UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Viscosity of secondary organic material and related atmospheric implications Grayson, James W. 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_may_grayson_james.pdf [ 7.58MB ]
Metadata
JSON: 24-1.0300661.json
JSON-LD: 24-1.0300661-ld.json
RDF/XML (Pretty): 24-1.0300661-rdf.xml
RDF/JSON: 24-1.0300661-rdf.json
Turtle: 24-1.0300661-turtle.txt
N-Triples: 24-1.0300661-rdf-ntriples.txt
Original Record: 24-1.0300661-source.json
Full Text
24-1.0300661-fulltext.txt
Citation
24-1.0300661.ris

Full Text

Viscosity Of Secondary OrganicMaterialAnd Related Atmospheric ImplicationsbyJames W GraysonM.Chem., The University of York, 2010A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinThe Faculty of Graduate and Postgraduate Studies(Chemistry)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)April 2016c© James W Grayson 2016AbstractAerosols are ubiquitous throughout the Earth’s atmosphere, and secondaryorganic material (SOM), which is produced from the oxidation of volatileorganic compounds, is estimated to constitute a significant fraction of at-mospheric aerosol mass. Furthermore, particles containing SOM can causenegative health outcomes, and affect Earth’s climate, both directly by scat-tering solar radiation, and indirectly by acting as nuclei for cloud droplets.Despite the importance of particles containing SOM, their physical prop-erties, such as viscosity, are poorly constrained. To address this knowledgedeficit, a technique to measure the viscosity of small samples of material,similar to that produced in environmental simulation chambers, was devel-oped and validated. This technique was then used to measure the viscosityof SOM produced via the ozonolysis of α-pinene in an environmental simu-lation chamber. The viscosity of this material was found to depend stronglyon the relative humidity (RH) used when measuring viscosity and the con-centration of SOM mass at which the SOM was produced. A differencebetween the viscosity of the water-soluble component of SOM and the totalSOM (water-soluble and water-insoluble components) was also observed.The viscosity of saccharides and a tetraol were subsequently measured,with these compounds serving as proxies of highly oxidized components ofSOM found in the atmosphere. For saccharides, viscosity was determinedto increase by at least four orders of magnitude as molar mass doubled. Inaddition, the tetraol was determined to have a viscosity at least two orders ofmagnitude lower than that of SOM produced via the oxidation of isoprene,in which the tetraol has been identified.Finally, literature viscosity data for organic compounds was used todemonstrate that saturation vapour concentration, the mass based equiv-iiAbstractalent of saturation vapour pressure, is a useful parameter for predictingviscosity, and better than elemental oxygen-to-carbon ratio or molar mass,at least for organic compounds containing only one or two functional groups.The results presented in this dissertation increase our knowledge of theviscosity of SOM, and its dependence on RH, the SOM mass concentrationat which the SOM is produced, number of hydroxyl functional groups in theorganic molecule, and molar mass.iiiPrefaceChapters 3-5 are co-authored peer-reviewed journal articles and Chapters 6and 7 are co-authored work in preparation for submission as peer-reviewedjournal articles. The details of my contributions to each research chapterare provided below.Chapter 3 (second author on a published journal article): Renbaum-Wolff, L., Grayson, J. W., Bateman, A. P., Kuwata, M., Sellier, M., Mur-ray, B. J., Shilling, J. E., Martin, S. T., and Bertram, A. K.: Viscos-ity of α-pinene secondary organic material and implications for particlegrowth and reactivity, Proc. Nat. Acad. Sci. USA, 110 (20), 8014-8019,doi:10.1073/pnas.1219548110, 2013• Assisted with research formulation and design.• Performed measurements using the poke-and-flow technique.• Assisted in data analysis.• Assisted in preparation of the manuscript.• Additional contributions from co-authors.• L Renbaum-Wolff., B. J. Murray, J. E. Shilling, S. T. Martin,and A. K. Bertram also designed research.• L. Renbaum-Wolff performed measurements using the bead-mobility technique.• A. P. Bateman, M. Kuwata, and J. E. Shilling produced andcollected samples.• L. Renbaum-Wolff and A. K. Bertram designed models usingCOMSOL.ivPreface• L. Renbaum-Wolff, M. Sellier, and A. K. Bertram also analyzeddata.• L. Renbaum-Wolff produced all figures in the manuscript.• L. Renbaum-Wolff, S. T. Martin, and A. K. Bertram also pre-pared the manuscript.Chapter 4 (co-first author on a published journal article): Grayson, J.W., Song, M., Sellier, M., and Bertram, A. K.: Validation of the poke-and-flow technique combined with simulations of fluid flow for determiningviscosities in samples with small volumes and high viscosities, Atmos. Meas.Tech., 8, 2463-2472, doi:10.5194/amtd-8-877-2015, 2015• Designed and formulated research aims with A. K. Bertram and M.Song.• Performed measurements on sucrose-water particles using the poke-and-flow technique.• Took photographs to determine the contact angle of particles com-prised of standard solutions.• Performed data analysis.• Prepared the manuscript with M. Song and A. K. Bertram.• Produced all figures in the manuscript.• Additional contributions from co-authors.• M. Song performed experiments on sucrose-water particles andparticles comprised of standard solutions using the poke-and-flow tech-nique.• M. Song also analyzed data.• M. Song supplied some images for Figure 4.1, and all images forFigure 4.3.vPrefaceChapter 5 (first author on a journal article under peer review): Grayson,J. W., Zhang, Y., Mutzel, A., Renbaum-Wolff, L., Bo¨ge, O., Kamal, S.,Herrmann, H., Martin, S. T., Bertram, A. K.: Effect of varying experimentalconditions on the viscosity of α-pinene derived secondary organic material,Atmos. Chem. Phys. Disc., 15 (22), 32967-33002, doi:10.5194/acpd-15-32967-2015, 2015• Designed and formulated research aims with A. K. Bertram and L.Renbaum-Wolff.• Assisted with sample production and collection.• Performed measurements using the poke-and-flow technique and mea-sured particle-substrate contact angles.• Analysed data.• Prepared manuscript, including all of the figures in the manuscript.• Additional contributions from co-authors.• Y. Zhang, A. Mutzel, and O. Bo¨ge produced and collected sam-ples.• S. Kamal assisted with contact angle measurements.• Y. Zhang, A. Mutzel, L. Renbaum-Wolff, S. Kamal, S. T. Martin,and A. K. Bertram also prepared the manuscript.Chapter 6• Designed and formulated research aims with A. K. Bertram.• Performed measurements of the tetraol compound using the bead-mobility technique.• Took photographs to determine the contact angle of particles com-prised of saccharide-water solutions.• Analysed data.viPreface• Prepared the writing and all figures in the chapter.• Additional contributions from co-authors.• M. A. Upshur synthesised the tetraol compound studied, andprovided a description of the synthesis and the tests of purity andstability.• M. Song performed experiments on saccharide-water particlesusing the poke-and-flow technique.• M. Song supplied the images for Figure 6.1.• M. Song also analyzed data.• A. K. Bertram also helped to write the chapter.Chapter 7• Designed and formulated research aims with A. K. Bertram.• Performed literature research.• Prepared the writing and all figures in the chapter.• Additional contributions from co-authors.• A. K. Bertram also helped to write the chapter.viiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviList of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxiList of Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxxiiList of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . .xxxiiiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . .xxxivDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxxvi1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Atmospheric aerosols . . . . . . . . . . . . . . . . . . . . . . 11.1.1 Formation, sources, and classification of aerosols . . . 11.1.2 Effects of aerosols . . . . . . . . . . . . . . . . . . . . 31.2 Formation of secondary organic material . . . . . . . . . . . 51.3 Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.3.1 What is viscosity? . . . . . . . . . . . . . . . . . . . . 71.3.2 Types of fluids . . . . . . . . . . . . . . . . . . . . . . 91.3.3 Commercially available methods to measure viscosity 11viiiTable of Contents1.3.4 Potential viscosities of SOM . . . . . . . . . . . . . . 131.3.5 The importance of the viscosity of SOM . . . . . . . 141.4 Previous research related to the viscosity of SOM . . . . . . 161.5 Overview of dissertation . . . . . . . . . . . . . . . . . . . . . 162 Experimental techniques . . . . . . . . . . . . . . . . . . . . . 182.1 Bead-mobility technique . . . . . . . . . . . . . . . . . . . . 182.2 Poke-and-flow technique combined with simulations . . . . . 232.2.1 Poke-and flow technique . . . . . . . . . . . . . . . . 232.2.2 Simulations using COMSOL Multiphysics . . . . . . . 272.2.3 Simulations of particles exhibiting flow . . . . . . . . 292.2.4 Simulations of particles that cracked when poked . . 322.3 Equilibration times . . . . . . . . . . . . . . . . . . . . . . . 353 Viscosity of water-soluble α-pinene derived SOM . . . . . 373.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 373.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . 393.2.1 Production of secondary organic material . . . . . . . 393.2.2 Bead-mobility technique . . . . . . . . . . . . . . . . 403.2.3 Poke-and-flow technique . . . . . . . . . . . . . . . . 403.2.4 Simulations of material flow at 40-70 % RH . . . . . 403.2.5 Simulations of material flow at 25-30 % RH . . . . . 433.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433.3.1 Experiments with the bead-mobility technique . . . . 433.3.2 Experiments with the poke-and-flow technique . . . . 443.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524 Additional validation of the poke-and-flow technique . . . 544.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 544.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . 564.2.1 Poke-and-flow technique . . . . . . . . . . . . . . . . 564.2.2 Simulations of fluid flow . . . . . . . . . . . . . . . . 574.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . 60ixTable of Contents4.3.1 Sucrose-water particles . . . . . . . . . . . . . . . . . 604.3.2 Particles of polybutene standards . . . . . . . . . . . 634.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655 Viscosity of α-pinene derived SOM . . . . . . . . . . . . . . . 675.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 675.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . 685.2.1 Generation of SOM in an environmental chamber . . 685.2.2 Generation of SOM in a flow tube . . . . . . . . . . . 695.2.3 Poke-and-flow technique . . . . . . . . . . . . . . . . 715.2.4 Simulations of fluid flow . . . . . . . . . . . . . . . . 725.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . 745.3.1 Effect on viscosity of RH . . . . . . . . . . . . . . . . 745.3.2 Effect on viscosity of production mass concentration . 775.3.3 Effect on viscosity of water-insoluble SOM . . . . . . 845.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876 Viscosity of a tetraol and saccharide-water mixtures . . . 896.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 896.2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . 916.2.1 Measurements of viscosity . . . . . . . . . . . . . . . 916.2.2 Predictions of viscosity using QSPR models . . . . . 946.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . 956.3.1 Measured viscosities of saccharides at a range of RHs 956.3.2 Measured viscosity of a tetraol . . . . . . . . . . . . . 996.3.3 Predicted viscosities of a tetraol . . . . . . . . . . . . 1016.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027 The relationship between viscosity and physical properties 1047.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 1047.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1077.2.1 Selection of compounds . . . . . . . . . . . . . . . . . 1077.2.2 Parameterisation of data . . . . . . . . . . . . . . . . 1087.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109xTable of Contents7.3.1 The relationship between viscosity and O:C . . . . . 1097.3.2 The relationship between viscosity and C* . . . . . . 1117.3.3 Relationship between viscosity and both O:C and C* 1117.4 The predicted viscosity of products of α-pinene ozonolysis . 1137.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1158 Conclusions and future work . . . . . . . . . . . . . . . . . . . 1178.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1178.1.1 Poke-and-flow technique: development and validation 1178.1.2 The viscosity of SOM and related compounds . . . . 1188.1.3 Predicting the viscosity of SOM . . . . . . . . . . . . 1198.2 Directions for future work . . . . . . . . . . . . . . . . . . . . 120Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122AppendicesA Appendix to chapter 3 . . . . . . . . . . . . . . . . . . . . . . 152A.1 Composition of SOM from ozonolysis of α-pinene . . . . . . 152A.2 Viscosity prediction using mixing rules . . . . . . . . . . . . 152B Appendix to chapter 5 . . . . . . . . . . . . . . . . . . . . . . 154B.1 Effect of carrier gas flow on SOM particle properties . . . . . 154B.2 Particle-to-particle and sample-to-sample variability . . . . . 155B.3 Calculated viscosity for prior studies of α-pinene derived SOM 155B.4 Simulations of fluid flow for particles that exhibit cracking . 157B.5 Tables and figures . . . . . . . . . . . . . . . . . . . . . . . . 158C Appendix to chapter 7 . . . . . . . . . . . . . . . . . . . . . . 165C.1 Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165xiList of Tables2.1 Summary of literature values of measured slip lengths for wa-ter and atmospherically relevant organic compounds. . . . . . 313.1 Mean bead speeds as a function of RH in water-soluble SOMfrom α-pinene ozonolysis and corresponding viscosities. Orig-inally published in Renbaum-Wolff et al. (2013a). . . . . . . . 443.2 Results from the poke-and-flow experiments. Originally pub-lished in Renbaum-Wolff et al. (2013a). . . . . . . . . . . . . 454.1 Experimental parameters used when simulating flow with COM-SOL for the sucrose-water experiments. Originally publishedin Grayson et al. (2015a). . . . . . . . . . . . . . . . . . . . . 584.2 Experimental parameters used when simulating flow with COM-SOL for experiments using polybutene standards. Originallypublished in Grayson et al. (2015a). . . . . . . . . . . . . . . 595.1 Conditions used for generating and collecting samples of SOMgenerated via the ozonolysis of α-pinene. The whole SOM(both water soluble and water insoluble component of theSOM) was collected. Originally published in Grayson et al.(2015b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 705.2 Physical parameters used when simulating particles that ex-hibited flow with COMSOL. Originally published in Graysonet al. (2015b). . . . . . . . . . . . . . . . . . . . . . . . . . . . 73xiiList of Tables5.3 Summary of τ exp,flow times and viscosities of whole SOM andwater-soluble SOM produced in the flow tube at a productionmass concentration of 14,000 µg m−3 and studied at <0.5 %RH. Originally published in Grayson et al. (2015b). . . . . . . 866.1 Properties of saccharide and tetraol compounds studied ex-perimentally. . . . . . . . . . . . . . . . . . . . . . . . . . . . 906.2 Physical parameters used to simulate the flow of material dur-ing poke-and-flow experiments where a half-torus geometrywas formed and the material subsequently observed to flow.R and r represent dimensions of a half-torus, with R repre-senting the radius of the tube of material, and r representingthe radius of the hole at the centre of the tube. . . . . . . . . 926.3 Physical parameters used to simulate a lower limit of viscosityfor poke-and-flow experiments where particles cracked whenimpacted by the needle, and no observable flow of materialwas observed over the course of the experiment. . . . . . . . . 936.4 Summary of experimental viscosity measurements using thebead-mobility and poke-and-flow techniques for the tetraoland the saccharide-water particles studied here, with resultsfrom individual particles grouped by RH. For experimentsusing the bead-mobility technique the mean is reported alongwith the 95 % confidence intervals. For experiments using thepoke-and-flow technique lower and upper limits of viscosityare reported, taking account of the 95 % confidence limitsof the simulated lower and upper limits of viscosity for thegroup of particles studied at each RH. N/A is reported for allexperiments performed using the poke-and-flow technique, forwhich no mean viscosity is calculated, and for upper limits ofviscosity for experiments with the poke-and-flow techniquewhere the particle cracked, as only a lower limit of viscositycan be calculated. . . . . . . . . . . . . . . . . . . . . . . . . . 97xiiiList of Tables7.1 A summary of the functional groups featured in the 156 com-pounds used to determine the relationship between viscosityand either O:C or C*. A compound with multiple oxygencontaining functional groups is counted once for each type offunctional group it contains, and once in the multiple oxygencontaining functional group row. As such, the total sums to>156. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1087.2 Summary of the parameterisations used in this study. In eachcase log(η / Pa s) is used as the dependent variable, and O:Cand log(C* / µg m-3) are the independent variables. RMSErefers to the root mean square error between the log(η / Pa s)predicted by the parameterisation for each of the compoundsin the study and the experimental viscosities. . . . . . . . . . 109B.1 Summary of τ exp,flow times and viscosities of sample analysedafter both 1 hour and 45 hours of exposure to a dry (<0.5 %RH) flow of Nitrogen gas. Originally published in Graysonet al. (2015b). . . . . . . . . . . . . . . . . . . . . . . . . . . . 158B.2 Experimentally determined contact angles for each of thesamples studied. The range of values represent the 95 % con-fidence intervals of the values measured for multiple particles.Originally published in Grayson et al. (2015b). . . . . . . . . 158xivList of TablesB.3 Summary of the percent relative standard deviation (% RSD)in τ exp,flow, lower limits of viscosity, and upper limits of vis-cosity for particles produced using equivalent conditions andstudied via the poke-and-flow technique in combination withsimulations of fluid flow at <0.5 % RH. Three samples werestudied per production mass concentration in the flow tube.Values prior to parentheses represent the relative standarddeviation between all particles studied that were producedat a given mass concentration, whilst the values inside eachparenthesis represent the average relative standard deviationbetween particles on the same substrate. Originally publishedin Grayson et al. (2015b). . . . . . . . . . . . . . . . . . . . . 159B.4 Summary of parameters used to estimate viscosity from liter-ature studies of SOM produced via the ozonolysis of α-pinene.Originally published in Grayson et al. (2015b). . . . . . . . . 160B.5 Physical parameters when simulating particles that don’t ex-hibit flow in COMSOL. Originally published in Grayson et al.