Review: Lee et al., “Effects of Aerosol Size and Coating Thickness on the Molecular Detection using Extractive Electrospray Ionization”

Lee et al. present a characterization study of extractive electrospray ionization (EESI) of particles which is a topic of great interest given increased use of EESI to examine particles and their components.  This EESI study of coated and uncoated particles of varied particle diameters is highly valuable given the range of results from existing EESI studies in the literature.  The variety of methods of particle generation and compositions lends insight into the mechanism of extraction over a range of conditions that are greatly informative to the EESI community.  The experimental techniques are robust and well-described and the paper is well written.  The authors have clearly taken care in their experimental techniques, one example being the use of a size selection method that does not suffer from multiply charged particles.

Data are presented to show that Brownian coagulation between the analyte particles and the electrospray (ES) droplets can explain the reported increase in sensitivity for small particles because of increased time for coagulation for those smaller diameters.  Brownian coagulation coefficients are shown to play a dominant role in controlling the sensitivity of EESI for SOA particles formed from pinene ozonolysis.  A variety of particle compositions are shown to be fully extracted but with size dependence.

Main comments and questions are listed below:

1.  The introduction mentions that the analyte particles contain <50% water. The experimental (section 2.3, line 118) describes the use of a silica gel dryer to dry particles, and section 2.4 (line 154) describes their humidification to 60% RH.  Can it be clarified how the 50% water content (by mass?) was determined?

Continuing the idea of particulate water content, this seems to be a large amount of water such that the particles are likely to have much greater internal mobility than dry, solid particles (even though many particles cannot be completely dried).  The presence of particulate water could have a significant effect on their timeframe for solubility or coalescence within the ES droplets.  Since results are contrasted with a comparison paper, Kumbhani et al. 2018, which reports a seemingly large effect of the presence of water in their particles during EESI analysis, it is important to clarify this.  It is not clear if all the particles generated in Lee et al. have this much water, but if so this could be a further difference between results here and in the comparison papers that may be worth noting.

2.  Is it expected that the inorganic salt, NH₄NO₃, would be detected as [NaNO₃+Na]⁺? It seems strange that no ammonium ions or ammoniated adducts are detected.  If there is ion exchange from NH₄NO₃ particles with the NaI that is added to the electrospray solution, this is consistent with full extraction of analyte particles by the electrospray droplets.  But related to the question above, if the NH₄NO₃ was initially dried to lower than 50% water content, do the authors believe this ion exchange would still occur?

3.  I’m afraid I did not follow at first what the Brownian coagulation was referring to, although it is an insightful calculation. May I suggest that the authors use the same or similar wording used in the conclusion earlier on in the manuscript?  Namely, include the description from line 276-277 “the coagulation duration between the ES parent droplets and the analyte particles” somewhere near line 198 where the Brownian coagulation coefficient was first described to clarify what interaction is being examined here.

4.  Line 216 – The sentence states that EESI source A provides a factor of 2 longer residence time but then the next sentence (line 217) says source B has twice the residence time as source A. Can you clarify which one has the longer residence time?

5.  Figure S4 mentions the use of ammonium sulfate particles being coated with pinene oxidation products. Should this be ammonium nitrate?

6.  Figure S6, line 83, has “BCC”, not sure if this is a typo? Also in Figure S6, the last sentence of the caption says that the mass concentrations for levoglucosan and NH₄NO₃ were measured by an LTOF-MS.  Perhaps it would not change the trend of S_{100 nm} values in Fig. S6, but are there ionization efficiencies that affect the mass concentrations of levoglucosan and NH₄NO₃?  There is limited description of how AMS signals were characterized in the main text and supplementary information.

7.  It is quite nice to experimentally change the electrospray parent droplet size. The parent droplet sizes (0.7 – 5.66 um) seem larger than in the papers the authors compare with but I’m not sure this is always true.  For example, Wang et al. (2012) points out that the ESI droplets are usually smaller than sample droplets and that this size is important in examining the mechanism.  Could this be another difference between your studies and comparison papers (the conclusion does not refer back to this difference in drawing on comparisons).

This paper presents a series of carefully planned and executed experiments to explore the sensitivity of the EESI source as a function of particle size and composition, as well as EESI operating conditions. It represents a useful contribution to our understanding of the EESI, especially as it is becoming more commonly used in the measurement of ambient aerosol particles. However, some of the conclusions are not supported by the data and this needs to be corrected before publication.

Specific comments:

Discussion of the trend in normalized sensitivity on pages 5 and 6:

Line 196: Why do you state that the sensitivity reaches a plateau only for EESI Source A?  The one set of data for EESI Source B shows a clear plateau.

