| 20 Nov 2025
Modelling gaseous and particulate secondary compounds formation during atmospheric degradation of 2-amino-2-methyl-1-propanol (AMP)
Abstract. The use of amines in carbon capture may contribute to high atmospheric emission. Amines have been shown to contribute significantly to the formation of carcinogenic compounds and aerosols. AMP (2-amino-2-methyl-1-propanol (CH3)2(CH2OH)CNH2) is a benchmark amine in carbon capture solvents, but the mechanisms of its contribution to particle formation are not yet well understood. This study aims to investigate the formation of aminium nitrates in the gas and particle phases. In order to model both the mass and number concentrations of the particles formed, reaction pathways for the formation of extremely low volatile products are proposed, as well as nucleation parameterisation. The model is constrained using an atmospheric chamber experiment, where AMP oxidation products were measured, as well as nitric acid (NA) and AMP in the particulate phase. Considering different molecular clusters of the type NAnAMPm allowed us to introduce low-volatility organic compounds likely to partition into the particulate phase. We present a first method considering the formation of a dimer, NA2AMP2, with low volatility and subject to nucleation parameterization, allowing for accurate reproduction of the evolution of the total number and particle size distribution compared to the experiment. This representation highlights an overestimation of the nitrate mass, emphasizing that the AMP/NA ratio in the clusters may not be equivalent to 1/1. A second method aims to correct this overestimation of nitrate mass by introducing NA1AMP2 clusters, which significantly improve AMP/nitrate partitioning in the particulate phase
Manuscript egusphere-2025-4427 by D. Ladet et al. investigates the mechanism of particle formation related to amines by modelling of a chamber photo-oxidation experiment on AMP, a benchmark amine for carbon capture and storage. In addition to the evaluation of modelled AMP oxidation products in the gas phase against measurements, the formation of particulate organic compounds is compared to measurements by CHARON for the respective size range. Total particle number, the mean particle diameter and particle size distributions measured by SMPS are further used to evaluate the simulated particle formation. The effect of different molecular clusters of type NAnAMPm from reactions of nitric acid and AMP is analyzed in two different simulations of the experiment. While exploring new mechanisms for secondary particle formation involving molecular cluster is appreciated, it needs to be emphasized better what is new about modelling in this study, since the same experiment, albeit only gas-phase products, was already modeled by Tan et al. (2021). This involves a clear outline of the new capabilities and aspects, as well as the limitations of the presented modelling. However, it may be argued that simulation of an additional experiment could lead to the need for further adjustments of the rate coefficients for cluster formation. Several omissions and imprecise statements in the current manuscript impede the complete evaluation. The authors should clarify and accommodate the following questions and concerns, before I can recommend publication of this work.
Specific Comments:
1.) Introduction: Please check the citations within the introduction. The order of multiple citations seems arbitrary (example: P2, L32). In case of citing several sources in one row, they should be either in alphabetical order or by year, depending on the journal’s formatting guidelines.
2.) Experimental context: rephrase first sentence on experiment (P3, L72-74). Tan et al. (2021) have modeled the same experiment (June 15, 2015). It needs to be stated why this experiment was chosen here and what information is missing in the simulation shown in Tan et al. (2021). Their work also presents five other photo-oxidation experiments, why have these not been simulated here?
3.) Are time stamps given as local time or UTC (e.g., 07:48:00 on P3, L82)? This must be indicated at all instances throughout the manuscript, including the figure plots.
4.) NA2AMP2 clusters (P7, L167-168): can you elaborate a bit more on "subject to a mechanistic representation" of the physical processes governing aerosol evolution? Maybe it should be stated somewhere that the initial aerosol is assumed to consist completely of NA2AMP2, as Figure 4b indicates. Is this assumption consistent in simulation 1 and 2?
5.) P4, L178: unfortunately, in contrast to what is stated here, the kinetic coefficients of R1 to R4 are not provided in Table 7.
