Upgrading CO2 from simulated power plant flue gas via integrated CO2 capture and dry reforming of methane using Ni-CaO

,


Introduction
Excessive CO 2 emission has been widely believed as the main cause of the greenhouse effect and climate issues [1].It is essential to develop carbon capture and storage or utilisation (CCS [2,3] or CCU [4,5]) processes to restrict carbon emissions before renewable energy dominates.CCS is promising to sequester massive amounts of carbon dioxide.However, there might be the potential risk of secondary leakage and the impact on the environment and ecology of the storage site [6], restricting its wide deployment.Therefore, converting CO 2 into valuable products attracts great interests in the field of carbon net-zero.
Upgrading CO 2 and CH 4 (natural gas) into syngas (CO + H 2 ) via a dry reforming process (Eq.( 1)) provides a valuable solution to utilise two greenhouse gases [16].It is promising to integrate CO 2 capture and dry reforming of methane (DRM) into one process (ICCU-DRM) to control carbon emissions via a profitable path.Ahmed et al. [17], Kim et al. [18] and Tian et al. [19] have explored the ICCU-DRM using Ni-CaO based DFMs.Other adsorbents, such as BaO [20], Li 4 SiO 4 [12] and SrCeO 3 [21], were also proven effective for ICCU-DRM with the assistance of Ni as the catalytic sites.Based on the Ni-CaO system, Hu et al. [22] achieved stabilised performance by introducing ZrO 2 and CeO 2 coating onto Ni-CaO particles.However, as an emerging field, all the existing research in ICCU-DRM were investigated at ideal CO 2 capture condition (CO 2 balanced in N 2 ) and neglected the effect of other components in the realistic power plant flue gas.
A few research groups have studied the effect of H 2 O and O 2 in flue gas on the performance of calcium looping [23,24] and other ICCU processes, including ICCU-CO 2 methanation [25,26] and ICCU-reverse water-gas shift reaction (ICCU-RWGS) [27].However, the influence of H 2 O and O 2 in realistic flue gas on the performance of ICCU-DRM process remains unknown and challenging.Furthermore, the carbon depositions in DRM would contribute to the formation of CO in the following CO 2 capture process, which will be diluted in CO 2 -lean exhaust gas and be a pollutant.To eliminate the CO production in CO 2 capture step, it is necessary to consume the carbon deposition on catalysts before contacting with flue gas.Gasifying carbon using H 2 O [28] (carbon steam gasification-CSG; Eq. ( 2)) would not only by-produce syngas, but also eliminate the undesirable CO generation in the following CO 2 capture step.Integrating CO 2 capture and DRM with CSG (Fig. 1) might represent one applicable process for environmentalfriendly ICCU-DRM with practical research significance.
Herein, we synthesised the Ni-CaO DFM and evaluated the ICCU-DRM at ideal (10 % CO 2 /N 2 ) and simulated (10.0 % CO 2 + 6.0 % H 2 O + 6.7 % O 2 /N 2 ) power plant flue gas conditions.The simulated flue gas component is designed based on conventional coal power plants (CO 2 : 5-11 %; O 2 : 4-7 % and H 2 O: 5-30 %), which contribute to ~ 30 % of worldwide electricity supply.The effect of H 2 O and O 2 in flue gas on ICCU-DRM was systematically investigated, and pivotal issues and solutions were presented.The effect of carbon removal and catalytic performance of the extra CSG process was also evaluated and analysed.Detailed characterisations, including XRD, SEM, TEM, Raman and TPR were applied to confirm the material structure, metal state and carbon deposition, and the mechanism of ICCU-DRM-CSG under simulated flue gas CO 2 capture condition was proposed.

Material characterisations
The phase information of original and reduced Ni 10 -CaO was determined by powder X-ray diffraction (XRD) using a PANalytical Empyrean series 2 diffractometer with a Cu Ka X-ray source (2 Theta = 5-80 • ).The H 2 temperature programmed reduction (H 2 -TPR) curve was collected on a Hi-Res TGA 2950 thermogravimetric analyser.Specifically, the Ni 10 -CaO DFM was pretreated at 800 • C for 10 mins and then equilibrated at 200 • C in N 2 .Subsequently, 5.0 % H 2 /Ar was introduced with a ramping rate of 10 • C min − 1 to 800 • C. The CH 4 temperature programmed reduction (CH 4 -TPR) and temperature programmed oxidisation (TPO) tests were operated using ~ 9.0 % CH 4 /N 2 and 21 % O 2 /N 2 with similar procedures to H 2 -TPR, respectively.The Raman spectrums were characterised on the WITec Alpha 300R Confocal Raman Microscope equipped with a 532 nm diode laser (10 mW).The Scanning Electron Microscopy (SEM) images were taken by FEI Quanta FEG-Environmental SEM Oxford Ex-ACT, and the Transmission Electron Microscopy (TEM) patterns were collected on JEM-1200EX.

