Cobalt Based Catalysts on Alkali-Activated Zeolite Foams for N2O Decomposition

In this work, we studied the effect of alkali-activated zeolite foams modifications on properties and catalytic activity of cobalt phases in the process of catalytic decomposition of N2O. The zeolite foam supports were prepared by alkali activation of natural zeolite followed by acid leaching and ion exchange. The cobalt catalysts were synthesised by a different deposition technique (direct ion exchange (DIE) and incipient wetness impregnation (IWI) method of cobalt on zeolite foams. For comparison, catalysts on selected supports were prepared and the properties of all were compared in catalytic tests in the pellet form and as crushed catalysts to determine the effect of internal diffusion. The catalysts and supports were in detail characterized by a variety of techniques. The catalyst activity strongly depended on the structure of support and synthesis procedure of a cobalt catalyst. Ion exchange method provided active phase with higher surface areas and sites with better reducibility, both of these factors contributed to higher N2O conversions of more than 80% at 450 ◦C. A large influence can also be attributed to the presence of alkali metals, in particular, potassium, which resulted in a modification of electronic and acid base properties of the cobalt oxide phase on the catalyst surface. The promotional effect of potassium is better reducibility of cobalt species.


Introduction
The alkali-activated materials have properties very similar to zeolites but do not form crystalline structures. These materials can be prepared by activating a variety of aluminosilicate components. Raw materials for the preparation of these materials, e.g., natural zeolites, most often occur as compact fine-grained rocks formed by sedimentary or volcanic activity, whose main component is zeolite clinoptilolite, belonging to the group of heulandite, accompanied by minerals from the group of feldspars, clays and mica. Due to their properties, natural zeolites are widely used as sorbents for water purification (Cs + , Rb + and NH 4+ ) or gases (NH 3 ) [1,2]. They do not have suitable properties for use in catalysis, which requires their modification by processes such as dealumination [3,4], desilication or In addition to zeolite catalysts, other types of catalysts prepared, for example, by impregnation of other supports, precipitated catalysts or catalysts prepared by other techniques based on the presence of spinel phases of cobalt or mixed Co-Mn-Al phases are studied for catalytic N 2 O decomposition [11,24,27,35]. Modification of the cobalt spinels with a small amount of alkali metals significantly increases the activity of the catalyst, alkali metal promoter lowers the work function of the cobalt spinel facilitating redox processes that occur between the catalyst surface and the reaction oxygen intermediates produced during the N 2 O decomposition leading to the significant increase in the catalytic activity [14,17,23,27,[35][36][37][38]. Synthesis conditions of the Co spinel oxide could significantly affect catalytic properties. Different methods were described for the preparation of cobalt oxide-based catalysts: (co)-precipitation from cobalt nitrate solution using different precipitation agents like Na 2 CO 3 [20,39], K 2 CO 3 [40,41], (NH 4 ) 2 CO 3 [16], KOH [42] and NH 3 ·H 2 O, NaOH [39], by thermal treatment of cobalt nitrate [24,43,44], cobalt carbonate [14,24] or cobalt benzoate-dihydrazinate complex [45] and by solution combustion synthesis from cobalt nitrate and urea [43]. Equally, the deposition of active phase on suitable support has a significant impact. Therefore, the selection of suitable support could lead to more efficient utilization of active phase. It is achieved by applying a thin active layer to the support material. Supported cobalt oxide N 2 O catalysts were studied mainly on different monoliths [46][47][48][49], sieves [32,50], tablets, pellets or extrudates [29,51] and ceramic foams [52].
The work aims to prepare cobalt catalysts based on alkali-activated zeolite foams and determine the influence of catalyst support modifications and cobalt oxide loading method on physic-chemical properties and catalytic activity in N 2 O decomposition. Acid leaching and ion exchange of alkali activated zeolite foam (AZF) were used to prepare catalyst supports, which were subsequently impregnated with cobalt nitrate, and after annealing, cobalt oxide catalysts were obtained. The same method was used to prepare comparative catalysts using alumina as a support. The effect of different preparation on the properties of the active phase was studied on catalysts prepared by direct ion exchange (DIE) of basic alkali activated zeolite foam, as well as the promotional effect of manganese in Co catalyst in catalytic decomposition of N 2 O. The effect of Mn, was monitored in our study because CoMnAl HTC 4:1:1 catalysts are often reported to have very good results.Studied catalytic materials showed significant differences in the textural properties and chemical composition of the active phase resulting in significant differences in activity depending on the chosen modification method. All catalysts were characterized by various instrumental techniques and tested in the N 2 O catalytic decomposition in inert gas. The obtained results showed the further possible use of catalysts based on alkali-activated zeolite foams in catalytic applications. In the field of industrial catalysis, in addition to the catalytic results, the ease of synthesis, affordability and the price of the catalysts used are also important, and therefore, these materials have a lot to offer.

Characterisation of the Catalysts and Supports
Zeolite foam used as catalyst support was prepared by foaming of alkali-activated natural zeolite mixture. Obtained porous solid foam material was post-synthesis modified (i) by acid leaching with 3M HCl or (ii) by ion-exchange treatment using 1M NH 4 NO 3 .
As-prepared zeolite foam (AA-S, Table 1) had a higher alkali metals content than natural zeolites due to the use of mixed alkaline activator KOH + Na 2 SiO 3 . Chemical composition of Al 2 O 3 can be seen in Appendix A Table A1. A part of the alkaline activator occurred in a form of mixed Na/K silicate, which can be seen in images from SEM equipped with EDS analyser as a leaf-like particles ( Figure 1). Ion-exchange treatment of AA-S sample using NH 4 NO 3 solution (AA-N) reduced predominantly content of K and Na cations. A part of K + and Na + ions was in "unchangeable positions" trapped inside of the zeolite and/or other aluminosilicates framework [4,8]. Results of chemical composition analysis (Table 1) showed a noticeable effect of HCl leaching (AA-D). HCl mostly reduced the alkali Catalysts 2020, 10, 1398 4 of 21 content (K, Na) in addition to other elements (Fe, Ca, Mg and especially Al), which resulted into higher Si/Al ratio (up to 11.8 from the original 5.8). Na) in addition to other elements (Fe, Ca, Mg and especially Al), which resulted into higher Si/Al ratio (up to 11.8 from the original 5.8).  * AM = alkali metals (Na + K), ** SSA = specific surface area (BET).
