Imholt 1. Optical properties of secondary organic aerosols using ultraviolet/visible spectroscopy. Felisha Imholt

Transcription

1 Imholt 1 Optical properties of secondary organic aerosols using ultraviolet/visible spectroscopy Felisha Imholt

2 Imholt 2 I. ABSTRACT It is well known that the increased warming effect due to greenhouse gases is a major environmental concern. While the amount of solar radiation absorbed by greenhouse gases is known to a high certainty, the amount absorbed by secondary organic aerosols (SOA) is not. Our study aimed to discover how much radiation SOA particles absorb between ~200 and 800 nm. The SOA were created in one of two temperature controlled Teflon chambers within Dr. John Shilling s lab at PNNL and were collected on 47 mm Teflon filters. We used an ultraviolet-visible (UV/Vis) spectrometer equipped with a liquid waveguide capillary flow cell to determine the amount of radiation SOA absorbs at various wavelengths. We developed and optimized UV/Vis procedures for SOA dissolved in three solvent systems: water, methanol, and 0.1 M hydrochloric acid. We also successfully determined the mass absorption coefficient values and imaginary refractive indices for the generated SOA. These values will be used in climate models developed at PNNL.

3 Imholt 3 II. INTRODUCTION Much uncertainty surrounds the impact of tropospheric aerosols on the enhanced greenhouse effect. Aerosols can affect the Earth s energy balance through direct and indirect radiative effects (Bluvshtein, Flores, Riziq, & Rudich, 2012). Currently, the amount of light absorbed and scattered by aerosols is uncertain. This is of high importance in current atmospheric research because of the impact light absorption can play on the atmosphere. Absorption by aerosols could raise the atmospheric temperature by absorbing light and radiating it back into the atmosphere in the form of heat. (Alexander, Crozier, & Anderson, 2008). These aerosols are created from both human and natural sources. A large fraction of anthropogenic aerosol mass comes from fossil fuel and biomass burning in the form of carbonaceous particles. Carbonaceous aerosols also come from natural biogenic emissions (Alexander, Crozier, & Anderson, 2008). Biogenic emissions result from the combustion or decomposition of biologically-based materials that are not fossil fuels. When these semi-volatile organics oxidize, the resulting condensed phase particles are termed secondary organic aerosols (SOA) (Kroll & Seinfeld, 2008). An increase in the emission of both anthropogenic and biogenic emissions would produce more SOA (Alexander, Crozier, & Anderson, 2008). Once formed, SOA can undergo physical and chemical changes in the atmosphere such as coagulation, structural rearrangement, and phase transition (Fuzzi et al., 2006). Several factors influence SOA formation including reactions in the presence of acid catalysts, relative humidity, and concentration of organics and oxidants (Kroll & Seinfeld, 2008). Some scientists suggest that 91% of SOA are caused by O 3 and OH oxidation (Chung & Seinfeld, 2002). Others propose that ozonolysis dominates SOA production compared to OH

4 Imholt 4 and NO x oxidation (Kanakidou et al., 2005). Organic material, which composes SOA, was previously thought to scatter light. However, SOA could be a type of light-absorbing carbon. Light-absorbing carbon (LAC) includes both black carbon (BC) and brown carbon (BrC) (Lambe et al., 2013). The optical properties of BC are known to absorb radiation at wavelengths above ~500 nm. BrC has been documented to absorb light at wavelengths lower than that of BC, at ~300 to 500 nm. This BrC has a different nature than black carbon (Alexander, Crozier, & Anderson, 2008). SOA absorbs light radiation between ~300 and 500 nm which indicates that it could be a form of BrC (Alexander, Crozier, & Anderson, 2008). Although SOA has been shown to absorb light, the amount absorbed is still unclear and is the topic of our research. Our research aimed to study SOA optical properties. By measuring the light absorbed by SOA, we calculated the mass absorption coefficient (MAC) value and imaginary refractive indexes (k values). These values will be used in climate models developed at PNNL. Our study will help to improve climate models and increase the understanding of SOA. III. EXPERIMENTAL A. Validation of UV/Vis Technique Using a 100-cm liquid waveguide capillary flow cell (LWCC) and an ultraviolet/visible spectrometer (UV/Vis), we determined the amount of ~ nm range light radiation absorbed by SOA. The UV/Vis had been used in the lab once prior to our experiment for the same purpose. We refined the previously developed technique here. The equipment was validated using Suwannee River fulvic acid (FA). Ghabbour and Davies (2009) as well as Lambe et al. (2007) have studied and described the MAC and k values characteristic of FA. By

