Glyoxal Production Pathways

1. Updates to the MCM Model

The last time these groups met, there were significant descrapancies between the Mozart T1 and Master Chemical Mechanism (MCM) in their formation of glyoxal. Specially, MCM made very little glyoxal and its diurnal pattern was considerably different compared to MCM. It was recently discovered that MCM within BOXMOX handles the photolysis of glyoxal considerably different. While the input files for photolysis rates only provided one rate for glyoxal, MCM was inputting the photolysis rate into three different reactions that were equal to each other. This led to the photolysis rate being 3 times greater than intended and thus affected the diurnal cycle. Mary reached out to Christoph Knote, the creator of BOXMOX, and confirmed the issue. Once correcting for this in MCM, the glyoxal mixing ratios look much better. Glyoxal mixing ratios remain greater early in the simulation within MCM, and reach a very similar peak compared to Mozart, and maintaining higher mixing ratios through the simulation.

2. Map of the Trajectories

This map shows the direction of the various trajectories colored by their initial starting height. Trajectories were started at on June 27th at 7:30 pm EST at 100, 200, and 500 meters, with new trajectories being launched every 30 minutes. The numbers on the plot represent the time marker in hours. Generally, trajectories launched at 100 meters turned Eastward, with many trajectories moving towards the Coast of New England while trajectories launched at 500 meters moving more Northeastern over Canada.

3. Planetary Boundary Layer Height Effects on Chemistry

In order to account The model uses a turbulence coefficient, Kturb, to estimate turbulent mixing with background air within the box. In previous experiments, Kturb was estimated using the fractional air/DT, where fractional air is defined as:

\[ \text{Fractional Air} = \frac{Observed - Previous}{Background - Previous} \]

However, this scheme was leading virtually every compound to become their background levels and stagnated the chemistry. Instead we made two assumptions:

1. Any trajectory that flies above the PBLH, emissions are immediately set to 0.
2. We assumed fractional = 0.7%, determined by PBL dilution experiments run previously. We don’t use the dilution scheme the whole time as it assumes that we are in boundary layer the whole time, which is very often not true in our case.

We then add the Kturb term to the convection equation within the box model. See the readme file for more information about the turbulent mixing schemes within boxmox (https://boxmodeling.meteo.physik.uni-muenchen.de/downloads/README.pdf).

Due to a lack of observational data for most species, we limited the background air to just CO, NO, NO₂,SO₂, and O₃.

4. OH Budget

Before examining the production of glyoxal, it’s important to look at the OH budget to ensure that simulations are working properly. For the production of OH, the major pathways are similar between MCM and Mozart. However, there seems to be a greater production of 5x10⁶ molecules cm^-3 s^-1 in Mozart compared to MCM. This may be due to less reactions that contain oxidation by OH in Mozart compared MCM. The major production pathways include photolysis of H₂O₂, HO₂+NO, HO₂+O₃, and O1D+H₂O.

There are virtually no differences in OH sinks between the two mechanism, with CO, CH₄ and CH₂O being the major loss pathways.

5. Production of Glyoxal

5.1 Mozart T1

Looking the production of glyoxal within Mozart T1, there are 4 major production pathways. Glycolaldehyde (GLYALD) is an important source that is largely produced by secondary sources but also has primary (mostly anthropogenic) emissions as well. However, there are few emissions of GLYALD in the majority of the trajectories, so mixing ratios of GLYALD are controlled by the initial conditions and chemistry. NC4CHO is an isoprene nitrate.This result is surprising as there is very little NO_x within the system and yet it seems to be a important production pathway for glyoxal. This could have important implications for future experiments that could have higher NO_x mixing ratios, as this was a very clean case. XO₂ is another isoprene derivitive, highlighting that biogenic influences on glyoxal are strong and to be expected. Lastly, TOLO₂ is toluene oxidation product, specifically a toluene bicyclic peroxy radical. This is a parameterization of the toluene reaction, as it’s the TOLO species degrading to glyoxal rather than the TOLO₂ directly forming glyoxal. The toluene alkoxy radical decays in 5 different ways, with a 60% yield of glyoxal (and interestingly, the other 40% decays into methylglyoxal). Even with low mixing ratios, anthropogenic VOCs have the potential to be important in the formation of precursor gases of aqueous chemistry.

