Head in a Cloud

troposphere and stratosphere meet blogosphere

Head in a Cloud random header image

Background on Rethinking Organic Aerosols: Semivolatile Emissions and Photochemical Aging by Robinson et al.

April 2nd, 2007 by Sean Davis · No Comments

Editor’s note: The following post was graciously contributed by Peter DeCarlo, a fellow Ph.D. candidate at the University of Colorado. His piece offers some background and commentary on the previous post.

Organic aerosols are studied in a variety of ways (field observations, lab studies, modeling, etc.) as most atmospheric constituents tend to be. With organic aerosols there have been considerable discrepancies in recent years between what is measured in the field, and what is modeled. Models assume a large fraction of the organic aerosol is primary based on early aerosol studies involving GC-MS (coupled gas chromatography and mass spectrometry). These studies found a significant portion of the resolvable aerosol to be primary.

Unfortunately, much of the secondary aerosol does not get through the GC column and this was skewing the data and showing too much primary aerosol (directly emitted) compared to secondary aerosol species (from e.g. gas-to-particle conversion). An additional source of error in the model input for primary aerosol is addressed in Figure 1A the Robinson et al. paper (Figure in Previous Post). Standard sampling of automotive exhaust is done in undiluted conditions. Exhaust is diluted by 1,000 to 10,000 on a timescale of seconds under ambient conditions. If one samples undiluted exhaust the emission factor for Primary Organic Aerosol is ~5 times too high. Obviously putting the undiluted Emission Factor into a model will result in an overestimate of primary organic aerosol.

On the other hand secondary organic aerosol (SOA) has been grossly underestimated in models. Field data suggests that the majority of organic aerosol is secondary in nature. The figure below (from Volkamer et al. GRL, 2006) plots measured SOA over Modeled SOA and shows that in Mexico City models under predicted Secondary organic aerosol by a factor of 8 on timescales of a day. Results from the TORCH campaign downwind of London and NEAQS 2002, show underestimates of a factor of 10, and The ACE-Asia campaign in 2001 shows underestimates of SOA up to a factor of 100. Clearly the models are missing something.


Now the models use explicit precursors to predict SOA. One of the higher yield types of species (as measured in the laboratory) from anthropogenic emissions is toluene and other aromatic hydrocarbons (referred to as high yield aromatics). These species are easily measured in the gas phase, the rates of reaction are fairly well known, and the yield of organic aerosol is also known. According to SOA models high yield aromatics should contribute ~70% of the modeled SOA mass. Accounting for these species in explicit chemistry models yields the underestimates in the above figure (underestimates of 8-100 for SOA). What the Robinson et al. Science paper does, is present SVOCs as precursors for a part of the missing secondary organic aerosol mass. Looking at data I have collected in the Mexico City region and doing a simple calculation, I find that including SVOCs does reduce the discrepancy between measured and modeled organics considerably, but does not complete the picture. There is still a significant amount of unaccounted for organic aerosol in the atmosphere.

SVOCs are primary emissions, however, classifying them as aerosol or gas is somewhat of a sticky question. The terminology as I understand it classifies these compounds as gas phase emissions, and not as primary aerosol emissions. This is because under atmospheric conditions they exist primarily in the gas phase, and not the particle phase. As these gas phase species get oxidized (and become less volatile) they partition into the particle phase as secondary organic aerosol. Under this paradigm primary organic aerosol from motor vehicle emissions can be considered static.

My final comment on SVOCs is that these are compounds that have yet to be measured and quantified. The Robinson et al. Science paper infers their existence on the basis of the volatility distribution needed to produce the measured SOA. I have no problem with this inference, and agree that it is a valid approximation of the expected volatility distribution. I think the challenge is for gas phase and aerosol scientists to measure and quantify these semi-volatile compounds. While these species will not contribute much to the total carbon mass in the gas phase (most of that has already been accounted for), the Robinson et al. paper shows that these species play a very important role in the relationship between gas-phase organics and aerosol phase organics.

Tags: aerosols · field measurements · modeling · troposphere

0 responses so far ↓

  • There are no comments yet...Kick things off by filling out the form below.

Leave a Comment