Evaluation of Energy-Saving Combo of MEA-EAE-AMP Tri-Solvent with Absorber and Desorber Catalysts in a Hot Oil-Based Bench-Scale Pilot Plant

Abstract

To mitigate the effects of climate change, novel carbon capture technologies need to be developed. Since 2020, a new solution has been to adopt an energy-efficient combination of “amine blend + heterogeneous catalysts” in large CCUS demonstration plants. This study adopted the specific tri-solvent MEA-EAE-AMP and solid catalysts CaSO₄, HND-580, and HND-8 in a novel bench-scale pilot plant with hot oil as the heat source. Three key parameters were investigated—absorption efficiency (AE), cyclic capacity (CC), and heat duty (HD)—to analyze the technology under a steady state. The results indicated that the solid acid significantly reduced α_lean and the solid base increased αrich, while the CC was increased and HD was reduced to its minimum at 2.47 GJ/tCO₂ and at optimized doses of the catalysts, 40 g CaSO₄ and 100 HND-580. These results verified another energy-efficient solution that could be further scaled up into an industrial amine scrubbing pilot plant.

1. Introduction

In order to reduce industrial CO₂ emissions and mitigate climate change, the Paris Agreement aimed at keeping global temperature rise to below 2 °C above pre-industrial levels before 2050 [1]. Among these greenhouse gasses, CO₂ emission control has attracted particular attention due to its increasing concentration in the atmosphere, primarily due to the combustion of fossil fuels in various industries [1,2,3,4,5]. Based on the current scale and growth of energy consumption, the supply capacity of non-carbon energy can hardly meet energy demands [4]. Therefore, the combustion of fossil fuel has maintained its major position in the energy structure for decades [5]. Because of this, carbon capture, utilization, and storage (CCUS) technology is regarded as an energetically effective technical approach to reducing CO₂ emissions. Among the various CCUS technologies, the post-combustion capture technology (PCC) [6] used on flue gas and based on an amine-based scrubbing process has exhibited great potential in the field of industrial applications due to its technological maturity, strong CO₂ processing capacity, high processing efficiency, and strong adaptability to coal-fired or natural gas-fired power plants [7].

PCC technology needs to be implemented into industries such as coal-fired power plants, cement plants, steel factories, etc. [6]. Besides the significant energy cost issue of heat duty reduction, another large gap in its implementation has been scaling up, which builds the bridge between fundamental research studies of CO₂ absorption and amine regeneration and the utilization of a large-scale CCUS demonstration pilot plant (to capture at least 1000 tCO₂/year) to model a real-world problem [6,8]. Most academic publications have reported experimental data from a batch or semi-batch process, while only a few studies have used a small bench-scale pilot plant to conduct steady-state analysis as a model of a CCUS demonstration plant [6,9,10,11,12]. Scheme 1 demonstrates the bridge between fundamental research and industrial application, which is bench-scale pilot plant studies. The published energy-efficient combinations that enable catalytic CO₂ desorption need to be adopted in pilot plant studies to then produce formula (the energy efficient combo + optimized operation conditions + optimized HD and CC) (Figure S1 in Supplementary Materials) [13]. Therefore, since 2017, a branch of research has focused on first designing and developing bench-scale pilot plants and then adopted published energy-efficient combinations (amine blends + heterogeneous catalysts) to evaluate AE, CC, and HD parameters, and especially heat duty (energy cost), to verify its potential for industrial application [9,14,15,16,17,18,19,20]. These studies cannot be ignored and have paved the way for the industrial application of CCUS demonstration plants in industry [11].

Most pilot plant studies adopt vapor steam as their heat source to mimic conventional CCUS plants, but some research groups have designed novel pilot plants with other heat sources since as early as 2017. Idem et al. invented a hot water-based bench scale process and published several studies with MEA, MEA-MDEA, MEA-DEAB, and BEA-AMP with solid acid–base catalysts [9,14,15,16,17,18]. In 2023, one of the authors (Idem’s Ph.D. student) developed a similar process to conduct studies on MEA-BEA-AMP with catalysts [19]. After 2024, the author’s group invented a novel bench-scale process with hot silicon oil as a heat source to analyze MEA-EAE-DEEA with various catalysts [20]. Such work was illustrative and possessed huge potential in the industrial application of CCUS plants constructed in desert areas such as the United Arab Emirates [20].

Amine blends have been studied for decades in order to develop energy-efficient combinations with reduced heat duty [21]. A small branch of the tri-solvents of amine A + B + C had drawn research interest since 2016 [19,22,23,24,25,26,27,28,29,30,31,32,33], and a mini review was published in 2022 [31]. The tri-solvent of MEA-EAE-AMP had been intensively studied by the authors’ group since 2022, as it enables non-catalytic and catalytic absorption–desorption [29,31]. Finally, 0.2 + 2 + 2 mol/L was used in this study, since its “coordinative effect” is consistent and optimized for both absorption and desorption [29].

