Understanding the Role of the Acid Sites in 5-Hydroxymethylfurfural Oxidation to 2,5-Furandicarboxylic Acid Reaction over Gold Catalysts: Surface Investigation on CexZr1- xO2 Compounds

1 Understanding the role of the acid sites in HMF oxidation to FDCA reaction over gold catalysts: surface investigation on CexZr1-xO2 compounds Cristina Megías-Sayago †,*, Kristina Chakarova ‡, Anna Penkova †,*, Alice Lolli ¥, Svetlana Ivanova †, Stefania Albonetti ¥, Fabrizio Cavani ¥, José Antonio Odriozola † † Departamento de Química Inorgánica e Instituto de Ciencia de Materiales de Sevilla, Universidad de Sevilla-CSIC, Américo Vespucio 49, 41092, Seville, Spain ‡ Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, Sofia 1113, Bulgaria ¥ Dip. di Chimica Industriale “Toso Montanari”, Università di Bologna, Viale Risorgimento 4, 40136 Bologna (BO), Italy ABSTRACT A series of CexZr1-xO2 supports with different Ce/Zr molar ratios were utilized for the preparation of gold catalyst used in the selective oxidation of 5-hydroxymethyl-2- furfural to 2,5-furandicarboxylic acid. The used method of gold deposition allows the preparation of homogeneous gold particles size and shape distribution, a formulation very useful for studies dedicated to reveal the support participation in the reaction. The supports are characterized by FTIR using CO as probe molecule and the sample catalytic activity is thereafter correlated to the support acid sites distribution. The possible participation of its Lewis/Bronsted acidity in the reaction mechanism is also proposed. 2 KEYWORDS: gold, 5-hydroxymethyl-2-furfural, oxidation, 2,5-furandicarboxylic acid, acid sites, CeZrO, CeO2, ZrO2 1. INTRODUCTION The economic and environmental impact derived from the use of fossil fuels is increasingly evident in nowadays society. The continuous exploitation of the nonrenewable feedstock led to its continuous depletion and increase of the oil price, urging the findings of renewable energy sources and/or materials [1–3]. Biomass has been proposed as the most promising alternative, allowing the production of fuels and chemicals from renewable raw materials such as lignocellulosic fraction [4,5]. The number of involved processes and technologies to transform the mentioned fraction into chemicals is increasing thanks to both, new catalytic technologies and synthetic routes recently discovered. An example of such a process is the efficient conversion of furfural derivatives, synthesized by means of hydrolysis and dehydration reactions from cellulose and hemicellulose [6,7]. In particular, the 5-hydroxymethylfurfural (HMF) is considered as one of the most important platform molecules derived from the lignocellulosic biomass since it can be transformed into several compounds used for fuels or chemicals synthesis [8]. The oxidation of both, aldehydic and the alcoholic functionalities of HMF leads to the formation of 2,5-furandicarboxylic acid (FDCA), a key monomer in the synthesis of new generation polymers and plastics [9,10] such as polyethylene furandicarboxylate (PEF), considered the most competitive substitute to the commonly used polyethylene terephthalate (PET) [11]. The general HMF oxidation network includes also 5-hydroxymethyl-2-furancarboxylic acid (HMFCA), 2,5- diformylfuran (DFF) and 5-formyl-2-furancarboxylic acid (FFCA) as reaction intermediates as presented in Scheme 1. 3 Scheme 1. General reaction network for HMF oxidation. In the production of the FDCA, noble metal catalyzed processes result in important product yields [12–15]. Concretely, gold catalysts demonstrated relevant activity in presence of oxygen [16,17] although the use of base as co-catalyst has been established as general requirement [18,19]. This actually is considered the main drawback of the process, causing reactor corrosion or active phase leaching. For this reason, the development of highly active gold catalysts and low base concentration as co-catalyst are still serious challenges. The supported metal nanoparticles could present different catalytic behavior depending on the used support; the interaction between gold and supports, such as CeO2, increased the intrinsic activity of the active phase both, in gas phase oxidation reactions [20–22] and in liquid phase alcohol oxidation processes [13–15]. It is usually considered that the support Lewis acid sites are responsible for the interaction of alcohol or the forming alkoxide species as main intermediates in this process with the catalyst [23,24]. Nonetheless, there is a lack of detailed studies correlating gold catalyst activity and CeO2 acidity and their synergy in the liquid phase selective oxidation reactions in presence of water and base. Therefore, the aim of this work is to study the influence of the support acidity in the HMF oxidation to FDCA. For this purpose, gold was deposited over a series of 4 CexZr1-xO2 supports with different Ce/Zr molar ratios. The physicochemical properties of the support and its Brönsted/Lewis acid sites interaction with CO as probe molecule were studied by FTIR spectroscopy and the results are correlated to the obtained final product yield. 