Guaiacol hydrodeoxygenation in hydrothermal conditions using N-doped reduced graphene oxide (RGO) supported Pt and Ni catalysts: Seeking for economically viable biomass upgrading alternatives

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in: APPLIED CATALYSIS A: GENERAL on 5 February 2021, available at https://doi.org/10.1016/j.apcata.2020.117977 .” Guaiacol hydrodeoxygenation in hydrothermal conditions using N-doped reduced graphene oxide (RGO) supported Pt and Ni catalysts: seeking for economically viable biomass upgrading alternatives S. Parrilla-Lahoz1, W. Jin1, L. Pastor-Pérez1, D. Carrales-Alvarado2, J.A. Odriozola1,3 A.B. Dongil2, T. R. Reina1,3* 1 Department of Chemical and Process Engineering, University of Surrey, Guildford GU2 7XH, UK. 2 Instituto de Catálisis y Petroleoquímica CSIC, Madrid, Spain 3 Departamento de Química Inorgánica e Instituto de Ciencia de Materiales de Sevilla (ICMS), Centro mixto CSIC-Universidad de Sevilla, Avda. Americo Vespucio 49, 41092 Sevilla, Spain Corresponding author: t.ramir[email protected] Abstract Herein we present an innovative route for model biomass compounds upgrading via “H2- free” hydrodeoxygenation (HDO) reactions. The underlaying idea is to implement a multifunctional catalyst able to activate water and subsequently use in-situ generated hydrogen for the HDO process. In this sense we have developed a series of effective Ni and Pt based catalysts supported on N-promoted graphene decorated with ceria. The catalyst reached commendable conversion levels and selectivity to mono-oxygenated compounds considering the very challenging reaction conditions. Pt outperforms Ni when the samples are tested as-prepared. However, Ni performance is remarkably boosted upon applying a pre-conditioning reductive treatment. Indeed, our NiCeO2/GOr-N present the best activity/selectivity balance and it is deemed as a promising catalyst to conduct the H2-free HDO reaction. Overall, this “proof-concept” showcases an economically appealing route for bio-compounds upgrading evidencing the key role of advanced catalysts for a low carbon future. Keywords: H2-free HDO, guaiacol, biomass, Pt, Ni, GOr, N-doped catalysts INTRODUCTION Nowadays fossil fuels supply more than 80 % of all the energy consumed in developed countries being also responsible for more than 70 % of anthropogenic greenhouse gas emissions [1]. The intensive use and consumption of petroleum derivates results in the diminishing fossil resources, this arising environmental and political concerns. Accordingly, solutions to mitigate climate change and reduce the consumptions of fossil fuel must be investigated [2]. In this sense, it is essential to develop renewable energy systems to alleviate the dependence on fossil fuels. Biomass, as the only renewable carbon source in nature, represents an optimal source of sustainable energy, which can be found in abundance with a low price and a wide distribution [3]. Biofuels derived from biomass are considered as carbon neutral fuels since plants could fix carbon dioxide by photosynthesis avoiding its discharge to the environment [4]. Biomass is mainly composed of lignin, hemicellulose and cellulose. Lignin owns the potential to produce high-value chemicals and transportation fuels. Lignin is a planted polymer composed of phenylpropanoid building units. Indeed, lignin is formed by multiple structures that are not repeated regularly, besides the compositions and structures of the lignin vary according to their origins [5]. Lignin-rich biomass is typically processed via thermochemical processes such as fast pyrolysis yielding bio-oil [6]. In general, bio-oils derived from fast pyrolysis of lignin contain large concentrations of oxygenated compounds leading to lowquality bio-fuels. More precisely the elevated oxygen concentration in bio-oil results in high viscosity, low solubility in other hydrocarbons, low volatility, poor stability, corrosiveness and insufficient heating value limiting its direct use as transportation/blending fuel [7]. Hence, oxygen-rich bio-oil must be upgraded to make it suitable for the biofuel industry [7]. Considering the complex structure of lignin, monomeric fragments are often used in catalytic studies to discern mechanistic aspects as well as fundamental parameters of the catalysts