Overall, I find this an interesting study that adds to our understanding of particle evaporation behavior and the influence of water on particle composition and evaporation behavior. My major comment relates to the definition of “factors” and how the concept of a factor can be consistent with shifting/evolving evaporation profiles between conditions.

L170: It is stated that the evaporation rate in dry SOA particles is slowest owing to “considerable kinetic limitations arising from high particle viscosity.” As currently written, this is stated as a categorical result. However, there has not yet been discussion of the extent to which chemical differences in the dry versus wet particles could lead to volatility changes separate from viscosity changes. In fact, later in the same paragraph the authors note that aqueous-phase processing occurs when the particles are exposed to 80% RH. I suggest it would be useful if the authors work to better separate results from conclusions to first demonstrate that one effect must be more important than another in determining the overall behavior.

L208: The authors state that “Each derived factor constitutes a group of organic compounds with the same thermal desorption behavior.” While this seems like an a priori true statement, I find it difficult to reconcile with the fact that the authors observe different thermal desorption profiles for the same factors between different conditions. Consider AV2 and AV4 in Fig. 3. Or any of the factors in Fig. 4. The peak desorption temperature and profile shapes change between the different conditions, even at the same RH. This is most evident at high RH. It would seem, therefore, that one factor can have more than one desorption profile and thus different thermal desorption behavior. It would be useful if the authors could provide further discussion regarding such differences for factors. I find it very interesting that the desorption profile for a given factor—presumably, a collection of molecules—should change so much.

L215: It would be useful to have further discussion of how the “background” factors were identified. The authors retain only 5 factors in their analysis, suggesting that there are 5 background factors. This seems excessive. What does it mean for these to be “predominately in filter blank measurements?”

Fig. S6: I find these “Kroll diagrams” a bit strange. These assume, presumably, that every ion is a unique molecule and not a fragment, correct? It could be more informative (or at least equally informative) to show one diagram that uses the average OsC and average Cnum for each factor.

Definition of factors as V or D: I find it helpful that the authors have worked to classify factors according to their thermal profiles. I think that this helps with understanding. But it would be useful to hear more about how they made decisions in what might be considered in between or marginal cases. For example, for the AV4 a-pinene factor, the desorption profiles are really, really broad and with much intensity remaining out to very high temperatures. And the fresh high RH SV4 SQTmix factor looks pretty similar to the SD5 factor for high RH RTC. As such, there seems to be some ambiguity in the definitions/assignment and it would be helpful if the authors were to address this further.

L263: The authors note that it is difficult to compare between experiments in terms of absolute signals owing to differences and uncertainty in collected mass. However, the authors presumably know the volume of material collected, and using reasonable estimates of density this should allow for comparability to well within a factor of two. Through normalization to the total signal, the authors do work to make the factor-specific observations quantitatively comparable. Use of a second method to estimate the total mass collected would help to validate the normalization method, as there is an implicit (but as best I can tell unstated) assumption that the CIMS response is the same for all factors.

Fig. 5 and 6: I understand the authors’ motivation to exclude Tdesorb and NCR values in cases for which the desorption profile is not well-formed owing to limited signal. However, for the NCR I would encourage the authors to put perhaps an open symbol at the lowest NCR value (0.1) as a visual indication to the reader that the signal is very small for that particular factor & condition.

L300: Regarding the authors’ discussion of SV4 as an ELVOC and the observation of a major shift in the NCR under humid conditions, I return to my previous comment regarding the identification of this factor as a “V” type to begin with. There is certainly, in my opinion, ambiguity in this identification, which affects the interpretation and determination that this observation is “surprising.” Further discussion here would be helpful. For SV2, I’m a little surprised myself to see this identified as being in the “ELVOC range.” The peak of this factor profile is fully in the “LVOC” range. Further, the volatility really exists as a continuum with the sharp lines drawn for convenience more than reality (although there is, of course, a tie to physical behavior that underlies these distinctions). Only once things go to high RH does the SV2 factor shift to the ELVOC volatility range. But this is a shift from the dry conditions, indicating that perhaps some chemical change has occurred.

