Winning Combination of Cu and Fe Oxide Clusters with an Alumina Support for Low-Temperature Catalytic Oxidation of Volatile Organic Compounds

A γ-alumina support functionalized with transition metals is one of the most widely used industrial catalysts for the total oxidation of volatile organic compounds (VOCs) as air pollutants at higher temperatures (280–450 °C). By rational design of a bimetal CuFe-γ-alumina catalyst, synthesized from a dawsonite alumina precursor, the activity in total oxidation of toluene as a model VOC at a lower temperature (200–380 °C) is achieved. A fundamental understanding of the catalyst and the reaction mechanism is elucidated by advanced microscopic and spectroscopic characterizations as well as by temperature-programmed surface techniques. The nature of the metal–support bonding and the optimal abundance between Cu–O–Al and Fe–O–Al species in the catalysts leads to synergistic catalytic activity promoted by small amounts of iron (Fe/Al = 0.005). The change in the metal oxide–cluster alumina interface is related to the nature of the surfaces to which the Cu atoms attach. In the most active catalyst, the CuO6 octahedra are attached to 4 Al atoms, while in the less active catalyst, they are attached to only 3 Al atoms. The oxidation of toluene occurs via the Langmuir–Hinshelwood mechanism. The presented material introduces a prospective family of low-cost and scalable oxidation catalysts with superior efficiency at lower temperatures.


XRD of samples calcined at 1000 o C
X-ray diffraction analysis of the samples showed a mixture of δ (PDF 00-046-1131) and θ (PDF 01-086-1410) Al2O3 phase. No copper or iron oxides were detected with XRD.

XRF analysis
Samples were dried and calcined prior to the analysis to determine the total weight loss of the material during the following sample preparation steps with Phoenix R sample preparation unit and lithium tetraborate flux (type LT100 Granular) addition in ratio 1:8 (sample: flux). Results were analyzed with Spectra plus Measurement and Launcher V 2.2.47 software and two different calculation methods were implemented to avoid any miscalculations. The loss of ignition was identified as water loss and accounted in the total mass of the sample for correct data analysis. Figure S5a presents the difference in strength of acid sites with increased calcination temperature (500 and 1000°C) in one iron free and one iron containing sample. Catalyst with Fe/Al molar ratio of 0.005 was chosen as it was the most active during the catalytic tests (vide S5 infra). In both cases, the first desorption peak of pyridine is observed at around 240°C and this peak is larger in the samples calcined at 500°C, revealing presence of a larger number of relatively weak acid sites. Low concentration of iron does not influence the strength of acid sites in most cases. Catalysts prepared at 1000°C contained stronger acid sites, based on more pronounced pyridine desorption at temperatures above 350°C, although their total concentration was reduced by about 50% in respect to those prepared at 500°C, which is likely related to specific surface area loss, which is also about 50% (Table 1).

Determination of surface acidic properties (pyridine-TPD)
The pyridine desorption profiles are compared for different Fe loadings on alumina calcined at 500°C with and without copper ( Figure S5b). Presence of only copper decreases the total acid site abundance, which is consistent with lower intrinsic acidity of CuO compared to Al2O3 (REFERENCE https://doi.org/10.1002/anie.200803837). However, when iron amount is gradually increased in the bimetallic Cu-Fe materials containing a constant Cu loading, we could observe a progressive trend in increasing weak (pyridine desorption peak at about 240°C) and strong acid.  Figure S6a shows type IV isotherms of the Fe-doped γ-alumina samples with similar hysteresis (H3) shape, indicating the presence of aggregates of plate-like particles with voids consisting of mesopores and macropores. On the other hand, two types of nitrogen isotherm hysteresis can be observed for samples containing iron and copper ( Figure S6b), e.g. H3 and H2a. The latter can be found for samples with higher iron amount (Fe/Al molar ratio 0.01 and 0.05). These isotherms are less intense and imply on presence of mainly mesopores. An apparent S7 macroporous structure (type II isotherm) transforms to a more mesoporous one (type IV isotherm) (REFERENCE https://doi.org/10.1515/pac-2014-1117 ). Figure S7 shows the pore size distribution (PSD) of samples with different Fe/Al ratios ( Figure   S7a), and their evolution during preparation of copper containing catalyst of the same origin ( Figure S7b). An evident broad pore size distribution in 0.01 and 0.05 Fe/Al containing samples changed to a structure with more uniform mesopores after copper loading. Copper deposition led to a decrease of larger mesopores and macropores (20-60 nm), for example at higher iron amount only mesopores below 16 nm can be observed. Pore size distribution is also significantly altered after copper introduction in sample D-500, where broad multimodal PSD with maxima at 30 and 44 nm changed to bimodal PSD with less intense maxima. Interestingly, the PSD of sample D-0.005Fe-500-Cu is broader in comparison to other presented materials with Cu and it remains preserved if compared to the material before copper impregnation.

