the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Aug 2023
SPARCLE 2: A new optical particle counter to measure particle size and refractive index
Abstract. In the last few decades, there has been an increasing need to improve in situ aerosol measurements to better understand the role of atmospheric aerosol on the Earth’s climate system and to assess the quality of ambient air. At the moment, in situ optically based aerosol photometers assume the refractive index of measured particles to give a size estimate. This assumption can result in large errors in estimated size. This study describes an instrument, called SPARCLE 2, that addresses this problem by simultaneously measuring the particle size and refractive index. SPARCLE 2 has two detectors to measure the pattern of aerosol-scattered light. The first is a detector with high sensitivity and high temporal resolution mainly to detect the presence of particles in the sensing volume. The second is a detector with high angular resolution to capture the pattern of scattered light. SPARCLE 2 is designed to measure particles whose diameters lie in the range 500 nm to 2,500 nm typical of the accumulation mode of troposphere aerosol. A theoretical lower particle size limit of 300 nm is determined by the optical and electronic noise. In practice, stray light limits the lower limit to a particle diameter of 800 nm. SPARCLE 2’s accuracy was tested using four monodisperse aerosols formed from non-absorbing polystyrene latex beads. The mean diameters of the test particles were 1,100 nm, 1,800 nm, 2,000 nm and 3,000 nm and their refractive index were 1.59. The standard deviation between SPARCLE 2’s measurement and the manufacturer’s stated size was 13 % for the 1,100 nm size particles and less than 3 % for the three larger sizes. The refractive index deviation was less than 1.25 % for all sizes. SPARCLE 2 was used to measure ambient Oxford (UK) city air. The size distributions measured by SPARCLE 2 were similar to those measured by a commercial optical particle counter. The refractive index distribution was consistent with the most abundant aerosol compositions around Oxford which are NO3−, NH4+ and SO42−.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1478 KB)
Interactive discussion
Status: closed
-
RC1: 'Comment on amt-2023-140', Anonymous Referee #1, 12 Sep 2023
Review of the manuscript titled: “SPARCLE 2: A new optical particle counter to measure particle size and refractive index”, by M. S. Romadhon, et al, submitted to the journal Atmospheric Measurement Techniques.
The authors report on the next generation of the SPARCLE instrument, an instrument that is designed to measure size and refractive index of individual aerosol particles. The refractive index of aerosol particles is an important property that affects the direct radiative effects but also hints at the particle chemical composition with further implications on aerosol chemistry, aerosol-cloud interactions and health effects. Measurements of this property are under-represented among atmospheric observations, mostly due to the difficulty of a precise retrieval. The main objective of the manuscript is therefore very relevant to the field of atmospheric research.
However, I can not recommend the manuscript for publication unless the following major concerns are addressed.
The authors present an instrument that claims to be able to measure size and refractive index of individual aerosol particles. The manuscript can be divided into three parts, an introduction to the instrument including a theoretical description of the retrieval process, the instrument's validation, and a measurement of ambient aerosol particles. I have no major concerns regarding the first part, but are convinced that the second part is incomplete and the last part is invalid in its present form.
Major concerns regarding part 2 (instrument validation):
- The instrument's main purpose is to determine the refractive index. However, the validation study determines this value of merely one material, which is polystyrene. The authors need to consider additional materials with a range of refractive indexes, e.g. Dioctyl sebacate (monodispersed with a DMA), Glass Particle Standards, etc. Ideally the authors would conduct an RH dependent observation of ammonium-sulfate (or similar) based aerosols, this would greatly benefit the ambient measurement in particular if the authors insist on conducting those on hygroscopically grown particles.
- The manuscript is lacking an uncertainty study. I believe the theoretical description of the retrieval can be utilized in a monte-carlo type uncertainty study. The manuscript would benefit from a conclusive comparison of SPARCLE2 to the other instruments that can retrieve the refractive index and which are mentioned in the manuscript’s introduction, to show where exactly SPARCLE2 excels over the others.
