Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Application of carbon nanotubes prepared from CH4/CO2 over Ni/MgO catalysts in CO2 capture using a BEA–AMP bi-solvent blend

Application of carbon nanotubes prepared from CH4/CO2 over Ni/MgO catalysts in CO2 capture using... Carbon nanotubes (CNTs) were synthesized by the chemical vapour deposition of methane and carbon dioxide over a Ni/MgO catalyst. The synthesized CNTs were then mixed with K/MgO catalyst at different ratios and used as the catalyst for CO absorption in butylethanolamine-2-amino-2-methyl-l-propanol bi-solvent blend. The catalysts were characterized using X-ray diffraction, scanning electron microscopy, butylethanolamine, thermal gravimetric analysis and temperature-programmed desorption of carbon dioxide in order to determine the characteristics responsible for good CO -absorption performance. The results showed that, with the addition of a catalyst into the amine solution, the amine reached equilibrium CO loading faster than without a catalyst. Also, the increase in the CNT content of the KMgO/CNTs catalyst made the CO absorption reach equilibrium much more quickly compared with just KMgO alone and without a catalyst. The KMgO/CNTs at a ratio of 1:4 yielded the fastest time to reach CO -loading equilibrium at 240 min, which was mainly due to the increase in strong basic sites as well as the highest total basic sites with an increase in CNT content. In addition, because of the extremely large specific surface area and pore volume generated due to the CNT, the number of exposed active centres per unit mass increased tremendously, leading to a corresponding tremendous increase in CO absorption. Keywords: carbon nanotubes; KMgO; catalyst; CO capture; absorption CO capture based on absorption using liquid amine solu- Introduction tions is currently the most mature and effective technology The continuous rise in CO concentration in the atmosphere applied for the removal of CO . The CO -absorption process 2 2 has generated worldwide concerns about global warming typically occurs via a continuous cycle, with CO absorption and climate-change issues. To control CO emission and slow 2 operating at around 40°C and simultaneous amine regener - down global warming, it is necessary to eliminate CO from ation upon heating the CO-loaded amine solution at a tem- major point sources such as fossil-fuelled power plants [ ]. CO 1 capture, utilization and storage (CCUS) have been proposed to perature of 120°C [4 , 5]. The solvents mainly employed in be a promising way to effectively control CO emissions [2, 3]. industrial processes include monoethanolamine (MEA) and Received: 29 March, 2019; Accepted: 27 April, 2019 © The Author(s) 2019. Published by Oxford University Press on behalf of National Institute of Clean-and-Low-Carbon Energy This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http:// creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, 251 provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 252 | Clean Energy, 2019, Vol. 3, No. 4 diethanolamine (DEA) but their weaknesses are corrosive- of K/MgO during the rate-determining step of CO absorp- ness and the high energy required for the process, especially tion in BEA–AMP as well as the high surface area provided during the regeneration of CO -loaded solvent (about 70–80% by K/MgO. Carbon-based materials such as activated carbon, of the running cost for a CO-capture plant) [, 67]. A  steric carbon nanofibers and carbon nanotubes (CNTs) are also hindrance amine such as 2-amino-2-methyl-l-propanol good capture materials due to their large surface area and (AMP) has been recommended to reduce the energy con- high-electron-density surface for facilitating gas absorption sumption. This is because AMP is a primary amine with OH [19,20,21]. The structure of CNTs consists of graphene layers, groups and a sterically hindered alkyl group. The steric c- har which is beneficial in terms of increasing the exposure of acter tends to form unstable carbamates whose hydrolysis amine molecules, thereby resulting in the enhancement leads to the formation of bicarbonates, which easily release of the CO -loading capacity of amine [20]. CNTs are typic- free amine molecules for further reaction with CO resulting ally prepared by chemical vapour deposition (CVD), arc dis- in increasing CO -absorption capacity [, 89]. Therefore, AMP charge and laser ablation. Among these three techniques, offers a lower energy requirement during the regeneration the CVD method has the advantage of being a simple pro- process. A bi-solvent blend has also been found to enhance cess that produces high purity of the CNTs and synthesizes the capture efficiency and reduce the energy of regener- high quantity at lower costs [22]. The production of CNTs ation. The combination of each amine offers the advantages via the CVD method is mostly accomplished using transi- of the high absorption capacity of a reactive amine (primary tion metals (Fe, Co and Ni) supported on different mater - or secondary amine) and low solvent regeneration cost of a ials, such as AlO , SiO and MgO in the presence of a carbon 2 3 2 tertiary or sterically hindered amine [10]. According to Idem source. Among these support materials, MgO has attracted et  al. [11], blending a primary amine (MEA) with a tertiary much attention because it can easily be removed during acid amine (MDEA) will significantly increase the absorption rate treatment. Various carbon sources have been used in CVD, and reduce the energy of regeneration (heat duty) when such as acetylene, propylene, ethylene, methane or carbon compared with the single-solvent system. Sakwattanapong monoxide [23, 24]. In this work, we prepared the CNTs using et al. [12] emphasized that the advantages of the bi-solvent CH /CO as a carbon source. The mixtures of CH/CO allows 4 2 4 2 blend depend on the concentration, mixing ratio and mo- adjusting the carbon yield, purity and stability of CNTs [25]. lecular properties of each individual solvent. Narku-Tetteh Furthermore, gaseous CH and CO are major contributors to 4 2 et  al. [13] studied the selection of amine components for the accumulation of greenhouse gases. Therefore, the trans- blending to determine an optimum formulated amine formation of two major greenhouse gases (CH and CO ) into 4 2 solvent for CO capture. The influence of various structures a useful and valuable product (CNTs) is one of the methods of amines on CO absorption/desorption was conducted to being pursued in the current research. develop rational criteria for selecting components to form As a result of its tremendous advantages in terms of an amine blend. A  high absorption parameter and an ex- electron density on its surface and the large surface area, cellent desorption performance were achieved by blending it was decided to apply CNTs for CO absorption in a BEA– AMP with butylethanolamine (BEA–AMP). AMP bi-blend. Consequently, various catalysts consisting A combination of heterogeneous catalysts and newly de- of CNTs, K/MgO and their mixtures in various ratios were veloped amine solvents has been developed in recent years. tested to determine their potential as CO -absorption cata- Idem et al. [14] were the first to report the application of solid lysts in a BEA–AMP bi-blend solvent in terms of the CO - catalysts in the amine-regeneration process. The use of solid absorption rate and approach to equilibrium loading in a catalysts significantly reduced the operating temperature for CO batch reactor. All these catalysts were compared to the the solvent regeneration process from 120–140°C to 75–95°C, scenario without a catalyst. All the catalysts were charac- and consequently decreased the heat duty. Therefore, the terized and analysed by means of X-ray diffraction (XRD), development of amine solvents and heterogeneous cata- Brunauer–Emmet–Teller (BET) analysis, Raman spectra lysts appears to be the new research trend to enhance CO analysis, temperature-programmed desorption of carbon loading capacity and reduce the energy requirement for dioxide (CO -TPD), field-emission scanning electron micros- regeneration [15]. The addition of solid basic catalysts (e.g. copy with energy-dispersive X-ray spectroscopy (FESEM– CaO, MgO) to the absorption process has been one of the EDS), transmission electron microscopy (TEM) and thermal promising approaches to increase the CO -absorption cap- gravimetric analysis (TG-DTG). All these results are pre- acity, as CO molecules comprise a Lewis acid that accepts sented and discussed in this paper. electrons from Lewis bases; therefore, the presence of basic sites on the surface of an alkaline metal can enhance the CO -absorption capacity [1617 , ]. Consequently, the addition 2 1 Experimental section of a solid base catalyst into the CO -absorption process re- 1.1 Synthesis of carbon nanotubes sulted in considerable enhancement of the CO -loading cap- acity of the amine. In a recent study [18], it was observed that The growth of CNTs was obtained by the CVD method of a there was a tremendous improvement in CO absorption mixture gas of methane and carbon dioxide (CH/CO ) over 2 4 2 when a Lewis base catalyst (K/MgO) was used to promote ab- Ni/MgO as a combination of catalyst and support. The Ni/ sorption. This was attributed to the electron-donating abilityMgO catal yst was prepared by the impregnation method Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 253 using Ni(NO ) .6H O and MgO as catalyst precursors. The Flow meter 3 2 2 prepared catalyst was dried at 110°C in an oven and cal- cined at 500°C for 2h. For the production of CNTs, the obtained catalyst (2 g) was packed into a stainless tube (re- actor) that was placed in the central region of a vertical Stainless furnace. The reactor was flushed with N to drive air out tube while it was heated at a 10°C/min rate until the desired CH temperature (500°C) was attained. After the attainment of CO Catalyst 2 g Bubble the desired temperature, 10% H /N was introduced into 2 2 flow the furnace at a 50-mL/min rate for 1 h so as to generate active Ni particles on the MgO support. After the reduc- tion, the furnace was heated up to 850°C and CH /CO was 4 2 introduced into the reactor at a 100-mL/min rate for 1  h. During the reaction, CNTs were formed and deposited onto the catalyst surface. For the final step, N was passed for Fig. 1: The schematic diagram of the CVD instrument a period of 3  h during which the furnace was allowed to cool down to room temperature. The schematic picture of the CVD instrument is presented in Fig. 1. The as-prepared CNTs were purified by washing in 5 M HCl for 4 h at 40°C H O out to remove Ni and other catalyst components present in the CNTs. The product was filtered and washed successively with distilled water, dried at 110°C and then calcined at 500°C for 2 h in N atmosphere. CO in 1.2 Absorption experiment First, magnesium hydroxide (Mg(OH), Alfa Aesar, H O in CO 95–100.5%) was calcined in air atmosphere at 600°C for 4  h to form MgO and then impregnated with KOH solu- tion followed by drying at 110°C and calcination at 500°C for 2  h. The KMgO/CNTs catalysts with different ratios (5:0, 4:1, 3:2, 2:3, 1:4 and 0:5) were prepared by physical mixing. The CO -absorption experiment was performed using Catalyst the experimental set-up shown in Fig. 2. The apparatus was Oil bath heater Magnetic stirrer made up of a three-necked round-bottomed flask equipped with a water-cooled condenser to recover the amine, a thermometer to measure the amine-solution temperature and a gas-dispersion tube for feeding the CO gas. In CO - 2 2 absorption experiments, 100  mL of fresh aqueous amine Fig. 2: Scheme diagram of the batch reactor solution containing 2 M butylethanolamine (BEA, Sigma Aldrich, >98%) and 2 M 2-amino-2-methyl-1-propanol (AMP, 1.3 Calculation Acros Organics, 99%) was injected into a round-bottomed flask. A total of 5 g of catalyst in a basket was added into The CO loading can be calculated in the liquid phase by the aqueous amine to compare that without a catalyst. the following equation: The round-bottomed flask containing the amine solution CO loading = mole of CO / mole of amine(1) 2 2 and catalyst was immersed into the oil bath using a mag- The CO -absorption rate can be calculated in the liquid netic stirrer. The oil-bath temperature was maintained at phase by the following equation: 40  ±  1°C. The 15% CO balanced with 85% N (purchased 2 2 from Praxair, Canada) was introduced to the reactor con- CO absorption rate = mole of CO / mole of amine • min 2 2 (2) tinuously through the dispersion tube at a flow rate 200 mL/ min. During the experiment, 1  mL of amine solution was The absorption efficiency can be calculated by the fol- taken every 30 min in order to measure the CO loading by lowing equations: titration using 1 N HCl until the CO loading in the amine 2 Absorption efficiency (%) = (X − X )/X CO CO CO 2,in 2,out 2,in solution was constant. The CO loading at equilibrium rep- (3) where X is the amount of CO (in feed, off gas) resents the equilibrium solubility of CO . 