2. EXPERIMENTAL SECTION 2.1. Materials Commercially available supports CeO2 (VP Ad Nano 90, Evonik) CeO2(25%wt)/ZrO2 and CeO2(50%wt)/ZrO2 (99%, Daiichi Kigenso Kagaku Kogio Co., Ltd) were used as received. For the preparation of the ZrO2 support, ZrN2O7·xH2O (99%, Sigma Aldrich) was used as precursor and ammonia (30%, Panreac) as precipitating agent. HAuCl4 (Johnson Matthey) was used as gold precursor. For the catalytic reaction the following products were used: sodium hydroxide (≥99%, Sigma Aldrich), 5- hydroxymethylfurfural (HMF) (≥99% Ava Biochem), 2,5-furandicarboxylic acid (97% FDCA), 5-hydroxymethyl-2-furancarboxylic acid (≥99% HMFCA), 5-formyl-2- furancarboxylic acid (≥99% FFCA) and 2,5-diformylfuran (97% DFF) (the last three from Toronto Research Chemicals). 2.2 Catalyst preparation ZrO2 support was prepared by precipitation [25]. Zirconium precursor was dissolved upon heating and the pH was fixed to 9 by using NH3. The obtained precipitated solid was aged for 4 h, filtered and dried at 110 ºC overnight before calcination in static air at 400 ºC for 1 h, with a heating ramp of 10 ºC/min. Gold (2 wt% nominal value) was deposited according to the Direct Anionic Exchange (DAE) method assisted by ammonia as reported elsewhere [26]. A solution of the gold 5 precursor (around 10-4M) was heated to 70 ºC, and then contacted with the support and 20 min later with NH3. The final solid was filtered, dried at 100 ºC overnight and calcined at 300 ºC for 4 hours. Ce, Ce25Zr, Ce50Zr, and Zr was the adopted nomenclature for CeO2, CeO2(25%wt)/ZrO2, CeO2(50%wt)/ZrO2 and ZrO2, respectively and preceded by Au for the catalysts. 2.3 Characterization Specific surface areas of the samples were measured by N2 physisorption equipment (Sorpty 1750 CE instruments) and the single-point BET method. All samples were previously outgassed at 150 ºC under vacuum. XRD measurements were performed at room temperature on Panalitycal X’Pert Pro diffractometer, equipped with Cu anode. All diffractograms were recorded in the 5-90º 2θ range, with 0.05 º step size and 300 s acquisition time. Gold contents and chemical compositions of solids were determined by X Ray Fluorescence (XRF) using Panalitycal AXIOS spectrometer with Rh tube of radiation. FTIR spectra were recorded with a Nicolet 6700 FTIR spectrometer equipped with an MCT detector. The measurements were performed accumulating up to 64 scans at a spectral resolution of 2 cm−1 and accuracy of 0.01 cm-1. Self-supporting wafers (≈ 10 mg cm-2) or pellets pressed on W-grid were prepared from the powdered samples and treated directly in a purpose-made IR cell allowing measurements at ambient and low temperatures. The cell was connected to a vacuum-adsorption apparatus with a residual pressure lower than 10−3 Pa. Prior to the adsorption measurements, the samples were activated by heating in air at 300 °C and subsequent evacuation at the same temperature. 6 2.4 Catalytic tests The oxidation of 5-hydroxymethylfurfural was carried out in a Parr autoclave reactor (100 mL) provided with a mechanical stirrer and temperature and pressure control. The reactor was charged with a 0.08M aqueous solution of HMF, NaOH and catalyst in a HMF:Au:NaOH molar ratio of 1:0.01:2. The metal loading experimentally obtained by XRF analysis was used to calculate the amount of catalyst according to the above mentioned ratio. Before test, the reactor was purged twice with pure O2 and finally pressurized at 10 bars. Then the temperature was increased to 70 ºC and the reaction mixture stirred at 400 rpm for 4 hours. The reaction starts (t = 0) when the temperature reached 70 ºC (in about 10 min). After 4 hours, the reactor was cooled down to room temperature in an ice bath and the reaction mixture centrifuged and filtered. A sample was taken and diluted before analysis carried out on Agilent Infinity 1260 liquid chromatograph equipped with Aminex HPX-87H 300 mm × 7.8 mm column using 0.005 M H2SO4 as eluent. Conversion, selectivity and carbon balance were calculated from the peak areas, after calibration using the commercially available samples as references. 