The Role of Ruthenium in CO 2 Capture and Catalytic Conversion to Fuel by Dual Function Materials ( DFM )

Development of sustainable energy technologies and reduction of carbon dioxide in the atmosphere are the two effective strategies in dealing with current environmental issues. Herein we report a Dual Function Material (DFM) consisting of supported sodium carbonate in intimate contact with dispersed Ru as a promising catalytic solution for combining both approaches. The Ru-Na2CO3 DFM deposited on Al2O3 captures CO2 from a flue gas and catalytically converts it to synthetic natural gas (i.e., methane) using H2 generated from renewable sources. The Ru in the DFM, in combination with H2, catalytically hydrogenates both adsorbed CO2 and the bulk Na2CO3, forming methane. The depleted sites adsorb CO2 through a carbonate reformation process and in addition adsorb CO2 on its surface. This material functions well in O2and H2O-containing flue gas where the favorable Ru redox property allows RuOx, formed during flue gas exposure, to be reduced during the hydrogenation cycle. As a combined CO2 capture and utilization scheme, this technology overcomes many of the limitations of the conventional liquid amine-based CO2 sorbent technology.


Introduction
As predicted by the Intergovernmental Panel on Climate Change (IPCC), carbon dioxide in the atmosphere may reach 570 ppmv by the year of 2100, accompanied by a rise in the mean global temperature of 1.9 • C and mean sea level by 38 cm [1][2][3].Excess carbon dioxide emission results primarily from anthropogenic activities by fossil fuel combustion for power generation, industrial manufacturing, and transportation [4].Due to the CO 2 accumulation and the rising demand of energy, sustainable energy development [5,6], and CO 2 capture, storage and utilization (CCSU) [7] are two efficient methods to combat green-house gas emissions and associated climate change.
For a decade, various processes and materials have been proposed for CO 2 capture, with extensive reviews available [8,9].The aqueous amine-based sorbent technology is still the most popular and developed approach toward capturing CO 2 from flue gas.However, the sorbent regeneration for separation of CO 2 and H 2 O from the amine solution by distillation is highly energy intensive.The separated CO 2 must then be transported to a recovery site for processing, which adds to the expense.Furthermore, the O 2 containing flue gas deteriorates the amine, limiting its ability to be recycled [10].
Recently, several unique material combinations and processes for upgrading CO 2 have been reported.A super-dry reforming process involving chemical looping in which CO 2 is utilized for enhancing CO production at 750 • C has been reported.Nickel catalyzes the dry reforming of CH 4 leading to a redox reaction of iron oxides.CO 2 is adsorbed on CaO/Al 2 O 3 , providing a series of equilibrium shifts enhancing CO production for ultimate use in synthesis gas reactions [11].Another interesting approach for enhanced CO production combines a reduced Fe, Cr, Cu reverse water gas shift (RWGS) catalyst with K 2 CO 3 as a CO 2 adsorbent.Approximately 6% CO 2 in N 2 is adsorbed on K 2 CO 3 at 550 • C. At the same temperature, injected H 2 reacts with adsorbed CO 2 generating CO and H 2 O via the RWGS reaction.A weakness in the process is the deterioration of performance in O 2 , limiting this to O 2 -free CO 2 streams [12].
Ruthenium is a well-known noble metal catalyst often used commercially for homogeneous or heterogeneous catalytic reactions requiring a high degree of product selectivity, including aromatic ring saturation to cyclo-hexyl derivatives, saturation of allyl compounds, organic acid hydrogenation to alcohols, hydrogenation of sugars [13], steam reforming of hydrocarbons for H 2 production [14], selective homogeneous catalysis for the oxidation of primary alcohols to aldehydes [15], and homogenous chiral synthesis [16].It is also used for ammonia synthesis [17], preferential oxidation of CO in H 2 [18], Fischer-Tropsch processes [19], electrochemical chlorine evolution [20], and methanation of CO and CO 2 [21][22][23].In representative reactions, such as those mentioned above, all require either the metal, its oxide, or an organo-metal complex form of Ru as the catalyst.A unique feature of Ru/RuO x is its redox chemistry of rapid and reversible transformation between its oxide and metallic states.
We are now reporting a work-in-progress combining a series of adsorption and catalytic reactions for converting adsorbed CO 2 emitted from power plant oxygen-containing flue gas where the redox property of Ru/RuO x is critical for successful use.The material used for this process is composed of a nano-dispersed [24] alkaline component in intimate contact with Ru on a high surface area carrier (Al 2 O 3 ), referred to as a Dual Function Material (DFM).It adsorbs CO 2 on its alkaline component and catalytically converts it to synthetic natural gas (i.e., CH 4 ) in the subsequent step over the Ru catalyst approaching a closed loop for carbon utilization.The sequential steps will be discussed below.
Alkaline carbonates have been reported to be efficient for CO 2 adsorbents [25].What is the most relevant to our new materials is the ability to adsorb CO 2 and then hydrogenate catalytically to methane.The catalytic hydrogenation of a series of carbonates was first reported by Tsuneto et al. in 1992 [26] and Yoshida et al. in 1999 [27].A variety of transition metal catalysts (but not Ru) were physically mixed with the carbonates and hydrogenation was conducted.The carbonate was completely hydrogenated but there was no attempt to reform the carbonate by CO 2 capture.
We have demonstrated that a DFM, containing 5% Ru, 10% Na 2 CO 3 /Al 2 O 3 , captures CO 2 both on the carbonate surface and depleted bulk carbonate sites, which are subsequently hydrogenated to CH 4 .In a typical implementation of this material, the methane produced is dried and compressed and recycled to the inlet of the power plant or integrated into the natural gas grid, ideally closing the carbon cycle.Thus, this material provides an additional source of sites for CO 2 capture by carbonate reformation, yielding an increase in subsequent methane production.For this scheme to be used, H 2 used for methanation needs to be generated from water electrolysis, the energy of which is provided by excess solar or wind energy [22].Installations for water electrolysis powered by solar energy for H 2 vehicle fueling stations has been announced between Guangdong Synergy Hydrogen Power Technology (Guangdong, China) and Proton On Site (Wallingford, CT, USA).Installations in Foshan and Yunfu, cities in China, will be initial locations [28].Thus it is expected that renewable H 2 will be available in selected locations.
The two-step sequential reactions involve first the selective adsorption of CO 2 from O 2 -containing flue gas on the DFM.During this step, the Ru in the DFM is oxidized to RuO x .Upon CO 2 saturation, the flue gas is diverted to an identical parallel reactor where CO 2 capture continues.Renewable H 2 is added to the reactor containing the saturated DFM where it rapidly reduces the RuO x to its metallic and active state, and methanation proceeds via CO 2 spill over to the Ru sites.All steps are conducted at the same temperature (~320 • C) [29].Therefore, the reversibility of the redox chemistry of Ru/RuO x makes it unique for this process relative to commonly used Ni, which when oxidized will not reduce under the conditions of this process [30][31][32].This paper reports cyclic aging studies from a simulated flue gas demonstrating stable repeatable cycles of CO 2 adsorption and methanation.The importance of H 2 exposure intensity will also be discussed.The price of Ru should not be confused the more expensive among precious metal catalysts.Its price per gram as of 21 December 2016, was Ru $1.25, Rh $24.53, Pt $28.40, and Pd $20.60 with all recycled for continued use.

