the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Determination of the Atmospheric Volatility of Pesticides using Chemical Ionisation Mass Spectrometry
Abstract. Pesticides have been found to be transported through the atmosphere away from fields on application. A key indicator of a pesticide’s likelihood to reside in the atmosphere is its vapour pressure. Within this study we evaluate a novel method, the Filter Inlet for Gases and AEROsols (FIGAERO) coupled with a chemical ionisation mass spectrometer using a set of calibration compounds, poly-ethylene glycols (PEGs). Two methods of compound delivery onto the filter have been tested: atomisation and syringe deposition. Delivery results are consistent with previous studies, highlighting the lack of suitability of the syringe method. The successful calibration using the atomisation method was then used to determine the vapour pressure of 6 pesticides. This is the first-time particle phase pesticides have been measured with online mass spectrometry. The pesticides have then been compared to widely accepted standard literature values used in industry as well as values determined by a common environmental model also used in industry to give an indication of pesticide volatilities. Results showed that measurements from the FIGAERO-CIMS were consistent with reported literature values for some compounds, others differed by up to 2 orders of magnitude. Determinations of Dicamba, MCPA and MCPP volatility using the FIGAERO-CIMS showed them to be semi-volatile in agreement with literature values to be within an order of magnitude. Mesostrione exhibited the largest difference in volatility with the FIGAERO-CIMS measuring a low volatility of 4.12x10-8 Pa at 298 K (compared to a literature value of 5.7x10-6 Pa). The difference for 2,4-D of one order of magnitude perhaps can be explained by the smaller particles deposited on the FIGAERO filter compared to the aerosolised PEG calibration particles, leading to evaporation at lower Tmax values and a lower measured vapour pressure. The atmospheric implications of the pesticide volatilities are also discussed. A pesticide’s volatility is often a key indicator of the likelihood of the potential for short- or long-range transport occurring, thus determining a pesticide’s fate in the atmosphere and potential for environmental pollution from transportation in the air.
- Preprint
(1358 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2380', Anonymous Referee #1, 30 Aug 2024
General Comments
Jackson et al. provides volatility measurements for a set of 6 pesticides in current use through application of a filter desorption method utilizing the Filter Inlet for Gases and AEROsols coupled to an iodide mode Chemical Ionization Mass Spectrometer (FIGAERO-CIMS). The authors discuss broad variability and inconsistency in measured and reported vapor pressures (volatility) for many compounds, dependent on measurement method. FIGAERO-CIMS volatility is derived from a relationship between temperature of maximum thermal desorption signal (Tmax) and volatility (increasing Tmax indicates lower volatility). This relationship is quantified by calibration of the technique against a set of polyethylene glycol (PEG) polymers of different lengths, which can be loaded onto the filter by syringe deposition or atomization. This calibration is then utilized to assess the volatility of the target pesticides when they are atomized in a mixture or individually.
Jackson et al. conclude that, in agreement with prior literature, use of a syringe for calibrant or sample delivery is less effective than atomization. However, as noted in specific comments below, greater consistency is achieved between literature PEG Tmax values using the syringe method than atomization. While arguments from prior literature are used to highlight the unsuitability of the syringe method, the data as presented does not provide clarity as to why syringe delivery is unacceptable. Additional quantification of thermogram peak width and subsequent resolution or overall variability of each measurement method would provide valuable insight into the failings of the syringe method for PEG calibration.
Pesticide volatilities are measured and reported for 6 pesticides of interest and compared to literature values, as well as values derived from 2 structure-activity relationship models. Here, Jackson et al. show close agreement between literature and FIGAERO-measured values for Dicambia, MCPA, and MCPP. Volatility of 2,4-D is measured as more than 1 order of magnitude higher than the literature value, explained here by differences in the PEG calibration particle size distribution. Volatility of mesotrione is measured as more than 2 orders of magnitude lower than the literature value, explained by error in the measurement of said literature value (measurement conducted at a higher temperature than this study). A Clausius-Clapeyron relationship is used to argue that observed and modeled volatilities are reasonable given the temperature difference between the two measurements. Trifluralin is noted as the highest volatility pesticide measured here. However, as noted in specific comments below, the data shown in Figure 6 seems to contradict this, since the Tmax observed there is substantially higher than the 27.7C reported in Figure 7. Furthermore, the desorption profile of Trifluralin has a uniquely large right tail and further assessment and analysis of this species behavior would be appreciated.
