Co-processing of fossil feedstock with lignin-derived model compound isoeugenol over Fe-Ni/H-Y-5.1 catalysts

Co-processing of n-hexadecane with lignin derived isoeugenol as a model compound was investigated in this work using low-cost mono-and bimetallic iron and nickel supported on H-Y-5.1 zeolite. Different Fe-Ni metal ratios in the catalyst led to different reaction rates of processes and product distribution. The presence of just 0.26 wt% isoeugenol in the mixture with n-hexadecane made hydroisomerization-hydro cracking of the latter two-fold less active. Catalysts with smaller metal particle sizes, lower than 6 nm were more efﬁcient pointing out on structure sensitivity. Extremely high activity in co-processing was obtained over 2 wt% Fe – 8 wt% Ni/H-Y-5.1 catalysts with the median metal particle size of 4.6 nm and metals-to-acid site ratio of 8.6. Fe catalyst were much less active in isoeugenol hydrodeoxygenation, while high cracking activity of hexadecane was observed in the presence of Ni. Alkylation of n-hexadecane was a feature of 8 wt% Fe – 2 wt% Ni/H-Y-5.1, whereas, over the 5 wt% Fe – 5 wt% Ni/H-Y-5.1 bifunctional catalyst no undesired oxygen-containing cyclic products were detected. This catalyst exhibited the highest hydrogen consumption according to temperature programmed desorption, which can serve as a marker for efﬁcient hydrodeoxygenation. The spent catalysts contained ca 40 wt% of coke with predominantly aliphatic species.


Introduction
Due to depletion of fossil-based feedstock and environmental concerns associated with fossil fuels, lignocellulosic biomass has sparked substantial attention as a source of renewable liquid fuels and chemicals [1][2][3].For decades, lignocellulosic biomass has been utilized in the Kraft process.The primary components of biomass, such as hemicellulose and lignin, are burned as a low-value fuel in this process, whereas cellulose is used to manufacture pulp.Due to its distinctive structure, lignin, which is currently underutilized, offers immense potential as a source of renewable fuels, chemicals, and aromatic compounds.Because lignin has a lower oxygen-tocarbon atomic ratio compared to carbohydrates, its derivatives could be used as fuel [2][3][4][5][6][7][8].
Fast pyrolysis and flash pyrolysis to convert lignocellulosic biomass to liquid fuels is often considered as a cost-effective method for biomass conversion requiring lower temperatures compared to e.g.gasification.Fe modified and acidic H-Y, H-Beta and H-Ferrierite microporous zeolite catalysts have been used for upgrading lignocellulosic biomass pyrolysis vapours to value added fuels and chemicals in a fluidized bed reactor [9].In the fast pyrolysis process, biomass is thermally degraded in the absence of oxygen at a moderate temperature ranging between 723 and 873 K with a residence time of less than two seconds.Through this procedure, the yield of condensed bio-oil including such compounds as aldehydes, ketones, alcohols, phenolics, and other phenolic compounds can reach up to 75%.In particular, guaiacol (ca.39%), phenol, anisole, cresol, syringol, and others are produced via lignin fast pyrolysis.Additionally, char, noncondensable gases such as carbon dioxide and methane, and water are also formed [2,10,11].
However, this bio-oil, as such, is unsuitable for the use as a liquid transportation fuel [12,13] due to low pH, instability as well as high O content and mediocre calorific value.A feasible technique for upgrading pyrolysis bio-oil pyrolysis for fuel production could be hydrodeoxygenation (HDO).In this reaction, oxygen-containing compounds are transformed to oxygen-free hydrocarbons and water in the presence of a catalyst at ca. 573 K and a high pressure of H 2 (up to 20 MPa) [2,14].Catalytic HDO reactions are often catalyzed by noble metals, transition metal carbides, bifunctional metals, and metal sulfides supported on metal oxides or carbon [4,[15][16][17][18][19].During catalytic HDO a variety of reactions occurs, including decarboxylation, hydrogenation, hydrocracking, hydrogenolysis, and dehydration, modifying the C-O-C, C-O-H, and C-C links [4].Because lignin-derived phenols contain a variety of functional groups such as alkyl, hydroxyl, alkoxy, olefinic double bonds, etc., there is a competition between HDO processes and double bond/aromatic ring hydrogenation reactions during upgrading of these compounds.Excessive ring hydrogenation is undesirable because of a significant hydrogen consumption in such reactions.
Several attempts have been made recently to promote the use of renewable energy and limit green-house gas emissions including energy and climate targets facilitating energy security to promote decarbonized economy and the use of biofuels of specific action plans have been established [41].One step forward to reach these targets is to utilize the existing infrastructure of the petroleum refineries via co-processing of renewable feedstock with fossil ones [42,43].Co-processing of bio-oils in fluid catalytic cracking has been intensively studied, however, several challenges have been identified due to a high water content in bio-oil [44].In order to understand the interactions between oxygenated and fossil feedstock during co-processing, also model compounds have been used [45], such as phenol, syringol, trimethoxybenzene with tetralin and benzene.Moreover, straight run gas oil (SRGO) has been coprocessed with e.g.guaiacol and anisole over CoMo/c-Al 2 O 3 catalyst in a fixed bed reactor [46].SRGO was also co-processed with different fractions of pyrolysis oil over Ru/C catalyst [47].Interestingly a competition between hydrodeoxygenation and hydrodesulphurization (HDS) steps when feeding oxygenated phenolic compounds with SRGO was observed, which was not, however, permanent, because with a neat SRGO feed HDS activity was recovered.
In the current work, hexadecane and isoeugenol were selected as a fossil and lignin derived model compounds, respectively.Hexadecane represents a model compound for hydrocracking and isoeugenol for hydrodeoxygenation.It should be pointed out here, that hexadecane can also originate from renewable sources, for example from non-edible oils [48], when it is not competing with the food chain.Hydrodeoxygenation of palmitic acid for production of hexadecane has been demonstrated over MoO 2 /CNT catalyst [49].In addition, hexadecane hydrocracking has been successfully demonstrated over several Ru-, and Ni catalysts for production of jet fuels [50][51][52].The reaction network of nhexadecane hydroisomerization-hydrocracking over bifunctional metal-acid catalysts [50,51,53,54] is shown in Fig. 1.
Isoeugenol was selected as a representative lignin-derived biooil model phenolic compound of oxygenated groups due to its molecular structure consisting of methoxy, hydroxy, and allyl groups in the aromatic ring [26].Isoeugenol is present in wood sawdust as a waste residue from oak and pine wood and, at the same time, hydrodeoxygenation of isoeugenol has been less intensively studied compared to other phenolic compounds.The reaction network of isoeugenol hydrodeoxygenation with potential side reactions based on the previous studies with Ni-, Co-, Pt-, Ir-, Re-supported catalysts [12,26,38,40] is shown in Fig. 2.
At 473 K and 3 MPa hydrogen, Bomont et al. [38] demonstrated that HDO of isoeugenol occurs promptly and completely, giving a high yield of the desired propylcyclohexane over Pt/H-Beta zeolites.At 523 K and 3-4 MPa hydrogen pressure, similar results were obtained in [40,55] over bimetallic Pt-Re and Ir-Re catalysts on alumina, a support bearing Lewis acidity.Disadvantages of these catalysts are the high prices of noble metals [4,19].In Ni, Co catalysts [12], no direct effect of the support material (zeolites, SiO 2 , Al 2 O 3 , TiO 2 ) on isoeugenol hydrodeoxygenation was observed.As a result, in the recent years, the research has mainly concentrated on the iron triad (Fe, Co, Ni) based catalysts as a less expensive option of the HDO process.Generally Fe-based catalysts show high activity in selective cleavage of C-OH, C-O bonds while exhibiting low activity in the ring hydrogenation.On the other hand, Co-and Ni-based catalysts demonstrated high activity in dihydrogen dissociation, and hydrogenation [12,26].Therefore, the bimetallic systems comprising Fe and Ni are potentially promising catalysts for co-processing.
The current work is focusing on the synergistic effect of Fe-Ni metals in the co-processing of hexadecane with isoeugenol.For this purpose, the monometallic catalysts and the bimetallic catalysts with a varied ratio between Fe and Ni were used with H-Y-5.1 zeolite as a support.The use of Ni-Fe is also encouraged by the recent studies demonstrating hydrodeoxygenation of waste cooking oil [56] and hydrocracking of coconut oil [57] over supported Ni-Fe catalysts.Taking into account that phenolic compounds are typically adsorbed stronger because of the aromatic ring, the straight chain alkane was used in a large excess.

