the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Jul 2026
Coupling Liquid Chromatography with Aerosol Mass Spectrometry to Enhance Chemically Resolved Aerosol Analysis
Abstract. We present a novel analytical method that links liquid chromatography with aerosol mass spectrometry (LC-AMS) to separate and chemically characterize soluble constituents in atmospheric samples. This method is demonstrated here using ion chromatography (IC) to resolve ionic species prior to AMS detection. As the IC effluent is analyzed by the AMS, quantifiable ion signals and reproducible mass spectra are obtained every second throughout the chromatographic run. These chromatographically resolved AMS data yield quantitative information and chemical fingerprints for distinct and partially overlapping chromatographic features. The method was evaluated using standard mixtures of inorganic anions and cations, as well as atmospherically significant organic compounds (e.g., levoglucosan and organic acids). The IC-AMS chromatograms exhibited stable background signals, which permitted reliable integration of peak regions for quantification of the AMS-derived ions. Isotopically labeled ammonium sulfate ((NH4)234SO4) was used as an internal standard to account for variations in aerosol generation efficiency and AMS sensitivity. Low-volatility compounds, such as ammonium sulfate and sucrose, showed recoveries close to 100 %. Application to an ambient aerosol sample collected during a wildfire event demonstrated that LC-AMS offers improved chemical resolution that can better characterize aerosol sources, atmospheric processing, and climate- and health-relevant properties.
- Preprint
(2385 KB) - Metadata XML
-
Supplement
(988 KB) - BibTeX
- EndNote
Status: open (until 13 Oct 2026)
- RC1: 'Comment on egusphere-2026-3483', Anonymous Referee #1, 31 Aug 2026 reply
-
RC2: 'Comment on egusphere-2026-3483', Anonymous Referee #2, 18 Sep 2026
reply
Review of “Coupling liquid chromatography with aerosol mass spectrometry to enhance chemically resolved aerosol analysis” by Farley et al., doi:10.5194/egusphere-2026-3483
This paper describes using an ion chromatography (IC) system to separate species prior to analysis with an aerosol mass spectrometer (AMS). The liquid chromatography with aerosol mass spectrometry (LC-AMS) technique is a novel application yet akin to downstream mass spectrometry analysis of any chromatography system. Descriptions of the method are fairly clear. However, the demonstrations of the effectiveness and quantitative ability of the LC-AMS system are deficient in details. In order for such a technique to be useful, the two properties of species separation and quantitation should be better established. The recommendations below should be made to the manuscript to make it more acceptable for publication in Atmospheric Measurements and Techniques.
With regard to separation, the three, individual organic standard species used in Figure 4 show a good IC separation between them. The m/z ions in Figure 4, however, are different and it is not obvious what the advantage is for separation. The mass spectra of the individual compounds in Figure S4 have different scales for m/z and do not have ion peaks identified, making it difficult to tell from Figure 4S where overlap in the mass spectra occurs and why separation is important. One feature of the AMS is that there is a lot of fragmentation in the instrument so that the mass spectra are sometimes very similar for different compounds. By eye, it looks like there are several overlapping m/z between the three pure organic compounds shown in this paper. For example, m/z 73 for C3H5O2+ is one of the two ion peaks shown in Figure 4 for levoglucosan and is clear in Figure S4. It is often used as a marker peak for the presence of levoglucosan in AMS ambient data. Yet, m/z 73 for the ion with two oxygens (C3H5O2+?) also appears as part of the peak at m/z 73 in the spectrum for glycolic acid. Is it an impurity or part of the spectrum for pure glycolic acid? Perhaps by looking at the retention times of m/z 73 (or C3H5O2+) it could be resolved? In addition, the ion peaks at m/z 29 and 31 as well as the grouping of m/z 43 – 46 seem to have varying ion peak intensities between the three compounds, making it difficult to identify which of the three compounds generates them. With a mixture of the three standards sampled without the IC separation, it’s expected that the spectrum in the AMS would be quite complicated. The case for separation by IC could be made stronger by showing this mass spectrum of the mixture (without the IC column), using the individual spectra in Figure S4 to point out some of the ions that are in common, then showing m/z retention time traces for these ions after separation with the IC column. This could be shown either similarly to Figure 5 where the mixture is put through the column or by adding the common ions for the individual compounds to Figure 4. Either way, it would better present the complexity of the unseparated mass spectrum for the three standards and how the common ions for the individual species are separated with the IC.
A more complete description of the changes to the solution composition during the chromatography process is needed. Since primarily anion exchange is discussed in this paper, an example would be of what happens to a solution composed of something like sodium nitrate compared to ammonium nitrate. Without the IC column, the response by an AMS would be different due to the different relative volatilities of these two compounds. For samples with the same concentrations of nitrate, the anion exchange should not be affected by the cation associated with nitrate. Does the cation eluting out of the column associated with nitrate change? Does it make it more or less detectable with the AMS? The paper implies that the species eluting is the same as what was in the solution and suggests that nitrate originally from the ammonium nitrate solution was volatilizing. Would sodium nitrate elute if was used instead of ammonium nitrate? Would calibration plots with sodium nitrate look the same as the ammonium nitrate in Figure 2 or more along the 1:1 line?