- The authors claim that the instrument is most efficient in the super-micron particle regime. Therefore the authors can not just refer to accumulation mode particles but need to consider coarse mode particles. This includes a discussion of effects from non-sphericity, preferably even an observation that shows the effect.
Major concerns regarding part 3 (ambient aerosol observations), which let me doubt the validity of this part of the manuscript and question its usefulness as is.
- During the observations RH is not well defined in the experiment which allows the authors to cherry pick the hygroscopic growth and thus the refractive index. In addition, the aerosol composition is diluted with a badly constrained amount of water making conclusion about the refractive index of the material in question impossible.
- Only 5 percent of particles that are sampled are actually processed of which more than 30 % give unrealistic results. Why that is and what it means regarding the quality of the SPARCLE2 obersvations is not critically discussed.
In the following I list further comments by their appearance in the manuscript.
- line 46 “The first goal is to provide size-resolved measurements of particles whose sizes range from 500 nm to 2,500 nm. This size range is within the range of the accumulation mode of tropospheric aerosol.”:
This sentence is misleading, often coarse mode particles (dust) are dominating the super-micron size range. Rephrasing required. - line 50 “SPARCLE 2 is suitable for ambient measurement since large proportions of ambient particles with sizes between 500 nm and 2,000 nm are either liquid droplets or particles with morphologies that have a dominant liquid phase (Prospero et al., 1983; Sullivan and Prather, 2005; Pinnick et al., 2011).”:
This statement is not true for accumulation mode particles under low humidity conditions and not true for many types of coarse mode particles under most conditions. Rephrasing required. - line 95 and 102 “... wept around a sphere …”: this seems a strange choice of words to me.
- line 98. Reference needed.
- Figure 3: I don’t see the need for this figure. It shows the same as Figure 6, doesn’t it? And the linear scaling is not beneficial.
- Figure 6: The usefulness of this figure is limited and missleading as the actual noise level is not shown. Also, please limit the plotting range to values relevant to this study.
- line 155 “Also shown in Figure 8.a, the background variation could be 0.4 V. Due to the variation, a threshold level of PMT outputs was set to be 0.4 V lower than the mean background to differentiate between background variation and scattering signals.“:
I do not understand what the authors want to say in these two sentences; please rephrase. - line 176 “Meanwhile, the shape of PMT responses are shown in the right of the figure.” This sounds like a very uncertain way to measure the extent of the beam. It depends on the knowledge of the particle speed, which I would assume varies a lot in particular since no sheath flow helps constrain the diameter of the sample flow. Why did the authors not use the TSL201R-LF?
- line 180: What is the “simple model” based on? Diffusion? Turbulence?
- Figure 12: Please discuss the origin of the interference pattern.
- Table 1: I believe it is relevant for this work (not just in table 1) to consider all PSL parameter, that is the nominal diameter (provided), the uncertainty of this diameter (this might be the std given in the manuscript) and the diameter distribution around the actual mean diameter (or this might be the std given in the manuscript). Please add and consider those parameters.
- line 211: I believe the manuscript would benefit if the authors would elaborate on this a bit more. What exactly is considered in y? What are the considered uncertainties? What goes into F(x)? I assume many of those things have been discussed above, but it is not clear how they are put together in this optimization procedure?
- line 223 “10,000 aerosols”: I believe this is supposed to mean “10,000 particles“?
- Figure 14: Please add uncertainties of the retrieved parameters (d and m) to this figure.
Citation: https://doi.org/10.5194/amt-2023-140-RC1 -
RC2: 'Comment on amt-2023-140', Anonymous Referee #2, 20 Sep 2023
The authors are reporting on an improved optical particle counting (OPC) instrument design attempting to retrieve both particle size and refractive index from ambient aerosol particles. Size-resolved particle refractive index measurements are very limited and a critical component necessary for a better understanding of aerosol radiative effects. This OPC design, designated “SPARCLE2”, is an improvement over a previous iteration of this instrument (SPARCLE1) and implements a PMT for sub-hemispherical integrated scattering (used for particle detection/triggering) and a linear CCD for capturing angularly resolved scattering, simultaneously. Scattering from the CCD is compared with a forward model to determine individual particle size and refractive index. The manuscript is separated into three sections, describing the instrument design, characterization, and results from ambient measurements collected at the University of Oxford.
