Study on Catalytic Soot Oxidation over Spinel Type ACo 2 O 4 ( A = Co , Ni , Cu , Zn ) Catalysts

In this study, a series of spinel type ACo2O4 (A = Co, Ni, Cu, Zn) catalysts were synthesized via a citric acid complex method for catalytic soot oxidation. The effects of A sites on the removal of soot particles and CO2 selectivity was investigated. The best performances for soot oxidation was achieved over ZnCo2O4 with the T50 of 569°C. The textural and redox properties of the ACo2O4 catalysts were tested using BET, XRD, H2-TPR, O2-TPD, XPS. The substitution of A site caused the change of Co/(Co + Co) ratios as well as the oxygen mobility in the spinel structure. Both Co and Co species were observed on the surface of ACo2O4 catalyst, and Co species were more easily to be reduced to lower states, thus improved the catalytic performance. Compared to other catalysts, the interactions between Zn and Co species led to larger surface area, higher relative concentration of surface active Co and more chemisorbed oxygen species. Additionally, the activity of the catalysts on soot oxidation were inhibited in the presence of sulfur dioxide and water vapor, while nitric oxide facilitated the catalytic activity.


INTRODUCTION
Soot particulate is one of the primary components of atmospheric aerosols, which is mainly emitted from diesel engines (Srivastava et al., 2011;Sitarski, 2012).In contrast to the majority of aerosol species, soot aerosols could act as the dominant absorber of solar radiation in the atmosphere leading to potential global warming (Kim et al., 2011;Fan et al., 2016).Furthermore, soot particles could directly serve as the condensation nuclei contributing to the droplet growth and the formation of haze (Ramanathan et al., 2001;Lohmann, 2002).As the rapid growth of transportation, the control of soot emission has become a hot spot in our society due to its severely negative impact on air quality and human health (Wallington et al., 2006;Fayad et al., 2015).The regeneration of diesel particulate filters (DPFs), a widely applied soot emission control technology for diesel engines, has drawn increasing attention since the DPF may be deactivated due to pore blockage by soot particles (Johansen, 2015).Considering the high burnt off temperature of direct soot combustion (> 600°C), finding effective catalysts for soot combustion in DPF regeneration at low temperatures was of great significance for industrial application (Wei et al., 2011).
Nowadays, spinel type oxides catalysts have become increasingly popular in thermal catalytic reactions (Kuo et al., 2015).By replacing with different transition metals, spinel type AB 2 O 4 oxides could gain remarkable redox properties and robust thermal stability for the unique crystalline structure (Shangguan et al., 1997;Chen et al., 2013), which made spinel type oxides much more appropriate for soot oxidation in diesel engines.Zawadzki et al. (2011) found that CoAl 2 O 4 catalyst showed an remarkable activity for soot oxidation similar to that of the Pt/Al 2 O 3 catalyst.Harrison et al. (2003) reported that diesel soot particles could be burnt over Co 3 O 4 catalyst at 683 K under tight contact, which was much lower than non-catalytic soot combustion.Niu et al. (2015) also brought forward the mesoporous Ce x Co 1-x Cr 2 O 4 spinel oxides as novel and highly stable catalysts for soot oxidation, which exhibited a catalytic activity with a 10% conversion of soot at 352°C.
Transition metal oxides, such as Mn, Cu, Cr, Fe, and Co oxides, played an important role in the activities of the spinel type catalysts.Among these metals oxides, cobalt (Co) oxide has been found to be one of the promising candidates for catalytic combustion of soot particulates due to its lower cost, relative abundant resources and comparable activity.Hong and Lee (2000) reported that the ignition temperature of soot combustion decreased in the order of Co > Mn > Fe over Cs-M catalysts.Dhakad et al. (2007) also reported that Ru-Co bimetallic clusters supported on zirconia showed remarkable improvement in catalytic soot oxidation performance, which was ascribed to Co doping.Hence, many cobalt-based spinel formulations with high activity of soot oxidation have been studied, such as CuCo 2 O 4 (Soloviev et al., 2015), CoFe 2 O 4 (Fino et al., 2006), CoMn 2 O 4 (Shangguan et al., 1996) and so on.However, the influence of the A sites in the cobalt-based spinel oxides on catalytic soot oxidation has not been mentioned in previous literatures.
In this study, a series of spinel type oxides with different A sites (A = Co, Ni, Cu, Zn) were prepared by a citric acid complex method.The removal of soot particles and CO 2 selectivity of the prepared catalysts with different A sites had been investigated.The textural properties were characterized by N 2 adsorption-desorption and X-ray diffraction (XRD).Redox properties of the samples were investigated by temperature-programmed reduction of hydrogen (H 2 -TPR) and oxygen temperature programmed desorption (O 2 -TPD).X-ray photoelectron spectroscopy (XPS) was performed to obtain the patterns of oxygen and metallic states on surfaces of the catalysts.The effects of feed gas components on the ACo 2 O 4 catalysts were also investigated.