Energy-efficient combinations of MEA-EAE-AMP + solid acid catalyst were studied and published in 2022 [31], using five types catalysts: blended γ-Al₂O₃/H-ZSM-5, Hβ, H-mordenite, HND-8, and HND-580. Afterward, HND-8 and HND-580 were selected as desorber catalysts for this study based on their superior performance [31], while CaSO₄ was selected as a new type of solid base catalyst for this study. The chemical properties of CaSO₄ were similar to those of CaCO₃, and CaCO₃ and verified to be effective in boosting CO₂ absorption as a Lewis base if installed into an absorber [19,34].

Therefore, this work conducted a pilot plant analysis (using hot silicon oil as heat source) with solid acid catalysts HND-8 and HND-580 installed in the desorber and the solid base catalyst CaSO₄ installed in the absorber, focusing on specifically 4.2 mol/L of the tri-solvent MEA-EAE-AMP. The CO₂ absorption–desorption performances of this energy-efficient combination were used to systematically evaluate the key parameters of AE, CC, and HD. The solid acid catalysts (HND-8 and HND-580) were investigated alone at first, and then the solid acid–base catalysts (HND-8 and CaSO₄) were studied to find the optimized W_cat with a bigger CC and smaller HD. This analysis verifies the industrial application of this combo in a CCUS demonstration plant.

2. Results and Discussion

This study concerned catalytic CO₂ absorption–desorption in a pilot plant based on key indicators such as CO₂ absorption efficiency (AE), cyclic capacity (CC), and heat duty (HD). The study consists of two parts: (1) A catalytic CO₂ desorption investigation was conducted with solid acid catalysts (HND-8 and HND-580) to verify heat duty reduction. (2) Catalytic absorption–desorption studies were carried out with a solid alkaline catalyst (CaSO₄) and solid acid catalyst (HND-8) to analyze the synergetic effects of the acid–base catalysts.

2.1. Effect of the Desorber Catalysts on Capture Performance

2.1.1. Effect of the Desorber Catalysts on Absorption Efficiency

Figure 1 illustrates the impact of various weights of two solid acid catalysts (HND-8 and HND-580) on the absorption efficiency (AE) when the inlet temperatures of the absorption tower are 25 °C and 30 °C. The addition of solid acid catalysts enhanced the AE of the amine solution compared to the non-catalytic experiment at both inlet temperatures. Figure 1a indicates that the AE increased with the weight of HND-8, with the W_cat reaching a maximum of 150 g. Figure 1b indicates that the addition of HND-580 also led to a similar trend, with a maximum W_cat of 100 g. This phenomenon is reasonable because the catalysts facilitate the desorption of CO₂ in the desorption tower, resulting in a smaller α_lean of the solvent entering the absorption tower. With more free amine molecules available for CO₂ absorption, the AE increased. After setting 150 g of HND-8 and 100 g of HND-580 as the optimized W_cat, the α_lean cannot reach any lower. Thus, the AE was still higher than 90% but smaller than the optimum value. Afterward, the inlet temperature was fixed at 30 °C since the AE was higher than that under the same conditions but at 25 °C.

2.1.2. Effect of the Desorber Catalysts on Cyclic Capacity

Figure 2 illustrates the CC of HND-8 and HND-580 at W_cat of 0–250 g. The cyclic capacity reflects the capability of the tri-solvent to hold up the absorption and desorption of CO₂ during each run. From Figure 2a, the addition of HND-8 increased the CC from 40.0 mmolCO₂/min without catalyst to a range of 42.5–50.0 mmolCO₂/min, reaching an optimum value at a W_cat of 150 g. Figure 2b indicates that the introduction of HND-580 enhanced CC to 43.8–47.5 mmolCO₂/min, with the maximum value at a W_cat of 100 g. Overall, the CC with both catalysts reached an optimum with increased W_cat and decreased a little at excessive masses of the catalysts.

The enhancement in CC was attributed to the addition of solid acid catalysts into the desorption tower. After analyzing the exact value of α_rich and α_lean (Table S1 in Supplementary Materials), it is evident the added Brønsted acids, HND-8, and HND-580 provided H⁺ protons directly, which accelerated the decomposition of carbamates without the amine protonation process. Thus, the desorption rate was enhanced. Additionally, the protons from the catalysts directly reacted with bicarbonate (HCO₃⁻) to produce CO₂. Consequently, compared to non-catalytic operations, greater nCO₂ is produced with the incorporation of catalysts, expediting the desorption process. This resulted in a lower α_lean and increased CC.