Preparation and characterization of the fresh catalysts
To study the influence of the metal on the co-processing of nhexadecane with lignin derived isoeugenol, a series of Fe, Ni, and Fe-Ni supported bifunctional catalysts on a commercial H-Y-5.1 zeolitic support have been prepared (Table 1).All catalysts were synthesized by the incipient wetness impregnation with an aqueous solution of the corresponding nitrate precursor, Fe(NO 3 ) 3 Á9H 2 -O (Sigma-Aldrich) or Ni(NO 3 )2ÁÁ6H 2 O (CJSC Souzchimprom), respectively.Characterization was made by several physicochemisorption methods as described below.
Textural properties were determined with nitrogen physisorption (Micromeritics 3Flex-3500) degassing the samples first exsitu under vacuum at 473 K for 8 h followed by pre-treatment in the physisorption equipment per se under vacuum for 5 h at 453 K.The specific surface area and the pore size distribution were calculated using the Dubinin-Radushkevich and the non-local density functional theory methods, respectively.
Qualitative and quantitative determination of Brønsted and Lewis acid sites was done using pyridine (!99%, Acros Organics) as a probe molecule by Fourier transform infrared spectroscopy (FTIR, ATI Mattson).After pretreatment of a thin self-supporting pellet inside the FTIR cell at 723 K under vacuum and decreasing temperature to 373 K, the background spectra were recorded.Adsorption of pyridine was done for 30 min at 373 K followed by measurements after desorption at various temperatures.The interval between 523 and 623 K reflects all sites, 623-723 K correspond to medium sites, while spectra after desorption at 723 K were attributed to strong sites.Quantification of the Brønsted and Lewis acid sites was based on the spectral bands at 1545 cm À1 and 1450 cm À1 .Concentration of metals in the bulk of the catalyst was performed by inductively coupled plasma -optical emission spectrometry (ICP-OES, PerkinElmer Optima 5300 DV instrument) digesting the catalyst in a mixture of 30% HCl (Sigma-Aldrich), 65% HNO 3 (Sigma-Aldrich) and 50% HBF 4 (Sigma-Aldrich) mixture.
The phase purity and identification of the crystal phases was performed by powder X-ray diffraction (XRD) for pre-reduced exsitu catalysts using D8 Advance diffractometer (Bruker, Germany) with Bragg-Brentano geometry, CuKa radiation (k = 1.5418Å) and a one-dimensional LynxEye detector.The 2h-angle was scanned from 15°to 70°with 0.05°step.
The spectra of 57 Fe Mössbauer spectroscopy in the transmission geometry were recorded at 295 K with a 18-month-old 57 Co:Rh source (Ritverc Co. 50 mCi June 2020) with a maximum Doppler velocity of 11.0 mm/s.
For temperature programmed reduction Microtrac Belcat II equipment was used pre-treating a catalyst sample at 473 K for 2 h in argon followed by cooling to 323 K and subsequent heating under 5 vol% H 2 and 95 vol% Ar flow to 1073 K with a 5 K/min ramp.
X-ray absorption spectroscopy was used to probe the bulkaveraged element specific local structure around Fe and Ni atoms.XAS spectra at Fe and Ni K absorption edges were recorded at the P65 beamline of PETRA III synchrotron radiation source (DESY, Hamburg) in transmission mode for calcined and reduced bimetallic catalysts.The k 2 -weighted EXAFS functions were Fourier transformed (FT) in the k range of 2-14 Å À1 and multiplied by a Hanning window with a sill size of 1 Å À1 .

Catalytic tests
The catalytic experiments of co-processing were carried out in a batch reactor (PARR instruments, 300 mL).Isoeugenol (98%, mixture of cis and trans, Sigma-Aldich) and n-hexadecane (!99%, Alfa Aesar) were employed as the reactants.Based on the corresponding TPR profiles, the catalyst was activated in the flow of hydrogen (40 mL/min) in a two-step reduction procedure: 298-523 K (hold 2 h) and 523-773 K (hold 2 h) with the temperature ramp of 2 °C /min.
The catalyst screening experiments were performed at 573 K for 4 h and 3 MPa of H 2 (99.999%,AGA) with 100 mL of hexadecane, 200 mg of isoeugenol and 100 mg of the catalyst.The catalyst sieved fraction was below 63 lm to eliminate internal masstransfer limitations and the mechanical stirrer of the reactor operated at 1000 rpm to overcome external mass-transfer limitations.

Characterization of the spent catalysts
Thermogravimetric analysis (TGA) was used to determine coke on the spent catalyst using SDT650 instrument (TA Instruments).
In air flow of 100 mL/min, ca 5 mg of catalyst, loaded into 90 lL alumina cup, was heated up to 800 °C (held 5 min) with the heating ramp 10 °C/min.Results obtained for the spent catalyst were compared to the results for the fresh one.
The soluble hydrocarbons trapped in the spent catalyst (coke) have been identified based on the often used procedure, namely, a soluble coke extraction with CH 2 Cl 2 [60][61][62], introduced by Guisnet et al. [63].Approximately, 20 mg of the spent catalyst in the polyethylene vial was dissolved in 2 mL of 40% HF (Merck KGaA) for 5 h at ambient temperature and for 2 h in an ultrasonic bath.In the second step, a solution was neutralized by ca. 30 mL of 1 M solution of NaOH.Finally, the organics were extracted overnight by adding 1 mL of dichloromethane (CH 2 Cl 2 !99.9%, GC, Sigma-Aldrich).Extracts (organic phase) were analyzed by NMR and GC-MS (Agilent GC/MS 6890N/5973) using the same column and the same temperature program (held for 10 extra minutes) as for the reaction product analysis.Before NMR analysis of the extracted coke, first, dichloromethane was evaporated and as a solvent CDCl 3 was used. 1 H NMR analysis in the solution state was performed using a Bruker Avance-III spectrometer operating at 500.20 MHz ( 1 H) equipped with a Smartprobe: BB/1H was used.The reported signals are referenced to an internal standard (TMS dH = 0.0 ppm).