These three types of support based on zeolite foams (AA-S, AA-N, AA-D) and one commercial supports (γ-Al2O3) were used to prepare cobalt catalysts with 5 wt.% Co content. The actual cobalt content ranged from 4.8 to 6.4 wt.%, as determined by XRF the slightly higher cobalt content was due to the non-stoichiometric water content of the cobalt nitrate hydrate. The last two catalysts tested (AA-IE-Co and AA-IE-CoMn) were prepared by a direct ion-exchange method using a solution of cobalt nitrate, and mixture of cobalt nitrate with manganese nitrate, respectively. The cobalt content of AA-IE-Co catalysts was 6.3 wt.% and the cobalt content of the AA-IE-CoMn catalyst was 5.4 wt.% and manganese 0.8 wt.%, respectively. The cobalt content was average for the whole pellet, so the cobalt content in the active layer of AA-IE-Co catalyst is higher than 6.3 wt.% (theoretically up to about 13 wt.% or about up 11 wt.% for AA-IE-CoMn catalyst). The molar ratio of Co/Mn is 6.5:1, the difference compared to the target ratio (Co/Mn 4:1) is due to the different course of precipitation of Co and Mn ions during the formation of the layer in the pellet. Catalysts prepared using this method had a significantly higher content of potassium and slightly higher content of sodium in comparison with a catalyst impregnated on AA-N support. This showed that cobalt precipitation precluded complete ion-exchange of Na + and K + ions by Co 2+ ion. Higher Na and K contents were observed only for the AA-S-Co catalyst, but the alkalis were present here in the form of Na/K silicates. As Table 1 shows, the presence of these components had a very negative effect on the specific surface area (SSA) because the silicates closed the pores. Additionally, due to pore opening and cleaning, there was also an increase in specific surface area (SSA) two times for (AA-N) and more than seven times for AA-D. SSA of catalysts These three types of support based on zeolite foams (AA-S, AA-N, AA-D) and one commercial supports (γ-Al 2 O 3 ) were used to prepare cobalt catalysts with 5 wt.% Co content. The actual cobalt content ranged from 4.8 to 6.4 wt.%, as determined by XRF the slightly higher cobalt content was due to the non-stoichiometric water content of the cobalt nitrate hydrate. The last two catalysts tested (AA-IE-Co and AA-IE-CoMn) were prepared by a direct ion-exchange method using a solution of cobalt nitrate, and mixture of cobalt nitrate with manganese nitrate, respectively. The cobalt content of AA-IE-Co catalysts was 6.3 wt.% and the cobalt content of the AA-IE-CoMn catalyst was 5.4 wt.% and manganese 0.8 wt.%, respectively. The cobalt content was average for the whole pellet, so the cobalt content in the active layer of AA-IE-Co catalyst is higher than 6.3 wt.% (theoretically up to about 13 wt.% or about up 11 wt.% for AA-IE-CoMn catalyst). The molar ratio of Co/Mn is 6.5:1, the difference compared to the target ratio (Co/Mn 4:1) is due to the different course of precipitation of Co and Mn ions during the formation of the layer in the pellet. Catalysts prepared using this method had a significantly higher content of potassium and slightly higher content of sodium in comparison with a catalyst impregnated on AA-N support. This showed that cobalt precipitation precluded complete ion-exchange of Na + and K + ions by Co 2+ ion. Higher Na and K contents were observed only for the AA-S-Co catalyst, but the alkalis were present here in the form of Na/K silicates. As Table 1 shows, the presence of these components had a very negative effect on the specific surface area (SSA) because the silicates closed the pores. Additionally, due to pore opening and cleaning, there was also an increase in specific surface area (SSA) two times for (AA-N) and more than seven times for AA-D. SSA of catalysts did not change significantly compare to original supports. The decrease of SSA was observed only for alumina and AA-D supported catalyst, which could have been caused by partially blocking the pores with cobalt, more precisely, its oxide. This decrease in AA-D was related to blockade of clinoptilolite micropores which have been released by acid leaching [4,8]. The catalysts AA-IE-Co and AA-IE-CoMn prepared in this way exhibit relatively high specific surface (SSA) areas compared to catalysts employing differently modified zeolite foam supports (Table 1). This was probably partially caused by the formation of porous cobalt and cobalt-manganese oxides originating from precipitated hydroxides on the zeolite foam surface, or by partial cleaning of the zeolite foam surface (an increase of crystallinity or intensity respectively was observed, Figure 2) and accessibility of pores. The increase in the surface area has not yet been satisfactorily explained but has been verified by repeated measurements. Rutkowska [53] states that the increase in surface area of the samples (Cu, Fe and Co beta zeolites) could be related to the secondary recrystallization of zeolites during the preparation of catalysts. In this case, it could also be caused by changes in the binder N/K-A-S-H phase or by the generation of additional porosity between the cobalt and manganese hydroxides deposited on the sample surface.
it could also be caused by changes in the binder N/K-A-S-H phase or by the generation of additional porosity between the cobalt and manganese hydroxides deposited on the sample surface.
