the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Jan 2024
A multi-instrumental approach for calibrating two real-time mass spectrometers using high performance liquid chromatography and positive matrix factorization
Abstract. Obtaining quantitative information from real-time soft-ionization aerosol instruments such as an Extractive Electrospray time-of-flight Mass Spectrometer (EESI) can be challenging, due to many individual species having different, and often hard to predict, sensitivities. Directly calibrating is time-consuming and relevant standards are often hard to obtain. In addition, the molecular identities of many of the sampled species may be ambiguous. Bulk OA sensitivities are sometimes used to estimate molecular sensitivities, but different types of OA can have bulk sensitivities that vary by a factor of ~10. A system to separate the compounds present in complex samples can enable their direct calibration. Here, high performance liquid chromatography (HPLC) followed by aerosol formation via atomization was combined with online, 1 Hz measurements to calibrate the EESI and a High Resolution Aerosol Mass Spectrometer (AMS) for compounds present in a secondary organic aerosol (SOA) mixture. Pure compounds were used to test the method and characterize its uncertainties. Pure compound calibration factors were consistent within ±20 % for direct atomization vs. HPLC separation, which is far superior to the orders of magnitude sensitivity differences that are possible with EESI. For species that were not well separated by chromatography, Positive Matrix Factorization (PMF) based on AMS spectra was used to test its ability to separate overlapping species. In two test cases, further separation was achieved using PMF, but derived sensitivities from direct and HPLC calibrations varied by up to a factor of 2.
- Preprint
(2266 KB) - Metadata XML
-
Supplement
(3431 KB) - BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on ar-2023-21', Anonymous Referee #1, 27 Jan 2024
This manuscript presents an extremely thorough study where the authors combine and compare several aerosol instruments together with an HPLC. While the technical aspects of the study, both instrumental and analytical (e.g. PMF analyses) show that the authors clearly know what they are doing, the paper at the same time becomes very technical, and as such, can be tough to follow at times. The target of the study relates to improving calibrations of complex organic aerosol mixtures using state-of-the-art mass spectrometers, which is a very important topic, in particular for the EESI system. However, while there is discussion on the needs for improved calibrations, I find the exact aim, and in particular the main conclusion, largely lacking. Overall, this manuscript presents detailed results from a good experimental study, and I find them suitable for publication in AR, but the manuscript still requires some amount of work to make it more readable and to make more clear what message the authors want to convey with this work. I list my main comments below.
- Purpose and main result. Both from the abstract, the introduction, and the conclusions, the sections largely state that existing calibration methods are imperfect, followed by what was (or will be) done in this study. Thus, a clear aim is not clearly outlined, and therefore the conclusions are missing the part where the authors state whether the study was successful or not. The conclusions now state things like “approximate calibration factors were obtained” and “we introduce a new technique for better quantifying instrument responses”, but it still remains unclear how much this (quite complicated) approach improved the quantification of OA mixtures, if at all.
- Both abstract and conclusions highlight that the new method matches within 20% with single compound standards, if measured from a simple mixture, but what is the added value then compared to simply running those standards? This 20% is compared to uncertainties of orders of magnitude, but this seems to be comparing quite different things, and should not be taken as a measure of the improvement provided by this new method. The orders of magnitude difference in EESI sensitivity is inherent to that method, and cannot be decreased by any kind of calibration.
- The word “real-time” is mentioned in the title and first sentence of the abstract, but this manuscript only shows offline analyses. It also remains unclear to me whether this offline analysis is meant to be a permanent installation to be used also during actual sampling with the EESI or AMS, or simply for calibration. And if only for calibration, would it still need to be included at e.g. a field site, with periodic switching to the HPLC? It seems that if the HPLC calibration is not done for every sample/mixture, the calibrations will have limited use, as the ions in the EESI spectra will represent different isomer distributions for every different mixture. Again, it remains unclear what the purpose and aimed use of this new method ultimately is. What do the authors recommend or expect for others to potentially use this approach in future?
