Manganese Based Perovskites in Soot Oxidation: Far from Noble Metals?

Mn-based catalysts for soot oxidation have been developed without noble metals. The compositions are LaMn0.9Co0.1O3, La0.9K0.1Mn0.9Co0.1O3 (LKMC), Sr0.9K0.1Mn0.9Co0.1O3 and 0.1 K/La0.9Mn0.9Co0.1O3: Mn provides stability to the structure both in oxidizing and reducing atmospheres and efficiently exchanges oxygen. Moreover, Co doping enhances soot oxidation. Adding La or Sr as A-site cation in the perovskite composition allows comparing the behaviors of the so-obtained perovskites, and K doping was chosen to increase catalytic activity both in soot and NOx removal. After the wet synthesis, the catalysts were tested for soot oxidation in presence of oxygen and nitrogen monoxide in overstoichiometric oxygen content. Temperature Programmed Oxidation tests were performed and double doping increases the oxidative catalytic activity: LKMC shows the lowest soot conversion temperature (306 °C, soot in tight contact with the catalyst). Sr doping results in worse performances, due to the formation of SrCO3. K incorporation helps oxygen vacancies formation, beneficial to the catalytic activity, through the Mars-van Krevelen mechanism.


Introduction
Soot is dangerous for human health and climate change [1]. To prevent soot emission different and complementary approaches are possible: the design of reactors or engines may be optimized, highly efficient filters can be applied and/ or catalysts can be used to lower the ignition temperature of soot, with regards to soot elimination by oxidation.
Commercially the current, and one of the first, solution employed for Diesel engines in automotive sector is the use of the so-called Diesel Particulate Filters (DPFs). They are fabricated in a similar way as Three Way Catalytic converters (TWC): they comprise a monolith (traditionally made of cordierite, or silicon carbide for higher mechanical resistance at high temperature) with a honeycomb structure with high surface areas. Through the mechanism called surface filtration or coke filtration, soot is deposited on the surface of the device and periodically combusted. Technologic progress in this field brought also the use of the so-called Continuous Regeneration Trap (CRT), patented by Johnson Matthey. This device combines a DPF filter with a Pt-based pre-oxidizer which converts CO, hydrocarbons and NOx to CO 2 , H 2 O and NO 2 . NO 2 is then used to oxidize soot to carbon dioxide. Later on a Four Way Catalyst was developed, again in the attempt of combining in a single device the TWC features with soot oxidation activity. The main drawback of such devices is still the massive use of noble metals in the active phase (e.g., Pt, Pd and Rh), and the present work is an endeavor to investigate the possibility of using non-noble metal based perovskites, already known for their TWC activity, in soot oxidation reactions. Other examples of this strategy are to be found in literature [2].
The use of noble metal free catalysts could allow to obtain low cost and sustainable soot abatement devices. Perovskites type oxides are promising catalysts [3] because of their capacity of achieving the simultaneous removal of NO x and diesel soot in the presence of oxygen [4][5][6][7]. Among the perovskites, LaMnO 3 -based ones are characterized by the higher efficiency. To improve the catalyst performances alkali and alkali-metals are exploited [8][9][10][11]: Including potassium in the structure seems to help in this senso [7,12].The mechanisms through which superior performances are obtained are not yet fully understood but a relevant role seems to be played by electron exchange, soot morphology modification and finally decreased stability of surface carbonates [13]. The contribution of all these phenomena can differ for the specific type of oxidic catalysts. In the present contribution we focused on cobalt-doped manganese-based perovskites. In particular, the effect of K on LaMn 0.9 Co 0.1 O 3 and SrMn 0.9 Co 0.1 O 3 is studied; moreover, the insertion of K into the crystalline cell and, on the other side, the deposition on the surface have been compared in terms of catalytic efficiency.
