An investigation into the adverse effects of O₂, SO₂, and NO_x on polyethyleneimine functional CO₂ adsorbents

Abstract

In this study, we investigated the influence of O₂, SO₂, and NO_x on branched and linear polyethyleneimine (PEI) functional silica CO₂ adsorbents (BPEI-SiO₂ and LPEI-SiO₂, respectively). O₂ was much more likely to oxidize BPEI-SiO₂, compared with LPEI-SiO₂, to form C=O and C=N groups and led to a 23.0% decrease in the CO₂ adsorption capacity after 990 min of cumulative contact with 10% O₂. In contrast, LPEI-SiO₂ lost only approximately 3.6% of its CO₂ adsorption capacity, although O₂ oxidized LPEI-SiO₂ to form C=O groups. SO₂ can cause severe degradation of BPEI-SiO₂ and LPEI-SiO₂ by forming heat-stable NH₃⁺—and/or NH₂⁺—containing adducts and by promoting the formation of urea linkages. After cumulative contact with 10, 50, and 200 ppm SO₂ for 990 min, BPEI-SiO₂ lost 18.2%, 61.4%, and 89.0% of its CO₂ adsorption capacity, and LPEI-SiO₂ lost 18.5%, 60.6%, and 78.5% of its CO₂ adsorption capacity, respectively. NO₂ at 10 ppm and NO at 200 ppm caused almost no loss in CO₂ adsorption capacity after cumulative contact for 990 min, but both led to degradation of adsorbents. NO₂ can cause irreversible formation of NH₃⁺—and/or NH₂⁺—containing adducts, acid products, N-nitro compounds (N–NO₂), C-nitroso compounds (C–N=O), and C-nitro (C–NO₂) compounds, and can promote the formation of urea linkages. NO can lead to the formation of NH₃⁺—and/or NH₂⁺—containing adducts and N-nitroso (N–N = O) compounds.

1 Introduction

CO₂ capture, utilization, and storage is a critical technology for realizing net-zero emissions [1]. Over the past decades, CO₂ adsorbents, such as zeolite [2, 3], porous carbon [4, 5], metal–organic frameworks [6, 7], calcium looping technology [8, 9], and solid amine adsorbents [10,11,12], have attracted much attention for capturing CO₂ from flue gas. Of the various adsorbents, solid amine has been considered a good choice for trapping CO₂ directly from flue gas due to its excellent CO₂ adsorption performance and low energy penalty [13,14,15,16]. To protect solid amine adsorbents from high concentrations of SO₂ (500–2500 ppm) and NO_x (1500–2500 ppm) [17,18,19,20,21,22,23,24] during post-combustion CO₂ capture, the best location of the CO₂ capture unit is after flue gas denitrification and desulfurization [23, 25,26,27]. However, a certain amount of SO₂ (50–200 ppm), NO_x (100–400 ppm) [16, 27,28,29,30,31,32], and O₂ (3–10%) [21, 33,34,35,36], which can cause degradation of solid amine sorbents, are still present in the flue gas after denitrification and desulfurization.

According to previous studies, O₂ can oxidize the organic components of solid amine sorbents by forming C=O [21, 37,38,39,40,41,42], N=O [40,41,42], C=N [37, 40, 42, 43], and aliphatic C=C/heterocyclic C-N/aromatic C=C [23]. Furthermore, SO₂ can lead to degradation of solid amine adsorbents by forming sulfate [17, 22, 44], sulfite [17, 22, 44, 45], bisulfite (under humid conditions) [45], and even nitro- and quinone-type compounds (aromatic amine) [22]. NO₂ can result in degradation of solid amine adsorbents by forming nitrate and nitro-compounds [46, 47]. Although it has been reported that NO has no apparent adverse effects on solid amine sorbents [27, 31, 44, 46, 48, 49], a few researchers found that NO can lead to the loss of CO₂ adsorption capacity in solid amine sorbents [28, 50].

Overall, the degradation of solid amine adsorbents induced by O₂, SO₂, and NO_x has been investigated relatively comprehensively by past studies. However, there is still a lack of information on the degradation of PEI functional adsorbents when interacting with O₂, SO₂, and NO_x, which is the focus of this study. BPEI and LPEI functional adsorbents were evaluated during long-term interactions with O₂, SO₂, and NO_x. Their degradation mechanisms were explored in more detail to clarify the degradation pathways of PEI molecules.

2 Experimental work

2.1 Chemicals

BPEI (molecular weight [MW] 25,000) and potassium bromide (KBr, IR grade) were purchased from Alfa Aesar (Tewksbury, MA, USA). LPEI (MW 25,000) was purchased from Polysciences (Warrington, PA, USA). The molecular structures of BPEI and LPEI are shown in Scheme S1 in the Supplementary Information. Methanol (HPLC grade) was purchased from Fisher Scientific (Waltham, MA, USA). Nano silica used as support was synthesized in lab, its surface area and pore volume respectively were 418 m²/g and 0.84 cm³/g [51]. All of the gases, including 99.999% N₂, 15% CO₂ balanced with N₂, 15% CO₂ and 10% O₂ balanced with N₂, 15% CO₂ and 200 ppm NO balanced with N₂, 15% CO₂ and 10 or 200 ppm NO₂ balanced with N₂, 15% CO₂ and 10, 50, or 200 ppm SO₂ balanced with N₂, were provided by ZG Special Gases (Beijing, China).

