the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Characterization of water-soluble brown carbon chromophores from wildfire plumes in the western US using size exclusion chromatography
Lisa Azzarello
Rebecca A. Washenfelder
Michael A. Robinson
Alessandro Franchin
Caroline C. Womack
Christopher D. Holmes
Steven S. Brown
Ann Middlebrook
Tim Newberger
Colm Sweeney
Abstract. Wildfires are an important source of carbonaceous aerosol in the atmosphere. Organic aerosol that absorbs light in the ultraviolet to visible spectral range is referred to as “brown carbon” (BrC), and its impact on Earth’s radiative budget has not been well characterized. We collected water-soluble brown carbon using a particle into liquid sampler (PILS) onboard a Twin Otter aircraft during the Fire Influence on Regional to Global Environments and Air Quality (FIREX-AQ) campaign. Samples were collected downwind of wildfires in the western United States from August to September 2019. We applied size exclusion chromatography (SEC) with ultraviolet-visible spectroscopy to characterize the molecular size distribution of BrC chromophores. The wildfire plumes had transport ages of 0 to 5 h and the absorption was dominated by chromophores with molecular weights <500 Da. With BrC normalized to a conserved biomass burning tracer, carbon monoxide, a consistent decrease in BrC absorption with plume age was not observed during FIREX-AQ. These findings are consistent with the variable trends in BrC absorption with plume age reported in recent studies. While BrC absorption trends were broadly consistent between the offline SEC analysis and the online PILS measurements, the absolute values of absorption and their spectral dependence differed. We attribute this difference to the organic modifier used in the chromatographic separation and demonstrate how this affects the molecular structure of the compounds comprising BrC, with implications for interpretation of absorption measurement of BrC field samples.
- Preprint
(2852 KB) - Metadata XML
-
Supplement
(2958 KB) - BibTeX
- EndNote
Lisa Azzarello et al.
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2023-1128', Anonymous Referee #1, 13 Jun 2023
The manuscript examines water-soluble light-absorbing compounds in biomass burning fire plumes using on-line particle into liquid sampler (PILS) and offline size exclusion chromatography (SEC), both coupled to spectrophotometric detectors. The main messages of the manuscript are: (1) absorption coefficient of the BBOA particles sometimes increases and sometimes decreases with the plume age; (2) PILS and SEC data do not agree with each other due to effects of solvents on the absorption spectra of the analyzed chromophores. While the manuscript is potentially publishable, I have two major comments about the manuscript that will likely require a major revision or an even more drastic action.
1). The section describing the solvent effects will need to be significantly revised. The authors have misinterpreted the change in the spectrum of nitrocatechol as the effect of solvent polarity. Instead, this is quite simply an acid-base equilibrium between nitrocatechol (C6H5NO4, absorption peak at around 350 nm) and its anion (C6H4NO4-, absorption peak at around 450 nm). Common nitrophenols have pKa of the order of 7, leading to large differences in the absorption spectra recorded during LC separation using an acidified and non-acidified eluent. For example, see Figure 5 of Cornard et al. (2005), https://doi.org/10.1016/j.chemphys.2004.09.020, for the comparison of absorption spectra of the nitrocatechol and its anion. Also, see Figure S2 in Lin et al. (2017), https://doi.org/10.1021/acs.est.7b02276, which shows how spectra of eluted nitrocatechol and other nitrophenols change depending on the eluent pH.
From what I can gather from Figure 4 in this manuscript, the ammonium acetate buffer that the authors used for their SEC ACN+buffer experiments was sufficiently basic to significantly deprotonate nitrocatechol. In contrast, in their DIW only experiment, nitrocatechol was only partly deprotonated (there is a shoulder there corresponding to the 450 nm band of the anion but the majority of the nitrocatechol is not deprotonated). Was the buffer prepared to provide buffering at pH 5 or pH 9 in this work? I presume it is the latter. For more on this buffer and its dual pH buffering nature, I would recommend reading Konermann (2017), https://doi.org/10.1007/s13361-017-1739-3.
2). The strong pH dependence of the absorption spectra of nitrophenols (and some other brown carbon compounds) and the different acidities of working solvents used for the PILS and SEC portions of this work, make it very hard to faithfully compare the results obtained by these two methods. I presume that the complete lack of correlation between the two methods in Figure 3 must be at least in part due to these solvent acidity effects. Broadly speaking, this manuscript shows that choosing an inappropriate solvent for the measurements will lead to questionable results. Is this self-evident conclusion really worth publishing? Would the agreement be better if a more acidic buffer was used for the SEC portion of the work? Given that the atmospheric particles that are commonly acidic, why was a basic buffer selected for the separation? In my opinion these questions need to be carefully addressed before the manuscript can proceed to a publication. Additional experiments (and possibly a full re-analysis of samples with a different solvent for SEC) may be necessary to address these questions.
Citation: https://doi.org/10.5194/egusphere-2023-1128-RC1 -
RC2: 'Comment on egusphere-2023-1128', Anonymous Referee #2, 20 Jun 2023
This paper focuses on brown carbon determined in samples collected from a PILS during aircraft measurements as part of the Twin Otter component of the FIREX study. The BrC reported here is the light absorption at 300 nm, a wavelength somewhat lower than 365 nm, which is what is typically reported for BrC. This is stated to be due to extensive dilution of the sample. These offline samples are run through SEC to assess the molecular weight of the chromophores, but the focus seems to be mainly on the change in the spectral properties of BrC with different solvents needed for the SEC analysis. By comparing BrC from an online instrument (PILS) to SEC, the latter involving the addition of acetonitrile and ammonium acetate buffer, differences are observed. They conclude that solvents can affect the spectral properties of BrC. Other studies have noted that solution pH and organic solvents can cause this issue. A major limitation in this work is that the authors never compared the un-altered offline samples to the online samples to make sure that dilution or differences in sample handler did not cause spectral shift issues. There are discrepancies between what this paper reports and other research comparing solvent and online measurements of BrC that could be discussed in more detail. Finally, the authors might consider what their reference BrC measurement really is; what is considered the correct BrC measurement that reflects the characteristics of actual particles?