Mul et al. [14] identified two different procedures to test soot oxidation in presence of a catalyst: loose and tight contact. Loose contact represents a situation closer to the real one, as catalyst and soot are simply mixed together with a spatula; tight mode prescribes instead a much more intimate contact between the two species, obtained by crushing them in a mortar.

Synthesis
The samples were prepared by citrate method [15] starting from La 2 O 3 (Sigma-Aldrich, pure), Sr(NO 3 ) 2 (Aldrich ≥ 98%), Manganese(II) acetate tetra hydrate (Sigma-Aldrich ≥ 99%), K 2 CO 3 (Aldrich 95%) and Co(NO 3 ) 2 ·6H 2 O (Sigma-Aldrich 98%). Carbonate and nitrate precursors were dissolved in an aqueous solution while lanthanum solution was obtained by mineralizing the corresponding oxide with nitric acid. A solution of monohydrate citric acid (Sigma Aldrich ≥ 99.0%) was added: its molar ratio was 1.9:1 with respect to the total amount of cations. Then solution is then heated up to 80 °C in air to promote water evaporation and to obtain a pink wet-gel. The gel is slowly heated up to 400 °C for 2 h in air to decompose the organic framework and the obtained powder samples were ground and calcined at 700 °C (LaMn 0.9 Co 0.1 O 3 -LMC, La 0.9 K 0.1 Mn 0.9 Co 0.1 O 3 -LKMC) and 800 °C (Sr 0.9 K 0.1 Mn 0.9 Co 0.1 O 3 -SKMC) for 6 h to obtain the perovskite. The nanocomposite, KO x /LaMn 0.9 Co 0.1 O 3 -K/LMC, was obtained by wet impregnation of the LMC perovskite powder with a solution of K + obtained by dissolving the carbonate in order to reach the same atomic composition of the LKMN catalyst. The suspension was stirred 24 h and then the solvent was slowly removed by heating at 90 °C. The final heat treatment was carried out at 500 °C for 3 h (the desorption of K was observed at temperatures higher than 600 °C).

Characterization
The XPS measurements were carried out with a Perkin Elmer Φ 5600ci Multi Technique System. The spectrometer was calibrated by assuming the binding energy (BE) of the Au 4f 7/2 line to be 84.0 eV with respect to the Fermi level. Both extended spectra (survey-187.85 eV pass energy, 0.8 eV step −1 , 0.05 s step −1 ) and detailed spectra (for Mn 2p, Co 2p, La 3d, Sr 2p, K 1s, O 1s and C 1s-23.50 eV pass energy, 0.1 eV step −1 , 0.1 s step −1 ) were collected with a standard Al Kα source working at 200 W. The standard deviation in the BE values of the XPS line is 0.10 eV. The atomic percentage, after a Shirley-type background subtraction [16], was evaluated by using the PHI sensitivity factors [17]. The peak positions were corrected for the charging effects by considering the C 1s peak at 285.0 eV and evaluating the BE differences [18].
The XRD analyses were performed with a Bruker D8 Advance diffractometer with Bragg-Brentano geometry using a Cu Kα radiation (40 kV, 40 mA, λ = 0.154 nm). The data were collected at 0.03°, with a counting time of 7s/step in the (2θ) range from 20° to 70°. The crystalline phases were identified by the search-match method using the ICDD (International Centre for Diffraction Data) database. Temperature Programmed Reduction (TPR), and specific surface area (BET, Brunauer-Emmett-Teller) measurements were performed with an Autochem II 2920 Micromeritics, equipped with a TCD detector. The H 2 -TPR measurements were carried out in a quartz reactor by using 50 mg of sample and heating from RT to 900 °C at 10 °C min −1 under a constant flow of H 2 5% in Ar (50 ml min −1 ). TPR samples were previously outgassed with He (50 ml min −1 ) at RT. In BET measurements 100 mg of sample were used; before measurement the sample was treated at 350 °C for 2 h under a constant flow of He (50 ml min −1 ); each surface area value obtained was the average of three consecutive measurements [19].