XAS
The catalyst samples were prepared in the form of homogeneous pellets, pressed from micronized sample powder, with the total absorption thickness (μd) of about 2.5 above the investigated Cu or Fe K-edge. Si (111) double crystal monochromator with about 1 eV resolution at 9 keV was used. The intensity of the monochromatic X-ray beam was measured by 30 cm long ionization chambers detectors filled with a N2, Ar, He gas mixture at 2 bars, such to have 20% of absorption for the I0 detector, 80% of absorption for the I1 detector and 95% of absorption for the I2 detector. Sample pellets were placed in the beam between first two ionization chambers detectors. For fluorescence detection, the samples were rotated 45 o with respect to the X-ray beam. Fluorescence signal was detected with SDD solid state fluorescence detector. The absorption spectra were measured in the energy region from -150 eV to +1200 eV relative to the Fe or Cu K-edge. In the XANES region equidistant energy steps of 0.3 eV were used, while for the EXAFS region equidistant k steps of 0.03 Å -1 were adopted, with an integration time of 2 s/step in transmission mode and 5s/ step in fluorescence mode. Three repetitions of the scans were superimposed to improve the signal-to-noise ratio. The exact energy calibration at Fe and Cu K-edge was established with simultaneous absorption measurement on a 5 µm thick Fe or Cu metal foil, respectively, placed between the second and the third ionization chamber. The first inflection point of the spectrum of the Fe foil at Fe Kedge was assigned at 7112 eV and that of the Cu foil at Cu K-edge was assigned at 8979.0 eV.
Absolute energy reproducibility of the measured spectra was ±0.01 eV.  Table S1. Parameters of the nearest coordination shells around Cu in Fe and Cu functionalized alumina samples: average number of neighbor atoms (N), distance (R), and Debye-Waller factor (σ 2 ). Uncertainty of the last digit is given in parentheses. A best fit is obtained with the S12 amplitude reduction factor S0 2 =0.80 and the shift of the energy origin ΔEo=-1(1) eV. The Rfactor (quality of fit parameter) is listed in the last column.

CHN elemental analysis
CHN elemental analysis was performed on spent samples D-500-Cu, D-1000-Cu and D-0.005Fe-1000-Cu to determine the potential coke formation and differences in catalytic activity among those samples. Table S3 shows that the spent samples D-500-Cu and D-0.005Fe-500-Cu contain 0.6 and 0.5 wt.% carbon, respectively. This indicates the formation of some organic species as side products, which remain adsorbed on the catalyst surface. Usually, coke formation leads to reduced activity after certain TOS. A significantly higher carbon deposition was observed on sample D-1000-Cu in comparison to other two samples, prepared at 500°C.  cupper-iron-alumina samples (D-0.005Fe-500-Cu, D-0.01Fe-500-Cu, D-0.05Fe-500-Cu).
Interesting observation was made when comparing the iron-alumina samples without cupper with iron-alumina samples with cupper. In all cases, except for the material D-0.005Fe-500-Cu, the band gap energy was reduced after introduction of Cu to the alumina or iron--alumina support. This reduction was intensified with increased Fe/Al ratio. In case of the sample with Fe/Al ration of 0.005, this was although not the case, and the band gap energy was increased.
The results show that the part of iron is embedded in the Al2O3 crystal lattice at Al sites and thus creates additional electronic states in the Al2O3 crystal structure within the energy gap below the conduction band. These electronic states extend over the entire crystal surface and act on Cu, which is bound to the Al2O3 surface. Cu binds to the surface after Fe is already incorporated into Al2O3.

Figure S12. Comparison of bandgap energies supporting electronic interactions between
CuOx clusters and iron-alumina-s