Limestone (97% CaCO₃) obtained from Yeongheung Power Station (Incheon, Republic of Korea) was ground and sieved to prepare samples of 325 mesh or less for use in this study. The organic acids used included formic acid (AR, ≥ 85.0%), acetic acid (AR, ≥ 99.5%), and propionic acid (AR, ≥ 99.0%) purchased from Daejung Chemicals (Siheung, Republic of Korea), and the limestone slurries to which these organic acids were added were denoted FS, AS, and PS, respectively. The ethanolamines used included monoethanolamine (MEA; AR, ≥ 99.0%), diethanolamine (DEA; AR, ≥ 99.0%), and triethanolamine (TEA; AR, ≥ 99.0%) purchased from Samchun Chemicals (Seoul, Republic of Korea), and the limestone slurries to which ethanolamines were added were denoted MS, DS, and TS, respectively. Limestone slurry without adding these additives or distilled water was denoted NS and DW, respectively.

A schematic diagram of the absorption experimental equipment is shown in Fig. S1 (as shown in supplementary data). The equipment consisted of a bubbling reactor, gas supplier, a slurry sample collection section, and analytical apparatus. The reactor had a double-jacket structure made of glass with a diameter of 13 cm, a height of 20 cm, and an internal volume of 1 L. The quantity of limestone slurry (1 wt%) within the reactor was 600 mL. When adding the additives to the limestone slurry, the respective additives were injected into the reactor at a concentration of 10 mmol/L and stirred for 20 minutes before starting the experiment. The SO₂/N₂ NO₂/N₂, O₂/N₂, and N₂ gases used in the experiment were each flow-controlled using MFC (5850E; Brook-MFC, Hatfield, PA, USA). Gas was injected through two bubble sprayers which were inserted into the limestone slurry in the reactor. The total flow rate of the gas injected into the reactor was 3 L/min. The concentrations of SO₂ and NO₂ were set at 500 and 200 ppm, respectively, based on the concentrations typically generated by coal-fired power plants in South Korea [17]. The oxygen concentration of the actual exhaust gas was less than 6% [23–24]. The O₂ concentration varied at 0, 3, and 6%, and was fixed at 3% in all experiments, excluding those investigating variations in O₂ concentration. The stirring speed was consistently maintained at 300 rpm to smoothly mix and diffuse the gas injected into the limestone slurry within the reactor.

The gas discharged from the reactor was injected into an analysis device after removing moisture using a cooling trap. SO₂ was measured using a Horiba gas analyzer (VS-3000; Horiba, Germany), and the NO₂ was measured using an Ecom gas analyzer (MK 3000; Ecom, Germany). The absorption experiment ended after 180 min for all samples. The removal rates of SO₂ and NO₂ in the reaction gas were calculated as follows.

where C_in refers to the gas concentration injected into the reactor, and C_out refers to the gas concentration discharged from the reactor.

During the reaction process, a predetermined amount of slurry was collected and vacuum-filtered to analyze four types of anions in the limestone slurry, SO₃²⁻ (including HSO₃⁻), SO₄²⁻, NO₂⁻, and NO₃⁻. During the filtration process, the oxidation of HSO₃⁻ and SO₃²⁻ to SO₄²⁻ was prevented by conducting the process within a nitrogen box. The anion analysis of the filtrate was performed using an ion chromatograph (Eco IC; Metrohm, Switzerland). The pH of the limestone slurry was continuously measured in the reactor using a pH meter (pH 340i; WTW, Germany).