Definitions
The conversion of isoeugenol (X IE ) was calculated by: where n IE,0 is the initial isoeugenol molar concentration and n IE,t is the isoeugenol molar concentration at time t.The conversion of dihydroeugenol (X DHE ) was calculated by: where n DHE,t is the dihydroeugenol molar concentration at time t.The conversion of n-hexadecane (X HXD ) was calculated by: where n HXD,0 is the initial n-hexadecane molar concentration and n HXD,t is the n-hexadecane molar concentration at time t.Conversion in co-processing (X IE+HXD ) was calculated in the following way: where x IE,0 is the initial mass fraction of isoeugenol and x IE,0 is the initial mass fraction of n-hexadecane.
The liquid phase mass balance of cyclic compounds (MB) was calculated by: where MB t is the sum of liquid-phase reactants and products at time t, and MB 0 is the sum of liquid-phase reactants and products at the beginning of the reaction (time 0).The compounds not included in Eq. ( 5) are gas-phase products, and heavy compounds adsorbed on the catalyst (coke).
The carbon balance (CB) was calculated by: where n i,t is molar concentration of i compound at time t, and CN is the carbon number.
Yields of products (Y i,t ) were calculated by: The reaction rates (r) and turnover frequency (TOF) per metals and acid sites are calculated as follows: Where Dn=Dt denotes reacted moles (e.g.hexadecane of dihydroeugenol) per time interval Dt (one to 30 min) in a batch reactor, m cat is the catalyst mass.In the calculations of TOF and TOF* it was considered that both metals are active and different options of acidity were tried (e.g.total, Brønsted, and Lewis acid sites; weak, medium, and strong sites), respectively, where n metals(surface) = (n (Fe) Â dispersion (Fe) /100 + n (Ni) Â dispersion (Ni) /100) is exposed moles of metals, n acid sites is moles of acid sites.

Catalyst characterization results
Reduction profiles of the catalysts determined by TPR, indicating interactions between iron and nickel for the three fresh catalysts are given in Fig. 3.In the reduced Fe 2 Ni 8 and Fe 5 Ni 5 catalysts, quasi in situ X-ray absorption spectra showed a tight contact between Fe and Ni with formation of surface Fe-enriched fcc FeNi random alloys.However, the presence of alloy was not clearly confirmed by XRD in the reduced Fe 5 Ni 5 catalyst (Fig. S1).A higher uptake of hydrogen in Fe 2 Ni 8 compared to monometallic counterparts is probably related to different iron and nickel oxide speciation in these catalysts.
The metal contents determined by ICP-EOS, metal particle sizes (by TEM) and acidity from pyridine adsorption-desorption are summarized in Table 2, as they have a major impact on hydrodeoxygenation activity and selectivity and will be discussed in connection with the catalytic results.Worth noting is that introducing 5 wt% Fe on H-5.1 zeolitic support bearing mainly Brønsted acidity led to a decrease of strong acid sites and at the same time increased mainly medium Brønsted acid site, while introducing 5 wt% Ni led to significantly elevated weak and medium Lewis acid sites as will be discussed in a separate paper reporting in depth catalysts characterization.
In addition, Table 2 also displays characterization of the spent catalysts by TEM, N 2 physisorption and TGA.Metal particle sizes distribution of Fe, Ni and Fe-Ni/H-Y-5.1 fresh and spent catalysts is shown in Fig. S2 while SEM images of the spent ones are displayed Fig. S3.
SEM images confirmed presence of well-defined zeolite crystallites, which retain their morphological structures after the metal introduction as well as after the reaction (4 h).For all catalysts, a comparison of the median particle size of both fresh and spent catalysts confirmed no sintering of metal particles during 4 h of coprocessing at 573 K and 3 MPa of hydrogen.However, significant changes were observed in terms of the textural properties (Table 2).The maximum decrease in the specific surface area (more than 97%) and the pore volume (more than 85%) was recorded for the monometallic Fe catalyst, and for Fe 5 Ni 5 , which was used in isoeugenol hydrodeoxygenation without n-hexadecane (marked as IE + cat, Table 3).Other spent catalysts used in co-processing had the specific surface area and pore volume lower by 37-41% than the fresh ones.A larger decrease in textural properties (by 58-68%) was observed for Fe 2 Ni 8 used in n-hexadecane hydroisomerization-hydrocracking experiment without the presence of isoeugenol (marked as HXD + cat, Table 3).
This decrease of the specific surface area of the catalyst was correlated with the coke formation, determined by TGA (m coke ).Catalysts showing a lower decrease of the surface area revealed a higher amount of coke (Fig. 4a).At the same time, coke formation was increased with increasing Ni concentration and decreasing Fe concentration in the catalyst (Fig. 4b).A direct correlation with the FeNi particle sizes could not be established.However, it should be also mentioned, that in TEM images, nanoparticles of Fe and Ni could not be separated from each other in bimetallic catalysts.The observed effect of the metal nature on the coke formation is in line with [64] dealing with coke formation during the steam cracking of propane on foils of nickel and on a series of Ni-Cr-Ir alloys, showing that a substantially higher rate of coke formation was observed for the pure nickel foils.It should be also mentioned that deposited coke in the spent catalyst also linearly increased with decreasing amount of Fe 3+ species determined in the bulk by Mössbauer spectroscopy in the fresh catalyst (Fig. S4a).
For Fe-based and Ni-based alloys [65,66] it was reported that carbon deposited over time is related to small metal particles (ca. 10 nm) generated through nickel disintegration, which occurs via a direct inward or internal growth of graphite (graphitisation).
According to the heat release during TGA, the coke was mostly oxidized in the 440-550 °C temperature region (Fig. S4b).In [67] a similar temperature range, 400-530 °C, was obtained pointing on the predominance of amorphous filamentous type coke formation for 2.5-10 wt% Ni-UGSO catalyst (Nickel-UpGraded Slug Oxides) in glycerol steam reforming.
In the current work, the extracted soluble coke species were identified by NMR (Fig. S5) and GC-MS (Fig. S5).Both methods confirmed predominated aliphatic coke (primary compounds: C10-C16 straight-chain alkanes) and a low amount of aromatics (primary component: dimethylethyl benzene).In the cases of the  monometallic Ni and bimetallic Fe 2 Ni 8 catalysts, mainly nhexadecane with a small amount of n-dodecane, dimethylethylbenzene, n-tridecane, n-tetradecane and methylpentadecane were detected (Fig. S5a).On the contrary in the cases of the bimetallic Fe 5 Ni 5 and Fe 8 Ni 2 catalysts, a slightly dominant component was aromatic dimethylethylbenzene with a broad distribution of aliphatic cokes (Fig. S5b).Analysis of the spent catalyst by the solid state 27 Al NMR illustrated also changes with aluminium (Fig. S7).A peak at about À75 ppm has appeared being almost as large as the tetrahedral alumina peak at 59 ppm.The exact reason for this peak is not clear, but most probably it is associated with deposited coke and formation of Al-CR 3 species as proposed in [68].

