Design Aspects of Doped CeO2 for Low-Temperature Catalytic CO Oxidation: Transient Kinetics and DFT Approach

CO elimination through oxidation over highly active and cost-effective catalysts is a way forward for many processes of industrial and environmental importance. In this study, doped CeO2 with transition metals (TM = Cu, Co, Mn, Fe, Ni, Zr, and Zn) at a level of 20 at. % was tested for CO oxidation. The oxides were prepared using microwave-assisted sol–gel synthesis to improve catalyst’s performance for the reaction of interest. The effect of heteroatoms on the physicochemical properties (structure, morphology, porosity, and reducibility) of the binary oxides M–Ce–O was meticulously investigated and correlated to their CO oxidation activity. It was found that the catalytic activity (per gram basis or TOF, s–1) follows the order Cu–Ce–O > Ce–Co–O > Ni–Ce–O > Mn–Ce–O > Fe–Ce–O > Ce–Zn–O > CeO2. Participation of mobile lattice oxygen species in the CO/O2 reaction does occur, the extent of which is heteroatom-dependent. For that, state-of-the-art transient isotopic 18O-labeled experiments involving 16O/18O exchange followed by step-gas CO/Ar or CO/O2/Ar switches were used to quantify the contribution of lattice oxygen to the reaction. SSITKA-DRIFTS studies probed the formation of carbonates while validating the Mars–van Krevelen (MvK) mechanism. Scanning transmission electron microscopy-high-angle annular dark field imaging coupled with energy-dispersive spectroscopy proved that the elemental composition of dopants in the individual nanoparticle of ceria is less than their composition at a larger scale, allowing the assessment of the doping efficacy. Despite the similar structural features of the catalysts, a clear difference in the Olattice mobility was also found as well as its participation (as expressed with the α descriptor) in the reaction, following the order αCu > αCo> αMn > αZn. Kinetic studies showed that it is rather the pre-exponential (entropic) factor and not the lowering of activation energy that justifies the order of activity of the solids. DFT calculations showed that the adsorption of CO on the Cu-doped CeO2 surface is more favorable (−16.63 eV), followed by Co, Mn, Zn (−14.46, −4.90, and −4.24 eV, respectively), and pure CeO2 (−0.63 eV). Also, copper compensates almost three times more charge (0.37e−) compared to Co and Mn, ca. 0.13e− and 0.10e−, respectively, corroborating for its tendency to be reduced. Surface analysis (X-ray photoelectron spectroscopy), apart from the oxidation state of the elements, revealed a heteroatom–ceria surface interaction (Oa species) of different extents and of different populations of Oa species.

High-resolution 3Flex Micromeritics (Atlanta, USA) adsorption apparatus with high-vacuum system and three 0.1 Torr pressure transducers was employed for textural measurements through the N 2 physisorption method at 77 K. A degassing process was carried out prior to the measurements to remove any residual moisture (degassing at 150 o C for 12 h) and weakly bound carbon dioxide. The BET surface area was calculated from the adsorption isotherm data in the relative pressure range of P/P 0 =0.04-0.25. Information regarding the pore size distribution and specific pore volume (cm 3   The materials were mounted on a U-shaped quartz sample tube and placed in a flow-through reactor. A gas flow of 5 vol% CO 2 /Ar (30 NmL/min) was passed over 0.2 g of the pre-calcined (20 vol%O 2 /He, 500 o C, 2 h) sample for 30 min. The temperature of the sample was then increased at the heating rate of 30 o C/min to 500 o C and kept there for 2 h, while the TCD signal was recorded continuously.
Information regarding surface elemental composition and chemical states were obtained using X-ray Photoelectron Spectroscopy (XPS, Thermo-Fisher Scientific East Grinstead, UK, Theta Probe spectrometer). A monochromatic Al Kα X-ray source (hν = 1486.6 eV) and an X-ray spot of radius ~ 400 μm were employed to record XPS spectra for the calcined materials. Survey and high-resolution spectra were acquired using pass energies of 300 eV and 50 eV, respectively. C 1s peak (285.0 eV) was taken into consideration in all spectra to account for charging effects during acquisition, and correct binding energies, accordingly. The non-linear (Shirley) background was eliminated from all spectra for the purpose of obtaining high resolution core level spectra for quantitative surface chemical analyses. The experiment was coupled with the manufacturer's Avantage software that acquires the appropriate sensitivity correction factors for the electron energy analyzer transmission function.
