the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 May 2024
Comparison of size distribution and electrical particle sensor measurement methods for particle lung deposited surface area (LDSAal) in ambient measurements with varying conditions
Abstract. During recent years it has become evident that additional metrics along the particle mass concentration together with dense air quality monitoring networks within cities are needed to understand the most efficient ways to tackle the health burden of particulate pollution. Particle lung deposited surface area (LDSAal) is a metric to estimate particle exposure in the lung alveoli, and it has gained interest as a parameter for air quality monitoring as it is relatively easy and cost-efficient to measure with electrical particle sensors. Also, various studies have indicated its potential as a health-relevant metric. In addition to the electrical particle sensors, LDSAal can be measured with various size distribution methods. However, different LDSAal measurement methods have fundamental differences in their operation principles e.g., related to the measurement size ranges, size-classification or conversion from the originally measured quantity into LDSAal. It is not well understood how these differences affect the accuracy of the measurement in ambient conditions where especially the particle effective density and hygroscopicity can considerably change the particle lung deposition efficiencies. In this study, the electrical particle sensor measurement (Partector) and two size distribution approaches (ELPI+, DMPS/SMPS) were compared in road traffic environments with different environmental conditions in Helsinki and Prague. The results were compared by utilising general assumptions of LDSAal measurement (spherical hydrophobic particles with the standard density) and by evaluating the effects of the particle effective density and hygroscopicity. Additionally, the Partector and ELPI+ were compared in various urban environments near road traffic, airport, river traffic and residential wood combustion. The results show that comparison of different LDSAal measurement methods can be complicated in ambient measurements. The challenges were especially related to the estimated lung deposition of accumulation mode particles roughly larger than 200–400 nm. On the other hand, the results suggest that the differences between the methods are reasonably low when considering only ultrafine and soot particles, highlighting the suitability of LDSAal as a monitored metric when estimating spatial differences in the particulate pollution within cities.
- Preprint
(1086 KB) - Metadata XML
-
Supplement
(865 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on ar-2024-13', Anonymous Referee #1, 23 Jun 2024
The manuscript presents a comprehensive comparison of different methods for measuring particle lung deposited surface area (LDSA) in ambient air, focusing on the challenges and uncertainties associated with these measurements. The study provides valuable insights into the performance of various LDSA measurement techniques under different environmental conditions and particle characteristics. The authors have conducted a thorough analysis, considering factors such as particle effective density and hygroscopicity, which are often overlooked in LDSA measurements.
General Comments:
- The abstract could benefit from a more detailed explanation of the challenges in estimating lung deposition of accumulation mode particles and the acceptable differences between methods when considering only ultrafine particles (UFP) and soot.
- Some statements in the introduction require revision or additional context to avoid oversimplification.
- The manuscript would benefit from more quantitative analysis to support some of the key conclusions, particularly regarding the effects of particle effective density and hygroscopicity corrections.
- Some technical aspects of the measurement devices and data processing methods need further clarification.
- The discussion of the impact of particle effective density and hygroscopicity on LDSA measurements needs to be supported by additional data analysis.
Specific Comments:
- Line 64: The statement about PM2.5 being relatively more harmful near local pollution sources like traffic is oversimplified and should be revised or removed.
- Line 70: The concept of LDSA should be more precisely defined, acknowledging that it can refer to different regions of the respiratory system, not just the lung alveoli. It should also be clarified that LDSA refers to a surface area concentration.
- Line 133: More information about the Partector's design, particularly regarding ion trapping, would be helpful to understand potential influences on the charging current and calculated LDSA. The authors should address how the extrinsic charging efficiency, which is affected by particle losses in the charger, impacts the measured charging current and, subsequently, the calculated LDSA. This is crucial because particle losses can significantly alter the relationship between the charging current and the actual LDSA, potentially leading to inaccuracies in the final LDSA measurements.