- EESI quantification. Perhaps this relates to the fact that this manuscript is very technical overall, but I was not able to follow how the EESI calibration factors (e.g. Table 2) were determined. For a well-separate ions, the SMPS signal can be compared to the EESI ion, as they both contain the total measured signal. But in the SOA mixture, only some 7 ions were selected for the comparison, but I would assume that the EESI sees hundreds of ions with many of them overlapping in this same range, and these 7 only contribute a fraction of the total signal. Still, they are compared to the PMF factors that together sum up to the entire signal from the SOA. This should be clarified better. I also urge the authors to consider which technical details are important to present in the main text, and what could potentially be moved to the SI, to make the overall story easier to follow.
Specific comments:
- All the signals in the AMS are very high. How much SOA was collected on the filter in the first place to produce such high concentrations after HPLC separation and atomization?
- Line 238: “atomic ratio of oxygen plus nitrogen to carbon (O+N:C)”. Should this be (O+N):C? If not, then I would not call it a ratio.
- Figure 6 K&L: There seems to be a distinct signal at m/z 60 in both these factors. Does that not suggest that part of the levoglucosan signal has been split into these factors?
- Line 507: The elemental formulas would be interesting to see also in the main text.
- Lines 545-555: Is all this really shown in Table S6? I don’t see anything about the sensitivities there.
Citation: https://doi.org/10.5194/ar-2023-21-RC1 - Purpose and main result. Both from the abstract, the introduction, and the conclusions, the sections largely state that existing calibration methods are imperfect, followed by what was (or will be) done in this study. Thus, a clear aim is not clearly outlined, and therefore the conclusions are missing the part where the authors state whether the study was successful or not. The conclusions now state things like “approximate calibration factors were obtained” and “we introduce a new technique for better quantifying instrument responses”, but it still remains unclear how much this (quite complicated) approach improved the quantification of OA mixtures, if at all.
-
RC2: 'Comment on ar-2023-21', Anonymous Referee #2, 30 Jan 2024
The article of Schueneman et al. introduces a new approach for calibrating two real-time mass spectrometers, EESI and AMS by using the liquid chromatograph and positive matrix factorization.
This method is innovative and provides a novel approach for the calibration of mass spectrometers for individual organic species that are difficult to calibrate by using a traditional atomizer, DMA and CPC setup. It seems that this method is more beneficial for the EESI but can also improve the data analysis of the AMS for the unknown organic species. This manuscript is well-written and fluent but it is rather demanding to read as it has so many technical details and the terminology is difficult to follow at times. This article is suitable for publication in Aerosol Research after minor revision.
General comments
I suggest making the article more concise ja check that the terminology is unambiguous. I also recommend summarizing in Conclusion section the benefits and the limitations of the method (for example this method may not be suitable for volatile species). As already mentioned, this method seems to be more advantageous for the EESI but what are the benefits for the HR-AMS. Does it enable the detection of species at a molecular level in real atmospheric/emission samples?
Specific comments:
Page 1 (abstract), line 17; “enable their direct calibration”, this direct calibration does not refer to same direct calibration as that used earlier in the abstract and later in the manuscript (monodisperse aerosol with an atomizer)?
page 5, line 117; “the SOA is unlikely to have changed over this period (one year)”, this is somewhat contradictory as it is said in Page 22 that the differences in the UV-Vis chromatograms could potentially rise due to the age of the SOA extract used here (~1 year)
page 8, lines 207-208; I suggest to change “calibration factor CFxE“ to “EESI calibration factor CFxE”. There is also one parenthesis missing.
page 9, lines 119-220; “background remained < 2 µg/m3”. That sounds quite large to me. How much was it relative to the peak concentrations? What are the detection limits of this method for the species? What was the source of the background signal, only solvent?
page 9, line 228-229; I suggest changing “response factor” to “AMS response factor”
page 9, line 230-231; “An RIE of 1.4 is typically applied to ambient aerosol.” I don’t see any reason to use a single RIE value for total ambient aerosol. Should this be “to ambient OA”?
page 11, lines 315-316; I suggest adding the instruments used to calculate composition-dependent density (SMPS, AMS) as they are mentioned in all the other cases.
Page 13, Section 3.1. Mass Balance of the Analyte in the Experimental System; this section could be moved to Supplemental material
Page 13-14; lines 358-363; The description of the direct calibration method should be in the methods section.
Page 14, Figure 3 caption; What is the difference between uncalibrated and raw data? “Monodisperse calibration factor” needs to be explained.