5 Imholt 5 using FA and comparing their MAC values with ours, we can determine the validity of the refined UV/Vis procedure. Three solvents were used: water, methanol, and 0.1M hydrochloric acid (HCl). Methanol was used as a second solvent in order to elucidate the effect of a different solvent on SOA. The third solvent, 0.1 M HCl, was chosen because of an experiment done to characterize the peroxide content in SOA. This study required the use of 0.1 M HCl as a solvent. The 0.1M HCl was created by mixing 50 ml 0.5M HCl and 200 ml ultrapure Milli-Q water (Millipore ultrapure water system, Milli-Q Advantage A10). In order to create the calibration curves for the water, methanol, and 0.1 M HCl solvent systems, stock solutions were created with FA. The FA/water stock solution was used to dilute FA to solutions of various concentrations in both water and 0.1M HCl. Another FA stock solution was made for the methanol experiment. The concentrations of the two stock solutions are shown in Table 1. Amount FA Amount solvent Final Concentration g 50.5 g water 120 mg/l FA in Water g 7.95 g MeOH 110 mg/l FA in MeOH Table 1. Stock solutions in two solvents: water and methanol (MeOH). Dilute solutions were created between the concentrations of 0.20 mg/l and 0.80 mg/l at increasing intervals of 0.15 mg/l. The solution concentrations were approximately 0.20, 0.35, 0.50, 0.65, and 0.80 mg/l. The solutions were then put into 10-mL glass syringes and injected into the 100-cm path length LWCC (World Precision Instruments model LWCC-3100). The absorbance of light ( nm) generated using a miniature deuterium-tungsten halogen light

6 Imholt 6 source (Ocean Optics Inc. model DT-MINI-2-GS) was measured using a UV/Vis spectrometer (Ocean Optics, Inc. Jaz model UX ). B. Generation of SOA Particles The SOA particles used in the experiment were produced in one of two 10.6 m 3 Teflon reaction chambers (Xu et al., 2014). Each chamber consists of a 10 x 5 x 7 Teflon reaction bag enclosed in a single temperature controlled room (15-45 o C). Each chamber can be operated independently in either a continuous flow mode or in the traditional batch mode (Shilling et al., 2009). Light is provided constantly by 110 UVA-340 fluorescent lamps which surround the chamber. Volatile organic compounds (VOC) were added to the chamber by injecting the VOC into a glass bulb with a syringe and then gently heating it. The vapors were then transported into the chamber via a flow of purified air. To initiate the SOA formation, a variety of oxidants were used including OH. OH is created by photolysis of either H 2 O 2 or HONO (depending on whether experiments were run under high-no x or low-no x conditions). Additional NO is occasionally injected from a pre-mixed gas cylinder to further control the NO x concentrations, which determines the fate of generated organic peroxy radicals. Analysis of the particle and gas phases was recorded in real time using a variety of analytical instrumentation including an Aerodyne HR-ToF-AMS, an Ionicon PTR-MS, and a TSI SMPS. After generation, the SOA samples were collected on 47-mm Teflon filters with collection times ranging from 2 to 6 hours. The filters were cut in half with a razor blade and the SOA was extracted from the filter via ultrasonication in 10-mL of methanol, water, or 0.1 M HCl. Each side of the filter was sonicated for 5 minutes. Raw data from the Scanning Mobility Particle Sizer (SMPS) gave the volume of particles per volume of sample collected (#/cm 3 and

7 Imholt 7 nm 3 /cm 3 ). Using this data, calculated the concentration of SOA particles on the filter in mg/l. The absorbance of the resulting SOA solution was measured using the above outlined UV/Vis procedure used for the FA solutions. Our experimental SOA were generated from two volatile organic compounds (VOC), 1,2,4-trimethylbenzene or toluene, under either high-no x or low-no x conditions. The SOA were collected at the end of the day before the chamber was injected with H 2 O 2 and again at the beginning of the day. The injected H 2 O 2 caused the SOA to become more oxidized and possibly change their absorbance. IV. RESULTS A. Fulvic Acid Trials Calibration curves were created from the three different solvent systems to ensure the UV/Vis was working properly before studying the SOA. Figures 1 through 3 are the calibration curves for water, methanol, and 0.1 M HCl respectively. Absorbance (arbitrary units) Absorbance of FA in H2O y = x R 2 = Concentration (mg/l) Figure 1. Absorbance of fulvic acid in water at 405 nm. 0.8

8 Imholt 8 Absorbance (arbitrary units) Abrobance of FA in MeOH y = 0.129x R 2 = Concentration (mg/l) Figure 2. Absorbance of fulvic acid in methanol at 405 nm. Absorbance (arbitrary units) Absorbance of FA in 0.1M HCl y = x R 2 = Concentration (mg/l) Figure 3. Absorbance of fulvic acid in 0.1 M HCl at 405 nm. 0.8 As the concentration of FA increased, the absorbance at 405 nm increased in a linear fashion. Mass absorption coefficient (MAC) values (in m 2 /g) for FA were calculated using an equation from Lambe et al. (2013): [ ] Equation 1.