2.2 Preparation of PEI functional adsorbents

First, BPEI or LPEI was dissolved into 25 ml methanol and stirred for 30 min using a magnetic mixer. Second, 2 g nano-silica was added to the solution. The silica had been dried in an oven at 105 °C under vacuum conditions (< 1 mm Hg) for 3 h. Third, another 5 ml methanol was added to the solution with stirring at ambient temperature until all of the methanol evaporated. Finally, the sample was dried at 50 °C under vacuum conditions (< 1 mm Hg) for 2 h. The products were named BPEI-SiO₂ and LPEI-SiO₂.

2.3 Characterizations

CO₂ cyclic adsorption–desorption tests were conducted using the TGA/DSC 2 thermogravimetric analyzer (Mettler Toledo, Greifensee, Switzerland). First, 15–20 mg adsorbents were placed into an aluminum oxide pan and pretreated at 120 °C for 30 min under a N₂ atmosphere. Second, the samples were cooled to 75 °C, and the N₂ gas was substituted with gas 1 (15% CO₂ balanced with N₂) for a 10-min adsorption period. The temperature was elevated, and gas 1 was switched to N₂ for 10 min desorption at 120 °C. Third, the samples were cooled to 75 °C, and the gas was switched from N₂ to gas 2 (15% CO₂ with some O₂, NO, NO₂, or SO₂ in N₂) for 10 min adsorption. The temperature was elevated, and gas 2 was switched to N₂ for 10 min desorption at 120 °C. Finally, the second and third steps were repeated 100 times. A reference test was also performed in which the second step was repeated 200 times. At the end of each test, the samples were stabilized for 2 h at 75 °C under a N₂ atmosphere. For the analysis, the CO₂ adsorption capacity of the 1st, 3rd, 5th, …, and 199th cycles were used. The samples used for cyclic adsorption–desorption tests under different conditions were denoted as BPEI-SiO₂ or LPEI-SiO₂, followed by the volumetric concentration of CO₂, O₂, SO₂, or NO_x in brackets. For example, BPEI-SiO₂(200 ppm NO) refers to BPEI-SiO₂ samples exposed to 200 ppm NO (15% CO₂ and 200 ppm NO balanced with N₂) in the adsorption–desorption cycles. BPEI-SiO₂(15% CO₂) or LPEI-SiO₂(15% CO₂) indicate that no O₂, SO₂, or NO_x was present during the adsorption–desorption tests.

Diffuse reflectance infrared Fourier transform (DRIFT) spectra for fresh adsorbents and samples from cyclic adsorption–desorption tests were collected by the Nicolet 6700 spectrometer coupled with an in situ reaction cell (Thermo Fisher Scientific, Waltham, MA, USA). The resolution and scan time were set as 4 cm⁻¹ and 32, respectively. The spectra were recorded in the range of 400–4000 cm⁻¹. The spectrum of KBr under N₂ was used as the background. In situ DRIFT spectra of BPEI-SiO₂ and LPEI-SiO₂ during interaction with different gas mixtures were also recorded using the Nicolet 6700 spectrometer. First, fresh samples of BPEI-SiO₂ and LPEI-SiO₂ were placed in the in situ reaction cell, the cell was sealed, and the sample was degassed for 2 h at 120 °C under N₂. The spectra were recorded and denoted as spectrum 1. Second, the samples were cooled to 75 °C under N₂, and the spectra were collected and used as the background for recording the test spectra in the presence of gases 1 or 2. Third, the N₂ was switched to gas 1 or gas 2, and simultaneously start to record infrared (IR) spectra at certain time points. Finally, after 10 or 24 h of interaction with gas 1 or gas 2, the temperature was increased and the atmosphere simultaneously switched to N₂. Samples were regenerated at 120 °C under N₂ for 1 h, and the IR spectra were recorded (using spectrum 1 as a background).

3 Results and discussion

3.1 The adverse effects of O₂

Figure 1 shows the CO₂ adsorption capacity for BPEI-SiO₂(10% O₂) and LPEI-SiO₂(10% O₂), as well as BPEI-SiO₂(15% CO₂) and LPEI-SiO₂(15% CO₂). BPEI-SiO₂(15% CO₂) and LPEI-SiO₂(15% CO₂) exhibited a relatively stable CO₂ adsorption performance during the tests. However, BPEI-SiO₂(10% O₂) lost 23.0% of its original CO₂ adsorption capacity by the 199th cycle. In contrast, LPEI-SiO₂(10% O₂) displayed almost the same stable performance as LPEI-SiO₂(15% CO₂) and only lost approximately 3.6% of its original CO₂ adsorption capacity by the 199th cycle. The CO₂ cyclic adsorption–desorption results demonstrate that LPEI is much more resistant to oxidation by O₂ than BPEI, which is consistent with previous research [37].