HR-TEM studies were performed over the solids with the use of an electron microscope (model Titan 80-300 ST, operated at 300 kV). The microscope was equipped with a spherical aberration (Cs) corrector for the image (CEOS CETCOR), and an energy filter (model GIF Quantum 963, Gatan, Inc.). Selected powdered specimens were dispersed in ethanol and drop casted onto carboncoated copper grid (200 mesh) to perform imaging, structure, composition, and elemental mapping analyses. The high-resolution imaging was performed in aberration corrected TEM mode, while the structure analysis was carried out using the selected area electron diffraction (SAED) technique of TEM. The composition of samples was determined using X-ray energy dispersive spectroscopy (EDX). However, nanoscale elemental quantification was carried out using scanning TEM (STEM) mode. For the elemental quantification, STEM-EDX experiments were performed by setting the same electron beam conditions for the specimens of each sample. The entire TEM data acquisition and post processing was done by utilizing Gatan Microscopy Suite (GMS, version 3.2).

S1.3. Catalytic CO oxidation performance studies
The catalytic oxidation of carbon monoxide over the various TM-doped CeO 2 materials was carried out in a differential packed-bed quartz U-tube flow reactor (i.d. = 9 mm) mounted in the middle of a programmable split tube furnace (Lindberg/Blue M Mini-Mite Tube Furnace -Thermo). A catalyst sample of 50 mg was packed between two thin layers of glass wool.
Thermocouples (K-type) were located at the inlet and outlet of the catalyst bed for gas-phase temperature monitoring, very close to that at the external surface of catalyst grains (powder form).
The signal from the thermocouple was acquired using USB-6008 Multifunction I/O and NI-DAQmx (National Instruments) data acquisition system. The composition of the exit gas stream from the micro-reactor was analyzed using an infrared gas analyzer (IR200, Yokogawa). The feed gas mixture consisted of 4 vol% CO, 20 vol% O 2 and He as balance gas. Unless otherwise mentioned, regulated gas mixture total flow -(use of digital thermal mass flow controllers (HI-TEC Model F-201CV-10K-AGD-22-V Multi-Bus DMFC; Bronkhorst) was set at 50 mL/min (STP) corresponding to a space velocity (SV) of  60,000 cm 3 g cat -1 h -1 . After stabilization, the temperature of the catalyst bed was raised at 5 o C min -1 in the flow of CO/O 2 /He gas mixture.
Labview-configured data acquisition software was used to record and store the vol% composition in CO, CO 2 and O 2 alongside the temperature throughout the CO activity vs T procedure. Further catalytic stability experiments were conducted on selected catalysts under continuous feed gas stream, ca. 20 h at 530 o C.

S1.4. DFT Calculations
The all-electron wave functions and the pseudo-potentials for the electron-ion interactions are described within the projector-augmented wave (PAW) 2  Usually, adsorption hysteresis is related to specific shaped pores, and H3-type indicates that the synthesized catalysts acquire a slit-shaped pores 7 . The low N 2 uptake at a low relative pressure is observed in all the mixed metal-oxides due to the lack of micropores (pore width < 2 nm). The surface area and cumulative specific pore volume were obtained through the BET and Barret-Joyner-Halenda (BJH) methods 8 , respectively. Variations in the BET area as well as pore volume are observed due to dopant-type effect (

S2.1.3. Elastic energy upon doping of ceria:
The elastic energy is given by the following formula: where G is the shear Modulus, and d o is the lattice parameter for the host oxide (CeO 2 ). G value of 92.5 GPa was used for CeO 2 21 . The elastic energy values for the Cu, Co, Ni, Zn, Fe, Mn were found to be 8.4, 5.2, 4.9, 6.5, 7.1 and 5.2 (x10 -9 ) N-Å.

S2.2. Oxidation states of TM-dopants in doped-CeO 2 catalytic materials
The oxidation states of transition metals on the surface of TM-doped CeO 2 were identified based on the core level binding energy signal and the presence or absence of shake-up satellites in the XPS spectra. The Cu 2p 3/2 peak is observed at  933.4 eV, and there is a broad shake-up satellite at a binding energy of 941-944 eV (Fig. S6E). The satellite peak is observed only for Cu 2+  The Co 2p 3/2 peak (Fig. S6F), occurs at a binding energy of 780.2 eV and there is a satellite peak at ~785 eV. These peak positions and the spectral shape recorded are in very good agreement with that of CoO reference material than the mixed oxidation states of Co 3 O 4 28 , demonstrating the predominance of Co 2+ oxidation state in the present mixed metal oxide Co-Ce-O solid composition.
