In their manuscript "A modular approach to volatile organic compound samplers for tethered balloon and drone platforms" (amt-2024-29), the authors present an instrument for collecting multiple sorbent tube samples that is suitable for drone- and balloon-based platforms. In general, this is a major topic in VOC sampling and the subject of many recent advances. The instrument described here is a nice addition to the suite of custom tools that have been developed throughout the research community. And the reported applications demonstrate the utility of these measurements To be suitable for publication, this manuscript should more fully examine the existing literature and explore what tools are out there. Many of the goals achieved in this work have been previously demonstrated. The authors are in a bit of a rock and a hard place, because many of the samplers previously demonstrated are not commercially available, so there is value to providing detail and demonstrating their own solution and I think it is still valuable to publish this work. However, with that in mind, I would have liked to see a little bit more of a demonstration of the science that will actually be achieved with this device and/or a demonstration of some of the next-step advances they describe as being possible and would advance the field. Specifically my two major concerns are described here:

(1) There are details missing on the technical aspects of this device. For example: it's overall weight and size are only in the abstract, the capacity of the battery is not provided, there is a lack of clarity about the control system (why a computer and a microcontroller?), whether power for the UAV version is being supplied by the UAV itself, how or if it is being integrated into the TBS payload, etc.

(2) The introduction overlooks prior work that demonstrated many of the achievements shown here, primarily multiple samples collected from a single airborne package. The authors elude to many promising features of their work, including communication with the ground, potential to integrate signals from other instruments, and complex sampling strategies. These advances would truly be novel and show the field what can be done, but are not really demonstrated here (and their theoretical possibility was mentioned in some of the previous literature). The authors should look a little more deeply into the literature (a few citations are provided, but I think there are others) and more thoroughly identify the gaps and how their work fits into that context.

Specific comments:

Line 68-70. There have been at least some efforts (including work published in this journal) on independent, drone-compatible VOC samplers, both for collection of multiple coordinated samples across multiple platforms (DOI: 10.5194/amt-16-4681-2023), and for sequential samples within the same platform (DOIs: 10.5194/amt-12-3123-2019; 10.1016/j.jes.2024.04.016) as is being reported here. So this statement is not really true and highlights some need for a bit more literature review to place this work in the context of prior work.

Line 86-87. It's not clear what the author's mean by this statement. For example, commercially available options can monitor trace gas levels and use them as triggers for sorbent tube collection (see for example the SENSIT SPOD), and canisters have regularly been collected alongside comprehensive instrument payloads, includign with an arbitrary trigger for collection (DOI: 10.5194/amt-10-291-2017). What do the author's mean by "integrated" in this context? 

Line 91: By "integration" in the ARM TBS payload, do author's mean they use the meteorological, aerosol, and ozone data to inform timing of sample collection, or just that they share a data aquisition system? The latter does not seem like a scientific advance, but rather solving a specific technoical issue relevant just to the TBS system. Throughout the work, it looks like this sampler only uses preset timing to determine sample intervals, so it's not clear to me that it is really "integrated" in any meaningful way.

Line 103-104. "uses lighter-than-air principles to obtain its initial lift" seems overly jargony/complex. I would say that the fact that helium-filled balloons rise can be considered common knowledge and does not need to be explained

Line 111. Authors should note the size range of particles measured by POPS, since later concentrations are provided just as total numbers

Line 131. The authors should clarify what they mean by modular. Which components can be separated and recombined in different ways?

Line 140. It's not really clear to me why there needs to be a computer on board. Why not just provide firmware to the Arduino and re-program it on the ground if necessary? Including the computer seems like extra weight and power. Or conversely, it looks like the UDOO Bolt has analog and digital I/O, so why include the microcontroller at all?

Line 147. How was flow controlled? Is flow checked on each tube, since their resistances could be different, or is it measured in real time? Also, no information about the pump is provided - what pump is being used? 

Line 204. What do the author's mean the code initiation was synced? That the start time was synced with the start time of the flight?

Line 236. More information about this pairing would be helpful. Is the power for the UAV being directly used to power this device? And what communication features are being used. In general, more description of the "modularity" of this device would improve understanding of its uniqueness and value. This is later described in more detail in lines 254-256 and I agree ground-based communication with the device is a big step forward that I'm not sure has been previously demonstrated; this should be described earlier and/or in the methods.

Line 250. Was the use of metadata used here or demonstrated? Complex sampling strategies have been previously shown to be possible on the other samplers mentioned above, but the inclusion of other sensors, though likely possible in the other systems, would to my knowledge be a new advance worth demonstrating explicitly.

Line 266. "1 A" is not a unit of power, do the authors mean 1 Ah? What is the size of the battery?

Line 343. Could the authors estimate increase in mass due to each additional tube? This would involve the tube, valve, and lines, are there other components that would also need to be scaled?