design [3,8]. Guaiacol is the most representative of phenolics derivatives and extensively used as lignin model compound [9]. Guaiacol contains two common oxygenates groups, methoxy (O-CH3) and phenolic (Ph-OH) groups offering a suitable simplified case to study catalytic oxygen removal [10]. In this sense, Hydrodeoxygenation (HDO) is a fundamental technology to upgrade bio-oils through catalytic processes to produce deoxygenated fuels or high-value chemicals [9]. Traditional HDO reactions requires highpressure hydrogen imposing severe limitations in terms of safety, manipulation and transportation [9]. In order to circumvent these limitations, our group has recently demonstrated the viability of “H2-free” HDO reactions using water as reaction media [11]. This is a very challenging approach involving a multifunctional catalyst able to conduct water activation and hydrodeoxygenation reaction simultaneously. Our “proof-of-concept” studies revealed the pivotal role of the active phase and the support being carbon-based systems an appealing option which deserves further investigation [11,12]. Among the different strategies to boost the performance of carbon-based catalysts altering its physicochemical properties using dopants such as nitrogen brings some advantages. For instance, there are evidences of increased nanoparticle dispersion as well as resistance to agglomeration of Platinum clusters when a N-doped carbon was selected as support [13]. The work by Groves et. al [14], uses density functional theory (DFT) to calculate the binding energy between Pt and five different graphene surfaces, one non-doped surface and four of them doped with different atoms. They found that the strength of the Pt–surface bond increased in the following order: undoped, nitrogen, boron, beryllium, and oxygen. Donghui Guao et. al, demonstrated that nitrogen-doped carbon material exhibit high electrocatalytic activity for the oxygen reduction reaction and high durability even under acidic conditions [15]. They also found that nitrogen-doped mesoporous carbon-support allows efficient mass transfer, suppress sintering and loss of the active phase during the HDO process [16]. Along with nitrogen, redox promoters such as ceria are identified as promising additives and/or catalyst supports its excellent oxygen storage capacity and its defect chemistry (i.e. concentration of oxygen vacancies, relative proportion Ce4+/Ce3+) which plays a beneficial role in dispersing metal nanoparticles [17]. As for the active phases noble and non-noble metals have been tested in the hydrothermal upgrading of HDO. For instance, Jin et al. demonstrated that Ru and Pt are promising materials leading to high selectivity towards partially de-oxygenated products [18]. Ni and combinations of Ni-CeO2 supported on activated carbon have also shown promising results representing an economically appealing alternative to the expensive noble metals. Still, noble metals outperform Ni in this reaction opening some room to further improve the design of Ni based catalysts for the H2-free HDO reaction. The use of graphene provides the benefits of carbon supports since they are very stable in water media, in both acid and basic conditions that are normally found in biomass feeds. In addition, graphene’s high surface area helps to better disperse the active phases. Also, when prepared from graphene oxide, nitrogen atoms can be incorporated easily using clean gas-phase procedures [19]. Herein we propose reduced graphene oxide (GOr) and N-doped reduced graphene oxide (N-GOr) as advanced supports to promote the activity of Ni based catalysts in this reaction. In this scenario, this work presents a comparison of Pt vs Ni catalysts supported multicomponent CeO2-promoted N-doped RGO as promising systems for biomass model compounds upgrading. Following the spirit of the special issue to honour the outstanding contributions of Prof. Parvulescu to the field of biomass conversion catalysis, we believe this piece of work showcases the pivotal role of applied catalysis for bio-resources conversion in the context of a circular economy. MATERIALS AND METHODS Supports In this study, Graphene (GOr) and N-doped Graphene (GOr-N) were used as supports. The preparation of both graphenes required graphite oxide (GO325) as a precursor. GO325 