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 254 | Clean Energy, 2019, Vol. 3, No. 4 before growing CNTs, after growing CNTs and after puri- 1.4 Catalyst characterization fication with acid. In the XRD pattern of the catalyst be- The surface morphology and actual elemental content of fore growing CNTs, only the XRD peaks of NiO-MgO at 2θ of the purified CNTs were examined via a FESEM equipped 37.1°, 42.9°, 62.5°, 75.2° and 78.9° were observed. The peaks with EDS and TEM using a JEOL JSM-7610F-XMaxN and JEOL of the NiO and MgO solids are clearly identified in the XRD 2100F instrument, respectively. Raman spectra of the CNTs spectrum. It should be noted that it is very difficult to de- were performed in a T64000 Raman spectrometer (Horiba termine the relative intensity of both NiO and MgO, as their JobinYvon, France) at 532-nm laser excitation. The thermo 2+ 2+ peaks overlap. This is because the Ni and Mg ions have gravimetric analysis (TG-DTG, METTLER TOLEDO STARe) 2+ similar valences, ionic radius values [r (Ni)  =  0.07  nm system operated with a flow of air at 10°C/min. The BET 2+ and r (Mg ) = 0.065 nm] and crystal cell dimensions. Thus, technique was employed to determine the specific surface the NiO and MgO components in the catalyst can easily area, pore volume and pore size of the catalysts. Prior to form a Ni Mg O solid solution [26]. For the catalyst after x 1-x each measurement, the catalysts were degassed under N growing CNTs, the XRD pattern shows that there are two flow for 2 h at 200°C using the Micrometrics FlowPrep060. intense peaks at the 2θ of 26.1° and 44.7°. Both the peaks The degassed catalysts were analysed at –196°C in the can be assigned with the graphitic phase of carbon [27 28 , ]. Micrometrics ASAP 2020 multi-point BET surface-area Moreover, the XRD peaks of metallic Ni were also observed analyser. The specific surface areas of the catalysts were at the 2θ of 44.5°, 52.1° and 76.6°, which were attributed determined by N adsorption/desorption. XRD patterns of to the reduction of NiO in the fresh catalyst. Then, the each catalyst were examined using a D8 ADVANCE X-ray catalyst was purified by washing with acid. The complete diffractometer (Bruker) equipped with Cu Kα radiation amorphous nature of the CNTs was observed. It means (k = 1.5406 A°) operated at 40 kV and 40 mA in the range of that the metals were removed completely. This result is 15°–80°. CO -TPD was performed in a ChemBET 3000 TPR/ also supported by the selected area of the catalyst surface TPD instrument. Typically, 0.05  g sample was pre-treated obtained from FESEM–EDS. in a quartz U-type reactor under a helium flow of 50  ml/ The FESEM images of the surface morphologies of min at 200°C for 1 h. The sample was then cooled down to the prepared CNTs are shown in Fig. 4. Fig. 4a shows the room temperature followed by the introduction of 3% (v/v) FESEM image of carbon nanotubes at low magnification, CO in N at a flow rate of 100 mL/min into the reactor for 2 2 while Fig. 4b shows the FESEM image of carbon nano- 4 h. The temperature was consequently gradually raised to tubes at high magnification. From these micrographs, it 900°C at a heating speed of 10°C/min under a helium flow can be observed that the CNTs are oriented randomly in of 60  mL/min. The CO desorbed was monitored using a the bending and twisting forms of tubes with few visible thermal-conductivity detector. defects. The prepared CNT surface was approximately smooth, having long and continuous tubes, which have an average diameter of 19 nm. EDS Fig. 4c was employed 2 Results and discussion to identify the concentration of catalyst embedded in the 2.1 Characterization of CNTs CNT. The EDS spectra of the synthesized CNTs indicate that the CNTs contain only carbon, indicating that the Carbon nanotubes were prepared by the thermal CVD metals were eliminated after washing with acid. method using Ni as a catalyst and CH /CO as a carbon 4 2 Thermogravimetric analysis (TGA) and the derivative source. MgO was used as a support for these catalyst thermogravimetric (DTG) curve of the weight-loss analysis nanoparticles.F ig. 3 shows the XRD patterns of catalysts are in most instances used to investigate the thermal sta- bility as well as the relative amounts of CNTs and catalyst NiO-MgO present in the product synthesized. The TG-DTG profile of the prepared CNTs before purification and after purifica- Ni tion are presented in Fig. 5. The TG-DTG curves show that Carbon the weight loss started at 480°C to 700°C, which was at- tributed to the oxidation of CNTs [29]. The temperature at (a) which the CNTs are oxidized is an index of its stability. In the case of the CNTs before purification, it is clear from (b) this curve that the sample is composed of CNTs but com- prises 40% catalyst particles at the end of the heating of the sample in air. This shows that the prepared CNTs (c) without purification contain catalyst particles that can be removed partly by washing the sample with HCl. 10 20 30 40 50 60 70 80 TEM and Raman spectra were utilized in order to verify 2θ (DEGREE) the characteristics of the CNTs. The micro images from TEM using 50 and 20  nm as the given scale (Fig. 6) indi- Fig. 3: XRD patterns of the samples (a) before growing CNTs, (b) after growing CNTs and (c) after purification cated that CNTs exhibit few compartments like a bamboo INTENSITY (A.U) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 255 AB 19 mm Spectrum 1 12 34 56 78 9 Full Scale 10371 cts Cursor: 0.000 keV Fig. 4: FESEM–EDS images of the purified CNTs structure characterized by outer and inner diameters of 9–12 and 3–6  nm, respectively. Fig. 7 shows the Raman spectra for the CNTs. It is obvious that there are two peaks –1 located at 1294 and 1546  cm , commonly known as the D −1 and G bands. The peak near 1294 cm is the D band, which was attributed to the presence of distortion carbon, while –1 the peak near 1546  cm is the G band, which is related to the graphitic layers with the high crystallinity of the 40 (a) carbon. Moreover, the relative intensity ratio of the D and G bands (I /I ) can be ascribed to the degree of graphitiza- D G tion and crystal quality of the carbon, where a lower /I I (b) D G value implies fewer defects of the carbon. The /I I ratio of D G 100 200 300 400 500 600 700 800 900 the CNTs was evaluated to be 0.71. Temperature (°C) Fig. 5: TG-DTG profile of the CNTs (a) before and (b) after purification 2.2 Characterization of the catalysts The XRD patterns of the fresh catalysts are shown in Fig. all of the catalysts are presented in Table 1. The specific 8. The sharp diffraction peaks centred at the 2θ of 36.9°, surface area and pore volume of K/MgO were found to be 2 3 42.8°, 62.3°, 74.8° and 78.6° were directly indexed to the 37.5 m /g and 0.177 cm /g, respectively. Besides, pure CNTs face-centred cubic-phase crystalline structure of MgO. The have the highest surface area and pore volumes of 122.4 2 3 broad diffraction peak at a 2θ of 26.1° and 44.7° were at- m /g and 0.177 cm /g, respectively. The binary systems of tributed to the graphitic carbon. Besides, when CNTs were KMgO/CNTs catalysts exhibited higher BET surface areas added, the diffraction peak of the CNTs appeared and the than the single KMgO catalyst, which could be attributed MgO intensity decreased with increasing CNT content. to the presence of CNTs. The total pore volume also fol- The BET surface areas, pore volume and pore size were lowed the same trend as the surface area. measured by the physical nitrogen adsorption–desorption The strength and amount of the basic surface sites using the Barett–Joyner–Halenda method. The results for on the catalysts were determined by CO -TPD based on Weight loss (%) DTG (a.u.) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 256 | Clean Energy, 2019, Vol. 3, No. 4 12 nm 50 nm 20 nm Fig. 6: TEM images of the CNTs Carbon MgO (a) I /I = 0.71 D G (b) (c) 20 (d) (e) (f) 500 1000 1500 2000 2500 3000 10 20 30 40 50 60 70 80 –1 Raman shift (cm ) 2θ (degree) Fig. 7: Raman spectra of the CNTs Fig. 8: XRD patterns of the fresh catalysts (a) KMgO, (b) KMgO/CNTs (4:1), (c) KMgO/CNTs (3:2), (d) KMgO/CNTs (2:3), (e) KMgO/CNTs (1:4) and the CO absorption–desorption process as shown in Fig. (f) CNTs 9. The basic sites or total basic sites over the catalyst are the number of exposed active centres per unit mass for the strong basic sites increased from 0 to 0.594  mmol/g, CO absorption. The temperature of desorption and the indicating that the nature of the basic sites had changed maximum desorbed CO are illustrative of the strength considerably. The numbers of strong basic sites and total (weak, medium and strong basic strength) and number basic sites are considered to be the main parameters in of basic sites, respectively. The basic strength and basic high CO absorption. Therefore, the relative number of sites of the fresh KMgO/CNTs catalysts with various ratios medium basic sites compared with strong basic sites was were compared. The results revealed that the pure KMgO used as one of the key parameters that can influence showed only one desorption peak for CO at the tempera- the catalytic activity of the catalyst for CO absorption in ture range of 250–470°C, which corresponded to the inter - amine solutions. action of CO with sites of medium basic strengths by the 2+ 2– presence of oxygen in Mg and O pairs. Besides, the pure 2.3 CO absorption CNTs showed only one desorption peak for CO at high temperatures in the range of 580–900°C (i.e. >500°C), which According to the literature, the absorption of CO using corresponded to strong basic sites. For the binary systems aqueous amine is a complex reaction that involves many of KMgO/CNTs catalysts, the results showed that two de- factors. The reaction between primary and secondary sorption peaks of medium basic sites and strong basic alkanolamines with CO can be described through two sites (small but tailing peaks at >500°C) were