ICCU-DRM evaluations
The ICCU-DRM evaluation system is illustrated in Fig. S1.Specifically, 0.3 g of original Ni 10 -CaO dual functional material (DFM) was fixed by quartz wool and placed in the middle of a quartz reaction tube (OD: 12.0 mm; ID: 10.5 mm; L: 65.0 cm).A thermocouple was placed at the position of the sample to monitor the temperature, and the reaction tube was fixed in the middle of a tube furnace (Elite TSH 12/50/300-2416CG).The inlet gases (CO 2 , air, N 2 ) were controlled by mass flowmeters (OMEGA FMA-A2306), which were calibrated by a bubble flowmeter.The outlet gases were analysed and recorded by a flue gas analyser (Enerac 700 AV; NDIR detector) and an H 2 analyser (CX-02; TCD detector).
A series of ICCU evaluations were carried out to investigate the ideal or simulated CO 2 conditions with or without extra carbon steam gasification (CSG).Typically, all the inlet gas flows were kept at 100 mL min − 1 and the outlet gas flows were assumed as constant.Initially, the Ni 10 -CaO DFM was in-situ reduced in 5.0 % H 2 /Ar at 650 • C for 1 h, and then equilibrated to test temperature in N 2 .For the CO 2 capture (step 1), 10.0 % CO 2 with or without 6.0 % H 2 O or 6.7 % O 2 balanced in N 2 was introduced for 20 mins at 650 • C. For the DRM (step 2), 9.0 % CH 4 /N was applied until there was no CO in the outlet gas at 650 • C. For some of the evaluations, an extra CSG (10.0 % H 2 O/N 2 ) (step 3) was carried out to remove the deposited carbon at 650 • C. The cycle studies were carried out by repeating the above steps.It is noted that there was a 3 mins' N purging step among each gas swing to avoid direct contact with reactant gases.The real-time signals of various gases were further integrated to evaluate the ICCU-DRM-CSG performance, such as CO, CO 2 , H 2 and CH yields.  in the diffusion-controlled stage, which might be attributed to the poorer mesoporosity due to the oxidisation of Ni (Figs.S2 and S3).

Results and discussions
Apart from the CO 2 capture performance related to CaO, the chemical state change of Ni was also noteworthy.Several reports had concluded that CO 2 would react with Ni at elevated temperatures [31,32].However, CO 2 oxidisation of Ni was very limited, as no distinct CO generation was detected during the stage of CO 2 capture (Fig. 2).The oxidisation of Ni with H 2 O was also not significant since there was no H 2 generation during CO 2 capture (Fig. 2b).However, there is no controversy on the Ni oxidisation by O 2 .As presented in the O 2 -containing condition test (Fig. 2c), there was a delay of the O 2 signal, proving the occurrence of Ni oxidisation.
Notably, CH 4 decomposition (Eq.( 3)) yielded carbon during the DRM step, resulting in a series of side reactions in the following CO 2 capture step.For ideal conditions, as shown in Fig. 2a, a large amount of CO (~16.0 mmol g − 1 ) was generated in the 2nd and 3rd CO 2 capture step via reverse Boudouard reaction (Eq.( 4)).The by-produced CO should not be treated as a valuable product [33] in the diluted exhaust gas.On the contrary, it should be eliminated as a bio-toxic pollutant [34,35] in the CO 2 capture step.The presence of H 2 O or O 2 would further promote the carbon consumption rate by reacting with the deposited carbon.As presented in Fig. 2b, apart from the reverse Boudouard reaction, the carbon could be gasified by H 2 O via carbon steam gasification (CSG) reaction (Eq.( 2)).O 2 could oxidise carbon (Eq.( 5)) with higher efficiency than H 2 O, reflected in the higher real-time CO generation concentration (~14.0 % and 11.0 % in the presence of O 2 and H 2 O, respectively).For ICCU-DRM at simulated flue gas conditions (Fig. 2d), the step of CO 2 capture was more complex, including combustion of carbon (Eq.( 5) and ( 6)), reverse Boudouard reaction (Eq.( 4)) and carbon steam gasification (Eq.( 2)).The co-existence of O 2 and H 2 O exhibit competitive reactivity to carbon, reflected in comparable CO generation to O 2 or H 2 O-containing conditions.Although some other side reactions, such as water gas shift reaction (Eq.( 7)), might be promoted by CaO (i.e.sorption enhanced WGS [36]) to consume CO, over 12.5 % CO is still present in the CO 2 -lean exhaust gas.Furthermore, some tandem reactions were even performed in simulated flue gas CO 2 capture, such as the formation of CH 4 via CSG or WGS and Sabatier reaction (Eq.( 8)) or CO methanation (Eq.( 9)).However, the byproduced H 2 or CH 4 from H 2 O and C would be then consumed by O 2 , which explained the O 2 signal would appear only when the carbon was thoroughly consumed. ) Therefore, CO generation during CO 2 capture is inevitable in the presence of carbon formed during the stage of CO 2 utilisation (DRM).It is thus necessary to introduce an extra process to pre-consume the deposited carbon before the CO 2 capture step to achieve environmentalfriendly carbon emission control performance.

