the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Dec 2024
Influence of ambient NO and NO2 on the quantification of total peroxy nitrates (∑PNs) and total alkyl nitrates (∑ANs) by thermal dissociation cavity ring-down spectroscopy (TD-CRDS)
Abstract. Measurement of total peroxy nitrates (∑PNs) and alkyl nitrates (∑ANs) by instruments that use thermal dissociation (TD) inlets to convert the organic nitrate to detectable NO2 may suffer from systematic bias (both positive and negative) resulting from unwanted secondary chemistry in the heated inlets. Here we review the sources of the bias and the methods used to reduce it and/or correct for it and report new experiments using (for the first time) atmospherically relevant, unsaturated, biogenic alkyl nitrates as well as two different peroxyacetyl nitrate (PAN) sources. We show that the commonly used commercial C3-alkyl-nitrate (isopropyl nitrate, IPN) inlet for characterising the chemistry of ANs is not appropriate for real-air samples that contain longer chain nitrates. ANs generated in the NO3-induced oxidation of limonene are strongly positively biased in the presence of NO. By detecting NOX rather than NO2, we provide a simple solution to avoid the bias caused by the conversion of NO to NO2 by primary and secondary peroxy radicals resulting from the complex chemistry in the thermal degradation of long-chain, alkyl nitrates in air at TD-temperatures. We also show that using a photochemical source of PAN to characterise the TD-inlets can result in a much stronger apparent bias from NO to NO2 conversion than for a diffusion source of synthesised (“pure”) PAN at similar mixing ratios. This is explained by the presence of thermally labile trace gases such as peracetic acid (CH3C(O)OOH) and hydrogen peroxide (H2O2).
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(789 KB)
-
Supplement
(329 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(789 KB) - Metadata XML
-
Supplement
(329 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-3694', Hans Osthoff, 08 Jan 2025
Wüst et al. provide an update on the quantification of total peroxy (ΣPN) and total alkyl nitrate (ΣAN) by thermal-dissociation cavity ring-down spectroscopy (TD-CRDS). New data are presented to show: (1) It is advantageous to oxidize NO to NO2 by adding (excess) O3 to the inlet after the thermal dissociation region. (2) Sampling synthetic PAN from a diffusion source will result in less inlet bias than sampling photochemically generated PAN (in the ΣAN channel). (3) Limonene derived nitrates suffer from different inlet biases than the commercially available isopropyl nitrate. (4) In high concentration, acetone generates an artifact in the ΣAN channel.
Regarding point (1): The authors do a good job convincing of the need to add O3 after the TD region (and measure everything as NOx) but this is not new. The first group to add O3 after the TD region were, as far as I am aware, Wild et al. (Environm. Sci. Technol., 48(16), 9609-9615, doi:10.1021/es501896w, 2014). Our group has implemented the addition of O3 since 2019 (see thesis by N. Gingerysty, doi: 10.11575/PRISM/38580 (2021)). Still, since there are groups quantifying ΣAN who are not adding O3 (e.g., Lin et al. Talanta 270, 125524, 2024) it is worthwhile publishing this.
Regarding point (2): it is well known in the community that photochemical PAN sources co-emit organic by-products (especially formaldehyde) and are “pure” only relative to other nitrogen oxides. The data presented in this manuscript represent a near-worst-case scenario for a photochemical source as a very large acetone concentration and a Hg pen ray lamp were used. I agree with the author's conclusion that PAN from a diffusion source should be used to characterize ΣAN inlet chemistry instead. For the ΣPN channel, photochemical PAN sources are fine, though, as long as acetone concentrations are kept low.
(3) The data on limonene derived nitrates are novel and interesting. I was wondering if the difference to isopropyl nitrate is partially caused by a matrix effect (see specific comments).
I would have liked to see an experiment in which PAN (from the diffusion source) is combined with the chamber/limonene output to verify that 2+2 indeed equals 4 (and not 3.5 or 2.5) in both the ΣPN and ΣAN channels.
(4) As an aside: We have observed a similar and more pronounced effect with acetone in our TD-CRDS NOy inlet (operated at ~850 K). Photochemical PAN sources are definitively no good for that application either!