This finding also indicates that Co 2+ species are more populated than the other possible oxidation states, Co 0 or Co 3+ 5 . However, an overlap between CoO, Co 3 O 4 , Co 2 O 3 and CoOOH exists at the same binding energy. Interestingly, it has been reported that Cu + and Co 2+ species are preferred over Cu 2+ and Co 3+ , respectively, during CO oxidation reaction 29,30 . This finding was correlated with the CO conversion performance of Cu-Ce-O and Co-Ce-O catalysts.
The Ce3d region (Fig. S6A) is a typical Ce 3d spectrum, as it is discussed in the main manuscript, where the 4+ valence state is predominant, with little evidence for 3+ cerium oxidation state, indicative of almost a fully oxidized CeO 2 . The Fe 2p region (not shown) is rather complicated by the presence of two Ce Auger peaks (M 4 N 4,5 N 6,7 at ~716 eV and M 5 N 4,5 N 6,7 at ~732 eV). The Fe 2p 3/2 and Fe 2p 1/2 peaks occur at the binding energies of 710.9 and 724.2 eV.
These binding energies agree very well with Fe in the Fe 3+ oxidation state 31 .
The Mn 2p spectra show the Mn 2p 3/2 peak at a binding energy of 641.7 eV, and a broad shakeup satellite peak at ~ 645 eV (Fig. S6D). The peak/satellite binding energies and shape are in best agreement with MnO (in comparison to other manganese oxides) 32 . Hence, Mn 2+ is the predominant oxidation state for this dopant in the Mn-doped CeO 2 structure.
In Fig. S6C, the Ni 2p 3/2 peak occurs at a binding energy of 855.0 eV. This binding energy is typical of Ni 2+ but the observed peak shape does not agree very well with either NiO or Ni(OH) 2 and the binding energy lies mid-way between that of NiO (854.7 eV) and Ni(OH) 2 (855.3 eV) 33 .
Thus, in the Ni-Ce-O mixed metal oxide, the Ni is predominantly present as Ni 2+ , but the XPS Ni 2p peak shape observed does not clearly correspond with that of NiO due to the presence of Ni(OH) 2 at the surface which can be due to the atmosphere exposure of the Ni containing catalyst.
Finally, Zn exists only in its 2+ state when oxidized. The Zn 2p 3/2 peak (Fig. S6B) shows a sharp peak at a binding energy of 1021.9 eV, which is in excellent agreement with ZnO recorded by Klein and Hercules 34 .    Fig. 4C and Fig. S8A, B suggest that the variations in the reduction behavior of the binary metal oxide systems are due to dopant-type as well as to the different dopant cation oxidation state (i.e. divalent, trivalent, and quadrivalent). According to the literature, pure ceria experiences two reduction peaks both appearing at temperatures higher than 350 °C 35,36 . In pure ceria, an initial reduction peak at  485 °C is attributed to surface shell reduction, and a high temperature reduction peak at  765 °C is attributed to bulk material reduction 37,38 . In microwave-assisted prepared ceria, the reduction peaks are observed at  247 o C (sub-surface), 467 o C and 1000 °C (bulk) 39 . The preparation route has been known to affect the reduction properties of ceria-based catalysts 40,41 .
Divalent cations: Fig. 4C shows the effect of divalent cations, Co 2+ , Cu 2+ , Ni 2+ and Zn 2+ on the doped-ceria reducibility. The Cu-Ce-O reduction profile presents an intense peak at very low temperatures, at  200 °C, with a shoulder peak at around 100 °C that extends up to 150 °C. These peaks correspond to the stepwise reduction of CuO, for example, Cu 2+ to Cu + , and Cu + to Cu 042 .