The individual absorptions of SO₂ and NO₂, as well as their simultaneous absorption in NS and DW, were conducted in a bubbling reactor. After the absorptions, the desulfurization and denitrification rates, and slurry pH changes were presented in Fig. 1. The main reactions during SO₂ absorption in limestone slurry were the SO₂ hydrolysis in Eq. (2–3), and the CaCO₃ dissolution in Eq. (4), while for NO₂ absorption, the NO₂ hydrolysis in Eq. (5–7) was the main reaction [25]. When SO₂ was absorbed into the NS sample, the desulfurization rate increased considerably compared with that following absorption into the DW sample owing to the pH buffering effect of the limestone; the desulfurization rate of reached 99.6% at the end of the 180 min absorption period (Fig. 1). When NO₂ was absorbed into the NS sample, slurry pH increased considerably compared with that following absorption into the DW sample; however, the denitrification rate only increased by approximately 3%, reaching approximately 22%. When simultaneously absorbing SO₂ and NO₂ into the NS sample, the slurry pH was slightly higher, and the desulfurization rate was slightly lower at approximately 98% compared with that for the absorption of SO₂ only. In contrast, while the slurry pH was lower compared with that of NO₂ absorption alone, the denitrification rate increased by approximately 12%, reaching approximately 35%. Fig. 2 presents the concentration of SO₃²⁻ (including HSO₃⁻), SO₄²⁻, NO₂⁻, and NO₃⁻, measured at 30 min intervals during the absorption of SO₂ and the simultaneous absorption of SO₂ and NO₂ in the NS shown in Fig. 1. When only SO₂ was absorbed in the NS sample, SO₄²⁻ and SO₃²⁻ (including HSO₃⁻) were detected. In contrast, when SO₂ and NO₂ were simultaneously absorbed, only SO₄²⁻ was detected along with NO₂⁻ and NO₃⁻, whereas SO₃²⁻ (including HSO₃⁻) was not detected. When SO₂ and NO₂ were simultaneously absorbed, the concentration of SO₄²⁻ did not substantially differ from the combined concentrations of SO₄²⁻ and SO₃²⁻ (including HSO₃⁻) when only SO₂ was absorbed. These results were attributed to the conversion of HSO₃²⁻ and SO₃²⁻, which were the products of the hydrolysis of SO₂ in limestone slurry, into SO₄²⁻ ions by reacting with NO_2(aq) as shown in Eq. (8–9) [26].

To examine the impact of O₂ concentration on the desulfurization and denitrification rates when SO₂ and NO₂ are simultaneously absorbed in the NS samples, the changes in the desulfurization and denitrification rates, and slurry pH with respect to O₂ concentration were presented in Fig. 3. As the O₂ concentration increased, slurry pH increased, however the desulfurization rate remained consistent between 96.3–97.9% regardless of the O₂ concentration at the end of the experiment. This result was attributed to the high concentration of limestone in the slurry compared with the SO₂ input concentration and the limitation of the absorption time to 180 min for the end of the experiment. In contrast, the denitrification rate decreased considerably with increasing O₂ concentration. In addition, at the end of the experiment, the denitrification rates were approximately 58, 33, and 26% for samples with O₂ concentrations of 0, 3, and 6%, respectively. The concentrations of the four types of anions comprising SO₃²⁻ (including HSO₃⁻), SO₄², NO₂⁻, and NO₃⁻ were measured at 30 min intervals in experiment of Fig. 3, and the SO₄²⁻ and NO₃⁻ fractions according to O₂ concentration are presented in Fig. 4. The SO₄²⁻ fraction in the 0% O₂ sample was approximately 0.5–0.7. However, the SO₄²⁻ fraction in the 3 and 6% O₂ samples was 1. The NO₃⁻ fraction in the slurry increased as the O₂ concentration increased. The ion chromatography analysis results suggest that the increase in O₂ concentration led to an increase in the oxidation of SO₃²⁻ (including HSO₃⁻) and NO₂⁻ by O₂, as shown in Eq. (10–12) [25–27]. Therefore, the increased HSO₃⁻ and SO₃²⁻ oxidation in the slurry resulted in a decrease in the reactions of Eq. (8–9), leading to a decrease in denitrification and an increase in slurry pH, as shown in Fig. 3.

Fig. 5 presents the desulfurization and denitrification rates, denitrification amount, and slurry pH when SO₂ and NO₂ are simultaneously absorbed in the limestone slurry with added formic, acetic, and propionic (FS, AS, and PS, respectively). In addition, these were compared with the NS sample. As shown in Fig. 5A, the desulfurization rates of the slurry samples were similar, ranging from 96.3 to 97.9%. The denitrification rate of each sample increased in the order of NS (33.5%) < PS (35.9%) < AS (38.7%) < FS (42.9%) by the end of the experiment. As shown in Fig. 5B, the denitrification amounts of each sample, increased in the order of NS (1.50 mmol) ≈ FS (1.51 mmol) < AS (1.59 mmol) < PS (1.85 mmol). Comparison with the NS, the denitrification amount increased by 6% for the AS and 23% for the FS. As shown in Fig. 5C, the slurry pH of each sample was in the order of FS < AS < PS < NS. In the simultaneous absorption of SO₂ and NO₂ in limestone slurry with organic acids, an additional reaction was the pH buffering of organic acids as shown in Eq. (13–15) [26,28]. In these equations, HA represents the organic acids. Since HSO₃⁻ is produced by the reaction shown in Eq. (14), the reactions shown in Eq. (8–9) could increase. In Fig. 5C, the slurry pH of the NS sample shows a trend of rapidly decreasing at the start of absorption and then gradually slowing down. In contrast, the slurry pH of the samples with organic acids displayed a gradual decrease from the start of absorption due to the organic acids being mixed in the limestone slurry before the absorption experiment. According to our previous research [16], in the absorption of SO₂ in limestone slurry, the more acidic the added organic acid, the more likely it is to inhibit the oxidation of SO₃²⁻ (including HSO₃⁻). As shown in Table S2, the slurry added formic acid, with low pKa, had a high SO₃²⁻ (including HSO₃⁻) oxidation inhibitory capacity. In contrast, the slurry with added acetic acid and propionic acid, both with high pKa, had a low SO₃²⁻ (including HSO₃⁻) oxidation inhibitory capacity, similar to the slurry without additives. Thus, as the acidity of the organic acid added to the limestone slurry increased, the denitrification amount of the sample increased, and the slurry pH decreased.

