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

Learn More →

Mass-transfer studies of solid-base catalyst-aided CO2 absorption and solid-acid catalyst-aided CO2 desorption for CO2 capture in a pilot plant using aqueous solutions of MEA and blends of MEA-MDEA and BEA-AMP

Mass-transfer studies of solid-base catalyst-aided CO2 absorption and solid-acid catalyst-aided... Mass-transfer studies of catalyst-aided CO absorption and desorption were performed in a full-cycle, bench- scale pilot plant to improve CO absorption using 5M MEA, 5M MEA-2M MDEA and 2M BEA-2M AMP. A solid-base catalyst, K/MgO, and an acid catalyst, HZSM-5, were used to facilitate absorption and desorption, respectively. Absorption and desorption mass-transfer performance was presented in terms of the overall mass-transfer coefficient of the gas side (K a ) and liquid side (K a ), respectively. For non-catalytic runs, the highest K a and G v L v G V Kmol 1 K a were 0.086 and 0.785 for 2M BEA-2M AMP solvent. The results showed 38.7% K a and 23.6% K a L V hr G v L v m .kPa.hr a and 45% K a increase for increase for 2M BEA-2M AMP with only HZSM-5 catalyst in desorber and a 95% K G V L V both K/MgO catalyst and HZSM-5 catalyst. This was attributed to the role of K/MgO in bonding loosely with CO and making it available for the amine reaction. Keywords: HZSM-5; K/MgO catalyst; CO removal efficiency; CO capture; mass-transfer coefficients of absorber 2 2 (K a ) and desorber (K a ) G V L V include monoethanolamine (MEA), diethanolamine (DEA), Introduction methyldiethanolamine (MDEA) and 2-amino-2-methyl-1- The process of post-combustion capture of CO with an propanol (AMP). Despite the high reactivity of MEA with amine solvent has been well accepted as the most re- , its use has been hampered by its high regeneration CO searched and cost-effective technology in the field of -absorption capacity heat. In contrast, MDEA has high CO capture [1, 2]. In this technology, the chemical reac- CO and low regeneration heat, but has slow reactivity with gas re- tion between an aqueous amine solvent and CO [3]. The process of improving the post-combustion CO sults in an enhanced mass transfer within the solvent. -capture process is divided into solvent formulation CO -absorption processes Common amines in commercial CO and process optimization [4]. Received: 29 March, 2019; Accepted: 13 May, 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, 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/3/4/263/5550982 by DeepDyve user on 10 December 2019 264 | Clean Energy, 2019, Vol. 3, No. 4 Under solvent formulation, studies on the blending of as the rate-determining step [, 89]), followed by the re- solvents are intended to obtain high-performance solv- action between the abstracted proton and carbamate to ents with higher CO cyclic capacity, lower solvent regen- form CO and free amine. Technologies employed to in- 2 2 eration energy requirement and faster absorption kinetics crease the rate of CO-amine absorption and desorption with favourable environmental characteristics. Recently, include the addition of either a homogeneous or heter - Narku-Tetteh et  al. [4] used study of the amine chemical ogenous component to the solvent as a catalyst to in- structure–activity relationships to develop the criteria crease the chemical reactivity between CO and amine. for selecting components to formulate high-performing Liu et al. [10] investigated CO absorption into MDEA and amine blends that resulted in the development of an effi- dimethylaminoethanol (DMEA) with the application of cient bi-solvent blend consisting of 2M butyl monoethanol carbonic anhydrase as a homogeneous catalyst. The re- amine (BEA)-2M AMP. Results from the work of Idem et al. sults showed an enhanced absorption of CO with a low [5] showed a lower heat duty for a blend consisting of 7 M concentration of carbonic anhydrase in the amine, which MEA-MDEA, in the ratio of 5:2, when compared with con- was attributed to its role in promoting the hydration of ventional 5M MEA. CO , the rate-determining step. Similar results have been The reaction steps for both amine CO absorption reported in the literature [11]. and desorption are shown in Figure 1. For the amine Idem et  al. [12] identified several catalysts for appli- CO -absorption process, two reactions occur according cation in the field of post-combustion capture. Recent to the zwitterion mechanism. These include CO reac- works [13, 14] have shown improvements in the regen- tion with the amine to produce the zwitterion molecule eration process with the introduction of a solid-acid in an electron-transfer process. The zwitterion molecule catalyst in the desorption unit for different solvent sys- then reacts with any base in a hydrogen ion (H proton) tems. Two major catalysts were studied, which included abstraction process to form a carbamate molecule. The HZSM-5, a Bronsted-acid catalyst; and gamma alumina mechanism is only different for tertiary amines where (ɣ-Al O ), a Lewis-acid catalyst. Results identified HZSM-5 2 3 the CO reacts with the amine and water in a single-step as the most suitable catalyst, which was attributed to process to form bicarbonate ions and water. The two-step the donation of its H proton for carbamate breakdown. zwitterion-reaction mechanism has been identified as a Zhang et  al. [15] also investigated the application of a base catalysed reaction []. 6 The electron-transfer process solid-acid catalyst including HZSM-5,-Al ɣ O , SAPO-34 2 3 2– of CO reacting with the amine has also been identified as and SO /TiO to increase the rate of CO regeneration 2 4 2 2 the rate-determining step []. 7 This makes the base cata- for loaded aqueous MEA, as well as a tri-solvent blend of lyst that is capable of donating a lone-pair electrons (i.e. MEA-AMP-PZ. Results also showed the enhanced amine a Lewis-base catalyst) the most suitable for increasing desorption rates and lower heat duty with the cata- the reactivity between the CO and amine. The Lewis- lyst, attributed to its high Bronsted/Lewis ratio and its base catalyst first donates a lone pair of electrons from mesoporous surface area. This promotes proton transfer its anion to the CO molecule by bonding loosely with to the carbamate (rate-determining step). The intro- CO and allowing more contact time for the reaction with duction of the desorption catalyst showed a significant amine. This makes the CO more susceptible to the amine. improvement in the heat duty; however, its indirect The desorption reaction also involves two main re- improvement on the CO-removal efficiency was only actions. The abstraction of a hydrogen (H) proton marginal. (deprotonation) from a protonated amine (identified This work intends to improve CO-removal efficiency by introducing a base catalyst and an acidic catalyst to enhance reaction kinetics in the absorption and desorp- AB tion columns, respectively. The overall mass-transfer coefficient represents the constant of proportionality be- Amine +CO +H O 2 2 Amine +CO tween the mass-transfer rate or flux and its driving force. A higher mass-transfer rate also results in a faster reaction rate. Faster mass transfer allows the operator to achieve the same CO -capture rate (or removal efficiency) with + – Amine COO +Amine a shorter column. Essentially, the heights of the absorp- (zwitterion formation) tion and desorption columns are inversely proportional to the overall mass-transfer coefficient, meaning that a high mass-transfer coefficient will result in a shorter column. + – Consequently, this will lead to smaller sizes of equipment, AmineH + HCO – + AmineCOO + AmineH and thus low capital costs. The results obtained from these (carbamate formation) mass-transfer studies for different catalyst configurations using 5M MEA, 5M MEA-2M MDEA and 2M BEA-2M AMP Fig. 1: Reaction pathways for amines. (a) Primary and secondary amines, (b) sterically hindered amines. solvents are presented in this paper. Absorption Desorption Absorption Desorption Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 265 average deviations (AADs) of 1.5 and 1.3%, respectively, 1 Experimental section which confirms the validity of the N O experiment (with 1.1 Materials and equipment repeatability less than 1%). The chemicals used and where they were obtained are as follows. Reagent-grade butylethanolamine (≥98%), 1.3 Amine structure 2-amino-2-methyl-1-propanol (≥99%), monoethanolamine (≥99%) and methyl diethanolamine (≥99%) were purchased The amines used for this study were monoethanolamine from Sigma-Aldrich Co., Canada. The 1 N hydrochloric acids (MEA), methyldiethanolamine (MDEA), butylethanolamine (HCl), used for titration to determine the concentration of (BEA) and 2-amino-2-methyl-1-propanol (AMP). Their the amines, was also purchased from Sigma-Aldrich. Pure chemical structures are given in Figure 2. CO and N gases were purchased from Praxair Inc., Regina. 2 2 The CO gas analyser was purchased from NOVA analytical system Inc. Gas-analyser calibrations were done using pure 1.4 K/MgO catalyst preparation N gas and 15% CO (N balance) also acquired from Praxair 2 2 2 The catalyst was prepared by the incipient wetness Inc., Regina. The stainless-steel LDX Sulzer structured method. Commercially available Mg(OH) purchased from packing with an outside diameter of 0.047 m was provided Sigma-Aldrich was used in the preparation of the sup- by Sulzer Chemtech Ltd. Temperatures during operation port. The Mg(OH) was first calcined at 600°C for 2 hours were recorded using J-type thermocouples from Cole to produce MgO, following which the catalyst was allowed Parmer, Canada. Commercially available Mg(OH) (≥98%) to cool for 1 hour. A  solution of KOH was prepared using and KOH (≥98%) were purchased from Sigma-Aldrich. the catalyst pore volume. The solution concentration was such that the final composition of K, the active metal, was 1 mol% on the support. The prepared slurry was dried for 1.2 Physical solubility of CO (N O analogy) 2 2 12 hours and later calcined at 600°C for 2 hours. The final The experimental setup used for determining N O solu- catalyst was pelletized to obtain particle sizes in the range bility was similar to one used by Sema [16]. The solu- of 4.2–4.7 mm. bility experiment was performed in a rotary-type 600-ml stainless-steel autoclave reactor (model Parr 5500, Parr Instrument Co., Moline, IL, USA) connected to a con- 1.5 Bench-scale pilot-plant experimental setup troller (model Parr 4843, Parr Instrument Co., Moline, IL, The setup describing the experimental system is shown USA). The reactor consisted of a variable-speed impeller, in Figure 3. The system comprises two lagged stainless- heating mantle, cooling coil, gas-feed port, thermo- steel columns, each measuring 3.5 ft (1.067 m) in height couple and pressure transducer. and having an internal diameter of 2 inches (0.0508 m). The Initially, the amine solution was degassed using an internal arrangements of both the absorber and desorber ultrasonic bath (VWR model 75D, VWR International, ON, columns with catalysts are shown in Figure 4. The de- Canada). Then, 300 ml of the degassed amine solution was sorber was filled from the bottom with a section of Sulzer introduced into the reactor. The desired temperature was LDX structured packing measuring 0.0508 m in diameter set and controlled by the controller, and then the vacuum and 0.18 m in height. On top of this was placed a 0.025-m pump was turned on. After shutting down the vacuum section of large inert marbles measuring 6 mm in diameter pump, the liquid was under a certain pressure due to li- to act as a support for the catalytic bed. A section of catalyst quid vaporization. The pressure, P (kPa), was measured at of 0.55 m in height was mixed with small, 3-mm-diameter vacuum. A certain amount of NO (n .; kmol) was fed to N O 2 2 the reactor. The amount of N O dissolved in the solvent was determined by measuring the pressure in the reactor be- CH fore N O injection (P ; kPa) and after NO injection (P ; kPa) 2 1 2 2 OH N (±6.25 kPa). These terms were correlated as in Equation (1): HO H N n =(P − P ) (1) MEA N O 2 1 Z RT MDEA N O OH where Vg is the volume of the gas container (m ), R is the gas constant (kJ/kmol.K) and Z is the compressibility N2O factor of NO. Equilibrium is reached when the NO pres- H C 2 2 OH sure does not change after 12 hours. HO CH The equipment and calculation procedure for the N O CH 2 3 NH AMP solubility were validated with a 5M MEA solution in the temperature range of 298  ±  1 to 343  ±  1  K. The experi- BEA mental results were in good agreement with the work by Ma’mun et  al. [17] and Tsai et  al. [18] with absolute Fig. 2: Chemical structure of amines used Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 266 | Clean Energy, 2019, Vol. 3, No. 4 CO Product Offgas Inlet hot water Lean Amine Hot water heat exchanger Outlet hot water Rich Amine Flue gas Lean Amine Rich Amine Lean amine cooler Rich-Lean Heat exchanger Amine tank Fig. 3: Schematic representation of the experimental setup for CO removal Adapted with permission from [19]. Copyright 2017 W. Srisang. drop, flooding and liquid carryover within the absorption 0.051 m column. Structured 0.18 m LDX packing Sulzer 0.025 m 0.0891 m 1.6 Bench-scale pilot-plant experimental