- Line 147: The statement about ELPI+ measurement requiring estimation of particle effective density needs clarification or correction. ELPI+ particle size distribution measurements are based on aerodynamic sizing, which inherently incorporates information about particle density. Therefore, it's not immediately clear why additional estimation of particle effective density would be required. The authors should explain this apparent discrepancy or revise their statement if it's not accurate. If there are specific reasons why effective density estimation is still necessary for LDSA calculations with ELPI+ data, these should be clearly explained.
- Line 155: A brief description of the ICRP model used for the lung deposition function should be included.
- More detailed information about the conversion factor/process for the Partector is needed better to evaluate the differences between its measurements and other methods.
- Table 1: The low PN concentration and high density reported for Prague require attention/explanation and careful interpretation.
- Lines 343-346: The conclusion about uncertainty related to particle effective density estimation needs stronger support from the data presented. Several issues should be addressed:
- The unusually low PN concentration in Prague needs explanation.
- The limitations of using an average constant effective density over a wide size range should be discussed more thoroughly.
- A sensitivity analysis showing how variations in effective density across different size ranges impact LDSA calculations would strengthen the argument.
- The significant variation between ELPI+ and SMPS size distributions in Prague, especially for accumulation mode particles, warrants a more detailed explanation, considering factors beyond effective density.
- Line 359: Quantitative analysis should be provided to support the conclusion about decreased differences after ρeff-correction.
- Figure 3: The opposite trends of overestimation and underestimation for ELPI+ and SMPS require more in-depth discussion.
- Lines 408-412: The statement about hygroscopicity-corrected size distributions for SMPS and DMPS should be reconsidered:
- The changes before and after corrections may be more related to particle size distribution than hygroscopicity or chemical composition.
- A more detailed analysis of how the PSD itself influences the observed changes after hygroscopicity correction is needed.
- The relative importance of hygroscopicity versus PSD in determining the final LDSA values should be discussed.
- Stronger quantitative support is needed for the conclusion about the agreement between methods after hygroscopicity correction.
- Line 497: The conclusion about neglected hygroscopicity not considerably changing the results due to balancing effects should be presented more cautiously, acknowledging that it may only be valid under certain conditions.
Technical Corrections:
- Line 312: Add the Wu et al. (2023) reference to the reference list.
- Check for consistency in terminology throughout the manuscript, particularly in the use of LDSA and LDSAal.
Citation: https://doi.org/10.5194/ar-2024-13-RC1 - AC1: 'Reply on RC1', Teemu Lepistö, 28 Jun 2024
-
RC2: 'Comment on ar-2024-13', Anonymous Referee #2, 08 Jul 2024
The manuscript describes a comprehensive study on the measurement of the alveolar lung deposited surface area (LDSA) concentration at different urban locations. LDSA is generally considered to be a more health relevant concentration metric than the usually measured mass or number concentrations. Since it can readily and relatively cheaply be measured by unipolar diffusion charging of particles, followed by a measurement of the current caused by the particle-borne charges, it has raised increased interest over the last years, among others in the long-term monitoring of atmospheric particles.
In the present study, LDSA concentrations were measured by using both an electrical diffusion charger device (Partector), which delivers LDSA concentrations directly as well as size distribution measurement devices (ELPI, SMPS, DMPS), which were used to calculate the LDSA concentrations from the measured size distributions. Due to concurrent measurements of the number size distributions based on the aerodynamic and mobility diameter, the authors obtained information on the effective density of the particles and used this to correct the measured data. Further, the authors investigated to what extent particle hygroscopicity may affect LDSA measurements. Measurement uncertainties caused by these two properties have thus far not been considered and provide useful added value to the scientific literature.
While the manuscript is well-written and timely, it unfortunately suffers from several shortcomings. I suggest that the manuscript should undergo a major revision before it can be accepted for publication.