The optimized W_cat for the system was mainly based on the αlean from the SI information. The αlean was the lowest at an optimized W_cat in the 0–250 g region. This optimal W_cat was attributed to the strong endothermic reactions of catalytic CO₂ desorption. The α_lean was strongly affected by two factors: T and W_cat. With an excess weight of solid acid, a massive amount of heat was absorbed by the desorption reaction and CO₂ was released, making the temperature profiles below 90 °C or even 85 °C throughout the desorber. Under these desorption conditions, the catalysis was inhibited by the low temperature, and the α_lean increased again. In fundamental studies of catalytic desorption, the desorption rate significantly declined at 70–85 °C [20,31,32].

Table 1. Properties of solid acid catalysts.

Catalyst	BET Surface Area (m2/g)	Average Pore Size (nm)	Acid Strength (mmol/g)	Pore Volume (cm3/g)
HND-8	>20	≥15	24.75	0.2–0.4
HND-580	≥20	≥15	≥4.95	0.2–0.45

compares the characteristics of the two catalysts. Their catalytic activity was influenced mainly by the strength of their acid, along with their porous physical structure and acidic chemical properties [13]. From Table 1, it is clear that both catalysts, HND-8 and HND-580, possessed huge acid strength. The acid strength of HND-8 was 24.75 mmol/g and that of HND-580 was ≥ 4.95 mmol/g. This indicated that both catalysts have a high capacity for proton (H⁺) donation, which was very effective in reaction Equations (11)–(13). Their acidic strength reflected their strong catalysis of CO₂ desorption, as confirmed by a review of larger publications [13]. The details of both catalysts are provided in the Supplementary Materials.

2.1.3. Effect of the Desorber Catalysts on Heat Duty

In Figure 3, the impact of HND-8 on the heat duty (HD) was plotted with W_cat of 0–250 g and absorber inlet temperatures of 25 °C and 30 °C. From Figure 3a, it is evident that the HD decreased with increases in the W_cat of HND-8, reaching its minimum (optimum) value of 2.70 GJ/tCO₂ at 150 g. The HD slightly increased at W_cat of 200 g and 250 g. The same trend is shown in Figure 3b, except that the minimized HD 2.60 GJ/tCO₂ was even smaller at a W_cat of 150g.

Figure 4 plots the HD of HND-580 within W_cat of 0–250 g and at absorber inlet temperatures of 25 °C and 30 °C. From Figure 4a,b, it is evident that the minimum HD achieved at 25 °C, 2.88 GJ/tCO₂, and at 30 °C, 2.74 GJ/tCO₂, was with a W_cat of 100 g. The trend was similar to Figure 3; the HD decreased to an optimum (minimum) value and then increased a little, except the optimized Wcat was 100 g of HND-580 instead of 150 g of HND-8.

This trend shares the same intrinsic reasoning as that in the previous section on CC, where the HD was determined by the CC and the α_rich–α_lean was the key factor, with its exact measurements given in Table S1 in the SI. However, since catalytic CO₂ desorption is a strongly endothermic process, more heat was absorbed with increased W_cat, which then led to an increase in CO₂ desorption rates and simultaneous a decrease in the desorption temperature. If the desorption temperature drops below 90 °C or even 85 °C, the desorption effect will be diminished. At an optimized W_cat for HD, the αlean was at its minimum. Similar effects have been discovered and published elsewhere [20].

Comparing Figure 3 and Figure 4, given the same W_cat added, the HD seen with an inlet temperature of 30 °C was consistently lower than that at 25 °C. The intrinsic reason for this is the basis of the HD calculation in Equation (16). HD is determined by the heat input and CO2 production (Q_input/nCO₂). The heat input is determined by the T_HMin (T inlet the desorber)—T_HMout (T out of the heat exchanger). The T_HMin was about 90 °C for different operations, and the T_HMout was higher when it left the base of the absorber and entered the heat exchanger. Higher temperatures require less energy to heat the amine solution, making the HD relatively lower. Similar trends have also been discovered by the author’s group [19,20]. Therefore, 30 °C was selected for the T_inlet and was adopted in Section 4.2.

Figure 5 plots the minimum HD of HND-8 and HND-580 with optimized W_cat under a T_inle of 30 °C compared with its non-catalytic counterpart. With the aid of solid acid catalysts, the HD decreased. Specifically, the incorporation of the HND-8 catalyst led to the largest reduction in HD to 2.60 GJ/t CO₂, representing a 20.2% decrease from the baseline value of 3.26 GJ/t CO₂ under non-catalytic conditions. The addition of the HND-580 catalyst reduced the HD to, at the lowest, 2.74 GJ/t CO₂, a 16% reduction from the non-catalytic scenario. This result indicated that the introduction of the catalyst contributed to significant energy-saving effects. This is mainly attributed to the catalyst’s ability to provide protons during the desorption process. Solid catalysts not only increased the pathways for the reaction, but also provided a pathway with a lower activation energy, so that more molecules could reach this reduced energy threshold before becoming activated molecules [9,16,18]. This promoted the kinetics of the reaction, improved desorption efficiency, shortened the heating time, and lowered the desorption temperature, thereby reducing the energy consumption of the entire heating process [16].

