Insect Infestation Increases Viscosity of Biogenic Secondary Organic Aerosol

Plant stress alters emissions of volatile organic compounds. However, little is known about how this could influence climate-relevant properties of secondary organic aerosol (SOA), particularly from complex mixtures such as real plant emissions. In this study, the chemical composition and viscosity were examined for SOA generated from real healthy and aphid-stressed Canary Island pine (Pinus canariensis) trees, which are commonly used for landscaping in Southern California. Healthy Canary Island pine (HCIP) and stressed Canary Island pine (SCIP) aerosols were generated in a 5 m3 environmental chamber at 35–84% relative humidity and room temperature via OH-initiated oxidation. Viscosities of the collected particles were measured using an offline poke-flow method, after conditioning the particles in a humidified air flow. SCIP particles were consistently more viscous than HCIP particles. The largest differences in particle viscosity were observed in particles conditioned at 50% relative humidity where the viscosity of SCIP particles was an order of magnitude larger than that of HCIP particles. The increased viscosity for the aphid-stressed pine tree SOA was attributed to the increased fraction of sesquiterpenes in the emission profile. The real pine SOA particles, both healthy and aphid-stressed, were more viscous than α-pinene SOA particles, demonstrating the limitation of using a single monoterpene as a model compound to predict the physicochemical properties of real biogenic SOA. However, synthetic mixtures composed of only a few major compounds present in emissions (<10 compounds) can reproduce the viscosities of SOA observed from the more complex real plant emissions.


Figure S2
. Example particle wall loss correction (kw = 0.0028 s -1 ) applied to healthy plant SOA trial. Correction was only applied during the SOA formation phase, and not during the SOA collection phase when the particle loss became faster. Note that the wall-loss rate constant may depend on RH in the chamber, but the SOA yields reported in Table 1 do not take this dependence into account. Ideally, the wall loss rates should be measured for each experiment but doing this would mean losing a lot of SOA mass before it is collected. Therefore, we relied on previously measured wall loss rates for ammonium sulfate particles at 50% RH. Figure S3. Particle formation as a function of time determined from SMPS data for SOA generated from HCIP1. Figure S4. a) picture of aphid-infested tree (SCIP4). b) Sample of aphid and aphid exuviae (exoskeleton) collected from pine needles onto a petri dish. C) Close up of collected aphid, identified as a light green pine needle aphid (Eulachnus brevipilosus).
Section S1: SOA yield calculation The SOA yield was calculated for each trial (CIP1-5) using the equation below.
The change in SOA mass concentration ( ) was determined from the SMPS data and was corrected for particle wall losses ( Figure S2). The SOA density in the SMPS was assumed to be 1.2 g cm -3 based on previous literature for biogenic SOA. [1][2][3] The change in VOC mass concentration ( ) was determined using a combination of PTR-ToF-MS data and TD-GC-MS data for the summed total mass concentration difference for monoterpene, oxygenated monoterpene, and sesquiterpene. The percent of reacted VOC was determined using PTR-ToF-MS data for total monoterpenes (m/z 137) and sesquiterpenes (m/z 205). As changes in oxygenated monoterpenes (OMT) could not be quantified, a range of reactivity was assumed for the OMT compounds. The low range for the SOA mass yield was determined assuming none of the OMT identified in the TD-GC-MS reacted away as a result of photooxidation whereas the high range for the SOA mass yield assumes that all (100%) of the OMT reacted away over the course of oxidation.
To convert percent reacted monoterpene, oxygenated monoterpene and sesquiterpenes into mass concentration to be used in Equation S.1, the percent reacted monoterpene, oxygenated monoterpene, and sesquiterpene was multiplied by the total mass concentration of these individual terpene categories based on the TD-GC-MS data.
SOA yield results are reported in Table 1 in the main text. SICP4 had the highest yield followed by HCIP3, HCIP2, HCIP1, and SCIP5. The yield of α-pinene photooxidation was previously reported as 26.7 ± 2.5 % for SOA by generation through photooxidation with no seeds with a comparable mass concentration of SOA of 66.8 ± 6.0 μg m -3 at low RH. 4 Most of the trials in this study had SOA yields >30%, meaning that α-pinene alone cannot replicate the higher SOA yield from real pine trees. Trials with high SOA yield correlated well with trees that had higher contribution of oxygenated monoterpene in the initial VOC profile in the chamber (Table S3), with SCIP4 having the largest followed by HCIP3, HCIP1, HCIP2, and HCIP5. However, it is worth noting that due to our assumptions with calculating the change in VOC mass concentration, it is possible there are some errors in our estimations of VOC yield and further investigation of how oxygenated monoterpene influences SOA yield from mixtures of biogenic volatiles is recommended. The high SOA yield for SCIP4 could also be attributed to increased fraction of germacrene D which is expected to have a higher yield due to more double bonds within the structure that are capable of oxidizing and forming lower volatility products. The yield for SCIP5 is not as high because the main stress SQT in   b This upper limit is consistent with surface tension measurements of SOA at RH ≲65% RH and surface tensions reported for alcohols, organic acids, esters, and ketones, as well as surface tension measurements of water solutions containing SOA products. [8][9][10][11] c Range based on measurements of the slip length of organic compounds and water on hydrophobic surfaces. [12][13][14][15][16][17][18][19][20][21][22][23][24] d Contact angles determined by measuring the height and radii of individual droplets using a confocal microscope following the method of Ref. 25 . Note: the simulated viscosities depend only weakly on the contact angle. Figure S6: The dependence of viscosity on conditioning time for all SOA types. A poke-flow experiment was performed at various conditioning times during which particles were exposed to an air flow up to 24 h at (a) 0%, (b) 25%, (c) 50%, and (d) 60% RH to ensure that the particles were at equilibrium. Black points correspond to those of healthy plant SOA and red points correspond to stressed plant SOA. Upward arrows indicate lower limits to viscosity. The area of a SOA particle was monitored for each SOA type over the course of up to 27 h. The particles were exposed to an air flow at (a) ≈ 0% RH and (b) 60% RH within the flow cell, and their area was tracked with the microscope. Error bars correspond to the standard deviation of repeated measurements of the area of the particle at a given time point. Black points correspond to those of healthy plant SOA and red points correspond to stressed plant SOA.    26