The paper is written well and very thoughtful. I recommend publication of this article once the authors have addressed my comments (above and below).
Specific comments
Line 14. Correct grammar (“show that the commonly used commercial C3-alkylnitrate (isopropyl nitrate, IPN) inlet for characterising”)
Line 18. Please rephrase “We also show that using a photochemical source of PAN to characterise the TD-inlets can result in a much stronger apparent bias from NO to NO2 conversion …”. As commented on lines 150 – 151, the photochemical source used in this work could have been operated more optimally such that the statement in the abstract comes across as too strong.
Lines 90 - 91. "partial detection of ANs in the PNs channel" It is suggested that glass beads catalyze (i.e., lower the temperature at which) ANs convert to NO2. Please comment on the potential effect of temperature gradients within the inlet (which can also lead to some overlap of the dissociation profiles and hence bias in TD-CRDS measurements of PNs and ANs).
Line 132. “effective absorption cross section”. Consider stating its value (and how it was determined).
Line 150. “4.6 % acetone”. Flocke et al. (2005) ”used a mixture of 10 ppmv acetone and 10 ppmv of CO … to keep the OH mixing ratio in the reaction vessel sufficiently low to almost completely suppress the formation of HNO3, but … found that using 20 ppmv acetone instead was acceptable”. If 4.6% acetone was used, it would not be surprising that this PAN photosource would contain a lot of impurities and result in bias.
Line 151. Phosphor-coated pen-ray lamps can get quite hot (which destroys PAN) and emit much radiation at 254 nm also that may drive unwanted photochemistry (see Furgeson et al. Atmos. Environm.45, 5025,doi: 10.1016/j.atmosenv.2011.03.072, 2011; Rider et al. Atmos. Meas. Tech. 8(7), 2737-2748, doi 10.5194/amt-8-2737-2015, 2015).
Line 175. Correct spelling (tridecane). A safety note regarding the potential explosive nature of PAN should be added also.
Line 192. Please change (CAN in 6 M HNO3) to (CAN) in 6 M HNO3 to improve clarity.
Line 200. “ANs mixing ratios between 0.47 and 4.4 ppbv”. How was this determined?
Line 201. “The nitrates generated in this manner from isoprene are a mixture of C5-nitrooxyhydroperoxides, C5-nitrooxycarbonyl, C5hydroxynitrate and also C10-nitrooxyperoxide (ROOR), with the relative concentrations depending on the fate of the initially formed nitrooxyperoxy radicals (IUPAC 2024).” I didn’t find IUPAC in the reference list (only in the SI) - there may be a more suitable reference for this statement anyways.
Lines 239-310 “At 648 K the chemistry is different”. What follows is interesting descriptions of possible reaction pathways. What is missing, though, are numerical simulations to constrain the impact of each of these side reactions. For example, I would be curious how much of the singlet α-lactone forms via (R29) and what the impact of this pathway could be (i.e., by how much does the product distribution shift when this reaction is “turned off” in the mechanism?). In other words, this section could be more quantitative (and thus appear less speculative).
Line 310. “In separate experiments, we observed that the thermal dissociation of acetone also accounts for a small fraction of the overestimation NO2 formed (400 pptv NO2 at 43 ppmv acetone and 16 ppbv NO), which is confirmed by its thermogram (Fig. S2).“ Carbon, hydrogen and oxygen atoms do not convert to nitrogen (and form nitrogen dioxide), so an explanation is needed what causes this effect. My hunch is that heating acetone to these temperatures generates α,β-dicarbonyls (methyl glyoxal or 2,3-butadione) which would absorb at 409 nm.
Line 322. “In summary, the photochemical PAN source is not well suited for characterising the inlet chemistry for alkyl nitrate detection via thermal decomposition to NO2 as it results in significant NO to NO2 conversion owing to the presence of thermally labile trace gases such as peracetic acid and hydrogen peroxide”. What if a lower acetone concentration had been used?
Line 330. “The IR spectrum (Fig. S4) of a gas sample eluted from the PAN diffusion source is in good agreement with the literature and does not reveal the presence of high levels of impurities.” Please add more information (to the experimental section or the SI) how this spectrum was obtained.
