The authors present two fully characterized complementary methods to quantify SOA marker compounds from filter samples. In this work, a number of target compounds that were previously analyzed in dedicated studies were combined and the instrument response was systematically investigated. The manuscript provides figures of merit for all investigated compounds and the instrument responses were well characterized. Such a combination of methods contributes to the better characterization of SOA composition and the comparable quantification of these marker compounds. The paper is logically structured and well written. I have, however, a few concerns about some of the numbers presented. In my opinion the paper can be suitable for publication in Atmospheric Measurement Techniques once a few issues have been addressed.

1. The authors applied two previously used LC methods and combined both to test a new method. Although the authors claimed that most of the target compounds eluted after 20 min of the second method, the new method was chosen to run for 45 min with a comparable gradient, which seems quite arbitrary. Did the authors test carry-over effects for the shorter LC methods or the effect of longer flushing times on TIC and repeatability?
2. More to that point, the authors describe in section 3.1.4, that the instrument signal response decreased during a continuous sequence by more that 70%. I am quite concerned about this finding, as the quantification is also depending on this response. Do the authors have an explanation for this behavior and were some parameters tested to improve on that? The authors claim that there is no carry-over effect, as none of the analytes were found in subsequent blank samples, but there seems to be some sort of matrix effect during the ionization. The authors need to provide more details here.
3. In line 329 the authors describe, that for analytes with higher variability, the third replicate was not considered due to loss over time, but it seems to me from figure 2, that due to the loss of signal, the second and third replicate of a sample show lower variability. Can the authors clarify how this effected the analysis?
4. The GC method applied by the authors was changed from previously used methods. Namely the SPE step was replaced by liquid extraction and direct injection, which resulted in higher LODs compared to the previous method. The authors note that in this work a wider range of compounds were analyzed, but did the authors test the markers investigated in this work with SPE injection?
5. For the GC/MS method, BSTFA is used as a derivatization agent for alcohols. Perhaps the authors could briefly discuss the possibility of using other derivatization agents. This could specifically improve the detection of other compound classes, such as carbonyls.
6. In the conclusion section the authors compare the LODs for the reported method with literature values. As many previously reported values are much lower, it would be worth highlighting the benefit of this combined method more, to emphasize the value from being able to analyze different compounds classes in one combined protocol.

Technical comments:

L.47 should read “nitrooxy organosulfate markers”

L.240 The authors should rephrase the last sentence of that paragraph

L.254 should read “the instrument variability between three randomly injected replicates of the mixture solution without filter extraction was less than 21% for all target compounds.”

L.274 should read “Signals were also higher than for sampled using inserts,…”

L.277 the authors should emphasize better that this was a hypothetical issue that they tested for and can decline it as a source of variability.

L.327 should read “is reported in Table 2”

L.359 unclear reference with “section 0”

L.381 ff The authors should rephrase that sentence as it is not clear which conditions are referred to

L.402 I suggest “best representation of real samples” instead of “the maximum equivalent time”

L.435 should read “ offered the advantage of detecting phenol compounds at higher sensitivity”

L.487 I would suggest “were about twice as high as”

L.502 The use of “systematically” in the context of three individual samples is a bit unclear

Review report for “Two optimized methods for the quantification of anthropogenic and biogenic markers in aerosol samples using liquid chromatography mass spectrometry and gas chromatography mass spectrometry”

Summary

This manuscript presented two analytical methods (UPLC-HRMS and GC-MS) to quantify 23 organic aerosol markers from biogenic and anthropogenic sources, and their application in ambient aerosol samples. The key finding of this study is that the two analytical methods were demonstrated to show complementary capability for comprehensive identification and quantification of many (23 compounds in this study) organic markers with diverse structures that covering both biogenic and anthropogenic sources. Five organic markers were identified by both analytical methods and showed some level of consistency in the ambient aerosol samples.

The text is quite easy to follow. The very unstable response from the UPLC-HRMS instrument is quite surprise for me, which is very usually. If this turns out to be an issue due to poor instrument performance specifically, then the authors should clarify this point and be a bit cautious for generalization. Overall, the two analytical methods used in this study are not new, but offer some insights for maximizing the identification capability of atmospheric organic markers and expanding their application, especially for some aerosol samples with mixing emission sources where many targeted organic markers are of interest. I would recommend this manuscript to be published in AMT after the following major concerns are addressed and incorporated, if these comments are helpful for improving the manuscript.

Major concerns

1. Usually the recovery range within 80-120% are good and 70-130% are overall acceptable. Can the authors discuss why we need to care about recovery and what are the major factors affecting recovery, and what is the outcome of recovery on quantification and therefore source analysis? By the way, I would suggest to use the ranges (e.g. 80-120% or 70-130%) for recovery throughout the manuscript, rather than recovery above a certain value

2. Line 65-75 A range of previous literatures were summarized in terms of LODs and recovery, and these are very boring. But the main motivation of this work is still not clear to me, especially I cannot see the uniqueness of this study. I would strongly suggest to add some information in the introduction to address why we need a new method as presented in this study.

