to investigate kinetics of guaiacol transformation over 40 wt% Mo2C/SBA-15 prepared from hexamethylenetetramine ammonium heptamolybdate as a precursor. Mesoporous silica SBA-15 with the cavity size larger than 6 nm [28] was selected as it can be expected that Mo2C would be well dispersed affording easier diffusion of the reacting molecules to the active sites. Mo2C/SBA-15 has been prepared previously from MoO3/SBA-15 via the temperature programmed carburization method [29].

The novelty in this work regarding the materials is that Mo2C/SBA-15 has been prepared using hexamethylenetetramine ammonium heptamolybdate complex as a precursor while this precursor has been earlier used in the synthesis of Mo2C [30]. According to our knowledge Mo2C/SBA-15 has not been investigated as a catalyst in HDO of phenolic lignin derived model compounds. As a benchmark catalyst Pt/C was used to compare its behavior with molybdenum carbide. In addition to kinetic studies also thermodynamics of guaiacol HDO, not previously reported, was explored using the Gibbs–Helmholtz equation [32].

2 Experimental

Mesoporous silica with the SBA-15 structure was synthesized following the method reported by Zukal et al. [33]. Tetraethyl orthosilicate (TEOS) was used as a silica source; amphiphilic triblock copolymer P123 was applied as a structure directing agent. Synthesis of SBA-15 was performed at 95 °C during 66 h. The resulting solid was recovered by filtration, extensively washed with distilled water and dried at 80 °C overnight. The template was removed by calcination in air at 540 °C for 8 h (with the temperature ramp of 1 °C/min). Calcined extrudates were crushed using a laboratory jaw crusher and sieved to obtain a fraction 560 to 850 µm (Retsch AS 300).

The supported catalyst was prepared by the incipient wetness impregnation method from hexamethylenetetramine molybdate complex (HMT–AHM). The latter was synthesized according to the method reported by Afanasiev [34], who dissolved 86 g of hexamethylenetetramine (HMT) and 50 g of ammonium heptamolybdate (AHM) in 400 mL and 300 mL distilled water, respectively. Thereafter, these solutions were mixed and the final solution was left in air at ambient conditions overnight resulting in precipitation of crystals. These crystals were filtered using a filter paper and washed with deionized water, followed by drying in air at ambient temperature for 3 days. The incipient wetness method was used to impregnate the HMT–AHM precursor on the support (2 g SBA-15) followed by drying at 120 °C overnight. This catalyst was carburized with the same method as described above.

Temperature programmed reduction was done in the flow of a gas mixture of 20 vol.% CH4 in H2 (the flow rate 75 mL/min) during 3 h at 700 °C (the temperature ramp of 10 °C/min) with further passivation of the final material with the mixture of 1 vol.% O2 in Ar (75 mL/min during 2 h).

40  wt% MoxC–SBA-15 was reduced in pure H2 (flow rate ~ 50 mL/min) at 450 °C (heating rate 5 °C/min) for 13 h and cooled down to 170 °C under H2 flow and flushed with Ar for 5 min. About 10–15 g of a solvent was introduced to the reduced catalyst under Ar flow to avoid oxidation. The catalyst was stored in the solvent overnight prior to its use.

2.1 Catalyst Characterization Methods

Thermogravimetric analysis of the fresh and the spent catalysts was performed under nitrogen using SDT Q600 (V20.9 Build 20) instrument. About 6–8 mg of the sample was weighed, put in a platinum pan and heated from room temperature to 625 °C with a 10 °C/min ramp. The purge gas feed rate into the system was 100 mL/min.