In this manuscript’s present form, I do not recommend it for publication without major revisions addressing the following concerns to improve the reader’s confidence in the measurements being made.
The following are major concerns with this manuscript:
- The authors indicate the design is capable of measuring accumulation mode aerosol (0.1 to 1 um diameter particles). However, the authors indicate a lower size limit for particles 800 nm in diameter and Figure 12 leads the reader to presume the instrument is not able to make reliable measurements below ~ 1um unless the refractive index is greater than approximately 1.60, which would be difficult to attain in an ambient atmosphere if the particles are aqueous or have a dominant liquid phase (as indicated in Section 1). It is requested that the authors clarify the limitations and the relevance of these particles in the atmosphere. The upper end of the effective range is stated to be designed to 2,500 nm and validation is carried out with particles of size up to 3,000 nm. Results appear to be extrapolated reported out to 10,000 nm. While these are corrected using the measurement efficiency from Figure 12, it would be supportive to extend the validation particles out to these larger sizes.
- A small range of sizes of latex spheres (having approximately the same refractive indices) were used in the validation of this measurement. Using particles of other known refractive indices (e.g., ammonium sulfate, sodium sulfate, and sodium chloride), sizes, and/or shape factors as part of the validation process would bring confidence to the results and clarify limitations. Even without size selecting the particles, the atomized distribution of a single salt species with known refractive index should yield a horizontal line in figure 15.
- How well conditioned is the forward modeled response from the CCD? Are there no two particle state cases where the forward model results in approximately the same response (e.g., is it possible to have different combinations of particle size, refractive index, and/or x-z position yield a similar result)? Sensitivity to this would be important to state in this study.
Additional comments:
- The use of “refractive index” throughout the manuscript should be clarified up front to indicate that it is the real portion of the refractive index and that absorption may have other impacts on this measurement.
- Line 30 (L30): One such instrument relevant to this manuscript is the Differential Aerosol Sizing and Hygroscopicity Spectrometer Probe (DASH-SP; Sorooshian et al., 2008, https://doi.org/10.1080/02786820802178506). It is used to measure particle size and refractive index using OPCs in conjunction with a differential mobility analyzer (DMA). The instrument isolates dried particles based on electrical mobility diameter and then measures light scattering in an OPC to determine an effective dry refractive index. The particles are then hydrated and measured with a second OPC to determine the wet diameter size using the refractive index of the dry particle and water.
- L50: At the diameters which SPARCLE2 can measure, there are frequent occurrences of non-spherical, non-aqueous, particles in the atmosphere, including dust, crystalized sea salt, and some soot emissions. It would be helpful to recognize these in this manuscript and address how SPARCLE2 measurements might be impacted by these.
- L105: How are the individual pixel efficiencies calculated? Figure 5 shows the expected responses, but what are the responses to these size particles? Do the responses need to be scaled per pixel? Is the Fresnel correction the only correction applied?
- L107/Figure 4: Even though the pixel size represents a small solid angle, how much impact does the particle’s position in the beam path have on the scattering angle in theta? It may be useful to overlay this on Figure 4.
- L122-125/Figure 6: It would be useful to have Figure 6 be a two-panel plot with the same style plot for the 90 deg scattering pixel of the CCD. Any indication of the variance in the SNR between pixels of the CCD would be beneficial. Reduce the axes limits of Figure 6 to only include the noise limits and above in the Y axes and above ~100 nm in the x axes.
- L133: Double check the math on this maximum flow rate. At a Reynold’s number of 40, the flow should be closer to 5E-7m^3/s. Possibly a typo in the exponent of “500 x 10^-19 m^3/s”.
- L141: It may be more appropriate to calculate the transmission efficiencies at the limits of the sampling envelope.
- Table 1: The caption states that the refractive indices are 1.59, but states in Figure 14 that they range from 1.57-1.60. Additionally, it was stated on line 202 that four monodisperse test particle solutions were used; however, Table 1 indicates that one solution contains 1,800-2,200 nm particles. Also, are the standard deviations listed in the table from the manufacturer?