Although HND-8 exhibited a superior catalytic performance and less heat duty than HND-580, the combination of HND-8 with 4.2 M of MEA-EAE-AMP tri-solvent produced crystalline precipitation in the amine solution after continuous operation for 30 h, after which the crystals clogged the pipeline and stopped the operation. In contrast, HND-580 avoided the problem of crystallization and had a satisfactory catalytic effect. Therefore, HND-580 was adopted in Section 4.2 to explore the synergistic effect of solid acid–base catalysts. As mentioned in Scheme 1 in Section 1, the pilot plant study was to verify the energy-efficient combos suitable for a steady-state process, as a preparation for scale up in a future CCUS demonstration plant. It is of the utmost importance to ensure long-term operations with a suitable catalyst used in a steady state. Therefore, the rest of the study selects “4.2 M MEA-EAE-AMP + HND-580” as the most energy-efficient combo instead of “4.2 M MEA-EAE-AMP + HND-8”.

2.1.4. Effect of the Desorber Catalyst on Temperature Profiles

Figure 6a plots the temperature distribution along the absorption tower with optimized HND-8 and HND-580 compared to its non-catalytic counterpart. From Figure 6a, it is clear that the temperature at the top, at the starting point, was almost the same, 30 °C, for three runs. The lean tri-solvent flowed from the top of the absorber to absorb CO₂ and released a massive amount of reaction heat inside the absorber, resulting in the rise in T. At the bottom of the absorber column, X_in reached its maximum as the mixed gas came, containing massive amounts of free CO₂ molecules as reactants, causing higher reaction rates and a stronger heat release. The accumulated heat of the reaction resulted in a highest T of 47.0 °C with no catalysts. Under the condition of 100 g of HND-580 and 150 g of HND-8, the highest temperatures were close to 49 °C. Since there were no catalysts installed in the absorbers, which were all the same, the effect of the solid acid catalysts in the desorber resulted in a lower α_lean of 0.294 mol/mol compared to the non-catalytic 0.314 mol/mol (SI). There were more free amine molecules in the absorber, such that the overall absorption reaction was enhanced with the extra heat release and higher T.

Figure 6b depicts the temperature distribution curve of the desorption tower. The pre-heated rich solution reached approximately 92.2 °C before entering the top of desorption tower to conduct CO₂ desorption reactions. The higher the temperature in the desorption column, the better the desorption performance is. Since the desorption process is a strong endothermic reaction, it was observed that the temperature decreased with the increase in the column height, eventually stabilizing at 86.1–86.9 °C at the bottom. The lines of catalytic desorption are to the right side of those of that were non-catalytic. The two reached similar values at the top, because initially the desorption was facilitated by a catalyst and the temperature then decreased faster than the non-catalytic case. The T decreased over the height of the tower, with the CO₂ desorption hindered and α_lean failing to decrease any further if the T was below 85 °C [30].

2.2. Effect of the Absorber–Desorber Catalysts on Capture Performance

In Section 4.1, HND-580 was selected for this system due to its ability to avoid the precipitation issue. That section established that 100 g of HND-580 should go in the desorber and installed 0–50 g of the solid alkaline catalyst CaSO₄ into the absorber. These solid base catalysts were adopted to enhance CO₂ absorption as well as the overall process, and the synergetic effects of the solid acid–base catalysts were investigated.

Table 2. Effect of the absorber–desorber catalysts (HND-580 + CaSO₄) (100 g + 0–50 g) on AE, CC, and HD.

Catalysts	AE (%)	CC (mmol CO2/min)	HD (GJ/tonne CO2)
100 g HND-580	94.5	47.5	2.74
100 g HND-580 + 10 g CaSO4	94.9	47.5	2.68
100 g HND-580 + 20 g CaSO4	95.1	50.0	2.60
100 g HND-580 + 30 g CaSO4	95.8	51.3	2.51
100 g HND-580 + 40 g CaSO4	95.9	52.5	2.47
100 g HND-580 + 50 g CaSO4	95.4	51.3	2.49

The bold was the optimized condition within the system.

demonstrates the impact of various W_cat of CaSO₄, from 0 to 50 g, in the absorption tower on its overall CO₂ capture performance, along with a fixed 100 g W_cat of HND-580 in the desorption tower. The three key indicators—adsorption efficiency (AE), cyclic capacity (CC), and heat duty (HD)—were developed and are categorized in Table 2. The optimized W_cat was determined mainly by the CC and HD. Under the optimal W_cat of 40 g CaSO₄ + 100 g HND-580, the AE increased to 95.9%, the CC rose to 52.5 mmol CO₂/min, and the HD reduced to 2.47 GJ/tCO₂. All these three parameters were the optimum among the rest seen under optimized W_cat.

