the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 02 Jun 2026
The influence of hydrogen addition on carbonaceous aerosols produced by an ethylene flame
Abstract. Combusting hydrogen alongside carbon-based fuels has been proposed to reduce CO2 emissions and combat climate change. However, combustion-generated aerosol particles can also cause a significant radiative forcing on climate. Since addition of novel fuels alters the combustion process, it also influences particle formation inside the fame and consequently the properties of the emitted aerosols. To investigate this, combustion-generated particles from various ethylene/hydrogen mixtures are sampled in the post-flame regime. The size distribution and light absorption properties of the particles are measured using a scanning mobility particle sizer (SMPS) and a multi-wavelength aethalometer. In addition, the particles are sampled on quartz-fiber filters and the mass concentrations of organic, elemental and total carbon (OC, EC, and TC) are measured using a thermo-optical OC-EC analyzer. The geometric mean diameter of the emitted particles decreased from 300 nm down to 150 nm upon increasing the hydrogen mole fraction in the fuel from 0 % to 50 %, while the EC/TC fraction decreased from 70 % to 35 %. The light absorption of methanol-dissolved OC were measured using UV-vis analysis, showing no dependence on flame parameters or fuel composition, and no significant light absorption at wavelengths larger than 500 nm. For combustion-generated particles, the mass absorption cross section σ of the total carbonaceous aerosol (i.e. the absorption coefficient normalized to TC mass concentration) is reported as a function of EC/TC ratio at wavelengths of 370, 590 and 880 nm. At a wavelength of 880 nm, σ is slightly higher than expected of an external mixture of OC and EC, indicating some absorption enhancement due to OC coating. At wavelengths of 590 and 370 nm, σ is much higher than that expected for a mixture of colorless OC and EC and this enhancement is attributed to light absorbing non-refractory species, also called brown carbon (BrC). The absorption Ångström exponent (370–660 nm) increased from 1.3 up to 3.8 with increasing hydrogen mole fraction in the fuel, especially at lower flame temperatures, indicating an increasing contribution of BrC to the light absorption of the emitted particles. It is concluded that BrC is a precursor to EC during particle formation, in line with the existing literature, and that it matures less efficiently into EC in the hydrogen containing flame.
- Preprint
(1240 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on ar-2026-21', Anonymous Referee #1, 13 Jul 2026
-
RC2: 'Comment on ar-2026-21', Anonymous Referee #2, 17 Aug 2026
Review of ar-2026-21
General comments
Van Rijn et al. investigated how adding hydrogen to ethylene changes the size, OC/EC fraction, and optical properties of the generated particles. The topic is interesting and relevant, and I agree that the manuscript fits the scope of Aerosol Research. My main concern is that several interpretations go beyond what the measurements can support. In particular, I am not convinced by the proposed condensation of OC after dilution, the evidence for internal mixing and lensing, or the conclusion that hydrogen inhibits the maturation of BrC into EC. Some of these interpretations may be possible, but they are not uniquely supported by the current data. I think the paper requires major revision before it can be accepted.
Major comments
- The comparison between sampling positions A and B is important to the paper, but these are not equivalent sampling conditions. At position A, the particles were collected at approximately 80 °C with relatively short sampling times and low flow rates. At position B, the aerosol was diluted, cooled, sampled for much longer, and collected at a much higher flow rate. A backup-filter correction was also applied only to the position-B samples. These differences could influence the measured OC and EC. The authors explain the lower EC/TC after dilution as condensation of gaseous organic compounds. I am not convinced that EC/TC alone can support this conclusion. Cooling may promote condensation, but dilution lowers the vapor concentration and usually shifts gas-particle partitioning toward evaporation. Particle and vapor losses in the dilution and sampling lines may also be important. Please show the absolute OC and EC concentrations at both positions, before and after correcting for the dilution factor. If the proposed explanation is correct, dilution-corrected EC should remain approximately constant while OC increases. Is that what the measurements show? The authors should also report the dilution ratio, residence time, tubing dimensions, and estimated particle losses. This issue also affects Figure 8. The α370–660 values were measured only after dilution, but the same optical measurements are compared with OC/TC measured at both positions A and B. Therefore, the shift between the two sets of points does not demonstrate that the additional OC at position B is colorless. Without optical measurements before dilution or a carbon mass balance, this conclusion should be removed or clearly identified as a hypothesis.