- L190: Further discussion on how the overall measurement efficiency can be improved with this technique would be helpful. How can the stray light and background signal be knocked down to improve the SNR for making measurements deeper into the accumulation mode?
- L227: It would be useful to put this 1.25% in terms of the atmospherically relevant range of refractive indices (~1.33-1.62).
- L233: Why were only <7% of the data analyzed with the fitting process?? How were these data selected over the other samples made?
- L241: A quick literature search resulted in multiple articles reporting refractive index distributions of ambient aerosol particles.
- Figure 15: Ambiguity using “The OPC” in the legend – label as “GRIMM OPC”?
Citation: https://doi.org/10.5194/amt-2023-140-RC2
Interactive discussion
Status: closed
-
RC1: 'Comment on amt-2023-140', Anonymous Referee #1, 12 Sep 2023
Review of the manuscript titled: “SPARCLE 2: A new optical particle counter to measure particle size and refractive index”, by M. S. Romadhon, et al, submitted to the journal Atmospheric Measurement Techniques.
The authors report on the next generation of the SPARCLE instrument, an instrument that is designed to measure size and refractive index of individual aerosol particles. The refractive index of aerosol particles is an important property that affects the direct radiative effects but also hints at the particle chemical composition with further implications on aerosol chemistry, aerosol-cloud interactions and health effects. Measurements of this property are under-represented among atmospheric observations, mostly due to the difficulty of a precise retrieval. The main objective of the manuscript is therefore very relevant to the field of atmospheric research.
However, I can not recommend the manuscript for publication unless the following major concerns are addressed.
The authors present an instrument that claims to be able to measure size and refractive index of individual aerosol particles. The manuscript can be divided into three parts, an introduction to the instrument including a theoretical description of the retrieval process, the instrument's validation, and a measurement of ambient aerosol particles. I have no major concerns regarding the first part, but are convinced that the second part is incomplete and the last part is invalid in its present form.
Major concerns regarding part 2 (instrument validation):
- The instrument's main purpose is to determine the refractive index. However, the validation study determines this value of merely one material, which is polystyrene. The authors need to consider additional materials with a range of refractive indexes, e.g. Dioctyl sebacate (monodispersed with a DMA), Glass Particle Standards, etc. Ideally the authors would conduct an RH dependent observation of ammonium-sulfate (or similar) based aerosols, this would greatly benefit the ambient measurement in particular if the authors insist on conducting those on hygroscopically grown particles.
- The manuscript is lacking an uncertainty study. I believe the theoretical description of the retrieval can be utilized in a monte-carlo type uncertainty study. The manuscript would benefit from a conclusive comparison of SPARCLE2 to the other instruments that can retrieve the refractive index and which are mentioned in the manuscript’s introduction, to show where exactly SPARCLE2 excels over the others.
- The authors claim that the instrument is most efficient in the super-micron particle regime. Therefore the authors can not just refer to accumulation mode particles but need to consider coarse mode particles. This includes a discussion of effects from non-sphericity, preferably even an observation that shows the effect.
Major concerns regarding part 3 (ambient aerosol observations), which let me doubt the validity of this part of the manuscript and question its usefulness as is.
- During the observations RH is not well defined in the experiment which allows the authors to cherry pick the hygroscopic growth and thus the refractive index. In addition, the aerosol composition is diluted with a badly constrained amount of water making conclusion about the refractive index of the material in question impossible.
- Only 5 percent of particles that are sampled are actually processed of which more than 30 % give unrealistic results. Why that is and what it means regarding the quality of the SPARCLE2 obersvations is not critically discussed.
In the following I list further comments by their appearance in the manuscript.