3. RESULTS AND DISCUSSION 3.1. XRD measurements The X-ray diffraction patterns of both, supports and catalysts are shown in Fig. 1. The CeO2 solid shows well-defined diffractions corresponding to the cubic fluorite structure (ICSD #00-034-0394). For the mixed oxides, the formation of homogeneous solid solutions and absence of segregated phases are confirmed on the basis of the diffraction patterns. In fact, all binary systems present typical diffraction patterns of CexZr1−xO2 solid solutions [27,28]. Neither Ce50Zr nor Ce25Zr show diffractions corresponding to 7 pure CeO2 and/or ZrO2. Upon 50% Zr addition the diffractions shift to higher 2  angles, attributed to ceria lattice shrinkage due to the replacement of Ce4+ (0.098 nm ionic radii) with smaller Zr4+cations (0.084 nm) and solid solution formation. On the contrary, for the Ce25Zr solid the addition of 75% Zr suggests zirconia lattice expansion due, in this case, to the replacement of Zr4+cations with bigger Ce4+ cations. Therefore, the Ce50Zr sample presents cubic structure (ICSD #00-028-0271), whereas Ce25Zr the tetragonal one (ICSD #01-080-0785). This is consistent with the reported literature [28]. The addition of Zr to ceria implies also a secondary effect, oxides particle size diminution, confirmed by the broadening of the diffractions for both mixed oxide samples. For the pure ZrO2 sample a mixture of cubic (ICSD #00-003-0640) and monoclinic (ICSD #00-007-0343) structures is observed. 20 30 40 50 60 70 80 90 AuZr Zr AuCe25Zr Ce25Zr AuCe50Zr Ce50Zr AuCe Au Intensity (a. u) o2 Ce Figure 1. XRD patterns of supports and catalysts. 8 As for the catalysts, gold diffractions are hardly observed, suggesting particle size below 4 nm as confirmed by the TEM analysis (Table 1). As a matter of fact, the application of this method of preparation results in a very homogeneous gold particles size distribution, especially, for the catalysts supported on bare or mixed oxides with isoelectric point between 5-8, the case of all studied supports in this work [29]. The same catalysts were employed for the glucose oxidation reaction and a detailed microscopic characterization revealed that indeed the obtained gold particle size is very similar with an average value between the samples 3.6  1nm [23], thus concluding that this synthetic approach is reliable and allows the preparation of small metal nanoparticles. The actual gold loadings, particle size and textural properties of the fresh catalysts are listed in Table 1. Except for the AuCe sample, gold loadings are close to the targeted value. Both gold particle size and specific surface area are comparable for all samples. The Zr4+ introduction into cerium oxide lattice or vice versa does not lead to any significant change of the textural parameters. Table 1. Characterization data of prepared catalysts. Sample % wt. Au Au Particle size (nm) Specific surface area (m2/g) AuCe 3.5 3.9 ± 1.5 57 AuCe50Zr 2.4 3.3 ± 1.4 55 AuCe25Zr 2.4 3.9 ± 1.2 57 AuZr 2.3 3.1 ± 1.5 70 9 3.2. Catalytic measurements In general, the incorporation of Zr4+ to CeO2 to form a mixed oxide leads to an increase of the metal center acidity, since the Zr-O bond is characterized by higher ionic character in comparison to the Ce-O bond [30]. For this reason, it is interesting to study in detail the influence of the support acidity in HMF oxidation reaction using as model system pure and mixed ceria-zirconia oxides (Figure 2). 0 10 20 30 40 50 60 70 80 90 AuCe AuCe50Zr AuCe25Zr AuZr Yields (%) HMFCA FDCA FFCA Figure 2. Products yields (%) over different gold supported catalysts in the HMF oxidation reaction. HMF conversion is complete in all cases (not plotted). Reaction conditions: HMF:NaOH:Au molar ratio 100:2:1, 10 bar O2, 70 ºC, 400 rpm, 4h. All studied catalysts present higher activity in comparison to previously Au-based reported systems. It is worth to mention that the catalysts are unable to convert HMF in base-free conditions making the presence of base obligatory for the reaction, as also reported in the literature[16,17,19]. The base equivalents varied a lot in all reported studies but never such a smaller quantity, like in our case, was reported (NaOH/HMF