Fig. 1

CO₂ cyclic adsorption–desorption results for a BPEI-SiO₂(15% CO₂) and BPEI-SiO₂(10% O₂) and b LPEI-SiO₂(15% CO₂) and LPEI-SiO₂(10% O₂)

Figure 2a shows the DRIFT spectra of BPEI-SiO₂, BPEI-SiO₂(15% CO₂), and BPEI-SiO₂(10% O₂), and Fig. 2b shows the DRIFT spectra of LPEI-SiO₂, LPEI-SiO₂(15% CO₂), and LPEI-SiO₂(10% O₂). Among the DRIFT spectra of BPEI-SiO₂, BPEI-SiO₂(15% CO₂), and BPEI-SiO₂(10% O₂), the most apparent difference is the peak at 1666 cm⁻¹. The weak peak in the case of BPEI-SiO₂ represents the C=O stretching vibration in carbamate and carbamic acid formed by adsorbing CO₂ from the atmosphere [13, 52,53,54,55,56]. The peak's absorption intensity in BPEI-SiO₂(15% CO₂) is slightly stronger than in BPEI-SiO₂ but weaker than in BPEI-SiO₂(10% O₂), mainly due to the C=O vibration in urea linkages [13, 57]. For BPEI-SiO₂(10% O₂), the peak at 1666 cm⁻¹ becomes very prominent and is likely associated with the oxidation of BPEI-SiO₂ by O₂ for various reasons. Bali et al. [38, 45] assigned a similar IR peak, located at 1693 cm⁻¹, as the C=O stretching vibration of amide, acid, and/or imide. Wang et al. [39, 58] assigned a similar band (1659 cm⁻¹) as the amide's C=O stretching vibration. Srikanth et al. [41, 53] assigned a similar peak at 1670 cm⁻¹ as the C=O stretching vibration in amide overlapping with the N=O stretching vibration in nitrites. Additionally, Yu et al. [40, 52] assigned a broad band at 1660–1680 cm⁻¹ as the C=O vibration in amide overlapping with the N=O vibration in nitrites. Gebald et al. [42, 54] assigned a similar peak (1670 cm⁻¹) as the C=N vibration in oxime/imine/nitrile and the C=O vibration in amide/imide. Calleja et al. [43] assigned a similar peak at 1667 cm⁻¹ as the C=N stretching vibration of imine, oxime, and/or nitrone. Assignment of this peak (1666 cm⁻¹) in the DRIFT spectra of BPEI-SiO₂ (10% O₂) is difficult based solely on relevant literature results. Therefore, further analysis was performed.

Fig. 2

DRIFT spectra of a BPEI-SiO₂, BPEI-SiO₂(15% CO₂), and BPEI-SiO₂(10% O₂) and b LPEI-SiO₂, LPEI-SiO₂(15% CO₂), and LPEI-SiO₂(10% O₂)

Figure 2b shows a weak peak at 1560 cm⁻¹ for LPEI-SiO₂, attributed to the COO⁻ stretching vibration in carbamate due to adsorption of CO₂ from the atmosphere [56]. However, the peak becomes more prominent for LPEI-SiO₂(15% CO₂) and LPEI-SiO₂(10% O₂) due to the C-N stretching vibration of urea linkages [53,54,55, 59]. In the DRIFT spectra of LPEI-SiO₂ and LPEI-SiO₂(15% CO₂), a peak at 1639 cm⁻¹, attributed to the N–H deformation vibration of the secondary amine in LPEI [60], was observed. However, in the IR spectra of LPEI-SiO₂(10% O₂), the peak at 1639 cm⁻¹ is obscured by a prominent peak at 1658 cm⁻¹, most likely the C=O stretching vibration from the oxidization of LPEI-SiO₂.

Figure 3a and b show the in situ DRIFT spectra of BPEI-SiO₂ and LPEI-SiO₂ when interacting with gas 1 (15% CO₂ balanced with N₂). No apparent changes in the DRIFT spectra were observed during the interaction. After regeneration, the flat line spectrum in Fig. 3a indicated no noticeable degradation induced by CO₂. In Fig. 3b, a negative peak at 1647 cm⁻¹ and a positive peak at 1604 cm⁻¹ were observed after regeneration, which is most likely due to the removal of chemically adsorbed H₂O in LPEI-SiO₂.

Fig. 3

In situ DRIFT spectra of a BPEI-SiO₂ and b LPEI-SiO₂ when interacting with gas 1 (15% CO₂ balanced with N₂)

Figure 4a and b show the in situ DRIFT spectra of BPEI-SiO₂ and LPEI-SiO₂, respectively, when interacting with 10% O₂ (15% CO₂ and 10% O₂ balanced with N₂). No apparent changes in the DRIFT spectra were observed during the interaction. After a 10-h interaction, no peaks indicating oxidation by O₂ were observed in the DRIFT spectra of regenerated BPEI-SiO₂. A weak peak at 1670 cm⁻¹ in the DRIFT spectra of regenerated LPEI-SiO₂ was attributed to the C=O vibration derived from oxidation by O₂. After a 24-h interaction, the peak size at 1670 cm⁻¹ increased in the DRIFT spectra of regenerated LPEI-SiO₂. In the DRIFT spectra of regenerated BPEI-SiO₂, two positive peaks appeared at 1670 cm⁻¹ and 1606 cm⁻¹ and two negative peaks at 2941 cm⁻¹ and 2817 cm⁻¹. The two negative peaks (2941 cm⁻¹ and 2817 cm⁻¹) corresponded to the C–H asymmetric and symmetric stretching vibration [56, 61, 62], indicating the loss of –CH₂– groups in BPEI. The positive peak at 1670 cm⁻¹ is likely due to the C=O stretching vibration, and the positive peak at 1606 cm⁻¹ indicates C=N vibration [37]. We concluded that O₂ oxidizes –CH₂– groups in BPEI to form C=O and can also oxidize –CH₂–NH– to form C=N groups. The C=O pathway seems to dominate based on the absorption intensity of the C=O peak (1670 cm⁻¹), which is much stronger than the C=N peak (1606 cm⁻¹). In LPEI, O₂ oxidizes −CH₂– groups to form a small number of C=O groups.