- line 46 “The first goal is to provide size-resolved measurements of particles whose sizes range from 500 nm to 2,500 nm. This size range is within the range of the accumulation mode of tropospheric aerosol.”:
This sentence is misleading, often coarse mode particles (dust) are dominating the super-micron size range. Rephrasing required. - line 50 “SPARCLE 2 is suitable for ambient measurement since large proportions of ambient particles with sizes between 500 nm and 2,000 nm are either liquid droplets or particles with morphologies that have a dominant liquid phase (Prospero et al., 1983; Sullivan and Prather, 2005; Pinnick et al., 2011).”:
This statement is not true for accumulation mode particles under low humidity conditions and not true for many types of coarse mode particles under most conditions. Rephrasing required. - line 95 and 102 “... wept around a sphere …”: this seems a strange choice of words to me.
- line 98. Reference needed.
- Figure 3: I don’t see the need for this figure. It shows the same as Figure 6, doesn’t it? And the linear scaling is not beneficial.
- Figure 6: The usefulness of this figure is limited and missleading as the actual noise level is not shown. Also, please limit the plotting range to values relevant to this study.
- line 155 “Also shown in Figure 8.a, the background variation could be 0.4 V. Due to the variation, a threshold level of PMT outputs was set to be 0.4 V lower than the mean background to differentiate between background variation and scattering signals.“:
I do not understand what the authors want to say in these two sentences; please rephrase. - line 176 “Meanwhile, the shape of PMT responses are shown in the right of the figure.” This sounds like a very uncertain way to measure the extent of the beam. It depends on the knowledge of the particle speed, which I would assume varies a lot in particular since no sheath flow helps constrain the diameter of the sample flow. Why did the authors not use the TSL201R-LF?
- line 180: What is the “simple model” based on? Diffusion? Turbulence?
- Figure 12: Please discuss the origin of the interference pattern.
- Table 1: I believe it is relevant for this work (not just in table 1) to consider all PSL parameter, that is the nominal diameter (provided), the uncertainty of this diameter (this might be the std given in the manuscript) and the diameter distribution around the actual mean diameter (or this might be the std given in the manuscript). Please add and consider those parameters.
- line 211: I believe the manuscript would benefit if the authors would elaborate on this a bit more. What exactly is considered in y? What are the considered uncertainties? What goes into F(x)? I assume many of those things have been discussed above, but it is not clear how they are put together in this optimization procedure?
- line 223 “10,000 aerosols”: I believe this is supposed to mean “10,000 particles“?
- Figure 14: Please add uncertainties of the retrieved parameters (d and m) to this figure.
Citation: https://doi.org/10.5194/amt-2023-140-RC1 -
RC2: 'Comment on amt-2023-140', Anonymous Referee #2, 20 Sep 2023
The authors are reporting on an improved optical particle counting (OPC) instrument design attempting to retrieve both particle size and refractive index from ambient aerosol particles. Size-resolved particle refractive index measurements are very limited and a critical component necessary for a better understanding of aerosol radiative effects. This OPC design, designated “SPARCLE2”, is an improvement over a previous iteration of this instrument (SPARCLE1) and implements a PMT for sub-hemispherical integrated scattering (used for particle detection/triggering) and a linear CCD for capturing angularly resolved scattering, simultaneously. Scattering from the CCD is compared with a forward model to determine individual particle size and refractive index. The manuscript is separated into three sections, describing the instrument design, characterization, and results from ambient measurements collected at the University of Oxford.
In this manuscript’s present form, I do not recommend it for publication without major revisions addressing the following concerns to improve the reader’s confidence in the measurements being made.
The following are major concerns with this manuscript:
- The authors indicate the design is capable of measuring accumulation mode aerosol (0.1 to 1 um diameter particles). However, the authors indicate a lower size limit for particles 800 nm in diameter and Figure 12 leads the reader to presume the instrument is not able to make reliable measurements below ~ 1um unless the refractive index is greater than approximately 1.60, which would be difficult to attain in an ambient atmosphere if the particles are aqueous or have a dominant liquid phase (as indicated in Section 1). It is requested that the authors clarify the limitations and the relevance of these particles in the atmosphere. The upper end of the effective range is stated to be designed to 2,500 nm and validation is carried out with particles of size up to 3,000 nm. Results appear to be extrapolated reported out to 10,000 nm. While these are corrected using the measurement efficiency from Figure 12, it would be supportive to extend the validation particles out to these larger sizes.