molar ratio of 2). It is very uncommon for monometallic Au supported catalyst, usually requiring higher amount of base. 16 studied. During evacuation, the band of the physisorbed CO (2141 cm−1) quickly disappears, while the band at 2153 cm−1 is more persistent and its maximum is gradually shifted to 2160 cm−1. As the CO coverage diminishes, a new low-frequency shoulder at 2154 cm−1 becomes visible. Both bands disappear at low temperatures suggesting electrostatic van der Waals interaction. A similar behavior was reported by Binet et al. [34,40]. They observed an intense carbonyl band at 2151 cm−1 and another weak one at 2162 cm−1 on activated CeO2, which disappear independently at coverage decreasing, and shift to 2157 and 2168 cm−1, respectively. The high-frequency band at 2168 cm−1 has been attributed to carbonyls bonded to Ce4+ cations, and the band at 2157 cm−1 is assigned to a weak CO interaction with surface OH groups. Moreover, in our previous work using the same oxide [33] activated in the same conditions, a band centered at 2158 cm−1 is attributed to CO hydrogen bonded to Ce4+ hydroxyls. Careful analysis of the OH stretching region (Fig. 5, right inset), shows that at high coverage the three Ce4+-hydroxyl species (isolated hydroxyls at 3721 cm−1 and the two different bridged hydroxyls at 3690 and 3666 cm−1) decrease and transform to H-bonded hydroxyls with bands centered at 3628 and 3600 cm−1. It is difficult to establish which OH group is the most affected by the CO adsorption, but one is sure, at least two H- bonded hydroxyls are observed and can be tentatively correlated to both carbonyl bands detected at 2160 and 2154 cm−1. About the band at 2095 cm−1, the natural abundance of 13CO isotope in carbon dioxide (ca. 1%) gives rise to corresponding 13CO-OH species. Since apparently only Brönsted acid sites were detected and no indication of the Lewis sites appear, the sample was re-activated at the same temperature (300ºC) under vacuum for 15 more minutes (Fig. 6). Logically, an additional dehydroxylation occurs resulting in lower-intensity OH bands. Although, these bands are not well resolved after CO addition (Fig. 6, spectrum b´), the same H-bonded hydroxyls are observed after the first 17 activation (Fig. 6, spectra b and b´ and 5, spectrum b). In the carbonyl region, at high coverage, in addition to the band attributed to CO hydrogen bonded to Ce4+-OH, shifted to lower wavenumbers at 2150 cm−1 (Fig. 6, spectrum b´), a band at 2167 with a shoulder at 2170 cm−1 (Fig. 6, spectrum b´ and Fig. 7, spectrum b) are observed and associated with two new Ce4+-CO species. Thus a presence of two different Ce4+ Lewis centers are confirmed upon additional reactivation [41]. 2180 2160 2140 2120 21003800 3700 3600 3500 -2150 a  b  a  b a -3721 -3690 -3665 -2153 -3628 -3600 Absorbance, a.u. Wavenumber, cm-1 0.02 2170 2167 b  b a 0.2 2141 Figure 6. Blue spectra: (a, solid line) background spectrum of activated sample; (b, hollow circles) in 5 mbar CO equilibrium pressure at 100 K. Magenta spectra: (a´, solid line) background spectrum of activated sample subjected to additional evacuation at 300oC; (b´ hollow circles) in 5 mbar CO equilibrium pressure at 100 K. When the CO coverage decreases during evacuation, a higher-frequency shift to 2175 cm−1 is observed (Fig. 7, left inset), attributed to the static shift and supporting the idea that both Ce4+centers are not isolated. Considering the stability of all species formed 18 upon evacuation, both groups of carbonyls disappear at low temperature, the Ce4+-CO species being more stable than the Ce4+-OH…CO ones. 2250 2200 2150 2100 2050 b 2155 2159 2175 2141 2095 2170 2153 Wavenumber, cm-1 2150 m a 0.1 3800 3700 3600 3500 3666 3688 - 3610 - 3635 a m 3585 3654 3705 - 0.02 m a 2170 2167 Figure 7. FTIR spectra of CO adsorbed on CeO2: activated sample subjected to additional evacuation at 300ºC, in 5 mbar CO equilibrium pressure at 100 K (a) and during outgassing at 100 K (b-m). The spectra are background and CO gas-phase corrected. 3.3.3. Mixed CeZr supports After activation of the mixed Ce50Zr sample several hydroxyl bands are detected, at 3782 and 3682 cm−1 along with another weak band at 3719 cm−1 (Fig. 8, spectrum a left inset). The latter is observed for ceria only and it is unambiguously assigned to Ce4+-OH groups. Concerning the other two bands, they can be attributed to the pure zirconia hydroxyls, as observed above for pure ZrO2. However, it cannot be excluded that the band at 3682 cm−1 could arise from hydroxyls bridging Zr and Ce cations, as a band at 3690 cm−1 is registered on pure ceria. 