Activity and selectivity of Fe, Ni and Fe-Ni/H-Y-5.1 catalysts
Co-processing of hexadecane with lignin derived isoeugenol as a model compound was investigated in this work using monoand bimetallic iron and nickel supported on H-Y-5.1 zeolite.In order to understand the co-processing behaviour and the role of each compound in the mixture, in addition to the catalyst screening experiments, a set of additional experiments were performed.They comprised thermal experiments without a catalyst and catalytic experiments with single compounds (Table 3).Thermal and catalytic results of such experiments were summarized in Table 4.
The thermal experiment with isoeugenol in n-hexadecane (HXD + IE) revealed almost zero conversion of n-hexadecane (X HXD = 0.03%) in the presence of even a small amount of isoeugenol (0.26 wt%, Table 3, Table 4, entry 3).On the contrary, hydrogenation of isoeugenol was accelerated in the diluted system at the expense of inhibited n-hexadecane transformations.However, it should be noted that a higher conversion of isoeugenol can be also related to a higher reaction temperature.The main products were the C8-C12 fraction (86%) consisting of 92% of dihydroeugenol and 8% of linear hydrocarbons.Selectivity to the isomerized hexadecane was only 9%.
The liquid phase mass balance and carbon balances of all thermal experiments without a catalyst was 100%.

Preliminary experiments: Catalytic experiments with single substrates
The results of preliminary catalytic experiments over catalyst are displayed in Table 4.After 4 h at 573 K of the solventless isoeugenol hydrodeoxygenation over Fe 5 Ni 5 catalyst (IE + cat), the total conversion of isoeugenol was achieved, wherein only 3% of the formed dihydroeugenol was further converted to propylphenol (Table 4, entry 4).Low activity of the deoxygenation step can be related to the high reactant-to-catalyst ratio (56, Table 3, entry 4) and also catalyst deactivation due to coke formation.The latter one is in line with the analysis of the catalyst textural properties.While the specific surface area of the fresh catalyst was 530 m 2 /g, it was only 15 m 2 /g for the spent one (Table 2).From the recording of the process data (Fig. 5), it can be concluded that the catalyst deactivation probably occurs already in the first seconds of the experiment, when the preheated reactant comes into contact with the hot dry catalyst.After injecting the reactant into the reactor with a fresh hydrogen feed, a temperature drop to 315 K was observed, followed by an increase to 333 K.
Hydroisomerization-hydrocracking of n-hexadecane over Fe 2 -Ni 8 catalyst (HXD + cat) revealed a relatively a high conversion, and the total yield of all products, 29%, with the content of branched and linear hydrocarbons 83 % vs 17 %, respectively (Table 4, entry 5).The main products were isomerized hexadecane (i-C16 = 48%, yield = 13.2%)followed by the jet fuels fraction of the cracked product (C8-C12 = 29%, yield = 8.6%).At the same conversion level of n-hexadecane, 30%, the same selectivity to i-C16, 49%, was obtained over 0.1 g Pt/ZSM-48 catalyst [54] and significantly lower, 20 % of i-C16 selectivity over Pt/ZSM-23 dual template [69] at the same temperature 573 K and a higher hydrogen pressure 4 MPa in a fixed-bed reactor.Over the monometallic 7.6 wt % Ni/H-MFI (18.5 nm) and 9.1 wt% Ni/Beta (11.4 nm) catalyst, selectivity to i-C16 was 0 and 56.6% at a significantly higher conversion of 95% and 64%, respectively, in the fixed-bed reactor at the same pressure 30 bar and lower temperature 548 K [70].For Ni/ZSM-48 catalysts exposed to different alkali treatment, selectivity to i-C16 was ca.25-35% at ca. 70-85% of conversion in the fixed-bed reactor at the same temperature [71].Compared to the thermal experiment, the cracked light hydrocarbons (<C8 of 17%) and small amounts of alkylated hexadecane (>C16 of 0.2%) were also detected in the current work.On the contrary, selectivity to the long-chain hydrocarbons decreased from 34% to 9%.Comparison with more expensive noble metal-supported catalysts [50] showed comparable results in terms of the yield of jet fuels fraction at the same conversion level of n-hexadecane, 30%.However, it should be noted that such comparison is not straightforward as the experiments were conducted under different conditions and the jet fuels fraction was defined by C9-C15 in [50].The experi-

Table 4
Catalytic results of preliminary experiments after 4 h.All data are presented in mol%.Notation: IE -isoeugenol, racemic mixture.DHE -dihydroeugenol, HXD -n-hexadecane, cat -catalyst, X -conversion, MB -the liquid phase mass balance, CB -carbon balance, > C16 -alkylated hexadecane, i-C16 -isomerized hexadecane, C13-C15 -long-chain hydrocarbons, C8-C12 -jet fuels fraction, < C8 -cracked light hydrocarbons, NL -nonlinear products (branched and cyclic), L -linear products.Co-processing of isoeugenol and n-hexadecane over Fe 2 Ni 8 catalyst (HXD + IE + cat) was performed with 0.1 g of Fe 2 Ni 8 catalyst (Table 3, entry 6).The amount of isoeugenol in n-hexadecane was the same as for the non catalytic experiment, i.e. 0.26 wt%.Full conversion of both isoeugenol and dihydroeugenol and 15.5% of n-hexadecane conversion was obtained after 4 h at 573 K and 3 MPa.In the case of hydrodeoxygenation, a significantly faster reaction can be mainly attributed to a lower ratio of the reactant-to-catalyst, 2, in the diluted system and negligible catalyst deactivation in the first seconds of the process.After 4 h, the specific surface area and the total pore volume of the catalyst decreased only by 37% and 41%, respectively (Table 2).In this case, both compounds (isoeugenol, n-hexadecane) and the catalyst were loaded into the reactor at room temperature and heated up to the reaction temperature with the heating rate of 10 K/min.This procedure was used for all catalyst screening experiments (Section 3.1.3).Just a half n-hexadecane conversion compared to the experiment without isoeugenol can be related to the strong adsorption of cyclic compounds on the surface of catalyst and to the faster isoeugenol hydrodeoxygenation in diluted system what points out on the competitive reactions in co-processing (Table 4, entry 5 and 6).This is also in line with results from the thermal experiment with 0.26 wt% of isoeugenol in n-hexadecane (HXD + IE).The ratio between nonlinear (branched and cyclic) and linear products was 85% vs 15%, respectively, in this catalytic co-processing.Selectivity based on the carbon number was close to the results for the experiment without isoeugenol, i.e.only slightly higher selectivity to alkylated and isomerized hexadecane and slightly lower selectivity to long-chain hydrocarbons was observed.

Entry
A lower liquid phase mass balance and carbon balances 93-99% could be related to the light cracked hydrocarbons in the gas phase, and adsorption of heavy compounds on the catalyst, coke formation.The latter one is in line with the textural properties of the spent catalyst (Table 2).Worth also noting that ca 1.8-fold higher deactivation of Fe 2 Ni 8 surface (Table 2) was observed in the experiment without isoeugenol, i.e.HXD + cat, compared to the one with this compound, HXD + IE + cat.This reflects 1.8-fold higher nhexadecane conversion for isoeugenol free experiment, HXD + cat (Table 4, entry 5 and 6).An analogous competition was found for tetralin and a phenolic compound in their co-processing over a commercial FCC catalyst [45].