Generally, Jackson et al. concludes that FIGAERO-CIMS measurements of volatility are valuable additions to pesticide environmental assessment to understand the fate of toxic pesticides in the environment. Such measurements can improve modeling of fate and transport and environmental persistence. The manuscript presents a unique set of measurements of a particular set of pesticides as an example of this technique and its reasonable agreement with other vapor pressure measurements. Furthermore, the manuscript identifies a series of key concerns in the vapor pressure measurement space, including poor documentation of measurement methodologies including temperature measurement.
The manuscript would benefit from additional quantification when making comparisons between measurements and measurement techniques, both in terms of quantified Gaussian goodness of fit and peak shape metrics, and in terms of presenting uncertainty in the form of standard deviations or propagated error. Furthermore, additional literature comparisons where possible would be valuable. As discussed in the technical comments, small updates to the presented figures would improve legibility, consistency, and clarity across the manuscript.
Overall, the manuscript provides a unique measurement of pesticide volatility and is an informative first step toward building a more consistent picture of pesticide fate and transport in the environment. I appreciate the author’s consideration of this reviewer’s comments and their ongoing work to illuminate this important area.
Specific Comments
Line 105-107: The authors describe how volatility measurements may be conducted at higher temperatures and be extrapolated to lower temperatures. However, later in the manuscript when discussing this effect with mesotrione, it seems that rather than being extrapolated to lower temperatures, the measurements at high temperatures are simply used at low temperatures (there is no particular extrapolation). The authors then describe using a Clausius-Clapeyron relationship as an appropriate adjustment to vapor pressure. How widespread is the practice of ascribing high temperature vapor pressure measurements to low temperature conditions?
Line 147-152: What are the implications of pesticides residing in a particular volatility class? What inference can you make about the fate and transport of that pesticide in the environment? Both here and elsewhere the authors comment on the volatility classes, but I think additional clarity in interpreting these classes would be valuable.
Line 165: Please note the thickness, pore size, and brand of the filter used in addition to the material. Similarly note purity and sources for all materials when described in the methods section.
Line 167: In what sense is the filter “backflushed”? The flow direction of heated nitrogen is the same as the initial particle sampling flow. (This may just be a filter sampling terminology I am unfamiliar with!)
Line 178-179: Here the authors mention Bannan et al. (2019). I would suggest including PEG Tmax data from that study for comparison in addition to Ylisirnio et al. (2021). In general, I think the manuscript would benefit from a broader range of comparison to FIGAERO data sets where possible to better capture the variability in available measurements and the corresponding uncertainty in the measurement here.
Line 190: Was precisely 1 ug of atomized particles deposited on the filter for each measurement in this study based on live SMPS data?
Repeatability: Were calibrations and measurements repeated in this study? I see that calibration measurements were repeated 3 times in the caption of figure 3. Please add to the body text as well. What is the distribution of measurements and can uncertainty be represented by error bars on many of the figures here?
Methodology: Why were separate calibration and pesticide volatility desorptions necessary? Could PEG be included in the pesticide solutions and atomized simultaneously for a concurrent volatility calibration during pesticide desorption?
Line 249-250, “Conversely it is also important to determine the volatility of pesticides thought to be involatile (i.e., no chance of volatilization in the atmosphere) to ensure that there is no potential for atmospheric presence thus no further risk assessment in air is required.”: Here, this argument is made the low volatility species have no atmospheric relevance. In other portions of the manuscript, the authors note that low volatility species can be sequestered in particles and avoid atmospheric oxidation and be transported long distances (Line 89) or that low volatility species can be resuspended in dirt or dust particles (Line 490). Is this line intended to capture the current state of environmental risk assessment policy? If so, that should be clear and distinct from the broader commentary the authors make on potential environmental fate based on the results of this study.
