Seth Chunn
University of California, San Diego
Scripps Institute of Oceanography
Russell Group

Abstract
  Atmospheric aerosol was collected in La Jolla, CA and separated into 1 μm and 180 nm size cutoffs.  The two subsets of data were then analyzed with Fourier transform infrared (FTIR) spectroscopy in order to analyze differences in organic functional group compositions for the two separate size cutoffs.  The results show clear differences in organic functional group composition between the size cuts.  The 180 nm cutoff showed an increased percentage of alkane groups when compared to the 1 μm cutoff, while some of the results varied in other organic functional group compositions such as, alcohols, organic acids, amines, and carbonyls.


I. Introduction
    Both urban and natural cycles and reaction have effects on the atmosphere, and in order to understand their characteristics chemical analysis must be carried out (Russell et. Al 2009).  Aerosol particles come in a variety of sizes ranging from coarse to fine aerosol particles (Jacobson et al. 2000).  Depending on the method used to collect aerosol particles, certain sizes can be collected separately for size-resolved chemical analysis.  By using a sharp-cut cyclone (1 μm cutoff) and Brenner impactor (180 nm cutoff) side by side, each with different size cutoffs, aerosol particles can be collected and sorted simultaneously by size into two separate subsets of data.  Through this sorting, the two different sets of aerosol can be analyzed for chemical composition
   
II. Methods
 In this experiment, a sharp-cut cyclone, 1 μm size cutoff, and a multistage Brenner impactor, 180 nm size cutoff, were used in parallel to collect aerosols within their respected size regions on Teflon filters.  The samples were collected in La Jolla, California on six separate days, October 10th, 16th, 17th, and 18th as well as November 1st and 4th of the year 2013. To determine the best time resolution needed to collect 180 nm particles above the detection limit for the FTIR, the duration of the samples   were varied from 3.75-6.5 hours. 
 The cyclone and Brenner impactor were installed behind a dryer and before Teflon filters to be collected simultaneously.  To reach the appropriate size cutoffs, the cyclone was run at 16.7 SLPM, with 7.0 LPM traveling through the Teflon filter, while the impactor was run at 46 SLPM, with 10.0 SLPM traveling through the filter.
Once collected, the filters were placed in a clean room and allowed to reach equilibrium with the environment at 20C and 55% RH.  After the filters equilibrated for 24 hrs, the samples were analyzed by Fourier transform infrared spectroscopy for organic functional groups.  Using a method from the Russell group, an algorithm was applied to baseline the spectra and quantify organic functional groups (alkane, alcohol, carboxylic acid, carbonyl and amine) (Russell et al. 2009). The composition was calculated as total mass (μg m-3) and mass percentages of alcohols, organic acids, carbonyls, amines, and alkane groups for each filter.
 When available, meteorological data (including wind speed and wind direction) was also taken from the Scripps Pier weather station (www.meteora.ucsd.edu/wx_pages/scripps.html).

III. Results
   
 
 

Figure 1 - Normalized spectra, total mass, and average organic functional group mass percentages for aerosol sampled with a cyclone (PM1, 1 μm) and Brenner impactor (180 nm) in October 2013. Pie charts represent functional group percentages: alcohol (pink), alkane (blue), carbonyl (light blue), amine (orange), and carboxylic acid (green).

 
Figure 2 – Wind rose plot showing the direction of wind on the two sampling days (10/10 and 10/18) and provides prospective of the location.  The widths of the orange and blue sections represent the standard deviation of the wind direction.