Catalyst Preparation
The spinel type ACo 2 O 4 (A = Co, Ni, Cu, Zn) catalysts were prepared via a citric acid complex method.Typically, 0.01 mol of zinc (or nickel, copper, cobalt) nitrate hydrates, 0.02 mol of cobalt nitrate hydrates and 0.045 mol of citric acid were dissolved in 40 ml deionized water under stirring at room temperature until the metal salt dissolved completely.Then, the solution was magnetically stirred and heated at 80°C until a porous wet gel was formed.The obtained gel was died at 110°C for 12 h in air followed by calcination at 700°C for 4 h with a heating rate of 10 °C min -1 in a muffle furnace.The prepared catalysts were crushed and sieved in 40-60 meshes.

Catalyst Characterizations
N 2 adsorption-desorption experiments were carried out to determine the specific surface area, pore size distribution and average pore diameter of the catalysts at 77 K using an Autosorb-1-C (Quantachrome Instrument Crop.).The specific surface area was computed using the Brunauer-Emmett-Teller (BET) method.
X-ray diffraction (XRD) measurements of the catalysts samples were carried out with a PANalytical X'Pert PRO XRD system with a Cu Kα radiation in the 2θ range from 10° to 80° and scanning rate of 4° min -1 .
X-ray photoelectron spectroscopy (XPS) data were obtained with a Thermo ESCALAB 250 using Al Kα X-ray (hν = 1486.6eV) as a radiation source at 150 W. Sample charging effects were eliminated by correcting the observed spectra with a C 1s binding energy (B.E.) value of 284.8 eV.
Temperature-programmed reduction of hydrogen (H 2 -TPR) experiments were carried out with a Micromeritics Autochem II 2920 equipped with TCD using approximately 50 mg of samples.Samples were pretreated at 300°C for 1 h in N 2 flow.After cooling to room temperature, the temperature was increased linearly from 50 to 800°C with a heating rate of 10 °C min -1 using 10% H 2 /Ar at a flow rate of 30 mL min -1 , while the amount of H 2 consumption was measured using a thermal conductivity detector (TCD).
The oxygen temperature programmed desorption (O 2 -TPD) curves were performed in the same apparatus as H 2 -TPR.50 mg catalyst were pretreated in 2% O 2 /He at 500°C for 1 h with a flow rate of 30 mL min -1 .After cooling to 50°C, the catalyst was then heated up to 900°C with a heating rate of 10 °C min -1 in He.The released O 2 was determined by the TCD.

