This manuscript presents aerosol chemical composition and hygroscopicity from offline measurements and comparison with online analysis in Okinawa based on a combination of instruments. The RH-dependent hygroscopicity of sub-micrometer aerosols and their chemical components are investigated. The measurements and data have been made carefully and then the interesting results are presented, especially the comparison between the two methods is helpful to understand whether the offline analysis can be used as an alternative method for aerosol hygroscopicity studies.

However, I have some major concerns about the bias of the offline analysis. And it could be more cautious and thorough when interpreting some results. In addition, there are also some editing issues need that have to be addressed. The authors thus need to make a careful revision and correction to improve the overall quality of the paper for publication in the journal. I would recommend the editor to reconsider the papers only after a major revision by the authors.

Major Comments:

1. Page 4, Line 30, the samples were stored in freezers after sample collection. How to ensure that compounds do not undergo some physical and chemical changes such as condensation and degradation during storage? Have you considered the offline analysis bias caused by storage?
2. Page 9, the authors concluded good agreement between offline and online analyses from the high positive correlations of mass concentrations, but the R² of the mass fractions of sulfate, organics, ammonium, and EC (BC) from offline and online analysis were low. The authors should try to explain this. In addition, the average mass concentrations of sulfate, organics, and ammonium from online measurements were lower, but BC from online analysis was almost equal to those from offline analysis. What is the reason for this?
3. Section 3.2, the authors should add more measured results from offline analysis and compare these with previous studies since many previous online analyses have tried to derive the hygroscopicity in a larger accumulation size above 300 nm. It is best to add more discussions in these paragraphs.
4. Page 13, Line 19, “Results obtained from … indicate that offline aerosol hygroscopicity analysis can be used as an alternative method”, but the coefficients of determination between offline and online results were low (Fig. 6), even less than 0.5. The authors should try to address this and make it convincing.
5. Fig. 7, comparing the hygroscopicity parameter of WSOM and EOM between this work and previous experiments needs to be shown to be much more rigorous. Unless the authors can explicitly show that the experiment setup and estimation method were identical between the different experiments, they cannot make a like-for-like comparison between the different experiments. In addition, the RH conditions of these previous studies should be included in the figure.Minor Comments:
6.  

1. Page 5, Line 30, The residence time in the heated region should be compared with that for other systems.
2. A schematic of the experimental set-up would help.
3. The retrieval method of the HTDMA data should be included in section 2.2.
4. Fig. 1, add the mass fractions from online analysis.
5. Fig. S6, looks like one point is missing in the Fig. S6(c).
6. I suggest adding more 'raw' measurement time series (e.g. in the supplementary information). For example, time series of hygroscopicity parameter over the experiment.
7. Page 10, I was lost from line 7 to 11 when the aerosol concentrations and fractions are discussed. Is the discussion still based on Fig. 2 or other figures in the manuscript?

This manuscript performed offline analysis of both chemical composition and hygroscopicity of submicrometer particles and then compared the results with those from online measurements. In traditional view, the offline filter analysis is tedious and time-consumed but this work presents a very interesting approach to compliment the online results. It is worthy of exploration since it might provide a more affordable way to measure the aerosol hygroscopicity, especially when online instruments are not available and offline analytical analysis can be easily accessible. Although it is found very interesting, the manuscript needs to be improved and the following issues should be fully resolved, before it can be publishable.

Major comments:

1. Although the authors made great efforts to elaborate how to avoid artifacts and those artifacts indeed can be minimized. However, the authors should also give some more details on how errors are propagated, for example, the errors generated from exaction of organics from filters and the nebulization of organics, subsequent growth measurement etc. To what extent, the errors are associated with the offline analysis presented in this study?
2. Recent literatures present a growing interest in the effects of surfactants on modification of surface tension of the particles and hence affect the hygroscopicity of the aerosol particles, including measurement inland, at sea, etc. The use of surface tension of pure water neglecting the above-mentioned effect and apparently it is problematic. How organics affect the surface properties of the particles and what surface tension range of values would be estimated from this study? Could the authors elaborate a bit more on this aspect?
3. It is quite interesting to see how the AMS spectra from extracted WSM are different from the average online AMS spectra for the same time period. Could the authors do some comparisons and present some results for the spectra between the two methods? What would make those differences?
4. Following the above question, how if the authors performed some factorization analyses based on the offline AMS method and then compared with those from online AMS data if available. Do they have the similar results?

Minor:

1. Quite a few instruments were employed to perform the offline analyses, corresponding to many chemical components as presented in the paper. It might be beneficial to the readers if the authors can provide a table showing all the measured components with their measurement techniques.
2. Line 19 on p2, “serve as”?
3. Line 11 on p7, “were obtained”?
4. Line 15 on p9, explain why values of 1.8 and 1.2 were used.
5. Line 29 on p11, “under highly acidic conditions…”?
6. Line 15 on p13, how large is large? Do you have a criterion?
7. Line 1 on p14, “at a supersite”; Line 4, “suggest the importance on considering” or something better instead of using “to”; Line 14, “estimated to be”? Lines 31-32, this sentence seems awkward, please change it.
8. Lines 3&18 on p15, using “as” here is not right, please rephrase the sentences. Line 30, “causal”?
9. Line 8 on p16, please rephrase “another important point is that….”

The manuscript by Deng et al. entitled reports the detailed comparison of online and offline analysis for submicron particles collected at Okinawa island, Japan. By comparing the online and offline analysis data, they demonstrated that the offline analysis of aerosol samples by the AMS can quantitatively be conducted. It is a good demonstration about the usefulness of offline AMS analysis. The research was carefully conducted. The manuscript is well organized. I suggest publication of the manuscript after addressing the following comments.

Chemical characteristics of OA

The manuscript compares mass concentrations of chemical species, especially focusing on compounds that are measurable by the AMS. I wonder if the mass spectra of the organic material for online and offline agree each other. The authors compare O:C ratios in Figure 3. It will also be useful if H:C ratios are provided.

Predictions hygroscopic properties

The authors employed E-AIM for predicting hygroscopic properties of particles, including phase transition phenomenon. Influence of organic compounds on hygroscopic growth is discussed in the main text. It will be good if the potential influence of organic compounds on deliquescence/efflorescence phenomena were also to be discussed.

Interpretation of hygroscopicity

In Figure 8, the authors compare the values of kappa with the mass fractions of wsom and ammonium-sulfate ratio. However, the data are scattered, especially at higher RH, suggesting that some other factors might also be influencing hygroscopicity. It would be better if the authors can provide some ideas on it.