Na 2 CO 3 Loading Study
Various carbonate-containing materials were evaluated on a thermal gravimetric analysis (TGA) for reversible CO 2 adsorption.The CO 2 adsorbed and methanation in mmol per gram on Al 2 O 3 supported K 2 CO 3 , Na 2 CO 3 , and MgO powder adsorbents with weight loadings from 5% to 20% were previously reported [30].In this study K 2 CO 3 and Na 2 CO 3 showed the largest CO 2 adsorption relative to Al 2 O 3 supported CaO and MgO.10% Na 2 CO 3 on Al 2 O 3 achieved the highest adsorption capacity (0.11 mmol/g), and was therefore selected for additional adsorption studies.Figure 1 shows TGA results for CO 2 adsorption and reversible desorption for dispersed Na 2 CO 3 on Al 2 O 3 at different loadings.This paper reports cyclic aging studies from a simulated flue gas demonstrating stable repeatable cycles of CO2 adsorption and methanation.The importance of H2 exposure intensity will also be discussed.The price of Ru should not be confused the more expensive among precious metal catalysts.Its price per gram as of 21 December 2016, was Ru $1.25, Rh $24.53, Pt $28.40, and Pd $20.60 with all recycled for continued use.

Na2CO3 Loading Study
Various carbonate-containing materials were evaluated on a thermal gravimetric analysis (TGA) for reversible CO2 adsorption.The CO2 adsorbed and methanation in mmol per gram on Al2O3 supported K2CO3, Na2CO3, and MgO powder adsorbents with weight loadings from 5% to 20% were previously reported [30].In this study K2CO3 and Na2CO3 showed the largest CO2 adsorption relative to Al2O3 supported CaO and MgO.10% Na2CO3 on Al2O3 achieved the highest adsorption capacity (0.11 mmol/g), and was therefore selected for additional adsorption studies.Figure 1 shows TGA results for CO2 adsorption and reversible desorption for dispersed Na2CO3 on Al2O3 at different loadings.Figure 1 indicates 10 wt % Na2CO3/Al2O3 obtained the highest capacity of CO2 adsorption (0.11 mmol/g) among other Na2CO3 loadings.The pure N2 purge verifies that Na2CO3 dispersed on γ-Al2O3 powder has the ability of desorb the CO2 reversibly in situ, consistent with dispersed CaO/Al2O3 substantiating that the binding interaction between CO2 and adsorbents is weak [33].This property insures the desorption and spill-over of CO2 to Ru catalytic sites for methanation.

Carbonate Hydrogenation Study
A H2 activation (reduction) step was conducted using unsupported Ru-Na2CO3 to understand the decomposition profile of Ru nitrosyl nitrate used in the powder preparation procedure.However, a greater weight loss than expected was observed indicating hydrogenation of the carbonate in the sample was also occurring.Results all conducted at 320 °C using the TGA is shown in Figure 2 in which the final step shows CO2 adsorption.Figure 1 indicates 10 wt % Na 2 CO 3 /Al 2 O 3 obtained the highest capacity of CO 2 adsorption (0.11 mmol/g) among other Na 2 CO 3 loadings.The pure N 2 purge verifies that Na 2 CO 3 dispersed on γ-Al 2 O 3 powder has the ability of desorb the CO 2 reversibly in situ, consistent with dispersed CaO/Al 2 O 3 substantiating that the binding interaction between CO 2 and adsorbents is weak [33].This property insures the desorption and spill-over of CO 2 to Ru catalytic sites for methanation.
Based on the methanation capacities, DFM composed of 5% Ru, 10% Na 2 CO 3 /Al 2 O 3 generated the most methane (1.05 mol CH 4 /kg of DFM) compared to other adsorbents at the same loading [30].Herein, further studies were conducted on Al 2 O 3 supported Ru-Na 2 CO 3 DFM.

Carbonate Hydrogenation Study
A H 2 activation (reduction) step was conducted using unsupported Ru-Na 2 CO 3 to understand the decomposition profile of Ru nitrosyl nitrate used in the powder preparation procedure.However, a greater weight loss than expected was observed indicating hydrogenation of the carbonate in the sample was also occurring.Results all conducted at 320 • C using the TGA is shown in Figure 2 in which the final step shows CO 2 adsorption.During the sample preparation, the 50 mg of 1.44% Ru/Na2CO3 was dried at 120 °C for 2 h and calcined at 250 °C for 2 h in air.When the precursor Ru(NO)(NO3)3 is heated, it has been reported that a mixture of oxides such as RuO, RuO2, RuO3, and RuO4 [33,34] form depending on the temperature.
After a one-hour N2 purge at 320 °C, the sample achieved stable weight as water and/or other volatile components evaporated.At this point, 2% H2/N2 was introduced.During the two-hour H2 exposure, Ru is expected to reduce, nitrate decompose, and some carbonate hydrogenate.The stable weight decreased by 1.763 mg.Assuming the Ru(NO)(NO3)3 remained in its original form as it is in the precursor, the expected weight loss would be 1.332 mg or 65.37% of the total mass loss.Therefore, the residual mass loss (an additional 34.63%) can be explained by carbonate hydrogenation catalyzed by the newly formed Ru metal.It is clear from Figure 2 that a continuing weight loss likely occurs due to the slow hydrogenation of the carbonate.After about 120 min of pretreatment, the sample showed an ability to re-adsorb CO2 in depleted sites, replacing some of that which was hydrogenated.