2.2.1. Effect of the Absorber–Desorber Catalysts on Absorption Efficiency

Figure 7 illustrates the impact of the solid acid–base catalysts on the AE. The non-catalytic process just met the industrial requirement of 90%, and HND-580 increased this to 94.5%, while synergetic catalysis increased this to 95.9%. The solid base catalyst enhanced the AE to its maximum, since it acted as an electron donor, efficiently supplying electrons to accelerate the formation of zwitterions.

2.2.2. Effect of the Absorber–Desorber Catalysts on Cyclic Capacity

Figure 8 illustrates the effect of both acid–base catalysts on CC. Compared to non-catalytic tests, HND-580 only enhanced the CC by 18.8%, while CaSO₄ + HND-580, the optimized W_cat, enhanced CC by 31.3%. The solid base CaSO₄ acted as a Lewis base, which facilitated CO₂ absorption by absorbing CO₂ molecules to its solid surface, thereby accelerating the CO₂ absorption process. Since the CC was determined by an operational region (α_rich–α_lean), its increase depended on various types of catalysts on both the absorption and desorption sides. From the α_rich and α_lean of three runs in Table S3 (SI), it is clear that the solid acid catalyst decreased the α_lean compared to the non-catalytic process, while the solid base catalyst increased the α_rich further and kept α_lean at low values. The synergy of these effects resulted in the maximum CC.

2.2.3. Effect of the Absorber–Desorber Catalysts on Heat Duty

Figure 9 illustrates the effect of both catalysts on the heat duty (HD). The heat duty (HD) was 3.26 GJ/t CO₂ for the non-catalytic tests, and it reduced to 2.74 GJ/t CO₂ with HND-580 at its optimized W_cat. Further, it reduced to 2.47 GJ/t CO₂ with both CaSO₄ and HND-580 at their optimized W_cat. The solid acid alone resulted in a 16.0% reduction in the HD, and both catalysts caused a 24.2% reduction compared to the non-catalytic operation. The trend in the HD reduction was exactly the opposite of that of the CC increase since HD was mainly affected by nCO₂ under similar Q_input, which was proportional to CC. With solid acid–base added, the (α_rich–α_lean) was increased to a maximum, so that CO₂ production was the largest within the three runs and thus the HD was at its minimum. The CaSO₄ packed in the absorber sped up CO₂ absorption and increased α_rich, while HND-580 facilitated the bond cleavage of carbamate by providing protons, simplifying CO₂ desorption and leading to a much lower energy cost [19].

Finally, in

Table 3. The key parameters of MEA-EAE-AMP and catalysts from other studies.

Absorbent Amin Blend	Heating Source	CA (mol/L)	Catalysts Acid–Base	Cyclic Capacity (mmol CO2/min)	Heat Duty (GJ/tCO2)	Group	Ref.
MEA	H2O	5	H-ZSM-5/γ-Al2O3 (2:1)	20.65	11.37	Idem	[18]
BEA + AMP	H2O	2 + 2	HZSM-5 + K/MgO	38.4–40.8	5.56	Idem	[14,15,17]
MEA + BEA + AMP	H2O	0.3 + 2 + 2	H-ZSM-5/γ-Al2O3 (2:1) + CaMg(CO3)2	58.1	2.40	Shi	[19]
MEA + EAE + DEEA	oil	0.5 + 2 + 2	HND-8 + CaMg(CO3)2	55.0	2.46	Shi	[20]
MEA + EAE + AMP	oil	0.2 + 2 + 2	HND-580 + CaSO4	52.5	2.47	Shi	This study

, this study compared the results of CC and HD with those from other studies considering bench-scale pilot plants. The HD was the key parameter used to compare various runs at a consistent level. The value of CC was not comparable due to various operation conditions such as the size of the absorber–desorber, F_G, F_L, and the heat source. Compared with other studies, the tri-solvent in this study contained a smaller HD than single MEA and a bi-blend of BEA + AMP due to the advantage of amine blends and solid acid catalysts. Within our own group, the HDs of our energy-efficient combo (tri-solvent with acid–base catalysts) were close to each other. The optimized HD needed to be found under optimized conditions that consider a combination of F_G, F_L, X_CO2, T, W_cat of abs, W_cat of des, etc., which await further study.