Fig. 4

In situ DRIFT spectra of a BPEI-SiO₂, and b LPEI-SiO₂ during interaction with 10% O₂ (15% CO₂ and 10% O₂ balanced with N₂)

These results demonstrate that the oxidization of BPEI-SiO₂ and LPEI-SiO₂ is a relatively slow process, but BPEI-SiO₂ is more readily oxidized than LPEI-SiO₂. We speculated that the CO₂ adsorption capacity of LPEI-SiO₂ would also gradually decrease if we increased the test duration.

3.2 The adverse effects of SO₂

Figure 5a and b show the CO₂ adsorption capacity of BPEI-SiO₂ and LPEI-SiO₂ after exposure to SO₂. SO₂ led to a severe decrease in CO₂ adsorption capacity. An almost linear decrease in CO₂ adsorption capacity was observed for both BPEI-SiO₂ and LPEI-SiO₂ after exposure to 10 or 50 ppm SO₂ (15% CO₂ and 10 or 50 ppm SO₂ balanced with N₂). The CO₂ adsorption capacity of BPEI-SiO₂ and LPEI-SiO₂ cumulatively decreased by 18.2% and 18.5% at 10 ppm SO₂ and by 61.4% and 60.6% at 50 ppm SO₂. When the level of SO₂ reached 200 ppm, the CO₂ adsorption capacity of BPEI-SiO₂ and LPEI-SiO₂ respectively lost 89.0% and 78.5%. And the decrease in the CO₂ adsorption capacity occurred mainly in the first 50–60 cycles in the 200 ppm SO₂ scenario. For example, the CO₂ adsorption capacity of BPEI-SiO₂ at the 60th cycle and of LPEI-SiO₂ at the 50th cycle decreased by 85.5% and 74.1%, respectively. During subsequent cycles, the decreasing CO₂ adsorption capacity reached a plateau. The stable CO₂ adsorption performance in the plateaus may be due to the residual isolated amino groups, which could adsorb CO₂ and, more importantly, could adsorb SO₂ reversibly [28, 44, 48, 52].

Fig. 5

CO₂ cyclic adsorption–desorption results for a BPEI-SiO₂(15% CO₂), BPEI-SiO₂(10 ppm SO₂), BPEI-SiO₂(50 ppm SO₂), and BPEI-SiO₂(200 ppm SO₂) and b LPEI-SiO₂(15% CO₂), LPEI-SiO₂(10 ppm SO₂), LPEI-SiO₂(50 ppm SO₂), and LPEI-SiO₂(200 ppm SO₂)

Figure 6a shows a peak at 1662 cm⁻¹ for each sample, typically associated with the C=O stretching vibration, but the absorption intensities differed significantly. For BPEI-SiO₂, the peak was derived from the C=O stretching vibration in carbamate and carbamic acid due to adsorption of CO₂ from the atmosphere [13, 52,53,54,55,56]. For BPEI-SiO₂(15% CO₂), the formation of urea linkages was most responsible for the peak [13, 57]. For BPEI-SiO₂(10 ppm SO₂), the peak’s adsorption intensity was similar to that for BPEI-SiO₂(15% CO₂). However, for BPEI-SiO₂(50 ppm SO₂) and BPEI-SiO₂(200 ppm SO₂), the peak’s intensity increased, likely due to the affinity of SO₂ to BPEI-SiO₂.

Fig. 6

DRIFT spectra of a BPEI-SiO₂, BPEI-SiO₂(15% CO₂), BPEI-SiO₂(10 ppm SO₂), BPEI-SiO₂(50 ppm SO₂), and BPEI-SiO₂(200 ppm SO₂) and b LPEI-SiO₂, LPEI-SiO₂(15% CO₂), LPEI-SiO₂(10 ppm SO₂), LPEI-SiO₂(50 ppm SO₂), and LPEI-SiO₂(200 ppm SO₂)

A similar phenomenon was observed at 1666 cm⁻¹ in Fig. 6b. Meantime, two other peaks at 1496 cm⁻¹ and 1560 cm⁻¹, most likely attributed to the C–N stretching vibration of urea linkages [53,54,55, 59], can also be observed in Fig. 6b. Thus, we postulate that the intense peak at 1666 cm⁻¹ for SO₂-exposed samples is due mainly to the C=O stretching vibration of urea linkages. The peak at 1662 cm⁻¹ in Fig. 6a may also represent the C=O stretching vibration of urea linkages. Therefore, we concluded that SO₂ promotes the formation of urea linkages when PEI functional adsorbents interact with CO₂ streams containing SO₂. Moreover, NH₂⁺ deformation vibrations were observed at 1616 cm⁻¹ in the DRIFT spectra of LPEI-SiO₂(10 ppm SO₂), LPEI-SiO₂(50 ppm SO₂), and LPEI-SiO₂ (200 ppm SO₂) [56, 61, 63] in Fig. 6b. These represent the formation of heat-stable NH₂⁺-containing adducts between SO₂ and LPEI-SiO₂.