\[ GlyAld + OH --> 0.2 Glyoxal \]

\[ NC_4CHO + OH --> Glyoxal + Products \]

\[ TOLO_{2} + NO --> 0.6 Glyoxal \]

\[ XO_2 + NO --> 0.1 Glyoxal \]

5.2 MCM 3.3.1

5.2.1 Lumped species

There are a few lumped species within Mozart, representing larger VOCs that are generally in low mixing ratios individually, and often have very ill-defined rate constants. In MCM, many of these specified within the mechanism, requiring the lumped species to be defined. The lumped species include BIGALK (alkanes > 3 carbons), BIGENE(alkene > 3 carbons), HYDRALD(lumped hydroxy carbonyl), and XYLENES(lumped xylene species). I used KORUS-AQ VOC whole air sample to speciate the lumped species. https://www-air.larc.nasa.gov/missions/korus-aq/. After determining which species belonged to lumped species , I then found the average fraction of each species, which was used to convert the Mozart lumped species to individual VOCs in MCM. I’m hoping to use more VOC data in the future, collected closer to New York for better estimates. However, when looking at Mozart simulations, lumped species play little role in the species that we are interested in for cloud water chemistry, so uncertainties in lumped species likely aren’t important for us.

5.2.2 MCM Results

Here we investigate the glyoxal production within MCM. MCM contains over 150 reactions that produce glyoxal. After looking each reaction, these are the reactions I found to be the most important in this simulation. I have a separate file that has all of the reactions that I will send with this writeup. The plot below shows many similarities to the Mozart T1 mechanism. For instance the oxidation of GLYALD (HOCH₂CHO within MCM) remains an important production pathway for glyoxal, though it represents a greater fraction in MCM. TLBIPERO and TLOBIPEROH are similar to the TOLO2 that is parameterized within Mozart T1, but represent a considerably smaller contribution of total glyoxal production. When looking at the mechanisms for toluene oxidization products in more detail, there are some missing photolysis rates that may lead to some underestimates of glyoxal production. This is something we’ll look into greater detail. Lastly, C510O represents an isoprene formed nitrate, similar to the NC4CHO within the Mozart mechanism. In total, the major production pathways in MCM are fairly similar to Mozart but have different relative contributions and the overall production rate within Mozart is greater.

6. Effect of Starting Height

There are important differences in glyoxal mixing ratios at the 3 different starting heights of the trajectories, specifically at 200 and 500 meters. In the first half of the simulation, glyoxal mixing ratios between the two mechanisms largely agree, but diverge in the second half.

If we look at the production rates by starting height, we find that GLYALD grows in importance and becomes the dominant production pathway in MCM. In fact, in the first half of the simulation, glyoxal production is greater at 500 meters than at 100 meters for MCM. These results suggest looking at GLYALD in more detail.

7. Glycolaldehyde

7.1 Influence on Glyoxal Production Rates

A closer look at GLYALD reveals some interesting results. At 100m, production of glyoxal from GLYALD are similar, with slightly higher production within Mozart, with the discrepancy growing towards the end of the simulation. However, trend is reversed with 500m trajectories, with MCM showing greater production of glyoxal from GLYALD. This difference is not driven by differences in OH mixing ratios as the mechanisms largely agree with at all launch heights, or rate constants, as the GLYALD+OH rate constant is equivalent in both models. This difference is driven by the diverging mixing ratios of GLYALD between the two models at the higher launch heights, implying that sources and/or sinks are different between the two mechanism. These results warrant a closer look at GLYALD chemistry within the simulation.

7.2 GLYALD Production

This section looks at the differences in sources and sinks of GLYALD between the two mechanisms. Similar to glyoxal, there were errors implementing the correct photolysis rate for GLYALD with MCM, so previous estimates of GLYALD were likely inaccurate, which has now been corrected. Both models largely agree with the major sinks of GLYALD, with OH being the dominant pathway, and minor removal from photolysis.