- I could not follow how C_TC in Eq. (5) was calculated. Was it obtained from the OC/EC filter measurements or from the SMPS? If it came from the filters, please explain how filter carbon loading was converted to aerosol mass concentration, including the sampled volume, filter area, blank correction, and dilution factor. If it came from the SMPS, please provide the assumed particle density and mobility-to-mass conversion.
- The methanol-extraction method is also not described in sufficient detail. Please provide the filter area, solvent volume, extraction time, extraction procedure, treatment of insoluble particles, blank correction, optical path length, and the equation used to convert solution absorbance to babs,dis. More importantly, the authors assume a methanol extraction efficiency of one. This is unlikely to be valid for all combustion-generated OC, particularly for darker BrC. Cheng et al. showed that the methanol-extraction bias increases as BrC becomes more strongly absorbing (https://doi.org/10.1080/02786826.2020.1820940). If σdis is normalized by total OC measured on the original filter, it is not strictly the mass absorption efficiency of methanol-dissolved OC. It is the extract absorption normalized by total filter OC. These are different quantities. The methanol-soluble OC mass should ideally be measured. If this is not possible, the terminology should be changed and the uncertainty associated with incomplete extraction discussed. OC1 + OC2, if that was used, should not be treated as equivalent to methanol-soluble OC without validation. How was α300–500 calculated? Since a full spectrum is available, it should be determined by fitting the power-law relationship over the selected wavelength range.
- The manuscript acknowledges that the default value may overestimate absorption, but then interprets relatively small deviations from the literature line as evidence of lensing. The magnitude of the proposed enhancement is comparable to the uncertainty associated with the Aethalometer correction. I suggest that the authors reprocess the raw DualSpot data using the internal-consistency diagnostic and alternative loading correction described by Poland et al. (https://doi.org/10.1080/02786826.2025.2587817). The attenuation range, loading-correction parameters, and data-screening criteria should be reported. A sensitivity analysis using reasonable C-factor values, for example 1.39, 2, and 3, would also show which conclusions remain valid. Unless the absorption measurement can be independently validated, the manuscript should refer to “apparent absorption enhancement” rather than treating the elevated σTC as direct evidence of lensing.
- The statement around lines 330–335 that a single-modal size distribution supports internal mixing is not correct. Externally mixed OC-rich and EC-rich particles can have overlapping mobility diameters and therefore appear as a single SMPS mode. An SMPS cannot determine composition-resolved mixing state. Likewise, the presence of an OC coating was not measured. Points above the external-mixture line could result from absorbing OC, variation in the intrinsic MAC of EC, Aethalometer artifacts, or actual lensing. Lensing is one possible explanation, but it is not demonstrated here. The authors should either provide direct coating or morphology evidence or using a thermodenuder or make the language much more cautious. The authors should also explain how the 550 nm literature MAC was converted to the external-mixture lines at 370, 590, and 880 nm, including the assumed wavelength dependence.
- I am concerned about the reliability of α calculated from only the 880 and 950 nm channels. These wavelengths differ by less than 8%, so relatively small channel-to-channel errors can produce a large error in α. For example, independent errors of only approximately 5% in the two absorption coefficients would result in an α uncertainty of roughly one. The manuscript reports a typical uncertainty of 10% for α880–950, but it is not clear how this was calculated. Please propagate the uncertainties in both absorption channels through Eq. (8). Confidence intervals should be shown for the individual data points. Based on the current analysis, I do not think an α880–950 value of 2.8 is sufficient evidence for methanol-insoluble OC absorption above 800 nm. Instrumental uncertainty, wavelength-dependent filter effects, particle morphology, and changes in EC optical properties are alternative explanations. This conclusion should be removed or substantially softened unless it can be supported by an independent optical measurement.