(2015b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161C.1 A list of compounds included in the study, along with selectedchemical and physical properties. . . . . . . . . . . . . . . . . 165xvList of Figures1.1 A summary of radiative forcing by atmospheric aerosols andother atmospheric constituents between 1750 and 2011. Un-certainties represent the 5-95 % confidence range, with soliderror bars representing the uncertainty in the magnitude ofeffective radiative forcing and dotted bars representing un-certainty in the magnitude of radiative forcing. Also shownis the level of confidence for each source of forcing, based onthe current level of scientific knowledge. Figure adapted fromFigure TS.6 of Stocker et al. (2013). . . . . . . . . . . . . . . 51.2 Chemical structures of isoprene (left) and (+)-α-pinene (right). 61.3 Laminar flow of a fluid between two 2D boundary plates - onestationary and one moving. . . . . . . . . . . . . . . . . . . . 81.4 The relationship between shear rate and shear stress for typesof fluids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101.5 Schematic of two examples of commercially available viscome-ters. Shown in (a) is an Ostwald viscometer. Sample is addedto side 1 until the sample in the capillary is level with lineA. Suction is subsequently applied at point 2, and the timetaken for the sample to flow between line B and line C ismeasured. Shown in (b) is a coaxial-cylinder viscometer. Theouter cylinder is rotated at a constant speed, and the angu-lar deflection of the inner cylinder is used to determine theviscosity of the sample. . . . . . . . . . . . . . . . . . . . . . . 12xviList of Figures1.6 (a) A scale showing part of the viscosity continuum, detailingthe regions typically defined to be bound by the solid, semi-solid, and liquid regimes, as well as the viscosities of somecommon substances (idea to show common substances is bor-rowed from Koop et al., 2011). The image of the pitch is a de-tail of a picture taken from Wikimedia Commons of the pitchdrop experiment Wikipedia page (GFDL, John Mainstone,University of Queensland, Australia). (b) A scale showingthe ranges of viscosity measurable using commercially avail-able instrumentation for measuring viscosity, as well as thepossible range of viscosities spanned by SOM. . . . . . . . . . 131.7 (a) Effect of particle viscosity on the mechanism of growthof SOM by semivolatile organic compound (SVOC) uptake.(b) Effect of particle viscosity on heterogeneous oxidation byozone. (c) Climate effects of particles and implications of highparticle viscosities on particle growth rates, particle mass, andheterogeneous oxidation by O3. The implications of particleviscosity on growth and heterogeneous oxidation in (a)-(c)assume a monodisperse particle population. Originally pub-lished in Renbaum-Wolff et al. (2013a). . . . . . . . . . . . . 152.1 Schematic representation of bead-mobility experimental setup.Originally published in Renbaum-Wolff et al. (2013b). . . . . 202.2 Illustration of the flow of the flow gas around a sample parti-cle. Originally published in Renbaum-Wolff et al. (2013b). . . 202.3 Internal circulation of beads within a particle of glycerol.The red lines overlaid on the image show the 2-D circula-tion of three beads within the particle. Originally publishedin Renbaum-Wolff et al. (2013b). . . . . . . . . . . . . . . . . 20xviiList of Figures2.4 Images from the bead-mobility studies. Images in (a) corre-spond to images recorded at 90 % RH and images in (b) corre-spond to images recorded at 70 % RH. Indicated in the figureare different beads monitored in the experiments and the x-ycoordinates of the beads. Originally published in Renbaum-Wolff et al. (2013a). . . . . . . . . . . . . . . . . . . . . . . . 212.5 Average bead speed (±1 σ) vs. viscosity for standard com-pounds. Originally published in Renbaum-Wolff et al. (2013b). 222.6 Schematic representation of poke-and-flow experimental setup.Originally published in Grayson et al. (2015a). . . . . . . . . 242.7 Deformation and recovery across time of poked SOM parti-cles. Prior to poking (pre-poke), the particle morphology canbe approximately described as a spherical cap. At higher RH(40-70 %) (rows a-c), geometries approximately described ashalf-torus are formed, and flow occurs. By comparison, forlow RH (≤30%) (row d) particles shatter and do not flowover a period of 8 h. The ring structures observed in the firstand last columns are an optical effect that arises from hemi-spherical diffraction. Originally published in Renbaum-Wolffet al. (2013a). . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.8 Details of half-torus model used to simulate the flow in exper-iments: (a) top view, where R and r are the notations usedhere to describe the dimensions of a half-torus geometry; (b)side view, where surface 1 represents the air-fluid interface,and surface 2 represents the fluid-substrate interface. Origi-nally published in Grayson et al. (2015a). . . . . . . . . . . . 30xviiiList of Figures2.9 The dependence of viscosity as (a) surface tension, (b) sliplength, (c) density, (d) contact angle, and (e) τmodel,flow, arevaried across a wide range of values for particles of dimensionsR0 = 16, r0 = 4 (filled symbols) and R0 = 14, r0 = 6 (opensymbols). For each simulation four of the five properties wereheld constant whilst the fifth was varied. Black squares: sur-face tension = 75.15 mN m−1, slip length = 10 µm, density= 1500 kg m−3, contact angle = 98.5 o, τmodel,flow = 1000seconds. Red circles: surface tension = 57.2 mN m−1, sliplength = 5 nm, density = 1500 kg m−3, contact angle = 98.5o, τmodel,flow = 0.1 seconds. . . . . . . . . . . . . . . . . . . . 322.10 Details of quarter-sphere model used to simulate flow for par-ticles that exhibit cracking behaviour and no discernible flowover subsequent hours of observation. Surfaces 1 and 2 rep-resent the external fluid interfaces, to which a relevant valueof surface tension from literature was assigned, and surface 3represents the particle-substrate interface, which was repre-sented as a Navier-slip boundary with a slip length of 10-17nm. Originally published in Renbaum-Wolff et al. (2013a). . . 333.1 Schematic detailing the production of SOM via the ozonolysisof α-pinene. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393.2 Calibration line from COMSOL simulations (black solid line).In the simulations used to generate this calibration a sur-face tension of 75 mN m−1 and contact angle of 90 ◦ wereused. The annotation in the figure illustrates how the cali-bration line would shift if a smaller surface tension, smallerequilibrium contact angle, smaller slip length, and/or surfacescratching was included in the model. An example line con-structed with lower surface tension (60 mN m−1) and lowercontact angle (70 ◦) is given to display the sensitivity of themodel to these parameters (grey dashed line). Originally pub-lished in Renbaum-Wolff et al. (2013a). . . . . . . . . . . . . 41xixList of Figures3.3 (a) Summary plot of the SOM viscosities determined by acombination of experiments using the bead-mobility technique(black empty squares and triangles for HEC and PNNL sam-ples, respectively, where the black bars represent the 95 %prediction intervals) and experiments using the poke-and-flowtechnique (where the blue bars represent the bounds of theviscosities). HEC refers to samples collected on quartz fiberfilters from the Harvard Environmental Chamber. PNNLrefers to samples collected on Teflon filters from the PacificNorthwest National Laboratory Continuous-Flow Environmen-tal Chamber. Various common substances have been placedalongside the diagram, along with their approximate viscosi-ties at room temperature, to provide points of reference fol-lowing the idea of Koop et al. (2011). The secondary y-axesshow (1) diffusion coefficients calculated using the Stokes-Einstein relation and (2) mixing times (τmixing) of the par-ticles due to bulk diffusion in 100 nm particles of the sameviscosity (see main text). The image of the pitch is a de-tail of an image from the pitch drop experiment (Wikime-dia Commons, GFDL, University of Queensland, Australia,John Mainstone). (b) Typical relative humidities observedin the planetary boundary layer (Hamed et al., 2011; Heldand Soden, 2000; Martin, 2000) and environmental chambers(Kostenidou et al., 2009; Tillmann et al., 2010). Originallypublished in Renbaum-Wolff et al. (2013a). . . . . . . . . . . 46xxList of Figures4.1 Optical images of sucrose-water particles poked at RHs of (a)48.8, (b) 52.7, and (c) 58.8 % recorded during typical poke-and-flow experiments. Images a1, b1, and c1 correspond tothe particles before they are poked. Images a2, b2, and c2 cor-respond to the first frame post-poke (i.e. the first frame afterthe needle has been removed). Images a3, b3, and c3 corre-spond to images of the experimental flow time, τ exp,flow, thepoint at which the equivalent area diameter of the hole at thecentre of the particle has decreased to 50 % of its original size.Images a4, b4, and c4 correspond to the final frame recorded,at which point each particle has re-attained its original spher-ical cap geometry. Scale bar: 20 µm. Originally published inGrayson et al. (2015a). . . . . . . . . . . . . . . . . . . . . . . 604.2 (a) τ exp,flow as a function of RH for individual sucrose-waterparticles. (b) Calculated viscosities for the individual sucrose-water particles in (a), where red bars represent the calculatedlower and upper limits of viscosity. (c) Lower and upper lim-its of viscosity for the particles shown in (b), grouped by RH.The error bars on the x -axis represent the range of RHs atwhich particles in the group were poked. Lower and upperlimits of viscosity were determined for each particle via sim-ulation, with the bottom of a bar on the y-axis representingthe lowest lower limit of viscosity for any of the particles inthe group, and the top of the bar representing the highestupper limit of viscosity for any of the particles in the group.Literature values (Power et al., 2013; Quintas et al., 2006;Swindells et al., 1958) are provided for comparison, with er-ror bars representing 1σ for Power et al. (2013) and 95 %confidence intervals for Quintas et al. (2006). Originally pub-lished in Grayson et al. (2015a). . . . . . . . . . . . . . . . . . 62xxiList of Figures4.3 Optical images of particles of polybutene standards (a) Stan-dard #1 (N450000), and (b) Standard #2 (N2700000), beingpoked at 0 % RH recorded during typical poke-and-flow ex-periments. Images a1 and b1 correspond to particles prior topoking. Images a2 and b2 correspond to the first frame post-poke (i.e. the first frame after the needle has been removed).Images a3 and b3 correspond to images of the experimentalflow time, τ exp,flow, the point at which the equivalent areadiameter of the hole at the centre of the torus has decreasedto 50 % of its original size. Images a4 and b4 correspondto the final frame recorded of each particle, at which pointeach particle has re-attained its original spherical cap geome-try. Scale bar: 20 µm. Originally published in Grayson et al.(2015a). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634.4 Viscosity as a function of temperature for experiments withthe polybutene standards. Results from standard #1 (N45000)are in black whilst results from Standard #2 (N2700000) arein red. Symbols represent values measured by Cannon Instru-ment Company using a manual capillary viscometer. Barsrepresent viscosities determined herein. For the bar that rep-resents each standard the bottom of the bar represents thelowest lower limit of viscosity of all the particles examined,whilst the top of the bar represents the highest upper limit ofviscosity of all of the particles examined. Originally publishedin Grayson et al. (2015a). . . . . . . . . . . . . . . . . . . . . 64xxiiList of Figures5.1 Optical images recorded during typical poke-and-flow experi-ments of whole SOM produced at a production mass concen-tration of 520 µg m−3 being poked at (a) <0.5 %, and (b) 50%, RH. Images a1 and b1 correspond to SOM prior to poking.Images a2 and b2 correspond to the first frame post-poke (i.e.the first frame after the needle has been removed). Images a3and b3 correspond to images of the experimental flow time,τ exp,flow, the point at which the diameter of the hole at thecentre of the torus has decreased to 50 % of its original size.Scale bar in Images a1 and b1: 20 µm. Originally publishedin Grayson et al. (2015b). . . . . . . . . . . . . . . . . . . . . 755.2 Summary of poke-and-flow experiments from <0.5 % to 50 %RH performed on samples of whole SOM produced at massconcentrations of 520 µg m−3 (Panels (a) and (c)) and 121µg m−3 (Panels (b) and (d)). Panels (a) and (b) show boxplots of observed τ exp,flow as a function of RH. Panels (c) and(d) show simulated lower (filled symbols) and upper (opensymbols) limits of viscosity. Y-error bars represent 95 % con-fidence intervals, and x-error bars represent the range of RHat which measurements were made. The shaded regions areincluded to guide the eye of the reader. The viscosities ofcommon substances at room temperature have been added to(d) to provide points of reference, as per Koop et al., 2011.The image of pitch is part of an image from the pitch dropexperiment (image courtesy of Wikimedia Commons, GNUFree Documentation License, University of Queensland, JohnMainstone). Originally published in Grayson et al. (2015b). . 76xxiiiList of Figures5.3 Optical images recorded during typical poke-and-flow experi-ments of particles of the whole SOM produced at productionmass concentrations of (a) 14,000 µg m−3, (b) 520 µg m−3,and (c) 121 µg m−3 being poked at <0.5 % RH. Images a1,b1 and c1 correspond to SOM prior to poking. Images a2,b2 and c2 correspond to the first frame post-poke (i.e. thefirst frame after the needle has been removed). Images a3, b3and c3 correspond to images of the experimental flow time,τ exp,flow, the point at which the diameter of the hole at thecentre of the torus has decreased to 50 % of its original size.Scale bar in Images a1, b1 and c1: 20 µm. Originally pub-lished in Grayson et al. (2015b). . . . . . . . . . . . . . . . . 785.4 Summary of poke-and-flow experiments performed on sam-ples of whole SOM at <0.5 % RH. Black symbols representresults from particles produced in a flow tube, whilst red sym-bols represent results from particles produced in a chamber.Panel (a) shows box plots of observed τ exp,flow times at differ-ent production mass concentrations for particles poked <0.5% RH. Boxes represent the 25, 50, and 75 percentiles, opencircles represent median values, and whiskers represent the5 and 95 percentiles. Panel (b) shows the simulated lower(filled squares) and upper (open squares) limit of viscosityfor particles at each production mass concentration poked at<0.5 %. Symbols represent mean values. The y error barsrepresent 95 % confidence intervals. The shaded regions areincluded to guide the eye of the reader. Also included in (b)are literature viscosities for SOM produced via the ozonoly-sis of α-pinene (Renbaum-Wolff et al., 2013a; Zhang et al.,2015). Originally published in Grayson et al. (2015b). . . . . 79xxivList of Figures5.5 Summary of poke-and-flow experiments performed on par-ticles of whole SOM at 30 % RH. Black symbols representresults from particles produced in a flow tube, whilst red sym-bols represent results from particles produced in a chamber.Panel (a) shows box plots of observed τ exp,flow times as afunction of SOM mass concentrations for particles studiedusing the poke-and-flow technique at 30 % RH. Panel (b)shows the simulated lower (filled squares) and upper (opensquares) limit of viscosity for particles at each SOM massconcentration studied using the poke-and-flow technique at30 % RH. Symbols represent mean values, whilst the y errorbars represent 95 % confidence intervals. The shaded region isincluded to guide the eye of the reader. Also included in (b)are literature viscosities from Renbaum-Wolff et al. (2013a)and Zhang et al. (2015), for SOM produced via the ozonolysisof α-pinene and studied at 30 % RH. Originally published inGrayson et al. (2015b). . . . . . . . . . . . . . . . . . . . . . . 82xxvList of Figures5.6 Optical images recorded during poke-and-flow experimentsusing particles consisting of (a) the water-soluble componentof the SOM and (b) the whole SOM (i.e., both the water-soluble and the water-insoluble components). In both exper-iments the SOM was produced using a mass concentration of14,000 µg m−3 and was poked at <0.5 % RH. Images a1 andb1 correspond to the SOM prior to being poked. The bright-ness in Image a1 is due to reflection of the source light by theneedle positioned just above the particle. Images a2 and b2correspond to the first frame post-poke (i.e. the first frameafter the needle has been removed). The particle comprisedof the water-soluble component of SOM exhibited crackingbehaviour and, as shown in Image a3, no change in the sizeor shape of the cracks can be observed 14 hours after the par-ticle has been poked. The particle comprised of whole SOMexhibited flow, and Image b3 corresponds to an image of theparticle at its experimental flow time, τ exp,flow, the point atwhich the diameter of the hole at the centre of the torus hasdecreased to 50 % of its original size. Scale bar in Images a1and b1: 20 µm. Originally published in Grayson et al. (2015b). 85xxviList of Figures6.1 Figure 1: Optical images recorded during poke-and-flow ex-periments using particles of (a) maltohexaose and (b) raffi-nose. Images a1 and b1 correspond to the particles prior tobeing poked, with the white haloes being an optical effect.Images a2 and b2 correspond to the first frame after the nee-dle has been remover. The particle composed of maltohexaoseand studied at 50 % RH exhibited cracking behaviour and,as shown in Image a3, no change in the size or shape of thecracks can be observed 3 h after the particle has been poked.The particle comprised of raffinose and studied at 54 % RHexhibited flow, and Image b3 corresponds to an image of theparticle at its experimental flow time, τ exp,flow, the point atwhich the diameter of the hole at the centre of the torus hasdecreased to 50 % of its original size. The scale bar in imagesa1 and b1 corresponds to 20 µm. . . . . . . . . . . . . . . . . 956.2 Plots of log10(viscosity) vs. (a) relative humidity and (b)both molar mass and number of saccharide units for glu-cose, sucrose, raffinose, and maltohexaose. Results deter-mined in the current study using the bead-mobility techniqueare shown using circle symbols, and those determined usingthe poke-and-flow technique are shown using squares, withfilled squares representing upper limits of viscosity and opensquares representing lower limits of viscosity, with y-errorbars representing 95 % confidence intervals for both tech-niques, as detailed for Table 3. Also included are literatureviscosity values for sucrose (Fo¨rst et al., 2002; Power et al.,2013; Quintas et al., 2006; Swindells et al., 1958; Telis et al.,2007) and glucose (Achard et al., 1992; Barbosa-Ca´novas et al.,2007; Haynes, 2015), with the viscosity of glucose at 47 %shown in (b) being determined using a polynomial fit of theliterature data. The viscosity of water is added to (a), andshaded regions are added to (a) and (b) to guide the readerseye. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98xxviiList of Figures6.3 A plot of log(viscosity) vs. number of hydroxyl functionalgroups added to 2-methylbutane. The symbols correspondto 2-methylbutane (square), 1-hydroxy-2-methylbutane (up-ward pointing triangle), and 2-methyl-1,2,3,4-tetraol (leftwardpointing triangle). The solid line is a fit of the data for 2-methylbutane and 1-hydroxy-2methylbutane. . . . . . . . . . 1006.4 Plot of log(viscosity) vs. number of hydroxyl groups addedto C3-C8 alkanes. Symbols correspond to alkanes (squares),monoalcohols (upward pointing triangles), diols (downwardpointing triangles), and triols (diamonds), with lines drawnbetween compounds with the same number of carbons. . . . . 1006.5 Plot of experimental vs. predicted log(viscosity) of C3-C8alkanes, monoalcohols, and polyols for the models derivedby (a) Sastri and Rao (1992) and (b) Marrero-Morejo´n andPardillo-Fontdevila (2000). Dashed 1:1 lines are shown oneach plot to guide the readers eye. . . . . . . . . . . . . . . . 1017.1 Plots detailing the relationship between viscosity (log(η / Pas)) and ((a) and (b)) O:C, and ((c) and (d)) log satura-tion vapour concentration (log(C*)) for a range of compoundscomprised only of carbon and hydrogen or carbon, hydrogen,and oxygen atoms, and containing atmospherically relevantfunctional groups. Lines of best fit are included on plots (a)and (c). 1:1 lines are included on plots (b) and (d) to guidethe eye of the reader. Statistical values from the plots arereported in Table 7.2. . . . . . . . . . . . . . . . . . . . . . . 110xxviiiList of Figures7.2 (a) is a 3-D contour plot with O:C and log10(C* / µg m−3) onthe x - and y-axes, respectively, and the colour scale on theplot indicating the viscosity predicted using an MLR withexperimental log(η / Pa s) as the dependent variable andO:C and log(C* / µg m−3) as the independent variables.(b)is a residual plot showing the difference for each compoundbetween its experimental viscosity and the viscosity predictedby the regression. (c) is a plot of the predicted viscosities foreach of the compounds vs. their experimental viscosities, witha 1:1 line to guide the readers eye. . . . . . . . . . . . . . . . 1127.3 Plot of log(η / Pa s) vs. log(C* / µg m−3) for the com-pounds detailed in Table C.1. The shaded pink region rep-resents the 95 % prediction interval of the linear regressionbetween log(C* / µg m−3) vs. log(η / Pa s). Shown onthe x -axis are the regions of the log(C*) scale attributedto semivolatile organic compounds (SVOCs), intermediatevolatility organic compounds (IVOCs), and volatile organiccompounds (VOCs). Also shown is the position of α-pinene(dark blue circle) (per Donahue et al., 2009), as well as firstgeneration products from α-pinene ozonolysis, cis-pinic acid(green), cis-pinonic acid (red), and pinonaldehyde (cyan) (C*values calculated from Hartonen et al., 2013). . . . . . . . . . 114B.1 A plot of particle volume vs. time for five particles exposedto a dry (<0.5 % RH) N2 gas flow. Dotted lines represent themeasured mean size of a particle. Error bars on the y-axisrepresent the uncertainty in measuring both the area of theparticle, and the equilibrium contact angle, at the particle-substrate interface. Originally published in Grayson et al.(2015b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162xxixList of FiguresB.2 (a) Schematic representation of instrumental setup for con-tact angle images. (b) Fluorescence image obtained of anSOM particle. The green overlay is used to determine the con-tact angle of the particle, in this case 60 0, and was producedusing the LB-ADSA plugin for ImageJ. Originally publishedin Grayson et al. (2015b). . . . . . . . . . . . . . . . . . . . . 163B.3 Plot of production mass concentration vs. viscosity for wholeSOM produced via the ozonolysis of α-pinene and studied at<5 % RH. Shown are the results determined here along withthose previously reported in literature (Abramson et al., 2013;Cappa and Wilson, 2011; Perraud et al., 2012; Renbaum-Wolff et al., 2013a; Robinson et al., 2013; Saleh et al., 2013;Zhang et al., 2015). Originally published in Grayson et al.(2015b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164xxxList of Symbolsβ slip lengthC* saturation vapour concentrationD diffusion coefficientη dynamic viscosityF forcekB Boltzmann constantl grid spacing in COMSOL simulationsm massν kinematic viscosityP vapour pressureR ideal gas constantr particle radiusρ momentumσ densityς activity coefficientT temperaturet timeτ shear stressτ exp,flow experimental flow time in poke-and-flow experimentsτmodel,flowmodeled flow time in simulations of poke-and-flowexperimentsu flow velocityυ velocityV volumexxxiList of Unitsatm atmospherescm centimetrem metresg gramsh hoursJ JoulesK Kelvinmins minutesµg microgramsµL microlitresnm nanometresPa PascalsPa s Pascal secondss secondssLpm standard Litres per minuteTg terragramsxxxiiList of AbbreviationsDRH deliquescence relative humidityELVOCs extremely low volatility organic compoundsIVOCs intermediate volatility organic compoundsMLR multiple linear regressionO:C elemental oxygen-to-carbon ratioPM2.5 particulate matter of diameter <2.5 µmRH relative humiditySOM secondary organic materialSVOCs semivolatile organic compoundsVOCs volatile organic compoundsxxxiiiAcknowledgementsThe work presented in this thesis is as much a reflection of the support ofthe people around me as it is of me.I’d like to begin by thanking my research supervisor, Professor AllanBertram, who I think is an excellent, principled scientist, and a conscientiousperson. Without his knowledge, ingenuity, and thoughtfulness, this body ofwork would be significantly less impressive. I believe I’m a better scientistand person as a result of having Allan as my supervisor. Thank you.I’d also like to extend my thanks to the former and current membersof the Bertram group, who fostered a friendly, helpful and encouraging at-mosphere, and made it easy to come to work: Richard, Donna, Michael,Jason, Yuan, Ryan, Meng, Yuri, Ce´dric, Vickie, Stephen, Mijung, Kaitlin,Adrian, Erin, Jacquie, Matt, Kristina, Maki, Amir, Robin, Pablo, Rachel,Ryan, Kenny. I’d especially like to thank Lindsay Renbaum-Wolff, withwhom I enjoyed working closely for three years and from whom I learned anenormous amount.This research has also been facilitated by a number of people outside ofthe Bertram group. Saeid Kamal has patiently and thoroughly taught methe theory and operation of the optical and fluorescence microscopes withinthe LASIR facilities at both UBC and Simon Fraser University, and hasbeen a joy to work with. The members of the Mechanical and Electronicsshops within the Chemistry department have a vast amount of knowledgeand expertise, which they’ve helpfully applied to assist my research on nu-merous occasions. The departmental administration staff and those in theChemistry Stores have each more than ably facilitated my work.I’ve been fortunate enough to perform experiments to support my re-search at laboratories at two institutions outside UBC, and would like toxxxivAcknowledgementsgive my thanks to both the Martin group at Harvard University, particu-larly Yue Zhang and Scot T. Martin, as well as the members of the TROPOSinstitute at the University of Leipzig, particularly Anke Mutzel, Maria Rodi-gast, Olaf Bo¨ge, and Hartmut Herrmann for their welcoming hospitality andhelpfulness during my visits. It was a pleasure to work with all of you andI wish you well for the future.I’m especially grateful to be surrounded by some wonderful people. Myfriends have been excellent sources of support and humour, and in no par-ticular order I’d like to express appreciation for time spent with Lindsay,Ryan, Ali, Niels, Emi, Scott, Joe, and Andreas, amongst countless others.Mum, Dad, and Mark, thank you for your love, encouragement, and un-derstanding, and for giving me every opportunity to achieve my goals. Thesame is true for my grandparents and extended family. Amy, you’re my bestfriend and I’m honoured to be your partner. This work has been so mucheasier with you by my side and I’m looking forward to sharing many moreadventures with you.Last, but by no means least, I wish to acknowledge the research andwork presented in this thesis was carried out on the traditional, ancestral,and unceded territory of the Musqueam people.xxxvTo my parents, Richard and Helen, my brother, Mark, and my partner, AmyxxxviChapter 1Introduction1.1 Atmospheric aerosols1.1.1 Formation, sources, and classification of aerosolsAtmospheric aerosols are liquid or solid particles suspended in the Earth’satmosphere. The formation of atmospheric aerosols may be primary, viathe direct emission of solid or liquid particulate matter from the Earth’ssurface, or secondary, as a result of gases reacting in situ in the atmosphere,eventually forming products of low volatility, which undergo gas-to-particleconversion (Seinfeld and Pankow, 2003).Aerosols range in diameter from nanometers to hundreds of microns,and are typically classified into four size modes (Seinfeld and Pandis, 2006,p.369). The smallest aerosols, of diameter <10 nm, belong to the nucleationmode, which is comprised of secondary aerosols. The Aitken mode, com-prised of particles ranging in diameter from 10 - 100 nm, is comprised of pri-mary aerosols recently emitted to the atmosphere and secondary aerosols re-cently produced via gas-to-particle conversion. Both nucleation and Aitkenmode aerosols rapidly coagulate to form accumulation mode particles, andhave atmospheric lifetimes on the order of tens of minutes. Ranging in di-ameter from 100 nm - 2.5 µm, accumulation mode particles also includeprimary aerosols emitted directly into the atmosphere, and have an atmo-spheric lifetime on the order of a week. Collectively, nucleation, Aitken,and accumulation mode aerosols are known as fine particles. The largestaerosols, those of diameter >2.5 µm, belong to the coarse mode, whichtypically consists mainly of primary particulate matter including dust fromnatural and anthropogenic sources, and salt particles produced by sea-spray.11.1. Atmospheric aerosolsDue to their size, only a small fraction of coarse-mode particles are entrainedinto the atmosphere and, as such, their lifetime is typically on the order ofhours to a few days.Aerosols are also often classified based on whether their emissions sourceis natural or anthropogenic. Natural sources include sea spray from oceans,mineral dust from deserts, and gaseous emissions from forests, whilst an-thropogenic sources include both industrial and agricultural activities, andthe combustion of fossil fuels. Sea spray and mineral dust combined accountfor the majority of the total mass of aerosols and aerosol precursors emittedto the atmosphere (≈45% and ≈30%, respectively) (Stocker et al., 2013),and all natural sources combined account for >90 % of the mass of aerosolsand aerosol precursors emitted. The single largest source of anthropogenicemissions is dust produced during industrial activities (Seinfeld and Pandis,2006, p.61).Vast spatial and temporal variations are observed in aerosol numberconcentrations, which can range from 102 - 108 cm−3 (Seinfeld and Pan-dis, 2006, p.350), although they typically fall between 103 and 105 cm−3(Pruppacher and Klett, 2010, p.252). Spatially, this dramatic spread is theresult of the short lifetime of Aitken mode particles that dominate aerosolnumber concentrations in the atmosphere. Areas located far from the emis-sion sources of Aitken mode particles, such as remote regions at the Earth’shigh latitudes, have the lowest number concentrations (Ottar, 1989), withurban regions close to industrial activity and transportation networks typi-cally having the highest concentrations (e.g. Hussein et al., 2005). Temporalvariations may be observed diurnally as a result of traffic patterns (e.g. Duet al., 2012), and references therein) and seasonally as a result of agriculturalactivities (e.g. Rastogi et al., 2016; Urban et al., 2016). Huge quantities ofcoarse mode mineral dust particles can be entrained during dust storms, alsoleading to short term temporal variations in particle number concentrations(e.g. Wehner et al., 2004).21.1. Atmospheric aerosols1.1.2 Effects of aerosolsThe presence, composition, and concentration of atmospheric aerosols im-pact the organisms inhabiting the Earth’s surface, and both directly andindirectly affect its atmosphere and climate.Aerosols can have numerous effects on health. For example, aerosolsfrom diesel soot, welding fumes, black carbon, and coal fly ash that reachthe lungs have the potential to cause oxidative stress, inflammation, andcancer (Donaldson et al., 2005), whilst prenatal exposure to polyaromatichydrocarbons present in fossil fuels and released during incomplete com-bustion of wood, coal, and diesel, has been linked to both lower IQ andasthma in infants, as well as an increase in the incidence of childhood anx-iety and depression (Perera et al., 2009, 2012). Farm workers have alsoreported coughing, wheezing, and breathlessness during grain harvesting,when concentrations of fungal spores are elevated (Darke et al., 1976). Fur-ther, exposure to secondary organic aerosols has been observed to result inan increase in the time needed to repair mammalian epithelial cells (Gaschenet al., 2010), as well as an increase in premature cell death (Baltenspergeret al., 2008).Aerosols affect the atmosphere directly through interactions with incom-ing solar radiation, and indirectly by acting as cloud condensation nuclei(Seinfeld and Pandis, 2006, p.1055). The direction and magnitude of theseeffects is quantified using radiative forcing. Radiative forcing, which is re-ported in Watts per metre squared (W m−2) and typically quantified at thetropopause, refers to a change in energy per unit area of the Earth due toa change in the concentration of a climate forcing agent. Effective radia-tive forcing is also used. Effective radiative forcing is similar to radiativeforcing, however, it determines the change in energy per unit area of theEarth caused by the change in concentration of a climate forcing agent afterallowing for atmospheric temperatures, water vapour, and clouds to adjustto the change in concentration of the climate forcing agent.The direct effects incorporate aerosol-radiation interactions, whereby in-coming solar radiation may be scattered and directed back into space by31.1. Atmospheric aerosolscompounds such as sulfates, or may be absorbed by black carbon, trappingthe energy of the radiation in the atmosphere (Stocker et al., 2013). Blackcarbon also covers the Earth’s surface at polar regions, absorbing incomingsolar radiation and accelerating the melting of snow and ice (Bond et al.,2013).The indirect effects of aerosols are more complex, with aerosols influenc-ing cloud formation, as well as the microphysical properties, amount, andlifetime of clouds in the atmosphere (Haywood and Boucher, 2000; Twomey,1991). Clouds are composed of droplets that scatter incoming radiation. Anincrease in the emission of aerosols from anthropogenic sources may lead toan increase in aerosol number concentration, which increases the concentra-tion of cloud condensation nuclei in the atmosphere and affects the compo-sition of clouds. For example, clouds formed in regions of elevated aerosolconcentrations can contain a greater concentration of smaller droplets com-pared with clouds of equivalent water and ice content formed in regions oflower aerosol concentration. Of the two, the cloud formed under a higheraerosol concentration will scatter a greater proportion of incoming solar ra-diation. This increase in scattered solar radiation is referred to as the albedoeffect, or the first indirect effect (Twomey, 1991). A second indirect effect,known as the cloud lifetime effect, arises as a result of the reduced propen-sity of these smaller droplets to precipitate, increasing cloud lifetime in theatmosphere (Seinfeld and Pandis, 2006).The extent of the radiative forcing attributed to direct and indirect forc-ing is detailed in Figure 1.1, along with the radiative forcing of other an-thropogenic and natural atmospheric constituents (Stocker et al., 2013).Both the direct and indirect effects of aerosols are observed to give rise to anegative radiative forcing, or a net cooling effect on the Earth-atmospheresystem. However, there are large uncertainties associated with each effect,along with a low level of scientific understanding, which suggests there ismuch to benefit from an improved understanding of the climate effects ofaerosols.41.2. Formation of secondary organic materialFigure 1.1: A summary of radiative forcing by atmospheric aerosols andother atmospheric constituents between 1750 and 2011. Uncertainties rep-resent the 5-95 % confidence range, with solid error bars representing theuncertainty in the magnitude of effective radiative forcing and dotted barsrepresenting uncertainty in the magnitude of radiative forcing. Also shownis the level of confidence for each source of forcing, based on the currentlevel of scientific knowledge. Figure adapted from Figure TS.6 of Stockeret al. (2013).1.2 Formation of secondary organic materialThis thesis focuses on atmospheric aerosols consisting of secondary organicmaterial (SOM). SOM forms as a result of the oxidation of gaseous volatileorganic compounds (VOCs) emitted from the Earth’s surface by both an-thropogenic and natural sources. Once in the atmosphere, highly volatileorganic compounds readily react with atmospheric oxidants such as ozone(O3), hydroxyl radicals (•OH), and nitrous oxides (NO or NO2, commonlyreferred to collectively as NOx), leading to the formation of semivolatile re-action products (Hallquist et al., 2009; Kanakidou et al., 2005; Seinfeld andPankow, 2003). Following numerous successive oxidation reactions, these51.2. Formation of secondary organic materialsemivolatile compounds have a sufficiently low vapour pressure to partitionto the particle phase (Donahue et al., 2011; Ehn et al., 2014; Jimenez et al.,2009; Schobesberger et al., 2013). SOM is estimated to account for 30 to70 % of the mass concentration of suspended submicron particles in mostregions of the atmosphere (Kanakidou et al., 2005).Anthropogenic sources of VOCs include industrial and agricultural activ-ities, as well as the intentional burning of biomass (e.g. Hodzic et al., 2009),and annual anthropogenic emissions of VOCs are estimated to exceed 100Tg (Stocker et al., 2013). By far the largest natural sources of VOCs are theEarth’s biogenic regions (Stocker et al., 2013), with annual emissions fromforested areas estimated to range from 500 to 820 Tg (Arneth et al., 2008),and inventories of biogenic sources suggesting they emit at least as great anamount of mass of VOCs into the Earth’s atmosphere as all anthropogenicactivities (Guenther et al., 2000). One class of compounds that account fora significant fraction of VOC emissions are terpenes (e.g. Guenther et al.,2000). Terpenes are emitted by biogenic sources for an array of reasons, suchas in response to insect damage in order to both attract insect predators totheir surface and communicate the damage to neighbouring plants (Pare´ andTumlinson, 1999). The structures of isoprene and α-pinene, which accountfor the majority of biogenic VOC emissions (Kanakidou et al., 2005), areshown in Figure 1.2.The number of competing oxidation pathways for VOCs and their oxida-tion products result in the formation of thousands of individual componentsthat have been observed in field studies to be largely comprised of carbon,hydrogen, and oxygen atoms (Farmer et al., 