Catalytic Activity test
The catalytic activity of ACo 2 O 4 in soot oxidation were measured in a fixed-bed quartz tube reactor from room temperature to 700°C with a heating rate of 5 °C min -1 .A standard gas of N 2 containing 10 vol.% O 2 was fed at the constant flow rate of 300 mL min -1 .In addition, 1000 ppm NO (when used), 5 vol.%H 2 O (when used) and 100 ppm SO 2 (when used) was also included in the feeding gas.The tubular quartz reactor was loaded with 100 mg of a 9:1 by weight mixture of powdered catalyst and carbon soot (Printex-U, Orion Engineered Carbons) along with 300 mg of SiO 2 granules (0.3-0.7 mm), which were added to prevent pressure drop and improved heat transfer.Catalytic soot oxidation tests were performed in loose soot-catalyst contact conditions.Blank experiments were performed under aforementioned experimental conditions without catalysts, that is, only in the presence of the soot.
The temperatures of the catalyst bed and tubular electric furnace were monitored automatically by E-type thermocouples.The outlet concentrations of CO 2 and CO were analyzed on-line by a FT-IR Gas Analyzer (Gasmet Dx4000, Finland).The CO 2 selectivity (S CO 2 ) was defined as follows: where C CO 2 and C CO are the outlet concentration of CO 2 and CO, respectively.Besides, X soot represented the conversion of soot particles.m tot represented the total mass of soot particles added to the reactor.M C represented the molecular weight of carbon.n CO 2 represented the converted amount of carbon dioxide.
The catalytic activity was evaluated by the values of T ig , T m , T 50 and T 90 .T ig represented the ignition temperature of catalytic soot oxidation, T m represented the temperature of the peaks in soot oxidation curves.T 50 and T 90 were defined as the temperatures at which 50% and 90% of the soot were oxidized during the experiment, respectively.
Fig. 2 showed the results of N 2 adsorption-desorption experiments.All the prepared catalysts exhibited the type II isotherms with H3 type hysteresis loop, which indicated that catalysts possessed a slit pore structure.S BET and pore size of the spinel type catalysts were shown in Table 1.Spinel oxides constituted by Ni and Co resulted in a smaller S BET compared to other catalysts, whereas the largest S BET was obtained over the ZnCo 2 O 4 catalyst with a value of 16.4 m 2 g -1 , which was comparable with the results of Abu-Zied et al. (2015).The interactions between Zn and Co species resulted in a relative larger surface area of the spinel type catalysts.

Redox Properties
H 2 -TPR results were shown in the Fig. 3.It could be seen that the nature of A sites had a great influence on the reduction profiles.For Co 3 O 4 catalyst, the main reduction peak centered around 411°C could be ascribed to the reduction of Co 3 O 4 to CoO (Zhou et al., 2015).The reduction profiles of NiCo 2 O 4 catalyst was more complicated, with a strong peak at 386°C and a weak peak at 547°C.The former peak corresponded to the reduction of NiCo 2 O 4 to NiCoO 2 , while the latter corresponded to the co-reduction of Co 2+ to Co 0 and Ni 2+ to Ni 0 (Wang et al., 2015c).In the case of CuCo 2 O 4 catalyst, we noted that three peaks between 200°C and 350°C were observed.The former shoulder peak at 208°C was assigned to the reduction of Cu 2+ to Cu 0 , whereas the second peak at 243°C was attributed to the overlap of the reduction of Cu 2+ to Cu 0 and fractional reduction of Co 3+ to Co 2+ (Tien-Thao et al., 2007).The last main peak at 302°C was assigned to fractional reduction of Co 3+ to Co 2+ .For ZnCo 2 O 4 catalyst, the TPR curves showed two reduction peaks attributed to two types of oxygen species.The shoulder at 364°C corresponded to octahedral domains, where Co 3+ was reduced to Co 2+ .The peak at 414°C could be due to the tetrahedral domains, where Zn 2+ was reduced to Zn 0 (Klissurski and Uzunova, 1990).We could conclude that the onset temperature of the reduction of Co 3+ species shifted to lower temperature after Ni, Cu or Zn substitution in A sites.H 2 consumptions of the prepared catalysts were shown in Table 1.It could be seen that the H 2 consumption of the spinel type catalysts followed the order as Co     The O 2 -TPD profiles of the spinel type catalysts were showed in Fig. 4. The α-O 2 peaks founded below 500°C stood for chemisorbed surface active oxygen species (O 2 -, O -) (Yoon et al., 2014), whereas the peaks above 750°C could be ascribed to the release of lattice oxygen (O 2-) (Yang et al., 2015).We could note that the amount of O 2 desorption from the spinel type catalysts varied in the order of NiCo 2 O 4 < Co 3 O 4 < ZnCo 2 O 4 < CuCo 2 O 4 , respectively.Magnification of the α-O 2 peaks were displayed in Fig. 4(b), a larger amount of the surface adsorbed oxygen species desorbed from the surface of ZnCo 2 O 4 (0.19 µmol g -1 , based on the curves below 500°C), compared to those of CuCo 2 O 4 (0.15 µmol g -1 ), NiCo 2 O 4 (0.12 µmol g -1 ), Co 3 O 4 (0.13 µmol g -1 ), which was well corresponding with the O 1s XPS results listed in Table 1.The substitution of A sites enhanced the active oxygen availability from catalyst surface, which was beneficial for soot combustion, especially for ZnCo 2 O 4 catalyst.
In order to obtain a better insight into the chemical states and local atomic environments of the spinel type catalysts, XPS technique was carried out.and Co 2p 1/2 .Both the spectra of Co 2p 3/2 and Co 2p 1/2 peaks contained a satellite peak of all the samples, which represented the existence of Co 2+ species (Chuang et al., 1976;Wang et al., 2015).The Co 2p spectra could be divided into four peaks.The peaks centered at 779.6 eV and 794.7 eV belongs to Co 3+ species, while the rest stood for Co 2+ species (Wang et al., 2015b).Co 3+ species were considered to play an efficient role in soot oxidation.Among the prepared samples, Co 3+ /(Co 2+ + Co 3+ ) ratios ranged from 24.3% to 28.0%, which suggested that transition metals substitution in A sites increased the amount of surface Co 3+ species.ZnCo 2 O 4 sample displayed the highest Co 3+ /(Co 2+ + Co 3+ ) ratio value of 28.0%.Furthermore, when cobalt was substituted with Zn, the binding energy of Co 2p 3/2 peak shifted towards higher binding energy value from 794.9 eV to 795.3 eV.The shifts to a higher binding energy demonstrated an increase of Co 3+ , causing by rearrangement of the cation distribution as well as the change of chemical environment around cobalt ions after taking place the A site.