L330: The authors note that chemical transformations could cause changes in Tdesorb and thermogram widths with RH for a given factor. But what is a factor even if the chemical composition has changed? Isn’t a factor presumably a collection of molecules (or at least ion signals) that are invariant/grouped together? If chemical changes occur, wouldn’t one expect this to end up as a different factor? Or is the argument here that chemical changes occur, but these chemical changes somehow lead to the same mass spectra/factor as before the chemical changes occurred? I suggest that further discussion is warranted. Similarly, on line 337 the authors note that “some of the compounds grouped into AV3 must have evaporated…or continued to react.” Can it really be “some” of the compounds if the factor itself is a stable/real thing? (Yes, factors are just mathematical constructs, but they still need to be stable, right?) How does one lose some compounds but not others from a factor and still have it come out as the same factor?

Section 3.3.3: Here, the authors aim to provide general understanding of the nature of the chemical processes or shifts in volatility that occur upon humidification. One thing that I think could be useful is if the authors were to further compare between the a-pinene and SQTmix. For example, it seems noteworthy that the “D” type factor identified for both exhibit completely distinct behavior. If anything, I would have intuitively thought that the “decomposition” factors would exhibit similar behavior between both chemical systems. Clearly, intuition is not serving me well here, which is why I think that further compare/contrast would help to strengthen the discussion further.

Li et al. conducted laboratory experiments to investigate the change in volatility and composition during the evaporation of SOA formed from alpha-pinene and a mixture of sesquiterpenes. They conducted two types of experiments, isothermal evaporation and thermo-desorption using the FIGAERO CIMS. They ran the experiments under different RH conditions to probe possible diffusive limitations and water-induced changes in composition. This study is well within the scope of the journal. However, there are some important missing pieces of information that affect the final conclusions. My comments are the following:

1. What were the mass concentrations of these experiments? How variable were the concentrations in dry and high RH experiments? It is unclear if the change in volatility (and composition) was due to the difference in mass loading which to the first order, determines the volatility distribution in the particles.

2. Here particles were generated using an OFR, and the average O/C value (from Table S1) is on the high end compared to the O/C from most of the alpha-pinene SOA generated in chamber experiments. In fact, many studies on SOA viscosity were done using chamber SOA. The authors should discuss the effects of the highly oxidized (and polar) nature of the particles on the evaporation behavior and how would this affect the comparison to other studies.

3. Did the presence of moist on the FIGAERO filter in high RH experiments affect the thermograms? Did the authors conduct any tests to make sure that for a single compound, or a mixture of known compounds, high RH did not change the shape of the thermograms?

4. How was the relationship between Tdesorp and volatility determined (Figure 2a)? Did the authors do any calibration using known compounds?

5. What was the vapor wall loss in the residence time chamber? How did vapor wall loss affect particle volatility and composition in the experiments?

6. Based on Figure 3b, it seems that the authors did not observe O3 and O4 species that are known to be major a-pinene oxidation products (e.g. pinic acid, pinoic acid). What is the reason for that?

This work by Li et al. investigates the changes in volatility and composition of SOA from the oxidation of α-pinene and a sesquiterpene mixture as a result of isothermal evaporation. It details laboratory experiments performed with a FIGAERO-CIMS to obtain volatility information and molecular formulas for species in the SOA. Results from the sesquiterpene mixture were compared to that from α-pinene, and also compared between dry (RH <7%) and wet (RH ~80%) conditions. Positive matrix factorization (PMF) was used to investigate the behavior of individual factors before and after evaporation at the two RH extremes. Unsurprisingly, one result is that the sesquiterpene SOA is more resistant to evaporation than α-pinene, implying lower volatility. They also found that under high RH, more of the signal evaporated. This work is novel, will be of interest to the readers of ACP, and provides valuable information to the community. After addressing the minor general and specific comments below, this manuscript will be suitable for publication in ACP. 

