the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Molecular and seasonal characteristics of organic vapors in urban Beijing: insights from Vocus-PTR measurements
Abstract. Understanding the compositions and evolution of atmospheric organic vapors is crucial for exploring their impact on air quality. However, the molecular and seasonal characteristics of organic vapors in urban areas, with complex anthropogenic emissions and high variability, remain inadequately understood. In this study, we conducted measurements in urban Beijing during 2021–2022 covering four seasons using a Vocus-PTR, an improved Proton Transfer Reaction-Mass Spectrometry (PTR-MS). During the measurement period, a total of 895 peaks are observed, and 543 of them can be assigned to formulas. The contribution of CxHyOz species is most significant, which compose up to 53.7 % of the number and 76.0 % of the mass of total organics. With enhanced sensitivity and mass resolution, various sub-ppt level species and organics with multiple oxygens (≥3) were discovered. When counting the species number, 42.2 % of the organics measured are at sub-ppt level and 37.8 % of the species contain 3–8 oxygens. Organic vapors with multiple oxygens mainly consist of intermediate volatility and semi-volatile compounds, and many of them are found to be the multi-generational oxidation products of various volatile organic precursors. In summer, the fast photooxidation process generates organic vapors with multiple oxygens, and leads to an increase in both their concentration and proportion. While in other seasons, the variations of organic vapors with multiple oxygens are closely correlated with those of organic vapors with 1–2 oxygens, which could be heavily influenced by primary emissions. Organic vapors with low oxygen content (≤ 2 oxygens) are comparable to the results obtained by traditional PTR-MS measurements in both urban Beijing and neighboring regions.
- Preprint
(1814 KB) - Metadata XML
-
Supplement
(4199 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
CC1: 'Comment on egusphere-2024-1325', Sachin Mishra, 31 May 2024
1. In Figure S3, the transmission efficiency of C8 aromatics (C8H11+) is greater than 1. Does the authors have any explanation for this?
2. What were the Limit of detection (LoD) values of the VOCs containing more than 6 oxygen atoms?
Citation: https://doi.org/10.5194/egusphere-2024-1325-CC1 -
AC1: 'Reply on CC1', Zhaojin An, 06 Sep 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1325/egusphere-2024-1325-AC1-supplement.pdf
-
AC1: 'Reply on CC1', Zhaojin An, 06 Sep 2024
-
RC1: 'Comment on egusphere-2024-1325', Anonymous Referee #1, 30 Jun 2024
Scientific significance:
In my opinion the significance of this paper is good. It shows the differences of VOC emissions for different seasons and also analyzes the influence of day and night times on VOC emissions.
Scientific quality:
The scientific quality is ok, however the errors of the measurements need to be included and better presented.
Presentation quality:
The presentation quality is good but can be improved by explaining what the influence of OSc and DBE is. And why it is important for this paper. It is explained in detail how it is calculated (which in my opinion can also be placed in the supporting material part) but the influence on the atmosphere is not made clear enough.
General comments:
First of all, you need to clarify the term concentration. This term isn’t used correctly in the whole paper. A concentration is defined as mass/volume what you measured with you Vocus is a mixing ratio (in this case a volume mixing ratio). You correctly used the unit ppt/ppb. However, to further clarify that you are talking about volume mixing-ratios you can either say it once and than say you will use ppb/ppt as a unit of volume mixing ratios or you can use.
Introduction
You don’t need to include all the different PTRs. Here just make clear what the advantage is (higher sensitivity and lower detection limits), how it is achieved (“incorporating radio frequency electric fields to focus ions”) and the disadvantage (lighter ions are cut off to protect the detector from overloading). Another disadvantage compared to GC and other methods is that you can’t be sure on the exact compound. You only get the information of the exact mass and from this you get information on a sum formular. However, you don’t know anything about the functional groups etc. (no chance to differentiate between ketones and aldehydes).
Why are you only comparing data from PTR, if you want to see higher oxidized compounds people used other methods. (Iodide-CIMS, nitrate-CIMS…). However I don’t know if studies were conducted with those techniques in the past in this way. If there were studies, just include some examples.
2.1.
How do you account for fragments, you seem to use a quite hard setting. What is your E/N? I know it is hard to calculate but you can find some help in this publication: Jensen, A., Koss, A. R., Hales, R., and de Gouw, J. A.: Measurements of VOCs in ambient air by Vocus PTR-TOF-MS: calibrations, instrument background corrections, and introducing a PTR Data Toolkit, Atmos. Meas. Tech., 16, 5261–5285, doi:10.5194/amt-16-5261-2023, 2023.