19 2250 2200 2150 2100 2050 c p 2143 2182 q 2192 2167 2175 Absorbance, a.u. Wavenumber, cm-1 2155 v b 0.05 3800 3700 3600 3500 b l 3618 3686 3789 0.05 3800 3700 3600 3500 o d a b 0.05 3630 3682 3719 -3782 Figure 8. FTIR spectra of CO adsorbed on Ce50Zr: activated sample (a), in 5 mbar CO equilibrium pressure at 100 K (b) and during outgassing at 100 K (c-p) and at increasing temperature (q-v). The spectra are background and CO gas-phase corrected except for the spectra in the left inset. After CO adsorption, the intensities of the hydroxyl bands diminish followed by a shift to 3618 cm−1 due to the formation of OHCO species (Fig. 8, spectrum b, left and right insets). Regarding the carbonyl region, the spectrum is very similar to those of CeO2 (see Fig. 5). An intense band at 2155 cm−1 slightly asymmetric around 2143 cm−1 and another one at 2175 cm−1 are registered. During the evacuation, the 2143 cm−1 band becomes hardly noticeable in the next spectrum, whereas the 2155 cm−1 band shifts to higher frequency showing a similar behaviour as the corresponding band in pure ceria. This band was attributed to Ce4+-OHCO species and the low frequency shoulder – to physisorbed CO. It should be underlined that the 2155 cm−1 band is less intense in 20 comparison to pure ceria, which suggests that the hydroxyl groups are also connected to zirconium cations. At low coverage, along with the band at 2175 cm−1 another band at 2167 cm−1 became visible. When the band at 2167 cm−1 is no longer evident (Fig. 8, spectrum o), the only band that remains in the spectrum is centered at 2182 cm−1. This band is stable at low temperature and shifts to higher frequency (2192 cm−1) when the temperature increases. Taking into account that the Zr4+-CO species are stable at 100 K and decomposes only with temperature increase, whereas Ce4+-CO complexes decompose quickly even at low temperatures, both types of carbonyls can easily be differentiated. The bands centered at 2175 and 2167 cm−1 disappear at low temperature, while the band situated at 2182 cm−1 is more stable. Therefore, the first two bands give notice of the existence of two different Ce4+sites, whereas the more stable band at 2182 cm−1 can be associated with Zr4+-CO species, indicating higher Lewis acidity of the Zr4+ sites. As for the Ce25Zr sample, the background spectrum shows the same hydroxyl bands detected for the Ce50Zr sample (Fig. 9, spectrum a, left inset), the band corresponding Ce4+-OH band (3719 cm−1) being less intense, due to the lower ceria content. When CO is adsorbed, the hydroxyl bands decrease in intensity and shift to lower wavenumbers (3616 cm−1). In the carbonyl region, the spectra appear to be different from the Ce50Zr sample and very similar to those of pure zirconia. The bands assignment, based on the samples studied above, is as follows: 2148 cm−1 band corresponds to Zr4+-OHCO species; the band at 2180 cm−1, stable at low temperature, is attributed to carbonyls formed with highly electrophilic and highly coordinatively unsaturated Zr4+ sites (converted into monocarbonyls at lower coverage); Ce4+ cations are detected as Ce4+-OHCO species (2159 cm−1) and as Ce4+-CO complexes (2171 21 cm−1), the latter being assigned according to their lower stability as compared to zirconia carbonyls (Fig. 9, spectrum p). 2250 2200 2150 2100 2050 p 2148 2193 2171 2180 Absorbance, a.u. Wavenumber, cm-1 2159 t b 0.05 3800 3700 3600 3500 b m 3616 3686 3789 0.05 3800 3700 3600 3500 h q c d a b 0.05 3630 3682 -3782 3719 Figure 9. FTIR spectra of CO adsorbed on Ce25Zr: activated sample (a), in 5 mbar CO equilibrium pressure at 100 K (b) and during outgassing at 100 K (c-p) and at increasing temperature (q-t). The spectra are background and CO gas-phase corrected except for the spectra in the left inset. 