Products distribution of the preliminary experiments
Distribution of all products according to the carbon number is presented in Fig. S8.Experiments with isoeugenol without nhexadecane (IE, IE + cat) showed a maximum at C10, which is obviously related to isoeugenol and dihydroeugenol.The thermal experiment of n-hexadecane (HXD) exhibited a maximum at C13 of linear tridecane, followed by the second maximum of methylnonane, C10.In n-hexadecane hydroisomerization-hydrocracking over Fe 2 Ni 8 (HXD + cat) the same shape of the product distribution of C3-C14 showing preferential cracking in the middle of the hexadecane hydrocarbon chain, i.e. symmetrical and centred at C8, has observed in line with the previous work with platinum based catalysts Pt/ZrO 2 [72], Pt/Al-MCM-48 [73], and Pt/H-Y-30A [73].In [73,74], selectivity toward cracking in the middle of the carbon chain was supported by small (2-8 nm) mesopores in the crystals.This selectivity shift was attributed to an enhanced mass transfer and faster desorption of the primary products due to a shorted diffusion path, limiting thus subsequent cracking.It should be noted, that, Fe 2 Ni 8 catalyst, in the current work, comprised mainly micropores (<2 nm, 73%).In the case of the monometallic Ni catalysts, the maximum selectivity for hydrocarbons below C15 was obtained for C5-C10 fraction, namely 60%, with 7.6 wt% Ni/H-MFI (<2 nm, 76%) and for C11-C15 fraction, 30%, with 9.1 wt% Ni/Beta (<2 nm, 23%) [70].
A similar profile was also observed in the co-processing (HXD + IE + cat) displaying a small additional maximum at C9 related to the products from isoeugenol hydrodeoxygenation.Overall, the highest maximum at C16 represents the highest selectivity to methyl-pentadecane of 45 and 48% in the experiment without (HXD + cat) and with isoeugenol (HXD + IE + cat), respectively.Distribution of the linear hydrocarbons according to the carbon number (Fig. 6) clearly showed that the presence of the catalyst led to significant n-hexadecane cracking to the light hydrocarbons.Fig. 7 displaying the linear product distribution as a function of reaction time, revealed only minor changes after 30 min.

Catalyst screening: conversion, reaction rates and turnover frequency
The synergistic effect of Fe and Ni supported on H-Y-5.1 zeolite in the co-processing of n-hexadecane with isoeugenol was investigated over five catalysts with different Fe/Ni ratios in a batch reactor at 573 K and 3 MPa of hydrogen (Fig. 9, Table 5).
The n-hexadecane conversion and the reaction rate decreased with the increasing metals particle size.Especially, the particle size smaller than 6 nm exhibited a positive influence on the reaction (Fig. 8a,b).In addition, the results of n-hexadecane hydroisomerization-hydrocracking pointed out on the structure insensitivity for the catalysts at least with the median Fe-Ni particle size larger than 5.5 nm as beyond this size the turnover frequency (TOF) was constant (Fig. 8c).
These results are in line with the literature [50], where the reaction rate of n-hexadecane (0.001-0.045 mmol/g/min) hydrocracking also decreased with increasing metal particle sizes independent on to the metal nature, namely Pt, Ni or Ru.A comparison of the monometallic Ni-catalysts showed that the reaction rate of 0.045 mmol/g/min over 5 wt% Ni/H-Beta-150 (9.4 nm) in an autoclave at a lower temperature of 483 K and a higher pressure of 4 MPa was comparable to the reaction rate of 0.04 mmol/g/min over Ni 5 (19.2 nm) in the current work in the presence of isoeu-genol.Simultaneously, the reaction rate of n-hexadecane in the presence of isoeugenol with Fe 5 Ni 5 (2.1Á10 -6 mol/g/s) was comparable with the reaction rate obtained in the fixed-bed reactor at a lower temperature of 513 K and a lower pressure of 2 MPa over Pt/H-Y30A (1.8Á10 -6 mol/g/s) and Pt/Al-MCM-48 (1.7Á10 -6 mol/g/ s) catalysts with a sieved fraction of 180-425 lm and Pt particle size of<3 nm [73].Compared to the monometallic Ni-catalyst in the current work, ca.two-fold lower TOF of n-hexadecane (29 1/ h) was obtained in the fixed-bed microreactor at a higher temperature of 613 K and 2 MPa over NiW/Ni-SAPO-11 catalysts with a sieved fraction of 400-840 lm, with the 5 and 15 wt% loading of NiO and WO 3 , respectively [75].
No clear correlation between the turnover frequency, normalized either per total amount of acid sites or sites of different strength on one hand and the acid site density on the other hand,  8. Co-processing isoeugenol with n-hexadecane: (a) conversion after 4 h, (b) reaction rate, (c) initial turnover frequency of n-hexadecane as a function of the median Fe-Ni particle size (determined by TEM).Conditions: 573 K and 3 MPa of hydrogen with 100 mL of the reaction volume, 0.1 g of catalyst, and the weight ratio of reactant-tocatalyst of 2 and 773 with respect to isoeugenol and n-hexadecane, respectively.Legend: no catalyst, n-hexadecane with isoeugenol (+), Fe 5 (black j), Fe 8 Ni 2 (red d), Fe 5 Ni 5 (blue N), Fe 2 Ni 8 (green .),Ni 5 (grey r).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)could be visible as follows from Fig. S9.Some reactions catalysed by solid acids, such as for example pentane cracking and dehydration of various alcohols over metal-free H-MFI catalysts with different Brønsted acid sites [76], displayed clear correlations between TOF and Brønsted acid site density.Apparently absence of correlations between TOF and the acid site density in the current work could be attributed to complexity of n-hexadecane hydroisomerization-hydrocracking reaction occurring over bifunctional/bimetallic catalysts and involving besides acid sites also the metal sites.
Overall, it can be concluded, that lower activity of n-hexadecane hydroisomerization-hydrocracking in co-processing was comparable with the results of the single processes obtained at milder reaction conditions, with a larger metal particle size or in the presence of mass transfer limitations.
The conversion data for catalytic and for comparison also noncatalytic experiments, of both processes, are displayed in Fig. 9. Except for the monometallic Fe catalyst, hydrogenation of isoeugenol was very rapid, and almost complete transformation to dihydroeugenol was already observed in the first minute of Table 5 Catalytic results of catalyst screening after 4 h.All data are presented in mol%.Notation: IE -isoeugenol, racemic mixture, DHE -dihydroeugenol, HXD -n-hexadecane, catcatalyst, X -conversion, MB -the liquid phase mass balance, CB -carbon balance, > C16 -alkylated hexadecane, i-C16 -isomerized hexadecane, C13-C15 -long-chain hydrocarbons, C8-C12 -jet fuels fraction, C8-C12* -jet fuels fraction without dihydroeugenol, < C8 -cracked light hydrocarbons, NL -nonlinear products (branched and cyclic), L -linear products, PC -phenolic compounds, OFBD -oxygen-free benzene derivatives, OFCAD -oxygen-free cycloalkane derivatives.Conditions: 573 K and 3 MPa of hydrogen with 100 mL of the reaction volume, 0.1 g of catalyst, and the weight ratio of reactant-to-catalyst of 2 and 773 with respect to isoeugenol and n-hexadecane, respectively.the co-processing with all catalysts.This result was analogous to the results reported for the supported Ni catalysts in [26].It should be also pointed out that the reaction proceeds via rapid hydrogenation to dihydroeugenol (Fig. 9a, Table 4), even in the absence of any catalyst, which is in line with the literature [40].Therefore, Fig. 9 also displays conversion of the consecutive step, i.e. conversion of dihydroeugenol.