Page 18, Figure 5; If this is a 6-factor solution, what is the 6th factor? Figure shows the time series and mass spectra only for 5 factors. Same comment for Figure 6, what are the other factors?
Page 22, Table 2; Column title “AMS PMF factor(s)” is a bit unclear, is it the number of PMF factors or Associated PMF factor(s)?
Page 24, line 548; Fig. S3, is this the correct figure number?
Citation: https://doi.org/10.5194/ar-2023-21-RC2 -
AC1: 'Comment on ar-2023-21', Melinda Schueneman, 01 Apr 2024
The comment was uploaded in the form of a supplement: https://ar.copernicus.org/preprints/ar-2023-21/ar-2023-21-AC1-supplement.pdf
Status: closed
-
RC1: 'Comment on ar-2023-21', Anonymous Referee #1, 27 Jan 2024
This manuscript presents an extremely thorough study where the authors combine and compare several aerosol instruments together with an HPLC. While the technical aspects of the study, both instrumental and analytical (e.g. PMF analyses) show that the authors clearly know what they are doing, the paper at the same time becomes very technical, and as such, can be tough to follow at times. The target of the study relates to improving calibrations of complex organic aerosol mixtures using state-of-the-art mass spectrometers, which is a very important topic, in particular for the EESI system. However, while there is discussion on the needs for improved calibrations, I find the exact aim, and in particular the main conclusion, largely lacking. Overall, this manuscript presents detailed results from a good experimental study, and I find them suitable for publication in AR, but the manuscript still requires some amount of work to make it more readable and to make more clear what message the authors want to convey with this work. I list my main comments below.
- Purpose and main result. Both from the abstract, the introduction, and the conclusions, the sections largely state that existing calibration methods are imperfect, followed by what was (or will be) done in this study. Thus, a clear aim is not clearly outlined, and therefore the conclusions are missing the part where the authors state whether the study was successful or not. The conclusions now state things like “approximate calibration factors were obtained” and “we introduce a new technique for better quantifying instrument responses”, but it still remains unclear how much this (quite complicated) approach improved the quantification of OA mixtures, if at all.
- Both abstract and conclusions highlight that the new method matches within 20% with single compound standards, if measured from a simple mixture, but what is the added value then compared to simply running those standards? This 20% is compared to uncertainties of orders of magnitude, but this seems to be comparing quite different things, and should not be taken as a measure of the improvement provided by this new method. The orders of magnitude difference in EESI sensitivity is inherent to that method, and cannot be decreased by any kind of calibration.
- The word “real-time” is mentioned in the title and first sentence of the abstract, but this manuscript only shows offline analyses. It also remains unclear to me whether this offline analysis is meant to be a permanent installation to be used also during actual sampling with the EESI or AMS, or simply for calibration. And if only for calibration, would it still need to be included at e.g. a field site, with periodic switching to the HPLC? It seems that if the HPLC calibration is not done for every sample/mixture, the calibrations will have limited use, as the ions in the EESI spectra will represent different isomer distributions for every different mixture. Again, it remains unclear what the purpose and aimed use of this new method ultimately is. What do the authors recommend or expect for others to potentially use this approach in future?
- EESI quantification. Perhaps this relates to the fact that this manuscript is very technical overall, but I was not able to follow how the EESI calibration factors (e.g. Table 2) were determined. For a well-separate ions, the SMPS signal can be compared to the EESI ion, as they both contain the total measured signal. But in the SOA mixture, only some 7 ions were selected for the comparison, but I would assume that the EESI sees hundreds of ions with many of them overlapping in this same range, and these 7 only contribute a fraction of the total signal. Still, they are compared to the PMF factors that together sum up to the entire signal from the SOA. This should be clarified better. I also urge the authors to consider which technical details are important to present in the main text, and what could potentially be moved to the SI, to make the overall story easier to follow.
Specific comments:
- All the signals in the AMS are very high. How much SOA was collected on the filter in the first place to produce such high concentrations after HPLC separation and atomization?
- Line 238: “atomic ratio of oxygen plus nitrogen to carbon (O+N:C)”. Should this be (O+N):C? If not, then I would not call it a ratio.
- Figure 6 K&L: There seems to be a distinct signal at m/z 60 in both these factors. Does that not suggest that part of the levoglucosan signal has been split into these factors?