IV. Discussion
 Apparent differences in the chemical compositions are seen between the different size cuts. The differences are evident in the overall shape of the normalized spectra and average organic functional group composition.  By viewing the masses of the PM1 and 180 nm filters, the PM1 filter had samples 2-10x the mass of the 180 nm filters.
 By comparing the weather data from Scripps Pier on October 10th and 18th with organic functional group composition between the size cuts, some interesting conclusions about source regions are reached (Figure 2).  On October 10th, 2013 the wind had an average direction of 238O from true north, meaning the wind was coming from the ocean and free of anthropogenic aerosol sources.  As a result, the filters had a significant alcohol composition (>45%) in both the 1μm and 180 nm size cuts, which is consistent with marine aerosol (Russell et al. 2010).  Marine aerosols are formed through natural processes, such as evaporation and bubbles caused by wave action (Finlayson-Pitts and Pitts 382).  There is a difference, however, in the alcohol percentage between the 1μm and 180nm cutoffs, with the 180nm level containing 15% less alcohol groups than 1μm aerosol.  This relationship would suggest that the marine aerosols are less present in the 180nm cutoff.
 In contrast to the primarily marine source, October 18th has a significantly smaller alcohol percentage at both the PM1 and 180nm levels.  When analyzing the wind direction, it was found that the wind was coming from a northern direction, 341O from true north.  North of La Jolla lies several major urban communities including the Los Angeles basin, Riverside, and Oceanside.  The increased composition of carboxylic acid and alkane groups are in agreement with urban, or anthropogenic aerosol sources.  Urban aerosol consists of a wide variety of groups that are created from combustion reactions used in manufacturing plants and gas powered vehicles, which includes enhanced alkane functional groups and typically strong ammonium signatures (LIU et al. 2012).  Comparing both the 1μm and the 180nm aerosol from October 18th, there is a clear difference in alkane composition between the size cuts, with 180nm containing more alkane functional groups (45%).  A similar difference was seen in the October 10th sample, suggesting that the 180nm size bin contains a larger percent of alkane than it’s 1μm counterpart. 
Another relationship amongst the samples is that the ammonium absorbance was higher in 180nm samples than the 1μm samples.  This trend can be explained because fossil fuel combustion creates ammonium functional groups.  The reason the ammonium aerosol is found in the 180nm size bin is due to the fact fresh combustions create small particles (Liu et al. 2012).
 When only the 180nm samples are compared, the composition is relatively invariant.  Williams et. al (2002) compared the residence time of aerosol particles as a function of time.  It was found that the 100-200nm particles exhibited the highest residence time in the lower atmosphere.  Since the particles had a residence time of around 2 days, the chemical composition observed in this experiment should not have drastic changes (Williams et al. 2002).  The consistency between the accurate samples suggests this is in agreement with the experiment.  PM1 particles have a shorter residence time due to their size, and therefore the PM1 size cut off is more susceptible to variability, also in agreement with the data.
   
V. Error Analysis/Conclusion
 The first inconsistency with the experiment can be seen by comparing the total organic mass acquired on the filters.  The 180 nm samples should have a lower mass in comparison to the 1μm filter.  The samples from 10/16 and 11/1 have masses that are too similar between the size cuts, within 0.1 μm.  The 180 nm impactor could be experiencing a bounce effect, where larger aerosols bounce off impaction stages and proceed to the downstream, smaller stages.  Bounce could eventually lead to these larger particles being accidentally sampled.  As a result, bounce can lead to an error in the total organic mass, as well as the organic composition.  Recognizing this possible error, a future experiment would include more sets of filters spread across more days.  In order to reduce the bounce effect, Vasoline could be added to the first stage to act as an adhesive.  However, this would introduce more error to the system because Vasoline is an organic and would influence the results.  In order to limit bounce and conserve the data, the films within the impactor will be replaced.

References

 Finlayson-Pitts, Barbara and Pitts, J.: Chemistry of the Upper and Lower Atmosphere. San Diego: Academic Press, 2000. pp. 382-383. Print.

Jacobson M. C., H.-C. Hansson, K. J. Noone, and R. J. Charlson. Organic Atmospheric Aerosols. 2000, Reviews of Geophysics pp. 267-294.

Liu, S., Ahlm, L., Day, D., Russell, L., Zhao, Y., Gentner, D., Weber, R., Goldstein, A., Jaoui, M., Offenberg, J., Kleindienst, T., Rubitschun, C., Surratt, J., Sheesley, R., and Scheller, S.: Secondary organic aerosol formation from fossil fuel sources contribute majority of summertime organic mass at bakersfield, Journal of Geophysical Research, 117, doi:10.1029/2012JD018170, 2012.

Russell, L.M., L.N. Hawkins, A.A. Frossard, P.K. Quinn and T.S. Bates, Carbohydrate-like composition of submicron atmospheric particles and their production from ocean bubble bursting, Proceedings of the National Academy of Sciences of the United States of America 107(2010), pp. 6652-6657.

Russell, L. M., Takahama, S., Liu, S., Hawkins, L. N., Covert, D. S., Quinn, P. K., and Bates, T. S.: Oxygenated fraction and mass of organic aerosol from direct emission and atmospheric processing measured on the r/v ronald brown during texaqs/gomaccs 2006, Journal of Geophysical Research-Atmospheres, 114, 10.1029/2008jd011275, 2009.

Williams J. et al. "Application of the variablility-size relationship to atmospheric aerosol studies: estimating aerosol lifetimes and ages" Atmos. Chem. Phys., 2, pp. 133–145, 2002. Web.