Since the samples are predominantly macroporous materials, textures of catalytic supports and prepared catalysts were determined via Hg porosimetry. The obtained results ( Figure 2) showed that other supports contain predominantly bigger pores. Zeolite foams have broad pore spectrum from mesopores to macropores to supermacropores with a diameter of tenths to hundreds of micrometres with corresponding pore mean size thousands nm. Total intrusion volume of zeolite foams varies from 0.65 (AA-S) to 0.75 mL/g (AA-D). Intrusion volume of their "mesopores" is from 0.06 mL/g (AA-S), 0.07 mL/g (AA-N) to 0.09 mL/g (AA-D). Increase in total intrusion volume and share of AA-D support mesopores was caused by cleansing the surface from the excess alkali activator and mineral sediments inside the pores of natural zeolite. In the case of cobalt catalysts as well of used catalysts (after catalytic reaction of N2O decomposition), there were no big changes (decreases) observed in intrusion volume. Changes were only due to blocking the pores with cobalt oxide. In contrast, an increase of the mesopore volume was observed in AA-IE-Co and AA-IE-CoMn catalysts (both 0.09 mL/g), which can be attributed to cobalt oxides formed during the catalyst synthesis. Alumina as comparative support is purely mesoporous material with the overall intrusion volume of 0.5 mL/g and mean pore diameter 5.6 nm, Co-Al2O3 catalyst shows a partial reduction in intrusion volume due to cobalt oxides deposited on the surface (Appendix A Figure A1).  Since the samples are predominantly macroporous materials, textures of catalytic supports and prepared catalysts were determined via Hg porosimetry. The obtained results ( Figure 2) showed that other supports contain predominantly bigger pores. Zeolite foams have broad pore spectrum from mesopores to macropores to supermacropores with a diameter of tenths to hundreds of micrometres with corresponding pore mean size thousands nm. Total intrusion volume of zeolite foams varies from 0.65 (AA-S) to 0.75 mL/g (AA-D). Intrusion volume of their "mesopores" is from 0.06 mL/g (AA-S), 0.07 mL/g (AA-N) to 0.09 mL/g (AA-D). Increase in total intrusion volume and share of AA-D support mesopores was caused by cleansing the surface from the excess alkali activator and mineral sediments inside the pores of natural zeolite. In the case of cobalt catalysts as well of used catalysts (after catalytic reaction of N 2 O decomposition), there were no big changes (decreases) observed in intrusion volume. Changes were only due to blocking the pores with cobalt oxide. In contrast, an increase of the mesopore volume was observed in AA-IE-Co and AA-IE-CoMn catalysts (both 0.09 mL/g), which can be attributed to cobalt oxides formed during the catalyst synthesis. Alumina as comparative support is purely mesoporous material with the overall intrusion volume of 0.5 mL/g and mean pore diameter 5.6 nm, Co-Al 2 O 3 catalyst shows a partial reduction in intrusion volume due to cobalt oxides deposited on the surface (Appendix A Figure A1).
SEM-EDS mapping of catalyst pellet cross-section ( Figure 3) showed identical macroporous structure of the catalysts (every first figure in the row) and homogeneous distribution of cobalt on the surface of the modified zeolite foam catalyst support in the case of AA-N-Co and AA-D-Co catalysts (Figure 3b,c). In case of impregnation of the AA-S support (Figure 3a), an area non-covered by cobalt was observable in the middle of the pellet, which could be caused by precipitation of cobalt hydroxide during impregnation process by alkali agents in the zeolite foam. This as-created layer was blocking further impregnation inside the pellet. Certainly, this will also be affected by the inhomogeneity of the pellets caused by the synthesis method used. A similar effect was observed during preparations via the direct ion-exchange method. As can be seen in Figure 3d,e cobalt was observed only on the outer surface of the pellet in the form of a ring. The depth of cobalt penetration reached approximately 0.7 mm. The difference in depth compared to AA-S-Co was due to different catalyst preparation conditions (solution concentration, temperature, etc.). Manganese distribution showed a decline in content from the surface to the core, but in comparison with cobalt, there was not a strong borderline. There was a significant decline in content of sodium and potassium (observable from grey and yellow colour intensity on Figure 3) in the following order: AA-S-Co, AA-IE-Co/AA-IE-CoMn, AA-N-Co and AA-D-Co and distribution in the whole cross-section is in agreement with results of chemical analysis (XRF).
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 22 surface of the modified zeolite foam catalyst support in the case of AA-N-Co and AA-D-Co catalysts (Figure 3b,c). In case of impregnation of the AA-S support (Figure 3a), an area non-covered by cobalt was observable in the middle of the pellet, which could be caused by precipitation of cobalt hydroxide during impregnation process by alkali agents in the zeolite foam. This as-created layer was blocking further impregnation inside the pellet. Certainly, this will also be affected by the inhomogeneity of the pellets caused by the synthesis method used. A similar effect was observed during preparations via the direct ion-exchange method. As can be seen in Figure 3d,e cobalt was observed only on the outer surface of the pellet in the form of a ring. The depth of cobalt penetration reached approximately 0.7 mm. The difference in depth compared to AA-S-Co was due to different catalyst preparation conditions (solution concentration, temperature, etc.). Manganese distribution showed a decline in content from the surface to the core, but in comparison with cobalt, there was not a strong borderline. There was a significant decline in content of sodium and potassium (observable from grey and yellow colour intensity on Figure  3) in the following order: AA-S-Co, AA-IE-Co/AA-IE-CoMn, AA-N-Co and AA-D-Co and distribution in the whole cross-section is in agreement with results of chemical analysis (XRF). The crystalline structure of materials was determined using X-ray powder diffraction (XRD). In case of zeolite foam supported catalysts, the analysis showed clinoptilolite to be a main crystalline phase The crystalline structure of materials was determined using X-ray powder diffraction (XRD). In case of zeolite foam supported catalysts, the analysis showed clinoptilolite to be a main crystalline phase with typical reflections at 2theta 9.9 • , 11.2 • , 17.5 • and 22.4 • (PDF 83-1261) along with other minor phases in the natural zeolite, belonging to the mineralogic group of clays and feldspars ( Figure 4a). The diffractograms showed the difference in crystallinity or intensities of main reflections of clinoptilolite respectively depending on the modification method used. The crystallinity of alkali-activated material was lower in comparison with the original natural zeolite [8]. This may be caused by a partial structure disruption during the alkaline activation. There was a significant increase in crystallinity after IE using NH 4 NO 3 caused by surface cleansing. In contrast, after leaching with 3M HCl, there was a decrease in crystallinity connected with disruption of the zeolite structure due to removing Al from the clinoptilolite crystalline matrix. During the modification process, there was a change in intensities at 2theta 11.2 • /9.9 • that corresponded to planes (020) and (200), caused by the difference in electron density in that plane. From the original ratio 0.5 in the natural zeolite, the value increases after alkali-activation to 0.8 and after leaching with HCl or IE treatment, it decreases to 0.4 [8]. Comparative catalysts showed typical broad reflections for γ-Al 2 O 3 (2theta 46.5 • and 67.3 • , PDF 74-2206) (Appendix A Figure A2).