The synthesised catalyst was characterized by X-ray powder diffraction (XRD) analysis performed with X-ray diffractometer D8 Advance Eco (Bruker AXS) equipped with SSD 160 detector using Cu Kα emission at λ = 1.54 Å. XRD patterns were collected in the range of 2θ values from 5° to 70° (0.021°/step, integration time of 0.5 s per step). The X-ray tube voltage was set to 40 kV and the current to 25 mA. Diffraction data were evaluated using the Diffrac.Eva V 4.1.1 software. Subsequently, crystalline phases were identified according to the Powder Diffraction database of the International Centre for Diffraction Data (ICCD PDF2).

The textural properties were characterized by N2-adsorption (BET) at 77 K performed with the gas sorption analyzer—Autosorb-iQ (Quantachrome Instruments).

In order to examine the catalyst morphology transmission electron microscopy (TEM) using JEM 1400 plus (JEOL) was applied. The acceleration voltage of 120 kV and the resolution of 0.98 nm for Quemsa II MPix bottom mounted digital camera were used. Scanning electron microscopy (Zeiss Leo Gemini 1530) was used to determine the crystal morphology.

Ammonia TPD was performed using Autochem 2910 with the following temperature program: heating to 450 °C for 1 h, flushing the catalyst at 450 °C with helium, cooling to 100 °C—ammonia adsorption at 100 °C for 60 min, flushing chemisorbed ammonia away with helium flow at 100 °C and then starting the temperature ramp 10 °C/min to 900 °C. The outlet was connected to MS recording the response from ammonia, water and CO.

Coke was extracted from the selected spent catalysts using heptane as a solvent similar to [35]. The extraction was performed by refluxing the catalyst, 10 mg in heptane solution, 10 mL, for 4 h under stirring.

2.2 Catalytic Tests and Analysis of Reaction Mixture

The reactions were carried out in 300 mL batch reactor (Parr Instruments) equipped with a stirrer and a sampling line (with a 5 μm filter in order to take only the liquid from the reactor) and cooling water circulation. The reactor was surrounded by an electrical heater. Argon (AGA 99.999%) and hydrogen (AGA 99.999%) gas bottles were coupled to the reactor system. The stirring speed during the reaction was 900 rpm to avoid external mass transfer limitations.

In a typical experiment the liquid volume was 50 mL, dodecane was used as a solvent together with 100 mg of the reactant and 50 mg of a catalyst. Two temperature levels, 200 °C and 300 °C were selected for investigations while the total pressure of 30 bar was used in all experiments. Hydrogen partial pressures were at 200 °C and 300 °C 29.3 bar and 24.8 bar, respectively.

40 wt% MoxC/SBA-15 was pre-reduced prior to experiments, while the commercial reference catalyst 5 wt% Pt/C (Degussa, F106, XKYF/W) was used without any prereduction. Total pressure of 30 bar was used also in the latter case.

The samples taken at different times from the reactor were analyzed by gas chromatography using DB-1 capillary column (Agilent 122-13e) with the length of 30 m, internal diameter of 250 µm and the film thickness of 0.50 µm. The flow rate of helium was 1.9 mL/min. The following temperature program was used: at 60 °C the temperature was held for 5 min, followed by ramp of 3 °C/min until 300 °C, where it was maintained for 1 min. The following chemicals were used in the experiments and quantification of the reaction products: guaiacol (Fluka, ≥ 98%), (1S,2S)-2-methoxycyclohexanol (Aldrich), methoxycyclohexane (Tokyo Chemical Industry Co., ≥ 98%), 2-methoxycyclohexanone (Tokyo Chemical Industry Co., > 95%), cyclohexanol (Sigma Aldrich, 99%), phenol (Sigma Aldrich, 99%), cyclohexane (Lab-Scan, 99%), cresol (Sigma Aldrich, > 98%), anisole (Sigma Aldrich, > 99%), 2-methoxy-4-methylphenol (Sigma Aldrich, > 98%) and benzene (Sigma Aldrich, > 99%). The unknown reaction products were identified with GC–MS using a similar column and a method used with GC.