6.) Simulation setup: at this point (before the simulation setup section) a clearly formulated section must be inserted that outlines the procedure of rate coefficient adjustments for simulation 1 and simulation 2 as comprehensible as possible. Answering the questions: what was the basis of adjusting rate constants before simulation 1 and what was the basis for adjustment of simulation 2. Were adjustments for simulation 2 decided in the beginning or after simulation 1 was performed?
7.) Figure 9 shows measured and simulated temporal concentration profiles of AMP oxidation products. Tan et al. (2021) have the same figure in their publication (Figure 9 therein). IPP in their simulation is not overpredicted. Please discuss the discrepancy based on differences in their gas-phase reaction mechanism for AMP.
8.) Formation of AMP-nitramine (AMPNO2) (P15/16): The conversion of 138% of reacted AMP to AMPNO2 (P16, L336) must be an error. I suggest comparing the modelled AMPNO2 yield against Braten et al. (2008) who suggest an AMP-nitramine yield of 0.4% of AMP reacted per ppbV NO2 (cited from Tan et al., 2021).
9.) Discuss the possibility of uptake of AMPNO2 to particles and heterogeneous surface reactions on particles.
10.) Despite simulation 1, which employs the low-volatility dimer NA2AMP2, tends to overestimate the nitrate mass, there are several reasons why simulation 1 should be preferred over simulation 2. What speaks against simulation 2 is that production of PAMP stops earlier than in simulation 1 (Figure 9) even though there is still sunlight to produce HNO3 (measured NO2 is still increasing). Also, I believe that measurements of the total particle number and the particle size distribution have a higher accuracy than the mass-based concentration measurements. Therefore, matching the size-resolved particle numbers is a stronger criterion for evaluation of the simulations than the more uncertain mass component measurements. Simulation 2 performs worse in comparison to SMPS size distribution than simulation 1. Furthermore, the mean diameter (Fig. 10a) in simulation 2 shows a sharp drop 15 minutes before this is observed by the SMPS measurement, indicating that the nitrate aerosol has too high volatility in this simulation.
11.) In addition to comparing the two simulations, it would be helpful to see the simulated particle formation for gradually changing the volatility of the molecular clusters, as a test of sensitivity towards the assumed or derived saturation vapor pressures.
12.) It seems that the reaction numbers (R9-R14) in Table 7 are not the ones that were intended for this table. I had rather expected a table with the rate coefficients of R1 to R6. Table 3 already provides rate coefficients for the chamber-specific reactions R10-R13 and particle loss (is that R14?). Reaction R9 denotes the reaction HO2+NO, which according to the NIST database is 8.91E-12 cm3/molecule s at 298 K (Atkinson, 2004) and certainly not uncertain by two orders of magnitude. At least the reasoning for changing rate constants of R9-R14 in simulation 2 needs to be justified.
13.) The broadening of the modelled number size distribution during growth indicates numerical diffusion. It is stated that increasing the number of size bins does not provide a better representation (P22, L403-405). Nevertheless, the possibility of numerical diffusion needs to be discussed, and the effect of changing the size representation on the width of the modelled size distribution should be documented in a supplement.
14.) Conclusions: atmospheric implications of the suggested formation of molecular nitric acid – AMP clusters in competition to nucleation of amines with sulfuric acid should be more quantitative. For the typical ranges of atmospheric mixing ratios of dimethyl amine (DMA) and sulfuric acid in the background boundary layer, which mixing ratios of AMP and nitric acid would be required to preferentially form NA2AMP2 clusters? The formation of aminium nitrate particles would probably only be relevant in the close vicinity of a capture plant. Such conclusions would obviously also depend on the assumed volatility of the nitric acid – AMP clusters. Another nucleation route may be acid-base reactions with methanesulfonic acid (MSA) when carbon capture is used in coastal environments (Perraud et al., 2024), which should be discussed in the implications section.
Technical corrections:
References:
Perraud, V., Roundtree, K., Morris, P. M., Smith, J. N., and Finlayson-Pitts, B.: Implications for new particle formation in air of the use of monoethanolamine in carbon capture and storage, Phys. Chem. Chem. Phys., 26, 9005−9020, https://doi.org/10.1039/D4CP00316K, 2024.