3.3.4 Au-containing sample In order to study the changes induced by the gold incorporation, the AuCe25Zr sample investigation followed the same methodology used with the supports (Fig. 10). In the OH region, in addition to the bands arising from the support – 3784 cm−1 (Zr4+ isolated hydroxyls) and 3676 cm−1 (Ce4+ and/or Zr4+ bridging hydroxyls) – a new intense band is observed at 3743 cm−1 (Fig. 10, spectrum a, left inset). This band could be ascribed directly to gold presence, i.e. to Au-OH species. It is important to underline that after CO adsorption, a bigger shift of the hydroxyl bands (to 3600 cm−1) is observed as 22 compared to the shift detected with the bare support, suggesting higher acidity of the OH species in the Au-containing sample. Regarding the carbonyl region, it is possible to observe the typical bands of Ce4+-OHCO (2161 cm−1) and Ce4+-CO species (observed at low CO coverage and stable at low temperature only) together with a germinal species formed on Zr4+ sites (2184 cm−1). On the other hand, a new band centered at 2102 cm−1 appears and it is attributed to gold carbonyls (Au0-CO) [42]. It should be noted that in this sample no Ce50Zr the Zr4+-OHCO species, characterized by the band at 2148 cm−1, were observed. However, we do not exclude the possibility of their superimposition with the intense Ce4+-OHCO band at 2161 cm−1. 2250 2200 2150 2100 2050 p 2120 2102 2197 2176 q 2184 Absorbance, a.u. Wavenumber, cm-1 2161 t b 0.05 3800 3700 3600 3500 3676 b l 3600 -3743 3786 0.05 3800 3700 3600 3500 c dd a b 0.1 3676 3784 3610 -3743 Figure 10. FTIR spectra of CO adsorbed on AuCe25Zr: activated sample (a), in 5 mbar CO equilibrium pressure at 100 K (b) and during outgassing at 100 K (c-p) and at increasing temperature (q-t). The spectra are background and CO gas-phase corrected except for the spectra in the left inset. 23 As a summary, Fig. 11 shows the comparison of the carbonyl region of all samples at maximal CO coverage (5 mbar equilibrium pressure). For the pure zirconia, Zr4+-OHCO (2149 cm−1) and Zr4+-CO species (2169 and 2182 cm−1) are detected, whereas for the ceria support only Ce4+-OHCO species (2153 cm−1) are observed after the initial activation. Ce4+-CO complexes (2167 and 2170 cm−1) are only visible after additional activation. For the mixed Ce50Zr sample, Ce4+-OHCO species (2155 cm−1), Ce4+-CO complexes (2167 and 2175 cm−1) and Zr4+-CO carbonyls (2182 cm−1) are detected. The Zr4+-OHCO species are not observed because of the presence of a significant ceria content and superimposition of the bands. When the Ce content diminish in the Ce25Zr sample, a predominant Zr4+ species are registered, Zr4+-OHCO (2148 cm−1) and Zr4+-CO (2180 cm−1). The Ce4+ sites are detected principally as Ce4+-OHCO species (2159 cm−1) and lesser amount of Ce4+-CO carbonyls (2171 cm−1). With the inclusion of gold to the AuCe25Zr sample, Ce4+-OHCO (2161 cm−1) and Au0-CO species (2102 cm−1) are detected. In addition, Ce4+ and Zr4+ carbonyls (2171 and 2184 cm−1, respectively) are observed. Considering the OH region, the bigger CO-induced shift of the hydroxyl band of the AuCe25Zr sample suggests a higher acidity of the corresponding hydroxyls. 24 2250 2200 2150 2100 2050 Ce4+-OH-CO 2143 Ce4+-CO Zr4+-CO Zr4+-OH-CO Au0-CO Ce4+-OH-CO AuCe25Zr Ce25Zr Ce50Zr CeO2 Absorbance, a.u. Wavenumber, cm-1 ZrO2 Figure 11. FTIR spectra of CO (5 mbar equilibrium pressure) adsorbed at 100 K on ZrO2, CeO2, Ce50Zr, Ce25Zr and AuCe25Zr. The spectra are background and CO gasphase corrected. Table 2 summarizes all IR bands observed in this study, where the OH bands are present on the samples before CO adsorption and all carbonyl bands appear after it. Table 2. Summary of the IR bands observed in this study Sample OH bands Bands observed after CO adsorption ZrO2 3784 cm-1 → terminal OH groups 3680 cm-1 → tri-bridged OH groups 2149 cm−1 → Zr4+-OH∙∙∙CO 2169 cm−1 → Zr4+-CO 2182 cm−1 → Zr4+(CO)2 2194 cm−1 → Zr4+-CO formed with highly electrophilic and highly coordinatively unsaturated Zr4+ CeO2 3721 cm−1 → isolated (terminal) OH groups 2153 cm−1 → Ce4+-OH∙∙∙CO 2167 and 2170 cm−1 → two types 25 3690 and 3665 cm−1 → bridging OH groups Ce4+-CO Ce50Zr 3782 and 3682 cm−1 → Zr4+-OH 3719 cm−1 → Ce4+-OH 2155 cm−1 → Ce4+-OH∙∙∙CO 2167 and 2175 cm−1 → two types Ce4+-CO 2182 cm−1 → Zr4+(CO)2 Ce25Zr 3782 and 3682 cm−1 → Zr4+-OH 3719 cm−1 → Ce4+-OH 2148 cm−1 → Zr4+-OH∙∙∙CO 2159 cm−1 → Ce4+-OH∙∙∙CO 2171 cm−1 → Ce4+-CO 2180 cm−1 → Zr4+(CO)2 AuCe25Zr 3784 cm−1 → Zr4+-OH 3743 cm−1 → Au-OH 3676 cm−1 → Ce4+ and/or Zr4+ bridging hydroxyls 2161 cm−1 → Ce4+-OH∙∙∙CO 2184 cm−1 → Zr4+(CO)2 2171 cm−1 → Ce4+-CO 2102 cm−1 → Au0-CO The oxidation states of the metal cations were additionally confirmed by XPS analysis (see IS). At first sight, it is not possible to establish a clear correlation between the CO-M4+ bands (M = Ce, Zr) and samples catalytic activity. FDCA yield increases in the following order AuZr < AuCe < AuCe50Zr < AuCe25Zr, while the detected shift of the CO band, due to its interaction with Ce4+/Zr4+ Lewis