Catalysts 2020, 10, x FOR PEER REVIEW 7 of 22 with typical reflections at 2theta 9.9°, 11.2°, 17.5° and 22.4° (PDF 83-1261) along with other minor phases in the natural zeolite, belonging to the mineralogic group of clays and feldspars (Figure 4a). The diffractograms showed the difference in crystallinity or intensities of main reflections of clinoptilolite respectively depending on the modification method used. The crystallinity of alkali-activated material was lower in comparison with the original natural zeolite [8]. This may be caused by a partial structure disruption during the alkaline activation. There was a significant increase in crystallinity after IE using NH4NO3 caused by surface cleansing. In contrast, after leaching with 3M HCl, there was a decrease in crystallinity connected with disruption of the zeolite structure due to removing Al from the clinoptilolite crystalline matrix. During the modification process, there was a change in intensities at 2theta 11.2°/9.9° that corresponded to planes (020) and (200), caused by the difference in electron density in that plane. From the original ratio 0.5 in the natural zeolite, the value increases after alkali-activation to 0.8 and after leaching with HCl or IE treatment, it decreases to 0.4 [8]. Comparative catalysts showed typical broad reflections for γ-Al2O3 (2theta 46.5° and 67.3°, PDF 74-2206) (Appendix A Figure A2). Cobalt catalysts prepared using these supports showed significant crystalline structures corresponding to Co3O4 (2theta 31.5°, 36.9°, 45.0°, 55.7°, 59.6° and 65.6°, PDF 65-3103). These were the most apparent in case of using alumina as support (Appendix A Figure A2). In case of other supports (Figure 4b), there was only one reflection visible (36.9° 2theta), while AA-D supported catalysts showed the reflections corresponding to clinoptilolite (2theta 9.9° and 11.2°) and other minor phases (clays and feldspars) with no significant intensities corresponding to cobalt oxide. This indicated homogeneous distribution of cobalt on the surface without the formation of a crystalline Co3O4 phase. Crystallite size of Co3O4 calculated using the Scherrer equation from the reflection 36.9° 2theta is 12 nm for catalysts with alumina support. It was not possible to evaluate the crystallite size in other catalysts due to overlap of cobalt oxide diffraction with line assigned to the clinoptilolite, quartz or anorthite.
Analysis of the used catalysts (Appendix A Figure A3), samples labelled with additional -U, catalysts after their usage for N2O decomposition test) did not show any significant changes in the crystalline structure compared with fresh catalysts. Powder diffraction patterns of Al2O3 can be seen in Appendix A Figure A2. After using the catalysts in the reaction test only slight decreases in intensity are noticeable. Only in the case of the AA-D-Co-U catalyst, there was a significant decrease in intensities 2theta 27.8°, 28.0° and others corresponding to anorthite (PDF 71-0748).
Temperature-programmed ammonia desorption (NH3-TPD) showed significant differences in the amount and strength of acid sites (Figure 5a) of each catalyst (Table 2). While the catalyst supported on industrial support (Al2O3-Co) showed low acidity (Appendix A Table A2, Figure A4a), the zeolite foam supported catalysts showed changes in acidity depending on the support modification method during catalyst preparation. In general, peaks at low temperatures, below 200 °C are attributed to weakly acidic sites or physisorbed ammonia, and peaks at higher temperatures in the range of 200-400 °C are associated with the interaction of NH3 with acidic Brønsted sites attributed to strongly acidic sites. Peaks Cobalt catalysts prepared using these supports showed significant crystalline structures corresponding to Co 3 O 4 (2theta 31.5 • , 36.9 • , 45.0 • , 55.7 • , 59.6 • and 65.6 • , PDF 65-3103). These were the most apparent in case of using alumina as support (Appendix A Figure A2). In case of other supports (Figure 4b), there was only one reflection visible (36.9 • 2theta), while AA-D supported catalysts showed the reflections corresponding to clinoptilolite (2theta 9.9 • and 11.2 • ) and other minor phases (clays and feldspars) with no significant intensities corresponding to cobalt oxide. This indicated homogeneous distribution of cobalt on the surface without the formation of a crystalline Co 3 O 4 phase. Crystallite size of Co 3 O 4 calculated using the Scherrer equation from the reflection 36.9 • 2theta is 12 nm for catalysts with alumina support. It was not possible to evaluate the crystallite size in other catalysts due to overlap of cobalt oxide diffraction with line assigned to the clinoptilolite, quartz or anorthite.
Analysis of the used catalysts (Appendix A Figure A3), samples labelled with additional -U, catalysts after their usage for N 2 O decomposition test) did not show any significant changes in the crystalline structure compared with fresh catalysts. Powder diffraction patterns of Al 2 O 3 can be seen in Appendix A Figure A2. After using the catalysts in the reaction test only slight decreases in intensity are noticeable. Only in the case of the AA-D-Co-U catalyst, there was a significant decrease in intensities 2theta 27.8 • , 28.0 • and others corresponding to anorthite (PDF 71-0748).