was synthesized from commercial natural flake graphite (G, 99.9 % purity) acquired from Alfa Aesar following a modified Brodie method [20]. The graphenes were obtained by thermal reduction of the GO325 carried out in a vertical cylindrical reactor. The reactor was loaded (350 mg GO325) and purged with N2 at room temperature for one hour at a flow of 100 cm3 (STP)/min. Subsequently, the de N2 flow was reduced to 87 cm3 (STP)/min and a flow of H2 and NH3 of 3 and 10 cm3 (STP)/min respectively, was fed to the reactor for the N-doped sample. Then, different heating ramps were employed. The first ramp was from room temperature to 100 °C with a heating rate of 5 °C/min. After, it was increased to 700 °C with at a rate of 5 °C/min and leave it at the final temperature for 5 min. Once the furnace heating was finished, the reactor was allowed to cool down to 400 °C, the H2/NH3 flows were cut off and the system was allowed to cool under N2 atmosphere. This graphene was designated as GOr-N. The reduction for the undoped sample was done following the same procedure described above, but without feeding NH3 to the reactor, this graphene was denoted as GOr. Catalyst Preparation The catalysts herein presented were prepared by sequential wet impregnation. Firstly, 10 wt.% of CeO2 was impregnated on GOr and GOr-N. For this, 0.0989 g Ce(NO)3·6H2O was dissolved in 50 mL of acetone and 350 mg of the corresponding support were added, keeping under stirring for 4 h. Subsequently, the solvent was evaporated, and the resulting sample was placed in the oven at 100 °C overnight. Then, the samples were calcined at 350 °C for 5 h with a temperature ramp of 1 °C/min. Then, impregnation of 1 wt.% Pt or 15 wt.% Ni was performed using H2Cl6Pt·6H2O or Ni(NO3)2·6H2O, as precursor salt, respectively and following the same procedure describe above. All reactants were purchased from Sigma- Aldrich. The samples were labelled PtCeO2/GOr, PtCeO2/GOr-N, NiCeO2/GOr and NiCeO2/GOr-N. To complete the series, Pt/GOr y Pt/GOr-N were prepared using the same procedure. Catalyst Characterization H2-Temperature-programmed reduction (TPR) analysis were carried out on the catalysts in a tube quartz reactor using a 5 % H2/N2 gas flow of 50 mL/ min, with a heating rate of 10 °C/min. Hydrogen consumption, water production and methane production were monitored using an online mass spectrometer (Pfeiffer, OmniStar GSD 301). X-Ray powder diffraction (XRD) patterns were recorded on a Bruker D8-Advance with a Goebel mirror and a Kristalloflex K 760-80F X-Ray generation system, fitted with a Cu cathode and a Ni filter. Spectra were registered between 10 and 80 (2θ) with a step size of 0.05 °and a time of 3 s per step. Information about the supported metal particles was acquired by TEM (Transmission electron microscopy) in a JEOL 2100F field emission gun electron microscope operated at 200 kV and equipped with an Energy-Dispersive X-Ray detector, XEDS. The sample was ground until powder and a small amount was suspended in acetone solution using an ultrasonic bath. Some drops were added to the copper grid (Aname, Lacey carbon 200 mesh) and the solvent was evaporated at room temperature before introduction in the microscope. XEDS-mapping analysis was performed in STEM mode with a probe size of 1 nm using the INCA x-sight (Oxford Instruments) detector. Catalytic reactions The HDO reaction was investigated in a high-pressure Parr Series 5500 HPCL batch reactor with a 4848-reactor controller. Approximately 0.5 g of guaiacol as a reactant, 49.5 g of water as solvent and hydrogen source, and 0.05 g of catalyst were loaded in a 300 ml glass-lined stainless-steel vessel. To avoid the air contamination, N2 was bubbled through the filled reactor for 5 min under a stirring speed of 100 rpm before closing the reaction system. Then the reactor was heated to the desired reaction temperature (250-300 °C) and stayed stable at this temperature for four hours under a stirring speed of 300 rpm. The pressure of the vessel is typically 50 bar when the temperature of the reactor is 250 oC and 98 bar at 300 oC. The spent catalyst was separated from the liquid by filtration and dried at 80 °C for 12 h. Non-catalytic HDO reaction (blank reaction) and the activity of fresh Pt/GOr and Pt/GOr-N catalysts were tested at 250 