Line 283-286: In this discussion of thermogram shape and uniformity, some quantification of the PEG desorption curves may be valuable. What is their goodness of fit to a gaussian and their full width at half maximum?
Figure 3a: Comparison between Ylisirnio et al (2021) using both methods here shows the increased Tmax associated with the syringe method over the atomization method which is well discussed here. However, Tmax data seems to be more consistent between the syringe method in both studies, with a bigger gap in Tmax observed between Ylisirnio and this study when atomization is used. Given that repeatability of vapor pressure measurements and consistency across measurement types is so poor, could you further explain/discuss if this apparent consistency when using the syringe method is desirable for vapor pressure estimation in targeted measurements?
Figure 3b: There are two more points represented in the syringe data set than in the atomization data set. Why are these measurements not represented in Figure 5 (where both methods begin showing data at PEG-5)?
Line 313-314: What differences are there between the size distributions used in this study and Ylisirnio et al (2021)?
Line 320-321 and Figure 4: Please provide some additional quantitative discussion of the particle size distributions observed (mode, total mass, spread, etc.) and if they are sufficiently similar to the PEG distribution for comparison and calibration per Ylisirnio et al (2021) as mentioned.
Figure 4: What density is assumed when calculated particle mass from SMPS particle volume distributions?
Figure 6: The shape of the pesticide desorption peaks seems, at least by eye, much sharper than the PEG desorption peaks. Can you provide quantitative information about peak shape statistics? How closely do the pesticides follow a Gaussian desorption curve?
Figure 6 and Figure 7 and associated trifluralin discussion and conclusions: The thermogram shape of trifluralin in particular is unique and worth discussion. It sharply appears at near 30 degrees and then peaks with a long tail to the right. Is this evidence of excessive trifluralin loading? Some type of thermal decomposition product? Furthermore, the Tmax for trifluralin in Figure 7 does not look correct based on Figure 6. There, the Tmax appears closer to 45 or 50 C, not the 27.7C reported. Please assess and correct as needed.
Line 363-365: This brief discussion of the potential bias in vapor pressure measurement of 2,4-D here is useful. Is there any method for extending this potential bias quantitatively to capture the anticipated error or uncertainty in the FIGAERO-CIMS volatility measurement?
Line 396 “This does not impact the FIGAERO-CIMS measurements which relies on thermal decomposition.”: What is meant here? FIGAERO-CIMS measurements rely on thermal desorption and thermal decomposition can and does occur in the FIGAERO, leading to multimodal thermograms which require additional processing to appropriately separate desorption and decomposition.
Line 478-479, “Unfortunately, the reports do not specify which particular method was utilized. Consequently, it cannot be assumed that the pesticide literature values can be completely reliably compared due to the substantial variations in these methods.”: Can the potential variation in methods be displayed for example in Figure 8 to capture the potential spread of literature values compared to the measurement conducted here? Are there other literature sources to display in Figure 8?
Line 496-498, “Despite this a study in Arctic monitoring stations found low levels of Trifluralin in arctic air(Balmer et al., 2019) and has now been predicted that small amount of Trifluralin may stick to aerosol particles and transported significant distances.”: This is inconsistent with your current observation of Trifluralin as an IVOC which would partition sparingly to aerosol particles. As per a comment above, please confirm that your data agrees with the Tmax and volatility characterization presented here. If it does, please discuss relevant literature and mechanisms that may allow Trifluralin to be transported to the Arctic even with a short atmospheric oxidation lifetime and high volatility.
Technical Corrections
Line 122: “maybe” to “may be”
Figure 2: Presenting these thermograms with the same horizontal axis in both panels will make the Tmax shifts more clear.
Figure 3: Increased marker size and consistency in marker choice for delivery method from this study across both panels would improve legibility. In 3b, I would suggest using log(vapor pressure) not ln(vapor pressure) for greater consistency across figures and more straightforward comparison to text values.