Conversion of guaiacol was defined as:

$$X = \frac{{c_{0,G } - c_{i,G } }}{{c_{0,G } }} \cdot 100\% ,$$
(1)

were X is conversion and c0,G and ci,G are the concentrations of guaiacol at the beginning and at time t. The product selectivity (or molar fraction) is defined as moles of a product P divided by the sum of all products visible in GC:

$$Molar\;fraction = \frac{{n_{P } }}{{\sum n_{Pi} }} \cdot 100\% .$$
(2)

It should be pointed out here that although guaiacol was in some cases converted rapidly, the products were not visible in chromatograms due to its oligomerisation. The sum of the masses of the reactant and products in the liquid phase (SMLP) determined by GC mass balance is defined as follows:

$$SMLP = \frac{\sum m_{{Gi + Pi_{i,G} }}} {{{m_{0,G } }}} \cdot 100\%$$
(3)

i.e. the sum of the masses of guaiacol and all products visible in chromatograms at time t divided by the initial mass of guaiacol.

3 Results and Discussion

3.1 Catalyst Characterization Results

The BET specific area of 40 wt% MoxC–SBA-15 equal to 340 m2/g is in the same range as reported for 30 wt% Mo2C–SBA-15, which exhibited BET surface area of 403 m2/g [29]. The BET for the parent SBA-15 was 645 m2/g. MoxC–SBA-15 was composed of both 50 wt% Mo2C and 50 wt% MoC phases, with MoC exhibiting peaks at 36.4°, 42.3°, 61.3° corresponding to (111), (200) and (220) faces [36] as denoted in XRD results (Fig. 1).

Fig. 1
figure 1

Diffractograms of a SBA-15 and b 40 wt% MoxC–SBA-15. Notation: (asterisks) Mo2C and (open circles) MoC phases. The diffraction peaks of Mo2C and MoC have been taken from PDFs 71-0242 and 89-4305, respectively

Full size image

TEM images of the fresh and spent 40 wt% MoxC–SBA-15 catalysts used in guaiacol HDO at 200 °C and 300 °C under 30 bar showed no sintering with the average crystallite particle sizes for both materials being in the range of 6.1–7.3 nm (Fig. 2). In the fresh 40 wt% MoxC–SBA-15 catalysts some separate clusters with the size of 140 nm outside SBA-15 matrix are visible, but their amount decreased in the spent catalyst, which was also indicated by energy-dispersive X-ray analysis (EDXA) analysis giving Mo/Si ratio (discussed below).

Fig. 2
figure 2

Transmission electron microscopy images and histograms of the fresh (a and b), the spent (c and d) 40 wt% MoxC–SBA-15 catalysts in the reaction of guaiacol at 200 °C under total pressure of 30 bar and 40 wt% MoxC–SBA-15 catalysts (e and f) in the reaction of guaiacol at 300 °C under total pressure of 30 bar. The scale bar is 100 nm

Full size image

The fresh and spent 5 wt% Pt/C catalysts contained well dispersed spherical Pt particles with the average particle size of 3.1 and 3.2 nm, respectively indicating absence of sintering (Table 3; Fig. 3).

Table 3 Characterization results of the fresh and spent catalysts
Full size table
Fig. 3
figure 3

Transmission electron microscopy images and histograms of the fresh (a and b) and spent (c and d) 5 wt% Pt/C catalysts in the reaction of guaiacol at 300 °C under 30 bar total pressure. The scale bar is 50 nm

Full size image

EDXA results showed that the Mo/Si mass ratio decreased from the fresh 40 wt% MoxC–SBA-15 from 1.0 to 0.88 when the catalyst was used in guaiacol HDO at 300 °C (Table 3). This might indicate that a part of molybdenum was removed from the catalyst by leaching.