Field emission-scanning electron microscopy and EDX measures were carried on a Zeiss SUPRA 40VP. Morphological analysis and EDX analysis were carried out setting the acceleration voltages at 20 kV.

Catalytic Activity Tests
The catalytic tests are carried out in a quartz reactor (6 mm ID) with a packed bed of powders; the temperature was monitored by a thermocouple right upstream of the bed. The inert carrier was always Ar. The flow rates were controlled by thermal mass flow meters (Vögtlin). The composition of the gas mixture (before and after reaction) was measured by Gas Chromatography (Agilent 7890A), with a TCD detector and 13X (60/80 mesh, 1.8 m) and Porapak Q (1.8 m) columns. Temperature ramp involved a fast heating (10 °C min −1 ) until 150 °C, which should favor CO 2 desorption from the sample, and a slow ramp (1 °C min −1 ) until 450 °C. Injection of the products at the GC was carried out every eight minutes from a continuous stream, allowing us to detect also the possible production of CO in competition to CO 2 from the combustion, an occurrence that, however, was never detected. Soot was added in 1:10 ratio in tight contact mode. If an oxygen-transfer step is involved, then the contact between catalyst and soot and, experimentally, the method used to bring them in contact, is crucial to the catalytic activity. Standard Printex-U carbon black is used to simulate soot. To enhance the catalyst answer and emphasize the difference in the behaviors, tight contact was preferred over loose one. This allows a better evaluation of the effect of the different A-site cation (K vs Sr) and of their insertion into the perovskite cell with respect to the surface deposition.
The OSC test was carried out in the same setup, employing 50 mg of catalyst, and composed of two steps. In the first one, the catalyst was subjected to 1 h of oxidation treatment in synthetic air (80% N 2 /20% O 2 ) at the fixed temperature of 300 °C. The second step, spaced with 2 min He flow, was a CO-TPR reaction at the same fixed temperature, using a 5%CO/He atmosphere. Gas flow was always kept at 100 ml min −1 .

Chemical and Structural Characterization
The materials prepared are listed in Tables 1and 2, together with the results of XRD and H 2 -TPR analysis. All diffraction peaks of the La-containing perovskites can be assigned to the rhombohedral geometry. (JCPDS 72-1156 (C)), (Fig. 1) while the hexagonal one is observed in SKMC. Doping with K(I) causes the shift of the whole XRD pattern toward lower angles (because of the difference in ionic radii K(I) = 1.38 Å, La(III) = 1.061 Å) [20] confirming the penetration of K into the perovskite cell. No secondary phases are observed.
BET surface area values are summarized in Tables 1 and  2. The values obtained in the catalysts containing K are only slightly lower than that of LMC. The specific surface area of SKMC is, in contrast, significantly lower; this result is probably a consequence of the higher calcination temperature (800 instead of 700 °C) necessary to obtain the perovskite phase.
The XP spectra are compared in Fig. 2. The Mn 2p XP peaks are centered at 642.2-642.5 and 653.6-654.0 eV for the 3/2 and 1/2 components, respectively, a position which is typical of Mn in oxides; it is difficult, however, to distinguish between Mn(III) and Mn(IV) [21]. Cobalt is rather absent in surface, in contrast, and is slightly more relevant in the region below; this suggests that eventual electronic interaction with Mn cannot be really significant. No significant differences are observed when K is also present or when Sr substitutes La. La 3d peak position (834.2-834. 6 and 851.0-851.4 eV, for 3d 5/2 and 3d 3/2 , respectively) and shape (shake-up contributions at 840 and 858 eV) are  consistent with how expected for La(III) in perovskites and are not modified by doping or by the deposition of K. Sr 3d peak shows two doublets whose 5/2 positions, at 132.8 and 133.6 eV, respectively, suggest the presence, beside the perovskite, of carbonate species, compatible with C1s spectra. Two main signals compose the O 1 s spectra, as obtained deconvoluting the spectra: the peak at 529.5-530.0 eV is due to the perovskite lattice oxygen whereas the one at about 532.0 eV at oxygen in hydroxides and carbonates; this last contribution is more evident in SKMC [18,[21][22][23][24][25][26]. The deconvolution procedure for O1s peaks is reported in the Supplementary Information file.
