the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Dec 2023
Estimating errors in vehicle secondary aerosol production factors due to oxidation flow reactor response time
Abstract. Oxidation flow reactors used in secondary aerosol research do not immediately respond to changes in the inlet concentration of precursor gases because of their broad transfer functions. This is an issue when measuring the vehicular secondary aerosol formation in transient driving cycles because the secondary aerosol measured at the oxidation flow reactor outlet does not correspond to the rapid changes in the exhaust flow rate. Since the secondary aerosol production factor is determined by multiplying the secondary aerosol mass with the exhaust flow rate, the misalignment between the two leads to incorrect production factors. This study evaluates the extent of the error in production factors due to oxidation flow reactor transfer functions by using synthetic and semi-synthetic exhaust emission data. It was found that the transfer function-related error could be eliminated when only the total production factor of full cycle was measured using constant volume sampling. For shorter segments within a driving cycle, a narrower transfer function led to smaller error. Even with a narrow transfer function, the oxidation flow reactor could report production factors that were more than 10 times higher than the true production factors if the segment duration was too short.
-
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
(2273 KB)
-
Supplement
(5991 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2273 KB) - Metadata XML
-
Supplement
(5991 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-2692', Anonymous Referee #1, 16 Feb 2024
The manuscript by Siomnen et al. discusses the influence of residence time distribution (RTD) of oxidation flow reactor (OFR) on SOA formation from car exhaust from some testing driving modes. They derived the equation for representing temporal variation in SOA concentration following OFR. Further analysis of experimental and synthetic data suggested that numerical deconvolution for the influence of VOC emission and RTD of OFR for observed temporal variation in SOA mass is needed, as the time scale for change in driving conditions is shorter than a residence time of typical OFRs. The topic is within the scope of the interest of readers of the journal. Some implicit assumptions in the manuscript will need to be clarified. Organization of the manuscript can be improved. In my opinion, the current quality of the manuscript does not meet with the criteria of the journal, even though the concept of the manuscript itself is interesting.
Â
Major comments
Influence of VOC oxidation kinetics.
Not all the VOCs would be consumed in the reactor if their oxidation time scale is longer than the mean residence time of the OFR. I wonder if exponential decay in VOC concentration ([VOC] = [VOC]0 * exp (-t/tau)) in the OFR should also be considered for deconvoluting the data. If the oxidation time scale in OFR is sufficiently short due to the high concentration of oxidants, it will need to be quantitatively discussed.
Â
Equation (5)
The equation (SOA = Y·HC) implies that all hydrocarbons in the OFR are oxidized. My understanding is that HC indicates the total amount of injected hydrocarbons, rather than reacted amount of hydrocarbons. Could the authors provide a justification for this assumption?
The equation also assumes that SOA yield does not change throughout the experiment, even though both gas phase chemical composition and SOA mass concentration in the OFR keeps changing. I personally think that Y should also be a function of t. Could the authors provide future detailed information/discussion about it?
Â
Discuss adsorption/absorption of VOCs/OVOCs on the wall
In most cases, the concept of RTD is employed by assuming that a flow pattern of fluid is the dominant regulator for determining the time scale for reactants to stay in a chemical reactor. However, in the case of VOCs/OVOCs for SOA precursors, absorption/desorption processes on walls of reactors are typically non-negligible for determining their actual residence time. This process slows down the response of a chemical reactor to the changes in operating conditions. It will be helpful if the authors could provide how the process influences responses of the OFR.
Â
Method section.
It is better to put the method section prior to the result section, as in the case of most of other publications in the journal. The manuscript indicates that some data sets were obtained using the constant volume sampler (CSV). The present description about the CSV is not sufficient to understand how the CSV works/why it is needed/what are the advantages and disadvantages to use it.
Â
Section S3. Synthetic driving cycle
Consider moving this section to the main text, as it is critical information for understanding the contents.
Â
Minor comments
Line 23
What does m-% mean?
Â
Figure 1
In this figure (and at many other parts in the manuscript), parameter t is employed for two meanings. One is the time after the start of a driving test, and another is time for residence time in the OFR. However, these two types of t do not correspond to each other except for the case of pulse injection. I suggest the authors to consider using different parameters for clarification.
Â
Line 117
I could not understand how the delay caused by the OFR can be calculated.
Â
Line 129
The concepts of [SOA]OFR and [HC]OFR are clear to me. However, I am wondering how these two metrics can simultaneously be measured during practical applications.
Â
Line 172
What does ‘transfer function standard deviation’ mean? Does it indicate width of the RTD? If so I wonder how the standard deviation was derived, as functional forms of RTD are not normal functions in many cases.