Temperature-programmed ammonia desorption (NH 3 -TPD) showed significant differences in the amount and strength of acid sites (Figure 5a) of each catalyst (Table 2). While the catalyst supported on industrial support (Al 2 O 3 -Co) showed low acidity (Appendix A Table A2, Figure A4a), the zeolite foam supported catalysts showed changes in acidity depending on the support modification method during catalyst preparation. In general, peaks at low temperatures, below 200 • C are attributed to weakly Catalysts 2020, 10, 1398 8 of 21 acidic sites or physisorbed ammonia, and peaks at higher temperatures in the range of 200-400 • C are associated with the interaction of NH 3 with acidic Brønsted sites attributed to strongly acidic sites. Peaks at temperatures above 400 • C correspond to the interaction of NH 3 with Lewis acids and are attributed to Co-O species that form very strongly acidic sites [54][55][56] AA-N-Co, but no strong acid sites.   1542  173  1313  85  390  229  15  ---AA-D-Co  593  158  446  75  255  70  12  382  77  13  AA-IE-Co  1446  178  1379  95  247  27  2  362  40  3  AA-IE-MnCo  1211  187  1211  100  ------* P = population of the relevant sites, index 1-first deconvolution peak, 2-second deconvolution peak and 3-third deconvolution peak. The basicity of the catalysts was determined by temperature programmed desorption of CO2 (CO2-TPD). The data obtained (Table 3, Figure 5b) showed significant low temperature peaks around 110 °C for ion-exchange treated (AA-N-Co) catalyst, direct ion-exchange treated (AA-IE-Co and AA-IE-CoMn) catalysts and a similar maximum is also observed for the Al2O3-Co catalyst (Appendix A Table A3, Figure A4b). The basicity is related to the alkali content, i.e., alkali metals (Na + K) in the sample ( Table  3). The AA-S-Co catalyst had the highest alkali content, but a large part of the alkali metals was inaccessible to the interaction (they were enclosed in the binder N(A)-A-S-H phase), while in the case of AA-IE-Co and AA-IE-CoMn catalysts which had also higher SSA, alkali metals were more available and the samples showed high overall basicity, especially also the content of weakly basic sites. On the other hand, the basicity of the samples may also be related to the adsorption of CO2 on the polar adsorption centres, i.e., on all coordination unsaturated cations present in the sample.   The highest total acidity exhibited catalyst prepared via impregnation of ion-exchange modified support (AA-N-Co), which contained a high population of strong acid sites, i.e., sites with high NH 3 desorption temperature. The acidity of AA-S-Co and AA-D-Co catalysts is about one-third in comparison and very similar to each other, which is probably due to poor availability of acid sites in the base zeolite-clinoptilolite for AA-S-Co and in case of AA-D-Co due to acid centre removal during acid leaching. Samples prepared using direct ion-exchange treatment showed slightly lower acidity than AA-N-Co, but no strong acid sites.
The basicity of the catalysts was determined by temperature programmed desorption of CO 2 (CO 2 -TPD). The data obtained (Table 3, Figure 5b) showed significant low temperature peaks around 110 • C for ion-exchange treated (AA-N-Co) catalyst, direct ion-exchange treated (AA-IE-Co and AA-IE-CoMn) catalysts and a similar maximum is also observed for the Al 2 O 3 -Co catalyst (Appendix A Table A3, Figure A4b). The basicity is related to the alkali content, i.e., alkali metals (Na + K) in the sample ( Table 3). The AA-S-Co catalyst had the highest alkali content, but a large part of the alkali metals was inaccessible to the interaction (they were enclosed in the binder N(A)-A-S-H phase), while in the case of AA-IE-Co and AA-IE-CoMn catalysts which had also higher SSA, alkali metals were more available and the samples showed high overall basicity, especially also the content of weakly basic sites. On the other hand, the basicity of the samples may also be related to the adsorption of CO 2 on the polar adsorption centres, i.e., on all coordination unsaturated cations present in the sample. * P = population of the relevant sites, index 1-first deconvolution peak, 2-second deconvolution peak and 3-third deconvolution peak.
The results of the H 2 -TPR analysis performed on the catalyst pellets show significant differences in H 2 consumption (Table 4), which corresponds to the content of various forms of Co oxides present on the surface of the catalyst support. Co 3 O 4 reduction takes place in two steps, most often parameters for reduction of Co 3+ to Co 2+ are the temperature range of 250-300 • C and higher than 400 • C for reduction of Co 2+ to Co 0 [9,10]. From the obtained TPR profiles ( Figure 6) it can be seen that the reduction took place in the two steps, but significant differences are observed for the individual catalysts. The results of the H 2 -TPR analysis performed on the Al2O3 can be seen in Appendix A Table A4 and Figure A5. Catalysts 2020, 10, x FOR PEER REVIEW 9 of 22  213  105  42  20  186  161  76  321  10  14  AA-N-Co  125  110  82  66  363  14  11  437  30  23  AA-D-Co  156  112  36  23  204  44  28  289  76  49  AA-IE-Co  306  111  215  70  338  91  30  ---AA-IE-MnCo  408  113  227  56  273  181  44 ---* P = population of the relevant sites, index 1-first deconvolution peak, 2-second deconvolution peak and 3-third deconvolution peak.
The results of the H2-TPR analysis performed on the catalyst pellets show significant differences in H2 consumption (Table 4), which corresponds to the content of various forms of Co oxides present on the surface of the catalyst support. Co3O4 reduction takes place in two steps, most often parameters for reduction of Co 3+ to Co 2+ are the temperature range of 250-300 °C and higher than 400 °C for reduction of Co 2+ to Co 0 [9,10]. From the obtained TPR profiles ( Figure 6) it can be seen that the reduction took place in the two steps, but significant differences are observed for the individual catalysts. The results of the H2-TPR analysis performed on the Al2O3 can be seen in Appendix A Table A4 and Figure A5. * P = population of the relevant sites, index 1-first deconvolution peak, 2-second deconvolution peak and 3-third deconvolution peak. The shift of the reduction peaks may occur due to (i) the presence of alkali [35] which generally lowers temperatures of reduction of cobalt species, or (ii) due to the interaction between the Co3O4 spin The shift of the reduction peaks may occur due to (i) the presence of alkali [35] which generally lowers temperatures of reduction of cobalt species, or (ii) due to the interaction between the Co 3 O 4 spin phase and the support used [51]. In case of the alumina supported catalyst, only one wide reduction peak was evident with two maxima at 435 • C and 479 • C, indicating cobalt species interaction with Al 2 O 3 to form Co aluminates with stronger oxygen bonds that decrease reducibility (Appendix A Table A4, Figure A5) [11]. The catalyst using the AA-S basic support exhibited a wide reduction peak associated with the interaction of Co 3 O 4 with alkali (Na/K), the high content of which comes from the excess alkali activator. AA-N-Co and AA-D-Co catalysts had a wide reduction band at about 300 • C, which in the case of AA-N-Co was accompanied by a second reduction band at a maximum of 483 • C corresponding to the less reducible Co species (bulk Co 3 O 4 ) [1]. The shift of reduction peaks towards the lower temperatures of AA-IE-Co and AA-IE-CoMn catalysts was caused by a different catalyst synthesis process, during which Co and CoMn hydroxides and consequently different cobalt species (CoO(OH)) were created on the support surface than in the IWI method [2]. These catalysts also contained a relatively high potassium content, which also significantly lowers the cobalt particle reduction temperature. The second maximum for the AA-IE-CoMn catalyst at 303 • C was related to the interaction of Co with Mn and the reduction of manganese (Mn 4+ to Mn 3+ ), which was reduced at about 300 • C [3]. TPR profiles showed that deposition of Co 3 O 4 on various supports significantly affects the reducibility, i.e., the strength of the binding of Co 3+ and Co 2+ to oxygen.