acid centers, follows a very different trend. The latter implies that supports Lewis acidity is not a factor controlling the catalytic performance. On the contrary, a clear tendency is achieved considering the CO-OH-M4+ bands of the supports (Zr, Ce, Ce50Zr and Ce25Zr) and FDCA yield in Figure 12. Is it true that the band shift is observed for the supports but we can extrapolate this tendency to the catalysts, as the Ce/Zr ratio is the variable and the Au loading is the constant (constant HMF/Au ratio during the reaction). The FDCA yield correlates linearly with the experimentally observed M4+-OH hydroxyl bands shift of supports (Table 3). Therefore, the Brönsted acidity of the OH groups is the major factor influencing the catalytic activity. It should also be underlined that, since zirconium is responsible for introducing 32 [9] Werpy, T.; Peterson, G. Top Value Added Chemicals from Biomass: Volume I - Results of Screening for Potential Candidates from Sugars and Synthesis Gas. United States: N. p., 2004. Web. doi:10.2172/15008859. [10] Bozell, J.J.; Petersen, G.R. Cutting-edge research for a greener sustainable future Technology development for the production of biobased products from biorefinery carbohydrates — the US Department of Energy’s “Top 10” revisited, Gre. 2010, 12, 539–554. [11] Gandini, A.; Silvestre, A.J.D.; Neto, C.P.; Sousa, A.F.; Gomes, M. The furan counterpart of poly(ethylene terephthalate): An alternative material based on renewable resources, J. Polym. Sci. A Polym. Chem. 2008, 47, 295–298. [12] Davis, S.E.; Houk, L.R.; Tamargo, E.C.; Datye, A.K.; Davis, R.J. Oxidation of 5- hydroxymethylfurfural over supported Pt , Pd and Au catalysts, Catal. Today. 2011, 160, 55–60. [13] Sahu, R.; Dhepe, P.L. Synthesis of 2,5-furandicarboxylic acid by the aerobic oxidation of 5-hydroxymethyl furfural over supported metal catalysts, Reac. Kinet. Mech. Cat. 2014, 112, 173–187. [14] Siyo, B.; Schneider, M.; Radnika, J.; Pohl, M.-M.; Langer, P.; Steinfeldt, N. Influence of support on the aerobic oxidation of HMF into FDCA over preformed Pd nanoparticle based materials, Appl. Catal. A Gen. 2014, 478, 107–116. [15] Ait Rass, H.; Essayem, N.; Besson, M. Selective Aerobic Oxidation of 5-HMF into 2 , 5- Furandicarboxylic Acid with Pt Catalysts Supported on TiO2- and ZrO2-Based supports, ChemSusChem. 2015, 8, 1206–1217. [16] Albonetti, S.; Lolli, A.; Morandi, V.; Migliori, A.; Lucarelli, C.; Cavani, F. Conversion of 5-hydroxymethylfurfural to 2 , 5-furandicarboxylic acid over Aubased catalysts : Optimization of active phase and metal – support interaction, 33 Appl. Catal. B, Environ. 2015, 163, 520–530. [17] Lolli, A.; Albonetti, S.; Utili, L.; Amadori, R.; Ospitali, F.; Lucarelli, C.; Cavani, F. Insights into the reaction mechanism for 5-hydroxymethylfurfural oxidation to FDCA on bimetallic Pd – Au nanoparticles, Appl. Catal. A, Gen. 2015, 504, 408– 419. [18] Zope, B.N.; Hibbitts, D.D.; Neurock, M.; Davis, R.J. Reactivity of the Gold / Water Interface During Selective Oxidation Catalysis, Science. 2010, 533, 74–79. [19] Pasini, T.; Piccinini, M.; Blosi, M.; Bonelli, R.; Albonetti, S.; Dimitratos, N.; Lopez-Sanchez, J. A.; Sankar, M.; He, Q.; Kiely, C. J.; Hutchings, G. J.; Cavani, F. Selective oxidation of 5-hydroxymethyl-2-furfural using supported, Green Chem. 2011, 13, 2091–2099. [20] Reina, T.R.; Alvarez, A.; Ivanova, S.; Odriozola, J.A.; Centeno, M.A. Influence of Vanadium or Cobalt Oxides on the CO Oxidation Behavior of Au/MOx/CeO2–Al2O3 Systems, ChemCatChem. 2012, 4, 512 – 520. [21] Reina, T.R.; Ivanova, S.; Domnguez, M.I.; Centeno, M.A.; Odriozola, J.A. Subambient CO oxidation over Au/MOx/CeO2-Al2O3 (M = Zn or Fe), Appl. Catal. A Gen. 2012, 58–66, 419–420. [22] Reina, T.R.; Ivanova, S.; Idakiev, V.; Delgado, J.J.; Ivanov, I.; Tabakova, T.; Centeno, M.A.; Odriozola, J.A. Impact of Ce–Fe synergism on the catalytic behaviour of Au/CeO2–FeOx/Al2O3 for pure H2 production, Catal. Sci. Technol. 2013, 3, 779–787. [23] Megías-Sayago, C.; Ivanova, S.; López-Cartes, C.; Centeno, M.A.; Odriozola, J.A. Gold catalysts screening in base-free aerobic oxidation of glucose to gluconic acid, Catal. Today. 2017, 279, 148–154. [24] Abad, A.; Concepción, P.; Corma, A.; García, H. A Collaborative Effect between 34 Gold and a Support Induces the Selective Oxidation of Alcohols, Angew. Chem., Int. Ed. 2005, 44, 4066–4069. [25] Letichevsky, S.; Tellez, C.A.; De Avillez, R.R.; Isabel, M.; Silva, P.; Fraga, M.A.; Appel, L.G. Obtaining CeO 2 – ZrO 2 mixed oxides by coprecipitation : role of preparation conditions, Appl. Catal. B Environ. 