9 Imholt 9 where A is the measured absorbance at a given wavelength, [FA] e is the concentration of FA in the solvent, and L is the LWCC path length (1-m). The imaginary refractive index (k) was calculated using equation 2: Equation 2. where MAC is the MAC value obtained from equation 1, λ is the wavelength at which the MAC value was calculated, and ρ is the density of the solution. Table 2 shows the average MAC and k values at 405 nm for the three solvent systems. We calculated MAC and k values for Solvent MAC k Methanol 0.14 ± ± the FA to validate our procedure. These Water 0.15 ± ± equations were also used for calculating the MAC and k values for SOA (replacing [FA] e 0.1 M HCl 0.13 ± ± Table 1 with [SOA] e in equation 1 where [SOA] e is the concentration of SOA particles in the solvent). The MAC and k values can be used in climate models for predicting the future impact of SOA on the atmosphere. The MAC value calculated for FA in water is in agreement with the literature value of 0.17 m 2 /g from Ghabbour and Davies (2009). Furthermore, MAC values of the three solvent standards are similar with almost equal standard deviations. Our MAC values for FA are consistent with the literature values. Therefore, we are confident that the experimental procedure to determine the optical properties for SOA in water, methanol, and 0.1 M HCl is valid.

10 Imholt 10 B. UV/Vis light absorbance by SOA The following graphs show the absorbance of SOA in two, two day collections ( to and to ) at various concentrations. We collected SOA on a Teflon filter on the first day and recorded the absorbance. We collected and analyzed the SOA again on a separate Teflon filter the following day after H 2 O 2 was added. We did this to see if the further oxidized SOA absorbed more or less light radiation. Figure 4 shows the absorbance when the SOA were dissolved in water and Figure 5 shows them in methanol. The solutions absorb most strongly between 280 and 400 nm. However, the SOA absorb less than the FA samples did at similar concentrations because this particular SOA is less absorbing than FA. Nevertheless, the SOA absorb a significant amount of visible light. Absorbance (arbitrary units) A B C A: 1.24 mg/l B: mg/l C: mg/l D: mg/l Zero (Baseline) D Wavelength (nm) Figure 4. Absorbance spectra of all concentrations of SOA particles in water. 800

11 Imholt Absorbance (arbitrary units) A B C D A: 1.21 mg/l B: mg/l C: mg/l D: mg/l Zero (Baseline) Wavelength (nm) Figure 5. Absorbance spectra of all concentrations of SOA particles in methanol. Because the collection times and amount of SOA created vary, the concentrations of SOA are different for each day. However, the MAC values take into account this difference and allow us to compare the absorbance of SOA at different concentrations. Table 3 shows the MAC values at 405 nm for the SOA particles shown in figures 4 and 5. The second day of SOA collected (more oxidized) absorb slightly more light radiation than the first day. This indicates that oxidation causes SOA to absorb more light. The average MAC Date MAC in H 2 O MAC in MeOH values range from to m 2 /g for the four SOA samples. Lambe et al. (2013) found the MAC values generated in their laboratory to be between and Our results fall into the same range as the reported literature Table 2

12 Imholt 12 further providing confidence that our procedure is valid. The methanol average MAC values have a greater range, from to However, dissolving SOA in methanol has not been tested prior to our experiment so our results for these trials are new. V. CONCLUSION Through our work, the UV/Vis technique was optimized for use in three different solvent systems: water, methanol, and 0.1 M HCl. The MAC and k values for FA in each solvent are consistent with the literature values. With our verified procedure, SOA particles were analyzed and found to absorb strongly between 280 and 400 nm. The MAC and k values agreed with the literature value in water and were also close with our new solvent, methanol. The two collections of SOA over two days with added peroxide showed that the more oxidized SOA absorb more light than the less oxidized. Our research will be used in further studies on SOA and in climate models developed at PNNL.