Thermogravimetry is one of the methods to investigate coke on the catalyst surface. Analysis of the spent Ru/C catalyst used in guaiacol HDO in [8] demonstrated maximally 35% weight loss from the spent catalyst in the temperature range of 100–1000 °C when the weight loss from the fresh catalyst was subtracted. In the current work the weight losses between 100 and 900 °C for the fresh and spent 40 wt% MoxC–SBA-15 used in guaiacol transformation at 200 °C and 300 °C were 3, 39 and 8%, respectively, clearly showing that especially at 200 °C a substantial amount of organic compounds was accumulated on the catalyst surface (Fig. 4a). In this case the catalyst was not very active, which will be further elaborated.

Fig. 4
figure 4

Thermogravimetrical analysis (TGA) of the fresh and spent a 40 wt% MoxC–SBA-15 and b 5 wt% Pt/C

Full size image

A commercial 5 wt% Pt/C (Degussa) was tested in this work for comparison. The metal dispersion and the cluster size were 36.2% and 3 nm, respectively. The catalyst is slightly acidic as pH of the catalyst slurry is 5.6 [37]. The weight loss from the fresh 5 wt% Pt/C catalyst, which contains also water release, is 14% in the temperature range between 100 and 900 °C, being 34% for the spent 5 wt% Pt/C catalyst applied in guaiacol transformation at 300 °C under 30 bar (Fig. 4).

In this work coke was extracted from the spent catalysts used in guaiacol transformation at 200 °C under total pressure of 30 bar hydrogen with refluxing heptane for 4 h and analyzed by SEC (Fig. 5). The amount of extracted oligomers from the spent 40 wt% MoxC–SBA-15 was high. When comparing the peaks in SEC chromatogram with the reference (polystyrene), it can be stated that the main oligomers over MoxC–SBA-15 contained mostly more than nine and five monomeric units.

Fig. 5
figure 5

Size exclusion chromatogram (SEC) of the extracted spent 40 wt% MoxC–SBA-15. The catalyst was used in guaiacol transformation at 200 °C under 30 bar total pressure

Full size image

Extensive coking of MoxC–SBA-15 can be explained by high temperature used in the current work for deoxygenation of guaiacol. Furthermore, the main product was benzene (see below). Demethoxylation and reductive hydroxylation can thus lead to catalyst coking.

3.2 Thermodynamic Analysis of Guaiacol Hydrodeoxygenation

In order to reveal thermodynamic consistency of the experimentally observed results calculations of the Gibbs free energy changes were made. Enthalpy (\(\Delta H_{r}^{0}\)) and Gibbs free energy (\(\Delta G_{r}^{0}\)) at standard conditions were calculated by following a thermodynamic approach, starting from the standard enthalpy (\(\Delta H_{f}^{0}\)) and Gibbs free energy (\(\Delta G_{f}^{0}\)) of formation from the elements derived from the database included in ChemCAD v.5.0 [38],

$$\Delta H_{r,j}^{0} = \sum\limits_{j} {\nu_{i,j} \cdot \Delta H_{f,i}^{0} } ,$$
(4)
$$\Delta G_{r,j}^{0} = \sum\limits_{j} {\nu_{i,j} \cdot \Delta G_{f,i}^{0} } .$$
(5)

The equilibrium constant of each reaction was calculated from its definition,

$$K_{j}^{0} = \exp \left( { - \frac{{\Delta G_{r,i}^{0} }}{RT}} \right).$$
(6)

The dependence of the reaction free Gibbs energy on temperature was included by implementing the Gibbs–Helmholtz equation valid at P = 1 bar (\(\Delta G_{r,j}^{\varPhi }\)).

$$\frac{{\Delta G_{r,j}^{\varPhi } (T)}}{T} = \frac{{\Delta G_{r,j}^{0} }}{{T^{0} }} + \Delta H_{r,j}^{0} \left( {\frac{1}{T} - \frac{1}{{T^{0} }}} \right).$$
(7)

Finally, to calculate the ΔGr,j at different pressures, the following equation was implemented.

$$\Delta G_{r,j} (P) = \Delta G_{r,j}^{\varPhi } + nRT\ln \left( {\frac{P}{{P^{0} }}} \right).$$
(8)