Was a heating installed around your inletline and was it kept constant? If not, this might also explain your observation of less IVOCs and SVOCs in winter times.
Were the meteorological parameters somehow included in your analysis?
2.2.
Your cut off is above the mentioned 35 amu. This needs to be told here. In the supplementary information the equations of linearity and transmission curve would also be a nice add on. Additionally, I don’t like the idea of using the mean of those three compounds (supplementary). I understand that you had to exclude the others due to your transmission curve. However, the error will be too low compared to the error that is expected if you only have 3 compounds with which you actually had to get the linearity alone. It is not always true that the offset is “0” which you claimed to be true. If you had used a softer setting you might not be able to detect compounds with a low k-rate. Therefore, the offset might even be negative. This error could be minimized by using compounds with higher and lower k-rates. Here however, all compounds had nearly the same k-rate. Luckily the k-rates were the k-rates that were by default anyhow used for most of the compounds. I would suggest to at least make this fit with those three compounds and if the error is high you need to show this.
The fragmentation of C10H17+ would be nice to see, the most abundant fragment is C9H6+. (to have an idea on the fragmentaion strength; however, this is only helpful not mandatory)
Make clear why you use DBE and OSc. What do you expect and what does it say? If you answer this, your analysis part will be easier to understand.
Line 220: there is an additional box in the text
Paragraph 328 ff
Isoprene is a bad example when ozone is present. In (https://amt.copernicus.org/articles/16/1179/2023/amt-16-1179-2023.pdf) it is described that oxidized compounds can fragment in the ion source of a PTR (also Vocus) and land on the exact mass as isoprene does. Therefore, the isoprene signal can be overestimated.
Paragraph 355 ff
Cold inlet line might also explain lower SVOC mixing ratios
“Day time cluster”
Do I understand correctly, all VOCs increase at 6 am? There should be seasonal changes (due to changing light conditions), or is there another source (e.g. traffic, factories?)
Is it possible to check for inversion layer and/or boundary layer. Especially, in winter this can decrease the efficiency of dilution. (meteorological data)
Line 457 style no “the” in front of winter
Line 459 style better: in winter in Beijing during the last few years
Line 466 just keep in mind that there will be fragments on the isoprene mass…
Supporting information
Figure S3 as mentioned above the compounds C8H11+, C9H13+ and C7H9+ would be needed to fit the k-rate to sensitivity line and not just the mean of all three slopes.
Figure S9 Which vmr is plotted here? The sum of all? What does this graph say? You see less when it’s cold? (explain your figure in a few sentences)
Figure S10 Which vmr is plotted against which? I assume it’s sum of nighttime cluster (x-axis) against sum of CxHyO1-2 compounds?
Table S2 under the table a "Benzene" is missing (1,3,-dichloro-) no “,” between 1,1-dichloro- and benzene (1,1-dichloro-benzene). Why are those compounds not included? They are quite heavy and it would definitively help to get a better idea on your k-rate to sensitivity plot and also in your transmission curve plot.
Citation: https://doi.org/10.5194/egusphere-2024-1325-RC1 -
AC2: 'Reply on RC1', Zhaojin An, 06 Sep 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1325/egusphere-2024-1325-AC2-supplement.pdf
-
AC2: 'Reply on RC1', Zhaojin An, 06 Sep 2024
-
RC2: 'Comment on egusphere-2024-1325', Anonymous Referee #2, 17 Jul 2024
This paper presents an analysis of highly oxygenated molecules measured by a Vocus PTR-ToF-MS in Beijing for one year. The authors present a seasonal analysis of their concentrations in addition to a cluster analysis which highlights which times of day certain types of highly oxygenated VOCs are present. They present relevant properties such as DBE, number of oxygens, and volatility.