Figure 4: Units on vertical axis.
Figure 7: Where error bars appear in a figure, what they represent should be listed in the figure caption. Vertical axis label should use a subscript for “max”
Figure 8: Increased marker size may be useful for legibility. In addition, error bars where possible should be included. Panels could be oriented side by side and use a shared legend. Finally, are multiple literature sources used in this figure? In that case, each literature source should have its own symbol and be clearly cited. If only one source is used, that should be listed in the legend and cited in the caption.
Figure 9: This figure is extremely difficult to read, though this may be due to some upload issues. In any case, I would suggest that the labels for the points not be cut off by the frame, and use of a legend might be more appropriate anyway. I would also suggest using log(Vapor Pressure) rather than ln(vapor pressure) to make the values in the text more easily read off of the chart and for consistency with earlier figures on a log scale. Finally, labels such as “stated literature value” are not useful. Indicate which specific literature source is used for that point, I believe this would be the University of Hertfordshire Pesticide Properties Database.
Table 2: Reported values should include uncertainties (standard deviation or similar).
Citation: https://doi.org/10.5194/egusphere-2024-2380-RC1 -
RC2: 'Comment on egusphere-2024-2380', Anonymous Referee #2, 04 Sep 2024
Jackson et al, present vapour pressures for six pesticides measured using the FIGAERO-ToF-CIMS analytical technique. There is a large discussion around the method of delivery of the pesticides onto the filter contrasting direct injection and atomisation. The method of calibration against a series of PEG compounds with known vapour pressures is also presented. The measured values are also contrasted with literature values and two SAR models.
This manuscript is of excellent scientific significance as it demonstrates the utility of the FIGAERO-ToF-CIMS measurement technique for investigating the volatility of pesticides with high environmental relevance. This manuscript should be considered for publication after considering several major and minor comments.
Major comments
The manuscript is extremely comprehensive regarding background, summarising the current state of knowledge and where these new finding fit in a literature and regulatory context.
While informative, large parts of the text are quite verbose and do not seem entirely relevant to the focus of the manuscript. Additionally, throughout the manuscript and specifically in section 4.1, there seems to be a lot of repetition. In some instances, the text does not match the section it is found in, for example there is some description of the experimental method at the end of the SAR section. I found the discussion on comparisons with the literature quite hard to follow, mainly due to the lack of clarity around which ‘literature’ is used for comparisons.
That being said, these issues concern the presentation of the study, rather than the study itself. I think the manuscript would benefit greatly from editing down to focus on the key message(s) and better connecting the bigger picture issues of regulation and reporting to the results. More efficient organisation of the text would also greatly improve its focus. This would also give the author the opportunity to check some grammar issues and typos.
Minor comments
Line 87 – “effectively shielded from degradative gas phase oxidation”. It would be nice to contrast this with condensed phase chemical processes. Is it possible to say something about the lifetime of the pesticide in the gas vs the condensed phase?
Line 93 – what is meant by “activated process”?
Line 95 – what is meant by “pesticide active substance”?
Line 125 (167) – what is meant by “reverse flushed”?
Line 146 – what “coefficient” is gamma?
Line 182 – flow of 20 Lm-1 of what gas?
Line 187 – Through -> although
Line 199 - I found the description of SAR a little confusing and not introduced particularly clearly. For example section 2.2 misses directly stating that SAR is used to predict vapor pressures, and there is no explanation of why the Nannoolal and MGM models are chosen specifically. I think this can be easily corrected by reformulating the text.
Line 200 – estimation methods of what?
Line 208 – “The Nannoolal model was chosen as training data for the model” doesn’t really make sense.
Line 263 – what is a PPP?
Line 283 – the commercial PEG solution is a great benefit. Can you explain a bit more about why you are using PEG-4 as your lower Tmax limit? Although the atomisation method is only useful down to PEG-5, it doesn’t look like the syringe method is great below that either? There is actually some discussion of this at line 338.