The calculated enthalpy and Gibbs free energy formation for each component are reported in Table 4. The data were retrieved from ChemCAD v.5.0 database directly. Starting from these values, the enthalpy and Gibbs free energy for each reaction (j) at standard conditions, equilibrium constants at standard conditions (\(K_{j}^{0}\)), enthalpy and Gibbs free energy at different temperatures and pressure were calculated. Two different temperatures were investigated (T1= 473.15 K, T2= 573.15 K) and the energy values were calculated also at P1= 25 bar.

Table 4 Enthalpy and Gibbs free energy for each reaction (j) (reaction numbers are shown in Scheme 1) at standard conditions, equilibrium constants at standard conditions (\(K_{j}^{0}\)), enthalpy and Gibbs free energy at different temperatures and pressures (T1 = 200 °C, T2 = 300 °C)
Full size table

The most feasible reaction is hydrogenation of guaiacol to 2-methoxycyclohexanol in Table 4 (step 12) at 200 °C under 1 bar hydrogen followed by reductive dehydroxylation of cyclohexanol and 2-methoxycyclohexanol (steps 8 and 11). The effect of hydrogen pressure on Gibbs free energy is clearly visible at both temperatures, i.e. the reactions became less favorable with increasing hydrogen pressure for all reactions. Hydrogenation of phenol and benzene became also thermodynamically not feasible at 300 °C under 25 bar (steps 7, 3). Experimental determination of 2-methoxycyclohexanone in the reaction at 300 °C under 30 bar total pressure in the current work agrees well with the thermodynamic calculations for step 10 showing that hydrogenation of 2-methoxycyclohexanone is not feasible at 300 °C under 25 bar. Table 4 also illustrates that alkylation of phenol with (step 4) is thermodynamically not feasibly, while there are no thermodynamic restrictions for formation of cresol by transalkylation of phenol with guaiacol.

3.3 Catalytic Results

Supported on SBA-15 molybdenum carbide was investigated in guaiacol transformation at 200 °C under 30 bar total pressure not, however, giving any products visible in GC chromatograms. This indicates strong adsorption of guaiacol on the catalyst surface as also confirmed by TGA and SEC analysis. The sum of the masses of reactant and products in the liquid phase determined by GC results was therefore equal to zero (Table 5).

Table 5 Experimental results for guaiacol HDO
Full size table

Supported 40 wt% MoxC–SBA-15 was at the same time active at 300 °C giving the final conversion levels of 62% after 240 min (Table 5). The sum of the masses of the reactant and products in the liquid phase determined by GC for 40 wt% MoxC–SBA-15 was only 37%. When the solid material determined in TGA was added to the sum of the masses of the reactant and products in the liquid phase determined by GC, it was still only 40%. These sums of the masses of reactant and products in the liquid phase determined by GC contain only compounds visible in the GC and the solid organic residue determined by TGA. In the extraction of the spent catalyst 40 wt% MoxC–SBA-15 from guaiacol transformation at 200 °C large amounts of oligomers were observed. The sum of the masses of the reactant and products in the liquid phase determined by GC obtained for molybdenum carbide catalyst supported on SBA-15 in the current work in guaiacol transformation at 300 °C under 30 bar total pressure is lower than the one reported by Jongerius et al. [13], namely 72% at 350 °C under 55 bar of hydrogen over Mo2C/CNF. Higher sums of the masses of the reactant and products in the liquid phase determined by GC were reported in [2] in the range of 74–93% in guaiacol HDO in water as a solvent over Mo2C supported on active carbon at 300 °C under 137 bar hydrogen. It has been reported that the mass balance closure increases with increasing hydrogen pressure [13], which is relatively low in the current work as compared to [13].