While I believe that the data is interesting and should be published, the paper and analysis requires major edits before it can be considered for publication in ACP. Most importantly, the authors should improve their quantification techniques so they are more applicable to the molecules in question and include a more thorough discussion of limitations and uncertainties. The paper is also lacking proper justification for studying these highly oxygenated molecules. I am left with key contextual questions such as (1) What fraction of the measured concentration are these highly oxygenated molecules? (2) What fraction of the calculated OH reactivity or SOA formation potential are these highly oxygenated molecules? (3) Why should we focus on them? A very quick google scholar search of ‘highly oxygenated molecules’ reveals that there is more research on these molecules and their importance in SOA formation that suggested in the introduction. I am not an expert on this topic, so I would like to see more justification for studying these species with such low concentrations. In addition to these technical and context-related concerns, there are many grammatical errors in the paper, so it should be carefully edited with that in mind, as well.
For the main reasons discussed in the prior paragraph, and after reviewing my general and specific comments below, I think the paper should be reconsidered after major revisions. I do encourage the authors to strongly consider my suggestions as well as the other reviewers suggestions for improving the paper and re-submit once they are addressed.
Abstract:
You switched from past to present tense a few times. Please stay consistent throughout the abstract and consider reporting your findings in the past tense in the abstract.
Introduction:
There is too much discussion on different PTR techniques for a paper whose results aren’t related to method development. Keep discussion on PTR methods in the introduction to one paragraph maximum, maybe moving some of the discussion on PTR to the methods.
Consider adding more background on oxygenated + highly oxygenated VOCs and why studying them is needed – this would frame the results of the paper more effectively. Specifically, I believe there should be more explanation in the introduction on why you are studying ‘organic vapors with low mixing ratios.’
There is far more recent published literature on urban VOCs than you suggest. See many recent papers published by Karl, Coggon, Pfannerstill, Gkatzelis, Acton, etc. etc.
Methods:
For a results section focused almost exclusively on highly oxygenated molecules, the methods section is lacking discussion of their quantification. How feasible is it to calibrate your highly oxygenated VOCs using the sensitivity of three aromatic VOCs? Have you attempted to quantify any highly oxygenated VOCs or at least determine their fragmentation ratios?
Do you expect any of your reported formulae are fragments or water clusters? The Vocus is prone to high levels of fragmentation, and this should be investigated for this dataset and discussed. See methods of Pfannerstill et al. 2023 ACP for a list of possible water clusters and fragments to look out for (https://doi.org/10.5194/acp-23-12753-2023).
There needs to be a discussion about the limits of detection for these highly oxygenated molecules.
Results:
Regarding section 3.1 on identifying VOC formulae, I believe more quality assurance should be included here since this is a central part of your results. Was the peak identification performed manually or automatically in Tofware or other program? Did you set a detection threshold for identifying peaks (i.e., what’s your limit of detection)? What maximum mass error was allowed for identifying peaks? How confident are you about formulae identifications, especially at high molecular weights where you might run into ambiguous assignments?
I think there should be more discussion on the context of the highly oxygenated molecules. You define this group of compounds to focus on as VOCs with 3 or more oxygen atoms. But what fraction of the total measured concentration does this group comprise? What fraction of the total OH reactivity and/or SOA formation potential? The paper would be much stronger with this added context and motivation.
Specific comments:
Abstract:
Line 28: Change ‘compositions’ to ‘composition’
Introduction:
Line 113-116: Add references
Lin 137: Delete ‘at molecular level’
Methods:
Line 144: Delete ‘traffic’ or ‘roads’
Line 163-164: How regularly was the filter changed?
Concerns about inlet design - Was your inlet heated and did you test any flow rates besides 3 LPM to assess adsorption of sticky VOCs and/or IVOC/SVOCs? How do you think the inlet impacts your measurements of lower volatility highly oxygenated VOCs?
Line 170: 5.6 km feels a bit far away, was this location usually downwind or upwind of your measuring site?
Line 184: What BSQ voltage was used? What was the lowest m/z you were able to detect at 100% transmission with that voltage? This m/z depends on the specific BSQ voltage used.
Line 190-193: Saying that you used the average of three VOC sensitivities would be more honest. Before seeing the SI figure, I interpreted this as if you used those three VOCs in the fit of sensitivity versus kPTR. Since there doesn’t appear to be a linear relationship between kPTR and sensitivity here, maybe try a different approach. Consider using an average of your well behaved calibrants (i.e., those outside of the BSQ filtering range that do not fragment or cluster to a large extent and the sum of monoterpene parent and major fragment [C10H17+C6H9])
Line 195: Note that Figure S3b is the transmission of VOCs through the BSQ, correct? Was this transmission curve used calibrate VOCs in the m/z region impacted by the BSQ?