observed for simultaneous reactions: nearly all the samples. Obviously, with a gradual increase + − Zwitterion formation : CO + RNH ↔ RNH COO 2 2 2 in CNT content, a broad peak at the high temperature ap- (4) pears, while the peak area of medium basic sites becomes + − Carbamate and protonated amine : RNH COO + RNH 2 2 smaller and shifts towards higher temperature ranges. − + The numbers of medium and strong base sites per square ↔ RNHCOO + RNH (5) metre were calculated and the results are listed in Table The reaction between a sterically hindered amine and CO 2. With the increase in the numbers of CNTs, the medium occurs through two simultaneous reactions: basic sites decreased from 0.771 to 0.238 mmol/g, whereas Intensity (a.u) Intensity (a.u) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 257 Table 1: Summarization of the specific surface area, pore volume and pore size of the fresh catalysts 2 3 Catalysts Surface area (m/g) Pore volume (cm /g) Pore size (nm) KMgO 37.5 0.177 18.86 KMgO/CNTs (4:1) 54.1 0.218 14.56 KMgO/CNTs (3:2) 69.5 0.277 12.26 KMgO/CNTs (2:3) 85.4 0.368 8.94 KMgO/CNTs (1:4) 107.4 0.479 5.90 CNTs 122.4 0.555 3.66 + − (6) Bicarbonate formation : CO + RNH ↔ RNH COO 2 2 2 + − + − (f) (7) RNH COO + H O ↔ RNH + HCO 2 2 3 3 (e) Bicarbonate formation by carbamate hydrolysis : CO + RNH 2 2 (8) (d) ↔ RNH COO + + − − (c) (9) RNH COO + NH ↔ RNHCOO + RNH 2 2 3 + + − (b) RNHCOO + H O + RNH ↔ RNH + RNH + HCO 2 3 2 3 3 (10) (a) The formation of bicarbonate occurs in the sterically hin- 100 200 300 400 500 600 700 800 900 dered amine due to the steric character caused by the Temperature (°C) bulky group adjacent to the amino group in the amine. Fig. 9: CO -TPD profile of the fresh catalysts (a) KMgO, (b) KMgO/CNTs Therefore, the carbamate can easily undergo hydrolysis, 2 (4:1), (c) KMgO/CNTs (3:2), (d) KMgO/CNTs (2:3), (e) KMgO/CNTs (1:4) and leading to the formation of bicarbonate and free amine (f) CNTs molecules [8 ]. For the catalyst-screening process for CO absorption, resulting in an enhanced CO loading capacity for CO ab- 2 2 15% CO was employed and the reaction temperature was sorption. This effect is combined with the increase in strong set at 40°C. Fig. 10a shows the profiles of CO loading in basic sites and total basic sites with the increase in the ratio BEA–AMP aqueous solutions with catalysts and without of CNTs in KMgO/CNT mixture .The overall implication is any catalyst. The overall period of absorption to reach CO - the generation of a large number of exposed active centres loading equilibrium takes 420  min in the case without a absorption. The catalyst provides more per unit mass for CO catalyst. With the addition of K/MgO, the overall time was electrons to be available to the amine, which facilitates CO reduced from 420 to 330 min. From the results, it is clear absorption, leading to an increase in the rate of carbamate that, at the same absorption time, the addition of the cata- and bicarbonate formation. A  detailed mechanism is ex- lyst into the amine solution led to increased CO loading plained in Scheme 1. Subsequently, a pair of electrons from and this reached equilibrium more quickly than without the amine is transferred to the catalyst−CO -formed species a catalyst. The addition of K/MgO catalysts to the absorp- (intermediate species) to restore the electrons of the catalyst tion process increased the CO-absorption capacity, as CO as well as to generate the zwitterion formation [18]. 2 2 molecules comprise a Lewis acid that accepts electrons Fig. 11 shows the CO -absorption rate in BEA–AMP from Lewis bases; therefore the presence of basic sites on aqueous solutions with catalysts and with no cata- the surface of metal oxide enhances the CO -adsorption lyst. The CO-absorption rate was calculated as the CO - 2 2 capacity. loading change per minute until it reached equilibrium The CO loading over the KMgO/CNTs catalysts with CO loading. From the results, it is clear that the introduc- various ratios was also determined. As can be seen from tion of the catalysts into the absorption system helped Fig. 10b, an increment in the CNT number increased the CO to increase the CO-absorption rate. The CO-absorption 2 2 loading at the same absorption time and reached equilib- rate increased from 0.0014  mol CO/mol amine/min in rium more quickly as compared with KMgO and without a the case without a catalyst to 0.0018 mol CO /mol amine/ catalyst. The KMgO/CNTs at the ratio of 1:4 had the highest min with the addition of KMgO catalyst. With the add- CO loading at the same absorption time and reached equi- ition of CNTs to K/MgO, the CO-absorption rate in- librium at 240 min. This can be explained on the basis of the creased from 0.0018 mol CO/mol amine/min in the case structure of CNTs, which consists of graphene layers with of the K/MgO catalyst to 0.0025  mol CO /mol amine/ a large BET specific surface area and large active area that min. The CO -absorption rate ranked as follows: KMgO/ help in increasing the exposure of the amine molecules, CNTs (1:4)  >  CNTs  >  KMgO/CNTs (2:3)  >  KMgO/CNTs Signall (a.u) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 258 | Clean Energy, 2019, Vol. 3, No. 4 OH NH H C H OH AMP O N C – – – e e e N H O N H BEA – – e e HO HO HO Carbamate Ammonium ion HO N H BEA O C HH N HO – – – C C N H N e e e AMP O – – e e HO HO HO Carbamate Ammonium ion Scheme 1: A possible catalytic CO -absorption pathway over the catalyst. Table 2: Basic sites of the catalysts Catalysts Medium sites (mmol 2/g) Strong sites (mmol /g) Total basic sites (mmol /g) CO CO2 CO2 KMgO 0.771 0 0.771 KMgO/CNTs (4:1) 0.451 0.117 0.568 KMgO/CNTs (3:2) 0.252 0.204 0.456 KMgO/CNTs (2:3) 0.251 0.303 0.554 KMgO/CNTs (1:4) 0.238 0.594 0.849 CNTs 0 0.657 0.657 (3:2) > KMgO/CNTs (4:1) > KMgO > no catalyst. The KMgO/ physical properties (surface area, pore volume and pore CNTs 1:4 catalyst shows the highest CO -absorption rate size) and chemical properties (total basic sites, strong (or increase of 44%) when compared with the system with basic sites and medium basic sites) of the solid catalysts. no catalyst. Fig. 12 shows the results for CO -absorption ef- The relationship between the absorption rate and the ficiency in BEA–AMP aqueous solutions with catalysts and physical and chemical properties of the solid catalysts are without any catalyst. From the results, the CO -absorption shown in Equations (11) and (12): efficiency with the addition of catalysts increased when Absorption rate = 0.003255 +(−0.000000059) ∗ A compared with no catalyst. The CO efficiency increased +(−0.00108) ∗ B +(−0.000067) ∗ C from 24% in the case without a catalyst to 27% with the (11) addition of K/MgO catalyst. With the addition of CNTs to where parameters A, B and C are surface area, pore volume K/MgO, the CO -absorption efficiency increased tremen- 2 and pore size, respectively. dously. The CO -absorption efficiency ranked as follows: Absorption rate = 0.002356 + 0.023692 ∗ D KMgO/CNTs (1:4) > CNTs > KMgO/CNTs (2:3) > KMgO/CNTs +(−0.02379) ∗ E +(−0.0244) ∗ F (3:2) > KMgO/CNTs (4:1) > KMgO > no catalyst. The KMgO/ (12) CNTs (1:4) catalyst show the highest CO -absorption ef- where parameters D, E and F are total basic sites, strong ficiency (or increase of 15.3%) when compared with no basic sites and medium basic sites, respectively. catalyst. The relationship between the CO-absorption efficiency In order to understand which properties of the solid and the physical and chemical properties of the solid cata- catalysts played the most significant role in promoting lysts are shown in Equations (13) and (14): the CO -absorption rate and absorption efficiency, a re- Absorption ef f iciency = 44.69023 +(−0.15283) ∗ A gression analysis was conducted and used to determine +(25.31378) ∗ B +(−0.891) ∗ C any possible correlation between absorption rate and ab- (13) sorption efficiency with properties of the catalysts. The Absorption ef f iciency = 31.13859 + 407.38 ∗ D CO -absorption rate and absorption efficiency obtained +(−402.877) ∗ E +(−413.286) ∗ F from the experiment were regressed against both the (14) e Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 259 A parity plot for the absorption rate and absorption effi- absorption rate and chemical properties (Fig.13a) of the solid ciency with physical and chemical properties is shown in catalysts showed a smaller value for AAD (1.05%) than the Fig. 13. As seen in Fig. 13, the experimental values versus correlations between the absorption rate and physical prop- those predicted using the corresponding correlations for erties (AAD = 2.59%). The correlations between the absorp- tion efficiency and chemical properties (Fig.13b) of the solid catalysts also showed a smaller value for AAD (0.48%) than (a) 0.6 the correlations between the absorption efficiency and phys- ical properties (AAD = 1.76%). This implied that the chemical 0.5 properties (number of basic sites and basic strength) played No catalyst a bigger role in the CO-absorption rate and absorption effi- 0.4 2 KMgO ciency by a solid basic catalyst than the physical properties. KMgO/CNTs(4:1) The degree of correlation obtained for the relationship be- 0.3 KMgO/CNTs(3:2) tween the CO -absorption rate and chemical properties of the KMgO/CNTs(2:3) 0.2 solid catalysts was 99%, which was higher than that obtained KMgO/CNTs(1:4) between the absorption rate and physical properties of the CNTs 0.1 solid catalysts (81%). Also, the CO -absorption efficiency cor - relations with chemical properties had a degree of correlation 2 2 30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 (R ) of 99%, whereas, with physical properties, R was 74%. The Time (min) catalyst appears to be working more on AMP, making it more effective than BEA. Based on its structure, AMP has a sterical (b) 0.6 hindered alkyl group, which tends to form unstable carba- mates whose hydrolysis leads to the formation of bicarbon- ates. The bicarbonates easily release CO to form free amine 0.55 No catalyst molecules for further reaction with incoming CO . KMgO The XRD patterns of the catalysts after 6 h of use for CO KMgO/CNTs(4:1) absorption are shown in Fig. 14. For all of the catalysts in use 0.5 KMgO/CNTs(3:2) with the presence of KMgO in the mixture, diffraction peaks KMgO/CNTs(2:3) at the 2θ° angle of 18.5°, 38.1°, 51.1°, 58.8°, 62.2°, 68.7° and KMgO/CNTs(1:4) 0.45 72.2° were observed, which were assigned to the crystal- CNTs line structure of Mg(OH). The presence of Mg(OH) peaks 2 2 is related to the reaction of MgO with water in the amine 0.4 180 210 240 270 300 330 360 390 420 450 480 solution during absorption [30]. Furthermore, a higher inten- Time (min) sity of MgCO patterns was also noticed at the 2 of 44.4° θ and 51.8°, when the CNTs were introduced to the KMgO. By Fig. 10: (a) CO -absorption profile of the 2 M BEA + 2 M AMP at 40°C with increasing the CNTs content, the diffraction peaks of MgCO and without catalysts based on (a) the entire run (30–450 min) and (b) a portion of the run (180–450  min) to show explicitly when each run intensity increased while diffraction peaks of Mg(OH) inten- reaches equilibrium. sity decreased. This was attributed to the strong basic sites 0.003 0.0025 0.002 0.0015 0.001 0.0005 no catalyst KMgO KMgO/CNTs KMgO/CNTs KMgO/CNTs KMgO/CNTs CNTs (4:1) (3:2) (2:3) (1:4) Catalyst Fig. 11: CO absorption rate of 2 M BEA + 2 M AMP solution with and without catalysts. Reaction conditions: 5.0 g of catalyst, BEA–AMP 100 mL, CO 2 2 flow rate at 200 mL/min and temperature 