Concentration dependencies for guaiacol HDO over 40 wt% MoxC–SBA-15 at 300 °C as well as selectivity dependence on conversion are given in Fig. 6. Guaiacol transformation was very rapid already during heating of the reaction mixture. Thereafter, the catalyst retained its activity during the course of the reaction.

Fig. 6
figure 6

HDO of guaiacol over 40 wt% MoxC–SBA-15 at 300 °C under 30 bar a concentration dependencies and b molar fraction of products as a function of conversion. Symbols: guaiacol (filled squares), phenol (filled triangles), benzene (filled circles), cyclohexanol (open squares), and cyclohexane (open circles)

Full size image

The main product in guaiacol transformation over 40 wt% MoxC–SBA-15 catalysts at 300 °C at the beginning of the reaction was benzene, whereas its selectivity declined with increasing conversion and at the same time selectivity to phenol increased (Fig. 6b). Formation of phenol as the main product was also seen for Mo2C/CNF [13]. On the other hand, cresol was the second major product [13], whereas no cresol was obtained in the current work over 40 wt% MoxC/SBA-15. Formation of cresol requires transalkylation over acidic sites [39], which were not detected in this catalyst by ammonia TPD (not shown).

Selectivity towards aromatic products, phenol, and benzene decreased slightly from 94 to 86% during 4 h reaction when selectivity to cyclohexanol was increasing to a minor extent.

As a comparison to supported molybdenum carbide guaiacol HDO was also investigated over 5 wt% Pt/C both at 200 °C and 300 °C. Guaiacol reacted at 200 °C very rapidly during heating of the reaction mixture, since its conversion was already 19% after 1 min. The reaction rate between 1 and 120 min was 0.02 mmol/min/gcat. after which the catalyst was completely deactivated and the final conversion remained at 46% level after 240 min (Fig. 7).

Fig. 7
figure 7

Guaiacol transformation over 5 wt% Pt/C catalyst at 200 °C under 30 bar total pressure, a concentrations of different compounds and b molar fraction as a function of guaiacol conversion, notation for a and b guaiacol (filled squares), 2-methoxycyclohexanol (times), cyclohexane (open circles), 2-methoxycyclohexane (open squares), cyclohexanol (filled circles)

Full size image

This result is in accordance with the literature [10], reporting rapid guaiacol transformations at 250 °C with 87% conversion over Pt/C under 30 bar hydrogen. For Pd/silica alumina catalyst there was no conversion of guaiacol at 200 °C under 30 bar hydrogen and this catalyst started to exhibit some activity at 230 °C. In addition, PtPd–Al–HMS (mesoporous aluminosilicate) was active in guaiacol HDO surprisingly already at 200 °C under 50 bar hydrogen in 3 h in methanol as a solvent giving 80% conversion [26]. This result is not directly comparative with the current results due to the presence of methanol as a solvent and lower pressure. Transformations of guaiacol at 300 °C proceeded also very fast already during heating giving conversion of 51% in 1 min (Fig. 8). After the first minute the reaction rate was also very high, 0.06 mmol/min/gcat. in the time range of 1 to 120 min. Noteworthy is also that even after 240 min traces of guaiacol were visible in the GC analysis indicating catalyst deactivation.

Fig. 8
figure 8

Guaiacol HDO over 5 wt% Pt/C at 300 °C under 30 bar total pressure in hydrogen: a concentration profiles and b molar fraction of products as a function of guaiacol conversion. Notation: guaiacol (filled squares), phenol (filled triangles), cyclohexanol (open squares), cyclohexane (open circles), methoxy-cyclohexanone (plus) and 2-methoxycyclohexanol (times) and 2-methoxycyclohexane (asterisks)

Full size image

The sums of the masses of the reactant and products in the liquid phase determined by GC obtained in guaiacol transformation at 200 °C and 300 °C over 5 wt% Pt/C were 91% and 95%, respectively, reflecting minor accumulation of some organic materials on the catalyst surface as also shown in TGA. When the sum of the masses of the reactant and products in the liquid phase determined by GC in guaiacol transformation over 5 wt% Pt/C was calculated taking into account both GC results from the liquid phase and TGA results, the sum of the masses of the reactant and products in the liquid phase determined by GC was 100%. In addition, accumulation of organic material was more prominent at lower temperature, being aligned with the literature data on guaiacol transformation over Mo2C/CNF catalyst [13]. For comparison, when Pt/C was used as a catalyst at 250 °C the guaiacol conversion was 87%, and only totally 80% of products were observed [10].