Figure 7a and b show in situ DRIFT spectra of BPEI-SiO₂ and LPEI-SiO₂, respectively, when interacting with 200 ppm SO₂ (15% CO₂ and 200 ppm SO₂ balanced with N₂) at 75 °C. In Fig. 7a, the initial peak at 3022 cm⁻¹, representing the NH₃⁺/NH₂⁺ vibration [56, 61, 63], clearly strengthened and gradually shifted to 3078 cm⁻¹ with prolonged interaction time. Similarly, the initial peak at 1628 cm⁻¹, representing the NH₃⁺ vibration [56, 61, 63], clearly strengthened and gradually shifted to 1647 cm⁻¹. After regeneration, the two peaks had a high absorption intensity. The peak at 2546 cm⁻¹, representing the NH₃⁺/NH₂⁺ vibration [56, 61, 64], and the peak at 2104 cm⁻¹, representing the NH₃⁺ vibration [56, 61, 64], were observed after regeneration. The two peaks at 1562 cm⁻¹ and 1500 cm⁻¹, which represent the COO⁻ stretching vibration in carbamate [52, 56, 61, 64,65,66], disappeared after regeneration, implying the release of the adsorbed CO₂. Therefore, the remaining NH₃⁺/NH₂⁺ groups must originate from the formation of heat-stable NH₃⁺/NH₂⁺-containing adducts between BPEI-SiO₂ and SO₂. Meanwhile, the peaks at 1018 cm⁻¹ and 966 cm⁻¹ most likely belong to the asymmetric and symmetric S=O stretching vibration [22, 45, 46, 63, 67], and the peak at 841 cm⁻¹ represents the S–O stretching vibration [46, 68], which all suggest the existence of sulfur-containing products. All of the above peaks are observed at similar locations in Fig. 7b.

Fig. 7

In situ DRIFT spectra of a BPEI-SiO₂ and b LPEI-SiO₂ during interaction with 200 ppm SO₂ (15% CO₂ and 200 ppm SO₂ balanced with N₂)

The above analysis demonstrates that SO₂ reacted with BPEI and LPEI to form irreversible NH₃⁺- and/or NH₂⁺-containing adducts. Previous studies reported that sulfites and/or sulfates formed between SO₂ and solid amine adsorbents [17, 44,45,46, 49, 68]. As H₂O and O₂ were free during the interaction processes in this study, we hypothesized that the following Eqs. (1) and (2) describe a possible mechanism for the reactions between SO₂ and amino groups [44, 69]. These equations are similar to the reaction between CO₂ and amino groups under dry conditions (Eqs. (3) and (4)):

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + SO}}_{{2}} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{SOO}}^{ - }$$

(1)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{SOO}}^{ - } \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}_{{2}}^{ + } {\text{ + R}}_{{1}} {\text{R}}_{{2}} {\text{NSOO}}^{ - }$$

(2)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + CO}}_{{2}} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{COO}}^{ - }$$

(3)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{COO}}^{ - } \leftrightarrow R_{{1}} {\text{R}}_{{2}} {\text{NH}}_{{2}}^{ + } {\text{ + R}}_{{1}} {\text{R}}_{{2}} {\text{NCOO}}^{ - }$$

(4)

$${\text{R}}_{{1}} {\text{, R}}_{{2}} {\text{ = H or CH}}_{{2}}$$

The peak at 1680 cm⁻¹ in Fig. 7a and at 1672 cm⁻¹ in Fig. 7b were both observed after regeneration. They most likely belong to the C=O stretching vibration of urea linkages. In previous research [13], we found that the in situ DRIFT peak representing C=O in urea linkages was extremely weak after 11 h of interaction between pure CO₂ and the BPEI (MW = 600 Da) functional adsorbent at 75 °C. Thus, the intense peaks observed after regeneration must be derived from the influence of SO₂ (Fig. 7a and b). As we had speculated, SO₂ can promote the formation of urea linkages between PEI functional adsorbents and CO₂.

The in situ DRIFT spectra of BPEI-SiO₂ or LPEI-SiO₂ in the presence of 10 ppm SO₂ (15% CO₂ and 10 ppm SO₂ balanced with N₂) are shown in Figures S7, S8, and S9 in the Supplementary Information. The spectra were similar to those in the 200 ppm SO₂ scenario, but the absorption intensity was much lower in the 10 ppm SO₂ scenario.

3.3 The adverse effects of NO₂

Generally, NO₂ accounts for only 5% or less of the total NO_x in flue gas [70]. In this study, we used a concentration of 10 ppm NO₂ (15% CO₂ and 10 ppm NO₂ balanced with N₂) to investigate the adverse effects of NO₂ on PEI functional adsorbents. As a reference, 200 ppm NO₂ (15% CO₂ and 200 ppm NO₂ balanced with N₂) was also investigated. CO₂ cyclic adsorption–desorption tests (Fig. 8a and b) showed excellent CO₂ adsorption stabilities for both BPEI-SiO₂ and LPEI-SiO₂ under the 10 ppm NO₂ scenario. However, under the 200 ppm NO₂ scenario, the CO₂ adsorption capacity of BPEI-SiO₂ and LPEI-SiO₂ showed an almost linear decline and were decreased by 49.6% and 49.5%, respectively.