Line 375. “0.47 or 4.4 ppbv isoprene nitrate” How were these mixing ratios determined?
Line 417-420. “when using biogenic nitrates derived from terpenoids.” The authors sampled biogenic nitrates in a complex mixture containing all kinds of compounds. The conclusions drawn may thus be more due to a matrix effect which would be absent when sampling IPN. In other words, would these results stand if a “pure” sample of limonene nitrate (isolated or synthesized) had been analyzed?
Figure S1. Please state in the caption for what temperature the simulations were run.
Citation: https://doi.org/10.5194/egusphere-2024-3694-RC1 - AC1: 'Reply on RC1 and RC2', Laura Wüst, 14 Feb 2025
-
RC2: 'Comment on egusphere-2024-3694', Anonymous Referee #2, 14 Jan 2025
General comments:
This paper describes a thorough set of experiments to quantify the effect of additional (i.e., other than that produced by thermal dissociation) NO and NO2 on the performance of a TD-CRDS instrument for the measurements of peroxy- and alkyl nitrates. It also provides a helpful overview of other studies that have characterized this secondary inlet chemistry, which serves as a nice review of the technique. This will be of great interest to the community using such thermal dissociation techniques for organic nitrate measurements. The paper is generally very well-written and clear. I recommend publication after some minor revisions to clarify especially the figures.
Specific comments:
Several times throughout the paper you use the term “eluted” for transfer of a liquid into the gas phase. I think “evaporated” would be more accurate; “eluted” typically refers to a separation of liquids.
Around lines 287-288: This is a very large range of thermal dissociation lifetimes! Maybe comment here on how well known these are, and how this means the organic composition in ambient samples will heavily influence the secondary chemistry. This can help anticipate the uncertainties in the simulations you show later.
Around line 302: You start using PAA later in this paragraph, so nice to introduce that acronym in the first line.
In the paragraph on lines 313-321, you describe the model simulations shown in Fig. S3. Since you do some experiments adding specific peroxides, can you comment on how well these simulated patterns match with your observations? Also, can you comments on how you arrived at the (very high) concentrations used in these simulations?
On line 337, you refer to “impurities” when I think you mean “additional organic species that can oxidize NO to NO2”
On line 359 “cause by the oxidation of NO2” – here, don’t you mean the recombination of NO2 with organics?
On line 373 you refer to “more relevant (unsaturated) alkyl nitrates”. This makes me wonder if the fact that first-generation terpenes alkyl nitrates are unsaturated is the key here? Below around line 390 you discuss the degradation of longer chain alky nitrates producing more organic peroxy radicals. Do you think the double bond plays a role here, or is it just carbon number?
On line 405 you say this means unreliable data unless NO levels are very low. Can you add a comment based on your ambient measurements how often / where this is case? I wonder if this is mainly a problem for urban / near-source datasets?
Figure 2: is a bit confusing because the lower panel is without NO in the PAN source, but you see the NO axis based on the added NO in the cavities. Maybe modify the last sentence to clarify that you in this case add only the organics and not the NO from the PAN source? The y-axis error bars are only visible on the panel b 648 K trace. Are they present on all of them and just so small to be invisible? If so, state this. But I don’t know why this would be so different e.g. form panel a to b 648 K traces. Finally, you refer to the errors including 1sigma NO2 mixing ratio std deviations of both 648 K and room temperature NO2 mixing ratios. Does this means you subtract the room temperature measurements in each case? That didn’t seem to be the case to me, since you’re not reporting sumANs but rather NO2 in the 648 K channel (sounds like it would be an unsubtracted quantity). And, do you do this subtraction only for the 648 k trace and not the 448K trace, since you don’t mention the latter? This caption raises a lot of questions that I think some revision could avoid.
In the later figures you also refer to the uncertainty in room temperature NO2 measurements. Maybe add to each caption what you subtract to get each trace, to clarify.
For parallel structure, could the x axis of Figure 6 be listed as “NO2 added (ppbv)”?