Â
L 188
I do not think that using the expression of ‘true SOA’ is very appropriate. The value was estimated from measured hydrocarbon concentration by assuming that SOA yield is always a constant. However, the validity of this assumption is unclear. It is not a good idea to use such an expression unless there is convincing evidence about how the true SOA mass should be.
Citation: https://doi.org/10.5194/egusphere-2023-2692-RC1 - AC1: 'Reply on RC2', Pauli Simonen, 29 Mar 2024
-
RC2: 'Comment on egusphere-2023-2692', Anonymous Referee #2, 23 Feb 2024
The paper by Pauli Simonen et al. presents an insightful exploration into the complexities of measuring vehicle secondary organic aerosol (SOA) production factors (PF) using oxidation flow reactors (OFRs). The authors' approach, particularly their consideration of OFR response time and its convolution with vehicle emissions during transient driving cycles, provides valuable insights into determining accurate PFs. By utilizing synthetic and semi-synthetic exhaust emission data, this study not only evaluates potential errors but also suggests methods for their mitigation, emphasizing the importance of constant volume sampling (CVS) for precise measurements. The research importantly points out the need for a thorough understanding of OFR transfer functions and response times in aerosol research. Before recommending acceptance of the paper, I have the following suggestions for the authors:
1. The paper defines the concept of SOA yield (Y) as the ratio of produced SOA to consumed hydrocarbon (HC). However, in the derivation process, "total" HC is used instead of "consumed" HC to calculate the SOA produced in OFR. Equation 5 implies a default assumption that all HC is completely oxidized in the OFR, which may not always be accurate in real scenarios.
2. In the absence of any sink within the OFR (such as wall loss or chemical reactions), the cumulative emissions measured before and after the OFR should match over time, since the CVS merely dilutes emissions without exhausting them. If this is not the case, the authors should present a clear mass balance scheme to explain the differences observed before and after the OFR.
3. Equation 14 in the paper derives the total cumulative SOA, but the methodology for deriving SOA(t) is not clearly explained. Clarification on this derivation would enhance the readers' understanding.
4. There is ambiguity between [HC]'OFR and [HC]OFR as mentioned in Lines 128-129.
5. The structure of the paper could be improved for better flow and coherence. Specifically, Section 5, which discusses Methods, should be relocated to an earlier part of the manuscript to enhance the logical progression of the paper.
Citation: https://doi.org/10.5194/egusphere-2023-2692-RC2 - AC1: 'Reply on RC2', Pauli Simonen, 29 Mar 2024
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2023-2692', Anonymous Referee #1, 16 Feb 2024
The manuscript by Siomnen et al. discusses the influence of residence time distribution (RTD) of oxidation flow reactor (OFR) on SOA formation from car exhaust from some testing driving modes. They derived the equation for representing temporal variation in SOA concentration following OFR. Further analysis of experimental and synthetic data suggested that numerical deconvolution for the influence of VOC emission and RTD of OFR for observed temporal variation in SOA mass is needed, as the time scale for change in driving conditions is shorter than a residence time of typical OFRs. The topic is within the scope of the interest of readers of the journal. Some implicit assumptions in the manuscript will need to be clarified. Organization of the manuscript can be improved. In my opinion, the current quality of the manuscript does not meet with the criteria of the journal, even though the concept of the manuscript itself is interesting.
Â
Major comments
Influence of VOC oxidation kinetics.
Not all the VOCs would be consumed in the reactor if their oxidation time scale is longer than the mean residence time of the OFR. I wonder if exponential decay in VOC concentration ([VOC] = [VOC]0 * exp (-t/tau)) in the OFR should also be considered for deconvoluting the data. If the oxidation time scale in OFR is sufficiently short due to the high concentration of oxidants, it will need to be quantitatively discussed.
Â
Equation (5)
The equation (SOA = Y·HC) implies that all hydrocarbons in the OFR are oxidized. My understanding is that HC indicates the total amount of injected hydrocarbons, rather than reacted amount of hydrocarbons. Could the authors provide a justification for this assumption?
The equation also assumes that SOA yield does not change throughout the experiment, even though both gas phase chemical composition and SOA mass concentration in the OFR keeps changing. I personally think that Y should also be a function of t. Could the authors provide future detailed information/discussion about it?