Catalyst
In the case of hydrodeoxygenation, the results clearly showed that the bifunctional nature of the metal-acid catalysts has a key role.The reaction rate of dihydroeugenol linearly increased with the increasing molar ratio of the metals-to-total acid site (c Fe-Ni /c TAS ).The exception was the monometallic Ni-catalyst (Fig. 10a), which could be explained by the absence of oxophilic iron in this catalyst compared to other ones.A significant role of catalyst bifunctionality is in line with [77], where the best catalytic behaviour in guaiacol hydrodeoxygenation exhibited by 15.7 wt% Ni/Beta-12.5 (9.2 nm), was attributed to the optimized balance between active hydrogenation (metal) and acid sites with c Ni /c TAS of 3.8.In the current work, for the bimetallic Fe-Ni catalyst, the turnover frequency, defined by exposed moles of metals, was independent on c Ni-Fe /c TAS and also on the metal particle size (Fig. 10b,  c), while the turnover frequency, defined by Brønsted and Lewis acid sites (total, respectively weak), decreased with increasing Brønsted and Lewis acidity (Fig. 10d,e,f).

Catalyst screening: Product distribution
Furthermore, the results from the catalyst screening experiments also clearly showed that not only the reaction rates, but also the product distribution strongly depends on the Fe-Ni metals ratio in the catalysts (Table 5).The reported results of product distribution reflect the simplified reaction schemes of both processes, n-hexadecane hydroisomerization-hydrocracking (Eq.( 9)) and isoeugenol hydrodeoxygenation (Fig. 11): where > C16 is alkylated hexadecane, HXD is n-hexadecane, i-C16 is isomerized hexadecane, C13-C15 are long-chain hydrocarbons with 13-15 carbon atoms, C8-C12 is jet fuels fraction, <C8 are cracked light hydrocarbons.
Extremely slow isoeugenol hydrodeoxygenation over the monometallic Fe catalyst resulted in only 27% of dihydroeugenol conversion and relatively rapid n-hexadecane hydroisomerization-hydrocracking with the second highest conversion of 3.3%.It should be noted that no other cyclic products were detected in the reaction mixture.This incomplete balance of cyclic products could be explained by a very low concentration of cyclic products (below the detection limit of GC-FID, MS) or strong adsorption of dihydroeugenol on the surface of the catalyst.The latter one is in line with the textural properties of the spent catalyst showing an extremely low specific surface area of 9 m 2 /g and the pore volume of 0.03 cm 3 /g (Table 2), i.e. by 98%, respectively, 85%, lower compared to the fresh catalyst.The rapid nhexadecane hydroisomerization-hydrocracking could be attributed to the extra framework Al species (determined by solid state 27 Al MAS NMR just for the monometallic Fe catalyst as a small peak at 0 ppm).This is in line with [78], where it was observed that isomerization, cracking and related coking processes were favoured with increasing extra framework Al, while octanoic acid conversion decreased in hydrodeoxygenation over 5 wt% Ni/Al-SBA-15 catalysts.
Overall, it can be concluded that activity of the monometallic Fe catalyst in isoeugenol hydrodeoxygenation step was low and the results were comparable with an experiment without the catalyst (IE, Table 4).This result is interpreted by an insufficient hydrogenation ability of iron [79].
On the contrary, very high activity of dihydroeugenol transformation and at the same time of n-hexadecane was observed over Fe 2 Ni 8 catalyst in the first minutes compared to other catalysts (Fig. 9).Almost complete transformation of dihydroeugenol (>98%) was obtained after 60 min, which could lead to acceleration of n-hexadecane hydroisomerization-hydrocracking (from 2.3% to 5.6% conversion per hour).After 4 h, the total conversion of isoeugenol, and dihydroeugenol, as well as 15.5% of n-hexadecane conversion was achieved with the highest total yield of 15.8% and the highest yield of jet fuels fraction without dihydroeugenol of 4.5%.However, it should be noted, that cyclic products still contained some of the undesired oxygen-containing phenolic compounds.Characterization data, namely EXAFS and TPR profiles of fresh catalysts, pointed out on Fe-Ni interactions in bimetallic catalysts, which can be involved in explanation of catalytic behaviour.Strong interactions, compared to other materials, were obtained especially for the most active Fe 2 Ni 8 catalyst, according to the largest deviations of H 2 -TPR profile (Fig. 1).
According to FT EXAFS Fe 2 Ni 8 spectra (Fig. S12) the interatomic distances and the mean-square deviations of interatomic distances around Fe and Ni atoms are identical being slightly lower compared to the bulk fcc Ni metal.This, and the fact that a bcc lattice is typical for pure Fe nanoparticles, can imply a nearly random fcc alloy comprising Fe and Ni.
Moreover, Fe 2 Ni 8 catalyst comprising more Ni than Fe exhibited the smallest Fe-Ni particle sizes (4.6 nm, Table 2), the lowest amount of Brønsted acid sites (109 lmol/g, Table 2), the highest specific surface area (660 m 2 /g) and the highest amount of Fe metallic species in the bulk (43%).The results for Fe 2 Ni 8 are in line with the mechanism for which hydrodeoxygenation of phenolic components starts with partial hydrogenation of the aromatic ring giving propylcyclohexenol followed by this enol transformations to propylcycohexanone and its subsequent hydrogenation to propylcyclohexanol, which can be easily dehydrated [26].Furthermore, the bond energy for breaking OH group from the phenyl ring is 472 kJ/mol [80], being higher than that of aliphatic C-O, which is 393 kJ/mol [81].
The other three catalysts (Fe 8 Ni 2 , Fe 5 Ni 5 and Ni 5 ) showed almost the same catalytic activity.Only slightly lower dihydroeugenol conversion was obtained over Fe 8 Ni 2 and a slightly lower n-hexadecane conversion over the monometallic Ni catalyst.On the contrary, the product distribution was significantly different for these catalysts compared to other materials.Selectivity to alkylated hexadecane, >C16, was 15% for Fe 8 Ni 2 , while for others it was close to zero.The main product of hexadecane alkylation was eicosane, C20.At the same time, this catalyst exhibited the highest ratio between nonlinear (branched and cyclic) and linear products (NL/L) equal to 6.2.In addition, exclusively for this catalyst, tetralin derivatives were produced, and in a significant amount (35% from all cyclic compounds, Table 5).This can be related to the ring-closure over the weak Lewis acid sites as was described previously for cyclization of 5-phenyl-1-entene to 1methyl-tetraline over FeCl 3 catalyst [82].Although the Fe 8 Ni 2 catalyst contained the lowest absolute amount of weak Lewis acid sites compared to others, this catalyst exhibited the highest fraction of weak Lewis sites (98%), i.e. with a negligible amount of medium and no strong Lewis acid sites (Table 2).At the same time, this catalyst had the highest concentration of Fe among the bimetallic ones.A similar mechanism of ring closing was suggested for the Friedel-Crafts reaction between benzene and succinic anhydride using aluminium chloride [83] or the intermolecular Friedel-Crafts reaction of benzyl carbinols mediated over stoichiometric amounts of Lewis and Brønsted acids [84].Tetralin derivatives were also formed in depolymerisation of polyethers with aromatic compounds using titanium-exchanged montmorillonite catalyst [85].