large inert (catalyst + marbles procedure empty At the start of the experiment, the amine was fed from space) 0.55 m catalytic bed the storage tank to the top of the absorber column via a 0.96 m (catalyst + variable-speed gear pump and left at the bottom of the LDX small inert marbles) Sulzer column. The amine was then pumped through a lean rich heat exchanger to exchange heat with the hot lean amine from the desorption column. The amine was further heated 0.025 m large inert by an external heater to the desired amine temperature for marbles desorption. The heating medium for this heater was gly- 0.18 m LDX cerol to heat up the rich amine prior to entering the de- Sulzer sorber. The amine in the desorption column exited as lean amine. This lean amine was pumped through the rich-lean Absorber Absorber Desorber heat exchanger and then through a cooler to further reduce with catalyst with catalyst its temperature prior to entering the absorption column. Once amine-solvent circulation was achieved, a mixture of Fig. 4: Column packing and catalyst-bed arrangement for absorber and and N gas at the appropriate CO concentration of 15% CO desorber 2 2 2 was introduced to the bottom of the absorber column and marbles to form the catalytic bed. On top of this was placed monitored by a gas-flow meter. The lean amine contacted a second 0.025-m section of large inert marbles, then an- the gas feed from the bottom of the column counter- other 0.18-m section of Sulzer LDX structured packing. currently. Treated gas left the top of the column while the The absorber column was arranged in an alternating bed rich amine solvent left the absorber bottom. The duration of Sulzer LDX structured packing and K/MgO (absorption of an experimental run was between 4 and 6 hours. catalyst). The absorber was filled from the bottom with a A steady-state operation of each experiment was indi- section of Sulzer LDX structured packing measuring 0.051 cated when there was no temperature change with time m in diameter and 0.051 m in height. On top of this was of different temperature points within the absorption placed the required catalyst weight in a section measuring column, indicating that the solution entering the absorber 0.089 m.  The catalyst filled that section, partly leaving a was at a constant CO concentration, amine concentration void space. This absorber arrangement was repeated in an and flow rate. At this point, the CO concentration in the alternating pattern. This was done to minimize pressure gas phase and temperature profile were measured along Absorber Stripper Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 267 the height of the column (error less than 1%) using the IR 2.1.1 Van’t Hoff equation fitting for He N O-amine gas analyser and thermocouples, respectively. Also, sam- Å ã Å ã ΔS −ΔH ples were taken from the bottom of both columns to de- He = exp (8) N O−BEA/AMP R R T g g termine the rich and lean CO loadings using a Chittick apparatus (error less than 1%). Tabl e 1 shows the operating where ΔS  =  the entropy change as the NO is physically Ä ä conditions used in the pilot-plant experiments (with re- absorbed by the amine , K. mol peatability less than 1%). ΔH = the enthalpy of NO physical absorption into the Ä ä Ä ä J J amine , R   =  the universal gas-rate constant mol K. mol and T = temperature (K). 2 Calculations Linearizing gives 2.1 Physical solubility Å ã ΔS −ΔH After the system had reached equilibrium, the equilibrium ln(He )=ln + (9) N O−BEA/AMP R R T g g pressure (P ; kPa) was measured. The number of moles of N O dissolved in the liquid phase at equilibrium n(   ; N O 2 2 kmol) can be determined using: 2.2 Mass-transfer characteristics l g (2) n = n − n N O N O N O 2 2 2 2.2.1 CO -absorption efficiency The moles of N O in the gas phase before (n ; kmol) and N O 2 2 The CO -removal efficiency is computed from three main g 2 after the experiment (n ; kmol) were calculated using N O processes: the ideal gas expression with compressibility factor. The concentration of NO at equilibrium C ( ; kmol/ (i) The CO removed from the gas N O 2 2 2 m ) is: in out l F − F n CO CO N O 2 2 × 100% (10) C = N O 2 in V F CO Then, the N O solubility (He ; kPa m /kmol) is defined as: N O 2 in where F is the flow rate of CO in the feed gas, given CO 2 2 out P − P T v as mol/s; F is the flow rate of CO in the exit gas; and CO He = 2 2 N O C in out N O 2 F − F represents the amount of CO removed from the CO CO 2 2 2 gas phase. The Henry’s Law constant for CO in the BEA-AMP solvent is computed from Equation 5, developed by Versteeg and (ii) The CO absorbed by the liquid solvent van Swaaij [20]: in out ï ò In this case, F − F would be the amount that the CO CO 2 2 He − H o CO 2 He − Amine = He − Amine CO N o amine has absorbed. 2 2 (5) He − H o N o 2 (iii) The amount of CO produced from the amine regeneration Table 1: Operating conditions used in the pilot-plant in out The F − F is obtained from the amount of CO pro- CO CO 2 2 2 experiments duced from the amine regeneration. The average of these three methods was taken as the Parameter Value CO -removal efficiency. The average deviation between the Solvent used 5M MEA, 7M MEA-MDEA, different methods of obtaining the CO -removal efficiency 2M BEA-2M AMP ranged between 0 and 5%. 3 2 Solvent flow rate 1.84 m /m h Feed-gas flow rate 18.5 kmol/m h CO in feed gas 15% Desorber amine inlet 87°C 2.2.2 Mass-balance error temperature A mass-balance error calculation was performed to de- Desorber catalyst HZSM-5 (Si/Al = 19) termine the validity of each run. The calculation com- Desorber catalyst weight 150 g pared the amount of CO removed from the gas phase to the CO amount added to the liquid phase. A value of 10% or less for an experiment was considered as a valid run. Table S1 in the online Supplementary Data ï ò −2284 shows the respective mass-balance errors associated (6) He = 8.55 × 10 exp N O−H O 2 2 with the experiments. ï ò −2044 6 CO removed from gas − CO absorbed by amine 2 2 (7) He = 2.82 × 10 exp Mass balance error%= × 100% CO −H O 2 2 CO removed from gas T Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 268 | Clean Energy, 2019, Vol. 3, No. 4 2.2.3 Catalyst-aided mass-transfer coefficient where L is the molar inert liquid flow rateρ, the molar Astarita [21] pioneered studies into the absorption of density of the solution,x the bulk liquid mole fraction of AL CO in aqueous solutions of MEA using a laminar jet ab- component A and x the bulk liquid mole fraction of com- 2 A sorber. For the absorption process, transfer begins in the ponent A in equilibrium with the bulk gas. gas phase. The concentration gradient created within the For the overall mass-transfer coefficient of the gas gas film serves as the driving force for the CO gas. The side in Equation 14, there are sample points on the ab- interface created between the gas and the liquid acts as a sorption column that help to measure the y and hence A,G driving force to allow the CO gas to cross the gas into the K a (overall volumetric mass-transfer coefficient of the 2 G V kmol liquid film, depending on its solubility. In the non-catalytic gas side, ) for a section of the column that minim- m . kPa pathway, reaction occurs within the liquid film [22]. izes the variation of y and y * (mole fraction of CO A,G A 2 For the catalytic pathway, the CO diffuses from the li- in the gas phase in equilibrium with that of CO in the 2 2 quid film to the interface between the liquid and the solid bulk liquid, mol/mol) with .Z This was further minim- due to the concentration gradient. The gas then diffuses ized by employing the trapezoidal rule for a step size ) (Z across the liquid–solid interface, through the pores of the of 0.00001 m.  A  logarithmic average of all K a values G V catalyst, and finally adsorbs on the surface of the active obtained (for each 0.00001-m step size) was computed. site for reaction. The amine diffuses through the catalyst For Equation 15 (Ka (overall volumetric mass-transfer L v pores to the active site of the catalyst. The products of the coefficient of liquid side, )), due to the lack of X hr A,L reaction desorb from the active site and diffuse through sample points on the desorber, a dilute concentration the catalyst pores to the bulk liquid. The solid, however, was assumed with a log mean average of X and X * at A,L A introduces a new resistance to mass transfer. the top of column and X and X * at the bottom column, A,L A The overall volumetric mass-transfer resistance in as expressed in Equation 16 developed by Osei et al. [14]: 1 1 ) and liquid side ( ) from the terms of the gas side ( n K a K a 1 G v L v Δx AL K a = x  (16) two-film theory modified to include the solid resistance is: L v ρ Z m (1 − x ) x − x AL Al i=1 A 1 1 He 1 = + + (12) A bar over a variable denotes the log mean average value. K a k a k a E k a G v G v L v s x is calculated from the Henry’s Law constant: 1 1 1 1 = + + (13) CO2 (17) = x K a Hk a k a E k a L v G v L v s A HeC where k a represents the individual gas-film resistance, G v C is the total concentration of the liquid components. k a represents the liquid-film resistance and k a repre- L v s sents the solid resistanceHe . and E represent the Henry’s Law constant and mass-transfer enhancement factor, 3 Results and discussion respectively. 3.1 Selection of absorber catalyst The inverse of the overall mass-transfer resistance rep- resents the overall mass-transfer coefficient—a measure of The base catalyst for the absorber was selected from a the mass-transfer performance. This can be computed by screening test performed on BaCO, Cs O/γ-Al O , Mg- Al 3 2 2 3 performing a mass balance around the absorption column Hydrotalcite, CaCO K/MgO, Cs O/α-Al O and Ca(OH) in 3 2 2 3 2 and the overall mass-transfer rate from the gas phase to a semi-batch reactor in terms of the CO -absorption rate liquid, using the overall gas-side mass-transfer coefficient, in comparison with no catalyst (blank) [26 27 , ]. The in- K a , as described in detail elsewhere [23]: G v fluence of the catalyst on the absorption of CO by the ˆ 2M BEA- 2M AMP was evaluated. The results presented 1 y A,G,2 dy A,G K a = χ (14) G v showed high absorption performance of K/MgO, CsO/α- PZ y (1 − y ) (y − y ) A,G,1 A,G A,G Al O and Ca(OH) . However, K/MgO was selected as the 2 3 2 where G is the molar inert-gas flow rate, P the column 1 most suitable solid-base catalyst due to its high mech- pressure, S the cross-sectional area of the column, y the A,G anical strength coupled with the high absorption per - bulk-gas mole fraction of component A (CO ), Z the differ- formance. This makes it the most stable for industrial ential height and y the bulk-gas mole fraction of compo- amine-absorption application [24]. nent A in equilibrium with the liquid-gas concentration of that component. 3.2 Physical solubility results Similarly, a material balance on the desorption column can be performed to yield the overall liquid side 3.2.1 Effect of temperature on the physical solubility of mass-transfer coefficient, K a , as described in detail L v 2M BEA- 2M AMP elsewhere [14]: The physical solubility of CO in the bi-blend solvent of A,L,2 BEA-AMP was evaluated and compared to the conven- dx AL tional solvent of 5M MEA. The results of the Henry’s Law K a = x (15) L v ρ Z m (1 − x ) x − x AL Al constant for NO and CO in water were obtained from the A,L,1 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 269 correlations developed by Versteeg and Swaaij [20]. The Å ã ΔS trend from Figures 57 - shows a decrease in the physical ln = 13.364 (intercept) solubility (increase in Henry’s Law constant) of the N O gas as the temperature increases. This can be explained by ΔH = 1552.1 (slope) the increase in the chemical activity as the temperature increases. The dissolution of the NO within the amine with R = 8.314 is an exothermic process, just like that of CO . In exo- K. mol thermic processes, increasing the temperature decreases J 13.364 ΔS= 8.314 × e = 5293238.7 the solubility of the solute. Increasing the temperature K. mol introduces more heat into the system. So, according to ΔH = 8.314 × 1552.1 = 12904.2 Le Chatelier’s Principle, the system will adjust to this ex- mol cess in heat energy by inhibiting the dissolution of the Substituting into Equation 8 gives: N O gas. The physical solubility of the gas was seen to be Å ã −1552.1 small (high Henry’s Law constant), hence the need for the He = 0.64 ∗ 10 exp N2O−BEA/AMP chemical reactivity with the amine. The experiment was validated with 5M MEA. The result showed good agree- ment with the work done by Ma’mun et  al. [17] with an 3.3 Pilot-plant results AAD of 1.5%. Figure 8 shows a logarithmic plot of the Henry’s Law 3.3.1 Solvent effect on mass-transfer performance constant of NO in the 2M BEA-2M AMP solution against The solvent screening experiment was performed to se- the negative reciprocal of the system temperature. From lect the best solvent for the CO -absorption unit using the Figure 8, a correlation was developed for the Henry’s Law lab-scale pilot plant. The solvents screened were: 5M MEA, constant of NO in the 2M BEA-2M AMP solution. The phys- 7M bi-blend of MEA-MDEA (5:2 ratio) and a 4M bi-blend of ical solubility of CO in the 2M BEA-2M AMP solvent was BEA-AMP (2:2 ratio). This