Â
Discuss adsorption/absorption of VOCs/OVOCs on the wall
In most cases, the concept of RTD is employed by assuming that a flow pattern of fluid is the dominant regulator for determining the time scale for reactants to stay in a chemical reactor. However, in the case of VOCs/OVOCs for SOA precursors, absorption/desorption processes on walls of reactors are typically non-negligible for determining their actual residence time. This process slows down the response of a chemical reactor to the changes in operating conditions. It will be helpful if the authors could provide how the process influences responses of the OFR.
Â
Method section.
It is better to put the method section prior to the result section, as in the case of most of other publications in the journal. The manuscript indicates that some data sets were obtained using the constant volume sampler (CSV). The present description about the CSV is not sufficient to understand how the CSV works/why it is needed/what are the advantages and disadvantages to use it.
Â
Section S3. Synthetic driving cycle
Consider moving this section to the main text, as it is critical information for understanding the contents.
Â
Minor comments
Line 23
What does m-% mean?
Â
Figure 1
In this figure (and at many other parts in the manuscript), parameter t is employed for two meanings. One is the time after the start of a driving test, and another is time for residence time in the OFR. However, these two types of t do not correspond to each other except for the case of pulse injection. I suggest the authors to consider using different parameters for clarification.
Â
Line 117
I could not understand how the delay caused by the OFR can be calculated.
Â
Line 129
The concepts of [SOA]OFR and [HC]OFR are clear to me. However, I am wondering how these two metrics can simultaneously be measured during practical applications.
Â
Line 172
What does ‘transfer function standard deviation’ mean? Does it indicate width of the RTD? If so I wonder how the standard deviation was derived, as functional forms of RTD are not normal functions in many cases.
Â
L 188
I do not think that using the expression of ‘true SOA’ is very appropriate. The value was estimated from measured hydrocarbon concentration by assuming that SOA yield is always a constant. However, the validity of this assumption is unclear. It is not a good idea to use such an expression unless there is convincing evidence about how the true SOA mass should be.
Citation: https://doi.org/10.5194/egusphere-2023-2692-RC1 - AC1: 'Reply on RC2', Pauli Simonen, 29 Mar 2024
-
RC2: 'Comment on egusphere-2023-2692', Anonymous Referee #2, 23 Feb 2024
The paper by Pauli Simonen et al. presents an insightful exploration into the complexities of measuring vehicle secondary organic aerosol (SOA) production factors (PF) using oxidation flow reactors (OFRs). The authors' approach, particularly their consideration of OFR response time and its convolution with vehicle emissions during transient driving cycles, provides valuable insights into determining accurate PFs. By utilizing synthetic and semi-synthetic exhaust emission data, this study not only evaluates potential errors but also suggests methods for their mitigation, emphasizing the importance of constant volume sampling (CVS) for precise measurements. The research importantly points out the need for a thorough understanding of OFR transfer functions and response times in aerosol research. Before recommending acceptance of the paper, I have the following suggestions for the authors:
1. The paper defines the concept of SOA yield (Y) as the ratio of produced SOA to consumed hydrocarbon (HC). However, in the derivation process, "total" HC is used instead of "consumed" HC to calculate the SOA produced in OFR. Equation 5 implies a default assumption that all HC is completely oxidized in the OFR, which may not always be accurate in real scenarios.
2. In the absence of any sink within the OFR (such as wall loss or chemical reactions), the cumulative emissions measured before and after the OFR should match over time, since the CVS merely dilutes emissions without exhausting them. If this is not the case, the authors should present a clear mass balance scheme to explain the differences observed before and after the OFR.
3. Equation 14 in the paper derives the total cumulative SOA, but the methodology for deriving SOA(t) is not clearly explained. Clarification on this derivation would enhance the readers' understanding.
4. There is ambiguity between [HC]'OFR and [HC]OFR as mentioned in Lines 128-129.
5. The structure of the paper could be improved for better flow and coherence. Specifically, Section 5, which discusses Methods, should be relocated to an earlier part of the manuscript to enhance the logical progression of the paper.
Citation: https://doi.org/10.5194/egusphere-2023-2692-RC2 - AC1: 'Reply on RC2', Pauli Simonen, 29 Mar 2024
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 268 | 89 | 22 | 379 | 34 | 18 | 19 |
- HTML: 268
- PDF: 89
- XML: 22
- Total: 379
- Supplement: 34
- BibTeX: 18
- EndNote: 19
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Miikka Dal Maso
Pinja Prauda
Anniina Hoilijoki
Anette Karppinen
Pekka Matilainen
Panu Karjalainen
Jorma Keskinen
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(2273 KB) - Metadata XML
-
Supplement
(5991 KB) - BibTeX
- EndNote
- Final revised paper