Specific Comments.
Typo line 102; edit: attribute assign.
The BrC of this study is defined somewhat differently than most other studies. Line 175 – to 177. This method focuses on light absorption in the 250 to 300 nm range since it is stated that the samples were too dilute to detect absorption above blanks at higher wavelengths. Does this affect the analysis? Is this a spectral range where BrC is optically important, from a climate/radiative forcing perspective, if not why concerned about it?
Analysis of BrC exclusively based on these low wavelengths is somewhat unusual. The authors have access to the online water-soluble species absorption data from the PILS. Have they compared the methods, ie run the offline samples prior to any alteration (addition of the buffer and acetonitrile) for the SEC analysis and compared the data to the PILS? This would provide a baseline, addressing possible issues such as differences in dilution, sampling handling, etc.
It seems that no offline measurement of BrC was made in this study without the addition of other solvents to the water samples? Is that correct? This should be clarified in the methods section. A plot like Fig 3 involving a direct comparison between the PILS and collected vials would be very informative and help interpret Fig 3.
Lines 180 to 203, one could also include the findings from the WeCan study, ie: Sullivan, A., R. P. Pokrhet, Y. Shen, S. M. Murphy, D. W. Toohey, T. Campos, J. Lindaas, E. V. Fischer, and J. L. Collett (2022), Examination of Brown Carbon Absorption from Wildfires in the Western U.S. During the WE-CAN Study, Atmos Chem Phys, 22, 13389-13406.
In Fig 2, define what absorption means. I assume it is the same as in Fig 1, light absorbance from the LWCC at 300 nm.
Related to the above discussion on lack of comparison between un-altered offline samples and PILS. Lines 282-283. This implies that there are no published comparisons between online and offline water-soluble BrC measurements. Is this true? I suggest a literature search. See for example Fig 8 in Zeng et al. (https://doi.org/10.5194/amt-14-6357-2021)
Related to the above is the question of published results comparing online vs solvent-extracted offline analysis, and the assertion that methanol can lead to artifacts (lines 307-309). It is noteworthy that this has not been seen in biomass burning plumes measured during FIREX, see Fig 5a (and supplemental Fig S5) in Zeng et al, https://doi.org/10.5194/acp-22-8009-2022. Maybe other solvents, such as those used in the SEC analysis, produce substantial changes; how does one explain these discrepancies?
A final point to consider is the idea that there is an ideal sampling method to measure BrC that does not alter the aerosol particle from its native state and so measured actual characteristics of BrC in an ambient particle. Taking pH as an example, the issue raised by the other reviewer, maybe solvent extractions give the best option since pH can be adjusted to that expected for the particles, whereas online methods, such as the PAS, dry the particles to reduce artifacts, which can drastically change particle pH. The question is, can the true spectral properties of an ambient particle be measured without alteration? If not, what is the reference that things should be compared to, or should the focus be on noting and understanding factors that can affect spectral properties? In this case, it seems the water-soluble BrC is the reference, but at what pH (dilute solution in equilibrium with air pH~5), which raises the issue if the PILS and vials collected gave similar results, as noted above.
Citation: https://doi.org/10.5194/egusphere-2023-1128-RC2 -
RC3: 'Comment on egusphere-2023-1128', Anonymous Referee #3, 21 Jul 2023
Review of “Characterization of water-soluble brown carbon chromophores from wildfire plumes in the western US using size exclusion chromatography” by Azzarello et al.
General comments: This manuscript presents results on a comparison of offline and online brown carbon (BrC) measurements for wildfire smoke collected from the Twin Otter aircraft during FIREX-AQ. The authors found that the BrC was dominated by smaller chromophores (<500 Da) and that there was not a consistent decrease in BrC with plume age (0-5 hours). They also found differences in the spectra between online and offline measurements and attribute this to solvent effects. Overall this is a very well written paper with a clear summary of the results. I have a few minor comments that should be addressed before this paper is accepted.
Specific comments:
- It is noted that there isn’t a consistent trend in BrC with plume age, as has been reported in other papers looking at these fires. Here the BrC is normalized to the average CO concentration. How much variation was observed in this CO mixing ratio during the sample collection period and how much variation was observed in general in the fire? How does this variation compare to previous studies using this normalization method?
- Looking at Figure 1, the black markers have a line fit to them that appears to extend beyond the data shown in the figure. Is there data not shown in the figure?
- For the solvent effects, pH has also been demonstrated to play an important role in shifts in the absorption of BrC. What was the pH (estimated) of the water for the PILS analysis. Do you expect a difference compared to the pH that is found in the water collected for offline analysis? The idea of pH for organics in organic solvents is complicated, but the pH of the water in the online measurements may play a role in the differences observed.
- Solvent effects can play a role in the position of the absorption, but I’ve seen less discussion around changing the overall intensity of the absorption. Do you think that the increase for the offline measurements can be attributed to this, or are there other possible reasons for that increase?
Citation: https://doi.org/10.5194/egusphere-2023-1128-RC3
Lisa Azzarello et al.
Lisa Azzarello et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 242 | 78 | 18 | 338 | 35 | 6 | 7 |
- HTML: 242
- PDF: 78
- XML: 18
- Total: 338
- Supplement: 35
- BibTeX: 6
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1