The broader narrow peak above 400 °C is attributed to the reduction of the remaining Ce 4+ 42 . CuO has been reported to have one reduction peak at  300°C 43 , where Cu-Ce-O Ce 0.8 Cu 0.2 O 2-δ is reduced at lower temperatures than both single oxides of CeO 2 and CuO, demonstrating the successful synergy achieved by introducing Cu 2+ into the CeO 2 lattice. From the synthesis perspective, it can be stated that microwave synthesis enhanced the homogeneous dispersion of Cu in the ceria lattice, since the low reduction temperature peak  100°C is absent for the same composition prepared via the co-precipitation method due to agglomeration. One main reduction peak also appears at ~ 250 °C 44 . Another study on 10 wt% Cu-doped ceria 35 reported the first (lowest temperature) reduction peak to occur at 245, 230 and 190 °C, respectively, for the same solid composition prepared via three different routes, namely, deposition-precipitation, coprecipitation, and complexation-combustion. The reduction of ceria surface is promoted with copper addition 35   On the contrary, Zn-Ce-O reduction profile (Fig. S9) is appeared rather suppressed (in terms of intensity), with Zn 2+ species being relative inactive, even though they are highly populated on the surface (as confirmed by XPS). Therefore, it can be concluded that the hydrogen reduction peaks are mainly due to the reduction of Ce 4+ species, where ZnO presence confirmed by XRD, contributes to the enhancement of the synergy with the mixed metal oxide which shifts the reduction peaks towards lower temperatures compared to those of pure ceria, in agreement with the literature 45,46 . It was reported that ZnO is not easily reducible, though Zn 2+ reduction can be facilitated when it is adjacent to a metal cation by a synergistic effect between the mixed metal cations 46 .
Co-Ce-O TPR profile demonstrates the stepwise reduction processes involved due to the highly dispersed cobalt species in ceria (Fig. 4C & Fig. S9). The addition of Co greatly boosts the surface reduction of Ce 4+ as indicated by the peak at 300 o C, whereas the formation of Ce 0.8 Co 0.2 O 2-δ solid solution implies that Co 3+ /Co 2+ /Co and Ce 4+ /Ce 3+ reduction processes can also take place. Namely, (a) reduction of surface oxygen and Ce 4+ , (b) Co 3+ to Co 2+ , (c) Co 2+ to Co and (d) reduction of surface ceria species (~ 545°C). This observation is in agreement with previous reported literature on cobalt-doped ceria 37 .  can be explained by Ni 2+ having lower cation charge than Ce 4+ which causes lattice distortion 47,48 .
The high temperature peak observed at 380-500 °C corresponds to the simultaneous reduction of NiO to Ni 0 and remaining Ce 4+ . Microwave-assisted synthesized Ni-Ce-O presented more mobile oxygen species than other similar in composition NiO-CeO 2 catalysts prepared by co-precipitation or sol-gel methods. In fact, the conventional methods reported reduction peaks for NiO-CeO 2 similar to that of pure NiO 49,50 . Thus, it can be concluded that Microwave synthesis strengthens the interaction between NiO and CeO 2 that allows reduction to commence at lower temperatures. In the case of Fe-Ce-O, an overlap between the reduction peaks of pure ceria as well as those of Fe 2 O 3 is expected. The H 2 -TPR trace (Fig. S8A) shows the first low temperature peak at 265 °C.
This peak is attributed to the reduction of surface Fe 3+ and Ce 4+ species. This behavior is reported also on previous studies on Fe-doped ceria 52 . The peak is shifted towards lower temperatures due to the synergy between the cation and pure ceria through solid solution formation (confirmed by XRD and Raman), which facilitates the reducibility of the mixed oxide as explained above. The high temperature peak is observed in the 450-600 °C range, which corresponds to reduction of Ce 4+ as well as of Fe 3 O 4 to FeO. The H 2 -TPR trace obtained for the microwave-assisted synthesized Fe-Ce-O shows resemblance to other studies on Fe-doped ceria 46 . However, a shift in the low-temperature peak is favored in microwave synthesized Fe-doped ceria compared to other traditional methods. The first reduction peak is observed at 384 °C for a sol-gel prepared 10 mol% Fe (~ 20 wt%) in ceria 46 , whereas microwave-assisted prepared Ce 0.8 Fe 0.2 O 2-δ provides the same peak at 265 °C. This behavior can be explained by taking into account that microwave synthesis enhances the dispersion of dopant cation on the surface, or/and promotes the formation of oxygen vacancies (confirmed by Raman) by allowing Fe 3+ to substitute for Ce 4+ , which introduces charge compensation.