40°C CO Loading (mol CO /mol amine) CO Loading (mol CO /mol amine) 2 2 2 2 CO absorption rate (mol CO /mol amine-min) 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 260 | Clean Energy, 2019, Vol. 3, No. 4 no catalyst KMgO KMgO/CNTs KMgO/CNTs KMgO/CNTs KMgO/CNTs CNTs (4:1) (3:2) (2:3) (1:4) Catalyst Fig. 12: CO -absorption efficiency of 2 M BEA + 2 M AMP solution with and without catalysts. Reaction conditions: 5.0 g of catalyst, BEA–AMP 100 mL, CO flow rate at 200 mL/min and temperature 40°C. Carbon Mg(OH) MgCO MgO 0.003 2 3 (a) 0.0025 (f) 0.002 (e) 0.0015 (d) Physical AAD (%) = 2.59 (c) 0.001 R = 81% (b) Chemical AAD (%) = 1.05 0.0005 R = 99% (a) 10 20 30 40 50 60 70 80 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 Experimental value 2θ (degree) (b) Fig. 14: Representative XRD patterns of the spent catalysts (a) CNTs, (b) KMgO, (c) KMgO/CNTs (4:1), (d) KMgO/CNTs (3:2), (e) KMgO/CNTs (2:3) and (f) KMgO/CNTs (1:4) for commercializing any catalytic process by reducing Physical AAD (%) = 1.76 the cost of production and for real application on a large R = 74% or industrial scale. In this research, the reusability of the Chemical AAD (%) = 0.48 10 2 catalyst was studied by calcination and without calcin- R = 99% ation (dried in the oven at 110°C overnight) as shown in Fig. 15. Calcination of the used catalyst was performed in 0 510 15 20 25 30 35 40 45 air at 300°C for 2 h to burn away any deposition of organic Experimental value contaminants. The results revealed that the CO -loading rate of the catalyst with calcination was not significantly Fig. 13: Parity chart of the predicted values versus experimental values of (a) the absorption rate and (b) the absorption efficiency changed and reached equilibrium at the same time as the fresh catalyst. On the other hand, the CO -loading rate of the catalyst without calcination decreased and reached of the CNTs aiding a strong interaction between MgO and equilibrium more slowly (360 min) when compared to the CO to form MgCO. This reaction hindered the formation of 2 3 fresh catalyst and the catalyst with calcination. This was Mg(OH) . Based on the literature, both Mg(OH) and MgCO 2 2 3 attributed to the surface active sites of the catalyst being are equally active phases for CO absorption in amines [31]. covered by organic contaminants, which could create a layer that covers the surface of MgO, Mg(OH) or MgCO 2 3 against CO . 2.4 Catalyst reusability and stability After the first run, the KMgO/CNTs (1:4) catalyst was An important characteristic feature of a catalyst is its collected from the reaction, dried in an oven at 110°C and reuse property. This reusability of the catalyst is a factor calcined in the atmosphere at 350°C for 2  h to obtain a Predicted value Predicted value CO absorption efficiency (%) Intensity (a.u) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 261 0.7 CNTs yielded very-high-performance catalysts for CO absorption with BEA–AMP solvent. 0.6 (3) The increase in the number of CNTs to KMgO reduces 0.5 the overall absorption periods to reach equilibrium as compared with KMgO and without a catalyst, mainly 0.4 due to the increasing numbers of strong basic sites. The KMgO/CNTs(1:4) 0.3 Reuse with calcined increased number of CNTs to KMgO also increased the Reuse without calcined number of exposed active centres per unit mass for CO 0.2 absorption. Furthermore, it also generated a large spe- 0.1 cific surface area and pore volume, which are beneficial in increasing the exposure of amine molecules. (4) The results obtained from this study are intended to 30 60 90 120 150 180 210 240 270300 330 360 Time (min) show clearly the relative performance of the basic cat- alyst. The best catalysts selected from this study will Fig. 15: Reusability of the catalyst be subjected to pilot plant studies. The addition of this catalyst into a commercially packed absorber is essen- tial towards the industrial application of CO capture. 0.003 2 0.0025 Acknowledgements 0.002 The financial support provided by Queen Elizabeth II Diamond Jubilee Scholarship (QES) and the Natural Science and Engineering 0.0015 Research Council of Canada is gratefully acknowledged. 0.001 References [1] Wang  Y, Zhao  L, Otto  A, et  al. A review of post-combustion 0.0005 CO capture technologies from coal-fired power plants. Energy Procedia 2017; 114:650–65. [2] Cuéllar-Franca  RM, Azapagic  A. Carbon capture, storage and 12 34 5 utilisation technologies: a critical analysis and comparison of Number of cycle (Times) their life cycle environmental impacts. J CO Util 2015; 9:82–102. Fig. 16: Recycling tests on the catalytic activity of KMgO/CNTs (1:4) for [3] Song C, Liu Q, Ji N, et al. Alternative pathways for efficient CO CO absorption. Reaction conditions: 5.0 g of catalyst, BEA–AMP 100 mL, 2 capture by hybrid processes—a review. Renew Sust Energ Rev CO flow rate at 200 mL/min and temperature 40°C 2 2018; 82:215–31. [4] Dutcher B, Fan M, Russell AG. Amine-based CO capture tech- nology development from the beginning of 2013—a review. regenerated catalyst. The activity property of the regener - ACS Appl Mater Interfaces 2015;7:2137–48. ated catalyst was tested under the same operational pro- [5] Wu X, Yu Y, Qin Z, et al. The advances of post-combustion CO cess as the fresh catalyst. The results for repeated usage capture with chemical solvents: review and guidelines. Energy (five times) are shown in Fig. 16. It was found that the CO- Procedia 2014; 63:1339–46. absorption rate only changed slightly after being repeat- [6] Lee CS, Ong YL, Aroua MK, et al. Impregnation of palm shell- based activated carbon with sterically hindered amines for edly used five times. CO adsorption. Chem Eng J 2013; 219:558–64. [7] Luis P. Use of monoethanolamine (MEA) for CO capture in a global scenario: consequences and alternatives. Desalination 3 Conclusions 2016; 380:93–9. [8] Bougie F, Iliuta MC. Sterically hindered amine-based absorb- (1) Carbon nanotubes have been synthesized successfully ents for the removal of CO from gas streams. J Chem Eng Data by the decomposition of CH/CO over Ni/MgO catalyst. 4 2 2012; 57:635–69. Raman spectral analyses as well as TEM images indi- [9] Zheng C, Tan J, Wang YJ, et al. CO solubility in a mixture ab- cate that CNTs were produced. The CNTs exhibit few sorption system of 2-amino-2-methyl-1-propanol with glycol. compartments like a bamboo structure characterized Ind Eng Chem Res 2012; 51:11236–44. [10] Nwaoha C, Saiwan C, Tontiwachwuthikul P, et al. Carbon di- by outer and inner diameters of 9–12 and 3–6  nm, oxide (CO ) capture absorption desorption capabilities of respectively. 2 2-amino-2-methyl-1-propanol (AMP) piperazine (PZ) and (2) The CO -absorption rate and absorption efficiency with monoethanolamine (MEA) tri-solvent blends. J Nat Gas Sci Eng catalysts and without catalyst were studied. With the 2016; 33:742–50. addition of a catalyst into the amine solution, the CO - [11] Idem R, Wilson M, Tontiwachwuthikul P, et al. Pilot plant studies absorption rate and absorption efficiency were higher of the CO capture performance of aqueous MEA and mixed than without a catalyst. The combination of KMgO and MEA/MDEA solvents at the University of Regina CO capture CO absorption rate CO Loading (mol CO /mol amine) (mol CO /mol amine-min) 2 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 262 | Clean Energy, 2019, Vol. 3, No. 4 technology development plant and the boundary dam CO cap- [22] Magrez A, Seo JW, Smajda R, et al. Catalytic CVD synthesis of ture demonstration plant. Ind Eng Chem Res 2006; 45:2414–20. carbon nanotubes: towards high yield and low temperature [12] Sakwattanapong R, Aroonwilas A, Veawab A. Behavior of reboiler growth. Materials (Basel) 2010; 3:4871–91. heat duty for CO capture plants using regenerable single and [23] Alvarez WE, Pompeo F, Herrera JE, et al. Characterization of blended Alkanolamines.Ind Eng Chem Res 2005; 44:4465–73. single-walled carbon nanotubes (SWNTs) produced by CO [13] Narku-Tetteh J, Muchan P, Saiwan C, et al. Selection of com- disproportionation on Co-Mo catalysts. Chem Mater 2002; ponents for formulation of amine blends for post combustion 14:1853–8. CO capture based on the side chain structure of primary, sec- [24] Kong J, Cassell AM, Dai H. Chemical vapor deposition of me- ondary and tertiary amines. Chem Eng Sci 2017; 170:542–60. thane for single-walled carbon nanotubes. Chem Phys Lett [14] Idem R, Shi H, Gelowitz D, et al. Catalytic method and appar - 1998; 292:567–74. atus for separating a gaseous component from an incoming [25] Corthals S, Noyen JV, Geboers J, et al. The beneficial effect of gas stream. 2017, US Patent 9,586,175. CO in the low temperature synthesis of high quality carbon [15] Shi  H, Naami  A, Idem  R, et  al. Catalytic and non catalytic nanofibers and thin multiwalled carbon nanotubes from CH solvent regeneration during absorption-based CO capture over Ni catalysts. Carbon 2012; 50:372–84. with single and blended reactive amine solvents. Int J Greenh [26] Jafarbegloo  M, Tarlani  A, Wahid  Mesbah  A, et  al. NiO–MgO Gas Con 2014; 26:39–50. solid solution prepared by sol–gel method as precursor for [16] Hamid  RR, Abdelhamid  S. A highly efficient CaO-based CO Ni/MgO methane dry reforming catalyst: effect of calcin- sorbent prepared by a citrate-assisted sol–gel technique. Chem ation temperature on catalytic performance. Catal. Lett 2016; Eng J 2015; 262:913–20. 146:238–48. [17] Li  L, Wen  X, Fu  X, et  al. MgO/AlO sorbent for CO capture. [27] Prasongthum N, Xiao R, Zhang H, et al. Highly active and stable 2 3 2 Energy Fuels 2010; 24:5773–80. Ni supported on CNTs-SO fiber catalysts for steam reforming [18] Narku-Tetteh J, Afari DB, Coker J, et al. Evaluation of the roles of ethanol. Fuel Process Technol 2017; 160:185–95. of absorber and desorber catalysts in the heat duty and heat of [28] Wang S, Yin Q, Guo J, et al. Influence of Ni promotion on liquid CO desorption from butylethanolamine−2 amino-2-methyl- hydrocarbon fuel production over Co/CNT catalysts. Energy 1-propanol and monoethanolamine−methyldiethanolamine Fuels 2013; 27:3961–8. solvent blends in a bench-scale CO capture pilot plant. Energy [29] Ezekiel ZD, Ntaote DS, Thema FT, et al. Raman and TGA study Fuels 2018; 32(9):9711–26. of carbon nanotubes synthesized over Mo/Fe catalyst on alu- [19] Khalil SH, Aroua MK, Daud WMAW. Study on the improvement minium oxide, calcium carbonate and magnesium oxide sup- of the capacity of amine-impregnated commercial activated port. Chem Sci Trans 2013; 2:1160–73. carbon beds for CO adsorbing. Chem Eng J 2012; 183:15–20. [30] Hanlon JM, Bravo Diaz L, Balducci G, et al. Rapid surfactant-free [20] Lu  C, Bai  H, Wu  B, et  al. Comparative study of CO capture synthesis of Mg(OH) nanoplates and pseudomorphic dehydra- 2 2 by carbon nanotubes, activated carbons, and zeolites. Ener gy tion to MgO. Cryst Eng Comm 2015; 17:5672–9. Fuels 2008; 22:3050–6. [31] Shi  H, Zhou  Y, Zuo  Y, et  al. Heterogeneous catalysis of CO - [21] Creamer  AE, Gao  B, Zhang  M. Carbon dioxide capture using diethanolamine absorption with MgCO and CaCO and com- 3 3 biochar produced from sugarcane bagasse and hickory wood. paring to non-catalytic CO -monoethanolamine interactions. Chem Eng J 2014; 249:174–9. React Kinet Mech Cat 2017; 122:539–55. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Clean Energy Oxford University Press

Application of carbon nanotubes prepared from CH4/CO2 over Ni/MgO catalysts in CO2 capture using a BEA–AMP bi-solvent blend

Loading next page...