The main product in guaiacol transformation over 5 wt% Pt/C was 2-methoxycyclohexanol both at 200 °C and 300 °C (Figs. 8, 9) analogously to the results in [10].

Fig. 9
figure 9

Comparison of experimental and calculated data for HDO of guaiacol over 40 wt% MoxC–SBA-15 at 300 °C under 30 bar

Full size image

From the time dependent concentration profiles, it can also be seen that deoxygenation occurs to a minor extent since only low amounts of cyclohexane were formed. Relatively large amounts of 2-methoxycyclohexanone were formed at 300 °C, whereas this intermediate was not visible at 200 °C. For formation of 2-methoxycyclohexanone two routes could be proposed, either starting from dehydrogenation of 2-methoxycyclohexanol or alternatively directly from guaiacol via reversible hydrogenation–dehydrogenation analogously to phenol–cyclohexanone route [40]. Experimental results show that as expected it is thermodynamically more difficult to hydrogenate 2-methoxycyclohexanone at 300 °C compared to 200 °C. Consecutive reactions, such as hydrogenolysis of 2-methoxycyclohexanol were not occurring to a large extent, which was also shown by thermodynamic analysis. The molar fraction of 2-methoxycyclohexanol remained at a constant level of 80% at 200 °C with increasing conversion (Fig. 7), whereas at 300 °C it decreased only slightly giving cyclohexanol via demethoxylation (Fig. 8; Scheme 1). At 200 °C selectivity to cyclohexanol was 2-fold higher compared to 300 °C indicating that demethoxylation occurs already at relatively low temperatures. Moreover, only minor amounts of cyclohexane (few %), a fully deoxygenated product, were formed over 5 wt% Pt/C catalyst. It can be concluded that a mildly acidic 5 wt% Pt/C promotes mainly hydrogenation of the phenyl ring thus consuming 3 mol of hydrogen per 1 mol of guaiacol and producing as the main product the non-oxygenated 2-methoxycyclohexanol.

A similar reaction pathway as mentioned above was observed for 40 wt% MoxC–SBA-15 catalysts at 300 °C when guaiacol very fast gave a constant concentration of benzene followed by a gradual transformation of guaiacol to 2-methoxycyclohexanol. This part of the overall reaction network can be thus essentially simplified leading to

$${\text{Guaiacol}}_{\text{ads}} \leftrightarrow {\text{Guaiacol}} \to 2{\text{ - methoxycyclohexanone}} \to 2{\text{ - methoxycyclohexanol}} .$$
(9)

In Eq. (9) it is suggested that there is strong adsorption of guaiacol on the catalyst forming some sort of its reservoir on the surface, which explains a lack of mass balance closure in the liquid phase at the beginning of the reaction and its increase when the reaction proceeds. In essence hydrogenation of guaiacol to 2-methoxycyclohexanone and 2-methoxycyclohexanol shifts this equilibrium involving strongly adsorbed guaiacol in favor of hydrogenation. At the beginning of the reaction guaiacol concentration decreased very rapidly giving benzene and strongly adsorbed guaiacol. With increasing reaction time 2-methoxycyclohexanol started to be formed with increasing the mass balance closure. The latter compound displays a clear S-shaped behavior pointing out that 2-methoxycyclohexanone was formed as an intermediate.