work selected the conventional compared to that of 5M MEA, as shown in Figure 9. The 5M MEA solvent to validate its performance from previous result shows a lower Henry’s Law constant at the various work [14] while providing a base case to select other suitable temperatures with 2M BEA-2M AMP. This is due to the low solvents. The 7M MEA/MDEA blend also served as a valid- solvent–solvent interaction of the BEA-AMP solvent re- ation from previous work [14] confirming higher perform- sulting in a higher solubility of the gas even though the ance in terms of faster reaction rate, faster mass transfer, viscosity of the 2M BEA-2M AMP is higher than that of 5M higher solvent capacity and lower heat duty. The selection MEA as reported in the literature [6]. Viscosity plays an im- of the 4M BEA/AMP blend was also based on previous work portant role in determining the physical solubility of a gas. showing an even faster reaction rate, faster mass transfer, The higher the viscosity, the lower the physical solubility. higher solvent capacity and lower heat duty over other The validity of the correlation developed from Figure 8 was solvents as reported by Narku-Tetteh et  al. [4]. The 4M confirmed with the experimental results in a parity plot BEA/AMP was then selected for a pilot-plant solvents test shown in Figure 10. The results show a strong correlation with the 7M MEA/MDEA and base 5M MEA. The average with the experimental work with an AAD of 1.0%. operating times for the different solvents are outlined in The He is fitted to the Van’t Hoff plot as given Table S2 in the online Supplementary Data. The tempera- N2O-BEA/AMP in Equation 8. ture profile for CO -amine absorption using the different From Figure 8, a plot of ln He as against solvents has been reported by Afari et al. [25] showing the N2O−BEA/AMP –1/T gives: highest reactivity for the 2M BEA-2M AMP bi-blend. The concentration profile (Figure 11) of the absorber for the dif- ferent solvent systems also indicates a high CO -removal efficiency for BEA-AMP, with the lowest concentration of the CO at the top of the absorption column. This reaffirms the high reactivity of the BEA-AMP solvent and relates the high reactivity to the high mass-transfer coefficient, as it reduces the free CO concentration in the liquid and cre- 2 2 ates a higher concentration gradient for mass transfer. This also corroborates the results of Narku-Tetteh et al. [4]. 4500 1 The performance can be attributed to the structure of the BEA-AMP solvent. The AMP is a sterically hindered amine that forms a very unstable carbamate from the re- action with CO. The unstable carbamate easily hydrolyses 3.05 3.1 3.15 3.2 3.25 3.3 3.35 to bicarbonate and, in so doing, frees up more amine for 1000/T, K reaction. This is the major reason for the high absorption Fig. 5: Henry’s Law constant of NO in water at different temperatures capacity of the amine given as 1 mol of CO per 1 mol of He in H O, Pa.m /mol N2O 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 270 | Clean Energy, 2019, Vol. 3, No. 4 3.05 3.1 3.15 3.2 3.25 3.3 3.35 1000/T, K Fig. 6: Henry’s Law constant of CO in water at different temperatures this work ref 1 (Ma'mun et al) ref 2 (Tsai et al ) 3.05 3.1 3.15 3.2 3.25 3.3 3.35 1000/T, K Fig. 7: Henry’s Law constant experimental setup validation plot (using 5M MEA) against 1000/T (at different temperatures) 8.6 3 8.55 8.5 8.45 2 8.4 8.35 8.3 8.25 8.2 8.15 8.1 –0.00335 –0.0033 –0.00325 –0.0032 –0.00315 –0.0031 –0.00305 –0.003 –1/T Fig. 8: A plot of the natural log of Henry’s Law constant for N O in 2M BEA-2M AMP against the negative reciprocal of temperature (ln He 2 N2O-BEA/AMP plot against –1/T) ln (He) Heco MEA, Pa.m /mol 2 3 HeCO in H O, Pa.m /mol 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 271 He CO2-BEA/AMP He CO2-MEA 300 305 310 315 320 325 Temperature, K Fig. 9: Henry’s Law constant for CO in 2M BEA-2M AMP compared with 5M MEA at different temperatures of the liquid-film resistance and an increase in the con- centration gradient. This is evident from the concentra- tion profile of the absorber in Figure 11, which shows the highest drop for the 2M BEA-2M AMP solvent over the 7M MEA-MDEA and 5M MEA. AAD 1.02% Under absorber performance, from Figures 12 and 13, the overall mass-transfer coefficient validates the improve- ment with the BEA-AMP solvent from a 101 and 239% in- a and K a , respectively, over the base 5M MEA crease in K G V L V He N2O in BEA/AMP solvent. This can be compared with the 7M MEA-MDEA 3500 y=x blend, which recorded a 7 and 8% increase, respectively, Linear (y=x) in K a and K a over the base 5M MEA solvent. Similar re- G V L V sults on the amine cyclic capacity and CO -removal effi- 3000 4000 5000 6000 7000 Predicted ciency for the different solvents were obtained on the pilot He N O -amine, Pa.m /mol plant as reported by Afari et  al. [25]. The desorption per - formance from Figure 14 showed a 139% increase in the Fig. 10: Parity plot for the Henry’s Law constant for N O in 2M BEA-2M desorption efficiency with the BEA-AMP blend, confirming AMP bi-blend solvent the superior performance of AMP in desorption. AMP [4]. AMP is also very reactive and hence produces 3.3.2 Effect of the HZSM-5 catalyst on mass transfer more unstable carbamates that eventually break down to The use of HZSM-5 as a desorption catalyst on post- form bicarbonates [4]. The unstable nature of the carba- combustion pilot-plant studies has been reported by mate also enhances desorption at a higher temperature. Akachuku [13]. Zeolite catalysts are known for their BEA has a long-chain alkyl group reported [4] to be re- unique pore size, which makes them suitable as a shape- sponsible for the high reactivity of the solvent. The alkyl selective catalyst. They have a high surface area as well group present in BEA increases the basicity of the solvent as being crystalline. The results from the overall mass- by inducing the transfer of electrons from the nitrogen ion transfer coefficient in Figures 12–14 show a significant im- to the alkyl group. provement. This can be explained by the HZSM-5 catalytic The MEA-MDEA blend, however, demonstrated better role in desorption in the mechanism shown in Figure 15. performance when compared to the conventional 5M MEA. ion first from its active The improvement in CO desorption is attributed to the The HZSM-5 catalyst donates its H site to the carbamate, which reacts to form free amine, easy breakdown of bicarbonates formed from the MDEA and releases the CO (Steps 1 and 2). The catalyst also do- as well as the role of MDEA in promoting the desorption + 3– nates part of its H protons to the breakdown of HCO as proton transfer, its catalytic role. The higher molarity of well [12]. The catalyst becomes a conjugate base after the MEA-MDEA blend was also a factor for the increased donating its proton, making it capable of accepting an H performance when compared against the 5M MEA, as more proton. The catalyst recovers its H ion from the proton- free amines were available for reaction. The issue with the ated amine, thereby breaking and lowering the energy high molarity was put to the test when comparing the 7M barrier for the release of CO . This results in higher CO MEA-MDEA blend with the 2M BEA-2M AMP. Mass transfer 2 2 desorption and translates to the much lower lean loading is enhanced with the chemical reaction by the reduction Experimental He N O -amine, Pa.m /mol He CO -amine, Pa.m /mol 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 272 | Clean Energy, 2019, Vol. 3, No. 4 0.9 0.8 0.7 0.6 5M MEA blank 0.5 0.4 7M MEA/MDEA blank 0.3 4M BEA/AMP blank 0.2 0.1 5.00 7.00 9.00 11.00 13.00 15.00 CO Conc, % (in gas) Fig. 11: Absorber concentration profiles of the different solvent systems for the non-catalytic (blank) process 0.14 0g HZSM5 with 150g HZSM5 0.12 0.1 0.08 0.06 0.04 0.02 4 M BEA/AMP (2:2) 7 M MEA/MDEA(5:2) 5M MEA Fig. 12: Catalyst (HZSM-5) effect on the overall mass-transfer coefficient (K a ) for absorption as a function of the solvent G V 1.2 0.8 0.6 0.4 0.2 4 M BEA/AMP (2:2) 7 M MEA/MDEA(5:2) 5M MEA 0g HZSM5 150g HZSM5 Fig. 13: Catalyst (HZSM-5) effect on the overall mass-transfer coefficient (K a ) for desorption as a function of the solvent L V shown in Table 2. The lean loading confirms the benefit of leading to a corresponding increase in the absorption per - HZSM-5 in improving the desorption rate. This explains formance with the addition of the HZSM-5 catalyst in the the increase in the desorption performance (Figure 14) desorber. column Height, m K a 1/hr L V K a , kmol/m .kPa.hr G v Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 273 45.00% 40.00% 35.00% 0g HZSM-5 150g HZSM-5 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% 4 M BEA/AMP (2:2) 7 M MEA/MDEA(5:2) 5M MEA Fig. 14: Catalyst (HZSM-5) effect on CO -desorption efficiency as a function of solvent Table 2: The desorption column lean amine loading (mol of Step 1: Deprotonation of HZSM-5 CO /mol of amine) – – – + HZSM – 5 + HCO /AMPCOO ⇔ AMPH COO + ZSM – 5 Solvent Non-catalytic HZSM-5 in desorber Step 2: Carbamate breakdown 5M MEA 0.42 0.41 7M MEA-MDEA 0.35 0.32 AMPH COO ⇔ AMP + CO 2M BEA-2M AMP 0.33 0.30 Step 3: Recovery of proton by ZSM-5 AMPH + ZSM – 5 ⇔ AMP + HZSM –5 catalyst. As explained earlier, the two reactions involved in the CO reaction with amine includes the electron-transfer Fig. 15: HZSM-5 AMP deprotonation mechanism step to form zwitterion (rate-determining step) and car - bamate formation from zwitterion deprotonation, which The catalytic contribution of HZSM-5 in improving the occurs spontaneously [8, 9]. As such, a Lewis-base catalyst overall mass-transfer coefficient for the different solvents capable of donating electrons is desired. was evaluated as shown in Figs 12 and 13. A  38.7% Ka The original role of the amine is to bond with CO and G v and 23.6% K a increase with the HZSM-5 catalyst on 2M break the carbon and oxygen double bond to form the L v BEA-2M AMP blend, 32.2% K a and 57.6% K a increase on zwitterion molecule. The catalyst therefore initiates the G v L v 5M MEA-2M MDEA blend, and 22.0% K a and 36.5% K a carbon double-bond breakage (an energy-requiring pro- G v L v increase on 5M MEA were obtained. Any percentage in- cess) leaving the amine with the single role of donating crease in the mass-transfer coefficient implies a reduction its lone-pair electrons. The K promoter destabilizes the in the size of the column by about the same margin. Also, bond between Mg and O, thereby increasing the amount 2– the increased rate of mass transfer as the concentration of O Lewis-basic sites. Also, magnesium oxide is an ac- gradient across the film increases results in a much lower tive catalyst for the hydrogenation of 1,3-butadiene. It has 2+ solvent loading. The presence of the HZSM-5 catalyst re- been reported that Mg has acid sites that are responsible sults in a lower solvent loading for the same heat input to for the reduction in the catalytic-base performance of the the desorber/reboiler. This is significant, since the energy K/MgO [26]. Thus, another role of K is to poison these acid cost accounts for more than 70% of the total CO -capture sites. The increase in the rate of reaction also results in plant cost [8, 9]. a higher driving force for mass transfer as the concentra- tion of free CO in the liquid reduces. All these result in a faster rate of mass transfer. The electron-donating ability 3.3.3 Effect of the presence of K/MgO and HZSM-5 of the K/MgO catalyst also results in a higher rich loading, catalyst on mass transfer as more CO gets absorbed into the amine. In addition, the The selected solid-base catalyst K/MgO was investigated in cyclic capacity also increases. its application in the absorption column. The desorption The temperature profile and concentration profile for column was loaded with 150 g HZSM-5 while the effect of the application of both catalysts on the 2M BEA-2M AMP introducing 150 g of K/MgO in the absorption column was were reported by Narku-Tetteh et  al. [27] with results evaluated. The results were evaluated in terms of the ab- demonstrating the progression of improved CO absorp- sorption efficiency, amine cyclic capacity and overall mass- tion when both catalysts are placed in the respective col- transfer coefficients, which showed improvement with the umns. The catalyst weight used was 150 g of HZSM-5 based use of a solid-base catalyst in conjunction with a solid-acid CO desorp. eff 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 274 | Clean Energy, 2019, Vol. 3, No. 4 on previous studies [12, 14]. Figure 16 shows the effect of K/ forming a bond between the catalyst and CO (Step 1). The MgO catalyst on rich amine loading. Results from Figure 16 bond formed violates the carbon octet rule, making it un- confirm improvements in CO absorption as well as an in- stable and causing the breakdown of one of the double crease in amine loading resulting from the K/MgO catalyst bonds between carbon and oxygen (C=O). This puts a par - influencing the absorption of CO . tial charge on the oxygen (from the C=O) and makes the For the application of both catalysts on the 2M BEA-2M CO molecule more reactive to the amine (Step 2). The AMP blend, the overall mass-transfer coefficient (Figures nitrogen of the amine now donates its lone electrons to 17 and 18), 38.7% K a and 23.6% K a increases were the catalyst–CO bond and eventually breaks the bond be- G v L v 2 obtained with HZSM-5 catalyst alone, while 95% K a and tween the catalyst and CO . The newly formed compound G v 2 45% K a increases were obtained with both K/MgO and between the amine and CO represents the zwitterion L v 2 HZSM-5, as compared with the base case of no catalyst in molecule (Step 3). both columns. 