- Line 507: The elemental formulas would be interesting to see also in the main text.
- Lines 545-555: Is all this really shown in Table S6? I don’t see anything about the sensitivities there.
Citation: https://doi.org/10.5194/ar-2023-21-RC1 - Purpose and main result. Both from the abstract, the introduction, and the conclusions, the sections largely state that existing calibration methods are imperfect, followed by what was (or will be) done in this study. Thus, a clear aim is not clearly outlined, and therefore the conclusions are missing the part where the authors state whether the study was successful or not. The conclusions now state things like “approximate calibration factors were obtained” and “we introduce a new technique for better quantifying instrument responses”, but it still remains unclear how much this (quite complicated) approach improved the quantification of OA mixtures, if at all.
-
RC2: 'Comment on ar-2023-21', Anonymous Referee #2, 30 Jan 2024
The article of Schueneman et al. introduces a new approach for calibrating two real-time mass spectrometers, EESI and AMS by using the liquid chromatograph and positive matrix factorization.
This method is innovative and provides a novel approach for the calibration of mass spectrometers for individual organic species that are difficult to calibrate by using a traditional atomizer, DMA and CPC setup. It seems that this method is more beneficial for the EESI but can also improve the data analysis of the AMS for the unknown organic species. This manuscript is well-written and fluent but it is rather demanding to read as it has so many technical details and the terminology is difficult to follow at times. This article is suitable for publication in Aerosol Research after minor revision.
General comments
I suggest making the article more concise ja check that the terminology is unambiguous. I also recommend summarizing in Conclusion section the benefits and the limitations of the method (for example this method may not be suitable for volatile species). As already mentioned, this method seems to be more advantageous for the EESI but what are the benefits for the HR-AMS. Does it enable the detection of species at a molecular level in real atmospheric/emission samples?
Specific comments:
Page 1 (abstract), line 17; “enable their direct calibration”, this direct calibration does not refer to same direct calibration as that used earlier in the abstract and later in the manuscript (monodisperse aerosol with an atomizer)?
page 5, line 117; “the SOA is unlikely to have changed over this period (one year)”, this is somewhat contradictory as it is said in Page 22 that the differences in the UV-Vis chromatograms could potentially rise due to the age of the SOA extract used here (~1 year)
page 8, lines 207-208; I suggest to change “calibration factor CFxE“ to “EESI calibration factor CFxE”. There is also one parenthesis missing.
page 9, lines 119-220; “background remained < 2 µg/m3”. That sounds quite large to me. How much was it relative to the peak concentrations? What are the detection limits of this method for the species? What was the source of the background signal, only solvent?
page 9, line 228-229; I suggest changing “response factor” to “AMS response factor”
page 9, line 230-231; “An RIE of 1.4 is typically applied to ambient aerosol.” I don’t see any reason to use a single RIE value for total ambient aerosol. Should this be “to ambient OA”?
page 11, lines 315-316; I suggest adding the instruments used to calculate composition-dependent density (SMPS, AMS) as they are mentioned in all the other cases.
Page 13, Section 3.1. Mass Balance of the Analyte in the Experimental System; this section could be moved to Supplemental material
Page 13-14; lines 358-363; The description of the direct calibration method should be in the methods section.
Page 14, Figure 3 caption; What is the difference between uncalibrated and raw data? “Monodisperse calibration factor” needs to be explained.
Page 18, Figure 5; If this is a 6-factor solution, what is the 6th factor? Figure shows the time series and mass spectra only for 5 factors. Same comment for Figure 6, what are the other factors?
Page 22, Table 2; Column title “AMS PMF factor(s)” is a bit unclear, is it the number of PMF factors or Associated PMF factor(s)?
Page 24, line 548; Fig. S3, is this the correct figure number?
Citation: https://doi.org/10.5194/ar-2023-21-RC2 -
AC1: 'Comment on ar-2023-21', Melinda Schueneman, 01 Apr 2024
The comment was uploaded in the form of a supplement: https://ar.copernicus.org/preprints/ar-2023-21/ar-2023-21-AC1-supplement.pdf
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 392 | 94 | 12 | 498 | 24 | 7 | 6 |
- HTML: 392
- PDF: 94
- XML: 12
- Total: 498
- Supplement: 24
- BibTeX: 7
- EndNote: 6
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1