The discussion of quantification was not very clear. It is mentioned that isotopically labeled 34S was used as a standard, and AMS peak areas for other species were essentially normalized with it for the response of the AMS to other species. Calibration plots of conductivity versus solution concentrations are typically quite linear. Calibration plots of AMS peaks are also linear with the mass sampled, as long as the mass sampled is well-known (e.g., after a particle sizer with a particle counter and there are no “doubly-charged” particles) and lens transmission is not an issue. So, in theory this is appropriate. However, I’m curious how this would work for varying concentrations of solutions, which might generate different size distributions in the nebulizer and potentially have AMS lens transmission differences. Is the isotopically labeled 34S concentration changed when the solution concentration is changed so that the relative amounts of each species are independent of the size distributions coming out of the nebulizer?
There was also a bit of vague text describing why the AMS concentrations for nitrate and sulfate in the wildfire sample were higher than measured with the IC. If the species were organic nitrate and sulfate, why would they elute from the column at the same time as the inorganic nitrate and sulfate? They are unlikely to be separated by the anion exchange column and with the AMS after the column the AMS ion peak areas should be only a measurement for the inorganic species. Without the column, the AMS mass concentrations for the organic + inorganic species could be higher than the inorganic species separated and measured with conductivity. There are some AMS/IC comparison papers discussing this. Hence, the reasoning for higher AMS sulfate and nitrate in the wildfire sample after the IC separation was not very robust.
Minor omission: The citations for Canagaratna et al., 2015 and Kiland et al., 2019 are not in the reference list.
Citation: https://doi.org/10.5194/egusphere-2026-3483-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 177 | 63 | 24 | 264 | 40 | 23 | 19 |
- HTML: 177
- PDF: 63
- XML: 24
- Total: 264
- Supplement: 40
- BibTeX: 23
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
In this manuscript, Farley et al. describe a method for coupling ion chromatography (IC) with aerosol mass spectrometry (AMS). Because AMS employs electron ionization and operates under high-vacuum conditions, it is inherently incompatible with direct liquid sample introduction. The authors address this challenge by incorporating a nebulizer, cyclone, and diffusion dryer to convert the liquid effluent into an aerosol stream suitable for AMS analysis. Through this approach, they achieve substantially improved chemical resolution, particularly for organic compounds, compared with conventional AMS measurements, which are generally limited to providing bulk chemical composition or broad chemical class information. This development represents a potentially valuable advance in extending the molecular level characterization capabilities of AMS-based analyses. While the manuscript falls within the scope of AMT, several major issues need to be addressed before I can recommend its publication. Most importantly, the mechanism by which seemingly non-ionic organic compounds are retained and separated on the ion chromatography column requires a much clearer explanation. As outlined below, the current discussion does not adequately justify the reported separations or their underlying chemistry.
Major comments
Line 203: I am a little surprised that the 14N15N+ ion can interfere with CHO+ ion at m/z 29, considering that the natural abundance of 15N is only about 0.37~0.38%. Where does it originate from? In my experience, column bleed typically produces a smoother and more consistent chromatographic background than the pulsed noise pattern shown in Fig. 4b, making quaternary ammonium groups derived from column materials a less likely source. Could the authors add a brief discussion of the possible origin of this interference and explain the observed background pattern?
Line 225: I wonder what type of organic nitrate the authors have in mind when they attribute the higher than expected nitrate signal in the IC-AMS to the co-elution of organonitrates with inorganic nitrate. Unlike inorganic nitrate, organonitrates generally do not ionize in solution, which makes their separation on an IC column difficult. Even if organonitrates are somehow ionized, their retention times are likely to be very different from that of inorganic nitrate. Could the authors explain the mechanism by which organonitrates could increase the nitrate signal in IC-AMS? In particular, it would be helpful to clarify how co-elution with inorganic nitrate is expected to occur and whether there is experimental evidence supporting this interpretation.
Line 239: I wonder how carbonyl compounds and esters can be separated using an anion-exchange column. Do these compounds form anionic species under the chromatographic conditions employed in this study? If so, could the authors clarify the chemical structures of the species being retained and separated?
Line 265: The coupling of LC with EI-MS predates the development of the widely used atmospheric pressure ionization techniques such as ESI and APCI. However, coupling LC to an EI source is intrinsically challenging because the high eluent flow rates and the relatively low volatility of typical reversed-phase and HILIC eluent and additives are fundamentally incompatible with the high vacuum conditions required for EI operation. Considerable efforts were made to develop LC/EI-MS interfaces, particularly up to the early 2000s. Nevertheless, the technique has seen limited acceptance as a practical analytical tool, largely because its advantages over ESI and APCI are confined to relatively niche applications.
I therefore suggest that the authors either provide a more comprehensive historical perspective on LC/EI-MS, including a discussion of previously developed interfaces and a clear explanation of how the approach described in this manuscript overcomes their limitations, or omit this section altogether. Without such context, the significance and novelty of the method are difficult to assess.
Technical corrections
Line 55: Electron Impact should be replaced with Electron Ionization, as the latter is the IUPAC-recommended term.
Line 149: A space is missing between "1" and "ppm" in NaBr (1ppm).
Line 196: It has one extra period.