2.2.4. Effect of the Absorber–Desorber Catalysts on Temperature Profiles

Figure 10 plots temperature distributions along absorber and desorber over three runs. In Figure 10a, temperature increases can be observed, with the following order: acid–base catalyst > base catalyst > non-catalyst. The reason for this was similar to that in the previous section, Section 2.1.4. The solid acid catalyst reduced α_lean and the solid base provided a Lewis base on the active sites to improve the CO₂ absorption process. The CaSO₄ installed in the absorption tower increased the rate and extent of CO₂ absorption. Since the CO₂ absorption reactions were exothermic, extra heat was released to increase the temperature profiles in the absorption tower.

From Figure 10b, it is clear that the initial T of each reaction was comparable and the endothermic desorption reactions absorbed the heat in the column and decreased the T gradually from 92 to 86~87 °C. Since catalytic desorption reactions are intensely endothermic processes, the energy cost of the desorption with solid acid catalysts was less than that of the non-catalytic tests. Then, the T of the catalytic tests was higher (on the right side) than the T of the non-catalytic tests. Later on, as the more CO₂ was released from the desorber with the large amount of heat consumed, the T decreased much faster compared to the non-catalytic tests and reached a value lower (on the left side) than the non-catalytic runs.

Afterward, our energy-efficient combination (tri-solvent of CaSO₄ + HND-580) was tested over five runs within 5 days. Each run took 8 h to reach a steady state and record data. The results for CC and HD were comparable, so the sustainability of the method was guaranteed. The figures of the results are provided in the Supplementary Information.

3. Theory

3.1. Main Reaction Schemes of CO₂ Absorption–Desorption Within Tri-Solvent

The main reactions for MEA-EAE-AMP’s absorption and desorption of CO₂ are shown in Equations (1)–(10) [29]. These reaction schemes have been published repeatedly [29,31,35]. MEA and EAE act as primary/secondary amines and AMP acts as a sterically hindered amine and similar to tertiary amine with unstable carbamate.

• Carbamate formation:

$$C{O}_2+R_1{R}_{2}NH\leftrightarrow R_1{R}_{2}N{H}^{+}CO{O}^{-}$$

(1)

$$R_1{R}_{2}N{H}^{+}CO{O}^{-}+R_1{R}_{2}NH\leftrightarrow R_1{R}_{2}NCO{O}^{-}+R_1{R}_{2}N{H}_2^{+}$$

(2)

$$R_1{R}_{2}NCO{O}^{-}+H_{2}O\leftrightarrow R_1{R}_{2}NH+HC{O}_3^{-}$$

(3)

• Carbamates breakdown:

$$R_1{R}_{2}NCO{O}^{-}+H_3{O}^{+}\leftrightarrow R_1{R}_{2}N{H}^{+}CO{O}^{-}+H_{2}O$$

(4)

$$R_1{R}_{2}N{H}^{+}CO{O}^{-}\leftrightarrow R_1{R}_{2}NH+C{O}_2$$

(5)

• AmineH⁺ deprotonation:

$$R_1{R}_{2}N{H}_2^{+}+H_{2}O\leftrightarrow R_1{R}_{2}NH+H_3{O}^{+}$$

(6)

$$R_1{R}_{2}N{H}_2^{+}+HC{O}_3^{-}\leftrightarrow R_1{R}_{2}NH+H_{2}C{O}_3$$

(7)

• Bicarbonate decomposition:

$$HC{O}_3^{-}+H_3{O}^{+}\leftrightarrow C{O}_2+2H_{2}O$$

(8)

$$HC{O}_3^{-}\leftrightarrow C{O}_2+O{H}^{-}$$

(9)

$$O{H}^{-}+H_3{O}^{+}\leftrightarrow 2H_{2}O$$

(10)

During the CO₂ desorption process, the decomposition of carbamates and the deprotonation of protonated amines are both strongly endothermic reactions [36]. The intrinsic reason for this is mainly due to the higher basicity of amines compared to H₂O; the transfer of protons from AmineH⁺ to H₂O can hardly occur automatically in Reaction (6). This effect resulted in a lack of available protons to participate in Reaction (4). Meanwhile, the breakdown of carbarmate resulted in a significant energy demand in the amine regeneration process [17].

With the aid of solid acid catalysts (Brønsted acid), the protons can be provided directly from the catalyst surface and thus facilitate CO₂ desorption. HND-8 and HND-580 are Brønsted acid catalysts, which provide protons to attach onto the N atom of carbamates (MEA and EAE) to facilitate the N-C bond stretch and bond cleavage of N-COO⁻. After bond breaking, the carbamate was decomposed and the CO₂ was desorbed. The catalytic reaction schemes are listed in Equations (11)–(13) [37] and have similar mechanisms [13]:

RR′NCOO⁻ + H-Cat ↔ RR′N-COOH + Cat⁻ ↔ RR′NH⁺COO⁻ ↔ RR′NH₂ +
CO₂ + Cat⁻

(11)

RR′NH₂⁺ + Cat⁻ ↔ RR′NH + Cat – H

(12)