Line 289 – I wonder if a 1:1 plot of atomisation vs syringe Tmax values for each PEG would give any insights into the systematic bias of the syringe technique?
Line 304 – is it possible to say more about the impact of the 0.1 g/L vs 0.5 g/L deposition on the filter on the Tmax?
Line 315 – “small but consistent repeatable effect”. This isn’t explained well here but I think this explained more fully on page 17 later. This is a good example of repeated information in two different places. It would be good to summarise this information in one place – maybe you can make a comment on the variability of particle diameter and how it affects Tmax in this instance – is it possible to have a metric like x degrees C / nm? Is the variation here significant?
Line 357 – “... highest tmax was for the least volatile pesticide ..” least volatile according to who? Are you referencing a literature value or the fact that the tmax is highest for mesotrione?
Line 364 – Figure 8. These need an (a) and (b). Which Literature Value is being referred to in the legend? It would be better if the legend handle was a bit more descriptive.
Line 376 – what is a “regulatory endpoint value”?
Line 382 – MGM vs MGM SAR these refer to the same thing but are given different names.
Line 385 – what is meant by comparison data? It would be better to be more specific with the literature value you are comparing to.
Line 386 – The University of Hertfordshire Pesticides Properties Database for Mesostrione is mentioned here, and then two lines later, the EFSA endpoints, from which the database draws its data. This section reads very narratively which is confusing, i.e., I am not sure it is necessary to explain that the database had a suspect value but on further inspection it was because the underlying EFSA data is measured at a different temperature. This just highlights an issue with the database.
It might be easier to standardise the way you refer to the different literature data earlier on to make this easier to follow. Is the “upper limit” value the green “stated literature” value in figure 9? Using these different terms is hard to follow.
Line 394 – I am not entirely sure why thermal decomposition is mentioned here. What does difference in the EFSA document mean?
Line 405 – Figure 9. What are the different “literature measurements at varying temperatures”? where do they come from? It would be more informative to have the sources rather than stating they are measured at various temperatures (this is what the x axis shows already). The ‘stated literature value’ is just plotted at the wrong 1/T value. If measured at 373K then this should be plotted at 1/373 = 0.000268 and so would appear to follow the trend.
Line 418 – “within an order of magnitude”. It is of course good the difference in values is less than an order of magnitude, but without uncertainty measurements it is difficult to assess how ‘good’ the agreement is. Either a measure of the variability or uncertainty.
Line 420 – Do you expect higher volatility with more functionalisation? Can you explain further?
Line 456 – “Lower volatility compounds are unlikely to be atmospherically relevant if applied to the surface in the liquid or solid phase”. I take it this refers to application of pesticide to the surface of a plant? Low volatility compounds are of course atmospherically relevant in an aerosol context.
Line 457 – I don’t understand the message of the last sentence of this paragraph.
Line 467 - “training dataset” what data does this refer to? This is the Nannool training data? I think this just requires more consistency in naming.
Line 470 – “fishtine factor” needs introducing earlier in the text.
Line 490 – “ … volatilisation is a major degradation mechanism .. ” volatilisation doesn’t degrade the active ingredient, I guess this means degradation of the pesticide product itself?
Citation: https://doi.org/10.5194/egusphere-2024-2380-RC2 -
RC3: 'Comment on egusphere-2024-2380', Anonymous Referee #3, 14 Sep 2024
This study examined the volatility of pesticides using a novel approach, the Filter Inlet for Gases and AEROsols (FIGAERO) coupled with a chemical ionization mass spectrometer. Two compound delivery methods were tested, and the results were compared with those from previous studies. Volatility values were also evaluated against literature data, with potential explanations provided for discrepancies between measured and reported values. The atmospheric implications of pesticide volatilities are further discussed.
The scientific method used in this study is sound, and the results are meaningful and show promise. Additionally, the topic is significant and relevant to the field. However, I believe this manuscript is still in the draft stage and requires improvement in several key areas before it can be considered for acceptance. Overall, I would recommend rejecting the manuscript in its current form.