2005, 58, 203–210. [26] Ivanova, S.; Petit, C.; Pitchon, V. A new preparation method for the formation of gold nanoparticles on an oxide support, Appl. Catal. A Gen. 2004, 267, 191–201. [27] Thammachart, M.; Meeyoo, V.; Risksomboon, T.; Osuwan, S. Catalytic activity of CeO2–ZrO2 mixed oxide catalysts prepared via sol–gel technique: CO oxidation, Catal. Today. 2001, 68, 53–61. [28] Gutiérrez-Ortiz, J.I.; De Rivas, B.; López-fonseca, R.; González-Velasco, J.R. Combustion of aliphatic C2 chlorohydrocarbons over ceria – zirconia mixed oxides catalysts, Appl. Catal. A Gen. 2004, 269, 147–155. [29] Ivanova, S.; Pitchon, V.; Petit, C. Application of the direct exchange method in the preparation of gold catalysts supported on different oxide materials, J. Mol. Catal. A Chem. 2006, 256, 278–283. [30] Fornasiero, P.; Balducci, G.; Di Monte, R.; Kaspar, J.; Sergo, V.; Gubitosa, G.; Ferrero, A.; Graziani, M. Modification of the Redox Behaviour of CeO2 Induced by Structural Doping with ZrO2, J. Catal. 1996, 164, 173–183. [31] Pushkarev, V. V.; Kovalchuk, V.I.; D’Itri, J.L. Probing Defect Sites on the CeO2 Surface with Dioxygen, J. Phys. Chem. B. 2004, 108, 5341–5348. [32] Tsyganenko, A.A.; Filimonov, V.N. Infrared spectra of surface hydroxyl groups and crystalline structure of oxides, J. Mol. Struct. 1973, 19, 579–589. [33] Kouva, S.; Honkala, K.; Lefferts, L.; Kanervo, J. Review: monoclinic zirconia, its surface sites and their interaction with carbon monoxide, Catal. Sci. Technol. 35 2015, 5, 3473–3490. [34] Morterra, C.; Cerrato, G.; Di Ciero, S. IR study of the low temperature adsorption of CO on tetragonal zirconia and sulfated tetragonal zirconia, Appl. Surf. Sci. 1998, 126, 107–128. [35] Hadjiivanov, K.; Vayssilov, G. Characterization of oxide surfaces and zeolites by carbon monoxide as an IR probe molecule, Adv. Catal. 2002, 47, 307–511. [36] Penkova, A.; Laguna, O.H.; Centeno, M.A.; Odriozola, J.A. CO-Induced Morphology Changes in Zn-Modified Ceria: A FTIR Spectroscopic Study, J. Phys. Chem. C. 2012, 116, 5747−5756. [37] Cerrato, G.; Bordiga, S.; Barbera, S.; Morterra, C. A surface study of monoclinic zirconia (m-ZrO2), Surf. Sci. 1997, 377–379, 50–55. [38] Zaki, M.I.; Hasan, M.A.; Al-Sagheer, F.A.; Pasupulety, L. In situ FTIR spectra of pyridine adsorbed on SiO2–Al2O3, TiO2, ZrO2 and CeO2: general considerations for the identification of acid sites on surfaces of finely divided metal oxides, Colloids Surf., A. 2001, 190, 261−274. [39] Azambre, B.; Idriss, I.; Bueno-Lpez, A.; García-García, A. Probing the Surface of Ceria−Zirconia Catalysts Using NOx Adsorption/Desorption: A First Step Toward the Investigation of Crystallite Heterogeneity, J. Phys. Chem. C. 2010, 114, 13300−13312. [40] Bensalem, A.; Bozon-Verduraz, F.; Delamar, M.; Bugli, G. Preparation and characterization of highly dispersed silica-supported ceria, Appl. Catal., A. 1995, 121, 81−93. [41] Romero-Sarria, F.; Penkova, A.; Martinez T., L.M.; Centeno, M.A.; Hadjiivanov, K.; Odriozola, J. A. Role of water in the CO oxidation reaction on Au/CeO2: Modification of the surface properties, Appl. Catal. B Environ. 2008, 84, 36 119‒124. [42] Penkova, A.; Chakarova, K.; Laguna, O.H.; Hadjiivanov, K.; Romero-Saria, F.; Centeno, M.A.; Odriozola, J.A. Redox chemistry of gold in a Au/FeOx/CeO2 CO oxidation catalyst, Catal. Commun. 2009, 10, 1196–1202. [43] Bolis, V.; Morterra, C.; Volante, M.; Orio, L.; Fubini, B.; Development and suppression of surface acidity on monoclinic zirconia: a spectroscopic and calorimetric investigation, Langmuir 1990, 6, 695–701. [44] Davis, S.E.; Zope, B.N.; Davis, R.J. Green Chemistry On the mechanism of selective oxidation of 5-hydroxymethylfurfural to 2 , 5-furandicarboxylic acid over supported Pt and Au catalysts, Green Chem. 2012, 14, 143–147. [45] Casanova, O.; Iborra, S.; Corma, A.; Biomass into Chemicals : Aerobic Oxidation of 5-Hydroxy-methyl-2-furfural into 2,5-Furandicarboxylic Acid with Gold Nanoparticle Catalysts, ChemSusChem. 2009, 2, 1138–1144. [46] Donoeva, B.; Masoud, N.; De Jongh, P.E.; Carbon Support Surface Effects in the Gold-Catalyzed Oxidation of 5‑Hydroxymethylfurfural, ACS Catal. 2017, 7, 4581−4591. [47] Song, H.; Bao, X.; Hadad, C.M.; Ozkan, U.S. Adsorption/Desorption Behavior of Ethanol Steam Reforming Reactants and Intermediates over Supported Cobalt Catalysts, Catal. Letters. 2011, 141, 43–54.