- Hydrogen addition changes more than hydrogen chemistry. It also changes the carbon input, precursor concentrations, burning velocity, modeled flame temperature, and probably the residence and coagulation histories of the particles. Changing vexit also changes both temperature and residence time. These effects are not separated in the present experiment. Please provide a table containing all fuel and air flow rates, carbon feed rates, modeled temperatures, and estimated residence times for each experiment. Total particle number and volume concentrations would also be helpful. The decrease in particle size could result from lower carbon availability or reduced coagulation, not necessarily delayed maturation. An inert-diluent experiment using He or N2 would be needed to separate the chemical effect of hydrogen from dilution and thermal effects. If such experiments are not available, the conclusions should be framed as correlations with hydrogen addition rather than a demonstrated chemical inhibition of maturation. Similarly, the conclusion that BrC is a precursor to EC is consistent with previous studies, but it is not directly demonstrated here. The particles were sampled at one post-flame height, and the study does not follow the same particles during maturation. Different-height sampling or independent structural measurements would be needed to demonstrate that transformation.
- The discussion in Section 4.2 suggests that “almost mature BC” can be classified as EC but have a lower σEC than fully mature BC. I find this interpretation problematic. EC measured using EUSAAR2 is an operationally defined quantity. It does not by itself demonstrate a graphitic structure or a particular degree of soot maturation. Please provide representative thermograms, filter loadings, pyrolyzed-carbon fractions, and information on the stability of the OC/EC split. This is particularly important for samples containing large amounts of tar-like or strongly absorbing OC. Without independent structural information, the authors should not use the thermal-optical result to identify “almost mature BC.”
Specific comments
- In Figure 3, is dg the mode of the fitted distribution or the calculated number geometric mean? Please clarify and report the geometric standard deviation.
- The full experimental matrix and number of replicates should be provided. Some figures include different vexit values, and Table 1 contains only a subset of the conditions.
- Statements that σdis and α300–500 do not change significantly with flame conditions require a statistical test. The individual spectra or values should be provided, at least in the supplement, rather than averaging over hydrogen fraction before demonstrating the absence of an effect.
- Line 306 appears to use 8 m s−1 rather than 8 cm s−1.
Citation: https://doi.org/10.5194/ar-2026-21-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 111 | 47 | 15 | 173 | 14 | 18 |
- HTML: 111
- PDF: 47
- XML: 15
- Total: 173
- BibTeX: 14
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of ar-2026-21
General Comments
The manuscript presents results from the investigation of the influence of hydrogen addition on the optical properties of generated combustion articles. The study investigates the effect of flame temperature and fuels mixtures of fossil fuel (ethylene) and hydrogen (H2) on the light absorbing properties of the generated combustion particles. Both Black and Brown Carbon are considered in the experimental approach.
The goal of the study is to examine potential impacts of combustion aerosol on the climate if fuel use switches from purely fossil to mixtures of fossil fuel and hydrogen. Since this mitigation strategy is one potential short-term option for reducing anthropogenic impacts on climate, the study makes a substantial contribution to an important research area.
Overall, the manuscript is well structured and fits very well into the scope of Aerosol Research. However, before being acceptable for publication it requires modifications of the presentation in general and of the explanation and discussion of results in particular. Suggested revisions are discussed in the following paragraphs. A separate language check is also recommended.
SPECIFIC COMMENTS
1| The terminology used here should be adjusted to the largely accepted terminology as described by Petzold et al. (2013). This can be introduced, e.g., in line 43 where the authors state that “BC is sometimes used as synonymous to EC”. Since this situation has been overcome by the introduction of the agreed terminology, this statement should be adjusted accordingly.
2| The Discussion section contains a part which might belong to the results, particularly the paragraph describing Fig. 8. Please consider rearranging the presentation.
3| The Conclusion section is very descriptive and summarises the findings. What is missing is a real conclusion by connecting the various findings into one “picture” of what has been gained in the study. In the current version, it is difficult to get the real take-home massages. Re-working the Conclusions section is strongly recommended.