Soot Oxidation Activity
The catalytic activities of the ACo 2 O 4 catalysts on soot oxidation were shown in Fig. 6 and Table 2.All the prepared catalysts decreased the temperature for soot oxidation compared to soot particles with no catalyst.From Fig. 6(a), we could found that Co 3 O 4 and NiCo 2 O 4 catalysts displayed a T ig of 500°C and 525°C, respectively.Moreover, ZnCo 2 O 4 and CuCo 2 O 4 catalysts was found to be easier to ignite with a T ig about 465°C and 470°C, which was obtained by extrapolating the steeply rising portion of the CO 2 curve to the axis of abscissas.For the blank test, the T ig of the soot with no catalyst appeared near 550°C, which was approximately 85°C higher than ZnCo 2 O 4 catalyst.For the peak temperature of the curves, soot oxidation with no catalyst exhibited a T m value as high as 624°C.As expected, the T m of the ZnCo 2 O 4 catalyst was much lower than the rest samples with a value of 575°C, and the CO 2 concentration at the T m was the highest, which meant the soot particles were much easier to turn into vigorous oxidation.Additionally, the width of the curve peaks could represent the time of catalytic soot oxidation process.It was obvious that ZnCo 2 O 4 spinel type catalysts cost the least time for soot conversion.In Fig. 6(b), the relationship between the removal of soot and temperature was much more clear.It was interesting that the activity of Co 3 O 4 catalyst was better than NiCo 2 O 4 catalyst.The T 50 and T 90 of Co 3 O 4 catalyst were 580°C and 626°C, respectively, which were both 5°C lower than NiCo 2 O 4 catalyst.Hence, replacing Co with Ni at A site was unfavorable for soot conversion.In contrast, when Cu and Zn substitution in A sites, the T 50 decreases from 580°C to 574°C and 569°C, respectively.Although the removal of soot on CuCo 2 O 4 catalyst exhibited a relative low value of T 50 , the T 90 was nearly 20°C higher than ZnCo 2 O 4 catalyst, which meant a slower reaction rate on CuCo 2 O 4 catalyst.Therefore, ZnCo 2 O 4 catalyst got the best catalytic activity over the removal of soot, and the efficiency of soot oxidation followed the order of ZnCo CO 2 selectivity of the spinel type samples were shown in Fig. 7, which mainly appeared an inverse correlation with the catalytic soot oxidation efficiency curves.The peaks of the soot oxidation efficiency curves corresponded to the valleys of CO 2 selectivity.Compared with the bare soot, all temperature programmed reactions with the prepared catalysts increased the CO 2 selectivity of catalytic soot oxidation.Below 590°C, ZnCo 2 O 4 catalyst presented a better CO 2 selectivity, while CuCo 2 O 4 catalyst overtook above 590°C.

