Interactive comment to "Identifying and correcting interferences to PTR-ToF-MS measurements of isoprene and other urban volatile organic compounds"

Matthew M. Coggon*, Chelsea E. Stockwell, Megan S. Claflin, Eva Y. Pfannerstill, Xu Lu, Jessica B. Gilman, Julia Marcantonio, Cong Cao, Kelvin Bates, Georgios I. Gkatzelis, Aaron Lamplugh, Erin F. Katz, Caleb Arata, Eric C. Apel, Rebecca S. Hornbrook, Felix Piel, Francesca Majluf, Donald R. Blake, Armin Wisthaler, Manjula Canagaratna, Brian M. Lerner, Allen H. Goldstein, John E. Mak, and Carsten Warneke

*Corresponding author: matthew.m.coggon@noaa.gov

https://doi.org/10.5194/egusphere-2023-1497

This work describes an analysis of measurement interference in PTR-ToF-MS observations of several key VOCs undertaken across 5 field campaigns in US urban areas employing 5 different PTR-ToF-MS systems. PTR-MS techniques are widely used in atmospheric measurements of VOCs and interference in the identification and quantification by non-target ions is a major source of uncertainty associated with this method. There has been significant evolution in the mass resolving power of PTR-MS techniques necessitating updated assessments of the contribution of this uncertainty and methods to address it. The methods for identification, quantification and correction of interference presented in this manuscript provide a significant contribution to the task of reducing uncertainty in PTR-ToF-MS observations which are widely used to validate emissions inventories, air-chemistry models and human exposure studies and are therefore highly relevant to the scope of AMT. The extent of the interference issue in urban air measurements is well evidence by the presentation of multiple datasets from co-located PTR-ToF-MS and GC instruments, collected across several US cities, on stationary and mobile platforms. The methods used are sufficiently outlined and references provided for further detail. Approaches to identification and correction based on the available data are sound, easily adopted for use by the wider PTR-ToF-MS community and result in substantial improvements in PTR-ToF-MS data quality. The manuscript is well-written and the results and conclusions are well presented.  Overall, the manuscript is of high quality and should be accepted for publication in AMT in its current form with only technical corrections.

That being said, the work presented would be strengthened if an estimated uncertainty was presented for each technique employed for the target species discussed (i.e. JCGM 100: 2008 Evaluation of Measurement Data – Guide to the Expression of Uncertainty in Measurement, BIPM, Sevres, France, 2008). This would allow us to see the progress in addressing uncertainties in commonly applied VOC measurement approaches that may currently be poorly understood/quantified.

The authors could also consider demonstrating the impact of these corrections by reporting the difference in estimated ozone or SOA formation potentials determined from corrected and uncorrected data.

Other general comments:

• Include info on type of regression analysis used (LLS, ODR, RMA). If LLS has been used consider other methods that account for random measurement error on both axis.
• If sufficient concentrations are present, consider applying the Method 2 benzene correction approach presented in Section 3.3.3 to isoprene using m/z 68? This could be of relevance for stand-alone PTR-ToF-MS measurements.

Technical corrections

• Spaces needed between references presented in brackets in text
• Suggest consistent colouring in figures. In Figures 3, 4, 6, 9 red is used to show interference whereas in Figures 5,7,8,11,13,14 red is used for the corrected concentration.
• Include actual value in brackets where slopes and interference ratios are reported in the text.

In this study, the authors combined data sets collected by various PTR-TOF-MS instruments (Vocus and traditional drift tube reactors) at different field campaigns in the USA over 2 years. Ionization by PTR is known to produce fragments complicating analysis of ambient measurements since lower molecular ions such as C5H8H+ (m/z 69) could form as a protonated parent molecule or as a fragment of a higher molecular species. The authors use GC-MS measurements to identify actual mixing ratios of those lower molecular ions to then build a mathematical function correcting the ambient mixing ratios detected by PTR-TOF-MS. These calculations are further supported by correlation plots. Overall, the publication contributes to the advancement and VOC measurement techniques by PTR-TOF-MS and is therefore  well suited for publication in AMT.  This manuscript will be an excellent resource of the PTR community.