13 Imholt 13 VI. ACKNOWLEDGMENTS Foremost, I would like to express my sincere thanks and gratitude to my faculty advisor Dr. Matthew Wise for this opportunity and for supporting me throughout the research. Thank you for patience, motivation, enthusiasm, jokes, and above all, guidance. I was never lost or confused with you as my advisor and when I was ready to work on my own, you let me branch out, make mistakes, and most importantly, succeed. Second, I would like to thank my partner Ryan Caylor for going on this journey with me. While we worked on our own projects, you still managed to help me with mine. I appreciate your comments and suggestions when I was trying to figure out a problem. Lab time was never dull with you and Dr. Wise. Many thanks go to Dr. John Shilling, our PNNL host, for the use of his lab. Thank you for inviting us to do this research and for helping us as we figured out what we were doing. You were always willing to answer questions, help us find a piece of equipment, let us try something new, and let us run our own experiment in one of the SOA chambers. We have learned so much about experimentation during our time here. Additionally, I would like to thank Pacific Northwest National Laboratory and the DOE for hosting me and providing support for research that I would not have been able to do otherwise. I d also like to thank SURI and the College of Theology, Arts, and Sciences at Concordia University for giving me this opportunity to grow and learn. This paper makes me want to shoot myself in the foot.

14 Imholt 14 VII. REFERENCES Alexander, D. T. L., Crozier, P. A., & Anderson, J. R. (2008). Brown carbon spheres in East Asian outflow and their optical properties. Science, 321(5890), doi: /science Bluvshtein, N., Flores, J. M., Riziq, A. A., & Rudich, Y. (2012). An approach for faster retrieval of aerosols' complex refractive index using cavity ring-down spectroscopy. Aerosol Science and Technology, 46(10), doi: / Chung, S. H., & Seinfeld, J. H. (2002). Global distribution and climate forcing of carbonaceous aerosols. Journal of Geophysical Research, 107(D19), 107. Fuzzi, S., Russell, L. M., Kulmala, M., Huebert, B. J., Andreae, M. O., Pöschl, U., Bond, T. C. (2006). Critical assessment of the current state of scientific knowledge, terminology, and research needs concerning the role of organic aerosols in the atmosphere, climate, and global change. Atmospheric Chemistry and Physics, 6(7), Ghabbour, Elham A. & Davies, Geoffrey (2009). Spectrophotometric analysis of fulvic acid solutions - A second look. Annals of Environmental Science: 3(10). Kanakidou, M., Balkanski, Y., Putaud, J. P., Swietlicki, E., Seinfeld, J. H., Pandis, S. N., Stephanou, E. G. (2005). Organic aerosol and global climate modelling: a review. Atmospheric Chemistry and Physics, 5(4), Kokkola, H., Yli-Pirila, P., Vesterinen, M., Korhonen, H., Keskinen, H., Romakkaniemi, S.,... Lehtinen, K. E. J. (2014). The role of low volatile organics on secondary organic aerosol formation. Atmospheric Chemistry and Physics, 14(3), doi: /acp

15 Imholt 15 Kroll, J. H., & Seinfeld, J. H. (2008). Chemistry of secondary organic aerosol: Formation and evolution of low-volatility organics in the atmosphere. Atmospheric Environment, 42(16), doi: /j.atmosenv Lambe, A. T., Cappa, C. D., Massoli, P., Onasch, T. B., Forestieri, S. D., Martin, A. T.,... Davidovits, P. (2013). Relationship between oxidation level and optical properties of secondary organic aerosol. Environmental Science & Technology, 47(12), doi: /es401043j Robinson, A. L., Donahue, N. M., Shrivastava, M. K., Weitkamp, E. A., Sage, A. M., Grieshop, A. P., et al. (2007). Rethinking Organic Aerosols: Semivolatile Emissions and Photochemical Aging.Science, 315(5816), Shilling, J. E., Chen, Q., King, S. M., Rosenoern, T., Kroll, J. H., Worsnop, D. R.,... Martin, S. T. (2009). Loading-dependent elemental composition of alpha-pinene SOA particles. Atmospheric Chemistry and Physics, 9(3), Tsigaridis, K., & Kanakidou, M. (2007). Secondary organic aerosol importance in the future atmosphere. Atmospheric Environment,41(22), doi: /j.atmosenv Xu, L., Kollman, M. S., Song, C., Shilling, J. E., & Ng, N. L. (2014). Effects of NOx on the volatility of secondary organic aerosol from isoprene photooxidation. Environmental Science & Technology, 48(4), doi: /es