Performance of CO 2 utilisation in ICCU-DRM
In ICCU-DRM, CO 2 in the flue gas was firstly fixed by CaO in the form of CaCO 3 (CO 2 capture step) and then utilised via DRM reaction by introducing CH 4 (CO 2 utilisation step).For ideal (Fig. 2a) and H 2 O-containing (Fig. 2b) flue gas, CH 4 decomposition and DRM would be triggered instantly, yielding syngas with an H 2 :CO ratio > 1.
As a comparison, the presence of O 2 in CO 2 capture significantly deteriorated the DRM catalytic performance (Fig. 2c and 2d).Firstly, the CH 4 could not instantly activate Ni at the testing conditions.As shown in CH 4 -TPR (Fig. S4), NiO requires a higher reduction temperature (~640 • C) in CH 4 than in H 2 (Fig. S5), which is close to the testing temperature (650 • C).Furthermore, the volume expansion of CaO carbonation might cover NiO and decrease its reducibility during the DRM process.As detailed in Fig. S6, the delay of NiO reduction was observed, and the coverage of CaCO 3 on NiO further hindered NiO reduction.The Ni particles in sol-gel derived Ni 10 -CaO were evenly dispersed in the CaO matrix (Fig. S7), explaining the longest prereduction delay in the 1st cycle evaluation.However, as shown in Fig. 2c and 2d, the pre-reduction delay became easier after 1-2 cycles, which might be attributed to the exposure of Ni particles.
Apart from the delay of catalyst reduction, the presence of O 2 in the CO 2 capture step also decreased the catalytic performance in the DRM step.Fig. 2a and 2d show that the initial CO concentration at ideal and simulated conditions ICCU were ~ 10.5 % and 8.0 %, respectively.Interestingly, the real-time CO 2 signals of various tests were different when the DRM is triggering (Fig. S8).It was confirmed that the reduction of NiO is accompanied by the formation of CO 2 , as illustrated in Fig. S9 and evidenced in Fig. S10.Furthermore, the contact between Ni and CH 4 would be hindered due to the nearby CO 2 -rich atmosphere, resulting in poor catalytic performance.

Performance of extra carbon steam gasification after ICCU-DRM
It is proposed that the formed carbon on Ni 10 -CaO from the stage of CO 2 utilisation (DRM) contributes to abundant side reactions in the following CO 2 capture step.However, the decomposition of CH 4 was inevitable in DRM.Furthermore, the reaction between CO 2 and C was believed as the key intermediate step for syngas production [37].Removing the deposited carbon via carbon gasification provided an effective path for an environmental-friendly ICCU-DRM process under simulated flue gas conditions.
As demonstrated in Fig. 3a, the CSG process exhibited effective carbon removal efficiency without affecting the DRM catalytic performance (detailed in Table S1).The yield of CO in the following CO capture step was decreased from ~ 16.0 to 0.3 mmol g − 1 compared to the experiment without CSG (Fig. 2a).Similar to DRM, CSG could produce equimolar CO and H 2 .However, the side reaction (Eq.( 10)) would be enhanced by the carbonation of CaO [38], resulting in a ~ 1.5 H 2 :CO ratio.Due to the co-existence of H 2 and CO 2 , Ni slightly catalysed CO methanation via tandem reaction.However, steam could not thoroughly consume the deposited carbon, as evidenced in the generation of CO in the CO 2 capture step in ideal conditions (Fig. 3a).Nevertheless, for simulated flue gas (Fig. 3b), the carbon residue after CSG could be preferentially consumed by O 2, generating CO 2, which would be recaptured by CaO.