 
/lp/oxford-university-press/application-of-carbon-nanotubes-prepared-from-ch4-co2-over-ni-mgo-qNq4VL0dQ1
Publisher
Oxford University Press
Copyright
© The Author(s) 2019. Published by Oxford University Press on behalf of National Institute of Clean-and-Low-Carbon Energy
ISSN
2515-4230
eISSN
2515-396X
DOI
10.1093/ce/zkz012
Publisher site
See Article on Publisher Site

Abstract

Carbon nanotubes (CNTs) were synthesized by the chemical vapour deposition of methane and carbon dioxide over a Ni/MgO catalyst. The synthesized CNTs were then mixed with K/MgO catalyst at different ratios and used as the catalyst for CO absorption in butylethanolamine-2-amino-2-methyl-l-propanol bi-solvent blend. The catalysts were characterized using X-ray diffraction, scanning electron microscopy, butylethanolamine, thermal gravimetric analysis and temperature-programmed desorption of carbon dioxide in order to determine the characteristics responsible for good CO -absorption performance. The results showed that, with the addition of a catalyst into the amine solution, the amine reached equilibrium CO loading faster than without a catalyst. Also, the increase in the CNT content of the KMgO/CNTs catalyst made the CO absorption reach equilibrium much more quickly compared with just KMgO alone and without a catalyst. The KMgO/CNTs at a ratio of 1:4 yielded the fastest time to reach CO -loading equilibrium at 240 min, which was mainly due to the increase in strong basic sites as well as the highest total basic sites with an increase in CNT content. In addition, because of the extremely large specific surface area and pore volume generated due to the CNT, the number of exposed active centres per unit mass increased tremendously, leading to a corresponding tremendous increase in CO absorption. Keywords: carbon nanotubes; KMgO; catalyst; CO capture; absorption CO capture based on absorption using liquid amine solu- Introduction tions is currently the most mature and effective technology The continuous rise in CO concentration in the atmosphere applied for the removal of CO . The CO -absorption process 2 2 has generated worldwide concerns about global warming typically occurs via a continuous cycle, with CO absorption and climate-change issues. To control CO emission and slow 2 operating at around 40°C and simultaneous amine regener - down global warming, it is necessary to eliminate CO from ation upon heating the CO-loaded amine solution at a tem- major point sources such as fossil-fuelled power plants [ ]. CO 1 capture, utilization and storage (CCUS) have been proposed to perature of 120°C [4 , 5]. The solvents mainly employed in be a promising way to effectively control CO emissions [2, 3]. industrial processes include monoethanolamine (MEA) and Received: 29 March, 2019; Accepted: 27 April, 2019 © The Author(s) 2019. Published by Oxford University Press on behalf of National Institute of Clean-and-Low-Carbon Energy This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http:// creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, 251 provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 252 | Clean Energy, 2019, Vol. 3, No. 4 diethanolamine (DEA) but their weaknesses are corrosive- of K/MgO during the rate-determining step of CO absorp- ness and the high energy required for the process, especially tion in BEA–AMP as well as the high surface area provided during the regeneration of CO -loaded solvent (about 70–80% by K/MgO. Carbon-based materials such as activated carbon, of the running cost for a CO-capture plant) [, 67]. A  steric carbon nanofibers and carbon nanotubes (CNTs) are also hindrance amine such as 2-amino-2-methyl-l-propanol good capture materials due to their large surface area and (AMP) has been recommended to reduce the energy con- high-electron-density surface for facilitating gas absorption sumption. This is because AMP is a primary amine with OH [19,20,21]. The structure of CNTs consists of graphene layers, groups and a sterically hindered alkyl group. The steric c- har which is beneficial in terms of increasing the exposure of acter tends to form unstable carbamates whose hydrolysis amine molecules, thereby resulting in the enhancement leads to the formation of bicarbonates, which easily release of the CO -loading capacity of amine [20]. CNTs are typic- free amine molecules for further reaction with CO resulting ally prepared by chemical vapour deposition (CVD), arc dis- in increasing CO -absorption capacity [, 89]. Therefore, AMP charge and laser ablation. Among these three techniques, offers a lower energy requirement during the regeneration the CVD method has the advantage of being a simple pro- process. A bi-solvent blend has also been found to enhance cess that produces high purity of the CNTs and synthesizes the capture efficiency and reduce the energy of regener- high quantity at lower costs [22]. The production of CNTs ation. The combination of each amine offers the advantages via the CVD method is mostly accomplished using transi- of the high absorption capacity of a reactive amine (primary tion metals (Fe, Co and Ni) supported on different mater - or secondary amine) and low solvent regeneration cost of a ials, such as AlO , SiO and MgO in the presence of a carbon 2 3 2 tertiary or sterically hindered amine [10]. According to Idem source. Among these support materials, MgO has attracted et  al. [11], blending a primary amine (MEA) with a tertiary much attention because it can easily be removed during acid amine (MDEA) will significantly increase the absorption rate treatment. Various carbon sources have been used in CVD, and reduce the energy of regeneration (heat duty) when such as acetylene, propylene, ethylene, methane or carbon compared with the single-solvent system. Sakwattanapong monoxide [23, 24]. In this work, we prepared the CNTs using et al. [12] emphasized that the advantages of the bi-solvent CH /CO as a carbon source. The mixtures of CH/CO allows 4 2 4 2 blend depend on the concentration, mixing ratio and mo- adjusting the carbon yield, purity and stability of CNTs [25]. lecular properties of each individual solvent. Narku-Tetteh Furthermore, gaseous CH and CO are major contributors to 4 2 et  al. [13] studied the selection of amine components for the accumulation of greenhouse gases. Therefore, the trans- blending to determine an optimum formulated amine formation of two major greenhouse gases (CH and CO ) into 4 2 solvent for CO capture. The influence of various structures a useful and valuable product (CNTs) is one of the methods of amines on CO absorption/desorption was conducted to being pursued in the current research. develop rational criteria for selecting components to form As a result of its tremendous advantages in terms of an amine blend. A  high absorption parameter and an ex- electron density on its surface and the large surface area, cellent desorption performance were achieved by blending it was decided to apply CNTs for CO absorption in a BEA– AMP with butylethanolamine (BEA–AMP). AMP bi-blend. Consequently, various catalysts consisting A combination of heterogeneous catalysts and newly de- of CNTs, K/MgO and their mixtures in various ratios were veloped amine solvents has been developed in recent years. tested to determine their potential as CO -absorption cata- Idem et al. [14] were the first to report the application of solid lysts in a BEA–AMP bi-blend solvent in terms of the CO - catalysts in the amine-regeneration process. The use of solid absorption rate and approach to equilibrium loading in a catalysts significantly reduced the operating temperature for CO batch reactor. All these catalysts were compared to the the solvent regeneration process from 120–140°C to 75–95°C, scenario without a catalyst. All the catalysts were charac- and consequently decreased the heat duty. Therefore, the terized and analysed by means of X-ray diffraction (XRD), development of amine solvents and heterogeneous cata- Brunauer–Emmet–Teller (BET) analysis, Raman spectra lysts appears to be the new research trend to enhance CO analysis, temperature-programmed desorption of carbon loading capacity and reduce the energy requirement for dioxide (CO -TPD), field-emission scanning electron micros- regeneration [15]. The addition of solid basic catalysts (e.g. copy with energy-dispersive X-ray spectroscopy (FESEM– CaO, MgO) to the absorption process has been one of the EDS), transmission electron microscopy (TEM) and thermal promising approaches to increase the CO -absorption cap- gravimetric analysis (TG-DTG). All these results are pre- acity, as CO molecules comprise a Lewis acid that accepts sented and discussed in this paper. electrons from Lewis bases; therefore, the presence of basic sites on the surface of an alkaline metal can enhance the CO -absorption capacity [1617 , ]. Consequently, the addition 2 1 Experimental section of a solid base catalyst into the CO -absorption process re- 1.1 Synthesis of carbon nanotubes sulted in considerable enhancement of the CO -loading cap- acity of the amine. In a recent study [18], it was observed that The growth of CNTs was obtained by the CVD method of a there was a tremendous improvement in CO absorption mixture gas of methane and carbon dioxide (CH/CO ) over 2 4 2 when a Lewis base catalyst (K/MgO) was used to promote ab- Ni/MgO as a combination of catalyst and support. The Ni/ sorption. This was attributed to the electron-donating abilityMgO catal yst was prepared by the impregnation method Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 253 using Ni(NO ) .6H O and MgO as catalyst precursors. The Flow meter 3 2 2 prepared catalyst was dried at 110°C in an oven and cal- cined at 500°C for 2h. For the production of CNTs, the obtained catalyst (2 g) was packed into a stainless tube (re- actor) that was placed in the central region of a vertical Stainless furnace. The reactor was flushed with N to drive air out tube while it was heated at a 10°C/min rate until the desired CH temperature (500°C) was attained. After the attainment of CO Catalyst 2 g Bubble the desired temperature, 10% H /N was introduced into 2 2 flow the furnace at a 50-mL/min rate for 1 h so as to generate active Ni particles on the MgO support. After the reduc- tion, the furnace was heated up to 850°C and CH /CO was 4 2 introduced into the reactor at a 100-mL/min rate for 1  h. During the reaction, CNTs were formed and deposited onto the catalyst surface. For the final step, N was passed for Fig. 1: The schematic diagram of the CVD instrument a period of 3  h during which the furnace was allowed to cool down to room temperature. The schematic picture of the CVD instrument is presented in Fig. 1. The as-prepared CNTs were purified by washing in 5 M HCl for 4 h at 40°C H O out to remove Ni and other catalyst components present in the CNTs. The product was filtered and washed successively with distilled water, dried at 110°C and then calcined at 500°C for 2 h in N atmosphere. CO in 1.2 Absorption experiment First, magnesium hydroxide (Mg(OH), Alfa Aesar, H O in CO 95–100.5%) was calcined in air atmosphere at 600°C for 4  h to form MgO and then impregnated with KOH solu- tion followed by drying at 110°C and calcination at 500°C for 2  h. The KMgO/CNTs catalysts with different ratios (5:0, 4:1, 3:2, 2:3, 1:4 and 0:5) were prepared by physical mixing. The CO -absorption experiment was performed using Catalyst the experimental set-up shown in Fig. 2. The apparatus was Oil bath heater Magnetic stirrer made up of a three-necked round-bottomed flask equipped with a water-cooled condenser to recover the amine, a thermometer to measure the amine-solution temperature and a gas-dispersion tube for feeding the CO gas. In CO - 2 2 absorption experiments, 100  mL of fresh aqueous amine Fig. 2: Scheme diagram of the batch reactor solution containing 2 M butylethanolamine (BEA, Sigma Aldrich, >98%) and 2 M 2-amino-2-methyl-1-propanol (AMP, 1.3 Calculation Acros Organics, 99%) was injected into a round-bottomed flask. A total of 5 g of catalyst in a basket was added into The CO loading can be calculated in the liquid phase by the aqueous amine to compare that without a catalyst. the following equation: The round-bottomed flask containing the amine solution CO loading = mole of CO / mole of amine(1) 2 2 and catalyst was immersed into the oil bath using a mag- The CO -absorption rate can be calculated in the liquid netic stirrer. The oil-bath temperature was maintained at phase by the following equation: 40  ±  1°C. The 15% CO balanced with 85% N (purchased 2 2 from Praxair, Canada) was introduced to the reactor con- CO absorption rate = mole of CO / mole of amine • min 2 2 (2) tinuously through the dispersion tube at a flow rate 200 mL/ min. During the experiment, 1  mL of amine solution was The absorption efficiency can be calculated by the fol- taken every 30 min in order to measure the CO loading by lowing equations: titration using 1 N HCl until the CO loading in the amine 2 Absorption efficiency (%) = (X − X )/X CO CO CO 2,in 2,out 2,in solution was constant. The CO loading at equilibrium rep- (3) where X is the amount of CO (in feed, off gas) resents the equilibrium solubility of CO . 