2010; Russell et al., 2011),and contain a wide range of functional groups including alkanes, alkenes,Figure 1.2: Chemical structures of isoprene (left) and (+)-α-pinene (right).61.3. Viscosityalcohols, carboxylic acids, aldehydes, ketones, esters, ethers, and acid an-hydrides, as well as both aromatic and non-aromatic cyclic molecules (As-chmann and Atkinson, 1998; Chan et al., 2010; Chen et al., 2011a; Christof-fersen et al., 1998; Day et al., 2009; Russell et al., 2011; Surratt et al., 2006).As a result of this complexity only ≈10 % of the individual components ofSOM have been identified (Hallquist et al., 2009). The average oxygen-to-carbon elemental ratio (O:C) of SOM ranges from approximately 0.25-1.0(Jimenez et al., 2009), though the O:C of individual compounds can exceed1.0 (Chen et al., 2011a; Ehn et al., 2014; Praplan et al., 2015), and whilsta wide range of molar mass is observed for individual components, the vastmajority are of molar mass <1000 g mol−1 (Hallquist et al., 2009; Praplanet al., 2015; Schobesberger et al., 2013).Due to the complexity of SOM, explicit description of its formation,evolution, and physical properties of SOM cannot currently be incorporatedinto large-scale models (Hallquist et al., 2009). As a result, researchers haveused simple methods to describe these processes and properties.1.3 Viscosity1.3.1 What is viscosity?The viscosity of a material is a measure of its resistance to deformationunder stress, with the resistance caused by the materials’ intermolecularforces, and the local structures they produce (Viswanath et al., 2007).Two measures of viscosity are commonly used, dynamic viscosity (η)and kinematic viscosity (υ), with kinematic viscosity incorporating both amaterials dynamic viscosity and its density (σ) (Equation 1.1),ν =ησ(1.1)Dynamic viscosity is the measure of the tangential force required to slideone layer of fluid against another (Viswanath et al., 2007), which produces amechanical non-equilibrium system, and is measured in Pascal seconds (Pas). The force is required to overcome intermolecular interactions between71.3. Viscosityneighbouring molecules, which may be electrostatic and/or steric in nature.This may be visualised as shown in Figure 1.3, for a sample between twoboundary plates, separated by a distance of y. One plate remains stationarywhilst the second plate moves, with velocity υ, imparting a shear stress, τ ,on the fluid. Resistance to this stress results in layers closer to the movingplate travelling with greater velocity than those far from the moving plate.From this picture, dynamic viscosity may be expressed mathematically as,η =τδυδy(1.2)Dynamic viscosity, rather than kinematic viscosity, is used throughoutthis thesis, and further mentions of viscosity in this thesis refer exclusivelyto dynamic viscosity. This is for two main reasons. Firstly, diffusion is anatmospherically important process, and diffusion and dynamic viscosity maybe readily converted using the Stokes-Einstein equation,η =kBT6piDr(1.3), where kB is the Boltzmann constant (1.38 x 10−23 J K−1), T is tempera-ture (K), D is the diffusion coefficient (m2 s−1), and r is the hydrodynamicradius (m) of the diffusing molecule. Secondly, the density of the materialunder investigation is not always known accurately.Figure 1.3: Laminar flow of a fluid between two 2D boundary plates - onestationary and one moving.81.3. ViscosityThat the Stokes-Einstein relationship successfully relates diffusion, whichis relevant on a molecular length scale, and viscosity, which is relevant on amacroscopic scale, may be surprising. However, whilst it is known to breakdown under certain conditions, such as close to a compounds glass transitiontemperature, it is nevertheless successful in correlating the temperature anddensity variations of diffusion and viscosity across a wide range of values forliquids (Corti et al., 2008; Stillinger and Debenedetti, 2005). An alternativemethod by which the transport of momentum, which gives to viscosity, maybe studied, is on the microscopic level via the Green-Kubo formalism. Theformalism may be expressed as,η =1kBTV∫ ∞0〈Jxy(0)Jxy(t)〉dt (1.4), where kB and T are as for Equation 2.6, V is the macroscopic systemvolume, the brackets 〈...〉 denote the ensemble average, and Jx,y is the x,ycomponent of a momentum flux tensor, which is expressed as:Jxy =N∑j=1(pjxpjym+ xjFjy) (1.5), where p represents the momentum of particle j, m represents the parti-cle’s mass, and Fjy represents the y component of the force experienced byparticle j.Whilst the Kubo-Green expressions provide an overview of the forcesexperienced by particles at the microscopic level, the more granular Stokes-Einstein equation is used throughout this thesis due to both the experimentalobservation that many liquids conform closely to Stokes-Einstein behaviour(Stillinger and Debenedetti, 2005) and its comparative simplicity.1.3.2 Types of fluidsThe relationship between shear rate and shear stress is used to broadly di-vide fluids into those that are Newtonian and those that are non-Newtonian91.3. Viscosity(Viswanath et al., 2007). Whilst Newtonian fluids exhibit a linear relation-ship between shear rate and applied shear stress, and thus have viscositiesthat are independent of applied shear stress, non-Newtonian fluids exhibit anon-linear relationship between shear rate and applied shear stress, meaningtheir viscosities are dependent upon applied shear stress. Examples of typesof non-Newtonian fluids include pseudoplastic, or shear-thinning, fluids thatexhibit a decrease in viscosity as shear rate is increased, and dilatant, orshear-thickening, fluids, that exhibit an increase in viscosity as shear rate isincreased. The relationship between shear rate and applied shear stress isshown in Figure 1.4 for Newtonian, pseudoplastic, and dilatant fluids.Though the relationship between shear rate and applied shear stress forsecondary organic material (SOM), which is the main focus of this thesis, hasyet to be determined, a linear relationship between shear rate and appliedshear stress has been determined for compounds that are typically used asproxies for SOM, including glycerol/water solutions (Hosny et al., 2013)and sucrose/water solutions (Saggin and Coupland, 2004). As such, SOMis assumed to be Newtonian in nature.Figure 1.4: The relationship between shear rate and shear stress for typesof fluids.101.3. Viscosity1.3.3 Commercially available methods to measure viscosityTwo commonly used families of commercially available viscometers are thosetypically classified as capillary viscometers and rotational viscometers.Capillary viscometers use either gravity or an applied pressure as theforce used to push or pull a sample through a capillary of known diameter. Aschematic of an Ostwald viscometer, an example of a capillary viscometer, isshown in Figure 1.5(a). A known volume of sample is added to the capillary(at point 1, filling the capillary to line A), and suction applied (at point 2).The time taken for the sample to flow between specified points (lines B andC) inside the capillary is measured, with the viscosity being calculated usingthe general equation,η = kt (1.6), where k is a constant that takes account of the capillary dimensions andthe suction applied to pull the sample through the capillary, and t is thetime taken for the sample to flow between lines B and C.Rotational viscometers determine the amount of rotational force thatmust be applied to a sample held in an outer cylinder in order to rotatean inner solid cylinder at the centre of the sample. One such design is thecoaxial-cylinder viscometer, a schematic of which is shown as Fig. 1.5(b).The angular deflection of the inner cylinder is a function of the viscosity ofthe sample, the dimensions of the inner and outer cylinders, and the speedof rotation of the outer cylinder.Figure 1.6(a) shows a scale outlining part of the viscosity continuum.Viscosity spans an enormous range for common substances, with at least 15orders of magnitude separating that of water from that of a glass.Commercially available viscometers are capable of measuring materialswith viscosities ranging from <10−3 - ≈106 Pa s, which corresponds tomaterials less viscous than water up to those of viscosity between that ofwindow putty and tar pitch (Fig. 1.6(b)).The design of many commercially available instruments such as capillary111.3. ViscosityFigure 1.5: Schematic of two examples of commercially available viscome-ters. Shown in (a) is an Ostwald viscometer. Sample is added to side 1until the sample in the capillary is level with line A. Suction is subsequentlyapplied at point 2, and the time taken for the sample to flow between line Band line C is measured. Shown in (b) is a coaxial-cylinder viscometer. Theouter cylinder is rotated at a constant speed, and the angular deflection ofthe inner cylinder is used to determine the viscosity of the sample.and rotational viscometers require samples on the order of grams or tens ofgrams of material in order to determine viscosities, though SOM is typicallyonly produced on the microlitre (µL) scale. Whilst methods are availablefor measuring viscosities on this scale, such as microfluidic capillary devicessimilar but smaller in design compared to their macro counterparts (Hanet al., 2007; Lin et al., 2007; Pipe and McKinley, 2009; Srivastava et al.,2005), the range of viscosities they can measure is limited to viscosities<10−1 Pa s, approximately equivalent to that of olive oil (Renbaum-Wolffet al., 2013b). The range of viscosities that may be measured for techniquesutilizing µL of material is illustrated in Fig. 1.6(b).121.3. ViscosityFigure 1.6: (a) A scale showing part of the viscosity continuum, detailingthe regions typically defined to be bound by the solid, semi-solid, and liq-uid regimes, as well as the viscosities of some common substances (idea toshow common substances is borrowed from Koop et al., 2011). The imageof the pitch is a detail of a picture taken from Wikimedia Commons of thepitch drop experiment Wikipedia page (GFDL, John Mainstone, Univer-sity of Queensland, Australia). (b) A scale showing the ranges of viscositymeasurable using commercially available instrumentation for measuring vis-cosity, as well as the possible range of viscosities spanned by SOM.1.3.4 Potential viscosities of SOMThe initial studies related to the viscosity of SOM have focused on discussionof the phase of the SOM, i.e., whether it is liquid, semi-solid, or solid, withthe terms solid and glass often being used interchangeably, as opposed to anumerical measure of its viscosity. As illustrated in Fig. 1.6(a), liquids areof viscosity <102 Pa s, semi-solids are of viscosities between 102 and 1012Pa s, and solids are of viscosity >1012 Pa s (Angell, 1995; Shiraiwa et al.,2011a).Initial studies by Virtanen et al. (2010) concluded SOM particles canadopt an amorphous solid - most probably glassy - state. As such, the vis-cosity of SOM may range from that of water (≈10−3 Pa s) to that of a131.3. Viscositysolid (>1012 Pa s) as illustrated in Fig. 4(b). Given the limited quantity ofSOM that can be generated on a reasonable timescale in the laboratory (onthe order of a few milligrams), currently available techniques for measur-ing viscosity are unlikely to be able to measure the full range of viscositiesexhibited by laboratory generated SOM.1.3.5 The importance of the viscosity of SOMThe viscosity of SOM is important for a number of reasons. First, theviscosity of SOM governs the rate at which organic molecules can diffusethrough particles, and knowledge of the viscosity is thus required to predictthe mechanism, rate of growth, total mass, and size of modelled particles(Figure 1.7(a)) (Riipinen et al., 2011; Shiraiwa and Seinfeld, 2012; Shiraiwaet al., 2011a, 2013; Zaveri et al., 2014). The reaction of oxidants within parti-cles comprised of SOM may also be inhibited at high viscosities (Fig. 1.7(b))(Shiraiwa et al., 2011a), as well as rates of both heterogeneous and photo-chemical processes (Houle et al., 2015; Kuwata and Martin, 2012; Lignellet al., 2014; Zhou et al., 2013).In addition, high viscosities in particles containing SOM may alter theirphysical properties after ice cloud processing (Adler et al., 2013; Robin-son et al., 2014), and high viscosities may also inhibit crystallization ofinorganic salts (Bodsworth et al., 2010; Murray and Bertram, 2008; Songet al., 2012), and the hygroscopic properties of particles (Bones et al., 2012;Hawkins et al., 2014; Lu et al., 2014; Price et al., 2014; Tong et al., 2011).Furthermore, if SOM particles are solid or ’glassy’ in phase under atmo-spheric conditions they may provide a surface for ice nucleation (Baustianet al., 2013; Berkemeier et al., 2014; Knopf and Rigg, 2011; Ladino et al.,2014; Murray et al., 2010; Schill et al., 2014; Wang et al., 2012), modifyingtheir ability to promote cloud formation. Viscosity in SOM particles is alsoimportant for predicting the long range transport of polycyclic aromatic hy-drocarbons, which are known to have a detrimental effect on human health(Zelenyuk et al., 2012; Zhou et al., 2012).141.3. ViscosityHigh viscosity particle Low viscosity particle SOM growth mechanism and rate (a) SOM SVOC High viscosity particle Low viscosity particle Heterogeneous oxidation by O3 (b) oxidation O3 O3 O3 O3 Direct effect  VOCs  Oxidation + Nucleation CCN Activation  particle mass may be lower if viscous Indirect effect  Growth (rates slower if viscous) Climate Effects Heterogeneous oxidation particles may be less oxidized if viscous (c) 1Figure 1.7: (a) Effect of particle viscosity on the mechanism of growth ofSOM by semivolatile organic compound (SVOC) uptake. (b) Effect of par-ticle viscosity on heterogeneous oxidation by ozone. (c) Climate effects ofparticles and implications of high particle viscosities on particle growth rates,particle mass, and heterogeneous oxidation by O3. The implications of par-ticle viscosity on growth and heterogeneous oxidation in (a)-(c) assume amonodisperse particle population. Originally published in Renbaum-Wolffet al. (2013a).151.4. Previous research related to the viscosity of SOM1.4 Previous research related to the viscosity ofSOMResearchers have traditionally assumed particles containing SOM to be oflow viscosity when modeling particle growth (Hallquist et al., 2009). How-ever, recent measurements have suggested that this may not be the caseunder certain conditions. Measurements that have suggested SOM can havehigh viscosities include (1) direct measurements of viscosity of SOM or prox-ies for SOM (Booth et al., 2014; Pajunoja et al., 2014; Song et al., 2015;Zhang et al., 2015), (2) measurements of diffusion rates and mixing timesin SOM (Abramson et al., 2013; Loza et al., 2013; Perraud et al., 2012), (3)bounce measurements off surfaces (Bateman et al., 2015; Kidd et al., 2014;Saukko et al., 2012; Virtanen et al., 2010, 2011), (4) measurements of theflatness of particles after impaction (O’Brien et al., 2014), (5) measurementsof rates of evapouration from SOM (Cappa and Wilson, 2011; Vaden et al.,2011), and (6) measurements of reactivity of SOM (Kuwata and Martin,2012; Wang et al., 2012, 2015). Nevertheless, the viscosities and diffusionrates of SOM are still a matter of debate (Price et al., 2015; Robinson et al.,2013; Saleh et al., 2013; Yatavelli et al., 2014).Given this array of observations, and the range of atmospheric processesaffected by the viscosity of SOM, it is important to determine the viscosityof SOM in order to improve our ability to predict the climate effects of SOM.1.5 Overview of dissertationChapter 1 (this chapter) provides an introduction to atmospheric aerosolsand secondary organic material as a well as a motivation for the rest of thethesis. Chapter 2 details two novel techniques used to measure a wide rangeof viscosities in small samples of organic compounds, the bead-mobility andthe poke-and-flow technique combined with simulations of fluid flow. Chap-ter 3 describes initial measurements of the viscosity of the water-solublecomponent of the SOM produced via the ozonolysis of α-pinene. Chapter4 details further validation of the poke-and-flow technique, which is subse-161.5. Overview of dissertationquently utilised in Chapter 5 to determine the viscosity of the total SOM(water-soluble and water-insoluble fractions) produced via the ozonolysis ofα-pinene, and in Chapter 6 to determine the viscosity of highly oxidizedcompounds that have been previously identified as components of SOM,and are used as proxies for highly oxidised SOM found in the atmosphere.In Chapter 7 the relationship between viscosity, elemental oxygen-to-carbonratio (O:C), molar mass, and saturation vapour concentration, the massbased equivalent of saturation vapour pressure, is investigated.17Chapter 2Experimental techniquesThe experimental results in this thesis are predominantly determined usingtwo novel techniques for measuring high viscosities in small samples, a bead-mobility technique and a poke-and-flow technique. These new techniques formeasuring viscosities were needed because conventional viscometers are notable to accommodate the small sample volumes associated with laboratory oratmospheric sampling of SOM particles (typically on the order of 1 µL) (seeSection 1.3.3), and because microviscometers that can accommodate smallsample sizes are often limited to measurements of low viscosities (<0.1 Pa s)(Han et al., 2007; Lin et al., 2007; Silber-Li et al., 2004; Srivastava and Burns,2006), whereas SOM is anticipated to have considerably higher viscosity, atleast at low RH. These newly introduced techniques may also find broaderfuture use in other disciplines that require viscosity measurements of smallsample volumes, e.g., due to cost or availability such as biological samples.The bead-mobility technique is used in Chapters 3 and 6, and the poke-and-flow technique is used in Chapters 3, 4, 5, and 6. Both of the techniques aredescribed in detail here, and these descriptions referred to in the relevantplaces in further Chapters.2.1 Bead-mobility techniqueThe bead mobility technique was first described, and validated, in Renbaum-Wolff et al. (2013b).Super-micron sized particles, typically 20-50 µm in diameter, were gen-erated on a hydrophobic slide surface by using a nebuliser (Meinhard, modelTR-30-A1, USA) to nebulise dilute aqueous solutions onto a hydrophobicglass slide or a Teflon slide. The glass slides used during experiments in182.1. Bead-mobility techniqueChapters 3, 4, and 6 were siliconised glass slides obtained commercially(12mm, Hampton Research, USA), whilst the glass slides used during ex-periments in Chapter 5 were produced by cleaning plain circle glass coverslides (12mm, Hampton Research, USA) using Piranha solution (three partsconcentrated sulfuric acid to one part 30 % hydrogen peroxide solution), andsubsequently coating the slides using a fluorinating agent (Trichloro-(1H,1H, 2H, 2H-perfluorooctyl) silane, Sigma Aldrich). A dilute aqueous sus-pension of 1 µm hydrophilic melamine beads (actual diameter 930 ± 50 nm(Sigma Aldrich, Cat# 86296)) was then nebulised over the slide containingthe super-micron sized particles, resulting in beads being incorporated intothe bulk of the particle. Melamine beads were chosen as they are preparedby the manufacturer without the use of surfactants and are not susceptibleto swelling or aggregation in solution. The slide containing the super-micronsized particles with bead inclusions was then placed in a flow cell with rela-tive humidity control.A schematic of the flow cell is shown as Figure 2.1. The flow cell wasmounted to an optical microscope (Zeiss, Axio Observer), and relative hu-midity inside the cell was controlled by passing a flow gas (N2) througha water bubbler located in a controlled-temperature bath. The dew pointtemperature of the gas was measured after the flow cell using a hygrometer(General Eastern, Model 1311DR). The temperature of the flow cell wasmeasured using a thermocouple probe. The hygrometer was calibrated atthe beginning of each set of experiments using the deliquescence of ammo-nium sulfate particles, with the uncertainty (1σ) of the hygrometer aftercalibration typically ± 0.5 % RH at 80.3 % RH. The flow gas (linear flowvelocity of 100 cm s−1), produced a shear stress on the particle surface(Figure 2.2), which resulted in internal circulations of material within theparticle. These circulations also carried the beads. Examples of the pathstravelled by the beads are shown in Figure 2.3, and observed to be roughlyhemispherical in nature, with the shear stress of the gas causing the beadsto travel around the outside of the particle, before travelling back throughits centre.192.1. Bead-mobility techniqueFigure 2.1: Schematic representation of bead-mobility experimental setup.Originally published in Renbaum-Wolff et al. (2013b).Figure 2.2: Illustration of the flow of the flow gas around a sample particle.Originally published in Renbaum-Wolff et al. (2013b).Figure 2.3: Internal circulation of beads within a particle of glycerol. Thered lines overlaid on the image show the 2-D circulation of three beads withinthe particle. Originally published in Renbaum-Wolff et al. (2013b).202.1. Bead-mobility techniqueOptical microscopy was used to track the movement of the beads overtime, with a frame collected every 0.2-40 s, depending on the rate of move-ment of the beads within the particles. In total a time series of 50 to 100optical images were recorded, and the movement of the beads was tracked asthey travelled all the way around the particle. Examples of optical imagesthat show the change over time in the x and y co-ordinates of beads withina particle are included in Figure 2.4.