Kinetic modeling of formation and evaporation of secondary organic aerosol from NO<sub>3</sub> oxidation of pure and mixed monoterpenes

Berkemeier, Thomas; Takeuchi, Masayuki; Eris, Gamze; Ng, Nga L.

doi:https://doi.org/10.5194/acp-20-15513-2020

Articles | Volume 20, issue 24

https://doi.org/10.5194/acp-20-15513-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/acp-20-15513-2020

© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 20, issue 24

Research article

|

14 Dec 2020

Research article |

| 14 Dec 2020

Kinetic modeling of formation and evaporation of secondary organic aerosol from NO₃ oxidation of pure and mixed monoterpenes

Thomas Berkemeier, Masayuki Takeuchi, Gamze Eris, and Nga L. Ng

Download

Final revised paper (published on 14 Dec 2020)
Supplement to the final revised paper
Preprint (discussion started on 27 Jan 2020)
Supplement to the preprint

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'difficult review', Anonymous Referee #1, 26 Feb 2020
RC2: 'Review of Berkemeier et al.', Anonymous Referee #2, 27 Feb 2020
SC1: 'Short comment on Berkemeier et al. “Kinetic modelling of formation and evaporation of SOA from NO3 oxidation of pure and mixed monoterpenes”.', Douglas Day, 23 Mar 2020
AC1: 'Author Comment and Responses', Thomas Berkemeier, 16 Jun 2020

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Thomas Berkemeier on behalf of the Authors (16 Jun 2020)

ED: Referee Nomination & Report Request started (16 Jul 2020) by Andreas Hofzumahaus

RR by Douglas Day (28 Aug 2020)

RR by Anonymous Referee #4 (10 Sep 2020)

Suggestions for revision or reasons for rejection

This manuscript presents a global optimization approach for assigning kinetic parameters to represent secondary organic aerosol (SOA) processes, including gas-phase chemistry, (equilibrium/non-equilibrium) gas/particle partitioning, and particle-phase chemistry. A number of additional processes that affect SOA formation and losses in chamber studies are represented, including diffusion and wall losses of gases and particles. The gas-phase chemical mechanism is based on the Leeds Master Chemical Mechanism, with addition of reactions/products to represent oxidation of the second double bond in limonene by NO3. Parameters are obtained using the Monte Carlo Genetic Algorithm through a two-step inverse fitting procedure. Given the extreme complexity and nonlinearity of the processes that directly and indirectly affect observed SOA mass formation and yields in chambers (and ambient) studies, the problem of representing SOA formation is well suited to optimization methods and statistical modeling. This paper is novel and represents an important advancement in diagnostic modeling of SOA formation.

While many of the represented processes are known to occur, at least under some conditions, many are also poorly constrained. It is thus understood that given the large number of processes represented and parameters optimized in the model, a number of assumptions were needed. That said, the manuscript could be improved by evaluating the model results (parameter ranges and observed sensitivities) more broadly in the context of published literature. That the chamber experiments are generally well represented by the model, and in some cases only well represented with inclusion of specific processes, is important. However, that alone does not necessarily demonstrate the validity of the model approach; i.e., that the derived parameters are reasonable and are in general agreement with other observations when such comparisons can be made. At times the observed sensitivities are attributed to one process or parameter in the model, but a systematic discussion of whether the parameter range(s) are reasonable and those conclusions are more broadly supported in the literature is lacking. Some specific examples are provided below. This is acknowledged to some degree at the end of the paper (the verification of parameter values), but I think could be mitigated with better comparison to existing literature.

While the manuscript is well written from a grammatical perspective, some of the technical writing needs improvement. Some specific examples are provided below. It is my general conclusion that this is an important paper that should be published in ACP, but some remaining technical issues need to be addressed.

Technical comments:
In line 280, the authors note that the number of parameter sets was likely not sufficient given the number of input parameters. What is rule for determining parameter sets, and what would be a more reasonable number?