N 2 O Catalytic Decomposition
The prepared cobalt catalysts were tested in laboratory units under defined conditions. The temperature dependence of N 2 O conversions using pelletized catalysts are shown in Figure 7a. The results showed significant differences in activity between catalysts due to different properties (chemical composition, textural and acid-base properties, reducibility). The catalyst prepared on a commercial Al 2 O 3 support (Al 2 O 3 -Co) exhibited very low N 2 O conversion (only 7% at 450 • C) (Appendix A Figure A6), even though it had the highest specific surface area of all catalysts tested. As shown by H 2 -TPR, the low catalytic activity was due to the formation of hardly reducible species which cannot participate in N 2 O decomposition. The conversion of N 2 O on cobalt catalysts supported on differently modified zeolite foams increased in following order: AA-D-Co → AA-S-Co → AA-N-Co with N 2 O conversions at 450 • C of 29%, 43% and 81% respectively. The order of zeolite foamed catalysts correlates with data obtained from TPR. Based on NH 3 -TPD analysis it can be observed that the activity was probably also influenced by the presence of acid centres, which in case of AA-N-Co was the highest of all catalysts tested, followed by AA-IE-Co and AA-IE-CoMn, whose activity at N 2 O decomposition was also high (conversion at 450 • C 63% and 79%, respectively). As Obalová et al. have shown, the influence of acidic and basic sites usually begins to apply only when measured under process conditions, which is in the presence of NO and H 2 O. For the decomposition of N 2 O on oxides, acidic and basic centres are generally not considered important [56]. Samples AA-IE-Co and AA-IE-CoMn show a relatively large effect of the manganese promoter on N 2 O decomposition (conversion at 450 • C increases from 63% to 79%). These catalysts contain a cobalt or cobalt-manganese phase in the form of hydroxides and a relatively high content of alkalis, especially potassium, which have a beneficial effect on N 2 O decomposition due to better reducibility of cobalt or cobalt-manganese species, as has also been demonstrated using catalysts based on mixed Co-Mn-Al phases doped by alkaline and prepared from hydrotalcite-like precursors [11,24,27,35].
The different shape of the conversion curves of the AA-S-Co catalyst showed a possible textural effect that has been verified via catalytic tests carried out in a small laboratory reactor using a crushed catalyst at a fraction of 160-315 µm. The order of N 2 O conversions of the catalysts obtained using the pellets (Figure 7a) and the crushed catalysts (Figure 7b) was different, indicating that the reaction kinetics were influenced by the textural properties of the catalysts and the internal diffusion had a significant effect. N 2 O conversion under the same reaction conditions was used to calculate reaction rates. Assuming that the rate of reaction on the crushed catalyst was not affected by internal diffusion, it can be used to compare the effect of diffusion in pellets [51]. The kinetic constant k (m 3 s -1 kg -1 ) for pellets and crushed catalyst was calculated from the conversion of N 2 O (X N2O ) at 450 • C, according to the relation (1): where m (kg) is the mass of the active phase in the pellet or crushed catalyst and . V (m 3 s −1 ) is the total gas flow. The k pellet /k grain ratio calculated from experimental data indicates the effect of internal diffusion. This ratio was only 0.12 for the AA-S-Co catalyst, indicating a significant internal diffusion effect. In contrast, catalysts on modified foamed supports basis AA-N-Co (k pellet /k grain ratio = 0.99), AA-D-Co (k pellet /k grain ratio = 1.06) showed that the effect of internal diffusion was very small or virtually none. The method of preparation of the AA-S-Co catalyst does not make it possible to eliminate these diffusion effects by changing the synthesis procedure. This can be achieved by further post-synthesis modification of the prepared alkali-activated zeolite foam by ion exchange and acid leaching. Although the AA-S-Co catalyst shows the best results for temperatures up to 350 • C, the AA-N-Co catalyst is the best for applications, followed by the AA-D-Co catalyst, even though they are less active in the kinetic mode. This confirms that it is not enough for the active catalyst to have only an effective active phase, but also the structural and textural properties, which are manifested only in shaped samples. The Al 2 O 3 -Co catalyst showed no activity under the selected conditions and therefore it was not possible to determine the effect of internal diffusion. As shown on SEM-EDX, another possible explanation may be the inhomogeneity in terms of the distribution of cobalt and potassium in the pellet. For some samples (e.g., AA-S-Co) the results may be skewed because after grinding only a small amount is used (the composition of the catalyst for pellet and powder tests may therefore differ).