oC. Succeeding, the fresh Pt/GOr, Pt/GOr-N, NiCeO2/GOr-N and PtCeO2/GOr-N were tested at slightly elevated temperature (i.e., 300 oC), and a non-catalytic experiment was tested at same reaction condition. All reduced catalysts, including Pt/GOr, Pt/GOr-N, NiCeO2/GOr, NiCeO2/GOr-N, PtCeO2/GOr, PtCeO2/GOr-N catalysts were tested at 300 oC. The catalysts were reduced for 2 h at 350 °C under a total gas flow of 100 mL/min (H2: N2 = 1:4). The liquid products were separated by a liquid-liquid extraction using ethyl acetate as solvent. The species of the organic compounds produced was identified by a gas chromatography-mass spectrometry (GC-MS). Quantitative analysis was performed with a gas chromatograph-flame ionisation detector (GC/FID). The GC injector temperature was 280 °C. The GC separation was performed using a Carboxen Packed Analytical Column (30 m×320 μm×0.25 μm). A split ratio of 8:1 was held. The column was firstly held at 50 °C for 1 min, then increased to 240 °C at a heating rate of 5 °C/min and held at 240 °C for 10 min. The conversion of guaiacol and selectivity (based on C mol) of the products are calculated using Equation 1 and 2. 𝑪𝒐𝒏𝒗𝒆𝒓𝒔𝒊𝒐𝒏 𝒐𝒇 𝑮𝒖𝒂𝒊𝒂𝒄𝒐𝒍 (𝒎𝒐𝒍%)= (𝒎𝑮𝒊𝒏 − 𝒎𝑮𝒐𝒖𝒕) ∗ 𝑵𝑮 𝑴𝑮 𝒎𝑮𝒊𝒏 𝑴𝑮∗ 𝑵𝑮 ∗𝟏𝟎𝟎 𝑬𝒒. 1 𝑺𝒆𝒍𝒆𝒄𝒕𝒊𝒗𝒊𝒕𝒚 𝒐𝒇 𝒑𝒓𝒐𝒅𝒖𝒄𝒕 𝒙 (𝒎𝒐𝒍%)= 𝒎𝒙∗ 𝑵𝒙 𝑴𝒙 (𝒎𝑮𝒊𝒏 − 𝒎𝑮𝒐𝒖𝒕) ∗ 𝑵𝑮 𝑴𝑮 ∗ 𝟏𝟎𝟎 𝑬𝒒. 2 mGin: Initial mass of guaiacol; mGout: Detected mass of guaiacol in the organic phase; mx: Mass of product x. MG: Molar mass of guaiacol; Mx: Molar mass of product x. NG: Number of Carbon in guaiacol; Nx: Number of Carbon in product x. RESULTS AND DISCUSSION CHARACTERIZATION XRD The X-ray diffraction patterns of the calcined catalysts are shown in Figure 1. All the samples display diffraction peaks centred at 2= 26.5 o and 44.0 o corresponding to the (002) and (100) planes of graphite respectively, this indicating the reduction of the GO matrix [21][22]. The peaks of graphite are sharper for the nitrogen-doped samples which seem to indicate that these materials recovered the graphitic structure to a higher extent [23]. The different wideness of this peak is related to the different interlayer distance generated during the thermal reducing treatment. The CeO2 peaks (JCPDS no. 03-065-5923) at 28.6o, 33.2o,47.6o, 56.5o and 76.8o, corresponding to the reflections in the (111), (200), (220), (311), (331) crystalline planes, respectively, of the cubic fluorite type phase can be clearly seen in XRD patterns of CeO2 based catalysts. As a general trend, cerium oxide presents smaller particle size and higher amorphous character when supported on carbon. The typical fluorite cell reflections are sharper and more intense for the ceria bulk materials compared to ceria supported on carbon [24]. This means that in these materials, good ceria dispersion and small crystal size is achieved. As for the XRD diffraction patterns of Ni-based samples, Ni0 peaks (JCPDS 01-070-1849) at 44.5o that overlaps with that of graphite, and at 51.9o corresponding to the (111) and (200) Figure 1. XRD patterns of the calcined samples. crystalline planes can be appreciated in all the samples. Interestingly, typical peaks ascribed to NiO, i.e. 37.2o, 43.3o and 62.9o, cannot be observed on the XRD patterns but NiO presence cannot be discarded as the main peak could be overlapped with the Ni metallic peak. Moreover, NiO reduction zones have been observed in the TPR obtained profiles discussed below. The results can be tentatively explained by the reduction of NiO into metallic Ni during the calcination process due to electrons transfer from the graphemic layers as previously reported [25]. It is worth noting that for the N-doped sample the peak of Ni at 2 of 44.5o seems to be wider compared to the same peak of NiCeO2/GOr indicating a greater degree of dispersion and smaller particle size of the Ni clusters for the N-doped samples. Pt diffractions peaks are not observed in any sample due to the high dispersion of the supported metals and the low loading of Pt (1 wt.