3.3.5 Effect of K/MgO catalyst weight on mass transfer 3.3.4 Proposed K/MgO catalytic-reaction mechanism Results from Figures 20 and 21 show an increase in the K/MgO catalyst has been reported as having Lewis-base weight of K/MgO, resulting in an increase in the mass- sites [26]. The proposed mechanism in which the K/MgO transfer performance. This can be attributed to the provides electrons is as a result of bond destabilizing be- increase in the reaction rate as the catalyst weight in- tween the magnesium and oxygen bond (Mg–O), as shown creases. The increase in the catalyst weight results in an in Figure 19. This makes electrons available for bonding increase in the total active catalyst sites available for re- with CO . K is more electropositive than Mg, hence it de- action. The number of active sites and total surface area stabilizes and weakens the bond between Mg and O. The of the active sites increase with the increased catalyst 2– O from the weakened bond donates electrons to CO , weight. This, therefore, results in an increase in the CO 0.6 0.58 0.56 0.54 0.52 0.5 0.48 0.46 0g K/MgO 150g K/MgO Fig. 16: Effect of K/MgO catalyst on rich amine loading 0.18 Solvent 0.16 Solvent+HZSM5 0.14 Solvent+HZSM5+K/MgO 0.12 0.10 0.08 0.06 0.04 0.02 0.00 Fig. 17: Effect of catalyst configuration on Ka of absorber G v Rich loading, mol CO /mol amine KGav, kmol/m .kPa.hr Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 275 1.40 Solvent 1.20 Solvent+HZSM5 Solvent+HZSM5+K/MgO 1.00 0.80 0.60 0.40 0.20 0.00 Fig. 18: Effect of catalyst configuration on Ka of desorber L v Step 1: Oxygen-containing catalyst donates a lone-pair electrons to CO +2 –2 Mg O Mg O OO C OO C Step 2: C = O bond breaking and electron transfer onto O Mg O Mg O O C O OO C Step 3: Nitrogen now donates its electrons to the carbon which breaks and forms N-C bond of zwitterion. Mg O C O = +2 –2 + Mg O NH O C O NH HO HO (Zwitterion) Fig. 19: Proposed K/MgO catalytic mechanism 1.2 0.2 1.15 0.15 1.1 0.1 1.05 0.05 0 50 100 150 200 0.95 K/MgO Catalyst weight, g 0.9 Fig. 20: Effect of K/MgO catalyst weight on K a of absorber 0 50 100 150 200 G v K/MgO weight, g and amine reactivity on the catalyst active site, as re- Fig. 21: Effect of K/MgO catalyst weight on K a of desorber L v ported elsewhere [24]. The increase in the reaction rate creates a higher con- decrease in the liquid-film resistance, seen from Equations centration gradient as free CO in the liquid is consumed at 10 and 11. The combined effect of an increase in the reac- a faster rate. This increased concentration gradient serves tion rate resulting in a decrease in the liquid-film transfer as the driving force for mass transfer and is directly propor - resistance, an increase in the interfacial area and an in- tional to the rate of transfer. Most importantly, the increase crease in the concentration gradient accounts for the in- in the reactivity results in an increase in the enhancement creased performance with the increase in catalyst weight. factor. The enhancement factor, which is a ratio of the ab- Figures 20 and 21 show the positive K/MgO catalyst weight sorption rate of CO due to reaction to the physical absorp- performance against Ka and K a . However, it needs to be tion of CO by the same solvent and conditions, results in a G V L V K a , kmol/m .kPa.hr G v KLav, 1/hr K a , 1/hr L V Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 276 | Clean Energy, 2019, Vol. 3, No. 4 pointed out that, above 150  g of K/MgO, the performance The mass flux of CO across the gas-liquid inter - does not change much across all results. kmol face, m s kPa. m He Henry’s law constant, kmol E enhancement factor 3.4 Industrial considerations y * mole fraction of CO in the gas phase in equilib- A 2 Coker [28] assessed the reduction in the capital expenses rium wh that of CO in the bulk liquid. mol/mol (CAPEX) as well as the energy required for operational ex- G molar inert flow rate (total gas without CO ) 1 2 S cross-sectional area of the column. penses (OPEX). Results showed a significant reduction in L molar inert liquid flowrate, . both the CAPEX and OPEX, which justified the use of cata- 1 ρ molar density of the solution, mol/m lysts in both the absorber and desorber. However, other m x bulk liquid mole fraction of component A, mol/ AL factors including pressure drop and liquid viscosity may mol determine the need for blower and higher pump power, x bulk liquid mole fraction of component A in equi- respectively. It is only after these studies are conducted librium with the bulk gas. mol/mol that a conclusive justification can be made. Therefore, fur - [A ] equilibrium solubility of CO in the amine, mol/m ther comprehensive studies on catalytic roles that include D physical diffusivity of CO in amine, m /s A 2 power consumption, pressure drop, how the absorber cata- k reaction rate constant between the CO and mn lyst reacts to species like SO , NO in flue gas and the poten- amine x x tial cost of replacing the catalyst if it becomes poisoned are m- order of reaction with respect to CO n order of reaction with respect to the amine needed to access the economic feasibility of this process. B concentration of unloaded amine, mol/m 4 Conclusions Acknowledgements This work has demonstrated in great measure the effect The financial support provided by the Natural Science and of a solvent, its chemistry and its catalytic effect on the Engineering Research Council of Canada (NSERC), the Canada mechanism and rate of mass transfer for CO absorp- Foundation for Innovation (CFI) and the Clean Energy Technologies tion and desorption. There are big increases in the mass- Research Institute (CETRI), University of Regina is gratefully transfer coefficients for CO absorption and desorption acknowledged. with the introduction of a catalyst to the capture process. This manifests in the significant increase in mass-transfer Supplementary data rates with the catalyst. The solvent BEA-AMP is seen to outperform the other solvents in mass transfer. The ex- Supplementary data is available at Clean Energy online. tensive study of the effect of catalyst weight shows an increase in the CO-absorption rate or mass-transfer per - References formance with an increase in the catalyst weight. This im- [1] Rao AB, Rubin ES. A technical, economic, and environmental plies that employing a catalyst in the CO -capture process assessment of amine-based CO capture technology for will introduce a significant increase in the mass-transfer power plant greenhouse gas control. Envir on Sci Technol 2002; 36:4467–75. performance, which will greatly reduce the heights of the [2] Houghton JT, Ding Y, Griggs DJ, et al. Climate change 2001. In: absorption and desorption columns. The catalyst introduc- Houghton JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, tion replaces part of the structured packing, and therefore Dai X, Maskell K, Johnson CA (eds). The Scientific Basis. New reduces the cost of the structured packing. York: University Press, 2001, 881. [3] Luis P. Use of monoethanolamine (MEA) for CO capture in a global scenario. Consequences and Alternatives 2016; 380:93–9. Nomenclature [4] Narku-Tetteh J, Muchanb P, Saiwanb C, et al. Selection of com- ponents for formulation of amine blends for post combustion K a Overall Volumetric mass transfer coefficient of gas G v CO capture based on the side chain structure of primary, sec- kmol side, ondary and tertiary amines. Chem Eng Sci 2017; 170:542–60. m . kPa K a Overall Volumetric mass transfer coefficient of li- [5] Idem  R, Wilson  M, Tontiwachwuthikul  P, et  al. Pilot plant L v studies of the CO capture performance of aqueous MEA and quid side, . hr mixed MEA/MDEA solvents at the University of Regina CO R universal gas constant, kJ/kmol.K capture technology development plant and the boundary AAD Absolute average deviation, % dam CO capture demonstration plant. Ind Eng Chem Res 2006; MEA Monoethanolamine 2 45:2414–20. MDEA Methyldiethanolamine [6] Danckwerts P. The reaction of CO with ethanolamines. Chem BEA Butyl Monoethanol amine Eng Sci 1979; 34:443–6. AMP 2- amino 2 methyl 1 propanol [7] Caplow M. Kinetics of carbamate formation and breakdown. J in F flowrate of CO in the feed gas, mol/s. CO 2 2 Am Chem Soc 1968; 90:6795–803. out F flowrate of CO in the exit gas, mol/s. CO2 2 [8] Shi H, Idem R, Naami A, et al. Catalytic solvent regeneration y mole fraction of CO in gas bulk, mol/mol A,G using hot water during amine based CO capture process. y mole fraction at gas interphase. mol/mol A,i Energy Procedia 2014; 63:266–72. Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 277 [9] Shi  H, Naami  A, Idem  R, et  al. Catalytic and non-catalytic [19] Srisang W, Pouryousefi Y, Osei PA, et al. Evaluation of the heat solvent regeneration during absorption-based CO capture duty of catalyst-aided amine-based post combustion CO cap- 2 2 with single and blended reactive amine solvents. Int J Greenh ture. Chem Eng Sci 2017; 170:48–57. Gas Con 2014; 26:39–50. [20] Versteeg GF, van Swaaij WPM. Solubility and diffusivity of acid [10] Liu B, Luo X, Liang Z, et al. The development of kinetics model gases (CO , N O) in aqueous alkanolamine solutions.J Chem 2 2 for CO absorption into tertiary amines containing carbonic Eng Data 1988; 33:29–34. anhydrase. AIChE J 2017; 63:4933–43. [21] Astarita  G, Savage  DW. Gas absorption and desorption with [11] Gladis  A, Gundersen  MT, Fosbøl  PL, et  al. Influence of tem- reversible instantaneous chemical reaction. Chem Eng Sci perature and solvent concentration on the kinetics of the en- 1980; 35:1755–64. zyme carbonic anhydrase in carbon capture technology. Chem [22] Levenspiel O. Chemical Reaction Engineering, 3rd edn. New York: Eng J 2017; 309:772–86. Wiley, 1999. [12] Idem R, Shi H, Gelowitz D, et al. Catalytic method and apparatus [23] Naami A, Edali M, Seema T, et al. Mass transfer performance for separating a gaseous component from an incoming gas stream. of CO absorption into aqueous solutions of 4-diethylamino-2- United States Patent: US 9,586,175 B2. March 7, 2017. 2017. butanol, monoethanolamine, and N-methyldiethanolamine. [13] Akachuku A. Kinetic study of the catalytic desorption of carbon di- Ind Eng Chem Res 2012; 51:6470–9. oxide (CO ) from CO -loaded monoethanolamine (MEA) and blended [24] Coker  J, Afari  DA, Narku-Tetteh  J, et  al. Synthesis and appli- 2 2 monoethanolamine-methyldiethanolamine (MEA-MDEA) during cation of solid base catalyst in a post combustion capture of post-combustion CO capture from industrial flue gases. MASc. CO - a mass transfer study. In: 25th Canadian Symposium on 2 2 Thesis. University of Regina, Canada, 2016. Catalysis Saskatoon, SK, 8–11 May, 2018. [14] Osei PA. Mass transfer studies on catalyst-aided CO desorption in [25] Afari  DA, Coker  J, Narku-Tetteh  J, et  al. Comparative kinetic a post-combustion CO capture plant. MASc. Thesis. University of studies of solid absorber catalyst (K/MgO) and solid desorber Regina, Canada, 2016. catalyst (HZSM-5)-aided CO absorption and desorption from [15] Zhang X, Zhang R, Liu H, et al. Evaluating CO desorption per - aqueous solutions of MEA and blended solutions of BEA-AMP formance in CO -loaded aqueous Tri-solvent blend amines and MEA-MDEA. Ind Eng Chem Res 2018; 57:15824–39. with and without solid acid catalysts. Appl Energy 2018; [26] Ono Y, Hattori H. Solid Base Catalysis. New York: Springer- 218:417–29. Verlag Berlin Heidelberg and Tokyo Institute of Technology [16] Sema T. Kinetics of carbon dioxide absorption into aqueous solu- Press, 2011. tions of 4-(diethylamino)-2-butanol and blended monoethanolamine [27] Narku-Tetteh J, Afari DA, Coker J, et al. Evaluation of the roles and 4-(diethylamino)-2-butanol. PhD. Thesis. University of Regina, of absorber and desorber catalysts in the heat duty and heat of Canada, 2011. CO desorption from butylethanolamine 2-amino-2-methyl- [17] Ma’mun  S, Svendsen  H. Solubility of NO in aqueous 1-propanol and monoethanolamine methyldiethanolamine monoethanolamine and 2-(2-aminoethyl-amino)ethanol so- solvent blends in a bench-scale CO capture pilot plant. Energy lutions from 298 to 343 K. Energy Procedia 2014; 1:837–43. Fuels 2018; 32:9711–26. [18] Tsai  T, Ko  J, Wang  H, et  al. Solubility of nitrous oxide in [28] Coker J. Mass transfer and emission studies on a catalyst-aided CO alkanolamine aqueous solutions.J Chem Eng Data 2011; absorption and desorption in a post combustion CO capture plant. 45:341–7. MASc. Thesis. University of Regina, Canada, 2018. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Clean Energy Oxford University Press

Mass-transfer studies of solid-base catalyst-aided CO2 absorption and solid-acid catalyst-aided CO2 desorption for CO2 capture in a pilot plant using aqueous solutions of MEA and blends of MEA-MDEA and BEA-AMP

Loading next page...