(a) 90 % RH t = 0 seconds t = 55 seconds(b) 70 % RH t = 0 seconds t = 70 minutesFigure 2.4: Images from the bead-mobility studies. Images in (a) correspondto images recorded at 90 % RH and images in (b) correspond to imagesrecorded at 70 % RH. Indicated in the figure are different beads monitored inthe experiments and the x-y coordinates of the beads. Originally publishedin Renbaum-Wolff et al. (2013a).212.1. Bead-mobility techniqueGiven the microscope was focused on a single plane in the z -axis, move-ment of beads in the z plane could not be quantified. Further, there is likelysome variation of bead speed at different heights within the particle and thuseffort was made to focus at the mid-height of each particle studied. Eachof these processes are potential sources of error for the technique. Above acertain viscosity the movement of the beads became too slow to quantify.The bead speed was converted to viscosity using a calibration curve such asthat shown in Figure 2.5. Such calibration curves typically gave rise to 95 %prediction limits that had lower/upper limits of viscosity that were withina factor of two of the best-fit function.The organic molecules used to generate Fig. 2.5 are listed on the fig-ure, and had O:Cs ranging from 0.1-1.0, molar masses ranging from 92-600g mol−1, surface tensions ranging from 32-75 mN m−1, and contact angleswith the slide substrate ranging from 58-95 ◦. The diameter of the parti-cles studied ranged from 20-50 µm. Within experimental uncertainty, therelationship between bead speed and viscosity was determined to be in-Figure 2.5: Average bead speed (±1 σ) vs. viscosity for standard com-pounds. Originally published in Renbaum-Wolff et al. (2013b).222.2. Poke-and-flow technique combined with simulationsdependent of these parameters within the ranges studied (Renbaum-Wolffet al., 2013b).The range of these physical properties is expected to extend beyond therange of that exhibited by SOM. For example, secondary organic particlesgenerated from the ozonolysis of α-pinene in environmental chambers havean average O:C of 0.3-0.4 (Aiken et al., 2008; Chen et al., 2011a; Heatonet al., 2007; Putman et al., 2012; Shilling et al., 2009). Although the surfacetension of secondary organic material collected in environmental chambers islargely unmeasured for the subsaturated RH regime, estimates of 40-75 mNm−1 have been made based on model compounds (Huff-Hartz et al., 2006;Hyva¨rinen et al., 2006; Tuckermann and Cammenga, 2004). The myriadcompounds present in the SOM have been estimated to have molar masslargely less than 600 g mol−1 (Gao et al., 2004, 2010; Putman et al., 2012).Measurements in Chapter 3 that were made using the bead-mobilitytechnique used the calibration curve shown in Fig. 2.5. Measurements inChapter 6 that were made using the bead-mobility technique occurred morethan a year after those in Chapter 3 and so a new calibration curve wasproduced to ensure any change in the instrumentation was accounted for.This calibration curve comprised of measurements of sucrose and glycerol,as it was demonstrated in Renbaum-Wolff et al. (2013b) that a calibrationcurve produced using just sucrose and glycerol predicted the same viscosityfor olive oil as the combination of all nine compounds included in Fig. 2.5.2.2 Poke-and-flow technique combined withsimulations of fluid flow2.2.1 Poke-and flow techniqueThe qualitative method of poking a particle to determine the particle phase(i.e., solid/semisolid vs. liquid) was introduced by Murray et al. (2012).This approach is expanded upon in this thesis by quantifying flow ratesafter poking and determining viscosities from simulations of flow.Super-micron sized particles were generated on a hydrophobic glass slide232.2. Poke-and-flow technique combined with simulations(Hampton Research, Canada) and the slide containing the super-micronsized particles was then placed in a flow cell with relative humidity control.A schematic of the flow cell is shown as Figure 2.6. The flow cell is similarto those described previously (Koop et al., 2000; Song et al., 2012; You et al.,2012), but with a small hole added at the top through which a needle couldbe inserted (Fig. 2.6). The flow cell was mounted to an optical microscope(Zeiss, Axio Observer), and relative humidity inside the cell was controlledby passing a flow gas (N2) through a water bubbler located in a controlled-temperature bath. The dew point temperature of the gas was measuredafter the flow cell using a hygrometer (General Eastern, Model 1311DR).The temperature of the flow cell was measured using a thermocouple probe.The hygrometer was calibrated at the beginning of each set of experimentsusing the deliquescence of ammonium sulfate particles, with the uncertainty(1σ) of the hygrometer after calibration typically ± 0.5 % RH at 80.3 % RH.The presence of the hole at the top of the flow cell upon RH in the cell wasdetermined by studying the deliquescence relative humidity of ammoniumsulfate and potassium carbonate particles with the hole open and the holeclosed. The DRH of ammonium sulfate particles differed by <0.2 % RH at80 % RH, whilst the DRH of potassium carbonate particles differed by <0.4% RH at 43 % RH.Figure 2.6: Schematic representation of poke-and-flow experimental setup.Originally published in Grayson et al. (2015a).242.2. Poke-and-flow technique combined with simulationsNeedles were used to poke the particles, with two different types of nee-dles used. Both types of needle had tips that were circular in geometry. Forparticles of lower viscosities a sterilized, sharp needle (0.9 mm x 40 mm)(Becton-Dickson, USA), with a tip diameter of ≈20 µm, was used. How-ever, at higher viscosity particles stuck to the needle and as a result wereremoved from the substrate. As such, a second set of needles (RS-6063,Roboz Surgical Instrument Co., USA), with a tip diameter of ≈10 µm, wasoccasionally used. These needles were coated with a hydrophobic Dursancoating (SilcoTek, USA) to prevent material sticking to the needles. In eachcase the needles were mounted to a micromanipulator (Narishige, modelMO-202U, Japan) and inserted through the hole at the top of the flow-cell.The micromanipulator was used to move the needle in the x -, y-, and z -,axes.For a given experiment, the tip of the needle was aligned vertically abovethe centre of a particle, and then moved down in the z direction, impactingthe particle at the peak of its spherical cap geometry. The poking andsubsequent behaviour of the particles was monitored during experimentsusing a reflectance optical microscope (Zeiss, Axio Observer), and recordedusing a CCD camera.For particles of lower viscosity, the needle penetrated the particle andsubsequently came in contact with the substrate beneath. After the needlewas removed the material of the particle was present in a non-equilibrium,half-torus, geometry, and began to flow in order to minimize the surfaceenergy of the system. Eventually the hole at the centre of the half-toruscompletely closed, and the particle returned to its original, spherical cap,morphology. Shown in Figure 2.7(a-c) are examples of particles before andafter being poked that exhibit this behaviour.Analysis of optical images captured during each experiment was per-formed using Zen software (Zeiss). The hole at the centre of the half-toruswas traced and the area of the hole calculated. An equivalent area diam-eter of the hole was calculated via the relationship d = (4A/pi)1/2, whered is the equivalent area diameter of a hole of area, A (Reist, 1992). Theexperimental flow time, termed τ exp,flow, was assigned as the time taken for252.2. Poke-and-flow technique combined with simulationspre-poke poke - 0 seconds 0.2 seconds 10 seconds 50 secondsa) 70% RH(HEC)pre-poke 1 second 20 seconds 5 minutesb) 50% RH(PNNL)pre-poke 1 second 3 minutes 20 minutesc) 45% RH(PNNL)pre-poke 2 seconds 30 minutes 8 hoursd) 30% RH(HEC)1Figure 2.7: Deformation and recovery across time of poked SOM particles.Prior to poking (pre-poke), the particle morphology can be approximatelydescribed as a spherical cap. At higher RH (40-70 %) (rows a-c), geometriesapproximately described as half-torus are formed, and flow occurs. By com-parison, for low RH (≤30%) (row d) particles shatter and do not flow overa period of 8 h. The ring structures observed in the first and last columnsare an optical effect that arises from hemi-spherical diffraction. Originallypublished in Renbaum-Wolff et al. (2013a).the equivalent area diameter to decrease to 50 % of its original value. Thisdefinition of experimental flow time was chosen as it allows experimentalflow time of more viscous particles, the holes of which may not fully closeon a laboratory timescale, to be measured.In some of the experiments a point around the inner edge of the half-torus appeared to be pinned to the hydrophobic surface. This behaviourmay suggest that in some cases the needle scratched the surface. Particlesthat exhibited this pinning behaviour were excluded from analysis in order262.2. Poke-and-flow technique combined with simulationsto avoid influencing the results. Pinning affected <20 % of all the particlesthat formed a torus geometry after poking.Particles of higher viscosity cracked when impacted by the needle, form-ing pieces with sharp, distinct edges. These particles were observed overa period of hours, with no flow discernible at the edges. An example of aparticle exhibiting this behaviour after upon and after being poked is shownin Figure 2.7(d).2.2.2 Simulations using COMSOL MultiphysicsSimulation of particles exhibiting flow during poke-and-flow experimentswas carried out using the laminar two-phase flow moving mesh mode withinthe Micro-fluidics module of COMSOL Multiphysics (version 4.3a), a finiteelement analysis software package.Finite element analysis employs the calculus of variations in order todetermine approximate solutions to boundary value problems for differentialequations. The method involves deducing simple equations to describe small,finite, regions, which are subsequently used to determine a more complexequation that approximates a larger domain.The Navier-Stokes equation, including surface tension, was used to de-scribe the transport of mass and momentum. The Navier-Stokes equationsdescribe viscous flow by applying Newton’s second law to fluid motion, whilstalso assuming that the stress in a fluid is the sum of a diffusing viscous termand a a pressure term.Newton’s second law,F =dpdt=d(mv)dt(2.1), states that the force (F) on an object is equal to the rate of changeof its linear momentum (p), where momentum = mass (m) multiplied byvelocity (v). As well as Newton’s first law, this relationship also implies aconservation of momentum whereby an object in motion will be of constantmomentum if no external forces are acting upon it.272.2. Poke-and-flow technique combined with simulationsIn an inertial frame of reference Newton’s second law can be appliedto produce a general form of the Navier-Stokes equations, which may beexpressed as:ρ(dudt+ u∇u)= −∇p+∇τ + f (2.2), where ρ represents density, u represents the flow velocity, p representspressure, τ represents a stress tensor, and f represents other forces actingon the material. For the simulations used herein one of these other forces isa frictional boundary force that arises at the fluid-substrate interface, andmay be expressed as:Ffr = − ηβu (2.3), where β represents slip length. The greater the slip length, the weakerthe frictional force at the fluid-substrate interface. Incompressible flow isassumed for the poke-and-flow experiments due to the low flow rate of theparticles with respect to the gas flow around them. In such instances thehis allows the Navier-Stokes equations to be expressed as:ρ(dudt+ u∇u)= −∇p+ η∇2u+ f (2.4)Division by ρ followed by rearrangement gives rise to the equation,dudt= u∇u− ∇pp+fp+ηp∇2u (2.5), which is of the form momentum/mass = mass + pressure + body force+ viscosity, consistent with the conservation of momentum.The evolution of the fluid as it flowed over time was tracked using theArbitrary Lagrangian Eularian (ALE) method. Within ALE, the initialflow is Lagrangian, whereby the mesh used to describe the materials initialgeometry moves with the material. However, over time this can lead tosignificant distortion of the mesh. In such instances a simulation using an282.2. Poke-and-flow technique combined with simulationsALE method will re-position distorted portions of the mesh. ALE provides acompromise between the large expense of models utilising a pure Lagrangianapproach whilst providing a more accurate solution than a pure Eularianapproach.2.2.3 Simulations of particles exhibiting flowSimulations were performed with a mesh that consisted of ≈5800 elementsand had a mesh spacing of 3.92-337 nm. A top view of the half-torus ge-ometry used to simulate the experimental observations is shown in Figure2.8(a), where the dotted line is at the midpoint between the inner and outeredges of the ring of material forming the torus. The initial radius of thehole at the centre of the half-torus is denoted as R0-r0, where R0 representsthe distance from the centre of the hole to the midpoint of the ring of ma-terial that creates the torus, whilst r0 represents the radius of the ring ofmaterial. The half-torus had two distinct surfaces (Figure 2.8(b)). Surface1 represents the air-fluid interface, which was allowed to undergo free defor-mation in all dimensions. Surface 2 is the fluid-substrate interface, whichwas allowed to undergo free deformation in the horizontal x-y plane, but notin the vertical, z, direction. In the simulations, the size of the hole in thehalf-torus geometry decreases in an axi-symmetric manner, as the materialflowed to attain a spherical cap geometry and thus minimize the surfaceenergy of the system. The time taken for R0-r0 to decrease to 50 % of theinitial value was assigned τmodel,flow. Flow occurred in the simulations in away that minimized the total surface energy.For a given simulation, values of R0, r0, surface tension, slip length, anddensity were required. The dimensions R0 and r0, and the equilibrium con-tact angle, were determined from measurements, and the values of surfacetension (Surface 1, Fig. 2.8(b)), slip length (which describes the interactionat Surface 2 in Fig. 2.8(b)), and the density of the material were deter-mined based on literature values. A summary of literature values for theslip length of water, as well as some atmospherically relevant compounds, isincluded as Table 2.1. Included as Figure 2.9 are the results from a range of292.2. Poke-and-flow technique combined with simulationsFigure 2.8: Details of half-torus model used to simulate the flow in experi-ments: (a) top view, where R and r are the notations used here to describethe dimensions of a half-torus geometry; (b) side view, where surface 1 rep-resents the air-fluid interface, and surface 2 represents the fluid-substrateinterface. Originally published in Grayson et al. (2015a).simulations, where surface tension, slip length, density, contact angle, andτmodel,flow are varied across a range greater than that expected for poke-and-flow experiments. The viscosities determined via simulation suggestthat the largest uncertainty associated with the technique is that due toslip length, with simulated viscosities varying by approximately two ordersof magnitude across the range of slip lengths outlined in Table 2.1. Acrossthe range of values studied the simulated viscosities were determined to beless dependent on surface tension and contact angle, and determined to beindependent of the density of the particle. As expected a linear relation-ship was observed between viscosity and τmodel,flow. The viscosity of eachparticle that was poked and formed a half-torus geometry, and was not sig-nificantly influenced by scratches (determined visually, as discussed above),was determined via simulations. Further details are provided in individualChapters.In Chapter 3, experimental flow times, τ (exp,flow), were converted toviscosities using the linear relationship between the modeled flow time,τ (model,flow), and viscosity for a particle of R0 = 20 µm and r0 = 7.5 µm,representative of the size of particles studied, whilst simulations of particles302.2.Poke-and-flowtechniquecombinedwithsimulationsTable 2.1: Summary of literature values of measured slip lengths for water and atmospherically relevant organiccompounds.Reference Solution Surface Slip length / nmSchnell (1956) Water Siliconised glass 2000-8000 aChuraev et al. (1984) Water, mercury Siliconised quartz 10-30Watanabe and Udagawa (1999) Water and glycerinAcrylic resin treated200-450awith hydrophobic silicaBaudry et al. (2001) Glycerol Gold and thiol 38Craig et al. (2001) Sucrose-waterGlass coated with 11-mercapto-<201-undecanol and 1-dodecanethiolTretheway and Meinhart (2002) Water Siliconised glass 1000Cheng and Giordano (2002)Water, hexane,Glass 20-90tetradecaneJin et al. (2004) WaterPolydimethylsiloxane<10microchannelJoseph and Tabeling (2005) Water Siliconised glass <100Choi and Kim (2006) Glycerol Superhydrophobic surface ≈50000Joly et al. (2006) Water BK7 glass 8-9Zhu et al. (2012) di-n-octyl phthalate Silanised silicon <400Li et al. (2014) Water Silicon <10a Slip lenghts inferred by Lauga et al. (2003) from reported