Catalysts 2020, 10, x FOR PEER REVIEW 11 of 22 where m (kg) is the mass of the active phase in the pellet or crushed catalyst and V̇ (m 3 s −1 ) is the total gas flow. The kpellet/kgrain ratio calculated from experimental data indicates the effect of internal diffusion. This ratio was only 0.12 for the AA-S-Co catalyst, indicating a significant internal diffusion effect. In contrast, catalysts on modified foamed supports basis AA-N-Co (kpellet/kgrain ratio = 0.99), AA-D-Co (kpellet/kgrain ratio = 1.06) showed that the effect of internal diffusion was very small or virtually none. The method of preparation of the AA-S-Co catalyst does not make it possible to eliminate these diffusion effects by changing the synthesis procedure. This can be achieved by further post-synthesis modification of the prepared alkali-activated zeolite foam by ion exchange and acid leaching. Although the AA-S-Co catalyst shows the best results for temperatures up to 350 °C, the AA-N-Co catalyst is the best for applications, followed by the AA-D-Co catalyst, even though they are less active in the kinetic mode. This confirms that it is not enough for the active catalyst to have only an effective active phase, but also the structural and textural properties, which are manifested only in shaped samples. The Al2O3-Co catalyst showed no activity under the selected conditions and therefore it was not possible to determine the effect of internal diffusion. As shown on SEM-EDX, another possible explanation may be the inhomogeneity in terms of the distribution of cobalt and potassium in the pellet. For some samples (e.g., AA-S-Co) the results may be skewed because after grinding only a small amount is used (the composition of the catalyst for pellet and powder tests may therefore differ). To compare and explain the differences in catalytic activities of pellets and powders, the rate constants for the individual catalysts and temperatures were calculated and then linearized using Arrhenius Equation (2): where y = a + bx the individual components of the equation correspond to the general regression line: lnk = y, lnA = a, −(EA/T) = b a 1/T = x, R is gas constant (J mol −1 K −1 ), A is pre-exponential factor and T is the temperature (K) [39]. From the values thus obtained, activation energies (EA) of N2O decomposition for individual catalysts were determined (Table 5). N2O is a stable molecule and decomposes only at high temperatures (above 1000 °C), the activation energy of thermal decomposition of N2O is about 250 kJ/mol [5]. Due to the catalyst, the decomposition temperature or EA (Table 5) drops very significantly and when using cobalt catalysts based on alkali-activated zeolite foams, it ranges from 1 /3 to ¼ of this To compare and explain the differences in catalytic activities of pellets and powders, the rate constants for the individual catalysts and temperatures were calculated and then linearized using Arrhenius Equation (2): where y = a + bx the individual components of the equation correspond to the general regression line: lnk = y, lnA = a, −(E A /T) = b a 1/T = x, R is gas constant (J mol −1 K −1 ), A is pre-exponential factor and T is the temperature (K) [39]. From the values thus obtained, activation energies (E A ) of N 2 O decomposition for individual catalysts were determined (Table 5). N 2 O is a stable molecule and decomposes only at high temperatures (above 1000 • C), the activation energy of thermal decomposition of N 2 O is about 250 kJ/mol [5]. Due to the catalyst, the decomposition temperature or E A (Table 5)    The ratio of activation energies of pellets and crushed catalysts (Table 5), which is in the range of 1.5 to 2.3 indicates that the effect of internal diffusion is dominant even in the case of the AA-S-Co catalyst, where the effectiveness factor was only 0.12. The obtained data confirmed our original assumption that the catalysts operate in the kinetic region and this discrepancy is rather due to the inhomogeneity of the sample in terms of the distribution of cobalt and potassium in the pellet. The inhomogeneity of the pellet and the small weight of catalyst used for the crushed catalyst tests led to skewed results.
The results obtained on cobalt catalysts using zeolite foam as support (AA-N-Co, AA-IE-Co, AA-IE-CoMn) are comparable to the results presented on other types of catalysts, e.g., CoMnAl mixed oxide catalysts prepared from a hydrotalcite precursor are very efficient and show similar results under the same conditions, but as other authors have shown in their studies of catalysts, other parameters such as Co/Mn ratio or type and content of alkalis (Na, K, Cs) need to be sensitively optimized [11,27]. In different studies, the same authors also dealt with catalysts using foam or ceramic supports.
The synthesized AA-IE-Co and AA-IE-CoMn catalysts prepared on zeolite foam have significantly higher N 2 O conversion (3×-4× higher at temperatures up to 390 • C and 2× higher at temperatures 420-450 • C) in comparison with catalysts prepared on SiC foams with an active layer consisting of CoMnAl mixed oxide (hydrotalcite precursor, Co content about 6 wt.%), under the same conditions [43].
Likewise, the cobalt phase on a cordierite monolith provided with a washcoat layer of CeO 2 -ZrO 2 showed about 2× lower conversion of N 2 O even though the Co content in the catalyst was almost doubled (13.5 wt.%) [49]. Similar results as AA-IE-Co (under the same test conditions) were shown by a cobalt catalyst prepared by repeated impregnation of a MgO/Al 2 O 3 support. Again, however, this catalyst had a significantly higher Co content (25 wt.%) [51]. Cobalt catalysts on beta zeolite (2 wt.% of Co) showed slightly better results under similar conditions, which is due to the better distribution of Co on the surface, as these catalysts have 4 to 6 times higher specific surface area (500-750 m 2 /g) [52].
Catalytic tests show that the active phase on a support with a high specific surface area (γ-Al 2 O 3 ) may not be the most advantageous solution due to the presence of poorly reducible cobalt species. A suitable alternative may be macroporous foam supports based on alkali-activated natural zeolites with suitable post-synthesis modification or a suitable process of catalyst synthesis. Advantage of these materials was also their ease of synthesis and low cost, meeting the industry's requirements for later large-scale application. In further research, it will be possible to increase the catalytic activity of the cobalt-based active phase applied to the foam support by other promoters such as caesium or by adding other elements such as manganese and cerium. It will be important to optimize their amount and, last but not least, to optimize the actual production processes of the zeolite foam to obtain an optimal distribution of the active phase on its surface. Subsequent long-term testing of the most active catalysts will, in addition to the stability of the catalyst, also monitor the process of deactivation and eventual reactivation of the catalyst and with the contribution of advanced analytical techniques such as X-Ray Photoelectron Spectroscopy (XPS), Transmission Electron Microscopy (TEM) or Auger Electron Spectroscopy (AES), it will be possible to characterize Co species in detail on the catalyst surface and elucidate reaction and deactivation mechanisms.
ZeoCem Eco 50 contains at least 85 wt.% of clinoptilolite and minor mineral components: clays, feldspar, mica. The chemical composition and specific surface areas are shown in Table 6.

Synthesis of Zeolite Foams
Zeolite foam pellets were obtained by the same procedure described in our previous publication [8]. Natural zeolite ECO 50 was activated with a mixed alkaline activator (KOH solution with a sodium silicate) and foamed by hydrogen peroxide solution. Mixture parameters are shown in Table 7. Shaped foam pellets (5 mm diameter) were activated for 48 h at 50 • C. Prepared pellets were aged under laboratory conditions in a closed plastic bag for 1 month prior to further processing. A prepared sample of alkali-activated zeolite foam (AZF) is designated "AA-S".