 
/lp/oxford-university-press/mass-transfer-studies-of-solid-base-catalyst-aided-co2-absorption-and-q1hesLk951
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/zkz015
Publisher site
See Article on Publisher Site

Abstract

Mass-transfer studies of catalyst-aided CO absorption and desorption were performed in a full-cycle, bench- scale pilot plant to improve CO absorption using 5M MEA, 5M MEA-2M MDEA and 2M BEA-2M AMP. A solid-base catalyst, K/MgO, and an acid catalyst, HZSM-5, were used to facilitate absorption and desorption, respectively. Absorption and desorption mass-transfer performance was presented in terms of the overall mass-transfer coefficient of the gas side (K a ) and liquid side (K a ), respectively. For non-catalytic runs, the highest K a and G v L v G V Kmol 1 K a were 0.086 and 0.785 for 2M BEA-2M AMP solvent. The results showed 38.7% K a and 23.6% K a L V hr G v L v m .kPa.hr a and 45% K a increase for increase for 2M BEA-2M AMP with only HZSM-5 catalyst in desorber and a 95% K G V L V both K/MgO catalyst and HZSM-5 catalyst. This was attributed to the role of K/MgO in bonding loosely with CO and making it available for the amine reaction. Keywords: HZSM-5; K/MgO catalyst; CO removal efficiency; CO capture; mass-transfer coefficients of absorber 2 2 (K a ) and desorber (K a ) G V L V include monoethanolamine (MEA), diethanolamine (DEA), Introduction methyldiethanolamine (MDEA) and 2-amino-2-methyl-1- The process of post-combustion capture of CO with an propanol (AMP). Despite the high reactivity of MEA with amine solvent has been well accepted as the most re- , its use has been hampered by its high regeneration CO searched and cost-effective technology in the field of -absorption capacity heat. In contrast, MDEA has high CO capture [1, 2]. In this technology, the chemical reac- CO and low regeneration heat, but has slow reactivity with gas re- tion between an aqueous amine solvent and CO [3]. The process of improving the post-combustion CO sults in an enhanced mass transfer within the solvent. -capture process is divided into solvent formulation CO -absorption processes Common amines in commercial CO and process optimization [4]. Received: 29 March, 2019; Accepted: 13 May, 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, 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/3/4/263/5550982 by DeepDyve user on 10 December 2019 264 | Clean Energy, 2019, Vol. 3, No. 4 Under solvent formulation, studies on the blending of as the rate-determining step [, 89]), followed by the re- solvents are intended to obtain high-performance solv- action between the abstracted proton and carbamate to ents with higher CO cyclic capacity, lower solvent regen- form CO and free amine. Technologies employed to in- 2 2 eration energy requirement and faster absorption kinetics crease the rate of CO-amine absorption and desorption with favourable environmental characteristics. Recently, include the addition of either a homogeneous or heter - Narku-Tetteh et  al. [4] used study of the amine chemical ogenous component to the solvent as a catalyst to in- structure–activity relationships to develop the criteria crease the chemical reactivity between CO and amine. for selecting components to formulate high-performing Liu et al. [10] investigated CO absorption into MDEA and amine blends that resulted in the development of an effi- dimethylaminoethanol (DMEA) with the application of cient bi-solvent blend consisting of 2M butyl monoethanol carbonic anhydrase as a homogeneous catalyst. The re- amine (BEA)-2M AMP. Results from the work of Idem et al. sults showed an enhanced absorption of CO with a low [5] showed a lower heat duty for a blend consisting of 7 M concentration of carbonic anhydrase in the amine, which MEA-MDEA, in the ratio of 5:2, when compared with con- was attributed to its role in promoting the hydration of ventional 5M MEA. CO , the rate-determining step. Similar results have been The reaction steps for both amine CO absorption reported in the literature [11]. and desorption are shown in Figure 1. For the amine Idem et  al. [12] identified several catalysts for appli- CO -absorption process, two reactions occur according cation in the field of post-combustion capture. Recent to the zwitterion mechanism. These include CO reac- works [13, 14] have shown improvements in the regen- tion with the amine to produce the zwitterion molecule eration process with the introduction of a solid-acid in an electron-transfer process. The zwitterion molecule catalyst in the desorption unit for different solvent sys- then reacts with any base in a hydrogen ion (H proton) tems. Two major catalysts were studied, which included abstraction process to form a carbamate molecule. The HZSM-5, a Bronsted-acid catalyst; and gamma alumina mechanism is only different for tertiary amines where (ɣ-Al O ), a Lewis-acid catalyst. Results identified HZSM-5 2 3 the CO reacts with the amine and water in a single-step as the most suitable catalyst, which was attributed to process to form bicarbonate ions and water. The two-step the donation of its H proton for carbamate breakdown. zwitterion-reaction mechanism has been identified as a Zhang et  al. [15] also investigated the application of a base catalysed reaction []. 6 The electron-transfer process solid-acid catalyst including HZSM-5,-Al ɣ O , SAPO-34 2 3 2– of CO reacting with the amine has also been identified as and SO /TiO to increase the rate of CO regeneration 2 4 2 2 the rate-determining step []. 7 This makes the base cata- for loaded aqueous MEA, as well as a tri-solvent blend of lyst that is capable of donating a lone-pair electrons (i.e. MEA-AMP-PZ. Results also showed the enhanced amine a Lewis-base catalyst) the most suitable for increasing desorption rates and lower heat duty with the cata- the reactivity between the CO and amine. The Lewis- lyst, attributed to its high Bronsted/Lewis ratio and its base catalyst first donates a lone pair of electrons from mesoporous surface area. This promotes proton transfer its anion to the CO molecule by bonding loosely with to the carbamate (rate-determining step). The intro- CO and allowing more contact time for the reaction with duction of the desorption catalyst showed a significant amine. This makes the CO more susceptible to the amine. improvement in the heat duty; however, its indirect The desorption reaction also involves two main re- improvement on the CO-removal efficiency was only actions. The abstraction of a hydrogen (H) proton marginal. (deprotonation) from a protonated amine (identified This work intends to improve CO-removal efficiency by introducing a base catalyst and an acidic catalyst to enhance reaction kinetics in the absorption and desorp- AB tion columns, respectively. The overall mass-transfer coefficient represents the constant of proportionality be- Amine +CO +H O 2 2 Amine +CO tween the mass-transfer rate or flux and its driving force. A higher mass-transfer rate also results in a faster reaction rate. Faster mass transfer allows the operator to achieve the same CO -capture rate (or removal efficiency) with + – Amine COO +Amine a shorter column. Essentially, the heights of the absorp- (zwitterion formation) tion and desorption columns are inversely proportional to the overall mass-transfer coefficient, meaning that a high mass-transfer coefficient will result in a shorter column. + – Consequently, this will lead to smaller sizes of equipment, AmineH + HCO – + AmineCOO + AmineH and thus low capital costs. The results obtained from these (carbamate formation) mass-transfer studies for different catalyst configurations using 5M MEA, 5M MEA-2M MDEA and 2M BEA-2M AMP Fig. 1: Reaction pathways for amines. (a) Primary and secondary amines, (b) sterically hindered amines. solvents are presented in this paper. Absorption Desorption Absorption Desorption Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 265 average deviations (AADs) of 1.5 and 1.3%, respectively, 1 Experimental section which confirms the validity of the N O experiment (with 1.1 Materials and equipment repeatability less than 1%). The chemicals used and where they were obtained are as follows. Reagent-grade butylethanolamine (≥98%), 1.3 Amine structure 2-amino-2-methyl-1-propanol (≥99%), monoethanolamine (≥99%) and methyl diethanolamine (≥99%) were purchased The amines used for this study were monoethanolamine from Sigma-Aldrich Co., Canada. The 1 N hydrochloric acids (MEA), methyldiethanolamine (MDEA), butylethanolamine (HCl), used for titration to determine the concentration of (BEA) and 2-amino-2-methyl-1-propanol (AMP). Their the amines, was also purchased from Sigma-Aldrich. Pure chemical structures are given in Figure 2. CO and N gases were purchased from Praxair Inc., Regina. 2 2 The CO gas analyser was purchased from NOVA analytical system Inc. Gas-analyser calibrations were done using pure 1.4 K/MgO catalyst preparation N gas and 15% CO (N balance) also acquired from Praxair 2 2 2 The catalyst was prepared by the incipient wetness Inc., Regina. The stainless-steel LDX Sulzer structured method. Commercially available Mg(OH) purchased from packing with an outside diameter of 0.047 m was provided Sigma-Aldrich was used in the preparation of the sup- by Sulzer Chemtech Ltd. Temperatures during operation port. The Mg(OH) was first calcined at 600°C for 2 hours were recorded using J-type thermocouples from Cole to produce MgO, following which the catalyst was allowed Parmer, Canada. Commercially available Mg(OH) (≥98%) to cool for 1 hour. A  solution of KOH was prepared using and KOH (≥98%) were purchased from Sigma-Aldrich. the catalyst pore volume. The solution concentration was such that the final composition of K, the active metal, was 1 mol% on the support. The prepared slurry was dried for 1.2 Physical solubility of CO (N O analogy) 2 2 12 hours and later calcined at 600°C for 2 hours. The final The experimental setup used for determining N O solu- catalyst was pelletized to obtain particle sizes in the range bility was similar to one used by Sema [16]. The solu- of 4.2–4.7 mm. bility experiment was performed in a rotary-type 600-ml stainless-steel autoclave reactor (model Parr 5500, Parr Instrument Co., Moline, IL, USA) connected to a con- 1.5 Bench-scale pilot-plant experimental setup troller (model Parr 4843, Parr Instrument Co., Moline, IL, The setup describing the experimental system is shown USA). The reactor consisted of a variable-speed impeller, in Figure 3. The system comprises two lagged stainless- heating mantle, cooling coil, gas-feed port, thermo- steel columns, each measuring 3.5 ft (1.067 m) in height couple and pressure transducer. and having an internal diameter of 2 inches (0.0508 m). The Initially, the amine solution was degassed using an internal arrangements of both the absorber and desorber ultrasonic bath (VWR model 75D, VWR International, ON, columns with catalysts are shown in Figure 4. The de- Canada). Then, 300 ml of the degassed amine solution was sorber was filled from the bottom with a section of Sulzer introduced into the reactor. The desired temperature was LDX structured packing measuring 0.0508 m in diameter set and controlled by the controller, and then the vacuum and 0.18 m in height. On top of this was placed a 0.025-m pump was turned on. After shutting down the vacuum section of large inert marbles measuring 6 mm in diameter pump, the liquid was under a certain pressure due to li- to act as a support for the catalytic bed. A section of catalyst quid vaporization. The pressure, P (kPa), was measured at of 0.55 m in height was mixed with small, 3-mm-diameter vacuum. A certain amount of NO (n .; kmol) was fed to N O 2 2 the reactor. The amount of N O dissolved in the solvent was determined by measuring the pressure in the reactor be- CH fore N O injection (P ; kPa) and after NO injection (P ; kPa) 2 1 2 2 OH N (±6.25 kPa). These terms were correlated as in Equation (1): HO H N n =(P − P ) (1) MEA N O 2 1 Z RT MDEA N O OH where Vg is the volume of the gas container (m ), R is the gas constant (kJ/kmol.K) and Z is the compressibility N2O factor of NO. Equilibrium is reached when the NO pres- H C 2 2 OH sure does not change after 12 hours. HO CH The equipment and calculation procedure for the N O CH 2 3 NH AMP solubility were validated with a 5M MEA solution in the temperature range of 298  ±  1 to 343  ±  1  K. The experi- BEA mental results were in good agreement with the work by Ma’mun et  al. [17] and Tsai et  al. [18] with absolute Fig. 2: Chemical structure of amines used Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 266 | Clean Energy, 2019, Vol. 3, No. 4 CO Product Offgas Inlet hot water Lean Amine Hot water heat exchanger Outlet hot water Rich Amine Flue gas Lean Amine Rich Amine Lean amine cooler Rich-Lean Heat exchanger Amine tank Fig. 3: Schematic representation of the experimental setup for CO removal Adapted with permission from [19]. Copyright 2017 W. Srisang. drop, flooding and liquid carryover within the absorption 0.051 m column. Structured 0.18 m LDX packing Sulzer 0.025 m 0.0891 m 1.6 Bench-scale pilot-plant experimental large inert (catalyst + marbles procedure empty At the start of the