The lowest NL/L ratio of 4.1 was obtained with the monometallic Ni catalyst.Selectivity to isomerized hexadecane in coprocessing (S iC16 = 47% at X = 0.3%, Table 5), was comparable with the monometallic 10 wt% Ni/H-Beta (S iC16 = 57% at X = 64%) and 10 wt% Ni/desilicated H-Beta (S iC16 = 44% at X = 83.3%)catalysts with the smaller particle sizes of Ni, 11.4 and 16.7 nm, respectively, investigated in n-hexadecane hydroisomerization-hydrocracking in the trickle-bed reactor under the same pressure and slightly lower temperature of 548 K [70].High conversion of hexadecane of 91% but a lower selectivity to i-C16 was obtained over 5 wt% Ni/WO 3 /ZrO 2 in the trickle-bed reactor under the same temperature of 573 K and a lower pressure of 2 MPa [86].This Ni 5 catalyst also exhibited the highest selectivity to the long-chain hydrocarbons, C13-C15 and 5% of oxygen-containing cyclic products from all cyclic products.
In contrast, no oxygen-containing cyclic products were detected for Fe 5 Ni 5 catalyst.The cyclic compounds were composed of 19% oxygen-free benzene derivatives and 81% oxygen-free cycloalkane derivatives.The same catalyst displayed the highest selectivity to jet-fuel fraction, C8-C12.Fe 5 Ni 5 catalyst containing equal amounts of Fe as Ni exhibited Fe and Ni particle sizes of < 3 and 9 nm, respectively, a low Brønsted to Lewis acid sites ratio of 1 (Table 2), the lowest blocking temperature respectively the highest magnetic saturation (8.5 Am 2 /kg,), and the highest hydrogen consumption in temperature programmed desorption (4.2 relative peak area) compared to other studied materials.
Overall, it can be concluded that dihydroeugenol conversion and selectivity to nonlinear cracked light hydrocarbons, <C8, increased with increasing Ni fraction in the catalysts.On the contrary, selectivity to a linear C14 hydrocarbon, n-tetradecane, decreased when there was more Ni in the catalyst.Except the monometallic Fe catalyst, the isomerized hexadecane, i-C16, was main product with selectivity of 44-48%.The fractions of nonlinear (branched and cyclic) and linear products were relatively close for all catalysts, namely 80-86% and 14-20%, respectively.Selectivity to isomerized hexadecane, i-C16, and to the jet-fuel fraction, C8-C12, as a function of conversion is displayed in Fig. 12. From the results obtained at low conversion, where selectivity was changed with conversion, a network of consecutive reactions is much more prominent.In contrast, from almost constant selectivity obtained for Fe 5 and Fe 2 Ni 8 catalysts at high conversion, a reaction network with parallel reactions can be expected.The highest initial selectivity to jet fuel products, S C8-C12 , among bimetallic Fe-Ni catalysts was obtained for Fe 5 Ni 5 , which exhibited no strong acid sites and the highest concentration of medium acid sites (Table 2).This is in line with [87] in which Ni-Mo/SAPO-11 with a moderate acidity strength displayed the best performance for producing a jet fuel fraction from long-chain hydrocarbons C15-C18.It should, however, be pointed out that Fe 2 Ni 8 was the most active forwards hydroisomerization of hexadecane giving the highest yield of i-C16 (Table 5).In the case of Fe 5 Ni 5 and Ni 5 catalysts, increased selectivity to i-C16, and at the same time, decreased selectivity to C8-C12 with increasing conversion indicate catalyst deactivation and suppression of the consecutive pathways.Increased selectivity of isomerized hexadecane with increasing conversion, was also observed for Pt/Beta, Pt/Al 2 O 3 -H-Beta (LL) and Pt/Al 2 O 3 -H-Beta (HL) catalysts with the Pt particle size of 3.1, 2.9 and 2.6 nm, respectively, in the fixed-bed reactor under the same pressure and a lower temperature of 493 K [88].On the  (c) the ratio of oxygen-containing cyclic compounds and oxygen-free cyclic compounds, (d) the yield ratio of nonlinear (branched and cyclic) and linear products, (e) the yield ratio of jet fuels fraction, C8-C12, and isomerized n-hexadecane, i-C16, as a function of the molar ratio of metals-to-acid site, (f) the yield ratio of long-chain hydrocarbons, C13-C15, and isomerized n-hexadecane, i-C16, as a function of Brønsted acid sites.Conditions: 573 K and 3 MPa of hydrogen with 100 mL of the reaction volume, 0.1 g of catalyst, and the weight ratio of reactant-to-catalyst of 2 and 773 with respect to the isoeugenol and n-hexadecane, respectively, after 4 h.Legend: Fe 5 (black j), Fe 8 Ni 2 (red d), Fe 5 Ni 5 (blue N), Fe 2 Ni 8 (green .),Ni 5 (grey r).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)contrary, decreased selectivity to isomerized hexadecane with increasing conversion, as for Fe 8 Ni 2 , was obtained in the fixedbed reactor at 543-583 K for Ni/ZSM-48 exposed to different alkali treatment [71].
Further correlation of the catalytic results with the catalyst characterization data revealed increased activity of benzene ring hydrogenation with an increasing fraction of Ni in the catalyst, leading to formation of cycloalkane derivatives (Fig. 13a,b).This is not surprizing considering well known activity of Ni in benzene hydrogenation [89][90][91].In the case of the bimetallic catalysts, the ratio of oxygen-containing cyclic compounds and oxygen-free cyclic compounds (OOC/OFC) increased with increasing the metals-toacid site molar ratio (c Fe-Ni /c TAS ).The same trend was observed for the yield ratio of nonlinear (branched and cyclic) and linear products (Y NL /Y L , Fig. 13d), while for the yield ratio of jet fuel fraction and isomerized n-hexadecane, it was opposite (Y C8-12 /Y i-C16 , Fig. 13e).In other words, decreasing c Fe-Ni /c TAS led to increased formation of OFC than OCC, more linear products than the branched ones and increased formation of jet fuel fraction than isomerized hexadecane.The yield ratio of long-chain hydrocarbons and isomerized n-hexadecane increased with increased Brønsted acid sites (BAS) for both monometallic and bimetallic catalysts (Y C13-15 /Y i-C16 , Fig. 13f).This is in line with [50,51] reporting that stronger acidity leads to more cracking, while mild acidity or BAS promote isomerization.It has been reported in the literature that the mechanism for hydroisomerization-hydrocracking of long-chain alkanes involves in the first step metal sites needed for dehydrogenation of an alkane forming an alkene.This is followed by formation of a carbenium ion on acid sites and subsequent isomerization to an isoolefin [51].
The distribution of all products according to the carbon number (Fig. S10) showed two maxima for all catalysts.As mentioned above, the first maximum was related to methyl-pentadecane, except for the monometallic Fe-catalyst giving dihydroeugenol.
The size of the second maximum, observed at a carbon number of 9 in all other cases, was dependent on the Fe-Ni ratio in the catalyst.The highest second maximum reaching 25 mol% was obtained for Fe 5 Ni 5 , giving a product mixture containing mainly propyl-cyclohexane and ethyl-methyl-cyclohexane.Lower amounts of product for the second maximum, 12 mol%, were obtained for Fe 8 Ni 2 , consisting of propyl-phenol, propylcyclohexane and butyl-cyclopentane.The second maximum for Fe 2 Ni 8 and Ni 5 was related to methyl-octane followed by propylcyclohexane, propyl-phenol and dimethyl-heptane.
The distribution of the linear hydrocarbons according to the carbon number (Fig. S11) pointed out on a dominant presence ntetradecane in the case of Fe 5 and n-pentadecane for other catalysts.A relatively equal distribution of linear hydrocarbons from C5 to C13 was observed for the monometallic Fe catalyst.For Fe 2 Ni 8 , the distribution was shifted to the lower linear hydrocarbons with the median at n-heptane, C7.Compared to the latter one, other catalysts showed a lower fraction of C4-C11 and higher for C12-C14 hydrocarbons.Higher amounts of cracked products formed over Fe 2 Ni 8 are linked with high activity giving 15.5% conversion of hexadecane combined with a consecutive reaction network [51].