Quadrivalent Cations: Zr-Ce-O TPR profile features two main asymmetric reduction peaks in the 100-400 o C range and at ~ 580 °C (Fig. S8B). The peak asymmetry indicates the overlap between surface and bulk reduction of Ce 4+ species. The reduction peaks of microwave prepared Ce 0.8 Zr 0.2 O 2-δ are shifted towards lower temperatures compared to those prepared via the coprecipitation method, according to Lan Li et al. 53 , who reported asymmetric TPR consumption peak at ~ 650°C. Zr-Ce-O TPR profile maintains the feature of CeO 2 with slight variation in peaks intensity and position. The first peak corresponds to reduction of the accessible Ce 4+ species, whereas the second peak corresponds to oxygen removal from the bulk 54 .
The H 2 -TPR profile of Mn-doped CeO 2 (Fig. S8B) shows reduction peaks at ~ 100, 180, 390 and 500 o C. The first peak could be referred to the reduction of highly dispersed MnO 2 /Mn 2 O 3 to Mn 3 O 4 , and the second peak can be assigned to the reduction of Mn 3 O 4 to MnO and surface Ce 4+ species. The third peak is associated with the reduction of remaining ceria and manganese oxides 55 .
The observed peaks are shifted towards lower temperatures compared to CeO 2 , which implies a synergistic interaction between ceria and manganese through the solid solution formation (confirmed by XRD). The latter promotes surface reduction of Mn-Ce-O. This result is in agreement with previous findings on Mn-doped ceria [55][56][57] . Moreover, the reduction peaks are shifted towards lower temperatures due to the catalyst preparation method effect (microwaveassisted synthesis). The first and second reduction peaks for 30 wt% Mn in ceria, prepared by the co-precipitation method, were found to be located at 315 and 410 °C 58 . This observation emphasizes the role of microwave-assisted synthesis in enhancing the reduction properties of ceria-based mixed metal oxides system.

S2.4. CO 2 -Temperature Programmed Desorption (CO 2 -TPD) on TM-doped CeO 2
CO 2 -TPD traces were obtained over the mixed metal oxides in order to study the surface-CO 2 interaction characterizing their surface basicity (Fig. S8C). As CO 2 is an acid molecule, it is expected to interact with basic surface sites 59 .
Divalent ions: Fig. S8C shows that all the mixed metal oxides demonstrate an initial CO 2 desorption peak at  100°C with a higher intensity than that of pure ceria. This observation indicates that transition metal doping of ceria strengthens the Lewis-type surface basic cites of ceria at the low-temperature region (50-150°C). The initial desorption peak intensity of all divalent  -O (Fig. S8C) shows the highest surface basicity compared to the other TM-doped ceria solids. CO 2 desorption starts at temperatures slightly below 100°C, which indicates basic centers enrichment of Ni-doped ceria. Similar CO 2 -TPD trace at low temperatures was obtained for the NiO-CeO 2 system prepared by the wet impregnation method 60 . However, with the present microwave-assisted synthesis, the basic centers are extended up to higher temperatures (400-600 °C), indicating a continuous desorption of CO 2 . Previous work conducted on NiO-CeO 2 prepared by molten-salt method showed the same CO 2 -TPD behavior as the one presented here. Mn 2+ -doping showed great modification of the basic centers of pure ceria at high temperatures. CO 2 -TPD trace of Mn-Ce-O (Fig. S8E) shows four desorption peaks with overlapping at 140, 250, 400 and 500 °C, which is in agreement with data reported in the literature 61 .
Trivalent ions: Pure iron (III) oxide is considered basic due to its metal ion electronegativity (El(Fe 3+ )= 12.6) 62 and according to the acid -base classification reported by Tanabe et al.. Doping of ceria has higher basic nature (e.g., El(Ce 4+ )=9.9) 62 and it is expected to increase the basicity of the mixed metal oxide system. According to the literature 62 , Fe 2 O 3 displays complex CO 2 -TPD trace with multiple desorption peaks (Fig. S8D). The most intense desorption peak is observed at low temperatures, ~ 70°C, whereas additional weak desorption peaks are observed at 250, 490 and 650 °C. Fig. S8D shows that doping of ceria with Fe 3+ causes the CO 2 -TPD trace to broaden over a wider temperature range compared to pure CeO 2 (Fig. S8C). It is noticed that the low desorption peak is maintained at similar intensity, however, the intensity of the high-temperature peaks is