Fig. 8

CO₂ cyclic adsorption–desorption results for a BPEI-SiO₂(15% CO₂), BPEI-SiO₂(10 ppm NO₂), and BPEI-SiO₂(200 ppm NO₂) and b LPEI-SiO₂(15% CO₂), LPEI-SiO₂(10 ppm NO₂), and LPEI-SiO₂(200 ppm NO₂)

In Fig. 9a and b, a sharp peak at 1666 cm⁻¹ was observed for both BPEI-SiO₂(200 ppm NO₂) and LPEI-SiO₂(200 ppm NO₂). The peaks are possibly associated with the C=O stretching vibration in urea linkages. However, in the presence of NO₂, it is difficult to exclude the N=O stretching vibration in nitrites and/or nitrates for this peak (1666 cm⁻¹) [63]. Furthermore, a peak at approximately 1361 cm⁻¹ was observed in the DRIFT spectra of BPEI-SiO₂(200 ppm NO₂) and LPEI-SiO₂(200 ppm NO₂), most likely associated with the formation of N-nitroso compounds [71, 72]. In Fig. 9a, the DRIFT spectra of BPEI-SiO₂(10 ppm NO₂) are similar to those of BPEI-SiO₂ (15% CO₂). In Fig. 9b, a peak at 1612 cm⁻¹, representing the NH₂⁺ deformation vibration [56, 61, 63], was observed for both LPEI-SiO₂(10 ppm NO₂) and LPEI-SiO₂(200 ppm NO₂), but not for LPEI-SiO₂ or LPEI-SiO₂(15% CO₂).

Fig. 9

DRIFT spectra of a BPEI-SiO₂, BPEI-SiO₂(15% CO₂), BPEI-SiO₂(10 ppm NO₂), and BPEI-SiO₂(200 ppm NO₂) and b LPEI-SiO₂, LPEI-SiO₂(15% CO₂), LPEI-SiO₂(10 ppm NO₂), and LPEI-SiO₂(200 ppm NO₂)

Figure 10a and b exhibit the in situ DRIFT spectra of BPEI-SiO₂ and LPEI-SiO₂ during interaction with 200 ppm NO₂ (15% CO₂ and 200 ppm NO₂ balanced with N₂). Figure 10a shows that the initial peak at 3022 cm⁻¹ gradually strengthened and shifted to 3074 cm⁻¹ during prolonged interaction time. This peak represents the NH₃⁺/NH₂⁺ stretching vibration [56, 61, 63] and was still prominent after regeneration. The peak at approximately 2505 cm⁻¹ represents the NH₃⁺/NH₂⁺ stretching vibration [56, 61, 64], and the peaks at approximately 2162 cm⁻¹ and 1631 cm⁻¹ represent the NH₃⁺ vibration [56, 61, 63, 64] and were observed after regeneration. The peak at 1651 cm⁻¹, most likely due to the N=O vibration in nitrites and/or nitrates [63], emerged and gradually strengthened with prolonged interaction time, and the sharp peak was still present after regeneration. Furthermore, the peak at 802 cm⁻¹, representing the C-N stretching vibration in nitrites [63], and the peak at 1126 cm⁻¹, representing N–N stretching vibration in nitrates [46, 63], both remained after regeneration.

All of the above peaks are observed at similar locations in Fig. 10b. Therefore, the formation of NH₃⁺ and/or NH₂⁺-containing nitrites and/or nitrates is an important route for the degradation of BPEI-SiO₂ and LPEI-SiO₂. Considering O₂ and H₂O are free during the interaction processes, we speculate that the possible formation mechanism of nitrites and nitrates is depicted in Eqs. (5) to (9) [73]:

Fig. 10

In situ DRIFT spectra of a BPEI-SiO₂ and b LPEI-SiO₂ when interacting with 200 ppm NO₂ (15% CO₂ and 200 ppm NO₂ balanced with N₂) and of c BPEI-SiO₂ and (d) LPEI-SiO₂ when interacting with 10 ppm NO₂ (15% CO₂ and 10 ppm NO₂ balanced with N₂)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + NO}}_{{2}} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{NOO}}^{ - }$$

(5)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{NOO}}^{ - } \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}_{{2}}^{ + } {\text{ + R}}_{{1}} {\text{R}}_{{2}} {\text{NNOO}}^{ - }$$

(6)

$${\text{NO}}_{{\text{2}}} ^{\cdot} {\text{ + NO}}_{{\text{2}}} ^{\cdot} \leftrightarrow {\text{N}}_{{\text{2}}} {\text{O}}_{{\text{4}}}$$

(7)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + N}}_{{2}} {\text{O}}_{{4}} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{NOOO}}^{ - } {\text{ + NO}}$$

(8)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{NOOO}}^{ - } \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}_{{2}}^{ + } + {\text{ R}}_{{1}} {\text{R}}_{{2}} {\text{NNOOO}}^{ - }$$

(9)

$${\text{R}}_{{1}} {\text{, R}}_{{2}} {\text{ = H or CH}}_{{2}}$$

In Fig. 10a, the peaks at 1525 cm⁻¹, 1396 cm⁻¹, and 1246 cm⁻¹ may represent different types of NO₂ stretching vibrations in N-nitro compounds (N-NO₂) [22, 63]. Corresponding peaks are observed at 1525 cm⁻¹, 1400 cm⁻¹, and 1242 cm⁻¹ in Fig. 10b. Meanwhile, the peak at 962 cm⁻¹ in Fig. 10a and the peak at 957 cm⁻¹ in Fig. 10b may represent the N–N stretching vibration in N-nitro compounds [63]. Moreover, in Fig. 10a, the two peaks at 1525 cm⁻¹ and 1396 cm⁻¹ may also represent NO₂ asymmetric and symmetric stretching vibrations in C-nitro compounds (C–NO₂) [63]. The peak at 1377 cm⁻¹ may represent the N=O stretching vibration in C-nitroso compounds (C-NO) [63], with the C-nitro and C-nitroso compounds arising due to the oxidation of NO₂ to a primary amine in BPEI. N₂O₄ may also act as an oxidizing agent.