Effect of Feed Gas Components
Since ZnCo 2 O 4 catalysts displayed best catalytic activity in our works, further tests were carried out to study the influence of feed gas components (i.e., SO 2 , H 2 O, NO).The effect of the SO 2 on the soot oxidation was shown in Fig. 8(a).100 ppm SO 2 was fed just after preheating.We noted that the presence of SO 2 had an obviously passive effect on soot oxidation.It could be seen that the T m for ZnCo 2 O 4 were 575°C while the T m increased to 621°C in the existence of SO 2 .The results revealed that SO 2 significantly abated the catalytic activity on soot oxidation.The inhibition of SO 2 might be caused by the formation of sulfate or sulfide on the surface of catalysts, which diminished the active sites and surface area (Li et al., 2006).
On the account of H 2 O existing in exhaust of diesel engines, catalytic soot oxidation experiments were carried out with 5% H 2 O in the feed gas.As reported in the literatures, H 2 O could play a role in facilitating the decomposition of the highly stable carbon surface species formed by the partial oxidation (Wang et al., 2015a).Nevertheless, H 2 O absorbed on the catalysts could also compete with the reactant for the active sites causing the inhibition on the catalytic activity (Dai et al., 2016).The results given in Fig. 8(b) exhibited the effect of H 2 O on the catalytic activity.ZnCo 2 O 4 catalysts displayed a remarkable activity decline around 500-650°C, which would have been the temperature range with most severe oxidation activity.The CO 2 concentration at the peaks of ZnCo 2 O 4 catalysts dropped nearly half of the initial value.Meanwhile, T ig for ZnCo 2 O 4 increased about 25°C with H 2 O in the feed gas.The influence of NO was carried out with NO/O 2 as oxidizing agents, as displayed in Fig. 9.The T ig for ZnCo 2 O 4 catalysts on catalyzed soot oxidation was 425°C in the feed gas NO + O 2 , which was 40°C lower than that in only O 2 /N 2 .It was observed that the presence of NO exhibited a promotional influence on the soot oxidation.The catalytic soot oxidation curves for ZnCo 2 O 4 shifted left when NO was used.Lin et al. (2009) suggested that C(O) intermediates were key precursors from soot oxidation and nitrates acted as the principal oxidants for C(O) oxidation.NO might first react with the catalyst to generate a much stronger oxidative NO 2 , and soot oxidation accomplished the catalysis cycle with the NO 2 -assisted mechanism.Hence, the asmentioned catalysts should have decent activity in the presence of NO, even better than O 2 /N 2 gas composition.