experimental data312.2. Poke-and-flow technique combined with simulationsFigure 2.9: The dependence of viscosity as (a) surface tension, (b) sliplength, (c) density, (d) contact angle, and (e) τmodel,flow, are varied acrossa wide range of values for particles of dimensions R0 = 16, r0 = 4 (filledsymbols) and R0 = 14, r0 = 6 (open symbols). For each simulation fourof the five properties were held constant whilst the fifth was varied. Blacksquares: surface tension = 75.15 mN m−1, slip length = 10 µm, density= 1500 kg m−3, contact angle = 98.5 o, τmodel,flow = 1000 seconds. Redcircles: surface tension = 57.2 mN m−1, slip length = 5 nm, density = 1500kg m−3, contact angle = 98.5 o, τmodel,flow = 0.1 seconds.in Chapters 4-6 used the dimensions of each particle. The viscosity usedin the simulations of a particle was varied until τmodel,flow agreed with theparticles τ exp,flow (to within 1 %). Further details are provided in individualChapters.2.2.4 Simulations of particles that cracked when pokedSimulation of particles that cracked when poked, and exhibited no dis-cernible flow during subsequent observations was also carried out using thelaminar two-phase flow moving mesh mode within the Micro-fluidics module322.2. Poke-and-flow technique combined with simulationsof COMSOL.For the initial conditions a quarter of a sphere was used (radius = 10µm) with one of the flat faces in contact with a solid substrate (see Figure2.10). In this case, there are three interfaces, labeled 1-3 in Fig. 2.10.Interfaces 1 and 2 represent external fluid interfaces and were assigned asurface tension that represented the lower limit of the estimated surfacetension of the particle and allowed to undergo free deformation. Interface 3represents the interface between the fluid and the solid slide substrate. Thisinterface was allowed to deform in the X-Y plane but not in the Z direction,and the equilibrium contact angle between the surface and the fluid wasassigned a value of 90 ◦ (see below for implications).The extent of movement of the sharp edge of the particle was indepen-dent of the equilibrium contact angle between 20 ◦ and 100 ◦. Interactionsbetween the fluid and the solid surface were described with a Navier slipwall boundary condition, with a slip length of 0.01*l, where l is the gridspacing of the mesh which ranged from 1-1.7 µm. Thus, the slip length wastypically between 10-17 nm, consistent with experiments of water on hy-drophobic substances at low shear rate (Churaev et al., 1984). The amountof movement of the sharp edge of the particle was only weakly dependentFigure 2.10: Details of quarter-sphere model used to simulate flow for parti-cles that exhibit cracking behaviour and no discernible flow over subsequenthours of observation. Surfaces 1 and 2 represent the external fluid interfaces,to which a relevant value of surface tension from literature was assigned, andsurface 3 represents the particle-substrate interface, which was representedas a Navier-slip boundary with a slip length of 10-17 nm. Originally pub-lished in Renbaum-Wolff et al. (2013a).332.2. Poke-and-flow technique combined with simulationson the slip length used, leading to less than a 10 % change in the observedmovement when the slip length was varied from 1 nm to 10 µm. A relevantvalue for density was assumed based on literature values. The viscosity ofthe fluid was adjusted until the sharp edge at the top of the particle, (viewedfrom above the quarter-sphere) moved by ≈ 0.5 µm over 8 h, a typical exper-imental time (see Fig. 2.7(b) and Fig. 2.10). The extent of movement, 0.5µm, was chosen because if movement of this magnitude occurs it is clearlydetectable in the microscope images.The viscosity determined as described above is a lower bound to viscosityfor the following reasons: (a) movement of 0.5 µm is an upper limit tothe movement observed in our experiments, (b) if a higher surface tensionwere used in the simulations the result would be a higher prediction of theviscosity.Simulations were initialized as a quarter sphere having one flat face incontact with a substrate. In a set of stepwise simulations, the viscosity wasdecreased until the maximum displacement at the corners of the quarter-sphere was 0.5 µm in 8 h, establishing a lower limit for viscosity of 5× 108Pa s for a particle that cracks and exhibits no discernible flow over the courseof 8 h (see Section 3.2.5).Further simulations were also carried out using other geometries to de-termine if the predicted lower bounds of the viscosities were sensitive to theinitial geometry. For example, simulations included using a) a 2-D squareas the initial geometry with lengths for the sides set to 20 µm and b) a3-D cylinder with a height and radius of 20 µm and 10 µm, respectively. Inthese two cases no interactions with a stationary surface were considered.Both the 2-D square model and the 3-D cylinder model resulted in 0.5 µmmovement in 8 h when a viscosity of 1× 109 Pa s was used, suggesting thequarter-sphere model used herein provides the most conservative lower limitto the viscosity of these models.342.3. Equilibration times2.3 Equilibration timesFor experiments using both the bead-mobility and poke-and-flow techniques,the RH inside the flow cell was initially set at 90 % for 30 min, after which theRH was decreased (at a rate of <0.5 % RH min−1) to the experimental value,at which point the RH was held constant to allow the particles to equilibrateonce more. To estimate the time for particles to come to equilibrium withthe water vapour the procedure outlined by Shiraiwa et al. (2011a) wasfollowed, which is based on percolation theory. Equilibration times for waterare expected to be significantly faster than equilibration times for largerorganic molecules because small molecules like water percolate (i.e., diffuse)more rapidly though the host matrix (Shiraiwa et al., 2011a).When calculating equilibration times, diffusion coefficients from Table 1in Shiraiwa et al. (2011a) and the particle diameters used in the poke-and-flow experiments were used with the equation,τmixing =d2p4pi2D(2.6), where dp is the particle diameter, D is the diffusion coefficient, andτmixing is the time taken for the concentration of a a representative moleculeanywhere in the particle to deviate by less than 1/e from the initial dise-quilibrium concentration. The particle size studied of 20-70 µm was usedin the poke-and-flow experiments, although this full range was not used ateach RH. Using this methodology, an estimate of the equilibration time wasdetermined to be <1 s for particles with a viscosity of 10−3 Pa s (roughlyequivalent to that of water), <1 minute for particles with a viscosity of 102Pa s (roughly equivalent to that of ketchup), and <20 minutes for particleswith a viscosity of 1012 Pa s (greater than that of tar pitch).Particles were allowed to equilibrate at a given RH for at least 30 minutesat all RHs for experiments performed using both the bead-mobility and poke-and-flow techniques. For poke-and-flow experiments the equilibration timewas extended as RH was decreased, reaching 120 minutes in some cases.As such, the equilibration times suggest the particles were in equilibrium352.3. Equilibration timeswith the surrounding water vapour when they were studied. Even if theequilibration time is longer than predicted by percolation theory (Boneset al., 2012), it is unlikely that the particles are far from equilibrium in thepoke-and-flow experiments considering both the slow RH ramp down timeand long equilibration times used in the experiments. Further, no strongdependence of the results on particle size was observed, suggesting thatnon-equilibration between the gas and the particle was not an issue.36Chapter 3Viscosity of α-pinenesecondary organic materialand implications for particlegrowth and reactivity3.1 IntroductionBiological sources (e.g., vegetation) emit copious quantities of volatile or-ganic compounds, such as α-pinene (Hallquist et al., 2009; Kanakidou et al.,2005). In the atmosphere, a complex series of chemical reactions oxidizesthese volatile compounds, to form semivolatile organic compounds (SVOCs)that condense to the particle phase (Hallquist et al., 2009; Kanakidou et al.,2005). These particles can influence climate by scattering and absorbingsolar radiation (direct climate effect) and by serving as nuclei for cloud for-mation (indirect climate effect), among other mechanisms (Solomon et al.,2007). They can also influence air quality and health (Baltensperger et al.,2008; Jang et al., 2006; US Environmental Protection Agency, 2008).Recently, molecular diffusion within SOM particles has become an areaof intense scientific interest (see Chapter 1). In many large-scale modellingstudies, it is often assumed that equilibrium is rapidly achieved between gas-phase organic compounds and the bulk of SOM particles (Hallquist et al.,2009; Kamens et al., 1999; Odum et al., 1994; Rounds and Pankow, 1990).This assumption implies that diffusion within the particles is fast whencompared to accommodation (Shiraiwa and Seinfeld, 2012). More recently,373.1. Introductionstrong evidence emerged that some SOM particles, including but not lim-ited to SOM from α-pinene ozonolysis, can behave as semisolids or solidsunder some conditions, such as low relative humidity (Abramson et al.,2013; Cappa and Wilson, 2011; Kuwata and Martin, 2012; Perraud et al.,2012; Saukko et al., 2012; Vaden et al., 2010, 2011; Virtanen et al., 2010,2011). Prior to this research, however, estimates available in the literaturefor molecular diffusivity inside SOM are limited to dry conditions (RH <3%)(Cappa and Wilson, 2011; Perraud et al., 2012) whereas atmospheric con-ditions may cover the full range of RH, with typical values ranging fromapproximately 20-100 % in the planetary boundary layer (Hamed et al.,2011; Held and Soden, 2000; Martin, 2000). As a result, diffusion rateswithin SOM particles under typical atmospheric conditions are uncertain,and this uncertainty implies concomitant uncertainty for predicting the im-pacts of SOM particles on air quality, visibility and climate (Koop et al.,2011).Our strategy for quantification of molecular diffusivity is to make ex-perimental measurements of viscosity, η. Viscosity and molecular diffusionrates are related by the Stokes-Einstein equation in the case of self-diffusion(i.e., similarly sized molecules) if the viscosity is not too near that of a glass,or by other approaches such as percolation theory for a diffusion of a smallmolecule in host matrix of large molecules (Koop et al., 2011; Zobrist et al.,2008). As is the case for molecular diffusivity, quantitative determinationof SOM viscosity is lacking for the range of relative humidity typical of theatmosphere.Herein, two newly developed techniques are used to measure at roomtemperature (293-295 K) the RH-dependent viscosity of the water-solublecomponent of SOM particles produced by α-pinene ozonolysis, a majorsource of SOM particles in the atmosphere, especially over boreal forests(Cavalli et al., 2006). The water-soluble component makes up the majorfraction of SOM particles from α-pinene ozonolysis (see Section A.1).383.2. Experimental3.2 Experimental3.2.1 Production of secondary organic materialSecondary organic material was formed via homogeneous nucleation duringthe dark ozonolysis of α-pinene in continuous-flow environmental chambersat Harvard University and Pacific Northwest National Laboratory (PNNL).A schematic detailing the steps during the production of SOM is includedas Figure 3.1, and the setup and experimental conditions were similar tothose employed by Shilling et al. (2008). The α-pinene (80-100 parts-per-billion by volume (ppbv)), ozone (300 ppbv), and 2-butanol (used as an OHscavenger) were introduced continuously to the chambers in a total flow of20 to 25 sLpm. The temperature and relative humidity inside the chamberswere maintained at 298 K and <5 %, respectively. The secondary organicmaterial was collected at the outlet of the Harvard University environmentalchamber on a quartz fiber filter (Whatman; 1851-047) and at the outlet ofthe PNNL chamber on a Teflon filter (Pall; R2PL037). During collectionthe composition of the SOM was monitored continuously and found to beconstant. The collection time for sampling was 48 h at a flow rate of 8.0sLpm.After collection, filters were stored at 263 K under dry conditions untilthey were extracted. The water-soluble species were extracted from the fil-ters through the addition of 20 mL of ultra-pure Millipore water (18.2 MΩcm). The extract solution was used within 14 days, and stored at 278 Kwhen not in use. Viscosities measured immediately after filter extractionFigure 3.1: Schematic detailing the production of SOM via the ozonolysisof α-pinene.393.2. Experimentaland 14 days after filter extraction were the same within experimental un-certainty. Super-micron sized particles were produced for the bead-mobilityand poke-and-flow experiments in this Chapter by nebulising the extractedSOM solution.3.2.2 Bead-mobility techniqueThe bead mobility technique is described in Section 2.1. The bead speedsdetermined here for the SOM were converted to viscosities using the cali-bration curve shown in Fig. 2.5.3.2.3 Poke-and-flow techniqueThe experimental procedure for experiments with the poke-and-flow tech-nique combined with simulations of fluid flow is detailed in Section 2.2.1.3.2.4 Simulations of material flow at 40-70 % RHBetween 40-70 % RH the needle penetrated the particle and caused thematerial to take on a half-torus geometry. Upon removal of the needle thehole at the centre of the half-torus was observed to decrease in size as thematerial flowed to re-attain a hemispherical geometry (see Fig. 2.7(a-c)for examples). This behaviour was simulated to determine upper limits ofviscosity as detailed in Section 2.2.3. Lower limits of viscosity were notdetermined.For the initial conditions a half-torus with fixed dimensions (R0 = 20 µm,r0 = 7.5 µm) was used, representative of the size of the particles studied.A surface tension of 75 mN m−1 was assigned to Interface 1 of Fig. 2.8,whilst a particle-substrate contact angle of 90 ◦ was assigned to Interface2 (see below for implications). Interactions between the fluid and the solidsurface were described with a Navier slip wall boundary condition, with aslip length of more than an order of magnitude greater than the particleheight. The effect of the slip length on the simulations is discussed below.For the fluid a density of 1.3 g cm−3 was assumed (Chen and Hopke, 2009;Nga et al., 2006; Saathoff et al., 2003).403.2. ExperimentalSimulations were performed as a function of viscosity to create a calibra-tion curve between viscosity and τ (model,flow) (Figure 3.2). This calibrationcurve was then used to convert values of τ (exp,flow) into viscosities.Viscosities determined as described above are upper limits for the follow-ing reasons: (a) The calibration curve was generated with a surface tensionof 75 mN m−1, which is most likely an upper limit to the surface tensionsin our experiments (Huff-Hartz et al., 2006; Hyva¨rinen et al., 2006; Tuck-ermann and Cammenga, 2004). If lower surface tensions were used in thesimulations the results would be lower predictions of viscosities for a givenτ (exp,flow). (b) The calibration curve was generated using a contact angle of90 ◦, which is greater than the upper limit to the contact angle measured100101102103104105106107108109101010-410-310-210-1100101102103104105106 90 o contact angle, 75 mN/m surface tension 70 o contact angle, 60 mN/m surface tensiondecreasing slip-lengthdecreasing surface tensiondecreasing equilibrium contact angleflow impeded due to surface scratchestmodel, flow (seconds)Viscosity (Pa s)1Figure 3.2: Calibration line from COMSOL simulations (black solid line).In the simulations used to generate this calibration a surface tension of75 mN m−1 and contact angle of 90 ◦ were used. The annotation in thefigure illustrates how the calibration line would shift if a smaller surfacetension, smaller equilibrium contact angle, smaller slip length, and/or sur-face scratching was included in the model. An example line constructed withlower surface tension (60 mN m−1) and lower contact angle (70 ◦) is given todisplay the sensitivity of the model to these parameters (grey dashed line).Originally published in Renbaum-Wolff et al. (2013a).413.2. Experimentalwith a confocal microscope. If a smaller contact angle is used in the simula-tions, the result would be lower viscosity predictions for a given τ (exp,flow).(c) A large slip length was used in the calculations (more than an order ofmagnitude greater than the height of the modelled particles). This is likelyan upper limit to the slip length in our experiments (Churaev et al., 1984).If a smaller slip length was used, the result would be lower predictions of vis-cosities for a given τ (exp,flow). (d) In some experiments the needle scratchesthe surface resulting in reduced flows of the particle on the scratched re-gions (observed visually). If this process was included in the model, thenthe result would be lower predictions of viscosity for a give τ (exp,flow). Seeannotations in Fig. 3.2 for the effect of decreased surface tension, decreasingcontact angle, decreasing slip length and effect of surface scratching on thecalibration curve that relates flow times to viscosities. The sensitivity of themodel to changes in surface tension and contact angle was tested by com-paring the τ (model,flow) values calculated for a contact angle of 90◦ and asurface tension of 75 mN m−1 (Fig. 3.2, solid line) to τ (model,flow) calculatedfor a contact angle of 70 ◦ and a surface tension of 60 mN m−1 (Fig. 3.2,dashed line). The small (20 - 30 %) change in the predicted τ (model,flow)values are predominantly due to the change in the surface tension.The upper limits to viscosity at 70 % RH estimated from the poke-and-flow results are consistent with the results obtained with the bead-mobilitytechnique (compare Tables 3.1 and 3.2). As a further test, poke-and-flowexperiments were performed using particles of sucrose. When sucrose parti-cles were poked at 56 and 60 % RH and half-torus geometries were formed,the resulting τ (exp,flow) values were 5 and 0.8 seconds, respectively. Usingthese experimental flow times and the calibration curve in Fig. 3.2, upperlimits to the viscosity of the sucrose particles were estimated to be 1× 104and 6 × 104 Pa s, at 60 % and 56 % RH respectively. At 60 % RH, theliterature viscosity of the sucrose/water mixture is approximately 9 × 102Pa s. At 56 % RH, the viscosity of the sucrose/water mixture is estimatedto be 8(+3−2)× 103 Pa s, based on an extrapolation of literature data using athird degree polynomial function (see Renbaum-Wolff et al., 2013b).423.3. Results3.2.5 Simulations of material flow at 25-30 % RHAt 25 and 30 % RH the particles were observed to crack upon impactionof the needle, forming pieces with sharp, distinct edges. No detectable flowwas discerned in the material during observations over the subsequent eighthours.Flow in these experiments was simulated as detailed in Section 2.2.4.To test the approach for determining lower limits to viscosities, particles ofraffinose were studied at 30 % RH using the poke-and-flow technique. Ac-cording to Zobrist et al. (2008), raffinose passes through a relative humidityinduced glass transition at ≈53 % RH at room temperature. At 30 % RH,therefore, raffinose particles are expected to be in a glass state, with