The rates for this scheme are given as

$$r_{G \to 2MCHO} = \rho_{B} k_{G \to 2MCHO} c_{G} ,$$
(10)
$$r_{2MCHO \to 2MCHL} = \rho_{B} k_{2MCHO \to 2MCHL} c_{2MCHO} ,$$
(11)
$$r_{G \to Gads} = \rho_{B} k_{G \to Gads} c_{G} ,$$
(12)
$$r_{Gads \to G} = \rho_{B} k_{Gads \to G} c_{Gads} ,$$
(13)

where ρB is the catalyst bulk density, G, Gads, 2MCHO and MCHL denote guaiacol, strongly adsorbed guaiacol, 2-methoxycyclohexanone and 2-methoxycyclohexanol.

The corresponding mass balances for each compound are given as

$$\frac{{dc_{G} }}{dt} = - r_{G \to 2MCHO} - r_{G \to Gads} + r_{Gads \to G} ,$$
(14)
$$\frac{{dc_{2MCHO} }}{dt} = r_{G \to 2MCHO} - r_{2MCHO \to 2MCHL} ,$$
(15)
$$\frac{{dc_{2MCHL} }}{dt} = r_{2MCHO \to 2MCHL} .$$
(16)

Because concentration of benzene did not change during the reaction it was included in modelling only indirectly. Namely the concentration of strongly adsorbed guaiacol was calculated by subtracting concentration of benzene, guaiacol, 2-methoxycyclohexanone and 2-methoxycyclohexanol from the initial concentration of guaiacol. The kinetic parameters were estimated using the backward difference method as a subtask to the parameter estimation with simplex and Levenberg–Marquardt methods implemented in software ModEst [41]. The objective function was defined as

$$\theta = \sum \left( {y_{i} - \hat{y}_{i} } \right)^{2}$$
(15)

and the coefficient of determination R2 is defined as

$$R^{2} = 1 - \frac{{\sum (y_{i} - \hat{y}_{i} )^{2} }}{{\sum (y_{i} - \bar{y}_{i} )^{2} }}$$
(16)

in which \(\bar{y}_{i}\) is the mean value of observations and \(\hat{y}_{i}\) denotes model estimation. The calculated values of parameters along with standard errors are given in Table 6.

Table 6 The estimated parameters for hydrogenation of guaiacol over 40 wt% MoxC–SBA-15 at 300 °C under 30 bar
Full size table

The model fit is shown in Fig. 9, illustrating a good overall description, which also follows from the value of the degree of explanation (97.22%).

4 Conclusions

Kinetics of guaiacol HDO was investigated in a batch reactor using 40 wt% MoxC–SBA-15 and 5 wt% Pt/C catalysts. Molybdenum carbide was prepared starting from hexamethylenetetramine molybdate as a precursor complex. Deposition of molybdenum carbide on SBA-15 resulted in both MoC and Mo2C phases and the specific surface area of 340 m2/gcat. being more than 100-fold higher than that for neat Mo2C. The average MoxC particle size in the supported catalyst was 7.3 nm.

Thermodynamic calculations according to the Gibbs–Helmholtz equation confirmed feasibility of guaiacol HDO to phenol in the range of 200–300 °C, used for experiments. Exothermic hydrogenation reactions in the reactions network being thermodynamically feasible at 200 °C became unfeasible at higher temperatures.

A comparative investigation using 5 wt% Pt/C and MoxC as catalysts clearly showed that a non acidic Pt/C catalyst was very active giving 98% conversion of guaiacol at 300 °C under 30 bar total pressure during 240 min. The main product with 5 wt% Pt/C was 2-methoxycyclohexanol indicating that it is not a suitable catalyst for production of hydrocarbons consuming also large amounts of hydrogen during hydrogenation of the aromatic ring. At the same time supported molybdenum carbide catalysts resulted in formation of phenol, benzene and cyclohexane in the liquid phase. Kinetic modelling was done quantitatively describing the concentration dependences with time.