4| The effect of the uncertainty of the Aethalometer C-factor is discussed e.g. on line 140 and following, and on line 352 and following but there is no reference to the review of Aethalometer inversion algorithms by Collaud Coen et al. (2010). It is highly recommended to analyse the potential effect of this landmark study on the presented results.
5| Equation 1 identifies the equivalence ratio by upper case Greek symbol Φ while in the text the same property is referred to by the lower-case Greek symbol φ. Please check for consistency. Furthermore, Equation 1 is not really presented as an equation. The terms should be explained, and a blank space should be added between “combustion” and “products”.
MINOR ISSUES:
1| In most of the figures the axis titles and colour bar descriptions contain the property and the unit in brackets. E.g., in Fig. 2 the y-axis title is “Tflame in [K]” and the description of the colour code of the lines is “vexit in [cm s-1]. Since this way of describing the axes is confusing one of the two option is suggested: (1) Tflame [K], or (2) ) Tflame in K. Personally, I would prefer type (1).
2| The last sentence of the abstract is too general. BrC can be a precursor of BC but is not always. Instead, it is another form of carbonaceous aerosol in the atmosphere. What is probably meant here is that BrC is a precursor to EC during combustion particle formation in high-temperature combustion processes. This is different to BrC formation, e.g., in smouldering processes during biomass burning events where lots of BrC is formed but only very minor BC because of the low flame temperatures and the different air to fuel mixture. I suggest a clarification.
3| In Figure 3, the inserted size distribution plot uses Dd for “diameter” while in all other figures and equations “dp” is used. This should be harmonized.
4| In Figure 5, the panel labels (a) and (b) should be more visible and may be moved to the top right corner of each panel.
5| In Figure 6, The symbols of the Mie- and Rayleigh-calculations should be enhanced. Actually, they are easily overlooked.
6| On line 420, it might help the reader to mention that higher values of vexit indicate higher flame temperatures which is the real process affecting particle maturation. On line 402, the authors say that Figure 5 indicates no significant light absorption at wavelengths short than 500 nm. However, shouldn’t it mean instead “longer than” 500 nm”?
TYPOS
Line11: Please correct to “flame”.
Line 67: It is suggested to write “In a combustion-generated aerosol, coated particles …” Now, there is a repetition of the term “particles”.
Line 82: It should read: “… as a function of the hydrogen content …”.
In Figure 1, the first sentence of the figure caption should read “Schematic representation of the experimental setup”.
Line 163: Please correct to “…, whereas the sampling times range from …”.
Line 173: Please correct to “… was negligible when operating …”.
Line 180: Please check the style of Equation 4.
Line 310: To increase readability, it is suggested to add “Values of” before a300-500 .
Line 376: The sentence starting with “Only under the conditions …” should be rephrased. A suggested modification is “Only under the conditions of γ = 0.5 and vexit = 6 cm s-1, σEC values were lower than …”.
Line 425: Please correct “The mass absorption cross section values of EC …”.
Line 435: The sentence is incomplete; it should be either “a multi-wavelength in-situ method …” or “multi-wavelength in-situ methods …”.
Line 477: It should read “thermo-optical”.
REFERENCES
Collaud Coen, M., Weingartner, E., Apituley, A., Ceburnis, D., Fierz-Schmidhauser, R., Flentje, H., Henzing, J. S., Jennings, S. G., Moerman, M., Petzold, A., Schmid, O., and Baltensperger, U.: Minimizing light absorption measurement artifacts of the Aethalometer: evaluation of five correction algorithms, Atmospheric Measurement Techniques, 3, 457-474, https://doi.org/10.5194/amt-3-457-2010, 2010.
Petzold, A., Ogren, J. A., Fiebig, M., Laj, P., Li, S.-M., Baltensperger, U., Holzer-Popp, T., Kinne, S., Pappalardo, G., Sugimoto, N., Wehrli, C., Wiedensohler, A., and Zhang, X.-Y.: Recommendations for reporting “black carbon” measurements, Atmospheric Chemistry and Physics, 13, 8365–8379, https://doi.org/10.5194/acp-13-8365-2013, 2013.