viscosi-ties ≥1012 Pa s. The poke-and-flow experiment was performed on raffinoseparticles at 30 % RH and the particles shattered into well-defined pieces withsharp edges. Movement was less than 0.5 µm over 8 h. The lower limit ofthe viscosity estimated with the quarter-sphere COMSOL model discussedin Section 2.2.4 was 5 × 108 Pa s, which is consistent with the establishedviscosity (≥1012 Pa s) for this material.3.3 Results3.3.1 Experiments with the bead-mobility techniqueShown in Fig. 2.4 are examples of optical images showing the change overtime in the position of beads within a particle. The average bead speedsand associated viscosities determined between 70 % and 90 % RH are sum-marised in Table 3.1 and plotted in Figure 3.3. At 70 % RH the viscosity(±95 % prediction limits) of the SOM was 791+1520−413 Pa s, comparable to thatof peanut butter. At 90 % RH the viscosity of the SOM was 6.25+11.8−3.35 Pa s,comparable to that of honey. Bead speeds were not determined for RH <70% because the rate of circulation became too slow to readily observe.433.3. ResultsTable 3.1: Mean bead speeds as a function of RH in water-solubleSOM from α-pinene ozonolysis and corresponding viscosities. Origi-nally published in Renbaum-Wolff et al. (2013a).ChamberRH / %Mean bead speed Viscosity (± 95 %sample a / µm ms−1 prediction limits) / Pa sHEC90 4.75× 10−5 6.25+11.8−3.3580 5.33× 10−6 61.8+117−32.770 4.67× 10−7 791+1520−413PNNL87 1.90× 10−5 16.3+30.7−8.6783 9.22× 10−6 34.8+66.7−18.579 3.90× 10−6 85.7+163−45.375 1.40× 10−6 251+481−13271 7.41× 10−7 488+938−256a HEC refers to samples collected on quartz fiber filters from the Harvard Envi-ronmental Chamber. PNNL refers to samples collected on Teflon filters fromthe Pacific Northwest National Laboratory Continuous-Flow EnvironmentalChamber.3.3.2 Experiments with the poke-and-flow techniqueIn experiments at RH ≥40 %, the needle penetrated the particle, resultingin a half-torus shape being generated after the needle was retracted (e.g.Fig. 2.7(a)-(c)). The half-torus reformed at an observable rate into a spher-ical cap, minimizing the surface energy of the system (e.g. Fig. 2.7(a)-(c)).The experimental flow time, τ (exp,flow), required for the inner diameter ofthe half-torus to decrease to 50 % of its initial diameter was determined byanalysis of a time series of images. The flow time increased from approxi-mately 10 s at 70 % RH to 4000 s at 40 % RH (Table 3.2).443.3. ResultsTable 3.2: Results from the poke-and-flow experiments. Orig-inally published in Renbaum-Wolff et al. (2013a).ChamberRH / %τ (exp,flow) Limits of viscosity / Pa ssample a / s b Lower c Upper dHEC 70 50.9 378 5.63× 105PNNL68 8.98 378 9.92× 10460 116.5 378 1.29× 10655 46.0 378 5.08× 10550 47.0 378 5.19× 10545 305 378 3.37× 10640 4090 378 4.53× 107a HEC refers to samples collected on quartz fiber filters from the HarvardEnvironmental Chamber. PNNL refers to samples collected on Teflonfilters from the Pacific Northwest National Laboratory Continuous-Flow Environmental Chamber.b The values of τ (exp,flow) can vary between experiments but the slowestobserved flows are reported here in order to calculate an upper limitto the viscosity.c A lower limit (378 Pa s) was established for 40-70% RH based on thelower 95 % prediction limit from the bead-mobility value at 70 % RH.d The upper limits to the viscosities were calculated using τ (exp,flow),and the calibration curve shown in Fig. 3.2.Simulations were carried out using the software package COMSOL Mul-tiphysics (see Section 3.2.4) for a particle of half-torus geometry havingsimilar physical dimensions as the experimental observations. Because ofassumptions used in the simulations, the viscosities determined for 40-70 %RH were upper limits (see Section 3.2.4).A lower limit (378 Pa s) was also established for 40-70 % RH based onthe lower 95 % prediction limit from the bead-mobility value at 70 % RH.The upper limits (based on the poke-and-flow technique) and lower limits(based on the bead-mobility technique) to the viscosities for 40-70 % RHare listed in Table 3.2 and plotted in Fig. 3.3(a). At 40 % RH, the limitsrange from 378 Pa s to 5 × 107 Pa s, i.e., from approximately the viscosity453.3. Results                                                                      honey  ketchup  peanut    butter  pitch  glass  A  Environmental    chamber  Planetary  boundary  layer  Typical  conditions  used    Typical  conditions  observed  B  1Figure 3.3: (a) Summary plot of the SOM viscosities determined by a com-bination of experiments using the bead-mobility technique (black emptysquares and triangles for HEC and PNNL samples, respectively, where theblack bars represent the 95 % prediction intervals) and experiments usingthe poke-and-flow technique (where the blue bars represent the bounds ofthe viscosities). HEC refers to samples collected on quartz fiber filters fromthe Harvard Environmental Chamber. PNNL refers to samples collected onTeflon filters from the Pacific Northwest National Laboratory Continuous-Flow Environmental Chamber. Various common substances have beenplaced alongside the diagram, along with their approximate viscosities atroom temperature, to provide points of reference following the idea of Koopet al. (2011). The secondary y-axes show (1) diffusion coefficients calculatedusing the Stokes-Einstein relation and (2) mixing times (τmixing) of the par-ticles due to bulk diffusion in 100 nm particles of the same viscosity (seemain text). The image of the pitch is a detail of an image from the pitchdrop experiment (Wikimedia Commons, GFDL, University of Queensland,Australia, John Mainstone). (b) Typical relative humidities observed in theplanetary boundary layer (Hamed et al., 2011; Held and Soden, 2000; Mar-tin, 2000) and environmental chambers (Kostenidou et al., 2009; Tillmannet al., 2010). Originally published in Renbaum-Wolff et al. (2013a).463.4. Discussionof peanut butter to that of pitch (Edgeworth et al., 2001).For RH values of 25 to 30 %, the particles shattered when poked witha needle. Moreover, restorative flow did not occur, at least for the exper-imental timescale (8-10 h) (e.g. Fig. 2.7(d)). The fragments had sharpwell-defined edges, and no smoothing of the edges was observed over thecourse of the experiments. In this case, modelling of the flow established alower limit to the viscosity. This lower limit is included in Fig. 3.3(a), witha value comparable to that of pitch (Edgeworth et al., 2001).3.4 DiscussionAmorphous solids have viscosities greater than 1012 Pa s, semi-solids suchas gels or ultra-viscous liquids have viscosities between 102 and 1012 Pa s,and liquids have viscosities less than or equal to 102 Pa s (Koop et al., 2011;Shiraiwa et al., 2011a). These phases are represented by different patternsin Fig. 3.3, together with the RH-dependent viscosities. The viscosities ofthe studied SOM correspond to liquid for RH ≥80 %, a semisolid for 40 %≤ RH < 80%, and a semisolid or solid for RH ≤30 %. These findings arein agreement with the results of other recent studies suggesting that certaintypes of SOM do not behave as liquids under some conditions, such as lowrelative humidity (Abramson et al., 2013; Cappa and Wilson, 2011; Kuwataand Martin, 2012; Perraud et al., 2012; Saukko et al., 2012; Vaden et al.,2010, 2011; Virtanen et al., 2010, 2011).The relationship between RH and viscosity in Fig. 3.3(a) can be ratio-nalized by considering the hygroscopic nature of the studied SOM. SOMfrom α-pinene ozonolysis is known to be hygroscopic, meaning the watercontent of the particles will increase as the RH increases (Cocker IIII et al.,2001; Saathoff et al., 2003; Varutbangkul et al., 2006; Virkkula et al., 1999).As the particles uptake water, the viscosity of the mixture is thus expectedto decrease, since the viscosity of pure water is low (1.002 ×10−3 Pa s at293 K (Kestin et al., 1978). This phenomenon may be described, at least toa first order approximation, by simple mixing rules (see Section A.2). Theapparent step in viscosity between 30-40 % RH may be due to a phase tran-473.4. Discussionsition, such as the formation of a glass or gel, or may be due to non-idealinteractions in the complex mixture (see Section A.2).Using the viscosities determined herein and an estimated hydrodynamicradius, the corresponding Stokes-Einstein-equivalent diffusion coefficients oforganic molecules in SOM, Dorg, have been calculated. These values corre-spond to the secondary y-axis of Fig. 3.3(a). The hydrodynamic radius wasapproximated as the molecular radius of 0.38 nm for a molar mass of 175g mol−1 (Cocker IIII et al., 2001), a density of 1.3 g cm−3 (Saathoff et al.,2003; Varutbangkul et al., 2006; Virkkula et al., 1999), and molecular spher-ical symmetry. The obtained Dorg values ranged from 10−9 to 10−11 cm2s−1 between 90 % and 70 % RH, from 10−11 to 10−16 cm2 s−1 between 70% and 40 % RH, and to <10−17 cm2 s−1 for RH ≤30 %. Three studies haveestimated the diffusivity of organic compounds within SOM. Two of thesestudies, one looking at partitioning of organic nitrates into SOM particles(Perraud et al., 2012) and the other looking at the evolution of the chemi-cal composition of SOM particles after they pass through a thermodenuder(Cappa and Wilson, 2011), have estimated an upper bound of Dorg <10−14cm2 s−1 for the diffusion coefficient of organic molecules in SOM producedby α-pinene ozonolysis under dry conditions (<3 % RH). These results areconsistent with the results reported herein for RH ≤30 % (Dorg <10−17 cm2s−1). The third study (Abramson et al., 2013) quantified the diffusivity ofpyrene in SOM under low RH conditions to be 2.5× 10−17 cm2 s−1, similarto, but slightly higher than, the diffusion coefficient measured herein for thewater-soluble component of α-pinene at ≤30 % RH (Dorg <10−17 cm2 s−1).The small discrepancy is largely attributed to the hydrodynamic radius ofpyrene used in the study by Abramson et al. (2013) compared to the hy-drodynamic radius of SOM material used to calculate molecular diffusioncoefficients from our viscosity data.Detailed modelling studies are required to fully explore the implicationsof the new viscosity data. Here the viscosity data is used together withprevious modelling studies or simple calculations to provide an initial as-sessment of the effect of the new viscosity data on (1) the mechanism ofgrowth of SOM particles, (2) predictions of particle mass, and (3) rates of483.4. Discussionreactions in SOM particles.Currently most models used for predicting the effect of SOM particles onair quality and climate assume that growth occurs by instantaneous equilib-rium partitioning of SVOCs into the bulk of the SOM particles (Fig. 1.7(a),top image). For this mechanism to be accurate the mixing time of SVOCswithin the particles must be short compared to the times scale for particlegrowth. The mixing times, τmixing, by diffusion of large organic moleculeswithin an SOM particle can be estimated from the viscosity data and theStokes-Einstein relation (Bones et al., 2012),τmixing =3d2ηr2pikT(3.1), in which d is the particle diameter, r is the hydrodynamic radius of a rep-resentative molecule of SOM within the SOM bulk matrix, k is Boltzmannsconstant, and T is temperature. After the mixing time, the concentrationof the representative molecule anywhere in the particles deviates by lessthan 1/e from the initial disequilibrium concentration (i.e., a homogeniza-tion process). It should be noted that Equation 3.1 may be inaccurate if theviscosity is close to that of a glass (Champion et al., 1997).Shown in Fig. 3.3(a) (right-hand y-axis) are mixing times within 100 nmparticles, calculated with Equation 3.1, the new viscosity data, and assuminga hydrodynamic radius of 0.38 nm. For an RH of 70 to 90 %, the mixingtimes are 0.01-1 s. In this case instantaneous equilibrium partitioning withinthe bulk of the SOM particles is likely a valid description of SOM growth(Fig. 1.7(a), top image). By comparison, the mixing times exceed 2.5 daysfor RH ≤30 %. At these relative humidities, gas-particle partitioning oflarge organic molecules such as pinonaldehyde (the most abundant productof α-pinene ozonolysis) may be effectively confined to the surface of theparticle (Fig. 1.7(a), bottom image). Hence, the mechanism of growthmay not occur by instantaneous equilibrium partitioning within the particlebulk, as suggested for α-pinene SOM studied under dry conditions (e.g.Perraud et al., 2012, and references therein). The mechanism of growthmay have important consequences for the particle size distribution (Riipinen493.4. Discussionet al., 2011), which may in turn affect the scattering and absorption of solarradiation and the ability of particles to act as nuclei for cloud condensation(Solomon et al., 2007).At ≤30 % RH, because the viscosity of the particles may allow parti-tioning of SVOCs only within the top few molecular layers of the particle,the particles will take up less SVOCs compared to low viscosity particleswhere instantaneous equilibrium partitioning within the bulk of SOM par-ticles may occur (Fig. 1.7(a)). As a result, models that assume equilibriumpartitioning at ≤30 % RH, may over predict SOM particle mass and underpredict gas-phase concentrations of SVOCs. Recently Shiraiwa and Sein-feld (2012) showed that when the viscosity of the particle is ≥5 × 106 Pas, the particle-phase mass concentrations of semivolatile and low-volatilityorganic compounds may be overestimated by at least an order of magni-tude compared to the assumption of instantaneous equilibrium partitioningwithin the bulk. The new viscosity data and the modelling results fromShiraiwa and Seinfeld (2012), suggest that at ≤30 % RH, the particle-massconcentrations of SVOCs and low-volatility organic compounds may be overpredicted by an order of magnitude with implications for predictions of airquality and visibility.Reactions between atmospheric oxidants, such as O3, and particle-phaseorganic molecules can lead to chemical aging of SOM particles, with possi-ble implications for particle hygroscopicity, optical properties, and toxicity(Shiraiwa et al., 2011b; Zahardis and Petrucci, 2007). The rates of these ag-ing reactions may depend on the diffusion coefficients of the oxidants, Dox,within the SOM (Pfrang et al., 2011; Shiraiwa et al., 2011a; Smith et al.,2002). The rates can also possibly depend on the diffusion coefficients of theorganic molecules, Dorg, in the case that the diffusion rates of the organicmolecules are slow (Pfrang et al., 2011; Shiraiwa et al., 2011a). Based onour viscosity data and the Stokes-Einstein equation, Dorg decreases by morethan eight orders of magnitude for a drop in RH from 90 % to 30 % RH.In organic matrices Dox is estimated based on previous measurements ofthe mobility of small molecules in different matrices to be ≈10−5 cm2 s−1for a liquid, 10−7 - 10−9 cm2 s−1 for a semi-solid, and 10−10 cm2 s−1 for a503.4. Discussionsolid (see Shiraiwa et al. (2011a) and references therein). Based on theseestimates and the phase information elucidated above, Dox is expected todecrease by at least two to five orders of magnitude for a change in RH from90 % to ≤30 % RH as the particles transition from liquid to semisolid/solid.These changes in diffusion coefficients imply that the rates of chemical ag-ing of SOM particles can depend strongly on RH, with a decrease in ratesexpected with a decrease in RH. In support of these implications, a recentstudy showed that the reaction rate of NH3 with α-pinene SOM particles de-creases significantly for a drop in RH from 95 % to 5% (Kuwata and Martin,2012).Applying the viscosity measurements from the present study to α-pineneSOM in the atmosphere or environmental chambers is subject to severalcaveats. First, the study focused on the water-soluble component of SOM.For α-pinene SOM, however, the water-soluble material constitutes the ma-jor mass fraction SOM (see Section A.1). Second, collection of SOM fromenvironmental chambers can be subject to both positive and negative sam-pling artifacts. Positive artifacts occur due to the adsorption of semivolatileorganic material on the filters and are expected to be much less with Teflonfilters than quartz fiber filters (Kirchstetter et al., 2001). Since the sameviscosity results were obtained within the uncertainty of the measurementsusing both Teflon filters and quartz filters (Tables 3.1 and 3.2, and Fig.3.3), positive artifacts seem unimportant in our experiments. Negative arti-facts can include the partial evapouration of semivolatile SOM. This processcannot be ruled out for our experiments. As a result the material studiedhere may be the less volatile component of the water-soluble secondary or-ganic material. These caveats notwithstanding, the results reported hereinprovide best estimates of α-pinene SOM viscosities over a wide range of at-mospherically relevant relative humidities. Information concerning viscosityand ultimately diffusion coefficients is needed for the accurate modelling ofheterogeneous chemistry as well as particle growth and evapouration.513.5. Summary3.5 SummaryTwo techniques, the bead-mobility technique and the poke-and-flow tech-nique combined with simulations of fluid flow have been used to experi-mentally determine the viscosity of particles containing the water-solublecomponent of SOM at a range of RHs.Firstly, the bead-mobility technique, which was initially described andvalidated previously (Renbaum-Wolff et al., 2013b), was used to determinethe viscosity of SOM particles at RHs ≥70 %. At ≥80 % RH the particleswere determined to have viscosities corresponding to that of a liquid, whilstat RHs between 70 and 80 % RH the particles were determined to haveviscosities corresponding to that of a semisolid (Koop et al., 2011; Shiraiwaet al., 2011a).Secondly a poke-and-flow technique was used to determine the viscosityof SOM particles at ≤70 % RH. At RHs between 40 % and 70 %, thematerial exhibited flow on an atmospheric timescale, whilst at RHs ≤ 30% the material was observed to crack, and no flow was observed over thesubsequent 8 h. Simulations of this behaviour suggest that the material isof viscosity corresponding to that of a semisolid between 40 and 70 % RH,and of viscosity corresponding to that of a semisoid or solid at ≤30 % RH(Koop et al., 2011; Shiraiwa et al., 2011a).These experimental findings are in agreement with the results of otherrecent studies suggesting that certain types of SOM do not behave as liquidsunder some conditions, such as low relative humidity (Abramson et al., 2013;Cappa and Wilson, 2011; Kuwata and Martin, 2012; Perraud et al., 2012;Saukko et al., 2012; Vaden et al., 2010, 2011; Virtanen et al., 2010, 2011).These results have important atmospheric implications. For example,if particles in the atmosphere are similar to those studied here then theirmixing times would be rapid (on the order of seconds) for particles at ≥70 %RH, but become increasingly longer as RH is reduced, being on the order ofdays at ≤30 % RH, affecting the mechanism and rate of growth of the par-ticles. Based on the viscosities and the phase information elucidated above,and for particles similar to those studied here, Dox is expected to decrease523.5. Summaryby at least two to five orders of magnitude for a change in RH from 90 % to≤30 % RH, as particles transition from liquid to semisolid/solid in phase,a direction of change in agreement with that suggested by prior research(Kuwata and Martin, 2012). These changes in diffusion coefficients implythat the rates of chemical