%). For clarification, the diffraction peak around 2 of 22o in the Pt/Gr pattern is due to the sample holder since the amount of the sample during the analysis was insufficient to completely cover the holder surface. H2-TPR The reducibility of the catalysts was studied to get further insights into their composition, metal interactions and redox properties. The H2-TPR profiles of all the samples are shown in Figure 2. Figure 2. H2-TPR profiles of all calcined samples. For Pt/GOr and Pt/GOr-N samples, two weak reductions zones are observed in H2-TPR profiles. The first at around 240 oC (very small), correspond to the reduction of Pt oxides species (PtOx) [26]. The second reduction zone between 600 oC and 800 oC is due to the hydrogen consumption by the reduction of oxygen superficial groups on the support. In the N-doped sample the amount of oxygen superficial groups decreases as a result of N substitution resulting in a smoother reduction event [14]. In general, all the samples present very low hydrogen consumption at low temperatures mainly due to the small number of expected oxides species, as can be observed in the XRD patterns, since part of the metal oxides have been reduced during the synthesis procedure. The H2-TPR profiles of PtCeO2/GOr and PtCeO2/GOr-N exhibit several reduction peaks. In both samples the reduction peak at low temperature, around 250 °C and 100 °C, indicates the reduction of PtOx and ceria adjacent to Pt species [27]. The high temperature peak at 700 -600 °C can be attributed mainly to the ceria oxide species that were not in contact with Pt particles [28], in addition to the reduction of carbon superficial groups and/or methanation of the support [29]. For the PtCeO2/GOr-N sample, the reduction peak is shown at slightly lower temperature as an indicative of a better Pt and CeO2 particle dispersion [23], which is consistent with the XRD results. The presence of N in the carbon structure may act as the preferential nanoparticles nucleation site depending on the type of nitrogen group [30]. The doping of nitrogen in the carbon structure enhances the interaction among Pt, CeO2 and the support. Potentially, such a strong metal-support interaction could improve the stability of this catalyst against possible sintering or leaching of metal particles [31]. Thereby, the doping of nitrogen results in smaller metal particle size and better dispersion since smaller particles showed the reduction processes at lower temperatures [23]. As for the Ni based catalysts the H2-TPR profiles of NiCeO2/GOr and NiCeO2/GOr-N reveal two reduction zones. The first one between around 240 oC can be ascribed to the reduction of NiO species in contact with CeO2 [17]. The wideness of this hydrogen consumption event points to the presence of nickel particles with different reducibility as well as an heterogenous interaction with the support [23]. The second peak between 400 oC and 800 oC is attributed to ceria surface reducible species [17], along with the hydrogenation of carbon atoms in graphite [28], since methane production has been observed in this temperature range. TEM-EDX Transmission Electron Microscopy (TEM) was used along with Energy dispersive X-ray fluorescence (EDX), to study the composition and distribution of the different elements in the synthesized reduced samples. Figure 3 (a-b) displays TEM images of Pt/GOr and Pt/GOr-N catalysts. The micrographs show clearly the graphene layers and some platinum particles. Generally speaking, N-doped sample displayed smaller Pt-particles size and better dispersion compared to unpromoted graphene. Nitrogen identification in micrographs is beyond the technique capabilities since nitrogen doping atoms are situated at the atomic level inside the support structure. Consequently, a surface mapping of the N-doped sample was conducted Furthermore, the conditioning step also improves the selectivity trends. Four monoaromatic compounds, including phenol, cresol, catechol and anisole were detected in the organic phase. Anisole production was not seen in the previous situation using the fresh ascalcined catalysts. Overall production of partial deoxygenated products (i.e. phenol + anisole) improved by a factor of three over the NiCeO2/GOr-N catalysts, indicating its superior ability for demethoxylation and/or dehydroxylation of Ni-based reduced catalysts compared to the fresh material. In fact, this sample presents the best activity/selectivity balance among all the studied samples in both activity tests (with and without activation). The degree of deoxygenation (accounted by the total production of mono-oxygenated compounds) is optimal for this