Technical corrections:
Line 166: “In section 4.1, we discuss the concentrations”
Line 175: “tridecane”
Line 218: “so that the cavities detected only NO2 (and not NOx).”
Line 311: “fraction of the excess of NO2 formed”
Line 317: “S3a), 88 ppbv PANs”
Line 320: “Fig. S3b), higher mixing”
Line 365: “we have shown that IPN”
Line 416: “we have quantified the potential bias”
Line 420: “as has been frequently assumed.”
Line 428 (a bit long sentence, maybe reword? Or at least): “to convert NO to NO2, corrections for”
Citation: https://doi.org/10.5194/egusphere-2024-3694-RC2 - AC1: 'Reply on RC1 and RC2', Laura Wüst, 14 Feb 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-3694', Hans Osthoff, 08 Jan 2025
Wüst et al. provide an update on the quantification of total peroxy (ΣPN) and total alkyl nitrate (ΣAN) by thermal-dissociation cavity ring-down spectroscopy (TD-CRDS). New data are presented to show: (1) It is advantageous to oxidize NO to NO2 by adding (excess) O3 to the inlet after the thermal dissociation region. (2) Sampling synthetic PAN from a diffusion source will result in less inlet bias than sampling photochemically generated PAN (in the ΣAN channel). (3) Limonene derived nitrates suffer from different inlet biases than the commercially available isopropyl nitrate. (4) In high concentration, acetone generates an artifact in the ΣAN channel.
Regarding point (1): The authors do a good job convincing of the need to add O3 after the TD region (and measure everything as NOx) but this is not new. The first group to add O3 after the TD region were, as far as I am aware, Wild et al. (Environm. Sci. Technol., 48(16), 9609-9615, doi:10.1021/es501896w, 2014). Our group has implemented the addition of O3 since 2019 (see thesis by N. Gingerysty, doi: 10.11575/PRISM/38580 (2021)). Still, since there are groups quantifying ΣAN who are not adding O3 (e.g., Lin et al. Talanta 270, 125524, 2024) it is worthwhile publishing this.
Regarding point (2): it is well known in the community that photochemical PAN sources co-emit organic by-products (especially formaldehyde) and are “pure” only relative to other nitrogen oxides. The data presented in this manuscript represent a near-worst-case scenario for a photochemical source as a very large acetone concentration and a Hg pen ray lamp were used. I agree with the author's conclusion that PAN from a diffusion source should be used to characterize ΣAN inlet chemistry instead. For the ΣPN channel, photochemical PAN sources are fine, though, as long as acetone concentrations are kept low.
(3) The data on limonene derived nitrates are novel and interesting. I was wondering if the difference to isopropyl nitrate is partially caused by a matrix effect (see specific comments).
I would have liked to see an experiment in which PAN (from the diffusion source) is combined with the chamber/limonene output to verify that 2+2 indeed equals 4 (and not 3.5 or 2.5) in both the ΣPN and ΣAN channels.
(4) As an aside: We have observed a similar and more pronounced effect with acetone in our TD-CRDS NOy inlet (operated at ~850 K). Photochemical PAN sources are definitively no good for that application either!
The paper is written well and very thoughtful. I recommend publication of this article once the authors have addressed my comments (above and below).
Specific comments
Line 14. Correct grammar (“show that the commonly used commercial C3-alkylnitrate (isopropyl nitrate, IPN) inlet for characterising”)
Line 18. Please rephrase “We also show that using a photochemical source of PAN to characterise the TD-inlets can result in a much stronger apparent bias from NO to NO2 conversion …”. As commented on lines 150 – 151, the photochemical source used in this work could have been operated more optimally such that the statement in the abstract comes across as too strong.
Lines 90 - 91. "partial detection of ANs in the PNs channel" It is suggested that glass beads catalyze (i.e., lower the temperature at which) ANs convert to NO2. Please comment on the potential effect of temperature gradients within the inlet (which can also lead to some overlap of the dissociation profiles and hence bias in TD-CRDS measurements of PNs and ANs).
Line 132. “effective absorption cross section”. Consider stating its value (and how it was determined).