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 254 | Clean Energy, 2019, Vol. 3, No. 4 before growing CNTs, after growing CNTs and after puri- 1.4 Catalyst characterization fication with acid. In the XRD pattern of the catalyst be- The surface morphology and actual elemental content of fore growing CNTs, only the XRD peaks of NiO-MgO at 2θ of the purified CNTs were examined via a FESEM equipped 37.1°, 42.9°, 62.5°, 75.2° and 78.9° were observed. The peaks with EDS and TEM using a JEOL JSM-7610F-XMaxN and JEOL of the NiO and MgO solids are clearly identified in the XRD 2100F instrument, respectively. Raman spectra of the CNTs spectrum. It should be noted that it is very difficult to de- were performed in a T64000 Raman spectrometer (Horiba termine the relative intensity of both NiO and MgO, as their JobinYvon, France) at 532-nm laser excitation. The thermo 2+ 2+ peaks overlap. This is because the Ni and Mg ions have gravimetric analysis (TG-DTG, METTLER TOLEDO STARe) 2+ similar valences, ionic radius values [r (Ni)  =  0.07  nm system operated with a flow of air at 10°C/min. The BET 2+ and r (Mg ) = 0.065 nm] and crystal cell dimensions. Thus, technique was employed to determine the specific surface the NiO and MgO components in the catalyst can easily area, pore volume and pore size of the catalysts. Prior to form a Ni Mg O solid solution [26]. For the catalyst after x 1-x each measurement, the catalysts were degassed under N growing CNTs, the XRD pattern shows that there are two flow for 2 h at 200°C using the Micrometrics FlowPrep060. intense peaks at the 2θ of 26.1° and 44.7°. Both the peaks The degassed catalysts were analysed at –196°C in the can be assigned with the graphitic phase of carbon [27 28 , ]. Micrometrics ASAP 2020 multi-point BET surface-area Moreover, the XRD peaks of metallic Ni were also observed analyser. The specific surface areas of the catalysts were at the 2θ of 44.5°, 52.1° and 76.6°, which were attributed determined by N adsorption/desorption. XRD patterns of to the reduction of NiO in the fresh catalyst. Then, the each catalyst were examined using a D8 ADVANCE X-ray catalyst was purified by washing with acid. The complete diffractometer (Bruker) equipped with Cu Kα radiation amorphous nature of the CNTs was observed. It means (k = 1.5406 A°) operated at 40 kV and 40 mA in the range of that the metals were removed completely. This result is 15°–80°. CO -TPD was performed in a ChemBET 3000 TPR/ also supported by the selected area of the catalyst surface TPD instrument. Typically, 0.05  g sample was pre-treated obtained from FESEM–EDS. in a quartz U-type reactor under a helium flow of 50  ml/ The FESEM images of the surface morphologies of min at 200°C for 1 h. The sample was then cooled down to the prepared CNTs are shown in Fig. 4. Fig. 4a shows the room temperature followed by the introduction of 3% (v/v) FESEM image of carbon nanotubes at low magnification, CO in N at a flow rate of 100 mL/min into the reactor for 2 2 while Fig. 4b shows the FESEM image of carbon nano- 4 h. The temperature was consequently gradually raised to tubes at high magnification. From these micrographs, it 900°C at a heating speed of 10°C/min under a helium flow can be observed that the CNTs are oriented randomly in of 60  mL/min. The CO desorbed was monitored using a the bending and twisting forms of tubes with few visible thermal-conductivity detector. defects. The prepared CNT surface was approximately smooth, having long and continuous tubes, which have an average diameter of 19 nm. EDS Fig. 4c was employed 2 Results and discussion to identify the concentration of catalyst embedded in the 2.1 Characterization of CNTs CNT. The EDS spectra of the synthesized CNTs indicate that the CNTs contain only carbon, indicating that the Carbon nanotubes were prepared by the thermal CVD metals were eliminated after washing with acid. method using Ni as a catalyst and CH /CO as a carbon 4 2 Thermogravimetric analysis (TGA) and the derivative source. MgO was used as a support for these catalyst thermogravimetric (DTG) curve of the weight-loss analysis nanoparticles.F ig. 3 shows the XRD patterns of catalysts are in most instances used to investigate the thermal sta- bility as well as the relative amounts of CNTs and catalyst NiO-MgO present in the product synthesized. The TG-DTG profile of the prepared CNTs before purification and after purifica- Ni tion are presented in Fig. 5. The TG-DTG curves show that Carbon the weight loss started at 480°C to 700°C, which was at- tributed to the oxidation of CNTs [29]. The temperature at (a) which the CNTs are oxidized is an index of its stability. In the case of the CNTs before purification, it is clear from (b) this curve that the sample is composed of CNTs but com- prises 40% catalyst particles at the end of the heating of the sample in air. This shows that the prepared CNTs (c) without purification contain catalyst particles that can be removed partly by washing the sample with HCl. 10 20 30 40 50 60 70 80 TEM and Raman spectra were utilized in order to verify 2θ (DEGREE) the characteristics of the CNTs. The micro images from TEM using 50 and 20  nm as the given scale (Fig. 6) indi- Fig. 3: XRD patterns of the samples (a) before growing CNTs, (b) after growing CNTs and (c) after purification cated that CNTs exhibit few compartments like a bamboo INTENSITY (A.U) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 255 AB 19 mm Spectrum 1 12 34 56 78 9 Full Scale 10371 cts Cursor: 0.000 keV Fig. 4: FESEM–EDS images of the purified CNTs structure characterized by outer and inner diameters of 9–12 and 3–6  nm, respectively. Fig. 7 shows the Raman spectra for the CNTs. It is obvious that there are two peaks –1 located at 1294 and 1546  cm , commonly known as the D −1 and G bands. The peak near 1294 cm is the D band, which was attributed to the presence of distortion carbon, while –1 the peak near 1546  cm is the G band, which is related to the graphitic layers with the high crystallinity of the 40 (a) carbon. Moreover, the relative intensity ratio of the D and G bands (I /I ) can be ascribed to the degree of graphitiza- D G tion and crystal quality of the carbon, where a lower /I I (b) D G value implies fewer defects of the carbon. The /I I ratio of D G 100 200 300 400 500 600 700 800 900 the CNTs was evaluated to be 0.71. Temperature (°C) Fig. 5: TG-DTG profile of the CNTs (a) before and (b) after purification 2.2 Characterization of the catalysts The XRD patterns of the fresh catalysts are shown in Fig. all of the catalysts are presented in Table 1. The specific 8. The sharp diffraction peaks centred at the 2θ of 36.9°, surface area and pore volume of K/MgO were found to be 2 3 42.8°, 62.3°, 74.8° and 78.6° were directly indexed to the 37.5 m /g and 0.177 cm /g, respectively. Besides, pure CNTs face-centred cubic-phase crystalline structure of MgO. The have the highest surface area and pore volumes of 122.4 2 3 broad diffraction peak at a 2θ of 26.1° and 44.7° were at- m /g and 0.177 cm /g, respectively. The binary systems of tributed to the graphitic carbon. Besides, when CNTs were KMgO/CNTs catalysts exhibited higher BET surface areas added, the diffraction peak of the CNTs appeared and the than the single KMgO catalyst, which could be attributed MgO intensity decreased with increasing CNT content. to the presence of CNTs. The total pore volume also fol- The BET surface areas, pore volume and pore size were lowed the same trend as the surface area. measured by the physical nitrogen adsorption–desorption The strength and amount of the basic surface sites using the Barett–Joyner–Halenda method. The results for on the catalysts were determined by CO -TPD based on Weight loss (%) DTG (a.u.) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 256 | Clean Energy, 2019, Vol. 3, No. 4 12 nm 50 nm 20 nm Fig. 6: TEM images of the CNTs Carbon MgO (a) I /I = 0.71 D G (b) (c) 20 (d) (e) (f) 500 1000 1500 2000 2500 3000 10 20 30 40 50 60 70 80 –1 Raman shift (cm ) 2θ (degree) Fig. 7: Raman spectra of the CNTs Fig. 8: XRD patterns of the fresh catalysts (a) KMgO, (b) KMgO/CNTs (4:1), (c) KMgO/CNTs (3:2), (d) KMgO/CNTs (2:3), (e) KMgO/CNTs (1:4) and the CO absorption–desorption process as shown in Fig. (f) CNTs 9. The basic sites or total basic sites over the catalyst are the number of exposed active centres per unit mass for the strong basic sites increased from 0 to 0.594  mmol/g, CO absorption. The temperature of desorption and the indicating that the nature of the basic sites had changed maximum desorbed CO are illustrative of the strength considerably. The numbers of strong basic sites and total (weak, medium and strong basic strength) and number basic sites are considered to be the main parameters in of basic sites, respectively. The basic strength and basic high CO absorption. Therefore, the relative number of sites of the fresh KMgO/CNTs catalysts with various ratios medium basic sites compared with strong basic sites was were compared. The results revealed that the pure KMgO used as one of the key parameters that can influence showed only one desorption peak for CO at the tempera- the catalytic activity of the catalyst for CO absorption in ture range of 250–470°C, which corresponded to the inter - amine solutions. action of CO with sites of medium basic strengths by the 2+ 2– presence of oxygen in Mg and O pairs. Besides, the pure 2.3 CO absorption CNTs showed only one desorption peak for CO at high temperatures in the range of 580–900°C (i.e. >500°C), which According to the literature, the absorption of CO using corresponded to strong basic sites. For the binary systems aqueous amine is a complex reaction that involves many of KMgO/CNTs catalysts, the results showed that two de- factors. The reaction between primary and secondary sorption peaks of medium basic sites and strong basic