CB6 mechanism regarding Semi-volatile POA

CMAQ Chemistry and Aerosols
1

Hi Everyone,

I have questions about the treatment of semi-volatile POA in the CMAQ_DESID file.

! Gas Species : VLVPO1 VSVPO1 VSVPO2 VSVPO3 VIVPO1
! Particle Species : ALVPO1 ASVPO1 ASVPO2 ASVPO3 AIVPO1
! C* (µg m⁻³) : 0.1 1 10 100 1000

‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VSVPO1’, ‘GAS’ ,0.045,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VSVPO2’, ‘GAS’ ,0.14 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VSVPO3’, ‘GAS’ ,0.18 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VIVPO1’, ‘GAS’ ,0.50 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘ALVPO1’, ‘FINE’,0.09 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘ASVPO1’, ‘FINE’,0.045,‘MASS’,‘a’,

  1. Gas and particle representation of SVPO1

    In the CMAQ_DESID file, both VSVPO1 and ASVPO1 are assigned a saturation concentration (C* ) of 1 µg m⁻³. However, the CMAQ User’s Guide (Appendix B.2.2.1) states that “Species with C* = 10 can generally have mass split between the gas and particle phases.”

    Question: Since SVPO1 has C = 1 µg m⁻³, why is it represented as both a gas species (VSVPO1) and a particle species (ASVPO1) in CMAQ DESID file? Does the statement in the User’s Guide simply describe a typical example, or is there another reason that species with C = 1 µg m⁻³* are also represented in both phases?

  2. Reversible gas-particle partitioning of POA emissions

    My understanding is that the vapor-particle pairs (e.g., VSVPO1 ↔ ASVPO1) undergo dynamic and reversible gas-particle partitioning within the CMAQ aerosol module. However, primary organic aerosol (POA) emissions are emitted directly into the aerosol species (ALVPO1, ASVPO1, ASVPO2, ASVPO3, and AIVPO1).

    Question: Does this mean that POA mass emitted directly into these aerosol species does not participate in reversible partitioning, while only the corresponding vapor species (VLVPO1, VSVPO1, VSVPO2, VSVPO3, and VIVPO1) can evaporate and condense? Or can the emitted aerosol species also evaporate into their corresponding vapor species and therefore participate in reversible partitioning?

Best regards,

Darby

2

Hi Darby,

  1. Saturation concentrations, C*, indicate how volatile a species is. Species of C* 1, 10, 100 ug/m3 often have substantial mass in both phases. Species of higher and lower volatility still theoretically exist in both phases, but they are generally predominantly in one phase. The gas-particle abundance depends on the concentration of organics in the environment. See equation 1 of Donahue et al. 2006: https://doi.org/10.1021/es052297c.
  2. A/V prepended species are paired and partition between phases. Something emitted in the aerosol (ALVPO1, etc) will later dynamically partition between phases based on environmental conditions.

Havala

3

@Havala.Pye Thank you. I just want to make sure I understand correctly.

  1. For the DESID table, are the fractions used to distribute POC emissions among the different C∗ bins and gas/particle phases a default emissions speciation assumption, while the actual gas particle distribution is subsequently calculated dynamically based on environmental conditions?

‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VSVPO1’, ‘GAS’ ,0.045,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VSVPO2’, ‘GAS’ ,0.14 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VSVPO3’, ‘GAS’ ,0.18 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘VIVPO1’, ‘GAS’ ,0.50 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘ALVPO1’, ‘FINE’,0.09 ,‘MASS’,‘a’,
‘EVERYWHERE’, ‘ALL’, ‘POC’, ‘ASVPO1’, ‘FINE’,0.045,‘MASS’,‘a’,

  1. For gas particle partitioning, do the A/V species partition only within their corresponding pairs, for example, ASVPO1 ↔ VSVPO1 and ASVPO2 ↔ VSVPO2? My understanding is that movement between different volatility bins, such as VSVPO1 → VSVPO2, would occur through chemical aging rather than gas particle partitioning.
  2. If so, does the same process apply to the CRACMM species below, for example, AROCN2OXY2 ↔ VROCN2OXY2, AROCP0OXY2 ↔ VROCP0OXY2, and AROCP1OXY1 ↔ VROCP1OXY1?

Best regards,

Darby

4

Yes! This is all correct.

5

@Havala.Pye

Thank you. I also have a question about how combustion related SOA is treated in CRACMM.

“The CRACMM mechanism does not use PCSOA or PCVOC because it provides a more mechanistic treatment of organic aerosol formation.”