In all catalysts, XPS quantitative analysis (Table 3) reveals the surface segregation of oxygen, consistent with the formation of hydroxyl and carbonate groups, as denoted by qualitative analysis. In LMC surface segregation of Mn is evident, and it is not significantly affected by the insertion of K. Also K shows an attitude to surface segregation. The propensity of K toward surface segregation is particularly evident when Sr substitutes La. It is noteworthy that, in K/ LMC, K is less present in surface than expected, consistently with its frequently reported desorption trend [27,28]. The "after reaction" compositional values are discussed below, in the section "Reactivity".
SEM images are grouped in Fig. 3. Qualitatively it can be appreciated the higher dimension of SKMC particles compared with samples from the same group, identified to reach about 100 nm. It is not clearly distinguishable the presence of K 2 O in K/LMC from the simple visual analysis, but the surface presents a higher variability and irregularity that can be a trace for the presence of different phases unhomogeneously distributed.
The H 2 -TPR curves of the different catalysts are compared in Fig. 4. In all cases the presence of two groups of signals centered around 300-400 °C (Mn(IV) to Mn(III)) and 600-800 °C (Mn(III) to Mn(II)) is evident. The reduction of cobalt is expected to contribute in a low amount to both the groups of signals (Co(III) to Co(II) around 440 °C and Co(II) to Co(0) around 500-600 °C). [29] Focusing on the lower temperature signals two main contributions are observed in LMC at 320 and 388 °C, whereas at higher temperature a broad signal is observed around 1 3 800 °C. The Gaussian fitting procedure on the TPR peaks reveals, in the low temperature signals group, minor contributions around 157, 218, and 270 °C. The comparison with literature data suggests to attribute these last signals to the reduction of surface oxygen species [30]. Calculating the expected moles of hydrogen consumed during the Temperature-Programmed Reaction (considering the stoichiometry of the perovskite) it was possible to estimate the percentage of each species reacting the in analyzed sample. Oxygen species contribute less than 10% to the total H 2 -consumption. The signals at 321 and 388 °C, consistent with the reduction of Mn(IV), contributes for about 30% of hydrogen consumption. The remaining consumption is observed around 800 °C and corresponds to the reduction of Mn(III) [31,32]. The H 2 -uptake corresponds to 0.24 mol H 2 mol −1 LMC at low temperature and 0.35 mol H 2 mol −1 LMC at higher temperature. According to literature, the hydrogen consumption below 500 °C was attributed to the reduction Mn(IV) to Mn(III) of some Mn(IV) present in the perovskite structure, however, based on the hydrogen consumption values, it is likely that the reduction process of Mn(III) to Mn(II) partially occurred in that temperature range. The effect of Co on enhancing the reducibility of Mn(III) cannot be excluded but this hypothesis is not supported by XPS data. The reducibility of Mn(III) at lower temperature was observed to depend on the amount of complexing agent (citric acid) and on the textural properties of the catalyst [33]. In particular, the Mn(III) reduction seems to be favored by higher specific surface area and lower size of the particles. The H 2 uptake between 150 and 400 °C (which can give a rough idea of the reducibility) is 0.24 mol H 2 mol −1 (Mn + Co) which corresponds to 40% of the Mn + Co ions in 1 mol of perovskite.