Major comment:

To make this resource most effective for the future reader, the text could be organized by *instrument* rather than by campaign. As a resource, I am much more interested in comparing my own results with the same instrument reported in the text. Then, I can see how the environment (urban vs rural for example) could impact the fragmentation. This proposed changed would involve some copy-pasting to rearrange the text so that each instrument is described in detail, and then following this description; its deployment is listed. Some field campaigns would then be reiterated, but the figures wouldn’t need to be updated. My comment here is really coming from the perspective of a PTR user and using this manuscript as a reference.

Scientific comments:

Generally, it seems that m/z  69 and isoprene are used interchangeably in this study (comparing Figure 5 and 7), but as the authors point out, they represent different values in this study: (1) m/z ratio obtained by the PTR or (2) parent ion of isoprene. Points to address to help bring consistency:

• L 737 “...m/z 69 varied between 200–1200 ppt…” It isn’t accurate to report a mixing ratio for a m/z ratio, because one can only calibrate for a specific VOC. So mixing ratios should be used when mentioning isoprene specifically.
• Would it then be worth reporting counts per second throughout the discussion on m/z 69?
• Some instruments have unit mass resolution (m/z 69) and some have high mass resolution (m/z 69.070). It would be worth being consistent throughout the text. (L 491-493)
• Table 1 could also be revised to include m/z ratios.

L 933-935, 1105 -1106: Great discussion on how instrument discrepancies could affect the correction factor. Thus, authors could consider discussing some of the statements (for example, RH dependencies, Charge transfer ionizations)  in light of instrument specificity. These factors will vary for different instruments.

Organization/formatting comments:

To help make this resource even more accessful to the future reviewer and reader, I can encourage the authors to consider:

1. adding more sub and subsubsections to their text. As it stands, there are sections that go on for ~5 pages (section 3.1).
2. adding to table 1 the years of the campaign and whether the instrument is from Ionicon or Tofwerk/Aerodyne or custom-built or another source. Since PTR instruments have differences in their drift tubes, highlighting this difference is necessary for the future user. This table would be an excellent summary and resource for the future reader, and could be as large as taking up one whole landscape page with all the details of each instrument. What I’m suggesting here is having a summary table for quick references on similariries and differences of instrument operation (arguably more important than the campaign results).
3. presenting the a merged methods and results sections, perhaps per instrument and then applied to a campaign.

Figure 13 and 14, the pie charts are concise demonstrations of the importance of applying the proposed corrections, maybe those could be added to the previous figures as well where applicable?

Some technical comments:

L50-53: Would be worth defining secondary ionization and help clarify how the drift tube designs make an important difference in how much charge transfer can occur vs proton-transfer. This point can be reiterated on lines 100-102 as this process is quite different between Ionicon vs Tofwerk/Aerodyne instruments.

L79,84, 90: the loss of OH2 through a “dehydration” process would involve a change in oxidation state at the carbon of an organic molecule. As a chemist, the process of dehydration is significantly different than the process of fragmentation. Do they authors want to clarify their point further? (also an aldehyde can be written as (C(O)H, not with a doublebonded H)

L81-82: would be worth adding a reference to support the statement on known fragmentation of alcohols and aldehydes through PTR ionization.

L123: VCPs should be defined as it’s the first use of this acronym.

L173: a 10m inlet might also play an important factor in observed signal by PTR. Can the authors comment on the sample line effects?

L203 “ Each flight covered approximately 5000 km of distance across the Los Angeles area,..” According to this website (https://man.fas.org/dod-101/sys/ac/uv-18.htm) the UV-18A Twin Otter has a range of 700 miles. Is it possible that the distance is instead 500 km (or there’s a typo?)? Also, who is the vendor of the NCAR TOGA-TOF?

L 348-349 “The Vocus was operated at a pressure of 2.2 mbar and axial voltage gradient of 600 V, corresponding to an E/N ratio of 125 Td.” What T was the reactor at? Including the temperature, pressure and electric field gradient for all instruments to streamline intercomparability is important for the future reader (that info should also be available for each instrument)

 L 464-473 This section seems to be a duplicate of L 442-455

 L 635 The Los Angeles or Caltech sites are the same correct (referring to the results in Fig 5)? Could be worth sticking to one name (to facilitate the control find function is a future reader is looking for LA info specifically within this resource)

I command the authors for this valuable future resource to the PTR community!