Hydrogen Exposure Time Study
50 mg of 5% Ru, 10% Na2CO3/Al2O3 powder sample was subjected to a one-hour pretreatment in 2% H2/N2 at 320 °C.In Figure 3, a rapid adsorption of CO2 up to 1.2 mg from a feed gas of 5% CO2 in N2 was observed.Upon the addition of 2% H2/N2 three distinct slopes of weight loss vs time are observed.The largest slope, between 30 and 35 min, was 0.086 mg/min.The second between roughly 35 and 70 min was 0.00952 mg/min and the third about 0.00195 mg/min from 70 to 660 min.The first loss, associated with the largest slope, occurs from 1.2 mg to 0.66 mg is likely due to methanation of the CO2 adsorbed on the Ru metal.This weight loss is approximately 50% of the Ru atoms present assuming 1 CO2 adsorbed per Ru site.It is also possible 1 CO2 adsorbs on two Ru sites for dissociative adsorption to CO and atomic O [23].The second slope has a weight loss of about 0.45 mg which is twice that observed for CO2 adsorbed on 10% Na2CO3/Al2O3 (see Figure 1) suggesting that additional sites for CO2 adsorption have become available due to the depleted sites generated by the hydrogenation of the bulk carbonate itself upon the addition of H2.The second slope is speculated to be the hydrogenation weakly adsorbed CO2 on the carbonate surface most of which has spilled over to the Ru sites where the heat liberated (not shown) is indicative of methanation.The weight loss associated with the lowest slope is less than 0.2 mg as is believed to be hydrogenation of the bulk carbonate.This is speculative, since Figure 1 was generated with a feed gas of only 0.5% CO2/N2 while Figure 3 was conducted with 5% CO2/N2.However, CO2 chemisorption reaches a plateau where it becomes independent of partial pressure.During the sample preparation, the 50 mg of 1.44% Ru/Na 2 CO 3 was dried at 120 • C for 2 h and calcined at 250 • C for 2 h in air.When the precursor Ru(NO)(NO 3 ) 3 is heated, it has been reported that a mixture of oxides such as RuO, RuO 2 , RuO 3 , and RuO 4 [33,34] form depending on the temperature.
After a one-hour N 2 purge at 320 • C, the sample achieved stable weight as water and/or other volatile components evaporated.At this point, 2% H 2 /N 2 was introduced.During the two-hour H 2 exposure, Ru is expected to reduce, nitrate decompose, and some carbonate hydrogenate.The stable weight decreased by 1.763 mg.Assuming the Ru(NO)(NO 3 ) 3 remained in its original form as it is in the precursor, the expected weight loss would be 1.332 mg or 65.37% of the total mass loss.Therefore, the residual mass loss (an additional 34.63%) can be explained by carbonate hydrogenation catalyzed by the newly formed Ru metal.It is clear from Figure 2 that a continuing weight loss likely occurs due to the slow hydrogenation of the carbonate.After about 120 min of pretreatment, the sample showed an ability to re-adsorb CO 2 in depleted sites, replacing some of that which was hydrogenated.

Hydrogen Exposure Time Study
50 mg of 5% Ru, 10% Na 2 CO 3 /Al 2 O 3 powder sample was subjected to a one-hour pretreatment in 2% H 2 /N 2 at 320 • C. In Figure 3, a rapid adsorption of CO 2 up to 1.2 mg from a feed gas of 5% CO 2 in N 2 was observed.Upon the addition of 2% H 2 /N 2 three distinct slopes of weight loss vs time are observed.The largest slope, between 30 and 35 min, was 0.086 mg/min.The second between roughly 35 and 70 min was 0.00952 mg/min and the third about 0.00195 mg/min from 70 to 660 min.The first loss, associated with the largest slope, occurs from 1.2 mg to 0.66 mg is likely due to methanation of the CO 2 adsorbed on the Ru metal.This weight loss is approximately 50% of the Ru atoms present assuming 1 CO 2 adsorbed per Ru site.It is also possible 1 CO 2 adsorbs on two Ru sites for dissociative adsorption to CO and atomic O [23].The second slope has a weight loss of about 0.45 mg which is twice that observed for CO 2 adsorbed on 10% Na 2 CO 3 /Al 2 O 3 (see Figure 1) suggesting that additional sites for CO 2 adsorption have become available due to the depleted sites generated by the hydrogenation of the bulk carbonate itself upon the addition of H 2 .The second slope is speculated to be the hydrogenation weakly adsorbed CO 2 on the carbonate surface most of which has spilled over to the Ru sites where the heat liberated (not shown) is indicative of methanation.The weight loss associated with the lowest slope is less than 0.2 mg as is believed to be hydrogenation of the bulk carbonate.This is speculative, since Figure 1 was generated with a feed gas of only 0.5% CO 2 /N 2 while Figure 3 was conducted with 5% CO 2 /N 2 .However, CO 2 chemisorption reaches a plateau where it becomes independent of partial pressure.For the 2nd cycle, the mass gain for CO2 adsorption was similar to the first cycle.However, with only 30 min of hydrogen exposure, the weight loss was significantly less than after the longer time exposure in cycle 1.In the third cycle it is likely only the Ru metal and the surface of Na2CO3 adsorb CO2 since insufficient hydrogenation of the bulk carbonate had occurred with only a 30-min exposure.It is clear that bulk carbonate hydrogenation is dependent on exposure time and hydrogen partial pressure.This will be demonstrated in the following studies.