RR′NH₂⁺ + HCO₃⁻ ↔ RR′NH + H₂CO₃ ↔ RR′NH + CO₂ + H₂O

(13)

3.2. Calculations of Absorption Efficiency, Cyclic Capacity, and Heat Duty

In the pilot plant process, an online analysis was carried out through the monitoring system. Several key variables were recorded or calculated: the liquid flow rate ($F_L$), inlet and outlet gas flow rates ($F_G$), CO₂ concentrations of the mixed gas inlets and outlets ($X_{in}$ and $X_{out}$), temperate profile ($T$) along the absorber and desorber, and CO₂ loadings of the rich/lean amine solution (${\alpha }_{rich}/{\alpha }_{lean}$). The loadings were determined using the Chittick titration technique [38].

The absorption efficiency (AE), cyclic capacity (CC), and heat duty (HD), as key parameters, were calculated according to Equations (14)–(16) in order to evaluate the performance of the operation at a consistent level [9,18,19,20]. The extra energy requirement of condensing water vapor was negligible compared to the HD analysis, so it was not calculated.

$$\mathrm{Absorption} \mathrm{Efficiency} (\%)={\displaystyle \frac{F_{{\mathrm{G}}_1}\times X_{\mathrm{in}}-F_{{\mathrm{G}}_2}\times X_{\mathrm{out}}}{F_{\mathrm{G}1}\times X_{\mathrm{in}}}}$$

(14)

$$\mathrm{Cyclic} \mathrm{Capacity}=C_A\times \left({\alpha }_{rich}-{\alpha }_{lean}\right)\times F_L$$

(15)

$$\mathrm{Heat} \mathrm{Duty}={\displaystyle \frac{m_{HM}\times C_{pHM}\times \left(T_{HM,in}-T_{HM,out}\right)}{F_L\times C_A\times \left({\alpha }_{rich}-{\alpha }_{lean}\right)\times M{W}_{C{O}_2}}}$$

(16)

where F_G1 and F_G2 represent the volumetric flow rates of the inlet feed gas and outlet off gas, $X_{in}$ and $X_{out}$ represent CO₂ compositions, and ${\alpha }_{rich}$ and ${\alpha }_{lean}$ represent the CO₂ loading of the rich and lean amine solution. The $C_A$ is the amine concentration, $F_L$ is the volume flow rate of the liquid, $m_{HM}$ is the mass flow rate of the hot medium (amine blend), and $C_{pHM}$ is the specific heat capacity of the hot medium (tri-solvent). T_HM,in and T_HM,out represent the inlet and outlet temperatures of the heating medium. $M{W}_{C{O}_2}$ is the molecular weight of CO₂.

4. Experimental Section

4.1. Materials

The chemicals MEA (≥99%), EAE (≥98%), and AMP (≥95%) were purchased from Shanghai Rin En Technology Development Co., Ltd., Shanghai, China. Catalysts HND-8 and HND-580 were purchased from Yinghe Zibo Catalyst Ltd., Shandong, China. The solid base CaSO₄ was purchased from Shandong Shanlan Environmental Group Co., Ltd., Shandong, China. CO₂ (99.9%) and N₂ (99.2%) were purchased from Huzhou Capital Gas Co., Ltd, Zhejiang, China. The silicon oil was purchased from Shanghai Yan Xin Resin Co., Ltd., Shanghai, China. The CO₂ loading tests were performed using a Chittick apparatus from the Association of Official Analytical Chemists (AOAC) [38], with methyl orange as the indicator and 1.0 mol/L HCl as the titration standard.

4.2. Experimental Apparatus and Procedures

This study utilized a bench-scale pilot plant that employs hot silicon oil as its heating medium [20]. Figure 11a is a photo of the experimental setup. The apparatus primarily consists of an absorption tower and a desorption tower, two amine tanks, two peristaltic pumps, a thermal silicon oil heat exchanger, a lean–rich amine heat exchanger, several gas flowmeters to CO₂ and N₂, and a portable CO₂ detector. A detailed explanation of the setup process has been published elsewhere [20].

The installment of the solid acid–base catalysts is depicted in Figure 11b. The absorption tower was filled with structured Sulzer packing, with the solid base CaSO₄ positioned at the bottom, producing a gas–liquid interface. Since the absorber was made of structured packing, the pressure drop was not significant. The desorption tower differs in that it was filled with catalytic bed in the form of random packing, and solid acid HND-8 and HND-580 were interspersed with a mixture of inert marbles (id 0.006 m) and stainless steel θ-rings (id 0.003 m). The catalytic bed layer is consistently situated between any two layers of inert packing. In the non-catalytic operation, the catalytic bed layer was filled with stainless steel θ-rings [20].