Post-Synthesis Modifications and Catalyst Synthesis
The foam pellets AA-S were first dried overnight at 120 • C, then, part of them was leached at 80 • C using 3M HCl for 6 h (AA-D). The second part of pellets was ion-exchanged by 1M solution of NH 4 NO 3 at 80 • C for 6 h (AA-N). The foam to solution ratio (g/mL) was 1:20 for both modified supports. The cobalt catalysts were synthesized by IWI method with aqueous cobalt nitrate solution on target cobalt content of 5 wt.% (AA-S-Co, AA-N-Co and AA-D-Co). The comparative Al 2 O 3 -Co catalyst using γ-Al 2 O 3 as catalyst support was prepared by the same methods. Other catalyst samples were synthesized by DIE method of AA-S pellets at 80 • C (6 hrs) by 0.25 M solution of cobalt nitrate (AA-IE-Co) or by 0.2 M cobalt nitrate with 0.05 M manganese nitrate respectively (AA-IE-CoMn). All samples (IWI + DIE) were dried overnight at 120 • C and calcined at 500 • C for 4 hrs (1 • C/min). An overview of prepared catalysts is given in Table 8.

Characterisation of Supports and Prepared Catalysts
Materials in form of pellets were used for all methods except the XRD. The X-ray diffraction patterns (XRD) were evaluated by D8 Advance ECO (Bruker AXS GmbH, Karlsruhe, Germany) instrument applying CuKα radiation (λ = 1.5406 Å) with the resolution of 0.02 • in the 2θ range of 5-70 • . The diffractograms were evaluated using the Diffrac.Eva software with the Powder Diffraction File database (PDF 4 + 2018, International Centre for Diffraction Data). The chemical composition of supports and catalysts was determined by X-ray fluorescence analysis (XRF) of powder materials using an S8 Tiger (Bruker AXS GmbH, Karlsruhe, Germany) with Rh cathode.
The N 2 physisorption and Hg porosimetry (MIP) were used for textural properties determination. Prior to the analysis, all samples were dried under vacuum in a glass-cell at 110 • C (200 • C for calcined materials) for 16 h. The specific surface area was determined from N 2 sorption isotherms at −196 • C, measured by Autosorb iQ (Quantachrome Instruments, Boynton Beach, FL, USA). Meso and macroporous properties of catalysts and supports were determined by MIP measurements performed on the AutoPore IV 9510 Hg porosimeter (Micromeritics Instrument Corporation, Norcross, GA, USA). The samples were prepared for the analyses in the same way as for the physisorption measurements.
For scanning electron microscopy (SEM) analysis, samples were coated with about 5 nm of gold to make them conductive under the microscope (JEOL JSM-IT500HR, Tokyo, Japan) accessorized with energy-dispersive X-ray spectroscopy (EDS) for elemental analysis and map analysis. Representative backscattered electron or secondary electron images of microstructures were taken in high vacuum mode using an accelerating voltage of 15 kV.
The reducibility of the cobalt species in the catalysts was determined by H 2 -TPR using an AutoChem 2950 HP (Micromeritics Instrument Corporation, Norcross, GA, USA) instrument. The pellets of the sample in a quartz U-tube were heated in He flow at 450 • C before the measurement. The reduction was carried out at t = 40-500 • C with a temperature ramp of 10 • C/minute under reducing gas flow (10 vol.% H 2 in Ar). The changes in hydrogen concentration were monitored by a TCD detector. Acid properties of the samples were characterized by NH 3 and CO 2 temperature-programmed desorption (NH 3 -TPD or CO 2 -TPD) using the same instrument as for TPR. Prior to TPD experiment, the sample was pre-treated by heating to 450 • C (heating rate 10 • C/min; kept at target temperature for 30 min) in the flow of He (flow rate 25 mL/min). After reaching 450 • C, the sample was cooled down to 100 • C (NH 3 -TPD) or 50 • C (CO 2 -TPD). Then, the sample was saturated with NH 3 or CO 2 by flowing the sample bed by a gas mixture containing 90 vol.% of He and 10 vol.% of NH 3 or CO 2 (flow rate 25 mL/min) for 30 min. Weakly bonded NH 3 or CO 2 molecules were removed by flushing the sample out under flow of helium (25 mL/min) for 15 min. Finally, NH 3 -TPD experiment was carried out by increasing the temperature from 100 to 500 • C (CO 2 -TPD of the temperature from 50 to 500 • C) with heating rate 10 • C/min. The desorbed NH 3 /CO 2 in the outlet gas was detected by TCD detector. The obtained curves were evaluated by deconvolution using a SW device.

Catalytic Tests
Catalytic decomposition of N 2 O was performed in integral stainless-steel fixed bed reactor at the temperature range from 300 to 450 • C and at atmospheric pressure. The feed introduced to the reactor contained 0.1 mol % N 2 O in N 2 and IR analyser (Sick) was used for N 2 O analysis. Testing of the grained catalysts was performed in a reactor with an internal diameter of 5.5 mm and shaped catalysts (foam and pellets) were tested in a reactor with an internal diameter of 50 mm. Grained catalyst bed contained 0.3 g of sample with a particle size of 160-315 µm and 100 mL/min, GHSV of 20 m 3 kg −1 h −1 were used. Pelleted catalyst bed contained 20 mL of sample and 1000 mL/min, GHSV of 3 000 m 3

Conclusions
Cobalt catalysts prepared by various techniques on alkali-activated zeolite foams have been characterized in detail and tested in catalytic decomposition of N 2 O. For comparison, catalyst was prepared on commercial γ-Al 2 O 3 support with identical content of cobalt phase (5 wt.%). Tests were also performed on crushed catalysts to determine the effect of internal diffusion. Despite the significantly higher specific surface area of the catalyst on γ-Al 2 O 3 support, catalysts on zeolite foams had significantly higher N 2 O conversions. The reason is that hardly reducible cobalt phases were formed on γ-Al 2 O 3 , while the zeolite foams did not have such a strong internal diffusion effect and besides, due to the presence of alkali, especially potassium, the cobalt phase had better N 2 O decomposition properties. Manganese doping of the cobalt catalyst further improved the catalyst properties and the N 2 O conversion increased by 25%. The results obtained show the great potential of these catalysts on alkali-activated zeolite foams in general and move them closer to further use in industrial applications.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. * AM = alkali metals (Na + K), ** SSA = specific surface area (BET). * P = population of the relevant sites, index 1-first deconvolution peak, 2-second deconvolution peak and 3-third deconvolution peak. Figure A1. Pore distribution of Al2O3 support, fresh and used cobalt catalyst. Figure A1. Pore distribution of Al 2 O 3 support, fresh and used cobalt catalyst.