Conclusions
The synergistic effect of Fe and Ni supported on H-Y-5.1 zeolite in the co-processing of n-hexadecane with isoeugenol was investigated in a batch reactor at 573 K and 3 MPa.The results from the catalyst screening experiments clearly showed that both the reaction rate and product distribution strongly depend on the Fe-Ni metal ratio in the catalysts.The spent catalysts contained ca 40 wt% of coke with predominantly aliphatic species.
In n-hexadecane hydroisomerization-hydrocracking without isoeugenol, the product distribution of C3-C14 exhibited a symmetrical shape, centred at C8, showing preferential cracking of the hexadecane hydrocarbon chain in the centre of the substrate.The presence of just 0.26 wt% isoeugenol made n-hexadecane hydroisomerization-hydrocracking two-fold slower.Lower n-hexadecane conversion compared to the experiment without isoeugenol could be related to strong adsorption of cyclic compounds on the surface of the catalyst and faster isoeugenol hydrodeoxygenation in a diluted system, which points out on the competitive reactions in co-processing.
Correlation of the catalytic results with the characterization data revealed a positive influence of the small medial metal particle size (<6 nm) in the case of n-hexadecane hydroisomerizationhydrocracking.At the same time, the experimental data clearly showed that bifunctionality of the metal-acid catalysts has a key role in the case of hydrodeoxygenation.A low activity towards isoeugenol hydrodeoxygenation was obtained over the monometallic Fe catalyst, while high cracking activity of hexadecane was observed in the presence of Ni.Significantly higher activity of coprocessing, compared to other catalysts, was obtained over 2 wt% Fe -8 wt% Ni/H-Y-5.1 catalysts with the metals particle sizes of 4.6 nm.High activity towards alkylation of n-hexadecane was obtained only for 8 wt% Fe -2 wt% Ni/H-Y-5.1.Over 5 wt% Fe -5 wt% Ni/H-Y-5.1 bifunctional catalyst no oxygen-containing cyclic products were detected.This catalyst exhibited the highest hydrogen consumption in temperature programmed desorption, which can serve as a marker for hydrodeoxygenation.
Overall, it can be concluded, that dihydroeugenol conversion, selectivity to nonlinear cracked light hydrocarbons and activity of benzene ring hydrogenation increased with increasing Ni fraction in the iron-nickel catalysts.Except the monometallic Fe catalyst, the isomerized hexadecane was the main product with selectivity of 44-48%.The ratio between nonlinear (branched and cyclic) and linear products was relatively close for all catalysts, with respective amounts 80-86% and 14-20%, respectively.

Fig. 3 .
Fig.3.H 2 -TPR profiles of Fe 5 (black), Fe 2 Ni 8 (green), Ni 5(grey).Conditions: to 1073 K with a 5 K/min ramp under 1.5 mL/min of hydrogen and 28.5 mL/min of argon (5 vol% H 2 and 95 vol% Ar).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)Table2Metalcontent (determined by ICP-EOS) of different Fe, Ni and Fe-Ni/H-Y-5.1 fresh catalysts, their concentrations of Brønsted and Lewis acid sites determined by pyridine adsorption-desorption, median metal particle sizes of the fresh and spent catalysts determined by TEM, and textural properties of the spent catalysts.Legend: BAS -Brønsted acid sites, LAS -Lewis acid sites, TAS -total acid sites, B/L -ratio of the Brønsted and Lewis acid sites, s (strong acid sites, data at 723 K), m (medium acid sites, data at 623 K minus data at 723 K), and w (weak acid sites, data at 523 K minus data at 623 K), cFe-Ni /c TAS -metals-to-total acid site, S -specific surface area, V -total pore volume, D -difference

Fig. 4 .
Fig. 4. Deposited coke in the spent catalyst as a function of: (a) decrease of the catalyst surface area, (b) metals concentration in the catalyst.

Fig. 5 .
Fig. 5. Solventless isoeugenol hydrodeoxygenation (a) full scale and (b) zoom-in of reaction conditions in time.Conditions: 52 mL of the reaction volume, 1 g of Fe 5 Ni 5 catalyst, and the weight ratio of reactant-to-catalyst of 56 at 473-573 K and 2 MPa of hydrogen.The preheated reactant was injected on the hot dry catalyst.The start of the reaction corresponds to the initiation of mixing.

Fig. 6 .
Fig. 6.Distribution of linear products (molar fraction) after 4 h as a function of the carbon number for experiments in Table 4: (a) full scale and (b) zoom-in.

Fig. 9 .
Fig. 9. Co-processing of isoeugenol with n-hexadecane: (a,b) full scale and zoom-in of n-hexadecane conversion, (c) isoeugenol conversion, (d) dihydroeugenol conversion, as a function of the reaction time.Conditions: 573 K and 3 MPa of hydrogen with 100 mL of the reaction volume, 0.1 g of catalyst, and the weight ratio of reactant-to-catalyst of 2 and 773 with respect to isoeugenol and n-hexadecane, respectively.Legend: no catalyst, n-hexadecane with isoeugenol (+), Fe 5 (black j), Fe 8 Ni 2 (red d), Fe 5 Ni 5 (blue N), Fe 2 Ni 8 (green .),Ni 5 (grey r).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 10 .Fig. 11 .
Fig.10.Co-processing of isoeugenol with n-hexadecane: (a) reaction rate, (b) initial turnover frequency of dihydroeugenol as a function of the molar ratio of metals-to-acid site, (c) initial turnover frequency of dihydroeugenol as a function of the median Fe-Ni particle size (determined by TEM), (d) initial turnover frequency of dihydroeugenol as a function of the total Brønsted acid sites, (e) initial turnover frequency of dihydroeugenol as a function of the total Lewis acid sites, (f) of the weak Lewis acid sites.Conditions: 573 K and 3 MPa of hydrogen with 100 mL of the reaction volume, 0.1 g of catalyst, and the weight ratio of reactant-to-catalyst of 2 and 773 with respect to isoeugenol and nhexadecane, respectively.Legend: no catalyst, n-hexadecane with isoeugenol (+), Fe 5 (black j), Fe 8 Ni 2 (red d), Fe 5 Ni 5 (blue N), Fe 2 Ni 8 (green .),Ni 5 (grey r).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 12 .Fig. 13 .
Fig. 12. Co-processing of isoeugenol with n-hexadecane: (a) selectivity to n-hexadecane isomers, (b) selectivity of jet-fuel hydrocarbons (C8-C12) as a function of coprocessing conversion (Eq.(4)).Conditions: 573 K and 3 MPa of hydrogen with 100 mL of the reaction volume, 0.1 g of catalyst, and the weight ratio of reactant-to-catalyst of 2 and 773 with respect to isoeugenol and n-hexadecane, respectively.Legend: no catalyst, n-hexadecane with isoeugenol (+), Fe 5 (black j), Fe 8 Ni 2 (red d), Fe 5 Ni 5 (blue N), Fe 2 Ni 8 (green .),Ni 5 (grey r).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1
List of catalysts, nominal loading, and the metal fraction determined by ICP-OES.