Major comments:
1. Many sentences in the manuscript are either grammatically incorrect or lead to confusion, making it difficult for readers to understand the content easily. For example, I’ve noted several specific instances, but I believe there are many others throughout the paper that require improvement. I recommend reviewing the entire manuscript to enhance clarity and readability.
Line 18-19: I feel it would be better if the sentence can be revised as “ The pesticide volatilities were compared with widely accepted standard literature values used in industry, as well as values derived from a common environmental model frequently employed in industrial applications.”
Line 32-33: It may be better like this “Pesticides are a group of compounds whose fate and behavior in the atmosphere are less studied and characterized compared to their behavior in soil, surface water, and groundwater environments.” The original sentence creates confusion by comparing the "environment" (which includes soil, surface water, and groundwater) with specific parts of the environment.
Line 1-83: This is verbose and needs to be more concise.
Line 97-106: Instead, it would be helpful to provide a more detailed introduction to the various methods used for vapor pressure measurement, as readers may be particularly interested in this aspect.
Figures: Figures are not well presented in this manuscript.
2. The workload of this study may not be sufficient for an ACP paper. I recommend expanding the scope by measuring more pesticides. For example, in Figure 6, we see that the Tmax for most compounds falls between 25-50°C.
There are also several alternative approaches the authors could take to make the paper more impactful and insightful:
- Conduct additional experiments if the current data is not ideal. For instance, with 2,4-D, if the one-order-of-magnitude difference is thought to be due to smaller particles, why not attempt to make the particle sizes more consistent with other measurements to confirm this assumption?
- Dive deeper into the experimental results. In Figure 2, why do the calibration curves for the syringe method show a more Gaussian distribution compared to the atomization method? In Figure 3b, why is linear fitting done using only a few points from Figure 3a? What are the R² values for those two fits? While I understand that the syringe method might shift Tmax due to droplet effects, does this mean the syringe method cannot provide a better linear relationship? As discussed in the paper, particle size also influences the derived vapor pressure. We really need to show more details of the experimental results before making a conclusion.
- Extend your experiments. How might organic mixtures affect volatility measurements? Traditionally, pure organics are used to assess volatility, but if we use mixtures at temperatures up to 200°C, will interactions between different compounds influence volatility? Additionally, is it possible to introduce these compounds into inorganic aerosol particles to examine how inorganics impact Tmax measurements?
3. There is no Supporting Information for this manuscript.
Minor comments:
Line 1: We can consider adding Figaero in the title. Make it more specific.
Line 17: I’m not sure if it’s appropriate to highlight "first time" here. I wouldn’t claim this is the first time particle-phase pesticides have been measured with mass spectrometry. Does the article below cover particle-phase measurements? Please verify this through a thorough literature review if you still wish to use “first time.” Additionally, this phrasing may cause some confusion, as it suggests pesticides were measured in field particles, whereas the study involves measuring compounds from generated particles. In other words, as long as the vapor pressure of these compounds can be measured accurately, whether the measurements are taken online or offline is not critical to this study.
Barker, Z., Venkatchalam, V., Martin, A. N., Farquar, G. R., & Frank, M. (2010). Detecting trace pesticides in real time using single particle aerosol mass spectrometry. Analytica chimica acta, 661(2), 188-194.
Line 26: A lower Tmax may be corresponded to higher measured vapor pressure?
Line 62: Please add a citation for the sentence “In terms of current EU regulatory context, ….”
Line 78-80: We need add some citations for the sentences “there has been relatively much less attention on the fate and behaviour…”, since I believe there should be some studies in this direction.
Line 141: Where did you find this 50%-50% definition for the C*. Please cite it. I feel this is not correct.
Line 196: A polonium source should be between CH3I flow and IMR.
Here I stopped looking for more minor comments and I would like to leave future work to authors.
Citation: https://doi.org/10.5194/egusphere-2024-2380-RC3
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
294 | 72 | 55 | 421 | 3 | 4 |
- HTML: 294
- PDF: 72
- XML: 55
- Total: 421
- BibTeX: 3
- EndNote: 4
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1