Line 150. “4.6 % acetone”. Flocke et al. (2005) ”used a mixture of 10 ppmv acetone and 10 ppmv of CO … to keep the OH mixing ratio in the reaction vessel sufficiently low to almost completely suppress the formation of HNO3, but … found that using 20 ppmv acetone instead was acceptable”. If 4.6% acetone was used, it would not be surprising that this PAN photosource would contain a lot of impurities and result in bias.
Line 151. Phosphor-coated pen-ray lamps can get quite hot (which destroys PAN) and emit much radiation at 254 nm also that may drive unwanted photochemistry (see Furgeson et al. Atmos. Environm.45, 5025,doi: 10.1016/j.atmosenv.2011.03.072, 2011; Rider et al. Atmos. Meas. Tech. 8(7), 2737-2748, doi 10.5194/amt-8-2737-2015, 2015).
Line 175. Correct spelling (tridecane). A safety note regarding the potential explosive nature of PAN should be added also.
Line 192. Please change (CAN in 6 M HNO3) to (CAN) in 6 M HNO3 to improve clarity.
Line 200. “ANs mixing ratios between 0.47 and 4.4 ppbv”. How was this determined?
Line 201. “The nitrates generated in this manner from isoprene are a mixture of C5-nitrooxyhydroperoxides, C5-nitrooxycarbonyl, C5hydroxynitrate and also C10-nitrooxyperoxide (ROOR), with the relative concentrations depending on the fate of the initially formed nitrooxyperoxy radicals (IUPAC 2024).” I didn’t find IUPAC in the reference list (only in the SI) - there may be a more suitable reference for this statement anyways.
Lines 239-310 “At 648 K the chemistry is different”. What follows is interesting descriptions of possible reaction pathways. What is missing, though, are numerical simulations to constrain the impact of each of these side reactions. For example, I would be curious how much of the singlet α-lactone forms via (R29) and what the impact of this pathway could be (i.e., by how much does the product distribution shift when this reaction is “turned off” in the mechanism?). In other words, this section could be more quantitative (and thus appear less speculative).
Line 310. “In separate experiments, we observed that the thermal dissociation of acetone also accounts for a small fraction of the overestimation NO2 formed (400 pptv NO2 at 43 ppmv acetone and 16 ppbv NO), which is confirmed by its thermogram (Fig. S2).“ Carbon, hydrogen and oxygen atoms do not convert to nitrogen (and form nitrogen dioxide), so an explanation is needed what causes this effect. My hunch is that heating acetone to these temperatures generates α,β-dicarbonyls (methyl glyoxal or 2,3-butadione) which would absorb at 409 nm.
Line 322. “In summary, the photochemical PAN source is not well suited for characterising the inlet chemistry for alkyl nitrate detection via thermal decomposition to NO2 as it results in significant NO to NO2 conversion owing to the presence of thermally labile trace gases such as peracetic acid and hydrogen peroxide”. What if a lower acetone concentration had been used?
Line 330. “The IR spectrum (Fig. S4) of a gas sample eluted from the PAN diffusion source is in good agreement with the literature and does not reveal the presence of high levels of impurities.” Please add more information (to the experimental section or the SI) how this spectrum was obtained.
Line 375. “0.47 or 4.4 ppbv isoprene nitrate” How were these mixing ratios determined?
Line 417-420. “when using biogenic nitrates derived from terpenoids.” The authors sampled biogenic nitrates in a complex mixture containing all kinds of compounds. The conclusions drawn may thus be more due to a matrix effect which would be absent when sampling IPN. In other words, would these results stand if a “pure” sample of limonene nitrate (isolated or synthesized) had been analyzed?
Figure S1. Please state in the caption for what temperature the simulations were run.
Citation: https://doi.org/10.5194/egusphere-2024-3694-RC1 - AC1: 'Reply on RC1 and RC2', Laura Wüst, 14 Feb 2025
-
RC2: 'Comment on egusphere-2024-3694', Anonymous Referee #2, 14 Jan 2025
General comments:
This paper describes a thorough set of experiments to quantify the effect of additional (i.e., other than that produced by thermal dissociation) NO and NO2 on the performance of a TD-CRDS instrument for the measurements of peroxy- and alkyl nitrates. It also provides a helpful overview of other studies that have characterized this secondary inlet chemistry, which serves as a nice review of the technique. This will be of great interest to the community using such thermal dissociation techniques for organic nitrate measurements. The paper is generally very well-written and clear. I recommend publication after some minor revisions to clarify especially the figures.