- A small range of sizes of latex spheres (having approximately the same refractive indices) were used in the validation of this measurement. Using particles of other known refractive indices (e.g., ammonium sulfate, sodium sulfate, and sodium chloride), sizes, and/or shape factors as part of the validation process would bring confidence to the results and clarify limitations. Even without size selecting the particles, the atomized distribution of a single salt species with known refractive index should yield a horizontal line in figure 15.
- How well conditioned is the forward modeled response from the CCD? Are there no two particle state cases where the forward model results in approximately the same response (e.g., is it possible to have different combinations of particle size, refractive index, and/or x-z position yield a similar result)? Sensitivity to this would be important to state in this study.
Additional comments:
- The use of “refractive index” throughout the manuscript should be clarified up front to indicate that it is the real portion of the refractive index and that absorption may have other impacts on this measurement.
- Line 30 (L30): One such instrument relevant to this manuscript is the Differential Aerosol Sizing and Hygroscopicity Spectrometer Probe (DASH-SP; Sorooshian et al., 2008, https://doi.org/10.1080/02786820802178506). It is used to measure particle size and refractive index using OPCs in conjunction with a differential mobility analyzer (DMA). The instrument isolates dried particles based on electrical mobility diameter and then measures light scattering in an OPC to determine an effective dry refractive index. The particles are then hydrated and measured with a second OPC to determine the wet diameter size using the refractive index of the dry particle and water.
- L50: At the diameters which SPARCLE2 can measure, there are frequent occurrences of non-spherical, non-aqueous, particles in the atmosphere, including dust, crystalized sea salt, and some soot emissions. It would be helpful to recognize these in this manuscript and address how SPARCLE2 measurements might be impacted by these.
- L105: How are the individual pixel efficiencies calculated? Figure 5 shows the expected responses, but what are the responses to these size particles? Do the responses need to be scaled per pixel? Is the Fresnel correction the only correction applied?
- L107/Figure 4: Even though the pixel size represents a small solid angle, how much impact does the particle’s position in the beam path have on the scattering angle in theta? It may be useful to overlay this on Figure 4.
- L122-125/Figure 6: It would be useful to have Figure 6 be a two-panel plot with the same style plot for the 90 deg scattering pixel of the CCD. Any indication of the variance in the SNR between pixels of the CCD would be beneficial. Reduce the axes limits of Figure 6 to only include the noise limits and above in the Y axes and above ~100 nm in the x axes.
- L133: Double check the math on this maximum flow rate. At a Reynold’s number of 40, the flow should be closer to 5E-7m^3/s. Possibly a typo in the exponent of “500 x 10^-19 m^3/s”.
- L141: It may be more appropriate to calculate the transmission efficiencies at the limits of the sampling envelope.
- Table 1: The caption states that the refractive indices are 1.59, but states in Figure 14 that they range from 1.57-1.60. Additionally, it was stated on line 202 that four monodisperse test particle solutions were used; however, Table 1 indicates that one solution contains 1,800-2,200 nm particles. Also, are the standard deviations listed in the table from the manufacturer?
- L190: Further discussion on how the overall measurement efficiency can be improved with this technique would be helpful. How can the stray light and background signal be knocked down to improve the SNR for making measurements deeper into the accumulation mode?
- L227: It would be useful to put this 1.25% in terms of the atmospherically relevant range of refractive indices (~1.33-1.62).
- L233: Why were only <7% of the data analyzed with the fitting process?? How were these data selected over the other samples made?
- L241: A quick literature search resulted in multiple articles reporting refractive index distributions of ambient aerosol particles.
- Figure 15: Ambiguity using “The OPC” in the legend – label as “GRIMM OPC”?
Citation: https://doi.org/10.5194/amt-2023-140-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 727 | 127 | 40 | 894 | 32 | 37 |
- HTML: 727
- PDF: 127
- XML: 40
- Total: 894
- BibTeX: 32
- EndNote: 37
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1