10-Cycle Test in CO2 Feed
Although the primary commercial target of our research is focused on power plant exhaust gases containing O2 and steam, there are applications where they are not present, such as fermentation, land fill, and cement manufacturing effluents.Therefore, an aging test consisting of 10 cycles using 7.5% CO2/N2 for the adsorption was conducted in our scaled up reactor with DFM pellets.This is more consistent with industrial processes since fine powdered materials create operational, technical and health problems for workers.For example, in a fixed bed reactor, powders have high pressure drop [35] and expose operators to possible risk of chronic human health issues such as respiratory problems and lung cancer [36].In order to avoid such issues and facilitate commercialization, a process was conducted with 10 grams of freshly prepared pelleted DFM (5 mm × 5 mm tablets of γ-Al2O3) and cycle tested in a larger scale fixed bed reactor.
The results for CO2 adsorption and methanation under O2-free conditions are shown in Figure 4. Given the slow kinetics for bulk carbonate hydrogenation experienced in Figure 3, the H2 was increased to 5% H2/N2.The Ru-Na2CO3 DFM achieved stable adsorption and methanation performance for 10 successive cycles and is expected to maintain this advantage over many more cycles.However, the average conversion efficiency of the adsorbed CO2 converted to methane was only 60.5%.A large amount of desorbed CO2 was observed at the reactor outlet.The 39.5% unreacted CO2 released is likely due to its premature desorption caused by the large exotherm of methanation.We intend to address this by performing parametric studies for feed gas components (i.e., H2 partial pressure) and process variations.It is also logical to locate a small Ru catalyst downstream to complete methanation For the 2nd cycle, the mass gain for CO 2 adsorption was similar to the first cycle.However, with only 30 min of hydrogen exposure, the weight loss was significantly less than after the longer time exposure in cycle 1.In the third cycle it is likely only the Ru metal and the surface of Na 2 CO 3 adsorb CO 2 since insufficient hydrogenation of the bulk carbonate had occurred with only a 30-min exposure.It is clear that bulk carbonate hydrogenation is dependent on exposure time and hydrogen partial pressure.This will be demonstrated in the following studies.

10-Cycle Test in CO 2 Feed
Although the primary commercial target of our research is focused on power plant exhaust gases containing O 2 and steam, there are applications where they are not present, such as fermentation, land fill, and cement manufacturing effluents.Therefore, an aging test consisting of 10 cycles using 7.5% CO 2 /N 2 for the adsorption was conducted in our scaled up reactor with DFM pellets.This is more consistent with industrial processes since fine powdered materials create operational, technical and health problems for workers.For example, in a fixed bed reactor, powders have high pressure drop [35] and expose operators to possible risk of chronic human health issues such as respiratory problems and lung cancer [36].In order to avoid such issues and facilitate commercialization, a process was conducted with 10 grams of freshly prepared pelleted DFM (5 mm × 5 mm tablets of γ-Al 2 O 3 ) and cycle tested in a larger scale fixed bed reactor.
The results for CO 2 adsorption and methanation under O 2 -free conditions are shown in Figure 4. Given the slow kinetics for bulk carbonate hydrogenation experienced in Figure 3, the H 2 was increased to 5% H 2 /N 2 .The Ru-Na 2 CO 3 DFM achieved stable adsorption and methanation performance for 10 successive cycles and is expected to maintain this advantage over many more cycles.However, the average conversion efficiency of the adsorbed CO 2 converted to methane was only 60.5%.A large amount of desorbed CO 2 was observed at the reactor outlet.The 39.5% unreacted CO 2 released is likely due to its premature desorption caused by the large exotherm of methanation.We intend to address this by performing parametric studies for feed gas components (i.e., H 2 partial pressure) and process variations.It is also logical to locate a small Ru catalyst downstream to complete methanation of the unconverted CO 2 .Conditions of our experiments were controlled to avoid CO formation.CO is only produced at higher temperatures and space velocities over Ru as shown in reference [23].

12-Cycle Test in Simulated Flue Gas
Performance of the pelleted Ru-Na2CO3 DFM in simulated O2-and-steam-containing flue gas is another key factor for its commercialization and application.A 12-cycle test was conducted in the fixed bed reactor with the performance shown in Figure 5.The results of BET specific surface area of the fresh and after-12-cycle Ru-Na2CO3 pelleted DFM sample is shown in Table 1, in which the small change is consistent with the stable performance of the sample.Table 1 reports the BET and Ru crystallite sizes for fresh and cyclic aged DFM.Some sintering is observed.However, given the low temperatures of the process, this is not expected to be a major source of deactivation.Extended cyclic aging will be necessary to more definitively determine the major sources of deactivation.These studies will be soon initiated.Crystallite sizes (d) were calculated from the equation d = (103/%D) where %D = % dispersion and 103 is the constant related to the number of Ru atoms present in the samples determined from room temperature H2 chemisorption [37].

12-Cycle Test in Simulated Flue Gas
Performance of the pelleted Ru-Na 2 CO 3 DFM in simulated O 2 -and-steam-containing flue gas is another key factor for its commercialization and application.A 12-cycle test was conducted in the fixed bed reactor with the performance shown in Figure 5.The results of BET specific surface area of the fresh and after-12-cycle Ru-Na 2 CO 3 pelleted DFM sample is shown in Table 1, in which the small change is consistent with the stable performance of the sample.Table 1 reports the BET and Ru crystallite sizes for fresh and cyclic aged DFM.Some sintering is observed.However, given the low temperatures of the process, this is not expected to be a major source of deactivation.Extended cyclic aging will be necessary to more definitively determine the major sources of deactivation.These studies will be soon initiated.Crystallite sizes (d) were calculated from the equation d = (103/%D) where %D = % dispersion and 103 is the constant related to the number of Ru atoms present in the samples determined from room temperature H 2 chemisorption [37].

12-Cycle Test in Simulated Flue Gas
Performance of the pelleted Ru-Na2CO3 DFM in simulated O2-and-steam-containing flue gas is another key factor for its commercialization and application.A 12-cycle test was conducted in the fixed bed reactor with the performance shown in Figure 5.The results of BET specific surface area of the fresh and after-12-cycle Ru-Na2CO3 pelleted DFM sample is shown in Table 1, in which the small change is consistent with the stable performance of the sample.Table 1 reports the BET and Ru crystallite sizes for fresh and cyclic aged DFM.Some sintering is observed.However, given the low temperatures of the process, this is not expected to be a major source of deactivation.Extended cyclic aging will be necessary to more definitively determine the major sources of deactivation.These studies will be soon initiated.Crystallite sizes (d) were calculated from the equation d = (103/%D) where %D = % dispersion and 103 is the constant related to the number of Ru atoms present in the samples determined from room temperature H2 chemisorption [37].Table 1.Brunauer-Emmett-Teller (BET) specific surface area of the fresh and flue gas cyclic aged Ru-Na2CO3 DFM pellets.Ru crystallite sized for fresh and cyclic aged samples based on H2 chemisorption assuming H/Ru = 1/1.