3. The sample extraction and chemical analysis for ambient aerosol samples presented in this study is done without considering the matrix effect. Would matrix effect be considered as a major issue especially the ambient aerosol samples are highly complex mixtures and may affect their LOD, recovery and quantification?  Could the calibration curves established from standards are actually not applicable for ambient aerosol samples due to matrix effect?

4. To minimize the matrix effect, a common pre-treatment for complex samples is to use solid phase extraction (SPE) to remove some inorganics and extremely high molecular weight compounds. Could some efforts be made to compare and evaluate the results for ambient aerosol samples that were treated with and without SPE?

5. I never work with ion mobility spectra (IMS) but my understanding is that IMS provides another dimension to distinguish isomers. Since UPLC is already able to separate isomers, I wonder whether ion mobility spectra are a bit redundant in this study, especially I only found two isomers among the 23 organic markers. I never found any figure related to IMS, and can any ion mobility spectra be displayed?  Even without CCS, relying on RT and m/z is enough for compound identification, isn’t it?

6. Line 210: The purpose of adding 0.1% formic acid can be added and explained. In addition, is mobile A referring to 0.1% formic acid in water and mobile B referring to 0.1% formic acid in methanol? Or mobile A referring to 0.1% formic acid in water and mobile B referring to only methanol? Please clarify.

7. Line 210-220 Two types of extraction solvents (50:50 acetonitrile and water; 50:50 methanol and water) and two types of mobile phase (water + acetonitrile; water + methanol), in total gives a combination of four. Would the extraction solvent (50:50 acetonitrile and water) be more suitable for water plus acetonitrile mobile phase combination rather than water plus methanol?

8. Line 230 the authors say “As the use of 100% organic solvent to prepare the standard solutions under analysis may negatively influence the peak shape and thus prevents proper quantification, we selected 50/50 ultrapure water/organic solvents for the preparation of standard solutions”. My experience is that this depends on injection volume, and injection of pure organic solvent does not affect the peak shape if the injection volume is not very high (e.g. 5 μL). In addition, the choice of extraction solvents either pure organic solvent or mixture of water and organic solvent are certainly feasible, which should be primary determined by the soluability and stablility of the sample compounds. Please clarify this point.

9. Line 230-235 What is the stability of these organic markers in extracted solvent (50:50)?

10. Sec 3.1.2 the variability among three replicates seems to be ramdom with and without extraction. However, their ratios (with / without extraction) should be compared in a separate figure, and the compound response seem lowering by one order of magnitude with filter extraction. I never have this kind of issue for filters spiked by standards. What is the reason for such large reduced response? Also, the ratio (with / without extraction) is used to determine the recovery for individual standard, isn’t it? But then the recovery is 10% and contradict with the summary table. Or the calibration curves were actually determined from filters spiked by standards following extraction procedure, which then used to determine the recovery? This is not a common procedure to determine recovery and quite confusing for me

11. Sect 3.1.3 I feel quite surprised for the large difference with and without insert, as well as the poor linear response as shown in Fig. S4. This usually should not be an issue as it is very common to use an insert in the vial for concentrated sample with low liquid volume. This is probably an issue dependent on specific user and instrument device. Could the material difference (wall of vial vs. insert) affect the ionization? Also, I notice some values are extremely low, is the instrument response in this study referring to fitted peak area from selected ion chromatogram?

12. Sect 3.1.4 Again, as an experience UPLC-HRMS user, I feel very surprising for the 70% signal loss after 25h for SST solution. This suggests the performance of the instrument is very unstable. This should be very rare and must be dependent by specific instrument. I wonder whether the authors test the standards and between every 24h, since these organic makers especially carboxylic acids are usually stable for at least weeks. Can the authors show time series of individual standards’ response over several days by repeating the injection?

13. Line the authors say “calibrations overtime where performed during analysis to account for the signal stability.” UPLC-HRMS can be operated 24/7 and calibration overtime might be a disaster for quantification. Probably this is a poor instrument that should not be used for accurate quantification purpose, I think

14. Sec 3.3.1 Those text that summarizing or repeating the LODs are really boring. I would suggest to improve the readability of this section and reduce a bit of the text

15. Among the 23 organic markers, only 14 were analyzed by UPLC-HRMS. Some other compounds (e.g. o-toluic acid) are expected to be detected by UPLC-HRMS. I wonder whether the authors made some efforts to analyze remaining 9 compounds by UPLC-HRMS?

Minor comments

1. For table 1, I would suggest to include UPLC and GC info for their target compounds.

2. Aromatics are large compound class. Can these aromatic SOA markers in table 1 be assigned to specific aromatic precursor, e.g. toluene and etc?

3. Can Figure 6 also be displayed in 1:1 plot (e.g. x axis vs. y axis referring to UPLC vs. GC method)? By the way, among the five common compounds determined by both methods, only pinic acid and pinonic acid found in ambient aerosol samples?

4. In Figure S2, it seems MBTCA showing two overlap runs in one figure? I would suggest to label all peaks for Fig. S1-S3 as I found some peaks are not labeled and I guess these could be labeled as impurity?

5. Line 140 “section 0” must be a typo

6. In Table S1, I found there are 24 organic makers. Since 23 compounds were studied in the main text, Is there any mistake here? Some text are not displayed in the middle in table S1, which should be fixed