experiment, the amine was fed from space) 0.55 m catalytic bed the storage tank to the top of the absorber column via a 0.96 m (catalyst + variable-speed gear pump and left at the bottom of the LDX small inert marbles) Sulzer column. The amine was then pumped through a lean rich heat exchanger to exchange heat with the hot lean amine from the desorption column. The amine was further heated 0.025 m large inert by an external heater to the desired amine temperature for marbles desorption. The heating medium for this heater was gly- 0.18 m LDX cerol to heat up the rich amine prior to entering the de- Sulzer sorber. The amine in the desorption column exited as lean amine. This lean amine was pumped through the rich-lean Absorber Absorber Desorber heat exchanger and then through a cooler to further reduce with catalyst with catalyst its temperature prior to entering the absorption column. Once amine-solvent circulation was achieved, a mixture of Fig. 4: Column packing and catalyst-bed arrangement for absorber and and N gas at the appropriate CO concentration of 15% CO desorber 2 2 2 was introduced to the bottom of the absorber column and marbles to form the catalytic bed. On top of this was placed monitored by a gas-flow meter. The lean amine contacted a second 0.025-m section of large inert marbles, then an- the gas feed from the bottom of the column counter- other 0.18-m section of Sulzer LDX structured packing. currently. Treated gas left the top of the column while the The absorber column was arranged in an alternating bed rich amine solvent left the absorber bottom. The duration of Sulzer LDX structured packing and K/MgO (absorption of an experimental run was between 4 and 6 hours. catalyst). The absorber was filled from the bottom with a A steady-state operation of each experiment was indi- section of Sulzer LDX structured packing measuring 0.051 cated when there was no temperature change with time m in diameter and 0.051 m in height. On top of this was of different temperature points within the absorption placed the required catalyst weight in a section measuring column, indicating that the solution entering the absorber 0.089 m.  The catalyst filled that section, partly leaving a was at a constant CO concentration, amine concentration void space. This absorber arrangement was repeated in an and flow rate. At this point, the CO concentration in the alternating pattern. This was done to minimize pressure gas phase and temperature profile were measured along Absorber Stripper Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 267 the height of the column (error less than 1%) using the IR 2.1.1 Van’t Hoff equation fitting for He N O-amine gas analyser and thermocouples, respectively. Also, sam- Å ã Å ã ΔS −ΔH ples were taken from the bottom of both columns to de- He = exp (8) N O−BEA/AMP R R T g g termine the rich and lean CO loadings using a Chittick apparatus (error less than 1%). Tabl e 1 shows the operating where ΔS  =  the entropy change as the NO is physically Ä ä conditions used in the pilot-plant experiments (with re- absorbed by the amine , K. mol peatability less than 1%). ΔH = the enthalpy of NO physical absorption into the Ä ä Ä ä J J amine , R   =  the universal gas-rate constant mol K. mol and T = temperature (K). 2 Calculations Linearizing gives 2.1 Physical solubility Å ã ΔS −ΔH After the system had reached equilibrium, the equilibrium ln(He )=ln + (9) N O−BEA/AMP R R T g g pressure (P ; kPa) was measured. The number of moles of N O dissolved in the liquid phase at equilibrium n(   ; N O 2 2 kmol) can be determined using: 2.2 Mass-transfer characteristics l g (2) n = n − n N O N O N O 2 2 2 2.2.1 CO -absorption efficiency The moles of N O in the gas phase before (n ; kmol) and N O 2 2 The CO -removal efficiency is computed from three main g 2 after the experiment (n ; kmol) were calculated using N O processes: the ideal gas expression with compressibility factor. The concentration of NO at equilibrium C ( ; kmol/ (i) The CO removed from the gas N O 2 2 2 m ) is: in out l F − F n CO CO N O 2 2 × 100% (10) C = N O 2 in V F CO Then, the N O solubility (He ; kPa m /kmol) is defined as: N O 2 in where F is the flow rate of CO in the feed gas, given CO 2 2 out P − P T v as mol/s; F is the flow rate of CO in the exit gas; and CO He = 2 2 N O C in out N O 2 F − F represents the amount of CO removed from the CO CO 2 2 2 gas phase. The Henry’s Law constant for CO in the BEA-AMP solvent is computed from Equation 5, developed by Versteeg and (ii) The CO absorbed by the liquid solvent van Swaaij [20]: in out ï ò In this case, F − F would be the amount that the CO CO 2 2 He − H o CO 2 He − Amine = He − Amine CO N o amine has absorbed. 2 2 (5) He − H o N o 2 (iii) The amount of CO produced from the amine regeneration Table 1: Operating conditions used in the pilot-plant in out The F − F is obtained from the amount of CO pro- CO CO 2 2 2 experiments duced from the amine regeneration. The average of these three methods was taken as the Parameter Value CO -removal efficiency. The average deviation between the Solvent used 5M MEA, 7M MEA-MDEA, different methods of obtaining the CO -removal efficiency 2M BEA-2M AMP ranged between 0 and 5%. 3 2 Solvent flow rate 1.84 m /m h Feed-gas flow rate 18.5 kmol/m h CO in feed gas 15% Desorber amine inlet 87°C 2.2.2 Mass-balance error temperature A mass-balance error calculation was performed to de- Desorber catalyst HZSM-5 (Si/Al = 19) termine the validity of each run. The calculation com- Desorber catalyst weight 150 g pared the amount of CO removed from the gas phase to the CO amount added to the liquid phase. A value of 10% or less for an experiment was considered as a valid run. Table S1 in the online Supplementary Data ï ò −2284 shows the respective mass-balance errors associated (6) He = 8.55 × 10 exp N O−H O 2 2 with the experiments. ï ò −2044 6 CO removed from gas − CO absorbed by amine 2 2 (7) He = 2.82 × 10 exp Mass balance error%= × 100% CO −H O 2 2 CO removed from gas T Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 268 | Clean Energy, 2019, Vol. 3, No. 4 2.2.3 Catalyst-aided mass-transfer coefficient where L is the molar inert liquid flow rateρ, the molar Astarita [21] pioneered studies into the absorption of density of the solution,x the bulk liquid mole fraction of AL CO in aqueous solutions of MEA using a laminar jet ab- component A and x the bulk liquid mole fraction of com- 2 A sorber. For the absorption process, transfer begins in the ponent A in equilibrium with the bulk gas. gas phase. The concentration gradient created within the For the overall mass-transfer coefficient of the gas gas film serves as the driving force for the CO gas. The side in Equation 14, there are sample points on the ab- interface created between the gas and the liquid acts as a sorption column that help to measure the y and hence A,G driving force to allow the CO gas to cross the gas into the K a (overall volumetric mass-transfer coefficient of the 2 G V kmol liquid film, depending on its solubility. In the non-catalytic gas side, ) for a section of the column that minim- m . kPa pathway, reaction occurs within the liquid film [22]. izes the variation of y and y * (mole fraction of CO A,G A 2 For the catalytic pathway, the CO diffuses from the li- in the gas phase in equilibrium with that of CO in the 2 2 quid film to the interface between the liquid and the solid bulk liquid, mol/mol) with .Z This was further minim- due to the concentration gradient. The gas then diffuses ized by employing the trapezoidal rule for a step size ) (Z across the liquid–solid interface, through the pores of the of 0.00001 m.  A  logarithmic average of all K a values G V catalyst, and finally adsorbs on the surface of the active obtained (for each 0.00001-m step size) was computed. site for reaction. The amine diffuses through the catalyst For Equation 15 (Ka (overall volumetric mass-transfer L v pores to the active site of the catalyst. The products of the coefficient of liquid side, )), due to the lack of X hr A,L reaction desorb from the active site and diffuse through sample points on the desorber, a dilute concentration the catalyst pores to the bulk liquid. The solid, however, was assumed with a log mean average of X and X * at A,L A introduces a new resistance to mass transfer. the top of column and X and X * at the bottom column, A,L A The overall volumetric mass-transfer resistance in as expressed in Equation 16 developed by Osei et al. [14]: 1 1 ) and liquid side ( ) from the terms of the gas side ( n K a K a 1 G v L v Δx AL K a = x  (16) two-film theory modified to include the solid resistance is: L v ρ Z m (1 − x ) x − x AL Al i=1 A 1 1 He 1 = + + (12) A bar over a variable denotes the log mean average value. K a k a k a E k a G v G v L v s x is calculated from the Henry’s Law constant: 1 1 1 1 = + + (13) CO2 (17) = x K a Hk a k a E k a L v G v L v s A HeC where k a represents the individual gas-film resistance, G v C is the total concentration of the liquid components. k a represents the liquid-film resistance and k a repre- L v s sents the solid resistanceHe . and E represent the Henry’s Law constant and mass-transfer enhancement factor, 3 Results and discussion respectively. 3.1 Selection of absorber catalyst The inverse of the overall mass-transfer resistance rep- resents the overall mass-transfer coefficient—a measure of The base catalyst for the absorber was selected from a the mass-transfer performance. This can be computed by screening test performed on BaCO, Cs O/γ-Al O , Mg- Al 3 2 2 3 performing a mass balance around the absorption column Hydrotalcite, CaCO K/MgO, Cs O/α-Al O and Ca(OH) in 3 2 2 3 2 and the overall mass-transfer rate from the gas phase to a semi-batch reactor in terms of the CO -absorption rate liquid, using the overall gas-side mass-transfer coefficient, in comparison with no catalyst (blank) [26 27 , ]. The in- K a , as described in detail elsewhere [23]: G v fluence of the catalyst on the absorption of CO by the ˆ 2M BEA- 2M AMP was evaluated. The results presented 1 y A,G,2 dy A,G K a = χ (14) G v showed high absorption performance of K/MgO, CsO/α- PZ y (1 − y ) (y − y ) A,G,1 A,G A,G Al O and Ca(OH) . However, K/MgO was selected as the 2 3 2 where G is the molar inert-gas flow rate, P the column 1 most suitable solid-base catalyst due to its high mech- pressure, S the cross-sectional area of the column, y the A,G anical strength coupled with the high absorption per - bulk-gas mole fraction of component A (CO ), Z the differ- formance. This makes it the most stable for industrial ential height and y the bulk-gas mole fraction of compo- amine-absorption application [24]. nent A in equilibrium with the liquid-gas concentration of that component. 3.2 Physical solubility results Similarly, a material balance on the desorption column can be performed to yield the overall liquid side 3.2.1 Effect of temperature on the physical solubility of mass-transfer coefficient, K a , as described in detail L v 2M BEA- 2M AMP elsewhere [14]: The physical solubility of CO in the bi-blend solvent of A,L,2 BEA-AMP was evaluated and compared to the conven- dx AL tional solvent of 5M MEA. The results of the Henry’s Law K a = x (15) L v ρ Z m (1 − x ) x − x AL Al constant for NO and CO in water were obtained from the A,L,1 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 269 correlations developed by Versteeg and Swaaij [20]. The Å ã ΔS trend from Figures 57 - shows a decrease in the physical ln = 13.364 (intercept) solubility (increase in Henry’s Law constant) of the N O gas as the temperature increases. This can be explained by ΔH = 1552.1 (slope) the increase in the chemical activity as the temperature increases. The dissolution of the NO within the amine with R = 8.314 is an exothermic process, just like that of CO . In exo- K. mol thermic processes, increasing the temperature decreases J 13.364 ΔS= 8.314 × e = 5293238.7 the solubility of the solute. Increasing the temperature K. mol introduces more heat into the system. So, according to ΔH = 8.314 × 1552.1 = 12904.2 Le Chatelier’s Principle, the system will adjust to this ex- mol cess in heat energy by inhibiting the dissolution of the Substituting into Equation 8 gives: N O gas. The physical solubility of the gas was seen to be Å ã −1552.1 small (high Henry’s Law constant), hence the need for the He = 0.64 ∗ 10 exp N2O−BEA/AMP chemical reactivity with the amine. The experiment was validated with 5M MEA. The result showed good agree- ment with the work done by Ma’mun et  al. [17] with an 3.3 Pilot-plant results AAD of 1.5%. Figure 8 shows a logarithmic plot of the Henry’s Law 3.3.1 Solvent effect on mass-transfer performance constant of NO in the 2M BEA-2M AMP solution against The solvent screening experiment was performed to se- the negative reciprocal of the system temperature. From lect the best solvent for the CO -absorption unit using the Figure 8, a correlation was developed for the Henry’s Law lab-scale pilot plant. The solvents screened were: 5M MEA, constant of NO in the 2M BEA-2M AMP solution. The phys- 7M bi-blend of MEA-MDEA (5:2 ratio) and a 4M bi-blend of ical solubility of CO in the 2M BEA-2M AMP solvent was BEA-AMP (2:2 