Specific comments:
Several times throughout the paper you use the term “eluted” for transfer of a liquid into the gas phase. I think “evaporated” would be more accurate; “eluted” typically refers to a separation of liquids.
Around lines 287-288: This is a very large range of thermal dissociation lifetimes! Maybe comment here on how well known these are, and how this means the organic composition in ambient samples will heavily influence the secondary chemistry. This can help anticipate the uncertainties in the simulations you show later.
Around line 302: You start using PAA later in this paragraph, so nice to introduce that acronym in the first line.
In the paragraph on lines 313-321, you describe the model simulations shown in Fig. S3. Since you do some experiments adding specific peroxides, can you comment on how well these simulated patterns match with your observations? Also, can you comments on how you arrived at the (very high) concentrations used in these simulations?
On line 337, you refer to “impurities” when I think you mean “additional organic species that can oxidize NO to NO2”
On line 359 “cause by the oxidation of NO2” – here, don’t you mean the recombination of NO2 with organics?
On line 373 you refer to “more relevant (unsaturated) alkyl nitrates”. This makes me wonder if the fact that first-generation terpenes alkyl nitrates are unsaturated is the key here? Below around line 390 you discuss the degradation of longer chain alky nitrates producing more organic peroxy radicals. Do you think the double bond plays a role here, or is it just carbon number?
On line 405 you say this means unreliable data unless NO levels are very low. Can you add a comment based on your ambient measurements how often / where this is case? I wonder if this is mainly a problem for urban / near-source datasets?
Figure 2: is a bit confusing because the lower panel is without NO in the PAN source, but you see the NO axis based on the added NO in the cavities. Maybe modify the last sentence to clarify that you in this case add only the organics and not the NO from the PAN source? The y-axis error bars are only visible on the panel b 648 K trace. Are they present on all of them and just so small to be invisible? If so, state this. But I don’t know why this would be so different e.g. form panel a to b 648 K traces. Finally, you refer to the errors including 1sigma NO2 mixing ratio std deviations of both 648 K and room temperature NO2 mixing ratios. Does this means you subtract the room temperature measurements in each case? That didn’t seem to be the case to me, since you’re not reporting sumANs but rather NO2 in the 648 K channel (sounds like it would be an unsubtracted quantity). And, do you do this subtraction only for the 648 k trace and not the 448K trace, since you don’t mention the latter? This caption raises a lot of questions that I think some revision could avoid.
In the later figures you also refer to the uncertainty in room temperature NO2 measurements. Maybe add to each caption what you subtract to get each trace, to clarify.
For parallel structure, could the x axis of Figure 6 be listed as “NO2 added (ppbv)”?
Technical corrections:
Line 166: “In section 4.1, we discuss the concentrations”
Line 175: “tridecane”
Line 218: “so that the cavities detected only NO2 (and not NOx).”
Line 311: “fraction of the excess of NO2 formed”
Line 317: “S3a), 88 ppbv PANs”
Line 320: “Fig. S3b), higher mixing”
Line 365: “we have shown that IPN”
Line 416: “we have quantified the potential bias”
Line 420: “as has been frequently assumed.”
Line 428 (a bit long sentence, maybe reword? Or at least): “to convert NO to NO2, corrections for”
Citation: https://doi.org/10.5194/egusphere-2024-3694-RC2 - AC1: 'Reply on RC1 and RC2', Laura Wüst, 14 Feb 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 216 | 65 | 11 | 292 | 76 | 12 | 11 |
- HTML: 216
- PDF: 65
- XML: 11
- Total: 292
- Supplement: 76
- BibTeX: 12
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Cited
Laura Wüst
Patrick Dewald
Gunther N. T. E. Türk
Jos Lelieveld
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(789 KB) - Metadata XML
-
Supplement
(329 KB) - BibTeX
- EndNote
- Final revised paper