The major pathways of GLYALD production within Mozart include:

\[ EO + O_2 --> GlyAld + Products \]

\[ MACRO_2 + CH_3O_2 -- > 0.26GlyAld + Products \]

\[ XO_2 + NO --> 0.25 GlyALD + Products \]EO is an alkoxy radical produced from the oxidation of C₂H₄. MACRO₂ is a lumped species that is produced by the oxidation of MACR and MVK by OH, which are important oxidation products of isoprene. Interestingly, NO_x levels are so low in this simulation, MACRO₂ is more likely to react with another peroxy radical. It’s important to note that MACR isn’t though to be an important production pathway for GLYALD, so production from MACRO₂ could be overestimated. Similar to MACRO₂, XO₂ is another isoprene oxidation lumped species, specifically oxidation products from IEPOX, HPALD, and ISOPOOH. At higher launch heights, as isoprene is depleted from the system, C₂H₄ becomes the most important production pathway., isoprene chemistry is the most important production pathway with important contributions from C₂H₄.

The major production pathways for GLYALD in MCM are as follows:

\[ HMVKBO -- > GlyAld +Products \]

\[ HYETHO_2H + OH --> GlyAld + Products \]

\[ HOCH_{2}CH_{2}O --> GlyAld +Products \]

These pathways share many similarities with Mozart. HMVKBO is an oxidation production of MVK (with no influence from MACR) and HYETHO2H and HOCH₂CH₂O are oxidation products of C₂H₄ and C₂H₅OH is very similar to EO within Mozart. While the major production pathways are the same between the two mechanisms, the contribution of anthropogenic vs biogenic is reversed, with greater biogenic contribution within Mozart and greater anthropogenic contribution with MCM. Removing potential contributions from MACR, there is a noticeable reduction in biogenic influence. In addition, pathways from oxidation of ISPOOH, HPALD, and IEPOX don’t directly become XO₂ species but several other compounds with various reaction rates, leading to further reduction in isoprene chemistry contributions. I’m currently investigating this exactly what the differences between the two mechanisms as these differences could be compounded in the dirtier simulations we want to run in the future.

The greater production of GLYALD at higher starting heights within MCM is driven by a combination of higher mixing ratios of the HYETHO2H (which makes sense for a hydroperoxide as they will have a longer life time) and higher mixing ratios OH.

8. Review of Other Important Cloud Water Chemistry Gases

Here I include some production plots of other important gases that are important to cloud water chemistry, including HCOOH, CH₃COOH, and Methylglyoxal. As seen in previous meetings, HCOOH is produced within MCM compared to Mozart. HCOOH is often underproduced within models, with theories including uncertainties within cregiee intermediates, missing isoprene chemistry, and biomass burning emissions.

While Mozart produces more HCOOH than MCM, after discussions with John Orlando of NCAR, HCOOH production likely isn’t correct as Mozart was not created with HCOOH in mind. It may be worth looking at other mechanisms with more updated isoprene chemistry to better investigate gas phase HCOOH production.

For CH₃COOH, Mozart and MCM agree almost entirely, with perhaps slightly more production within MCM compared to Mozart. Both models seem to show quite a bit of prodution in the first half of the simulation, particularly at the 100 meter starting height.

There is a small divegence between the mechasims at higher starting heights, which is likely related to the differences in CH₃CO₃ oxidation by HO₂ between the two mechanims. The sources of CH₃CO₃ will be investigated in the future.

Lastly, there are some slight differences in methyglyoxal production between the two mechanisms. There seems to some nocturnal production of methylglyoxal early in the simulation within MCM. At 100 meters, production resumes, with overall mixing ratio increasing over Mozart. There are about 190 reactions that produce methylglyoxal in MCM, so a future detailed investigation is warranted.

9. Conclusions

It’s clear that an error in how MCM handles photolysis of glyoxal was the major cause for the large disagreements of glyoxal between the two mechanisms. Once this issue was corrected, there was much better agreement. Both models had important contributions from GLYALD, Toluene oxidation products, and isoprene nitrate oxidation products, though GLYALD is more important within MCM. When looking at the GLYALD production C₂H₄ and MVK are important precurors but biogenic precursors are more important in Mozart while anthropogenic sources are more important in MCM. Differences also remain between the mechanisms for HCOOH, CH₃COOH, and methylglyoxal, with differences varying with different start heights of the trajectories. Work will continue to better understand the controlling factors of the important aqueous chemistry VOCs.