Fig. 6 presents the concentrations of four types of anions, SO₃²⁻ (including HSO₃⁻), SO₄²⁻, NO₂⁻, and NO₃⁻ and the SO₄²⁻ and NO₃⁻ fractions measured at 30 min intervals for the slurry samples shown in Fig. 5. As shown in Fig. 6A, SO₃²⁻ (including HSO₃⁻) was not detected in all FS, AS, and PS samples, but only SO₄²⁻ was detected, whereas NO₂⁻ and NO₃⁻ were detected in all three samples. The X-ray diffraction analysis confirmed the formation of gypsum in all samples. In Fig. 6B, the fraction of SO₄²⁻ in the NS, FS, AS, and PS samples, was 1, and the fraction of NO₃⁻ was in the order of NS < FS < AS < PS. These results imply an increase in the oxidation of NO₂⁻ to NO₃⁻ in the same order. It is consistent with PS < AS < FS in the order of the inhibitory capacity for the SO₃²⁻ (including HSO₃⁻) oxidation of each sample that contributed to the denitrification amounts of the samples shown in Fig. 5B. However, the reason for the higher NO₃⁻ fraction in the FS sample, which had the highest oxidation inhibitory capacity, than in the NS sample could be explained by the reaction shown in Eq. (15), wherein the NO₃⁻ fraction increased, aside from the reactions shown in Eq. (8–9). Therefore, combining the denitrification amounts of NS ≈ PS < AS < FS in Fig. 5B and the NO₃⁻ fraction order of NS < FS < AS < PS in Fig. 6B, it was presumed that the ratio of the Eq. (15) reaction in the NO₂ absorption reactions of Eq. (8–9), and (15) also gradually increased as the acidity of the organic acid and the denitrification of the samples increased.

Fig. 7 presents the desulfurization and denitrification rates, denitrification amount, and slurry pH when SO₂ and NO₂ were simultaneously absorbed in limestone slurry with added MEA, DEA, and TEA (MS, DS, and TS, respectively). As shown in Fig. 7A, the desulfurization rates of the samples were similar, ranging from 97.2% to 98.2%. The denitrification rates of the samples increased in the order of MS (30.7%) < DS (31.1%) < NS (33.5%) < TS (40.1%) by the end of the experiment. In contrast, the denitrification rates showed a different order of NS < MS < TS < DS until approximately 50 min into the absorption reaction. As shown in Fig. 7B1, the denitrification amounts of the samples were in the order of NS (1.50 mmol) < MS (1.53 mmol) < DS (1.86 mmol) < TS (1.94 mmol) by the end of the experiment. Compared with the NS, the DS and TS increased by 24% and 29%, respectively. As shown in Fig. 7C, the slurry pH of the samples, displayed an order of NS < TS < DS < MS within 50 min of absorption, subsequently exhibiting an order of TS < NS < DS ≈ MS afterwards. The reaction between SO₂ and ethanolamine was an acid-base reaction due to the basicity of ethanolamine, as shown in Eq. (16–18) [29]. Since HSO₃⁻ and SO₃²⁻ were produced by the reactions shown in Eq. (16–17), the reactions shown in Eq. (8–9) could increase. As shown in Fig. 7C, the slurry pH of each sample followed the order MS < TS < DS within 50 min of absorption, corresponding to the order of ethanolamine basicity as shown in Table S2. Nevertheless, the order of the denitrification rate of the samples differed, being in the order of MS < TS < DS (Fig. 7A).