Main criticism:
- The authors correctly point out that the LDSA-metric suffers from the lack of a proper definition. The lung deposition efficiency depends on various breathing parameters, as well as age and sex of an individual. Furthermore, different models as well as approximations to the model results exist. To the best of my knowledge, the only device, for which all parameters have been fully disclosed, is the TSI NSAM, which has been calibrated to mimic the breathing parameters of a “reference worker”, applied in the ICRP model (Fissan et al., 2007). Here, the authors applied different breathing parameters to an approximation of the same model, published in the Hinds textbook. However, the textbook claims that this approximation is only accurate to within +/- 0.03, which means that the approximation is particularly inaccurate in the important size range of the accumulation mode, where the deposition efficiencies are rather low. Why has this approximation still been used? Why not the KDEP computer model, which is a free software, that calculates the values of the ICRP model much more accurately?
- How were the breathing parameters for the approximation chosen?
- On another note: the relatively new draft CEN/TS 18073 finally defines a convention for the breathing parameters and lung deposition values to be applied in LDSA devices of the future.
- It is not clear to me, how the correction for the effective density was applied. Only to transfer from aerodynamic to mobility diameter and then calculate the surface area (and lung deposition efficiency?) based on this value? The original ICRP model allows for differentiating between the diffusional deposition, for which the mobility diameter would be of relevance, and the mass-driven deposition, for which the aerodynamic diameter is of relevance. Has this difference been considered?
- Similarly, I wonder how the hygroscopicity correction has been applied. Which rh has been assumed? 100% as in the respiratory tract? Was the grown diameter only been used to determine the lung deposition or also the surface area?
In fact, I have been wondering for a relatively long time, how relevant the effect of hygroscopicity is in general for the relevant surface area. All the relevant toxicological studies that I know have shown that surface area is a good predictor for health outcomes of non-soluble particles only, whereas soluble particles would dissolve in the lung and consequently their mass would be the relevant metric. Would thus the surface area of a hygroscopically grown particle be of relevance? - The comparisons of the LDSA values are only based on averages and 25th/75th percentiles, which hides a lot of information. I think it would be more informative, if the data (e.g. 1 h averages) are presented as scatter plot diagrams.
- Why are the BC and NO measurements mentioned in the text and data shown on Table 1, but not further dicussed? Either add a discussion (if they provide a benefit to this study) or remove any text on these two metrics.
- Each of the measurement campaigns has been relatively short (few weeks at most) and all were during winter/early spring. I therefore wonder, how representative the data can be. A disclaimer should be added, mentioning this shortcoming and that more research over longer periods and covering different seasons is needed.
Specific points:
Line 93/94: The electrical particle sensors do not need to assume a size distribution due to the similarity of the charging efficiency and the lung deposited surface area per particle (see e.g. Todea et al., 2015), which only holds in a size range from approximately 20 to 400 nm. Note that the analysis in this paper is based on the ICRP model with the breathing parameters used by TSI for NSAM.
Line 100 ff.: Whereas the effect of hygroscopicity and effective density on the lung deposition are extensively discussed here, a discussion on the effect of (individual) breathing parameters is missing.
Line 140 ff.: Did you try to use the total current (e.g. in size range 20-400 nm), measured by ELPI directly to determine the LDSA concentration instead of calculating it from the size distribution? Since ELPI uses a unipolar diffusion charger with in principle similar charging characteristic as the Partector (I assume), this may directly yield the LDSA concentration.
Line 171: “…converted from the number concentration” should read “…converted from the number size distribution”.
Line 173: The Partector does not assume a size distribution (see above).
Line 302/303/Figures S8-S11: Why are these graphs termed “deviation” plots? Aren’t these simply histograms, showing how often (relatively) certain concentration ranges occurred?
Line 465/Figure 7: It is striking that the Partector/ELPI ratio is always below 1, showing that the difference must be systematic
Citation: https://doi.org/10.5194/ar-2024-13-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 193 | 52 | 21 | 266 | 32 | 8 | 10 |
- HTML: 193
- PDF: 52
- XML: 21
- Total: 266
- Supplement: 32
- BibTeX: 8
- EndNote: 10
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1