Mechanism of ICCU-DRM-CSG under simulated flue gas condition
As demonstrated in Fig. 1, the ICCU-DRM-CSG process included CO capture, methane dry reforming and carbon steam gasification steps.The phase change and deposited carbon during this process were monitored by ex-situ XRD (Fig. 4a), Raman spectrum (Fig. 4b) and SEM (Fig. 4d to 4g).
As shown in Fig. S5, NiO on Ni 10 -CaO exhibited two reduction peaks, which could be related to the free NiO (474 • C) and interacted (543 • C) NiO [19], respectively.NiO could be thoroughly reduced in the pretreatment before ICCU evaluation, confirmed in Fig. S11.After carbonation in simulated flue gas, Ni was oxidised into NiO, and CO 2 in  the flue gas was captured by adsorbents in the form of CaCO 3 (grey line in Fig. 4a).Subsequently, as shown in Fig. 4a and 4b, the carbonates could be thoroughly consumed after DRM, and distinct carbon was detected from XRD and Raman analysis.After consuming the deposited carbon via CSG, the C characteristic peak (2 Theta = 26 • ) in XRD disappeared; however, it could still be identified by the Raman spectrum (blue line in Fig. 4b).The total carbon amount was then tested by TPO (Fig. 4c), further confirming the carbon residual after the CSG step.
There are three bands characterising carbon species with various orderliness, including the D band at ~ 1357 cm − 1 from amorphous or disordered carbon [39], the G band at ~ 1579 cm − 1 from sp 2 stretching vibration of graphitic carbon and the G' band from the second-order scattering of two phonons [40].The intensity ratio of D and G bands (I D :I G ) could be utilised to evaluate the graphitisation of carbons [41], while the amorphous carbon was reported not to have a G' band [42].As confirmed in Fig. 5f, many carbon nanotubes formed after DRM, accompanied by a low I D :I G ratio (0.65) in the Raman spectrum.Applying steam to gasify carbon significantly changed the carbon species on Ni 10 -CaO DFM.The carbon nanotubes disappeared (Fig. 4g), and the I D :I G ratio increased to 1.07, accomplished with the decrease of I G' : I G .These characteristics indicated that the formed ordered deposited carbon was damaged into fragments [33].Nevertheless, as shown in Fig. 3b, O 2 in flue gas could be utilised to effectively convert the trace carbon residual into CO 2 instead of CO.
In short, ICCU-DRM possessed high sensitivity to H 2 O and O 2 in simulated flue gas conditions.The CaO-based adsorbents are the most potential materials for high-temperature CO 2 capture due to their low price, accessibility and excellent capture performance [43,44].However, the poor reducibility of NiO and massive carbon deposition might be the restriction of this process.Therefore, improving the reducibility of catalytic sites and inhibiting the carbon deposition formation in the DRM process or introducing an extra carbon gasification step before CO 2 capture are highly recommended for ICCU-DRM, particularly when realistic flue gas is used as the carbon source.Considering the CO 2 capacity and the consumed time of the overall ICCU-DRM cycle, ~1.8 tons Ni-CaO DFMs are needed to capture and convert over 700 tons CO 2 emissions per month.Reducing the cost of materials is also urgent for deploying this process economically.

Conclusion
In this study, we investigated the integrated CO 2 capture and dry reforming of CH 4 using Ni 10 -CaO dual functional material under ideal and simulated flue gas conditions.The carbonation kinetics and cycle stability were promoted by H 2 O during the CO 2 capture step.However, O 2 in flue gas would thoroughly oxidise Ni into NiO, and the coverage of

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 3 .
Fig. 3. Three cycles of evaluation using Ni 10 -CaO under (a) ideal and (b) simulated flue gas conditions with extra carbon steam gasification (CSG) step.(The whole process was operated at 650 • C isothermally under atmospheric pressure; all the flow rate was constant at 100 mL min − 1 ).

Fig. 4 .
Fig. 4. Ex-situ XRD (a) and Raman spectrums (b) of Ni 10 -CaO during ICCU-DRM-CSG and TPO (c) of Ni 10 -CaO after DRM (red line) and CSG (blue line); SEM images of Ni 10 -CaO after pre-reduction (d), simulated flue gas carbonation (e), DRM reaction (f) and CSG (g).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) CaCO 3 on NiO further deteriorated NiO reduction, resulting in a timeconsuming pre-reduction delay before triggering DRM.The triggered reduction of NiO by CH 4 would generate CO 2 and potentially heat the nearby CaCO 3 , resulting in an instant CO 2 release and obstructing the contact of CH 4 and Ni.In cyclic ICCU-DRM, the deposited carbon was gasified into toxic CO via the reverse Boudouard reaction during CO 2 capture.Introducing an extra carbon steam gasification step after DRM demonstrates effective carbon removal performance with the byproduction of syngas and without negatively affecting the overall ICCU-DRM performance.CRediT authorship contribution statement Shuzhuang Sun: Conceptualization, Methodology, Formal analysis, Investigation, Data curation, Writingoriginal draft.Yingrui Zhang: Methodology, Investigation, Formal analysis.Chunchun Li: Methodology, Investigation, Formal analysis.Yuanyuan Wang: Methodology, Investigation, Formal analysis.Chen Zhang: Methodology, Investigation, Formal analysis.Xiaotong Zhao: Methodology, Investigation, Formal analysis.Hongman Sun: Formal analysis, Resources, Writingreview & editing, Supervision.Chunfei Wu: Conceptualization, Writing review & editing, Supervision, Project administration, Funding acquisition.