Did you correct for ToF transmission?
Line 202: Wrong unit – need molecule in the denominator.
Line 235: ‘Square Euclidean’ Instead of ‘Sqeuclidean’?
Please include uncertainty estimates for calibrated and non-calibrated VOCs.
Please include your E/N ratio and see the supplement of Coggon et al. 2024 AMT (https://doi.org/10.5194/amt-17-801-2024) for a discussion on how other parameters (i.e., skimmer gradient) can also impact fragmentation in the vocus.
Results:
Line 242: I think you mean number of molecular formulae instead of ‘number of organics’?
Line 250-252: Please indicate what time resolution you are referring to for these concentrations.
Line 258: Consider saying ‘identified’ instead of ‘discovered’
Line 268: Replace ‘individually’ with ‘individual’
Line 297: Change ‘organic vapors species’ to ‘organic species’ or something else
Line 309: too many significant figures, consider rounding to 2800, same with ‘2352’ in line 312.
Line 328-329: I don’t see evidence to support this claim.
Line 332-333: Do you have evidence that these formulas are the oxidation products you think they might be (e.g., GC-MS)? With Vocus, you can only measure the formula which could have multiple isomers. Please correct the grammar here, too.
Line 344: I don’t think you have evidence to support the claim that this may explain the missing source of OH radicals. Also, what do you mean missing source?
Line 349: I think grammar needs to be adjusted here. This phrase is unclear ‘…supplement the missing VOCs when calculating OH reactivity…’
Line 352-354: Is this rate constant reasonable? You seem to have arbitrarily chosen a rate constant that’s on the order of many terpenes, which are considered to react very quickly with OH. And again, what do you mean by missing OH reactivity? Was that measured here? Need more context and especially need proper evidence. Please consider using a more methodical approach for estimating what the kOH should be for each formula and scaling each concentration by each kOH to get your total calculated OH reactivity and report how important these highly oxygenated molecules are relative to the other VOCs are on this scale. I am having a hard time contextualizing these low-concentration highly oxygenated molecules.
Line 357: Reduce number of significant figures
Line 370: ‘Significant’ - if you don’t mean statistically significant, change word to ‘substantially’ or something similar.
Line 377: Statistically significant? If so, please indicate p value?
Line 418-419: I feel like you may be lacking evidence here. How do you know they are more influenced by secondary sources or that they even have primary sources.
Line 465: Overserved? --> observed
Line 466: Did you investigate fragmentation onto isoprene’s parent mass? See Coggon et al. 2024 AMT https://doi.org/10.5194/amt-17-801-2024
Line 503: Do you mean up to 230 m/z were observed in this study? Did you look into siloxanes like D4, D5, etc.?
Line 508: Do you mean ‘urban areas’?
Line 511: don’t need to say both dominant and main
Line 513: urban areas
Line 532: change discovered to identified or something else.
Line 533: content instead of contents
Line 537: urban area to urban areas
Line 562: change measurement to measurements
Figure 1: Add labels to pie charts in (b) and (c). In the caption, (b) should be number of organic formulae I think?
Figure 2: To make this figure more clear, please add label to legend in (a) saying something like ‘number of oxygen atoms.’ Please add labels to pie charts in so you can see clearly what each one is plotting, i.e., ‘concentration of CHO species’.
Figure 3: Final sentence of caption – change ‘Y axials’ to ‘Y axes’
Figure 4: Please indicate if the molecular formula in figure 4 are protonated or not. If they are the protonated ion formulas, please indicate they are charged.
Figure 5: (a) I can barely see the dot distribution to the left of the box plot for ‘spring’. (d) Y axis label should read ‘Fraction of’ instead of ‘Fraction to’
Figure 6: Please add a legend for cluster 1 and cluster 2 in the plots to make this clearer. Shading of percentiles is too light in c, f, h, I, k, and l.
Citation: https://doi.org/10.5194/egusphere-2024-1325-RC2 -
AC3: 'Reply on RC2', Zhaojin An, 06 Sep 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1325/egusphere-2024-1325-AC3-supplement.pdf
-
AC3: 'Reply on RC2', Zhaojin An, 06 Sep 2024
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
520 | 157 | 84 | 761 | 60 | 18 | 25 |
- HTML: 520
- PDF: 157
- XML: 84
- Total: 761
- Supplement: 60
- BibTeX: 18
- EndNote: 25
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1