The peaks at 1680 cm⁻¹ in Fig. 10a and 1676 cm⁻¹ in Fig. 10b were observed after regeneration. The two peaks are most likely due to the C=O stretching vibration in urea linkages. These two peaks were mutually corroborative with the two peaks at 1666 cm⁻¹ in Fig. 9a and b. Therefore, similar to SO₂, NO₂ also promotes the formation of urea linkages between PEI functional adsorbents and CO₂.

At 3215 cm⁻¹ in Fig. 10a and 3190 cm⁻¹ in Fig. 10b, apparent variations are observed. These two peaks gradually strengthened and can be observed after regeneration. A similar observation was noted at 3224 cm⁻¹ in Fig. 7a. These peaks most likely represent the O–H vibration in acid. We hypothesized that the following Eqs. (10) to (15) might explain the appearance of the peak:

$${\text{RNH}}_{{2}} {\text{ + CO}}_{{2}} \leftrightarrow {\text{RNH}}_{{2}}^{ + } {\text{COO}}^{ - } \leftrightarrow {\text{RHNCOOH}}$$

(10)

$${\text{RNH}}_{{2}} {\text{ + SO}}_{2} \leftrightarrow {\text{RNH}}_{{2}}^{ + } {\text{SOO}}^{ - } \leftrightarrow {\text{RHNSOOH}}$$

(11)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH + NO}}_{{2}} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{NOO}}^{ - } \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NNOOH}}$$

(12)

$${\text{NO}}_{{\text{2}}} ^{{\cdot}} {\text{ + NO}}_{{\text{2}}} ^{{\cdot}} \leftrightarrow {\text{N}}_{{\text{2}}} {\text{O}}_{{\text{4}}}$$

(13)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}_{{2}} + {\text{ N}}_{{2}} {\text{O}}_{{4}} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{NOOO}}^{ - } {\text{ + NO}}$$

(14)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}^{ + } {\text{NOOO}}^{ - } \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NNOOOH}}$$

(15)

$${\text{R}}_{{1}} {\text{, R}}_{{2}} {\text{ = H or CH}}_{{2}}$$

Figure 10 c and d show in situ DRIFT spectra during interaction with 10 ppm NO₂ (15% CO₂ and 10 ppm NO₂ balanced with N₂) for BPEI-SiO₂ and LPEI-SiO₂. The spectra are similar to those in Fig. 10a and b, but with a much weaker absorption intensity. Therefore, 10 ppm NO₂ could also lead to degradation of BPEI-SiO₂ and LPEI-SiO₂ via similar mechanisms. The degradation induced by 10 ppm NO₂ was very slight, and therefore no pronounced decrease in the CO₂ adsorption capacity was observed. However, when the CO₂ cyclic adsorption–desorption cycles were increased, both BPEI-SiO₂ and LPEI-SiO₂ encountered a loss in their CO₂ adsorption capacity under the 10 ppm NO₂ scenario

3.4 The adverse effects of NO

As mentioned above, NO typically accounts for over 95% of the total NO_x in flue gas. We only investigated the impact of 200 ppm NO on BPEI-SiO₂ and LPEI-SiO₂. Figure 11a and b show the changes in the CO₂ adsorption capacity of BPEI-SiO₂ and LPEI-SiO₂ after exposure to 200 ppm NO (15% CO₂ and 200 ppm NO balanced with N₂). The CO₂ adsorption performance of BPEI-SiO₂ and LPEI-SiO₂ was very stable during the whole process.

Fig. 11

CO₂ cyclic adsorption–desorption results for a BPEI-SiO₂(15% CO₂) and BPEI-SiO₂(200 ppm NO) and b LPEI-SiO₂(15% CO₂) and LPEI-SiO₂(200 ppm NO)

In Fig. 12a and b, the DRIFT spectrum of BPEI-SiO₂(200 ppm NO) is similar to that of BPEI-SiO₂(15% CO₂), while the DRIFT spectrum of LPEI-SiO₂(200 ppm NO) is similar to that of LPEI-SiO₂(15% CO₂). Compared with the DRIFT spectra of BPEI-SiO₂, the absorption intensity of the peak at approximately 1666 cm⁻¹ was stronger in the DRIFT spectra of BPEI-SiO₂(200 ppm NO) and BPEI-SiO₂ (15% CO₂). The sharp peak at 1558 cm⁻¹ was far more prominent for LPEI-SiO₂(200 ppm NO) and LPEI-SiO₂(15% CO₂) than for LPEI-SiO₂. These two peaks at 1666 cm⁻¹ and 1558 cm⁻¹ represent the C=O stretching vibration and the C-N stretching vibration in urea linkages. This is due mainly to the formation of urea linkages during the CO₂ cyclic adsorption–desorption processes. The DRIFT spectra provide no information on the degradation induced by NO.