Using the method introduced by Donahue et al., VBS fits are only well constrained for bins within 1 order of magnitude of the observed SOA levels. For observed SOA mass levels of ~50-80 ug/m3, the assigned bins are outside this range. Is there something unique about the optimization and assignment method used that supports the bin ranges assigned and parameters obtained? The standard deviations shown in Fig. S4 suggest that the distributions vary widely between model runs (and thus are not well constrained). Further, if I understand Figs. 2b,c, the model suggests nearly all of the SOA mass is oligomeric and in the lowest volatility bin. This may explain the relative insensitivity to the initial volatility distribution (i.e., the similar measurement-model agreement achieved despite the significant differences in the volatility distributions obtained). The discussion of Fig. 2c in the text (327) is confusing, as it states that Fig. 2c shows the “actual” volatility distribution in the particle phase. This suggests observational results, but the figure caption indicates model results. Further, the use of “actual” implies a disagreement between Fig. S4 and 2c., but they really represent different processes.

The discussion of the kinetic modeling results for alpha-pinene starting on line 104 is very hard to follow. I think the confusion stems from the statement that the gas-phase dimer “content” is increasing. The mass amount looks like it is staying the same, while the monomers are decreasing (evaporating), and thus the gas-phase dimers represent a greater fraction of the total condensed mass. That is more consistent with partitioning theory and Fig. 3b.

In lines 365, the authors are comparing yields as a function of temperature with experiments from a prior publication (presumably under the same conditions and in the same chamber). It isn’t clear that all of the kinetic processes are the same at 5 deg C as they are at 25 deg C. The rate constants (for oxidation, oligomerization, etc.) should have a dependence on temperature (as noted in the manuscript). Diffusion rates/viscosity may also be affected by the lower temperatures. Thus, unless everything is understood to be the same between this study and the Boyd study, it is misleading to state that the results are not in line with partitioning theory. It Is more accurate to say that considering gas/particle partitioning alone, and the decrease in vapor pressure with temperature, higher SOA yields at lower temperatures would be expected.

One of the things that is notably absent in the discussion of measured yields in this work, including comparisons with previously measured yields, is the role of gas-phase chemistry. The papers out of the Fry group (cited here and also Draper at al. ACP 2015) provide analysis of experimental monoterpene + NO3 SOA data from a mechanistic perspective, and should be used here to strengthen the discussion regarding the measurement-model comparison (particularly lines 385-389). Are the oligomerization rates and product C* values consistent with those summarized in Barsanti et al. J. Phs. Chem. 2017?

In line 386, it is confusing to say that the yield is lower in the -pinene experiment because more precursor was added. The yield is lower because -pinene does not produces as much SOA as limonene under the same conditions, and thus to achieve the same mass loading for a compounds with a lower yield, more precursor had to be added.

Editorial comments:

43: “sink” should be “sinks”

60: The latter “…is missing completely…” is repetitive; suggest to remove that part of the sentence.

83: “allows to infer” should be “allows inference of”

96: “pinene” in “alpha-” should be capitalized

141-142: Description of sequential precursor oxidation experiments is awkward and unclear as written; needs revision.

257: “residue” should be “residual”

321: the second “non-nitrated” should be “nitrated”

343: “constant drift” suggests a physical process, and not a chemical thermodynamics process

Figure S4: This figure was difficult for me to understand. I did not see what it added to the paper.

Figure S5: This figure is confusing as labeled. The insets are labeled “SOA”, but the volatility distributions clearly include products that reside only in the gas-phase.

Figure S6: There is a mismatch between the legend and the traces (black line appears to be total).

Hide

ED: Reconsider after major revisions (10 Sep 2020) by Andreas Hofzumahaus

AR by Thomas Berkemeier on behalf of the Authors (19 Oct 2020) Author's response Manuscript

ED: Publish as is (20 Oct 2020) by Andreas Hofzumahaus

AR by Thomas Berkemeier on behalf of the Authors (26 Oct 2020) Author's response Manuscript

Download

Article (3180 KB)
Full-text XML

Short summary

This paper presents how environmental chamber data of secondary organic aerosol (SOA) formation can be interpreted using kinetic modeling techniques. Utilizing pure and mixed precursor experiments, we show that SOA formation and evaporation can be understood by explicitly treating gas-phase chemistry, gas–particle partitioning, and, notably, particle-phase oligomerization, but some of the non-linear, non-equilibrium effects must be accredited to diffusion limitations in the particle phase.

Kinetic modeling of formation and evaporation of secondary organic aerosol from NO3 oxidation of pure and mixed monoterpenes

Download

Interactive discussion

Peer-review completion

Kinetic modeling of formation and evaporation of secondary organic aerosol from NO₃ oxidation of pure and mixed monoterpenes