Figure 11c shows the hot silicon oil batch, which is the same as in previous publications [20]. The reboiler was no longer used and replaced with a heat exchanger [9,16,18,19]. This study uses thermal silicon oil as the heat source instead of water, which is different from Idem’s study [14,15,16,17]. The advantage of hot oil as a heating source is that its temperature can exceed 100 °C, while the max temperature of hot water is capped at 97 °C. Above 97 °C, the water vapor was overwhelming and led to a huge heat loss. This innovation is illustrative for new CCUS demonstration plants in water-scarce regions, reducing massive amounts of water loss and water evaporation [20].

The operational process was similar to our recent publications [19,20] with operational conditions grouped into

Table 4. The major operation variables of the pilot plant study.

Operation Parameter	Value
tri-solvent	4.2 M MEA-EAE-AMP
liquid flow rate (mL/min)	60
feed gas flow rate (SLPM)	6.5
CO2 concentration	15% CO2 balance with 85% N2
lean amine inlet temperature, Tinlet	25 °C, 30 °C
absorber catalysts	CaSO4 (0–50 g)
desorber catalysts	HND-8, HND-580 (0–250 g)
pressure in both towers	1 atm
hot silicon oil and Cp a	1.63 kJ/(kg·k) (100 °C)
operation temperature of hot oil	100.5 °C

^a the Cp was slightly different for silicon oil under different operation temperatures.

. To introduce it briefly: (1) 8 L of tri-solvent was pumped into the pilot plant from the amine solution tank at F_L of 60 mL/min. (2) After the tri-solvent began circulating stably through the pipeline, a mixed gas of 15% CO₂ and 85% N₂ was introduced from the bottom of the absorption column and exited the top of absorber. The mixed gas underwent a counter-current mode with lean amine solution in the absorber. The CO₂-enriched amine solution exited the bottom of the absorption column. (3) Meanwhile, the thermostatic oil bath was set at 100.5 °C to heat the rich amine solution to a desired temperature of 90–92 °C before it entered the desorption column. The rich amine solution was first preheated via a lean–rich amine heat exchanger, then heated in a thermal oil heat exchanger, and then entered the top of the desorption column at 90 ± 2 °C. The solvent evaporation was negligible since the top of the column had a condenser installed and the temperature was below 100 °C. (4) Rich tri-solvent passed through the catalytic desorption bed layer of the desorber, where CO₂ was desorbed out of the solvent, with amine regenerated simultaneously. The desorbed CO₂ carrying water and amine vapors was condensed at the condenser at top of the tower to prevent solvent loss. (5) The lean amine solution then exited from the bottom of the desorber and was cooled and pumped back to the absorber to complete a circulation cycle. (6) This entire process reached a steady state approximately 6 h after the beginning of the experiment, and then the experimental data were recorded for analysis. The process was continued for 1–2 h to ensure the data sets recorded every 30 min were comparable to each other, and thus that a steady state was reached. The CO₂ loading was measured using a Chittick apparatus and X_in and X_out were tested with an infrared (IR) gas analyzer.

5. Conclusions

This study investigated a pilot plant using CO₂ capture technology and the synergetic effects of solid acid–base catalysts (HND-580, HND-8, CaSO₄) added into the desorber and absorber. A specified tri-solvent of 4.2 mol/L of MEA-EAE-AMP (0.2 + 2+2 mol/L) was used. This specific “energy-efficient combination” was developed as 0.2 + 2+2 mol/L MEA-EAE-AMP with 40 g CaSO₄ and 100 g HND-580, which could be a potential candidate for CCUS demonstration plants after the scale up process, and this was evaluated through a comprehensive analysis of the AE, CC, and HD to evaluate its feasibility.

For solid acid catalysts, the W_cat contributed to a significant CC increase and HD reduction. With 150 g of HND-8 and 100 g of HND-580 installed, as the optimized W_cat, the CC was increased to 25% for HND-8 and 18.8% for HND-580; the heat duty decreased by 20.2% for HND-8 and 16% for HND-580, because both HND-8 and HND-580 facilitated the breakdown of carbamate and reduced the α_lean of the tri-solvent as a Brønsted acid.

The steady-state operation of the system was further enhanced by the solid acid–base combination of CaSO₄ and HND-8 with an amine solution at a T_inlet of 30 °C in the absorber. Under an optimized Wcat of 40 g of CaSO₄ + 100 g of HND-580, with 6.5 L/min of F_G and 60 mL/min of F_L, the HD further reduced to 24.2% compared to the non-catalytic test. The role of the solid acid was decreasing the α_lean, while the role of solid base was increasing the α_rich. Two catalysts worked on either side independently.

Another energy-efficient combo of 4.2 M MEA-EAE-AMP + HND-8 was proven to be unsuitable for a pilot plant study due to a solvent precipitation issue; further studies may be conducted with different amine blending ratios to develop a different combo that is ready for a pilot plant study.