Sample
BET Specific Surface Area (m 2 /g) Ru Crystallite Sizes (nm)  The CO 2 adsorption capacity was reduced compared to the O 2 free process of Figure 4, because Ru was inactivated-for CO 2 adsorption-to RuO x by O 2 in the simulated flue gas.Both adsorption and methanation capacity remained relatively stable until the 9th cycle, when reduction in both CO 2 adsorption and CH 4 produced was noted.Since both CO 2 adsorption and methanation are linked to Ru activity, it was reasonable to assume that the RuO x was not sufficiently reduced, causing a loss in both performances.
Before the 11th cycle, the material was given an extra 150-min of hydrogen exposure, resulting in a recovery of both CO 2 adsorption and methanation.When the extra hydrogen exposure was extended to 48 h before the 12th cycle, both capacities reached even higher levels.This apparent deactivation was not observed in the O 2 free cycle tests of Figure 4.This again demonstrated the importance of H 2 partial pressure in maintaining Ru reduced while enhancing the rate of bulk carbonate hydrogenation.
The average conversion of adsorbed CO 2 converted to methane during the 12-cycle test was 89.6%, which is a great improvement compare to the 10-cycle test in the dry CO 2 feed (60.5%).Though the overall CO 2 adsorption of the 12-cycle test in the simulated flue gas was lower than in the 10-cycle test, the conversion efficiency of 12-cycle test is preferred.Water generated at the surface during methanation may oxidize the surface of Ru.Therefore, a minimum H 2 partial pressure is required to maintained Ru in the reduced and active state.We have found that H 2 partial pressure is a critical parameter necessary for reducing the RuO x formed during CO 2 adsorption (Figure 5) and methanation (Figure 6).This will be confirmed in the future.

Effect of H 2 Partial Pressure on CO 2 Hydrogenation and RuO x Reduction
As observed for Figure 5, the loss in performance was reversed by increasing the exposure time of the DFM to H 2 .It is reasonable to assume that increasing the partial pressure of H 2 would have the same effect by maintaining the Ru in the metallic and active state while increasing the extent of carbonate hydrogenation.This was demonstrated in Figure 6 where 10% H 2 /N 2 dramatically increased the amount of methane produced (131 mL) compared to 5% H 2 /N 2 (58.7 mL).This also points to the importance of the redox chemistry of Ru/RuO x .It is logical to assume that during the 12-cycle simulated flue gas test (Figure 5), Ru was oxidized and there was insufficient H 2 to re-reduce it.
Catalysts 2017, 7, 88 7 of 13 The CO2 adsorption capacity was reduced compared to the O2 free process of Figure 4, because Ru was inactivated-for CO2 adsorption-to RuOx by O2 in the simulated flue gas.Both adsorption and methanation capacity remained relatively stable until the 9th cycle, when reduction in both CO2 adsorption and CH4 produced was noted.Since both CO2 adsorption and methanation are linked to Ru activity, it was reasonable to assume that the RuOx was not sufficiently reduced, causing a loss in both performances.
Before the 11th cycle, the material was given an extra 150-min of hydrogen exposure, resulting in a recovery of both CO2 adsorption and methanation.When the extra hydrogen exposure was extended to 48 h before the 12th cycle, both capacities reached even higher levels.This apparent deactivation was not observed in the O2 free cycle tests of Figure 4.This again demonstrated the importance of H2 partial pressure in maintaining Ru reduced while enhancing the rate of bulk carbonate hydrogenation.
The average conversion of adsorbed CO2 converted to methane during the 12-cycle test was 89.6%, which is a great improvement compare to the 10-cycle test in the dry CO2 feed (60.5%).Though the overall CO2 adsorption of the 12-cycle test in the simulated flue gas was lower than in the 10-cycle test, the conversion efficiency of 12-cycle test is preferred.Water generated at the surface during methanation may oxidize the surface of Ru.Therefore, a minimum H2 partial pressure is required to maintained Ru in the reduced and active state.We have found that H2 partial pressure is a critical parameter necessary for reducing the RuOx formed during CO2 adsorption (Figure 5) and methanation (Figure 6).This will be confirmed in the future.

Effect of H2 Partial Pressure on CO2 Hydrogenation and RuOx Reduction
As observed for Figure 5, the loss in performance was reversed by increasing the exposure time of the DFM to H2.It is reasonable to assume that increasing the partial pressure of H2 would have the same effect by maintaining the Ru in the metallic and active state while increasing the extent of carbonate hydrogenation.This was demonstrated in Figure 6 where 10% H2/N2 dramatically increased the amount of methane produced (131 mL) compared to 5% H2/N2 (58.7 mL).This also points to the importance of the redox chemistry of Ru/RuOx.It is logical to assume that during the 12-cycle simulated flue gas test (Figure 5), Ru was oxidized and there was insufficient H2 to re-reduce it.The rate and extent of RuOx reduction is critical for the flue gas operation.In TGA experiments the rate and extent of RuOx reduction was to understand the minimum H2 content needed for rapid reduction of the oxides of Ru. Results are shown in Figure 7.It is clear the reduction rate (slope) increases with H2 content (0.069 mg/min at 1.30% H2 to 0.370 mg/min at 2.86% H2).The most dramatic The rate and extent of RuO x reduction is critical for the flue gas operation.In TGA experiments the rate and extent of RuO x reduction was to understand the minimum H 2 content needed for rapid reduction of the oxides of Ru. Results are shown in Figure 7.It is clear the reduction rate (slope) increases with H 2 content (0.069 mg/min at 1.30% H 2 to 0.370 mg/min at 2.86% H 2 ).The most dramatic increase occurs for H 2 > 1.3%.Equally as significant is the decrease in time for initiation of reduction as the H 2 content increases.This points to the reducibility of RuO x to Ru, making it more suitable for the DFM flue gas process relative to NiO x to Ni metal [32] that requires temperatures in excess of 500 • C for reduction.
Catalysts 2017, 7, 88 8 of 13 as the H2 content increases.This points to the reducibility of RuOx to Ru, making it more suitable for the DFM flue gas process relative to NiOx to Ni metal [32] that requires temperatures in excess of 500 °C for reduction.