The insertion of K into the perovskite cell determines a slight increment of the surface oxygen species which contribute to the total consumption for 12.5%. The H 2 -uptake at higher temperature decreases from 0.35 to 0.29 mol H 2 mol −1 perovskite. Also the total uptake decreases from 0.59 to 0.54 mol H 2 mol −1 (Mn + Co). The H 2 uptake at lower temperature is 37% of the expected. The reducibility decreases even more when K is deposited on the surface of LMC (28%). The insertion of Sr inside the perovskite cell in substitution of La is expected to favor the formation of Mn(IV) and thus the reducibility. Consistently the H 2 -uptake at low temperature is 0.78 mol H 2 mol −1 (Mn + Co) and the reducibility is 74%. Noteworthy in the K/LMC and SKMC catalysts no traces of surface oxygen species (reduced at  Tables 1 and 2, and the relative amount of consumed hydrogen for each reduction process is included.
A theoretical average oxidation state for Mn could be also calculated with the H 2 -TPR data, taking into account the charge balance in the perovskite formula. The so-obtained values are reported in the following Table 4. Consistently with how expected the oxidation state is higher and near to 4 when A-cation is Sr and 3 when A is La.

Catalytic Activity
The catalytic behavior in the soot oxidation is compared in Fig. 5. The temperature of maximum conversion of soot and the CO 2 production curve shape underline that the insertion K into the perovskite cell improves the performance of the catalysts toward soot oxidation. The worst results are obtained with the un-activated LMC. In all cases, however, the presence of a catalyst favors the soot oxidation; as demonstrated by blank tests in literature [34].
The soot oxidation in absence of catalyst is suggested to proceed through the following steps: firstly, molecular oxygen is chemisorbed onto the soot surface; the dissociative chemisorption allows oxygen atoms to interact with the surface carbon atoms and form a complex solid [35]. At the end the solid complex desorbs subtracting a carbon atom from the soot particle [36].
The adsorption of the oxidant on the surface active site of carbon and the desorption of the products strongly depend on the oxidant concentration, on the amount of active sites, on the coverage degree and sticking fault. It is reasonable to consider the dissociative chemisorption of oxygen as the rate determinant step. The reactivity of NO 2 is higher with respect to that of O 2 ; it is suggested that NO 2 can form surface complex molecules of the type C-NO 2 and C-ONO capable of decomposing with increasing temperature forming CO, CO 2 and NO [36].
In presence of oxide-based catalysts the reaction mechanism is more complicated and involves the interaction of perovskite and soot particles. It is suggested that several steps follow: adsorption of oxygen on the catalyst, exchange of activated oxygen species (such as O − and O 2− ) from catalyst to the carbon surface, formation of surface oxygen containing (SOC) complex molecules, desorption of SOCs. The catalytic activity is, in this case, affected by the capability of the catalyst to promote the formation and exchange of SOCs [37].
Li et al. suggest four reaction mechanisms that simultaneously occur during soot combustion: soot oxidation in the soot/catalyst interface carried out by the active species O 2− and O − formed by interaction of oxygen with the surface oxygen vacancies; catalytic oxidation of NO to NO 2 on the catalyst surface; N-containing species adsorb on the catalyst surface and can migrate toward the soot particles; NO oxidation to nitrates and nitrites that can oxidize soot giving rise to CO 2 and N 2 ; adsorption of NO 2 on the soot surface and direct reaction (of NO 2 or NOx resulting from decomposition) with surface carbon [38].  All catalysts are active in the soot oxidation: the temperature of maximum combustion, considered as a value to compare the catalytic activity, T m , decreases of, at least, 60 °C with respect of the case without catalyst. The more active catalyst, however, is LKMC with a T m of 306 °C whereas the worst is LMC. The substitution of La with Sr moves toward lower values (around 250 °C) the temperature at which the catalysts start its activity but the T m is similar to that of LMC (323 vs 330 °C). It is interesting to observe that when K is deposited on the LMC surface the catalyst seems to activate at a temperature similar to that of the LKMC but the T m is almost coincident to that of LMC).