sample. Catalysts conversion varies between 28 % and 31 %. These results represent around 40% increase in conversion compared to the non-catalytic process and are quite acceptable since the hydrogen supplier in our reaction system is water. A numerical table expressing the data observed in the bar graph can be found in the electronic supporting information. It is assumed that the calcined Ni-based samples were not completely reduced in the calcination process even though the presence of nickel oxide was not presented in the XRD patterns. Besides, we could verify the latter through TPR profiles in which a certain amount of NiO was reduced at low temperatures. The complete reduction of Ni on this sample was beneficial for the catalytic activity. In contrast, the reduction treatment is not particularly favourable for Pt catalysts since there is not much progress in terms of conversion and selectivity. From the process perspective adding a pre-conditioning step for the Pt-based materials would represent extra cost and no major benefits in terms of performance. In fact, according to the TPR experiments platinum species are almost completely reduced after calcination, what explains its superior behaviour when the fresh non-activated samples are tested (Figure 7). The conditioning in H2 at 350oC could potentially foster a certain degree of nanoparticles sintering thus explaining the absence of a tangible performance improvement upon pre-activation within the Pt family (Figure 8). The appearance of anisole when the NiCeO2/GOr-N catalyst is implemented reflects the hydrogenation of C (sp2)- OH bonds. Anisole and catechol can be further hydrodeoxygenated into phenol. Nevertheless, anisole is a more desired intermediate since it is a more advanced deoxygenated product (only one oxygen in its composition) than catechol. A simplified reaction pathway for guaiacol HDO is proposed in Figure 9 which is inspired in previous studies and matches well with our observations. Anisole and catechol are intermediates products from dehydroxylation (DHO) and demethylation (DME) reactions respectively. Both of them should lead to phenol, which is the most advanced product we have obtained, and it is just the previous step before accomplishing full oxygen removal to yield benzene. Generally speaking all our catalysts yielded a good proportion of catechol indicating that the DME routes prevail over the DHO. This is logical in principle given the C-O bond-breaking capacity of the studied samples and the ability to protonate the reaction intermediates in the hydrothermal media leading to catechol (doubly hydroxylated compounds). Catechol production proceeds via C(sp3)- O bond scission which is the weakest interaction among the possible bond-breaking options in the reaction scheme thus facilitating the DME pathway [32]. Interestingly the NiCeO2/GOr-N is the only system generating anisole which indicates its ability to proceed via DHO routes. Finally, the presence of o-cresol corresponds to a transalkylation and a subsequent DHO process as previously observed elsewhere [33]. SPENT SAMPLES CHARACTERISATION Given the relatively aggressive reaction conditions (i.e. high temperature and high pressure) the spent samples characterisation becomes necessary to ascertain potential changes in the material’s structure. XRD Figure 9. Proposed reaction pathways of guaiacol HDO over noble metal catalysts (DMO: Demethoxylation DME: Demethylation; DHO: Dehydroxylation) [7]. The XRD patterns of spent reduced catalysts, including Pt/GOr, Pt/GOr-N, PtCeO2/GOr-N and NiCeO2/GOr-N are presented in Figure 10. These samples were selected for the postreaction study given their promising performance in the catalytic tests. In comparison to the calcined samples, some news diffractions peaks have been detected for the NiCeO2/GOr-N catalyst. Diffractions peaks at 37.3o, 43.3o and 63 o are observed, corresponding to the (111), (200) and (220) planes of the NiO fcc phase respectively, in agreement with JCPDS no. 00- 047-1049. The latter points out oxidation of the Ni metallic phase which is somehow expected given the selected reaction media (H2O). On the other hand, such oxidation phenom is not appreciated within the Pt based samples reflecting the noble character of this metal. However, PtOx presence cannot be