Speculated Functional Pathway of the Ru-Na2CO3 DFM
The speculated pathways for each step and component in DFM are summarized in Table 2.The RuOx, as prepared, must be activated to its active metallic state for methanation.For feeds containing O2, it will be oxidized and lose its ability to adsorb CO2.However, during methanation it will be reduced and re-activated for methane production.For O2-free feed gas, Ru is both active for CO2 adsorption and methanation.For Ru-Na2CO3 DFM, bulk carbonate hydrogenation occurs during pretreatment, producing NaOH and CH4 with CO2 spill-over to Ru sites.Upon exposure to CO2, it is reformed to a bulk carbonate species where CO2 adsorbs on the surface Na2CO3⋯CO2 and reforms bulk carbonate.The slowest step is the hydrogenation of the reformed bulk carbonate.Table 2 is still speculative and is the focus of on-going kinetic and characterization studies.

Material Synthesis
The powder materials tested were prepared via incipient wetness method on a high surface area γ-Al2O3 (SBA-150, BASF, Iselin, NJ, USA) support using ruthenium(III) nitrosylnitrate (Ru(NO)(NO3)3, Ru 31.3% min, Alfa Aesar, Tewksbury, MA, USA) and Na2CO3 (Sigma-Aldrich, St.The speculated pathways for each step and component in DFM are summarized in Table 2.The RuO x , as prepared, must be activated to its active metallic state for methanation.For feeds containing O 2 , it will be oxidized and lose its ability to adsorb CO 2 .However, during methanation it will be reduced and re-activated for methane production.For O 2 -free feed gas, Ru is both active for CO 2 adsorption and methanation.For Ru-Na 2 CO 3 DFM, bulk carbonate hydrogenation occurs during pretreatment, producing NaOH and CH 4 with CO 2 spill-over to Ru sites.Upon exposure to CO 2 , it is reformed to a bulk carbonate species where CO 2 adsorbs on the surface Na 2 CO 3 • • • CO 2 and reforms bulk carbonate.The slowest step is the hydrogenation of the reformed bulk carbonate.Table 2 is still speculative and is the focus of on-going kinetic and characterization studies.

Material Synthesis
The powder materials tested were prepared via incipient wetness method on a high surface area γ-Al 2 O 3 (SBA-150, BASF, Iselin, NJ, USA) support using ruthenium(III) nitrosylnitrate (Ru(NO)(NO 3 ) 3 , Ru 31.3% min, Alfa Aesar, Tewksbury, MA, USA) and Na 2 CO 3 (Sigma-Aldrich, St. Louis, MO, USA) as the precursors for Ru and Na 2 CO 3 , respectively.Repeated impregnations were conducted to reach the designed loadings if necessary.The Al 2 O 3 supported alkaline materials (Na 2 CO 3 ) were dried at 120 • C for 2 h and then calcined at 400 • C for 4 h in air.For Ru deposition, the material was dried at 120 • C for 2 h followed by calcination at 250 • C for 2 h in air after metal catalyst (Ru) impregnations.
The pellet materials (5 mm × 5 mm tablets) used for the cyclic tests were prepared via a precipitation method with the target composition of 5% Ru, 10% Na 2 CO 3 /Al 2 O 3 (TH100, SASOL, Hamburg, Germany).The principle procedure of the precipitation method was derived from the work done by Zheng et al. [31], but the impregnation order, selected precursors, and calcination programs were modified (see Figure 8 for the preparation process flow).Ruthenium(III) chloride hydrate (RuCl 3 •xH 2 O, 99.9% PGM basis, Ru 38% min, Alfa Aesar, Tewksbury, MA, USA) and sodium carbonate (Sigma-Aldrich) were the chosen precursors for Ru and Na 2 CO 3 , respectively.Using RuCl 3 as the Ru precursor for pellet sample instead of Ru(NO)(NO 3 ) 3 achieved more uniform penetration within the pellets, consequently improved dispersion of Ru, and avoided residual of precursor in the sample.The highest calcination temperature was 250 • C to avoid generation of high temperature volatile ruthenium oxides (>300 • C in air).The water wash was introduced, after precipitation of water-insoluble Ru(OH) x to remove unwanted chlorides.The carbonate salts were then impregnated and processed as shown in Figure 8.
Catalysts 2017, 7, 88 9 of 13 conducted to reach the designed loadings if necessary.The Al2O3 supported alkaline materials (Na2CO3) were dried at 120 °C for 2 h and then calcined at 400 °C for 4 h in air.For Ru deposition, the material was dried at 120 °C for 2 h followed by calcination at 250 °C for 2 h in air after metal catalyst (Ru) impregnations.The pellet materials (5 mm × 5 mm tablets) used for the cyclic tests were prepared via a precipitation method with the target composition of 5% Ru, 10% Na2CO3/Al2O3 (TH100, SASOL, Hamburg, Germany).The principle procedure of the precipitation method was derived from the work done by Zheng et al. [31], but the impregnation order, selected precursors, and calcination programs were modified (see Figure 8 for the preparation process flow).Ruthenium(III) chloride hydrate (RuCl3•xH2O, 99.9% PGM basis, Ru 38% min, Alfa Aesar, Tewksbury, MA, USA) and sodium carbonate (Sigma-Aldrich) were the chosen precursors for Ru and Na2CO3, respectively.Using RuCl3 as the Ru precursor for pellet sample instead of Ru(NO)(NO3)3 achieved more uniform penetration within the pellets, consequently improved dispersion of Ru, and avoided residual of precursor in the sample.The highest calcination temperature was 250 °C to avoid generation of high temperature volatile ruthenium oxides (>300 °C in air).The water wash was introduced, after precipitation of water-insoluble Ru(OH)x to remove unwanted chlorides.The carbonate salts were then impregnated and processed as shown in Figure 8.

Thermogravimetric Analysis (TGA)
Thermogravimetric analysis were conducted using either a TG 209 F1 Libra NETZSCH or SAT 449 F3 Jupiter NETZSCH unit for the Na2CO3 loading, carbonate hydrogenation, and hydrogen exposure time study.For each test, 50 mg of the prepared powder sample was placed in an alumina crucible and located in the instrument.The temperatures for carbonate hydrogenation and nitrate decomposition, CO2 adsorption/desorption, and CO2 methanation, were all conducted at 320 °C, while the feed gases and corresponding time periods were varied.Table 3 shows the programs in detail for each test.The total flow rates were always 40 mL/min for each feed gas.Before each test, the sample was heated to 320 °C in pure nitrogen until a stable final weight was observed.Hydrogen was introduced in pre-treatments to decompose the ruthenium nitrate to active metallic ruthenium while decomposing the remaining nitrate compounds until stable mass was observed.Pure nitrogen was used only to de gas each sample.