LKMC is the catalyst characterized by the higher surface segregation of the B-Cations; moreover, the TPR reveals the higher relative amounts of species reduced at lower temperatures (280-320 °C) so suggesting the contribution of surface oxygen active species to switch the reaction and of Mn(IV) species. Mn(IV) is present in higher amount in the SKMC catalysts but the reduction temperature is shifted at higher temperature (around 400 °C). In this last case the almost complete absence of oxygen surface active species is underlined by a Gaussian fitting procedure of H 2 -TPR peaks.
The XPS analysis carried out on the catalysts after reaction reveals a significant surface segregation of Co in LMC; in the other catalysts the surface segregation of K is present. The surface segregation of cobalt in SKMC is less marked while no particular difference is observed between LKMC and K/LMC. A possible explanation of the phenomenon of Co segregation after reaction is that Co, being more prone to reduction, is drawn to the surface by the reducing surface potential imposed by contact with the soot. Figure 6 shows the carbon dioxide profiles obtained on the K-doped catalysts tested for their Oxygen Storage Capacity (OSC).
Hydrogen-TPR had denoted a faster reduction kinetic when K was present in the structure. In this particular test, the first maximum, sampled at 12 min from the beginning of the test, is related to the most reactive oxygen, and the degree of convergence towards 0 yield can instead give insights on the transport and active replenish of the surface oxygen species. We already argued that K insertion mainly affects surface reactivity, increasing oxygen retention on the surface, whereas Sr employment can help in stabilizing the structure and increase oxygen supply to the reaction sites.
LKMC significantly outperforms SKMC, suggesting that with combined K-doping, helping chemisorption of oxidant species, and Co, that has already proven to induce higher oxygen mobility and structural modifications, the mechanism of the reaction can find a good balance between the two paths of activated oxygen species creation. Indeed surface chemisorption of oxygen and transport of oxygen from the bulk to the surface are the two driving forces of the reaction of soot oxidation in this process. Oxygen transport contribution, intimately related to ion mobility within the structure, becomes more important when doping with Sr. Cobalt was found by Royer et al. to present a good redox activity even at low temperatures, favoring ion mobility in the bulk [39]. Finally, SKMC benefits from the presence of Co, but its transport activity is undermined by K presence.

Conclusion
The scope of this manuscript is the investigation and development of innovative perovskite catalysts with the aim of a possible application in the abatement of carbon soot from diesel engines in the automotive sector. The employment of novel materials, such as the ones suggested from the present work, with a reduced content of noble metals is a challenging yet urgent task in this technologic branch. Therefore, the materials here proposed have been tested for soot oxidation in presence of oxygen and NO, with an excess of oxygen to emulate the conditions of a real diesel engine exhaust mixture, and for their oxygen storage capacity, to get a deeper insight on their mechanistic aspects.
The compositions chosen arise from redox considerations: Mn, stable both in reducing and oxidizing environments, is a good oxygen donating element in perovskite oxides. In the B-site, furthermore, Co is known for soot and NO oxidation. In the A-site the K-doping resulted in an increased catalytic activity both for soot oxidation and  1 3 NO x conversion. K deposition was also investigated to shed light on its significance when outside the structure.
Temperature Programmed Oxidation test (i.e., catalytic activity assessment) in presence of NO, oxygen and soot showed that the double doping (Co and K) increases notably the catalytic activity of the material in the oxidation reaction: LKMC is able to fully oxidize soot at 306 °C, a considerably lower temperature than without any catalyst. Sr-doped samples are still quite active, but less than the counterpart without Sr. Another interesting result is the different behavior for the K-doped sample and the K-deposited one: it is likely that K-doping only is able to promote oxygen vacancies formation.
Considering all these aspects, both from compositional points of view and from catalytic and OSC tests, LKMC seems to be a good trade-off between surface activity guaranteed by K and the importance of oxygen from the lattice thanks to the presence of Co.
Funding Open access funding provided by Università degli Studi di Padova within the CRUI-CARE Agreement.

Conflict of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.