DISCUSSIONS
All the prepared ACo 2 O 4 (A = Co, Ni, Cu, Zn) catalysts displayed spinel type structures.For an ideal spinel oxide, cations were arranged in a cubic close-packed lattice, where trivalent cations occupied half the octahedral holes and bivalent cations took place one-eighth of tetrahedral holes (Chellam et al., 2000).It was reported that cobalt containing spinel catalysts were ideal spinel structures, in which all the octahedral sites were occupied by cobalt species.Since the A site was replaced by Ni, Cu and Zn, respectively, distortion occurred in the spinel crystals.Octahedral sites tended to stretch or flatten with the variation of composition, while A sites with tetrahedral structure only changed in size (Robinson et al., 1971).Hence, the Co-O band and the chemical states of cobalt species appeared in varied forms, which have a significant impact on performance of soot oxidation.
Due to catalytic soot oxidation is a heterogeneous catalytic process, the interaction between soot and catalyst is very important.As presented in Table 1 and Fig. 6, the catalytic performance was extremely related to the specific surface area of the prepared samples.It had been reported the interaction between soot and catalyst was enhanced with larger specific surface area.Similar results were also obtained in the literatures, in which tight contact conditions provided substantial increases in activity for catalytic soot oxidation compared to loose contact (Gardini et al., 2016).Despite the fact that BET surface area affects catalytic activity, the redox properties of the catalysts were more decisive factors than S BET .In order to gain high thermal stability and form spinel structures, all the prepared samples were calcined at a relatively high temperature (700°C), which resulted in little difference in the value of S BET .Nevertheless, the catalytic performance of spinel type oxides varied a lot.
Depending on the XPS and O 2 -TPD results, ZnCo 2 O 4 samples had a comparatively highest ratio of O ads /(O lat + O ads ) about 64.8%.O ads species were assumed to be very active to react with soot at the catalytic surface.O ads /(O lat + O ads ) ratio could also be considered as a mirror to reflect the amount of oxygen vacancies (Chen et al., 2013).It had been reported that oxygen vacancies play a key role in the adsorption and dissociation of oxygen species.On the other hand, Bueno-López et al. (2005) reported the significance of the lattice oxygen in the soot oxidation reactions.The equilibrium was reached between the gas phase and solid oxides, so the mobility of oxygen could be reflected by desorbed active oxygen species.Piumetti et al. (2016) demonstrated that oxygen storage and release had to carry out under the diffusion of oxide ions.Fig. 4 showed the O 2 -TPD profiles of the prepared catalysts.The oxygen desorption peaks areas were corresponding to the mobility of the oxygen species.The amount of oxygen desorption under 500°C varied in the order of ZnCo 2 O 4 > CuCo 2 O 4 > Co 3 O 4 > NiCo 2 O 4 .We proposed that lattice oxygen participated in the soot oxidation reaction at the beginning (Levasseur et al., 2009).Oxygen vacancies formed in the same position where was occupied by lattice oxygen involved into the reaction.Then surface adsorbed oxygen species transferred to supply the oxygen species, which was considered the rate determining step in the reaction (Fino et al., 2003).Hence, oxygen mobility was a vital factor in the catalytic removal of soot particles.
In addition, according to H 2 -TPR and XPS results, the catalysts with more Co 3+ cations tended to be more effective in soot oxidation.In consideration of the onset temperature of the reduced peaks, Co 3+ ions had more active redox properties, which was agreed with the catalysts activity profiles.Omata et al. (1996) and Zafeiratos et al. (2010) also reported the similar results.It could be seen in Fig. 5(a) that ZnCo 2 O 4 spinel catalyst possessed a Co 3+ /(Co 2+ + Co 3+ ) ratio of 28.0%, which was the highest among the prepared samples.The spinel structure of cobalt based catalysts disordered with Zn substitution in A site, and the high valence state tended to occupy the octahedral holes in the spinel oxides.Therefore, we could consider that Zn had the strongest influence on increasing the amount of Co 3+ , which was considered to be crucial in catalytic soot oxidation reactions.

CONCLUSIONS
In summary, a series of spinel type ACo 2 O 4 catalysts were successfully obtained using a citric acid complex method.The effect of A sites on catalytic soot oxidation over ACo 2 O 4 catalysts had been investigated as far as removal efficiency and CO 2 selectivity be concerned.Among all the prepared catalysts, ZnCo 2 O 4 catalyst showed the best performance for the catalytic removal of soot particles, which reached a 50% conversion at 569°C and 90% conversion at 602°C.When Zn substitution in A sites, the spinel structure displayed better oxygen mobility and higher concentration of surface active Co 3+ , which could be considered as major factors in catalytic soot oxidation.The effects of feed gas components were also studied in this work.Sulfur dioxide and water vapor were found to have significantly passive impacts on the catalytic activities of ACo 2 O 4 for soot oxidation, whereas nitric oxides facilitated the catalytic performance.Deep insight to cobalt-containing spinel catalysts provided the rational design of catalysts for soot combustion and other heterogeneous catalytic processes.
3 O 4 > CuCo 2 O 4 > ZnCo 2 O 4 > NiCo 2 O 4 , which could be due to the change of A sites.

Fig. 2 .
Fig. 2. N 2 adsorption-desorption isotherms and the corresponding pore size distribution curves of the spinel type catalysts.
Fig. 5 displayed the O 1s and Co 2p XPS spectra of the ACo 2 O 4 (A = Co, Ni, Cu, Zn) catalysts.The Co 2p spectra of the spinel type samples were shown in Fig. 5(a).The binding energy values were split into two main peaks corresponding to Co 2p 3/2

Fig. 7 .
Fig. 7. Catalytic soot oxidation efficiency of the spinel type with different cation A sites.
H 2 O had an opposite influence on catalytic activity of ZnCo 2 O 4 .