Thermogravimetric Analysis (TGA)
Thermogravimetric analysis were conducted using either a TG 209 F1 Libra NETZSCH or SAT 449 F3 Jupiter NETZSCH unit for the Na 2 CO 3 loading, carbonate hydrogenation, and hydrogen exposure time study.For each test, 50 mg of the prepared powder sample was placed in an alumina crucible and located in the instrument.The temperatures for carbonate hydrogenation and nitrate decomposition, CO 2 adsorption/desorption, and CO 2 methanation, were all conducted at 320 • C, while the feed gases and corresponding time periods were varied.Table 3 shows the programs in detail for each test.The total flow rates were always 40 mL/min for each feed gas.Before each test, the sample was heated to 320 • C in pure nitrogen until a stable final weight was observed.Hydrogen was introduced in pre-treatments to decompose the ruthenium nitrate to active metallic ruthenium while decomposing the remaining nitrate compounds until stable mass was observed.Pure nitrogen was used only to de gas each sample.

Conclusions
The bulk carbonate present in Ru-Na2CO3 DFM can adsorb and convert CO2 to CH4.Additionally, the bulk carbonate can also be catalytically hydrogenated to CH4 providing "depleted" sites for additional CO2 adsorption from flue gases.Ru plays a key role as a methanation catalyst for O2-containing flue gas due to its ease of reduction once oxidized.Once in its reduced state it provides an additional site for CO2 capture as well as effective methanation.This technology overcomes some of the limitations of aqueous amine CO2 scrubbing by having the dual capability of both capturing CO2 in O2-flue gas and subsequently methanating the CO2 at one temperature using renewable H2.The methane produced is dried, compressed, and recycled to the front of the plant ideally closing the carbon balance, with conversion of 89.6% of the captured CO2 in simulated flue gas.
The Ru-Na2CO3 DFM demonstrated stable performance for at least 10 cycles of CO2 adsorption and methanation in an O2-free CO2 feed gas.In simulated flue gas, due to the presence of oxygen, part of active metallic Ru is oxidized and loses its activity for CO2 adsorption and methanation.Stable performance was observed in the flue gas up to the point where Ru becomes less effective due to oxidation.This apparent deactivation can be reversed by prolonged exposure to H2 or preferably by a high H2 content.The reversibility is due to the ease of reduction of RuOx to Ru.The 12-cycle test was performed in simulated natural gas fired power plant flue gas, containing 7.5% CO 2 , 4.5% O 2 , 15% H 2 O and balance N 2 .A 3-min N 2 purge between CO 2 adsorption and methanation and the reverse was added to avoid contact of H 2 and O 2 .The first ten cycles were operated continuously and identically, while the last two cycles required a prolonged H 2 purge before CO 2 feed was introduced.The 11th cycle underwent a 150 min 5% H 2 /N 2 exposure and the 12th for 48 h.The prolonged H 2 exposure was needed to regenerate the re-oxidized ruthenium.

Conclusions
The bulk carbonate present in Ru-Na 2 CO 3 DFM can adsorb and convert CO 2 to CH 4 .Additionally, the bulk carbonate can also be catalytically hydrogenated to CH 4 providing "depleted" sites for additional CO 2 adsorption from flue gases.Ru plays a key role as a methanation catalyst for O 2 -containing flue gas due to its ease of reduction once oxidized.Once in its reduced state it provides an additional site for CO 2 capture as well as effective methanation.This technology overcomes some of the limitations of aqueous amine CO 2 scrubbing by having the dual capability of both capturing CO 2 in O 2 -flue gas and subsequently methanating the CO 2 at one temperature using renewable H 2 .The methane produced is dried, compressed, and recycled to the front of the plant ideally closing the carbon balance, with conversion of 89.6% of the captured CO 2 in simulated flue gas.
The Ru-Na 2 CO 3 DFM demonstrated stable performance for at least 10 cycles of CO 2 adsorption and methanation in an O 2 -free CO 2 feed gas.In simulated flue gas, due to the presence of oxygen, part of active metallic Ru is oxidized and loses its activity for CO 2 adsorption and methanation.Stable performance was observed in the flue gas up to the point where Ru becomes less effective due to oxidation.This apparent deactivation can be reversed by prolonged exposure to H 2 or preferably by a high H 2 content.The reversibility is due to the ease of reduction of RuO x to Ru.

Figure 1 .
Figure1.Thermal Gravimetric Analysis (TGA) adsorption studies conducted at 320 °C for 50 mg of powder.Before the adsorption the samples were degassed in 2% H2/N2 for 2 h, then exposed to 0.5% CO2 in N2 and subsequently pure N2.

Figure 1 .
Figure 1.Thermal Gravimetric Analysis (TGA) adsorption studies conducted at 320 • C for 50 mg of powder.Before the adsorption the samples were degassed in 2% H 2 /N 2 for 2 h, then exposed to 0.5% CO 2 in N 2 and subsequently pure N 2 .

Figure 2 .
Figure 2. TGA result of Ru salts and carbonate hydrogenation study operated at 320 °C for 50 mg of 1.44% Ru/Na2CO3 powder sample.

Figure 2 .
Figure 2. TGA result of Ru salts and carbonate hydrogenation study operated at 320 • C for 50 mg of 1.44% Ru/Na 2 CO 3 powder sample.

Figure 3 .
Figure 3. TGA result of the 3-cycle test for 50 mg of 5% Ru, 10% Na2CO3/Al2O3 at 320 °C.The plot excluded the over-night activation step at 320 °C in 2% H2/N2.The long tail of low weight loss in the first cycle was compressed indicating an overnight (10.5 h) hydrogen exposure at 320 °C.

Figure 3 .
Figure 3. TGA result of the 3-cycle test for 50 mg of 5% Ru, 10% Na 2 CO 3 /Al 2 O 3 at 320 • C. The plot excluded the over-night activation step at 320 • C in 2% H 2 /N 2 .The long tail of low weight loss in the first cycle was compressed indicating an overnight (10.5 h) hydrogen exposure at 320 • C.