ratio). This work selected the conventional compared to that of 5M MEA, as shown in Figure 9. The 5M MEA solvent to validate its performance from previous result shows a lower Henry’s Law constant at the various work [14] while providing a base case to select other suitable temperatures with 2M BEA-2M AMP. This is due to the low solvents. The 7M MEA/MDEA blend also served as a valid- solvent–solvent interaction of the BEA-AMP solvent re- ation from previous work [14] confirming higher perform- sulting in a higher solubility of the gas even though the ance in terms of faster reaction rate, faster mass transfer, viscosity of the 2M BEA-2M AMP is higher than that of 5M higher solvent capacity and lower heat duty. The selection MEA as reported in the literature [6]. Viscosity plays an im- of the 4M BEA/AMP blend was also based on previous work portant role in determining the physical solubility of a gas. showing an even faster reaction rate, faster mass transfer, The higher the viscosity, the lower the physical solubility. higher solvent capacity and lower heat duty over other The validity of the correlation developed from Figure 8 was solvents as reported by Narku-Tetteh et  al. [4]. The 4M confirmed with the experimental results in a parity plot BEA/AMP was then selected for a pilot-plant solvents test shown in Figure 10. The results show a strong correlation with the 7M MEA/MDEA and base 5M MEA. The average with the experimental work with an AAD of 1.0%. operating times for the different solvents are outlined in The He is fitted to the Van’t Hoff plot as given Table S2 in the online Supplementary Data. The tempera- N2O-BEA/AMP in Equation 8. ture profile for CO -amine absorption using the different From Figure 8, a plot of ln He as against solvents has been reported by Afari et al. [25] showing the N2O−BEA/AMP –1/T gives: highest reactivity for the 2M BEA-2M AMP bi-blend. The concentration profile (Figure 11) of the absorber for the dif- ferent solvent systems also indicates a high CO -removal efficiency for BEA-AMP, with the lowest concentration of the CO at the top of the absorption column. This reaffirms the high reactivity of the BEA-AMP solvent and relates the high reactivity to the high mass-transfer coefficient, as it reduces the free CO concentration in the liquid and cre- 2 2 ates a higher concentration gradient for mass transfer. This also corroborates the results of Narku-Tetteh et al. [4]. 4500 1 The performance can be attributed to the structure of the BEA-AMP solvent. The AMP is a sterically hindered amine that forms a very unstable carbamate from the re- action with CO. The unstable carbamate easily hydrolyses 3.05 3.1 3.15 3.2 3.25 3.3 3.35 to bicarbonate and, in so doing, frees up more amine for 1000/T, K reaction. This is the major reason for the high absorption Fig. 5: Henry’s Law constant of NO in water at different temperatures capacity of the amine given as 1 mol of CO per 1 mol of He in H O, Pa.m /mol N2O 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 270 | Clean Energy, 2019, Vol. 3, No. 4 3.05 3.1 3.15 3.2 3.25 3.3 3.35 1000/T, K Fig. 6: Henry’s Law constant of CO in water at different temperatures this work ref 1 (Ma'mun et al) ref 2 (Tsai et al ) 3.05 3.1 3.15 3.2 3.25 3.3 3.35 1000/T, K Fig. 7: Henry’s Law constant experimental setup validation plot (using 5M MEA) against 1000/T (at different temperatures) 8.6 3 8.55 8.5 8.45 2 8.4 8.35 8.3 8.25 8.2 8.15 8.1 –0.00335 –0.0033 –0.00325 –0.0032 –0.00315 –0.0031 –0.00305 –0.003 –1/T Fig. 8: A plot of the natural log of Henry’s Law constant for N O in 2M BEA-2M AMP against the negative reciprocal of temperature (ln He 2 N2O-BEA/AMP plot against –1/T) ln (He) Heco MEA, Pa.m /mol 2 3 HeCO in H O, Pa.m /mol 2 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 271 He CO2-BEA/AMP He CO2-MEA 300 305 310 315 320 325 Temperature, K Fig. 9: Henry’s Law constant for CO in 2M BEA-2M AMP compared with 5M MEA at different temperatures of the liquid-film resistance and an increase in the con- centration gradient. This is evident from the concentra- tion profile of the absorber in Figure 11, which shows the highest drop for the 2M BEA-2M AMP solvent over the 7M MEA-MDEA and 5M MEA. AAD 1.02% Under absorber performance, from Figures 12 and 13, the overall mass-transfer coefficient validates the improve- ment with the BEA-AMP solvent from a 101 and 239% in- a and K a , respectively, over the base 5M MEA crease in K G V L V He N2O in BEA/AMP solvent. This can be compared with the 7M MEA-MDEA 3500 y=x blend, which recorded a 7 and 8% increase, respectively, Linear (y=x) in K a and K a over the base 5M MEA solvent. Similar re- G V L V sults on the amine cyclic capacity and CO -removal effi- 3000 4000 5000 6000 7000 Predicted ciency for the different solvents were obtained on the pilot He N O -amine, Pa.m /mol plant as reported by Afari et  al. [25]. The desorption per - formance from Figure 14 showed a 139% increase in the Fig. 10: Parity plot for the Henry’s Law constant for N O in 2M BEA-2M desorption efficiency with the BEA-AMP blend, confirming AMP bi-blend solvent the superior performance of AMP in desorption. AMP [4]. AMP is also very reactive and hence produces 3.3.2 Effect of the HZSM-5 catalyst on mass transfer more unstable carbamates that eventually break down to The use of HZSM-5 as a desorption catalyst on post- form bicarbonates [4]. The unstable nature of the carba- combustion pilot-plant studies has been reported by mate also enhances desorption at a higher temperature. Akachuku [13]. Zeolite catalysts are known for their BEA has a long-chain alkyl group reported [4] to be re- unique pore size, which makes them suitable as a shape- sponsible for the high reactivity of the solvent. The alkyl selective catalyst. They have a high surface area as well group present in BEA increases the basicity of the solvent as being crystalline. The results from the overall mass- by inducing the transfer of electrons from the nitrogen ion transfer coefficient in Figures 12–14 show a significant im- to the alkyl group. provement. This can be explained by the HZSM-5 catalytic The MEA-MDEA blend, however, demonstrated better role in desorption in the mechanism shown in Figure 15. performance when compared to the conventional 5M MEA. ion first from its active The improvement in CO desorption is attributed to the The HZSM-5 catalyst donates its H site to the carbamate, which reacts to form free amine, easy breakdown of bicarbonates formed from the MDEA and releases the CO (Steps 1 and 2). The catalyst also do- as well as the role of MDEA in promoting the desorption + 3– nates part of its H protons to the breakdown of HCO as proton transfer, its catalytic role. The higher molarity of well [12]. The catalyst becomes a conjugate base after the MEA-MDEA blend was also a factor for the increased donating its proton, making it capable of accepting an H performance when compared against the 5M MEA, as more proton. The catalyst recovers its H ion from the proton- free amines were available for reaction. The issue with the ated amine, thereby breaking and lowering the energy high molarity was put to the test when comparing the 7M barrier for the release of CO . This results in higher CO MEA-MDEA blend with the 2M BEA-2M AMP. Mass transfer 2 2 desorption and translates to the much lower lean loading is enhanced with the chemical reaction by the reduction Experimental He N O -amine, Pa.m /mol He CO -amine, Pa.m /mol 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 272 | Clean Energy, 2019, Vol. 3, No. 4 0.9 0.8 0.7 0.6 5M MEA blank 0.5 0.4 7M MEA/MDEA blank 0.3 4M BEA/AMP blank 0.2 0.1 5.00 7.00 9.00 11.00 13.00 15.00 CO Conc, % (in gas) Fig. 11: Absorber concentration profiles of the different solvent systems for the non-catalytic (blank) process 0.14 0g HZSM5 with 150g HZSM5 0.12 0.1 0.08 0.06 0.04 0.02 4 M BEA/AMP (2:2) 7 M MEA/MDEA(5:2) 5M MEA Fig. 12: Catalyst (HZSM-5) effect on the overall mass-transfer coefficient (K a ) for absorption as a function of the solvent G V 1.2 0.8 0.6 0.4 0.2 4 M BEA/AMP (2:2) 7 M MEA/MDEA(5:2) 5M MEA 0g HZSM5 150g HZSM5 Fig. 13: Catalyst (HZSM-5) effect on the overall mass-transfer coefficient (K a ) for desorption as a function of the solvent L V shown in Table 2. The lean loading confirms the benefit of leading to a corresponding increase in the absorption per - HZSM-5 in improving the desorption rate. This explains formance with the addition of the HZSM-5 catalyst in the the increase in the desorption performance (Figure 14) desorber. column Height, m K a 1/hr L V K a , kmol/m .kPa.hr G v Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 273 45.00% 40.00% 35.00% 0g HZSM-5 150g HZSM-5 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% 4 M BEA/AMP (2:2) 7 M MEA/MDEA(5:2) 5M MEA Fig. 14: Catalyst (HZSM-5) effect on CO -desorption efficiency as a function of solvent Table 2: The desorption column lean amine loading (mol of Step 1: Deprotonation of HZSM-5 CO /mol of amine) – – – + HZSM – 5 + HCO /AMPCOO ⇔ AMPH COO + ZSM – 5 Solvent Non-catalytic HZSM-5 in desorber Step 2: Carbamate breakdown 5M MEA 0.42 0.41 7M MEA-MDEA 0.35 0.32 AMPH COO ⇔ AMP + CO 2M BEA-2M AMP 0.33 0.30 Step 3: Recovery of proton by ZSM-5 AMPH + ZSM – 5 ⇔ AMP + HZSM –5 catalyst. As explained earlier, the two reactions involved in the CO reaction with amine includes the electron-transfer Fig. 15: HZSM-5 AMP deprotonation mechanism step to form zwitterion (rate-determining step) and car - bamate formation from zwitterion deprotonation, which The catalytic contribution of HZSM-5 in improving the occurs spontaneously [8, 9]. As such, a Lewis-base catalyst overall mass-transfer coefficient for the different solvents capable of donating electrons is desired. was evaluated as shown in Figs 12 and 13. A  38.7% Ka The original role of the amine is to bond with CO and G v and 23.6% K a increase with the HZSM-5 catalyst on 2M break the carbon and oxygen double bond to form the L v BEA-2M AMP blend, 32.2% K a and 57.6% K a increase on zwitterion molecule. The catalyst therefore initiates the G v L v 5M MEA-2M MDEA blend, and 22.0% K a and 36.5% K a carbon double-bond breakage (an energy-requiring pro- G v L v increase on 5M MEA were obtained. Any percentage in- cess) leaving the amine with the single role of donating crease in the mass-transfer coefficient implies a reduction its lone-pair electrons. The K promoter destabilizes the in the size of the column by about the same margin. Also, bond between Mg and O, thereby increasing the amount 2– the increased rate of mass transfer as the concentration of O Lewis-basic sites. Also, magnesium oxide is an ac- gradient across the film increases results in a much lower tive catalyst for the hydrogenation of 1,3-butadiene. It has 2+ solvent loading. The presence of the HZSM-5 catalyst re- been reported that Mg has acid sites that are responsible sults in a lower solvent loading for the same heat input to for the reduction in the catalytic-base performance of the the desorber/reboiler. This is significant, since the energy K/MgO [26]. Thus, another role of K is to poison these acid cost accounts for more than 70% of the total CO -capture sites. The increase in the rate of reaction also results in plant cost [8, 9]. a higher driving force for mass transfer as the concentra- tion of free CO in the liquid reduces. All these result in a faster rate of mass transfer. The electron-donating ability 3.3.3 Effect of the presence of K/MgO and HZSM-5 of the K/MgO catalyst also results in a higher rich loading, catalyst on mass transfer as more CO gets absorbed into the amine. In addition, the The selected solid-base catalyst K/MgO was investigated in cyclic capacity also increases. its application in the absorption column. The desorption The temperature profile and concentration profile for column was loaded with 150 g HZSM-5 while the effect of the application of both catalysts on the 2M BEA-2M AMP introducing 150 g of K/MgO in the absorption column was were reported by Narku-Tetteh et  al. [27] with results evaluated. The results were evaluated in terms of the ab- demonstrating the progression of improved CO absorp- sorption efficiency, amine cyclic capacity and overall mass- tion when both catalysts are placed in the respective col- transfer coefficients, which showed improvement with the umns. The catalyst weight used was 150 g of HZSM-5 based use of a solid-base catalyst in conjunction with a solid-acid CO desorp. eff 2 Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 274 | Clean Energy, 2019, Vol. 3, No. 4 on previous studies [12, 14]. Figure 16 shows the effect of K/ forming a bond between the catalyst and CO (Step 1). The MgO catalyst on rich amine loading. Results from Figure 16 bond formed violates the carbon octet rule, making it un- confirm improvements in CO absorption as well as an in- stable and causing the breakdown of one of the double crease in amine loading resulting from the K/MgO catalyst bonds between carbon and oxygen (C=O). This puts a par - influencing the absorption of CO . tial charge on the oxygen (from the C=O) and makes the For the application of both catalysts on the 2M BEA-2M CO molecule more reactive to the amine (Step 2). The AMP blend, the overall mass-transfer coefficient (Figures nitrogen of the amine now donates its lone electrons to 17 and 18), 38.7% K a and 23.6% K a increases were the catalyst–CO bond and eventually breaks the bond be- G v L v 2 obtained with HZSM-5 catalyst alone, while 95% K a and tween the catalyst and CO . The newly formed compound G v 2 45% K a increases were obtained with both K/MgO and between the amine and CO represents the zwitterion L v 2 HZSM-5, as compared with the base case of no catalyst in molecule (Step 3). both columns. 