fully disregarded, as the low metal loading in the sample hampers its detection via XRD. TEM-EDX TEM images of the most actives samples (PtCeO2/GOr-N and NiCeO2/GOr-N) were presented in Figure 11 and 12. The micrographs of the spent reduced PtCeO2/Gr-N sample (Figure 11) show that, in comparison with the calcined samples the size of the particles slightly increased due to the sintering of the metal particles. This is ascribed to the reduction treatment and the reaction conditions used. Figure 10. XRD patterns of the spent reduced samples. The dispersion of the active phase in spent reduced NiCeO2/GOr-N can be observed in Figure 12. Compared to calcined sample, few agglomerate clusters are observed and the particle distribution still remains homogenous. This is indeed a remarkable observation since it showcases the structural and morphological stability of the NiCeO2/GOr-N sample for the hydrothermal upgrading of lignin model compounds despite the very demanding reaction conditions. CONCLUSIONS This work proposes an innovative catalytic route for biomass upgrading based on a “H2-free” HDO strategy. More precisely, we demonstrate the viability of hydrothermal deoxygenation of guaiacol as lignin model compound suppressing external hydrogen input. For such challenging process we have developed multicomponent catalysts able to activate water and conduct the subsequent HDO reaction. Our catalysts based on Ni or Pt nanoparticles supported on N-doped graphene decorated with ceria particles result effective in the partial oxygen removal from the initial feedstock. Figure 11. Micrographs of the PtCeO2/GOr-N spent reduced sample. Figure 12. TEM images of the NiCeO2/GOr-N spent reduced sample. Both Pt and Ni systems present well-dispersed active phases and an intimate metal-support contact which tunes the catalytic properties and favours the overall activity. The systems doped with N displayed better dispersion and optimal properties to tackle the HDO reaction. Pt based catalysts are superior to Ni ones when the reaction is conducted with the as-prepared non-activated materials. Nevertheless, a catalysts’ pre-conditioning in a reductive environment changes the situation resulting on the superior behaviour of NiCeO2/GOr-N. This sample is the only one leading to anisole – an advanced reaction product which eventually leads to phenol and benzene triggering the complete oxygen removal. Very importantly the samples are stable under the reaction conditions (high pressure, high temperature in a hydrothermal media) as evidenced in the post reaction XRD and TEM analysis. Overall the catalytic performance achieved with the engineered catalysts may be deemed as modest in comparison to that exhibited by other catalysts in the HDO reaction. Nevertheless, the unique advantage of our process is the suppression of external hydrogen input, hence although still preliminary this work should spark new research attempts to further refine the catalytic formulation and boost the performance. This way we will be showcasing the pivotal role of heterogeneous catalysis for bio-resources upgrading to facilitate the transition towards a low carbon economy. ACKNOWLEDGEMENTS Financial support for this work was provided by the Department of Chemical and Process Engineering at the University of Surrey and the EPSRC grant EP/R512904/1 as well as the Royal Society Research Grant RSGR1180353. LPP acknowledge Comunitat Valenciana for her APOSTD2017 fellowship. This work was also partially sponsored by the CO2Chem through the EPSRC grant EP/P026435/1 and the Spanish Ministry of Science and Innovation through the project RYC2018-024387-I. DHCA thanks the Consejo Nacional de Ciencia y Tecnologia, CONACyT, Mexico for a postdoctoral fellowship grant number 2018-000022-01EXTV-00519, 2019-000029-01EXTV-00057. REFERENCES [1] A.R. Hoffman, "Water, Energy, and Environment – A Primer ", IWA publishing, 2019 pp. 61–83. https://doi.org/10.2166/9781780409658_0061. [2] F. Cherubini, Energy Convers. Manag. 