Figure 5 .
Figure 5. CO2 adsorption and methanation using 10 grams of the Ru-Na2CO3 DFM (5% Ru, 10% Na2CO3/Al2O3) pellet for the 12-cycle test conducted in the fixed bed reactor in simulated flue gas at 320 °C.Pretreatment of the DFM was conducted with 5% H2/N2 for 150 min before the first cycle.For cycles 2-10, hydrogen exposure was 30 min.Before cycle 11 another 150-min 5% H2/N2 exposure was added, and for cycle 12, 48 h H2 exposure.The volumes shown in the figure are for CO2 adsorbed and CH4 produced.

Table 1 .
Brunauer-Emmett-Teller (BET) specific surface area of the fresh and flue gas cyclic aged Ru-Na2CO3 DFM pellets.Ru crystallite sized for fresh and cyclic aged samples based on H2 chemisorption assuming H/Ru = 1/1.

Figure 4 .
Figure 4. CO 2 adsorption and methanation cycles (10) using 10 grams of the Ru-Na 2 CO 3 DFM (5% Ru, 10% Na 2 CO 3 /Al 2 O 3 ) pellets in a fixed packed bed reactor with 7.5% CO 2 /N 2 at 320 • C. A 150-min of 5% H 2 /N 2 exposure was used during pretreatment followed by 10 identical CO 2 adsorption-methanation cycles.5% H 2 /N 2 for 30 min was used for the methanation cycles.The volumes of CO 2 adsorbed and CH 4 produced are shown in the figure.

Figure 5 .
Figure 5. CO2 adsorption and methanation using 10 grams of the Ru-Na2CO3 DFM (5% Ru, 10% Na2CO3/Al2O3) pellet for the 12-cycle test conducted in the fixed bed reactor in simulated flue gas at 320 °C.Pretreatment of the DFM was conducted with 5% H2/N2 for 150 min before the first cycle.For cycles 2-10, hydrogen exposure was 30 min.Before cycle 11 another 150-min 5% H2/N2 exposure was added, and for cycle 12, 48 h H2 exposure.The volumes shown in the figure are for CO2 adsorbed and CH4 produced.

Figure 5 .
Figure 5. CO 2 adsorption and methanation using 10 grams of the Ru-Na 2 CO 3 DFM (5% Ru, 10% Na 2 CO 3 /Al 2 O 3 ) pellet for the 12-cycle test conducted in the fixed bed reactor in simulated flue gas at 320 • C. Pretreatment of the DFM was conducted with 5% H 2 /N 2 for 150 min before the first cycle.For cycles 2-10, hydrogen exposure was 30 min.Before cycle 11 another 150-min 5% H 2 /N 2 exposure was added, and for cycle 12, 48 h H 2 exposure.The volumes shown in the figure are for CO 2 adsorbed and CH 4 produced.

Figure 6 .
Figure 6.Methane production comparison of 5% H 2 /N 2 and 10% H 2 /N 2 .The test was operated using 10 grams of 5% Ru, 10% Na 2 CO 3 /Al 2 O 3 pellets in the large fixed bed reactor.The operational conditions were at 320 • C, 300 mL/min of 7.5% CO 2 /N 2 for CO 2 adsorption.The volume of methane produced are shown.

Figure 7 .
Figure 7. TGA results of RuOx reduction on 50 mg 5% Ru/Al2O3 powder samples H2 partial pressures from 1.3% to 2.86% at 300 °C.Only reduction sections are shown in the figure.

Figure 7 .
Figure 7. TGA results of RuO x reduction on 50 mg 5% Ru/Al 2 O 3 powder samples H 2 partial pressures from 1.3% to 2.86% at 300 • C.Only reduction sections are shown in the figure.

Figure 8 .
Figure 8. Ru-Na2CO3 DFM pellet preparation process flow using the precipitation method.A slight excess of NaOH was used with a ratio of Ru:OH = 1:4.5 to precipitate Ru(OH)x.A water wash and filtration (40 times of water immersion and separation) was used to remove any NaCl.Drying and calcination steps were conducted in air.

Figure 8 .
Figure 8. Ru-Na 2 CO 3 DFM pellet preparation process flow using the precipitation method.A slight excess of NaOH was used with a ratio of Ru:OH = 1:4.5 to precipitate Ru(OH) x .A water wash and filtration (40 times of water immersion and separation) was used to remove any NaCl.Drying and calcination steps were conducted in air.

Figure 10 .
Figure 10.Process flow for the two different cyclic tests operated on the scaled-up fixed bed reactor.(a) O2-and steam-free CO2 adsorption and hydrogenation for the 10-cycle test.Pre-reduction was utilized only once before the first cycle.The subsequent ten cycles were conducted repeatedly under identical experimental conditions; (b) Simulated flue gas 12-cycle test with the pretreatment used only before the first cycle followed by 10 identical cycles.A N2 purge was needed between each cycle to avoid H2-O2 contact.Before the 11th and the 12th cycle, a prolonged H2 exposure was utilized while subsequent steps remained the same as the first ten cycles.

Figure 10 .
Figure 10.Process flow for the two different cyclic tests operated on the scaled-up fixed bed reactor.(a) O 2 -and steam-free CO 2 adsorption and hydrogenation for the 10-cycle test.Pre-reduction was utilized only once before the first cycle.The subsequent ten cycles were conducted repeatedly under identical experimental conditions; (b) Simulated flue gas 12-cycle test with the pretreatment used only before the first cycle followed by 10 identical cycles.A N 2 purge was needed between each cycle to avoid H 2 -O 2 contact.Before the 11th and the 12th cycle, a prolonged H 2 exposure was utilized while subsequent steps remained the same as the first ten cycles.

Table 1 .
Brunauer-Emmett-Teller (BET) specific surface area of the fresh and flue gas cyclic aged Ru-Na 2 CO 3 DFM pellets.Ru crystallite sized for fresh and cyclic aged samples based on H 2 chemisorption assuming H/Ru = 1/1.

Table 2 .
Speculated function pathway of Ru and Na2CO3 in the Ru-Na2CO3 DFM.Na2CO3 bulk and Na2CO3 surface are the same component but bulk and surface are distinguished.

Table 3 .
Programs of TGA tests at 320 °C related to the results discussed in this paper.
1The step 2 shows in the table was for the first cycle of the test only.For the 2nd and the 3rd cycle,

Table 3 .
Programs of TGA tests at 320 • C related to the results discussed in this paper.No N 2 purge between adsorption and methanation was needed since no O 2 was present in this series of tests.