3.3.5 Effect of K/MgO catalyst weight on mass transfer 3.3.4 Proposed K/MgO catalytic-reaction mechanism Results from Figures 20 and 21 show an increase in the K/MgO catalyst has been reported as having Lewis-base weight of K/MgO, resulting in an increase in the mass- sites [26]. The proposed mechanism in which the K/MgO transfer performance. This can be attributed to the provides electrons is as a result of bond destabilizing be- increase in the reaction rate as the catalyst weight in- tween the magnesium and oxygen bond (Mg–O), as shown creases. The increase in the catalyst weight results in an in Figure 19. This makes electrons available for bonding increase in the total active catalyst sites available for re- with CO . K is more electropositive than Mg, hence it de- action. The number of active sites and total surface area stabilizes and weakens the bond between Mg and O. The of the active sites increase with the increased catalyst 2– O from the weakened bond donates electrons to CO , weight. This, therefore, results in an increase in the CO 0.6 0.58 0.56 0.54 0.52 0.5 0.48 0.46 0g K/MgO 150g K/MgO Fig. 16: Effect of K/MgO catalyst on rich amine loading 0.18 Solvent 0.16 Solvent+HZSM5 0.14 Solvent+HZSM5+K/MgO 0.12 0.10 0.08 0.06 0.04 0.02 0.00 Fig. 17: Effect of catalyst configuration on Ka of absorber G v Rich loading, mol CO /mol amine KGav, kmol/m .kPa.hr Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 275 1.40 Solvent 1.20 Solvent+HZSM5 Solvent+HZSM5+K/MgO 1.00 0.80 0.60 0.40 0.20 0.00 Fig. 18: Effect of catalyst configuration on Ka of desorber L v Step 1: Oxygen-containing catalyst donates a lone-pair electrons to CO +2 –2 Mg O Mg O OO C OO C Step 2: C = O bond breaking and electron transfer onto O Mg O Mg O O C O OO C Step 3: Nitrogen now donates its electrons to the carbon which breaks and forms N-C bond of zwitterion. Mg O C O = +2 –2 + Mg O NH O C O NH HO HO (Zwitterion) Fig. 19: Proposed K/MgO catalytic mechanism 1.2 0.2 1.15 0.15 1.1 0.1 1.05 0.05 0 50 100 150 200 0.95 K/MgO Catalyst weight, g 0.9 Fig. 20: Effect of K/MgO catalyst weight on K a of absorber 0 50 100 150 200 G v K/MgO weight, g and amine reactivity on the catalyst active site, as re- Fig. 21: Effect of K/MgO catalyst weight on K a of desorber L v ported elsewhere [24]. The increase in the reaction rate creates a higher con- decrease in the liquid-film resistance, seen from Equations centration gradient as free CO in the liquid is consumed at 10 and 11. The combined effect of an increase in the reac- a faster rate. This increased concentration gradient serves tion rate resulting in a decrease in the liquid-film transfer as the driving force for mass transfer and is directly propor - resistance, an increase in the interfacial area and an in- tional to the rate of transfer. Most importantly, the increase crease in the concentration gradient accounts for the in- in the reactivity results in an increase in the enhancement creased performance with the increase in catalyst weight. factor. The enhancement factor, which is a ratio of the ab- Figures 20 and 21 show the positive K/MgO catalyst weight sorption rate of CO due to reaction to the physical absorp- performance against Ka and K a . However, it needs to be tion of CO by the same solvent and conditions, results in a G V L V K a , kmol/m .kPa.hr G v KLav, 1/hr K a , 1/hr L V Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 276 | Clean Energy, 2019, Vol. 3, No. 4 pointed out that, above 150  g of K/MgO, the performance The mass flux of CO across the gas-liquid inter - does not change much across all results. kmol face, m s kPa. m He Henry’s law constant, kmol E enhancement factor 3.4 Industrial considerations y * mole fraction of CO in the gas phase in equilib- A 2 Coker [28] assessed the reduction in the capital expenses rium wh that of CO in the bulk liquid. mol/mol (CAPEX) as well as the energy required for operational ex- G molar inert flow rate (total gas without CO ) 1 2 S cross-sectional area of the column. penses (OPEX). Results showed a significant reduction in L molar inert liquid flowrate, . both the CAPEX and OPEX, which justified the use of cata- 1 ρ molar density of the solution, mol/m lysts in both the absorber and desorber. However, other m x bulk liquid mole fraction of component A, mol/ AL factors including pressure drop and liquid viscosity may mol determine the need for blower and higher pump power, x bulk liquid mole fraction of component A in equi- respectively. It is only after these studies are conducted librium with the bulk gas. mol/mol that a conclusive justification can be made. Therefore, fur - [A ] equilibrium solubility of CO in the amine, mol/m ther comprehensive studies on catalytic roles that include D physical diffusivity of CO in amine, m /s A 2 power consumption, pressure drop, how the absorber cata- k reaction rate constant between the CO and mn lyst reacts to species like SO , NO in flue gas and the poten- amine x x tial cost of replacing the catalyst if it becomes poisoned are m- order of reaction with respect to CO n order of reaction with respect to the amine needed to access the economic feasibility of this process. B concentration of unloaded amine, mol/m 4 Conclusions Acknowledgements This work has demonstrated in great measure the effect The financial support provided by the Natural Science and of a solvent, its chemistry and its catalytic effect on the Engineering Research Council of Canada (NSERC), the Canada mechanism and rate of mass transfer for CO absorp- Foundation for Innovation (CFI) and the Clean Energy Technologies tion and desorption. There are big increases in the mass- Research Institute (CETRI), University of Regina is gratefully transfer coefficients for CO absorption and desorption acknowledged. with the introduction of a catalyst to the capture process. This manifests in the significant increase in mass-transfer Supplementary data rates with the catalyst. The solvent BEA-AMP is seen to outperform the other solvents in mass transfer. The ex- Supplementary data is available at Clean Energy online. tensive study of the effect of catalyst weight shows an increase in the CO-absorption rate or mass-transfer per - References formance with an increase in the catalyst weight. This im- [1] Rao AB, Rubin ES. A technical, economic, and environmental plies that employing a catalyst in the CO -capture process assessment of amine-based CO capture technology for will introduce a significant increase in the mass-transfer power plant greenhouse gas control. Envir on Sci Technol 2002; 36:4467–75. performance, which will greatly reduce the heights of the [2] Houghton JT, Ding Y, Griggs DJ, et al. Climate change 2001. In: absorption and desorption columns. The catalyst introduc- Houghton JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, tion replaces part of the structured packing, and therefore Dai X, Maskell K, Johnson CA (eds). The Scientific Basis. New reduces the cost of the structured packing. York: University Press, 2001, 881. [3] Luis P. Use of monoethanolamine (MEA) for CO capture in a global scenario. Consequences and Alternatives 2016; 380:93–9. Nomenclature [4] Narku-Tetteh J, Muchanb P, Saiwanb C, et al. Selection of com- ponents for formulation of amine blends for post combustion K a Overall Volumetric mass transfer coefficient of gas G v CO capture based on the side chain structure of primary, sec- kmol side, ondary and tertiary amines. Chem Eng Sci 2017; 170:542–60. m . kPa K a Overall Volumetric mass transfer coefficient of li- [5] Idem  R, Wilson  M, Tontiwachwuthikul  P, et  al. Pilot plant L v studies of the CO capture performance of aqueous MEA and quid side, . hr mixed MEA/MDEA solvents at the University of Regina CO R universal gas constant, kJ/kmol.K capture technology development plant and the boundary AAD Absolute average deviation, % dam CO capture demonstration plant. Ind Eng Chem Res 2006; MEA Monoethanolamine 2 45:2414–20. MDEA Methyldiethanolamine [6] Danckwerts P. The reaction of CO with ethanolamines. Chem BEA Butyl Monoethanol amine Eng Sci 1979; 34:443–6. AMP 2- amino 2 methyl 1 propanol [7] Caplow M. Kinetics of carbamate formation and breakdown. J in F flowrate of CO in the feed gas, mol/s. CO 2 2 Am Chem Soc 1968; 90:6795–803. out F flowrate of CO in the exit gas, mol/s. CO2 2 [8] Shi H, Idem R, Naami A, et al. Catalytic solvent regeneration y mole fraction of CO in gas bulk, mol/mol A,G using hot water during amine based CO capture process. y mole fraction at gas interphase. mol/mol A,i Energy Procedia 2014; 63:266–72. Downloaded from https://academic.oup.com/ce/article-abstract/3/4/263/5550982 by DeepDyve user on 10 December 2019 Coker et al. | 277 [9] Shi  H, Naami  A, Idem  R, et  al. Catalytic and non-catalytic [19] Srisang W, Pouryousefi Y, Osei PA, et al. Evaluation of the heat solvent regeneration during absorption-based CO capture duty of catalyst-aided amine-based post combustion CO cap- 2 2 with single and blended reactive amine solvents. Int J Greenh ture. Chem Eng Sci 2017; 170:48–57. Gas Con 2014; 26:39–50. [20] Versteeg GF, van Swaaij WPM. Solubility and diffusivity of acid [10] Liu B, Luo X, Liang Z, et al. The development of kinetics model gases (CO , N O) in aqueous alkanolamine solutions.J Chem 2 2 for CO absorption into tertiary amines containing carbonic Eng Data 1988; 33:29–34. anhydrase. AIChE J 2017; 63:4933–43. [21] Astarita  G, Savage  DW. Gas absorption and desorption with [11] Gladis  A, Gundersen  MT, Fosbøl  PL, et  al. Influence of tem- reversible instantaneous chemical reaction. Chem Eng Sci perature and solvent concentration on the kinetics of the en- 1980; 35:1755–64. zyme carbonic anhydrase in carbon capture technology. Chem [22] Levenspiel O. Chemical Reaction Engineering, 3rd edn. New York: Eng J 2017; 309:772–86. Wiley, 1999. [12] Idem R, Shi H, Gelowitz D, et al. Catalytic method and apparatus [23] Naami A, Edali M, Seema T, et al. Mass transfer performance for separating a gaseous component from an incoming gas stream. of CO absorption into aqueous solutions of 4-diethylamino-2- United States Patent: US 9,586,175 B2. March 7, 2017. 2017. butanol, monoethanolamine, and N-methyldiethanolamine. [13] Akachuku A. Kinetic study of the catalytic desorption of carbon di- Ind Eng Chem Res 2012; 51:6470–9. oxide (CO ) from CO -loaded monoethanolamine (MEA) and blended [24] Coker  J, Afari  DA, Narku-Tetteh  J, et  al. Synthesis and appli- 2 2 monoethanolamine-methyldiethanolamine (MEA-MDEA) during cation of solid base catalyst in a post combustion capture of post-combustion CO capture from industrial flue gases. MASc. CO - a mass transfer study. In: 25th Canadian Symposium on 2 2 Thesis. University of Regina, Canada, 2016. Catalysis Saskatoon, SK, 8–11 May, 2018. [14] Osei PA. Mass transfer studies on catalyst-aided CO desorption in [25] Afari  DA, Coker  J, Narku-Tetteh  J, et  al. Comparative kinetic a post-combustion CO capture plant. MASc. Thesis. University of studies of solid absorber catalyst (K/MgO) and solid desorber Regina, Canada, 2016. catalyst (HZSM-5)-aided CO absorption and desorption from [15] Zhang X, Zhang R, Liu H, et al. Evaluating CO desorption per - aqueous solutions of MEA and blended solutions of BEA-AMP formance in CO -loaded aqueous Tri-solvent blend amines and MEA-MDEA. Ind Eng Chem Res 2018; 57:15824–39. with and without solid acid catalysts. Appl Energy 2018; [26] Ono Y, Hattori H. Solid Base Catalysis. New York: Springer- 218:417–29. Verlag Berlin Heidelberg and Tokyo Institute of Technology [16] Sema T. Kinetics of carbon dioxide absorption into aqueous solu- Press, 2011. tions of 4-(diethylamino)-2-butanol and blended monoethanolamine [27] Narku-Tetteh J, Afari DA, Coker J, et al. Evaluation of the roles and 4-(diethylamino)-2-butanol. PhD. Thesis. University of Regina, of absorber and desorber catalysts in the heat duty and heat of Canada, 2011. CO desorption from butylethanolamine 2-amino-2-methyl- [17] Ma’mun  S, Svendsen  H. Solubility of NO in aqueous 1-propanol and monoethanolamine methyldiethanolamine monoethanolamine and 2-(2-aminoethyl-amino)ethanol so- solvent blends in a bench-scale CO capture pilot plant. Energy lutions from 298 to 343 K. Energy Procedia 2014; 1:837–43. Fuels 2018; 32:9711–26. [18] Tsai  T, Ko  J, Wang  H, et  al. Solubility of nitrous oxide in [28] Coker J. Mass transfer and emission studies on a catalyst-aided CO alkanolamine aqueous solutions.J Chem Eng Data 2011; absorption and desorption in a post combustion CO capture plant. 45:341–7. MASc. Thesis. University of Regina, Canada, 2018.

Journal

Clean EnergyOxford University Press

Published: Dec 6, 2019

References