Line 204: What are you referring to as a longer period for coagulation? Do you mean the difference in length of interaction region in Source A and in Source B? If so, please state this more clearly. Source A has a 1 mm interaction region and Source B has 0.5 mm, so Source A has the longer period for coagulation. By your argument, Source A should have less of a size dependent sensitivity. The data in Figure 2 shows exactly the opposite. The blue points (Source A) have, taken as a whole, a steeper slope than the red points (Source B). Please draw a conclusion that it is supported by the data.

Lines 206-211:  None of these conclusions about plateaus or deviations are supported by the data. First, rephrase the sentence on lines 206 to 208 to make it clear that it is the calculation that suggests the sensitivity plateaus when the particles are close in size to the ES droplets.  The data does not suggest this at all.  In Figure S6, the data for the largest ES droplets plateaus at the smallest diameter, exactly the opposite of what the calculations suggest. In addition, the sensitivity changes for two components of a single particle, e.g., levo and NO3 or sucrose and NO3, are very different on the sensitivity vs size graphs. For example, NO3 plateaus at ~ 250 nm and turns back up while levo from the same particles does not plateau at all. Sucrose plateaus at ~ 300 nm while NO3 from the same particles does not plateau at all. I don’t think you can draw any conclusions about particle size/droplet size relationships from the data. For the claim about the high deviation in the data above 100 nm, there is no discernible pattern as a function of ES droplet size in Figure S6. Therefore, it does not make sense to attribute the scatter in the data to ES droplet size. Or maybe you are saying that you have no idea what the droplet size is in any experiment. If that is the case, then state that.

Discussion of residence time on page 6:

                Line 218: This statement that Source B has twice the residence time of Source A is not consistent with the schematic in Figure S1 and directly contradicts the preceding sentence. Residence time is not the explanation for the shallower sensitivity dependence of Source B. In addition, this (incorrect) residence time argument was already made in the previous paragraph and there is no need to repeat it here. Please remove this discussion.

Update the abstract:

                Lines 25-26: This sentence needs to be updated once you have revised the discussion sections. You do not demonstrate that the sensitivity dependence varies with ES droplet size or with residence time.

Minor comments:

Lines 58-59: This sentence about water content is confusing – are the ES droplets really >90% water if you are using 50:50 H2O:ACN for the solution? You could just say you are going to call the analyte droplets particles to avoid confusion with the ES droplets. No need to invoke water content.

Line 64: What do you mean by “fragmentation of the analyte”? Isn’t the point of EESI that it does not fragment the analyte so that you get molecular information.

Line 135: The figures should be in the same order in the SI as you call them out in the text. Here you have S5 before S3 or S4.

Line 160: What do you mean by depending on conditions?  What conditions and how do you know what morphology you have? You also say that the morphology will not affect your conclusions, “as discussed below” but you do not discuss morphology at all in Section 3.2. Please add a sentence or two about morphology to the discussion in Section 3.2.

Line 181: The figures should be in the same order as they are called out in the text.  Here you have Figure 3a before Figure 2.

Line 269: What do you mean by dissolution period?

Figure 2 caption: In the text, Source A and B are the TOF and Orbitrap, respectively. Please correct.

Figure S1: Is the only difference between Source A and Source B the length of the gap between the ESI capillary and the transfer tube? Since you don’t do any experiments with the Orbitrap, why show the schematic of it? I would move the inset for Source B to part A of the figure and delete part b of the figure.

Figure S2: In the schematic, you show a denuder and a HEPA filter, but you do not mention the use of the HEPA filter in the description of the experiments. You also don’t mention bypassing the denuder as is shown in the schematic. Maybe you could simplify the red part of the schematic to match the text. Is the red arrow next to the HEPA filter in Figures S2 and S4 going in the wrong direction? Finally, please label the EESI.

Figure S6: Use the same symbols in 6d as in 6c for the same particles.  In the caption, delete the extra “BCC” after the second sentence. Move the sentence about what A and B denote after the sentence describing panels b-d. It would be much easier to compare the data in these four panels if you use the same range on the x and y-axes.

Table S2: You have reversed the Source labels A and B.  Please correct. Why do you have two separate rows for experiments with mixed particles? For example Levo7 and AN2 are the same experiment, so just use one row.

Figure S8: What is the point of this figure? How is Figure S8 related to S9? I think you could skip S8.

Figure S12: This figure is not referenced in the main text. I think you could skip it.

Typographical errors:

                There are many, many typographical errors (missing words, random extra words, misspellings, etc.). The authors should proofread much more carefully before submitting an article.