alkanolamines with CO can be described through two sites (small but tailing peaks at >500°C) were observed for simultaneous reactions: nearly all the samples. Obviously, with a gradual increase + − Zwitterion formation : CO + RNH ↔ RNH COO 2 2 2 in CNT content, a broad peak at the high temperature ap- (4) pears, while the peak area of medium basic sites becomes + − Carbamate and protonated amine : RNH COO + RNH 2 2 smaller and shifts towards higher temperature ranges. − + The numbers of medium and strong base sites per square ↔ RNHCOO + RNH (5) metre were calculated and the results are listed in Table The reaction between a sterically hindered amine and CO 2. With the increase in the numbers of CNTs, the medium occurs through two simultaneous reactions: basic sites decreased from 0.771 to 0.238 mmol/g, whereas Intensity (a.u) Intensity (a.u) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 257 Table 1: Summarization of the specific surface area, pore volume and pore size of the fresh catalysts 2 3 Catalysts Surface area (m/g) Pore volume (cm /g) Pore size (nm) KMgO 37.5 0.177 18.86 KMgO/CNTs (4:1) 54.1 0.218 14.56 KMgO/CNTs (3:2) 69.5 0.277 12.26 KMgO/CNTs (2:3) 85.4 0.368 8.94 KMgO/CNTs (1:4) 107.4 0.479 5.90 CNTs 122.4 0.555 3.66 + − (6) Bicarbonate formation : CO + RNH ↔ RNH COO 2 2 2 + − + − (f) (7) RNH COO + H O ↔ RNH + HCO 2 2 3 3 (e) Bicarbonate formation by carbamate hydrolysis : CO + RNH 2 2 (8) (d) ↔ RNH COO + + − − (c) (9) RNH COO + NH ↔ RNHCOO + RNH 2 2 3 + + − (b) RNHCOO + H O + RNH ↔ RNH + RNH + HCO 2 3 2 3 3 (10) (a) The formation of bicarbonate occurs in the sterically hin- 100 200 300 400 500 600 700 800 900 dered amine due to the steric character caused by the Temperature (°C) bulky group adjacent to the amino group in the amine. Fig. 9: CO -TPD profile of the fresh catalysts (a) KMgO, (b) KMgO/CNTs Therefore, the carbamate can easily undergo hydrolysis, 2 (4:1), (c) KMgO/CNTs (3:2), (d) KMgO/CNTs (2:3), (e) KMgO/CNTs (1:4) and leading to the formation of bicarbonate and free amine (f) CNTs molecules [8 ]. For the catalyst-screening process for CO absorption, resulting in an enhanced CO loading capacity for CO ab- 2 2 15% CO was employed and the reaction temperature was sorption. This effect is combined with the increase in strong set at 40°C. Fig. 10a shows the profiles of CO loading in basic sites and total basic sites with the increase in the ratio BEA–AMP aqueous solutions with catalysts and without of CNTs in KMgO/CNT mixture .The overall implication is any catalyst. The overall period of absorption to reach CO - the generation of a large number of exposed active centres loading equilibrium takes 420  min in the case without a absorption. The catalyst provides more per unit mass for CO catalyst. With the addition of K/MgO, the overall time was electrons to be available to the amine, which facilitates CO reduced from 420 to 330 min. From the results, it is clear absorption, leading to an increase in the rate of carbamate that, at the same absorption time, the addition of the cata- and bicarbonate formation. A  detailed mechanism is ex- lyst into the amine solution led to increased CO loading plained in Scheme 1. Subsequently, a pair of electrons from and this reached equilibrium more quickly than without the amine is transferred to the catalyst−CO -formed species a catalyst. The addition of K/MgO catalysts to the absorp- (intermediate species) to restore the electrons of the catalyst tion process increased the CO-absorption capacity, as CO as well as to generate the zwitterion formation [18]. 2 2 molecules comprise a Lewis acid that accepts electrons Fig. 11 shows the CO -absorption rate in BEA–AMP from Lewis bases; therefore the presence of basic sites on aqueous solutions with catalysts and with no cata- the surface of metal oxide enhances the CO -adsorption lyst. The CO-absorption rate was calculated as the CO - 2 2 capacity. loading change per minute until it reached equilibrium The CO loading over the KMgO/CNTs catalysts with CO loading. From the results, it is clear that the introduc- various ratios was also determined. As can be seen from tion of the catalysts into the absorption system helped Fig. 10b, an increment in the CNT number increased the CO to increase the CO-absorption rate. The CO-absorption 2 2 loading at the same absorption time and reached equilib- rate increased from 0.0014  mol CO/mol amine/min in rium more quickly as compared with KMgO and without a the case without a catalyst to 0.0018 mol CO /mol amine/ catalyst. The KMgO/CNTs at the ratio of 1:4 had the highest min with the addition of KMgO catalyst. With the add- CO loading at the same absorption time and reached equi- ition of CNTs to K/MgO, the CO-absorption rate in- librium at 240 min. This can be explained on the basis of the creased from 0.0018 mol CO/mol amine/min in the case structure of CNTs, which consists of graphene layers with of the K/MgO catalyst to 0.0025  mol CO /mol amine/ a large BET specific surface area and large active area that min. The CO -absorption rate ranked as follows: KMgO/ help in increasing the exposure of the amine molecules, CNTs (1:4)  >  CNTs  >  KMgO/CNTs (2:3)  >  KMgO/CNTs Signall (a.u) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 258 | Clean Energy, 2019, Vol. 3, No. 4 OH NH H C H OH AMP O N C – – – e e e N H O N H BEA – – e e HO HO HO Carbamate Ammonium ion HO N H BEA O C HH N HO – – – C C N H N e e e AMP O – – e e HO HO HO Carbamate Ammonium ion Scheme 1: A possible catalytic CO -absorption pathway over the catalyst. Table 2: Basic sites of the catalysts Catalysts Medium sites (mmol 2/g) Strong sites (mmol /g) Total basic sites (mmol /g) CO CO2 CO2 KMgO 0.771 0 0.771 KMgO/CNTs (4:1) 0.451 0.117 0.568 KMgO/CNTs (3:2) 0.252 0.204 0.456 KMgO/CNTs (2:3) 0.251 0.303 0.554 KMgO/CNTs (1:4) 0.238 0.594 0.849 CNTs 0 0.657 0.657 (3:2) > KMgO/CNTs (4:1) > KMgO > no catalyst. The KMgO/ physical properties (surface area, pore volume and pore CNTs 1:4 catalyst shows the highest CO -absorption rate size) and chemical properties (total basic sites, strong (or increase of 44%) when compared with the system with basic sites and medium basic sites) of the solid catalysts. no catalyst. Fig. 12 shows the results for CO -absorption ef- The relationship between the absorption rate and the ficiency in BEA–AMP aqueous solutions with catalysts and physical and chemical properties of the solid catalysts are without any catalyst. From the results, the CO -absorption shown in Equations (11) and (12): efficiency with the addition of catalysts increased when Absorption rate = 0.003255 +(−0.000000059) ∗ A compared with no catalyst. The CO efficiency increased +(−0.00108) ∗ B +(−0.000067) ∗ C from 24% in the case without a catalyst to 27% with the (11) addition of K/MgO catalyst. With the addition of CNTs to where parameters A, B and C are surface area, pore volume K/MgO, the CO -absorption efficiency increased tremen- 2 and pore size, respectively. dously. The CO -absorption efficiency ranked as follows: Absorption rate = 0.002356 + 0.023692 ∗ D KMgO/CNTs (1:4) > CNTs > KMgO/CNTs (2:3) > KMgO/CNTs +(−0.02379) ∗ E +(−0.0244) ∗ F (3:2) > KMgO/CNTs (4:1) > KMgO > no catalyst. The KMgO/ (12) CNTs (1:4) catalyst show the highest CO -absorption ef- where parameters D, E and F are total basic sites, strong ficiency (or increase of 15.3%) when compared with no basic sites and medium basic sites, respectively. catalyst. The relationship between the CO-absorption efficiency In order to understand which properties of the solid and the physical and chemical properties of the solid cata- catalysts played the most significant role in promoting lysts are shown in Equations (13) and (14): the CO -absorption rate and absorption efficiency, a re- Absorption ef f iciency = 44.69023 +(−0.15283) ∗ A gression analysis was conducted and used to determine +(25.31378) ∗ B +(−0.891) ∗ C any possible correlation between absorption rate and ab- (13) sorption efficiency with properties of the catalysts. The Absorption ef f iciency = 31.13859 + 407.38 ∗ D CO -absorption rate and absorption efficiency obtained +(−402.877) ∗ E +(−413.286) ∗ F from the experiment were regressed against both the (14) e Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 259 A parity plot for the absorption rate and absorption effi- absorption rate and chemical properties (Fig.13a) of the solid ciency with physical and chemical properties is shown in catalysts showed a smaller value for AAD (1.05%) than the Fig. 13. As seen in Fig. 13, the experimental values versus correlations between the absorption rate and physical prop- those predicted using the corresponding correlations for erties (AAD = 2.59%). The correlations between the absorp- tion efficiency and chemical properties (Fig.13b) of the solid catalysts also showed a smaller value for AAD (0.48%) than (a) 0.6 the correlations between the absorption efficiency and phys- ical properties (AAD = 1.76%). This implied that the chemical 0.5 properties (number of basic sites and basic strength) played No catalyst a bigger role in the CO-absorption rate and absorption effi- 0.4 2 KMgO ciency by a solid basic catalyst than the physical properties. KMgO/CNTs(4:1) The degree of correlation obtained for the relationship be- 0.3 KMgO/CNTs(3:2) tween the CO -absorption rate and chemical properties of the KMgO/CNTs(2:3) 0.2 solid catalysts was 99%, which was higher than that obtained KMgO/CNTs(1:4) between the absorption rate and physical properties of the CNTs 0.1 solid catalysts (81%). Also, the CO -absorption efficiency cor - relations with chemical properties had a degree of correlation 2 2 30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 (R ) of 99%, whereas, with physical properties, R was 74%. The Time (min) catalyst appears to be working more on AMP, making it more effective than BEA. Based on its structure, AMP has a sterical (b) 0.6 hindered alkyl group, which tends to form unstable carba- mates whose hydrolysis leads to the formation of bicarbon- ates. The bicarbonates easily release CO to form free amine 0.55 No catalyst molecules for further reaction with incoming CO . KMgO The XRD patterns of the catalysts after 6 h of use for CO KMgO/CNTs(4:1) absorption are shown in Fig. 14. For all of the catalysts in use 0.5 KMgO/CNTs(3:2) with the presence of KMgO in the mixture, diffraction peaks KMgO/CNTs(2:3) at the 2θ° angle of 18.5°, 38.1°, 51.1°, 58.8°, 62.2°, 68.7° and KMgO/CNTs(1:4) 0.45 72.2° were observed, which were assigned to the crystal- CNTs line structure of Mg(OH). The presence of Mg(OH) peaks 2 2 is related to the reaction of MgO with water in the amine 0.4 180 210 240 270 300 330 360 390 420 450 480 solution during absorption [30]. Furthermore, a higher inten- Time (min) sity of MgCO patterns was also noticed at the 2 of 44.4° θ and 51.8°, when the CNTs were introduced to the KMgO. By Fig. 10: (a) CO -absorption profile of the 2 M BEA + 2 M AMP at 40°C with increasing the CNTs content, the diffraction peaks of MgCO and without catalysts based on (a) the entire run (30–450 min) and (b) a portion of the run (180–450  min) to show explicitly when each run intensity increased while diffraction peaks of Mg(OH) inten- reaches equilibrium. sity decreased. This was attributed to the strong basic sites 0.003 0.0025 0.002 0.0015 0.001 0.0005 no catalyst KMgO KMgO/CNTs KMgO/CNTs KMgO/CNTs KMgO/CNTs CNTs (4:1) (3:2) (2:3) (1:4) Catalyst Fig. 11: CO absorption rate of 2 M BEA + 2 M AMP solution with and without catalysts. Reaction