4. Chamber modelling of aromatic system using MCMv3.1.

Task 4.1. Simulation of toluene chamber experiments

• Explore the toluene mechanism on the MCM website using the search tool
• Extract the mechanism using the subset mechanism extractor. Select FACSIMILE format and check the box to include inorganic reactions.
• Copy and paste additional reactions for dilution and the chamber mechanism from this file to the end of the mechanism file
• Create an initial concentrations file containing the concentrations listed in Table 5. Remember concentrations must be converted to molecules cm^-3
• Set the appropriate variables using the values from Table 5
• Using data from tol_in.xls constrain the model to measured j(NO₂), relative humidity and temperature.

• Create a concentration output file to output the following species: TOLUENE, O3, NO, NO2 and OH.
• Run the model and download the output files
• Compare the results with the experimental data contained in tol270901.xls by pasting your results into the 'MCMv31' worksheet.

Q1.     How do the concentration profiles compare to the measurements?

In general, MCMv3.1 shows improved ability to simulate some of the EXACT observations and represents our current understanding of aromatic degradation. However, significant discrepancies remain concerning ozone formation potential and oxidative capacity of aromatic hydrocarbon systems:

• Peak O₃ is simulated well for benzene but over estimated for the substituted aromatics
• OH radical production is too low to account for the OH inferred from the rate of loss of the parent aromatic.
• For the majority of the systems the NO oxidation rate is under-predicted. This parameter is linked to the production of O₃ and the oxidative capacity of the system.

Ideas and strategies for resolving these issues have been suggested and additional laboratory and smog chamber experiments are required in order to investigate them further (see Bloss et al. 2005 for more details).

A number of possible mechanistic fixes that have been investigated include:

• A fast reaction of O₃ with an intermediate to produce OH would result in improved model-measurement agreement. However, a very large rate coefficient and substantial intermediate concentrations would be required to give good agreement.
• Conversion of NO₃ to HO₂ by a dummy reaction in the model decreases O₃ yields but does not significantly affect the aromatic decay rate.
• A modelled conversion of NO₂ to HONO on the secondary organic aerosol has been shown to improve model-measurement agreement for O₃, NO_x and toluene decay. However, the relatively high HONO concentrations generated have not been observed experimentally, and the reactive uptake coefficient required is much higher than the upper limits for such processes suggested in the literature.
• An increase in OH yield without additional NO to NO₂ conversion would improve the simulated concentration time profiles.

As mentioned above, one means of decreasing ozone production while increasing OH concentrations is to implement a regeneration of OH without conversion of NO to NO₂ (and hence O₃). One such OH regeneration route postulated is a H shift in the peroxy radical formed in the peroxy bicyclic route, which is shown in Figure 1 below:

Figure 1. Schematic of MCMv3.1 peroxy bicyclic route for toluene oxidation and postulated OH regeneration pathway (Bloss et al. 2005b).

Introduce this OH regeneration route into the MCMv3.1 toluene model. Open your toluene mechanism file and add the following OH regeneration route to the end of the existing reactions

• % KDEC*0.167 : TLBIPERO2 = GLYOX + C5DICARB +OH ;
• % KDEC*0.167 : TLBIPERO2 = GLYOX + TLFUONE + OH ;
• % KDEC*0.25 : TLBIPERO2 = MGLYOX + MALDIAL + OH ;
• % KDEC*0.167 : TLBIPERO2 = GLYOX + C4MDIAL + OH ;
• % KDEC*0.25 : TLBIPERO2 = MGLYOX + BZFUONE + OH ;

This scheme replaces the usual peroxy bicyclic ring opening route. Therefore, comment out the following reactions (insert a "*") in front of each line:

• % KRO2NO*0.889 : TLBIPERO2 + NO = NO2 + TLBIPERO ;
• % KRO2NO*0.111 : TLBIPERO2 + NO = TLBIPERNO3 ;
• % KRO2NO3 : TLBIPERO2 + NO3 = NO2 + TLBIPERO ;
• % KRO2HO2*0.820 : TLBIPERO2 + HO2 = TLBIPEROOH ;
• % 8.80D-13*RO2*0.60 : TLBIPERO2 = TLBIPERO ;
• % 8.80D-13*RO2*0.20 : TLBIPERO2 = TLBIPER2OH ;
• % 8.80D-13*RO2*0.20 : TLBIPERO2 = TLOBIPEROH ;

Save the mechanism as 'tol_ohregen.fac'. Run the model and compare the results with the experimental data contained in tol270901.xls by pasting your results into the 'OH regen' worksheet.

Q2 How do the new concentration profiles compare to the measurements and MCMv3.1 model runs?

The OH regeneration step currently seems like the most favoured explanation of the under prediction of the radicals but over prediction of ozone from the updated aromatic degradation in MCMv3.1. High level ab-initio calculations on aromatic systems are being used to investigate the likelihood of such an OH regeneration route occurring under tropospheric conditions and a similar step has been added to the SAPRC mechanism which is the most used tropospheric chemistry mechanism incorporated into air quality models in the US.

Task 4.2. Additional Exercise (for you to carry out at home): Simulation of a cresol chamber experiments

The chemistry of hydroxyarenes was updated following new product studies indicating higher ring retaining products in aromatic systems than previously assumed (see Bloss et al. 2005a for more details). As part of the EXACT campaign a photo-smog experiment on cresol (a subset of the toluene mechanism) was performed on the 04/10/2001.

• Open the initial concentrations file created for the toluene experiment and update it to the initial condition for the cresol experiment performed on the 04/10/2001 as listed in Table 6. NOTE: remove the initial concentration of TOLUENE or replace it with CRESOL.
• Start the model at the appropriate time and output every 5 minutes until the end of the experiment.
• Constrain the model to measured water, j(NO₂) and temperature data(cresol data).
• Modify the output to calculate the following species: CRESOL, O3, NO, NO2 and OH.

Run the model and compare the results with the experimental data and a similar model run using the MCMv3 toluene mechanism contained in cresol041001.xls by pasting your results into the 'MCMv31' worksheet.

Q3      How do the new concentration profiles compare to the measurements and the MCMv3 results?