Does this mean that the organic compounds previously represented by PCVOC/PCSOA are now represented through CRACMM ROC species, such as VROCP3–6ALK and VROCP5–6ARO, which can undergo oxidation and contribute to SOA formation?

Or are PCVOC/PCSOA represented by a different set of CRACMM species and chemical pathways?

Best regards,

Darby

6

Hi Darby,

In CRACMM, we built each SOA system from the bottom up. So either OA was directly emitted (generally with a specified volatility) or SOA was parameterized from a gas-phase precursor. For example, in the case of aromatics, we used a semiexplict chemistry, MCM, and then updated based on literature. Any semi/low volatility species were given a C*; glyoxal products were subject to heterogeneous chemistry. We also added new SOA paths from furans, oxygenated VCP emissions, and phenolic compounds that create SOA that is not created in aero7. We never intended to add back in pcSOA. After all the bottom up construction, we found CRACMM1 could produce OA levels on par with aero7 including pcSOA which confirmed we did not need to add any missing precursor or SOA path. Havala

7

Hi @Havala.Pye ,

Thank you for your explanation

I have more questions:

  1. I would like to confirm my understanding of the VROCALK oxidation pathway in CRACMM.

It appears that only the gas phase VROCALK species react with atmospheric radicals and form more oxygenated products such as VROCOXY

  • AROCALK + VROCALK → gas phase VROCALK oxidation → VROCOXY → repartitioning → secondary/aged organic aerosol

Also, VROCOXY species contribute to SOA only after partitioning into the particle phase as the corresponding AROCOXY species, but I cannot see any of AROCALK or AROCOXY species in the Figure below.

  • So, would the following pathway be correct?

VROCALK → oxidation → VROCOXY → gas particle partitioning → AROCOXY → SOA

image
image837×428 89.1 KB

  1. About the aircraft gas turbine POA volatility distribution used in CRACMM.

In the CRACMM speciation file using Lu’s paper, the mass allocation factors are 0.149, 0.130 + 0.130, 0.384, and 0.208 for the corresponding volatility bins. These sum to 1.001.

image
image680×225 4.68 KB

In Lu et al. (2020), Table 3 reports the gas turbine volatility distribution as 0.15, 0.26, 0.38, and 0.21, which sums to exactly 1.00.

image
image494×215 19.7 KB

The CRACMM values appear to be higher-precision versions of these values, but could you clarify where the additional decimal precision in the CRACMM factors (e.g., 0.149, 0.384, and 0.208) came from? Were these obtained from the raw data used by Lu et al. rather than directly from the rounded values in Table 3?

Thanks,
Darby

8

Hi Havala,

I have one more question about the autoxidation treatment in CRACMM. The CRACMM paper states that:

“Autoxidation, a gas phase reaction particularly effective in producing SOA, was added for C10 and larger alkanes, aromatic hydrocarbons, sesquiterpenes, and monoterpene systems…”

Is there a comparable autoxidation pathway for C10 and larger alkanes in CB6, or is this chemistry specific to CRACMM?

My understanding is that CB6 represents many alkanes through carbon bond species such as PAR, whereas CRACMM retains more information about carbon number and volatility, allowing C10+ alkanes to undergo explicit autoxidation and form lower volatility products. Is that interpretation correct?

Thank,

Darby

9

Hello, I’ll do each item one at a time

  1. ROCALK Chemistry

This flow is correct–note that these are species names/labels and the underlying ROCALK chemistry is based on GECKO-A with autoxidation added.

The figure you show is Figure 5 from Pye et al. (2023) ACP. For any figures taken from other sources, please add a citation so the source and context are documented. The figure is for the aromatic-phenolic-furan system (not the alkane aka ROCALK system).

10
  1. Aircraft volatility profile

The DESID file sum to 1.001 looks like a rounding error. The Lu et al. (2020) work was done is collaboration with EPA–it looks like the published table rounded values more than the DESID file. Note that starting with CRACMM2 CMAQ model-ready emissions, the volatility profiles were moved into the SMOKE inputs using S2S-Tool.

11
  1. Alkane chemistry in CB6 and alkane SOA aero6/7

You are correct that CB6 treats alkanes (and alkane functionality) as “PAR”. AERO6/7 does have alkane SOA–it is based on empirical fits to lab data (Pye and Pouliot, 2012) whereas CRACMM is based on GECKO-A+autoxidation and the corresponding volatility of those products (Pye et al., 2023 ACP). PAR chemistry (for ozone) and AERO6/7 SOA are computed in parallel.