conditions: 5.0 g of catalyst, BEA–AMP 100 mL, CO 2 2 flow rate at 200 mL/min and temperature 40°C CO Loading (mol CO /mol amine) CO Loading (mol CO /mol amine) 2 2 2 2 CO absorption rate (mol CO /mol amine-min) 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 260 | Clean Energy, 2019, Vol. 3, No. 4 no catalyst KMgO KMgO/CNTs KMgO/CNTs KMgO/CNTs KMgO/CNTs CNTs (4:1) (3:2) (2:3) (1:4) Catalyst Fig. 12: CO -absorption efficiency of 2 M BEA + 2 M AMP solution with and without catalysts. Reaction conditions: 5.0 g of catalyst, BEA–AMP 100 mL, CO flow rate at 200 mL/min and temperature 40°C. Carbon Mg(OH) MgCO MgO 0.003 2 3 (a) 0.0025 (f) 0.002 (e) 0.0015 (d) Physical AAD (%) = 2.59 (c) 0.001 R = 81% (b) Chemical AAD (%) = 1.05 0.0005 R = 99% (a) 10 20 30 40 50 60 70 80 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 Experimental value 2θ (degree) (b) Fig. 14: Representative XRD patterns of the spent catalysts (a) CNTs, (b) KMgO, (c) KMgO/CNTs (4:1), (d) KMgO/CNTs (3:2), (e) KMgO/CNTs (2:3) and (f) KMgO/CNTs (1:4) for commercializing any catalytic process by reducing Physical AAD (%) = 1.76 the cost of production and for real application on a large R = 74% or industrial scale. In this research, the reusability of the Chemical AAD (%) = 0.48 10 2 catalyst was studied by calcination and without calcin- R = 99% ation (dried in the oven at 110°C overnight) as shown in Fig. 15. Calcination of the used catalyst was performed in 0 510 15 20 25 30 35 40 45 air at 300°C for 2 h to burn away any deposition of organic Experimental value contaminants. The results revealed that the CO -loading rate of the catalyst with calcination was not significantly Fig. 13: Parity chart of the predicted values versus experimental values of (a) the absorption rate and (b) the absorption efficiency changed and reached equilibrium at the same time as the fresh catalyst. On the other hand, the CO -loading rate of the catalyst without calcination decreased and reached of the CNTs aiding a strong interaction between MgO and equilibrium more slowly (360 min) when compared to the CO to form MgCO. This reaction hindered the formation of 2 3 fresh catalyst and the catalyst with calcination. This was Mg(OH) . Based on the literature, both Mg(OH) and MgCO 2 2 3 attributed to the surface active sites of the catalyst being are equally active phases for CO absorption in amines [31]. covered by organic contaminants, which could create a layer that covers the surface of MgO, Mg(OH) or MgCO 2 3 against CO . 2.4 Catalyst reusability and stability After the first run, the KMgO/CNTs (1:4) catalyst was An important characteristic feature of a catalyst is its collected from the reaction, dried in an oven at 110°C and reuse property. This reusability of the catalyst is a factor calcined in the atmosphere at 350°C for 2  h to obtain a Predicted value Predicted value CO absorption efficiency (%) Intensity (a.u) Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 Natewong et al. | 261 0.7 CNTs yielded very-high-performance catalysts for CO absorption with BEA–AMP solvent. 0.6 (3) The increase in the number of CNTs to KMgO reduces 0.5 the overall absorption periods to reach equilibrium as compared with KMgO and without a catalyst, mainly 0.4 due to the increasing numbers of strong basic sites. The KMgO/CNTs(1:4) 0.3 Reuse with calcined increased number of CNTs to KMgO also increased the Reuse without calcined number of exposed active centres per unit mass for CO 0.2 absorption. Furthermore, it also generated a large spe- 0.1 cific surface area and pore volume, which are beneficial in increasing the exposure of amine molecules. (4) The results obtained from this study are intended to 30 60 90 120 150 180 210 240 270300 330 360 Time (min) show clearly the relative performance of the basic cat- alyst. The best catalysts selected from this study will Fig. 15: Reusability of the catalyst be subjected to pilot plant studies. The addition of this catalyst into a commercially packed absorber is essen- tial towards the industrial application of CO capture. 0.003 2 0.0025 Acknowledgements 0.002 The financial support provided by Queen Elizabeth II Diamond Jubilee Scholarship (QES) and the Natural Science and Engineering 0.0015 Research Council of Canada is gratefully acknowledged. 0.001 References [1] Wang  Y, Zhao  L, Otto  A, et  al. A review of post-combustion 0.0005 CO capture technologies from coal-fired power plants. Energy Procedia 2017; 114:650–65. [2] Cuéllar-Franca  RM, Azapagic  A. Carbon capture, storage and 12 34 5 utilisation technologies: a critical analysis and comparison of Number of cycle (Times) their life cycle environmental impacts. J CO Util 2015; 9:82–102. Fig. 16: Recycling tests on the catalytic activity of KMgO/CNTs (1:4) for [3] Song C, Liu Q, Ji N, et al. Alternative pathways for efficient CO CO absorption. Reaction conditions: 5.0 g of catalyst, BEA–AMP 100 mL, 2 capture by hybrid processes—a review. Renew Sust Energ Rev CO flow rate at 200 mL/min and temperature 40°C 2 2018; 82:215–31. [4] Dutcher B, Fan M, Russell AG. Amine-based CO capture tech- nology development from the beginning of 2013—a review. regenerated catalyst. The activity property of the regener - ACS Appl Mater Interfaces 2015;7:2137–48. ated catalyst was tested under the same operational pro- [5] Wu X, Yu Y, Qin Z, et al. The advances of post-combustion CO cess as the fresh catalyst. The results for repeated usage capture with chemical solvents: review and guidelines. Energy (five times) are shown in Fig. 16. It was found that the CO- Procedia 2014; 63:1339–46. absorption rate only changed slightly after being repeat- [6] Lee CS, Ong YL, Aroua MK, et al. Impregnation of palm shell- based activated carbon with sterically hindered amines for edly used five times. CO adsorption. Chem Eng J 2013; 219:558–64. [7] Luis P. Use of monoethanolamine (MEA) for CO capture in a global scenario: consequences and alternatives. Desalination 3 Conclusions 2016; 380:93–9. [8] Bougie F, Iliuta MC. Sterically hindered amine-based absorb- (1) Carbon nanotubes have been synthesized successfully ents for the removal of CO from gas streams. J Chem Eng Data by the decomposition of CH/CO over Ni/MgO catalyst. 4 2 2012; 57:635–69. Raman spectral analyses as well as TEM images indi- [9] Zheng C, Tan J, Wang YJ, et al. CO solubility in a mixture ab- cate that CNTs were produced. The CNTs exhibit few sorption system of 2-amino-2-methyl-1-propanol with glycol. compartments like a bamboo structure characterized Ind Eng Chem Res 2012; 51:11236–44. [10] Nwaoha C, Saiwan C, Tontiwachwuthikul P, et al. Carbon di- by outer and inner diameters of 9–12 and 3–6  nm, oxide (CO ) capture absorption desorption capabilities of respectively. 2 2-amino-2-methyl-1-propanol (AMP) piperazine (PZ) and (2) The CO -absorption rate and absorption efficiency with monoethanolamine (MEA) tri-solvent blends. J Nat Gas Sci Eng catalysts and without catalyst were studied. With the 2016; 33:742–50. addition of a catalyst into the amine solution, the CO - [11] Idem R, Wilson M, Tontiwachwuthikul P, et al. Pilot plant studies absorption rate and absorption efficiency were higher of the CO capture performance of aqueous MEA and mixed than without a catalyst. The combination of KMgO and MEA/MDEA solvents at the University of Regina CO capture CO absorption rate CO Loading (mol CO /mol amine) (mol CO /mol amine-min) 2 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/251/5531680 by DeepDyve user on 10 December 2019 262 | Clean Energy, 2019, Vol. 3, No. 4 technology development plant and the boundary dam CO cap- [22] Magrez A, Seo JW, Smajda R, et al. Catalytic CVD synthesis of ture demonstration plant. Ind Eng Chem Res 2006; 45:2414–20. carbon nanotubes: towards high yield and low temperature [12] Sakwattanapong R, Aroonwilas A, Veawab A. Behavior of reboiler growth. Materials (Basel) 2010; 3:4871–91. heat duty for CO capture plants using regenerable single and [23] Alvarez WE, Pompeo F, Herrera JE, et al. Characterization of blended Alkanolamines.Ind Eng Chem Res 2005; 44:4465–73. single-walled carbon nanotubes (SWNTs) produced by CO [13] Narku-Tetteh J, Muchan P, Saiwan C, et al. Selection of com- disproportionation on Co-Mo catalysts. Chem Mater 2002; ponents for formulation of amine blends for post combustion 14:1853–8. CO capture based on the side chain structure of primary, sec- [24] Kong J, Cassell AM, Dai H. Chemical vapor deposition of me- ondary and tertiary amines. Chem Eng Sci 2017; 170:542–60. thane for single-walled carbon nanotubes. Chem Phys Lett [14] Idem R, Shi H, Gelowitz D, et al. Catalytic method and appar - 1998; 292:567–74. atus for separating a gaseous component from an incoming [25] Corthals S, Noyen JV, Geboers J, et al. The beneficial effect of gas stream. 2017, US Patent 9,586,175. CO in the low temperature synthesis of high quality carbon [15] Shi  H, Naami  A, Idem  R, et  al. Catalytic and non catalytic nanofibers and thin multiwalled carbon nanotubes from CH solvent regeneration during absorption-based CO capture over Ni catalysts. Carbon 2012; 50:372–84. with single and blended reactive amine solvents. Int J Greenh [26] Jafarbegloo  M, Tarlani  A, Wahid  Mesbah  A, et  al. NiO–MgO Gas Con 2014; 26:39–50. solid solution prepared by sol–gel method as precursor for [16] Hamid  RR, Abdelhamid  S. A highly efficient CaO-based CO Ni/MgO methane dry reforming catalyst: effect of calcin- sorbent prepared by a citrate-assisted sol–gel technique. Chem ation temperature on catalytic performance. Catal. Lett 2016; Eng J 2015; 262:913–20. 146:238–48. [17] Li  L, Wen  X, Fu  X, et  al. MgO/AlO sorbent for CO capture. [27] Prasongthum N, Xiao R, Zhang H, et al. Highly active and stable 2 3 2 Energy Fuels 2010; 24:5773–80. Ni supported on CNTs-SO fiber catalysts for steam reforming [18] Narku-Tetteh J, Afari DB, Coker J, et al. Evaluation of the roles of ethanol. Fuel Process Technol 2017; 160:185–95. of absorber and desorber catalysts in the heat duty and heat of [28] Wang S, Yin Q, Guo J, et al. Influence of Ni promotion on liquid CO desorption from butylethanolamine−2 amino-2-methyl- hydrocarbon fuel production over Co/CNT catalysts. Energy 1-propanol and monoethanolamine−methyldiethanolamine Fuels 2013; 27:3961–8. solvent blends in a bench-scale CO capture pilot plant. Energy [29] Ezekiel ZD, Ntaote DS, Thema FT, et al. Raman and TGA study Fuels 2018; 32(9):9711–26. of carbon nanotubes synthesized over Mo/Fe catalyst on alu- [19] Khalil SH, Aroua MK, Daud WMAW. Study on the improvement minium oxide, calcium carbonate and magnesium oxide sup- of the capacity of amine-impregnated commercial activated port. Chem Sci Trans 2013; 2:1160–73. carbon beds for CO adsorbing. Chem Eng J 2012; 183:15–20. [30] Hanlon JM, Bravo Diaz L, Balducci G, et al. Rapid surfactant-free [20] Lu  C, Bai  H, Wu  B, et  al. Comparative study of CO capture synthesis of Mg(OH) nanoplates and pseudomorphic dehydra- 2 2 by carbon nanotubes, activated carbons, and zeolites. Ener gy tion to MgO. Cryst Eng Comm 2015; 17:5672–9. Fuels 2008; 22:3050–6. [31] Shi  H, Zhou  Y, Zuo  Y, et  al. Heterogeneous catalysis of CO - [21] Creamer  AE, Gao  B, Zhang  M. Carbon dioxide capture using diethanolamine absorption with MgCO and CaCO and com- 3 3 biochar produced from sugarcane bagasse and hickory wood. paring to non-catalytic CO -monoethanolamine interactions. Chem Eng J 2014; 249:174–9. React Kinet Mech Cat 2017; 122:539–55.

Journal

Clean EnergyOxford University Press

Published: Dec 6, 2019

References