Fig. 8 presents the concentrations of the four types of anions, SO₃²⁻ (including HSO₃⁻), SO₄²⁻, NO₂⁻, and NO₃⁻ and the SO₄²⁻ and NO₃⁻ fractions measured at 30 min intervals for the slurry samples shown in Fig. 7. As shown in Fig. 8A, the SO₃²⁻ (including HSO₃⁻) concentration was the highest in the order of MS < DS < TS during the absorption experiment. This indicates that the inhibitory capacity for SO₃²⁻ (including HSO₃⁻) oxidation was higher in the same order of samples. The X-ray diffraction analysis confirmed the formation of gypsum in all samples. In Fig. 8B, the SO₄²⁻ and NO₃⁻ fractions were lower in the order of NS > MS > DS > TS. Wang et al. [30] reported that in Wet FGD, the calcium sulfite oxidation, in the presence of ethanol in limestone slurry, is controlled by the reactions shown in Eq. (19–20). In these equations, ·SO₃⁻ acts as a radical, with HSO₃⁻ becoming an intermediate in the oxidation process to SO₄²⁻. Eq. (21) is the application of MEA with a 2-hydroxyethyl group to Eq. (19). If the reaction shown in Eq. (21) is assumed to be reversible, the inhibitory capacity of the HSO₃⁻ oxidation of ethanolamines depends on the number of 2-hydroxyethyl groups in the ethanolamine, and it could be inferred that it increases in the order of MEA < DEA < TEA. Therefore, the results in Fig. 7A, which show the order of denitrification rate by sample as MS < TS < DS, can be attributed to the complex effects of the basicity differences in ethanolamine and the inhibitory capacity for the SO₃²⁻ (including HSO₃⁻) oxidation of ethanolamines.

Fig. 7B2 presents the denitrification amount of each sample from 60 to 180 min, excluding the denitrification amount under 60 min. The denitrification amount of slurry with added ethanolamine presented a linear increase over time, similar to that of slurry without the addition of ethanolamine. This indicates a considerable reduction in the acid-base reaction due to the basicity of the ethanolamine. As shown in Fig. 7B2, the denitrification amounts of the samples were in the order of MS < NS ≈ DS < TS. It could be presumed that for the MS and DS samples, the added ethanolamines were consumed, and the inhibitory capacity for SO₃²⁻ (including HSO₃⁻) oxidation had mostly disappeared. Choi et al. [31] reported that among the amines used in CO₂ absorption liquids, primary and secondary amines have a fast CO₂ absorption rate but a low absorption amount, while tertiary amines have a slow absorption rate but a high absorption amount. Also, some researchers have reported that the CO₂ absorption capacity of ethanolamine is in the order of MEA > DEA > TEA [32–33]. In Fig. 7A, the denitrification rate of the slurry with added ethanolamine decreased after approximately 10 min of absorption. The denitrification rates of the MS with added primary amine, MEA, and the DS with added secondary amine, DEA, were lower than that of the NS from approximately 50 and 100 min of absorption, respectively. In contrast, the TS with added tertiary amine, TEA, maintained a higher denitrification rate than that of the NS until the end of the experiment at 180 min. These results are attributed to the decrease in the reactions shown in Eq. (16–17) due to the increased degradation of MEA and DEA in the slurry, which rapidly reacted with the CO_2(aq) generated from the reactions shown in Eq. (4) and (18). As shown in Fig. 8A, the total concentrations of NO₂⁻ and NO₃⁻ in the TS and DS samples, which had high denitrification amounts, were lower than those of NO₂⁻ and NO₃⁻ in the FS sample, which had a low denitrification amount as shown in Fig. 6A. The total concentrations of NO₂⁻ and NO₃⁻ in the MS sample were also lower than those of the NS and PS samples, which had a similar denitrification amount. The reaction of ethanolamine and the hydrolysis product of NO₂ are represented in Eq. (22–26), where R represents the 2-hydroxyethyl group. In the case of MEA and DEA, n-nitrosoethanolamine and n-nitrosodiethanolamine are generated in the reactions shown in Eq. (22) and (23) respectively, through which MEA and DEA are degraded [34]. For TEA, triethanolamine nitrite and nitrate are formed as shown in Eq. (24–25), and degradation of TEA can also occur through Eq. (26) [15,35]. Although nitrosamine can be formed in the reaction between TEA and NO₂, this reportedly occurs when TEA is not sufficiently hydrated [36]. Therefore, the lower total concentrations of NO₂⁻ and NO₃⁻ in the slurries with added ethanolamines than those in slurries with added organic acids in Fig. 8A are attributed to ethanolamine degradation during the absorption process of NO₂. Moreover, the total concentrations of NO₂⁻ and NO₃⁻ in the DS and TS samples appear similar. It is presumed that if the ethanolamine degradation by NO₂ does not greatly differ, ethanolamine is degraded more in the DS with added DEA, which has a faster CO₂ absorption rate than that of TEA.