General comments:

1. The use of so many acronyms make the manuscript difficult to follow at times. Some examples include:

• the abbreviation of α-pinene to αpin is unnecessary
• changing STG to ∑TG could make it easier for the reader to remember what this stands for
• RTC for residence time chambers is used to refer to both the physical chamber (lines 94-96, 105, 111) and the experiment type (lines 130, 160, 182, etc) which is confusing. When referring to experiment type does it always refer to a specific evaporation time length?
• CNerror is only used once (line 155).

2. This study could benefit from a more robust discussion of, and comparison to, the several previous studies measuring the evaporation of α-pinene SOA

Specific comments:

Line 78: I think it could be helpful to the reader to 1) specify the mixture of sesquiterpenes, at least with respect to the most dominant species present, and 2) show their structures somewhere.

Paragraph starting line 78: please add your SOA loadings to this paragraph. If there was much more than typical ambient (>~50 ug/m3), can you please comment in the results and discussion section how your results may or may not be affected from partitioning more SVOC to the SOA than would be observed in ambient?

Line 81: Do you think forming the SOA at one RH and evaporating it at a different RH will affect your results? Do you think there is a long or short equilibration time between formation and evaporation RH?

Line 93-96:

• Was there anything in these evaporation chambers to absorb desorbing vapors (for example activated charcoal)? If not, do you think there’s a role for re-partitioning?
• The workflow isn’t obvious—were all particles sent sequentially through each stage laid out in i-iii, or did particles only go into one evaporation chamber? If particles only went into one type of reaction chamber, was the reasoning behind having 3 types to achieve different lengths of evaporation time?
• Do you expect the same evaporation behavior and results in each chamber (i.e. no impacts from air volume to wall surface area, etc)?

Line 94: “a 25 L… chambers”— was one chamber used or multiple? If multiple, were they all identical?

Line 115: saturation vapor pressure or concentration?

Line 120: instead of “the appearance”, “the shape” might be more easily understood, but ok as-is.

Line 145: half the sum thermogram signal or mass?

Line ~190: Can you please discuss why the overall volatility is lower under wet versus dry conditions, yet under these low-volatility wet conditions more signal is lost after evaporation?

Line 200: Couldn’t this be tested by comparing the signal to mass concentration under dry versus wet conditions? If the mass concentration is equivalent, you could determine what percent of signal was lost presumably due to evaporation of high volatility material during the SOA collection time. Or if the mass concentration was different but the signal was the same, you’d have a hint that the SOA under the two conditions may have been made via two different pathways (e.g. reactive uptake at high RH vs. vapor-pressure driven condensation at low RH).

Line 239: I think it makes sense to reference some of the papers on thermal decomposition upon heating with a FIGAERO here (or above).

Line 255: Would help to directly reference the figure in place of saying “above”

Line 265: excl = excluding? Please write out complete word for clarity

Line 270: rational should be rationale

Line 285: “decreases with evolving isothermal…” or “with increasing isothermal…”?

Line 287: if the NCR doesn’t decrease with decreasing VFR couldn’t this also mean that nothing is happening to the compounds, that they are neither being lost or produced?

line 303: can you speculate on other possible loss mechanisms?

Line 307: Is it possible that there is different gas-phase chemistry or SOA formation pathways (i.e. reactive uptake vs. condensation) at high RH, so these species don’t necessarily have to be produced in the particle phase?

Line 319: figure 2, panels a & b only, correct?

Line 331: figure 3, panel a only, correct? And figure 6 panel a?

Line 333: is it possible that there are more compounds grouped into the factor under wet conditions, instead of more signal of the same number of compounds?

Line 393: Should have a reference for α-pinene having largest emissions globally

Line 397: sentence “These findings are generally…” is redundant

Line 740/ Figure 2: Putting the y-axis in panels a & b on the same scale would make these easier to compare.

Key figure: is time on the x-axis for both the left and right figure/schematic?