This article describes the conceptual design and deployment of a battery-powered sampler that was used for the collection of volatile organic compounds onto multi-stage solid adsorbent cartridges from a tethered balloon and a drone platform. The sampler was deployed at two sampling locations in Texas and data from these campaigns are presented. 

While I appreciate the effort that was put into the study and preparation of the manuscript, in my opinion this work has crucial deficiencies.

There are quite a few other previous tethered balloon and drone VOC sampling systems and deployments that have been reported in the literature.  There is relatively little novelty in the particular sampling approach and deployment presented here.  Furthermore, I consider the data rather questionable for the reasons described below. 

The analytical experiments do not specify the accuracy and precision of the sampling and analysis method.

The effect of the downwash from the UAV rotors on the actual effective sampling height is not addressed.

It appears that sampling tubes do not seem to have shutoff valves on the inlet side but are always open to the outside air during the flight, which allows them to passively sample VOCs during the entire deployment.  These types of adsorbent tubes have been found to have passive uptake rates of approximately 0.5 ml/min in typical deployments at the surface (e.g. [Mowrer et al., 1996; Walgraeve et al., 2011; Markes, 2021]).  This sampling rate might be higher under the highly turbulent conditions during the flight deployment.  This will add substantial additional sampling VOCs during times when there is no active (pumped) sampling.  For instance, if a deployment would be 2 hours from the time of installation to the removal, this would add on the order or 60 ml of sampling to the 1 l of sampling at the deployment altitude.  VOCs might be higher at the surface than aloft, so one will never know exactly now much of the analysis results actually reflects the VOCs mole fraction at the balloon sampling height.

The manuscript does not clearly state that samples are collected sequentially.  The profiling data do not truly represent vertical profiles as intended in this sampling, but also need to consider that atmospheric VOCs may change rather rapidly in plumes in an urban or suburban environment.  This temporal aspect of the sampling is not recognized in the manuscript.

It has long been known [Goldan et al., 1995] [Pollmann et al., 2005] that unsaturated VOCs, in particular biogenic VOCs such as isoprene, undergo rearrangement and loss during atmospheric sampling with prefocusing techniques from reaction with ozone in ambient air.  A series of approaches have been researched and used by researchers over the years to mitigate this artifact [Helmig, 1997].  It does not appear that the sampler had any sort of scrubber for selective removal of ozone in the sampling flow path. It is therefore questionable if and what fraction of the actually present VOCs were captured by the sampling protocol. 

Calibrating the adsorbent sampling with diluted liquid solutions is far from ideal as it does not reflect the actual air sampling.  Utilization of certified VOC compressed gas standards is a much more widely used and accepted way for calibration in atmospheric VOCs monitoring.

I could not find the Dieu Hein , 2019, reference (line 55) in the references list.

The sampling only captures a subset of VOCs present in the atmosphere, probably well below half of the total VOC ppbC.  This needs to be realized.  The term ‘Total VOC concentration’ (line 286) is not quite accurate in this context.

Results are presented in ppb, which is a mole fraction unit.  Calling this a concentration (e.g. line 288) is not correct.

The url that is provided for the data availability statement only leads to the archive portal, but not to the particular data that were generated in the study.

Information provided in Table S2 is not clear.  What does the product number refer to?  What are the numbers in columns 4 and 5?  Significant figures are inconsistent.

References

Goldan, P. D., W. C. Kuster, F. C. Fehsenfeld, and S. A. Montzka (1995), Hydrocarbon measurements in the southeastern United States: The Rural Oxidants in the Southern Environment (ROSE) program 1990, Journal of Geophysical Research-Atmospheres, 100, 25945-25963, doi:10.1029/95jd02607.

Helmig, D. (1997), Ozone removal techniques in the sampling of atmospheric volatile organic trace gases, Atmospheric Environment, 31, 3635-3651.

Markes (2021), Listing of uptake rates for axial passive samplers, https://markes.com/content-hub/application-notes/application-note-001.

Mowrer, J., P.-A. Svanberg, A. Potter, and A. Lindskog (1996), Diffusive monitoring of C6-C9 hydrocarbons in urban air in Sweden, Analyst, 121, 1295-1300.

Pollmann, J., J. Ortega, and D. Helmig (2005), Analysis of atmospheric sesquiterpenes: Sampling losses and mitigation of ozone interferences, Environmental Science & Technology, 39, 9620-9629.

Walgraeve, C., K. Demeestere, J. Dewulf, K. Van Huffel, and H. Van Langenhove (2011), Diffusive sampling of 25 volatile organic compounds in indoor air: Uptake rate determination and application in Flemish homes for the elderly, Atmospheric Environment, 45, 5828-5836, doi:10.1016/j.atmosenv.2011.07.007.