Fig. 12

DRIFT spectra of a BPEI-SiO₂, BPEI-SiO₂(15% CO₂), and BPEI-SiO₂(200 ppm NO) and b LPEI-SiO₂, LPEI-SiO₂(15% CO₂), and LPEI-SiO₂(200 ppm NO)

Figure 13a and b show the in situ DRIFT spectra for BPEI-SiO₂ and LPEI-SiO₂ during interaction with 200 ppm NO (15% CO₂ and 200 ppm NO balanced with N₂). No apparent changes were observed for BPEI-SiO₂ or LPEI-SiO₂ during the interaction processes. But some peaks remained after regeneration in Fig. 13a, for example, the peaks at 2985 cm⁻¹ and 2509 cm⁻¹ denoting the NH₃⁺/NH₂⁺ vibration [56, 61, 64], the peak at 1657 cm⁻¹ representing the N=O vibration [63], the peak at 1606 cm⁻¹ most likely denoting the NH₂⁺ vibration, and the peak at 1011 cm⁻¹ likely representing the N–N stretching vibration. All of these peaks were observed at similar locations in Fig. 13b. Equations (16) to (18), shown below, may provide an explanation for these peaks [69]. The reaction in Eq. (16) limited the formation of R₁R₂NH₂⁺ R₁R₂NHN₂O₂⁻ and further confined the degradation of BPEI-SiO₂ and LPEI-SiO₂ induced by NO.

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH }} + {\text{ NO }} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NHNO }}\left( {{\text{slow}}} \right)$$

(16)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NHNO }} + {\text{ NO }} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NHN}}_{{2}} {\text{O}}_{{2}} \left( {{\text{fast}}} \right)$$

(17)

$${\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NHN}}_{{2}} {\text{O}}_{{2}} {\text{ + R}}_{{1}} {\text{R}}_{{2}} {\text{NH }} \leftrightarrow {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NH}}_{{2}}^{ + } {\text{R}}_{{1}} {\text{R}}_{{2}} {\text{NHN}}_{{2}} {\text{O}}_{{2}}^{ - } \left( {{\text{fast}}} \right)$$

(18)

$${\text{R}}_{{1}} {\text{, R}}_{{2}} {\text{ = H or CH}}_{{2}}$$

Furthermore, the peak at 1361 cm⁻¹ in Fig. 13a and the peak at 1360 cm⁻¹ in Fig. 13b may be associated with the formation of N-nitroso compounds (N–N=O) [71, 72]. These results demonstrate that NO can lead to degradation of PEI functional adsorbents by forming R₁R₂NH₂⁺ R₁R₂NHN₂O₂⁻ and N-nitroso compounds. However, the degradation induced by NO was extremely low and would not cause an obvious decrease in the CO₂ adsorption capacity during the limited duration of CO₂ cyclic adsorption–desorption tests.

Fig. 13

In situ DRIFT spectra of a BPEI-SiO₂, and b LPEI-SiO₂ during interaction with 200 ppm NO (15% CO₂ and 200 ppm NO balanced with N₂)

4 Conclusions

In this study, the adverse effects of simulated flue gas on BPEI and LPEI functional adsorbents were investigated. The results showed that O₂, SO₂, NO₂, and NO all caused degradation of PEI functional adsorbents. After contact with 10% O₂ (15% CO₂ and 10% O₂ in N₂) for 990 cumulative minutes, BPEI-SiO₂ lost 23.0% of its original CO₂ adsorption capacity. On the other hand, LPEI-SiO₂ maintained a stable CO₂ adsorption performance during the adsorption process and only lost approximately 3.6% of its original adsorption capacity. However, the IR spectra demonstrated that 10% O₂ also damages LPEI-SiO₂. Three concentration gradients were considered for SO₂, namely 10, 50, and 200 ppm SO₂ (15% CO₂ and 10, 50, or 200 ppm SO₂ in N₂). Higher concentrations of SO₂ would lead to more severe and quicker degradation of BPEI-SiO₂ and LPEI-SiO₂. After contact with 10, 50, or 200 ppm SO₂ for 990 cumulative minutes, BPEI-SiO₂ lost 18.2%, 61.4%, and 89.0% of its original CO₂ adsorption capacity, and LPEI-SiO₂ lost 18.5%, 60.6%, and 78.5% of its original CO₂ adsorption capacity, respectively. As for NO_x, 10 ppm NO₂ (15% CO₂ and 10 ppm NO₂ in N₂) and 200 ppm NO (15% CO₂ and 200 ppm NO in N₂) caused almost no decrease in the CO₂ adsorption capacity after 990-min interactions, but as shown by the IR spectra, both concentrations induced degradation of BPEI-SiO₂ and LPEI-SiO₂. We also investigated 200 ppm NO₂ (15% CO₂ and 200 ppm NO₂ in N₂) and observed a 49.6% and 49.5% decrease in the original CO₂ adsorption capacity of BPEI-SiO₂ and LPEI-SiO₂, respectively.

Further exploration of the degradation mechanism demonstrated that O₂ oxidized the –CH₂– and –CH₂–NH– groups of BPEI-SiO₂ to form C=O and C=N groups, and C=O formation seems to be the primary pathway. For LPEI-SiO₂, only C=O formation via oxidation of –CH₂– was observed. SO₂ can react with the amine groups of BPEI-SiO₂ and LPEI-SiO₂ to form heat-stable NH₃⁺—and/or NH₂⁺—containing adducts. SO₂ can promote the formation of urea linkages between PEI functional adsorbents and CO₂. Similar to SO₂, the presence of NO₂ can lead to the formation of heat-stable NH₃⁺—and/or NH₂⁺—containing adducts and promote urea linkage formation. Furthermore, NO₂ can result in the formation of heat-stable acid adducts and, likely, N-nitro (N–NO₂), C-nitroso (C–N=O), and C-nitro (C–NO₂) compounds. NO can lead to the formation of heat-stable NH₃⁺—and/or NH₂⁺—containing adducts, as well as N-nitroso (N–N=O) compounds.