51 (2010) 1412–21. https://doi.org/10.1016/j.enconman.2010.01.015. [3] W. Jin, L. Pastor-Pérez, D.K. Shen, A. Sepúlveda-Escribano, S. Gu, T. Ramirez Reina, ChemCatChem 11 (2019) 924–960. https://doi.org/10.1002/cctc.201801722. [4] C.B. Field, J.E. Campbell, D.B. Lobell, Trends Ecol. Evol. 23 (2008) 65-72. https://doi.org/10.1016/j.tree.2007.12.001. [5] M. Qingqing, S. Xiaojun, L. Huizhen, H. Buxing, CCL 30 (2019) 15-24. https://doi.org/10.1016/j.cclet.2018.04.032. [6] L. Fan, Y. Zhang, S. Liu, N. Zhou, P. Chen, Y. Cheng, M. Addy, Q. Lu, M.M. Omar, Y. Liu, Y. Wang, L. Dai, E. Anderson, P. Peng, H. Lei, R. Ruan, Bioresour. Technol. 241 (2017) 1118-1126. https://doi.org/10.1016/j.biortech.2017.05.129. [7] H. Shafaghat, J.M. Kim, I.G. Lee, J. Jae, S.C. Jung, Y.K. Park, Renew. Energy 144 (2019) 159-166. https://doi.org/10.1016/j.renene.2018.06.096. [8] X. Wang, Q. Fan, S. Yu, Z. Chen, Y. Ai, Y. Sun, A. Hobiny, A. Alsaedi, X. Wang, Chem. Eng. J. 287 (2016) 448-455. https://doi.org/10.1016/j.cej.2020.124272. [9] W. Jin, J.L. Santos, L. Pastor‐Perez, S. Gu, M.A. Centeno, T.R. Reina, ChemCatChem 11 (2019) 4434–4441. https://doi.org/10.1002/cctc.201900841. [10] D. Gao, C. Schweitzer, H.T. Hwang, A. Varma, Ind. Eng. Chem. Res. 53 (2014) 18658– 18667. https://doi.org/10.1021/ie500495z. [11] W. Jin, L. Pastor-Pérez, J.J. Villora-Picó, A. Sepúlveda-Escribano, S. Gu, T.R. Reina, ACS Sustain. Chem. Eng. 7 (2019) 16041-16049. https://doi.org/10.1021/acssuschemeng.9b02712. [12] A. Bjelić, M. Grilc, M. Huš, B. Likozar, Chem. Eng. J. 359 (2019) 305-320. https://doi.org/10.1016/j.cej.2018.11.107. [13] W. Jin, L. Pastor-Pérez, J.J. Villora-Pico, M.M. Pastor-Blas, A. Sepúlveda-Escribano, S. Gu, N.D. Charisiou, K. Papageridis, M.A. Goula, T.R. Reina, Energies 13 (2019). https://doi.org/10.3390/en13010132. [14] M.N. Groves, C. Malardier-Jugroot, M. Jugroot, J. Phys. Chem. C 116 (2012) 10548– 10556. https://doi.org/10.1021/jp203734d. [15] R. Shibuya, T. Kondo, J. Nakamura, Carbon-Based Met. Catal. Des. Appl. 1–2 (2018) 227–249. https://doi.org/10.1002/9783527811458.vol1-ch8. [16] G.H. Wang, Z. Cao, D. Gu, N. Pfänder, A.C. Swertz, B. Spliethoff, H.J. Bongard, C. Weidenthaler, W. Schmidt, R. Rinaldi, F. Schüth, Angew. Chemie - Int. Ed. 55 (2016) 8850–8855. https://doi.org/10.1002/anie.201511558. [17] L.C. Loc, P.H. Phuong, D. Putthea, N. Tri, N.T. Thuy Van, H.T. Cuong, Int. J. Nanotechnol. 15 (2018) 968–982. https://doi.org/10.1504/IJNT.2018.099935. [18] W. Jin, J.L. Santos, L. Pastor-Perez, S. Gu, M.A. Centeno, T.R. Reina, ChemCatChem 11 (2019) 4434–4441. https://doi.org/10.1002/cctc.201900841. [19] E. Lam, J.H.T. Luong, ACS Catal. 4 (2014) 3393-3410. https://doi.org/10.1021/cs5008393. [20] F. Barroso-Bujans, Á. Alegría, J. Colmenero, J. Phys. Chem. C 114 (2010) 21645-21651. https://doi.org/10.1021/jp108905j. [21] R. Kumar, M. Mamlouk, K. Scott, Int. J. Electrochem. 2011 (2011) 1–7. https://doi.org/10.4061/2011/434186. [22] N. Wang, Z. Yang, F. Xu, K. Thummavichai, H. Chen, Y. Xia, Y. Zhu, Sci. Rep. 7 (2017) 1– 9. https://doi.org/10.1038/s41598-017-12200-1. [23] A.B. Dongil, L. Pastor-Pérez, A. Sepúlveda-Escribano, P. Reyes, Appl. Catal. A Gen. 505 (2015) 98–104. https://doi.org/10.1016/j.apcata.2015.07.036. [24] L. Pastor-Pérez, A. Sepúlveda-Escribano, Appl. Catal. A Gen. 529 (2017) 118–126. https://doi.org/10.1016/j.apcata.2016.10.022. [25] G. Liu, Q. Chen, E. Oyunkhand, S. Ding, N. Yamane, G. Yang, Y. Yoneyama, N. Tsubaki, Carbon 130 (2018) 304-314. https://doi.org/10.1016/j.carbon.2018.01.015. [26] X. Qian, H. Qin, T. Meng, Y. Lin, Z. Ma, Materials 7 (2014) 8105-8130. https://doi.org/10.3390/ma7128105. [27] B. Chen, B. Wang, Y. Sun, X. Wang, M. Fu, J. Wu, L. Chen, Y. Tan, D. Ye, Catalysts 9, no.2 (2019) 1-18. https://doi.org/10.3390/catal9010002. [28] C. Zhang, W. Lv, Q. Yang, Y. Liu, Appl. Surf. Sci. 258 (2012) 7795–7800. https://doi.org/10.1016/j.apsusc.2012.03.131 [29] C. Price, L. Pastor-Pérez, E. le Saché, A. Sepúlveda-Escribano, T.R. Reina, Int. J. Hydrog. Energy 42 (2017) 10747-10751 . https://doi.org/10.1016/j.ijhydene.2017.02.013. [30] C. Ozga, K. Jänkälä, P. Schmidt, A. Hans, P. Reiß, A. Ehresmann, A. Knie, J. Electron Spectros. Relat. Phenomena 207 (2016) 34-37. https://doi.org/10.1016/j.elspec.2018.02.001. [31] W. Xia, Catal. Sci. Technol. 6 (2016) 630-664. https://doi.org/10.1039/c5cy01694k. [32] Z. Lin, R. Chen, Z. Qu, J.G. Chen, Green Chem. 20 (2018) 2679–2696. https://doi.org/10.1039/c8gc00239h. [33] W. Jin, L. Pastor-Pérez, J. Yu, J.A. Odriozola, S. Gu, T.R. Reina, Curr. Opin. Green Sustain. Chem. 23 (2020) 1-9. https://doi.org/10.1016/j.cogsc.2019.12.008.