aging of SOM particles can depend strongly onRH, with possible implications for particle hygroscopicity, optical properties,and toxicity (Shiraiwa et al., 2011b; Zahardis and Petrucci, 2007).53Chapter 4Additional validation of thepoke-and-flow techniquecombined with simulations offluid flow for determiningviscosities in samples withsmall volumes and highviscosities4.1 IntroductionThe importance of viscosity and diffusion within SOM particles is outlined inChapter 1. The viscosity, η, of SOM may span multiple orders of magnitude,from 10−3 to >1012 Pa s, across the ambient relative humidity range in theatmosphere through the uptake and release of water (Koop et al., 2011;Kuwata and Martin, 2012, and Chapter 3). Measuring such a wide range ofviscosities presents a challenge, made more difficult by the small, milligram,scale of SOM samples typically collected in the atmosphere or chambers usedto simulate atmospheric conditions. Currently, there is no commerciallyavailable technique capable of quantifying the viscosity of SOM samplesacross the entire viscosity range important in the atmosphere. However, a544.1. Introductionfew techniques have recently been developed, each of which is capable ofcovering at least part of the range of interest.Renbaum-Wolff et al. (2013b) developed a bead-mobility technique, whichcan determine the viscosities of SOM samples with masses between 1-5 mgand viscosities between 10−3 and 103 Pa s. This technique consists of deter-mining the speed of circulation of micrometer sized beads within a particleas a shear stress is applied to the particle. In Chapter 2 the development ofa poke-and-flow technique combined with simulations of fluid flow to con-strain the viscosities of samples of 1-5 mg in mass is detailed. The techniqueconsisted of generating a hole in a supermicron sized particle suspended ona surface and determining a characteristic time taken for the hole to close.In Chapter 3 simulations of fluid flow were performed to determine upperlimits of viscosity for a particle based upon the time taken for the hole at itscentre to close, and the measured upper limits of viscosity were consistentwith literature values up to at least 104 Pa s.Power et al. (2013) used holographic optical tweezers to coalesce twosuspended particles with a combined volume of <500 femtolitres. By mea-suring the time taken for the resulting particle to relax to a spherical shape,viscosities of sucrose-water or sucrose-salt-water particles were quantifiedacross the range of 10−3-109 Pa s. In a subsequent publication, Power andReid (2014) further outlined the application of optical tweezers for rheolog-ical measurements. In a similar vein, Pajunoja et al. (2014) used scanningelectron microscopy images to determine the viscosity of secondary organicaerosols by studying the time taken for multiple particles to coalesce. Hosnyet al. (2013) observed the behaviour of molecular rotors using fluorescencelifetime imaging microscopy to determine the viscosity of sodium chlorideand sucrose-water particles. In addition, Kidd et al. (2014) estimated somelimits to viscosities of particles from the extent to which material collectedin the centerline of an impactor spreads under high airflow.In Chapter 3, only a preliminary validation of the poke-and-flow tech-nique combined with simulations of fluid flow was carried out for viscosities<108 Pa s due to the lack of suitable standards for validation at the time ofpublication. Specifically, in Chapter 3, sucrose-water particles over a narrow554.2. Experimentalrange of relative humidities to validate the approach for viscosities <108 Pas. In addition, as mentioned above, for viscosities <108 Pa s, the work inChapter 3 only showed that the approach was able to provide upper limitsto the viscosity of the particles. No attempt was made to determine lowerlimits to the particle viscosity using the poke-and-flow technique combinedwith simulations of fluid flow when the viscosity was <108 Pa s.The following expands on the initial validation and characterization ofthe poke-and-flow technique combined with simulations of fluid flow. First,the approach is used to determine the viscosity of sucrose-water particlesover a wider range of relative humidities than previously done in Chapter3. These results are compared to recent results published by Power et al.(2013) who reported viscosities of sucrose-water particles ranging from 10−3- 109 Pa s, and Quintas et al. (2006) who measured a viscosity of 103 Pas at 54 % RH using a rotational controlled stress rheometer. Second, theapproach was used to determine the viscosity of two polybutene standards,and the results compared with viscosities measured by the manufacturerusing a commercially available viscometer. The results for both the sucrose-water particles and the polybutene standards shows this approach is capableof providing both lower and upper limits of viscosity that are consistent withliterature or measured values for particles of material that range in viscosityfrom ≈5 x 102 - ≈3 x 106 Pa s.4.2 Experimental4.2.1 Poke-and-flow techniqueThe poke-and-flow technique combined with simulations of fluid flow is de-tailed in Section 2.2. Super-micron sized sucrose particles were produced onhydrophobic glass substrates by nebulising sucrose-water solutions. Super-micron sized particles consisting of polybutene standards (N450000 andN2700000; Cannon Instrument Company, USA) were prepared on hydropho-bic glass substrates using a pipette. The samples were heated to 60 - 70 ◦Cover ≈30 minutes to aid the production of particles by reducing the vis-564.2. Experimentalcosity of the material, with material being picked up on a pipette tip andthe pipette being ’flicked’ towards the substrate, resulting in particles be-ing formed on the substrate. After particle production, the particles wereallowed to equilibrate under dry (<0.5 % RH) conditions at room tempera-ture for 60 minutes. Particles consisting of polybutene standards that werepoked ranged in diameter from 40-70 µm.4.2.2 Simulations of fluid flowSimulations of fluid flow were performed as detailed in Section 2.2.3.The physical properties used in the simulations are detailed in Tables4.1 (for particles of sucrose-water) and 4.2 (for particles of polybutene stan-dards). The dimensions of R0 and r0 and the equilibrium contact angle weredetermined from measurements, and the values of surface tension (Surface1, Fig. 2.8(b)), slip length (which describes the interaction at Surface 2in Fig. 2.8(b)), and the density of the material were determined based onliterature values. For each particle that was poked and formed a half-torusgeometry, and was not significantly influenced by scratches (determined vi-sually, as discussed in Section 2.2), lower and upper limits of viscosity weredetermined via simulations. Values from row 2 of Tables 4.1 and 4.2 wereused for simulations of the lower limit of viscosity for a particle, whilst val-ues from Row 3 of Tables 4.1 and 4.2 were used for simulations of its upperlimit of viscosity.A proportion (≈30 %) of the sucrose-water particles that were poked haddimensions where r0/R0 > ≈0.4. Simulations of the lower limit of viscosityfor many of these particles gave rise to a non-physical geometry whereby theinner edge of the half-torus geometry appeared jagged and wavy - a phe-nomena not observed during experiments. Further study of this phenomenarevealed that this was the result of the stretching of the mesh elements atthe moving front of the particle, and use of a finer mesh did not prevent thisfrom occurring. As such, all sucrose-water particles of dimensions r0/R0 >≈0.4 were removed from the study. No particles of the polybutene standardfell into this size range.574.2.ExperimentalTable 4.1: Experimental parameters used when simulating flow with COMSOL for the sucrose-water experi-ments. Originally published in Grayson et al. (2015a).Surface tension Slip length Density Contact angle(mN m−1) (m) (kg m−3) (degrees)Range of values 57.2 - 75.15 a5 x 10−9 - 1490 -94.8 - 102.2 d1 x 10−5 b 1520 cValue used to calculatelower limit of viscosity57.2 5 x 10−9 150094.8 for particles of (R0−r0)/r0<2102.2 for particles of (R0−r0)/r0>2Value used to calculateupper limit of viscosity75.15 1 x 10−5 1500102.2 for particles of (R0−r0)/r0<294.8 for particles of (R0−r0)/r0>2a MacDonald et al. (1996); Power et al. (2013)b This range is based on the literature values included in Table 2.1.c Tong et al. (2011); Zobrist et al. (2008). A density of 910 kg m−3 was used for all simulations as density was found to haveno effect on simulated viscosities.d Contact angles were determined by photographing a series of five sucrose-water particles, each on a separate hydrophobicslide. The contact angle was measured at the particle-substrate interface of both the right and left edges of the particle usingImageJ software. The mean contact angle was determined to be 98.5 ◦, and the lower and upper limits of contact angle weredetermined to be 94.8 and 102.2 ◦ (98.5 ◦ ± 1 σ). The relationship between the simulated viscosity of a particle and its contactangle is dependent upon the dimensions of the particle, more specifically the value of the ratio (R0−r0)/r0 . The lower limit ofcontact angle gave rise to the lower limit of viscosity for particles where (R0−r0)/r0>2, whilst the upper limit of contact anglegave rise to the upper limit of viscosity for particles where (R0−r0)/r0<2.584.2.ExperimentalTable 4.2: Experimental parameters used when simulating flow with COMSOL for experiments using polybutenestandards. Originally published in Grayson et al. (2015a).Surface SlipDensityContact angle (degrees)tension length(kg m−3)Standard #1 Standard #2(mN m−1) (m) (N450000) (N2700000)Range of values 25 - 50 a5 x 10−9 - 910 -53.6 - 66.4 d 48.8 - 57.4 d1 x 10−5 b 913 c25 5 x 10−9 91053.6 for particles of 48.8 for particles ofValue used to calculate (R0−r0)/r0<2 (R0−r0)/r0<2upper limit of viscosity 66.4 for particles of 57.4 for particles of(R0−r0)/r0>2 (R0−r0)/r0>250 1 x 10−5 91066.4 for particles of 57.4 for particles ofValue used to calculate (R0−r0)/r0<2 (R0−r0)/r0<2upper limit of viscosity 53.6 for particles of 48.8 for particles of(R0−r0)/r0>2 (R0−r0)/r0>2a Five studies have examined the surface tension of the polybutene, with the reported values ranging from 29 - 34.3 mN m−1 (Blunkand Wilkes, 2001; Jeong and Moffatt, 1992; Lewandowski and Dupuis, 1994; Mewis and Metzner, 2006; Roe, 1968). As the polybutenestandards studied here are unlikely to differ much from those studied elsewhere, a conservative value of 25 mN m−1 has been used asthe lower limit of surface tension. Blunk and Wilkes (2001) studied three different polybutene resins of differing viscosities. Measuredsurface tension values suggested a direct, though weak, relationship between surface tension and viscosity (surface tension increasedfrom 29.3 - 30.0 mN m−1 as the viscosity of the resins increased from 4 to 16.4 Pa s). As the resins studied by Blunk and Wilkes(2001) were two orders of magnitude less viscous than those measured herein, a conservative upper estimate of 50 mN m−1 has beenused in simulations for the surface tension of the polybutene standards.b This range is based on the literature values included in Table 2.1.c Measured by Cannon Instrument company. A density of 910 kg m−3 was used for all simulations as density was found to have noeffect on simulated viscosities.d Contact angles were determined as for sucrose-water particles (Table 4.1). For Standard #1 (N450000) the lower and upper limits ofcontact angle were determined to be 53.6 and 66.4 ◦ (i.e. 60.0 ◦ ± 1 σ), whilst the corresponding values for Standard #2 (N2700000)were determined to be 48.8 and 57.4 ◦ (i.e. 53.1 ◦ ± 1 σ).594.3. Results and discussion4.3 Results and discussion4.3.1 Sucrose-water particlesShown in Figure 4.1 are examples of optical images of sucrose-water par-ticles at 48.8, 52.7 and 58.8 % RH recorded during typical poke-and-flowexperiments. Prior to being poked the particles may be described geomet-rically as a spherical cap (Fig. 4.1; Panels a1, b1 and c1). Just after beingpoked the geometry of the particles can be described as a half-torus - a ringof material with a hole at its centre (Fig. 4.1; Panels a2, b2, and c2), whichFigure 4.1: Optical images of sucrose-water particles poked at RHs of (a)48.8, (b) 52.7, and (c) 58.8 % recorded during typical poke-and-flow exper-iments. Images a1, b1, and c1 correspond to the particles before they arepoked. Images a2, b2, and c2 correspond to the first frame post-poke (i.e.the first frame after the needle has been removed). Images a3, b3, and c3correspond to images of the experimental flow time, τ exp,flow, the point atwhich the equivalent area diameter of the hole at the centre of the particlehas decreased to 50 % of its original size. Images a4, b4, and c4 correspondto the final frame recorded, at which point each particle has re-attained itsoriginal spherical cap geometry. Scale bar: 20 µm. Originally published inGrayson et al. (2015a).604.3. Results and discussionis energetically unfavourable compared to that of a spherical cap. For theparticles in Fig. 4.1, τ exp,flow was determined to be 11.25, 3.75, and 1.25seconds at 48.8 %, 52.7 %, and 58.8 % RH, respectively (Fig. 4.1; Panels a3,b3, and c3). Following τ exp,flow the material continued to flow, and eventu-ally re-attained its initial, energetically favourable, spherical cap geometry(Fig. 4.1; Panel a4, b4, and c4).The τ exp,flow values of each of the individual sucrose-water particlespoked and analysed is shown in Figure 4.2(a). Experimental flow times in-creased from ≈150 milliseconds at 59 % RH to ≈40 minutes at 37 % RH. Themillisecond time resolution of the camera precluded experiments being per-formed at RH >60 % as the closure time for the sucrose-water particles wastoo fast to measure. Lower and upper limits of viscosity were determined foreach individual particle using their dimensions and τ exp,flow (Figure 4.2(b)).Between 59 % RH and 36 % RH, the viscosities for individual sucrose-waterparticles range from 1.0 x 101 - 6.6 x 106 Pa s with the upper limit ofviscosity for a given particle being a factor of 16 to a factor of 140 largerthan the corresponding lower limit of viscosity, with the uncertainty mainlydue to the uncertainties of the values of the physical properties used in thesimulations.In Figure 4.2(c), the viscosities of individual particles are grouped by RH,and previously reported values of the viscosity of sucrose-water particles areincluded for comparison (Power et al., 2013; Quintas et al., 2006; Swindellset al., 1958). Viscosities have been determined by grouping particles basedupon RH, with lower and upper limits of viscosity from particles in the groupreported. The lower and upper limits of viscosity of the group of particlespoked at ≈59 % RH are 1.0 x 101 Pa s and 1.6 x 104 Pa s, whilst at 37 % RHthe corresponding values are 7.2 x 104 and 4.7 x 106 Pa s, respectively. Asshown in Fig. 4.2(c), the results are in good agreement with Quintas et al.(2006) who measured a viscosity of 103 Pa s at 54 % RH using a rotationalcontrolled stress rheometer, and Power et al. (2013) who recently reportedmean measured viscosities of ≈5 x 102 Pa s to ≈3 x 106 Pa s between 60 %RH and 37 % RH using holographic optical tweezers.614.3. Results and discussionFigure 4.2: (a) τ exp,flow as a function of RH for individual sucrose-waterparticles. (b) Calculated viscosities for the individual sucrose-water parti-cles in (a), where red bars represent the calculated lower and upper limitsof viscosity. (c) Lower and upper limits of viscosity for the particles shownin (b), grouped by RH. The error bars on the x -axis represent the range ofRHs at which particles in the group were poked. Lower and upper limits ofviscosity were determined for each particle via simulation, with the bottomof a bar on the y-axis representing the lowest lower limit of viscosity for anyof the particles in the group, and the top of the bar representing the highestupper limit of viscosity for any of the particles in the group. Literaturevalues (Power et al., 2013; Quintas et al., 2006; Swindells et al., 1958) areprovided for comparison, with error bars representing 1σ for Power et al.(2013) and 95 % confidence intervals for Quintas et al. (2006). Originallypublished in Grayson et al. (2015a).624.3. Results and discussion4.3.2 Particles of polybutene standardsFigure 4.3 shows examples of optical images of particles of the polybutenestandards recorded during poke-and-flow experiments at <0.5 % RH. As forsucrose-water particles, the geometry of the particles of standard solutioncould be described as a spherical cap prior to being poked (Fig. 4.3; Panelsa1 and b1), and a half-torus after being poked (Fig. 5; Panels a2 andb2). Upon removal of the needle the material flowed, with the size of thehole at the centre of the half-torus geometry decreasing over time. For theparticles in Fig. 4.3, τ exp,flow was determined to be 2.50 and 6.00 seconds,respectively (Fig. 4.3; Panels a3 and b3). The particle continued to flowafter τ exp,flow and eventually re-attained it’s initial, energetically favourable,spherical cap geometry (Fig. 4.3; Panels a4 and b4). The mean experimentalflow times, τ exp,flow, were 2.79 (σ = 0.39) seconds for the lower viscosityFigure 4.3: Optical images of particles of polybutene standards (a) Stan-dard #1 (N450000), and (b) Standard #2 (N2700000), being poked at 0% RH recorded during typical poke-and-flow experiments. Images a1 andb1 correspond to particles prior to poking. Images a2 and b2 correspondto the first frame post-poke (i.e. the first frame after the needle has beenremoved). Images a3 and b3 correspond to images of the experimental flowtime, τ exp,flow, the point at which the equivalent area diameter of the holeat the centre of the torus has decreased to 50 % of its original size. Imagesa4 and b4 correspond to the final frame recorded of each particle, at whichpoint each particle has re-attained its original spherical cap geometry. Scalebar: 20 µm. Originally published in Grayson et al. (2015a).634.3. Results and discussionstandard #1, and 7.86 (σ = 1.65) seconds for the higher viscosity standard#2.As for the sucrose-water experiments, τ exp,flow values from individualparticles were converted into viscosities using simulations. To determineupper and lower limits for the viscosities of the standards, the upper andlower limits of the viscosities determined from the individual particles wereused. Using this approach, and the parameters listed in Table 4.2 the simu-lated lower and upper limits of viscosity for the particles poked were 2.0 x 102and 1.2 x 104 Pa s for standard #1 (N450000), and 3.1 x 102 and 2.4 x 104Pa s for standard #2 (N2700000). These values are in good agreement withthose reported by the manufacturer, Cannon Instrument Company (Figure4.4).Figure 4.4: Viscosity as a function of temperature for experiments with thepolybutene standards. Results from standard #1 (N45000) are in blackwhilst results from Standard #2 (N2700000) are in red. Symbols representvalues measured by Cannon Instrument Company using a manual capillaryviscometer. Bars represent viscosities determined herein. For the bar thatrepresents each standard the bottom of the bar represents the lowest lowerlimit of viscosity of all the particles examined, whilst the top of the barrepresents the highest upper limit of viscosity of all of the particles examined.Originally published in Grayson et al. (2015a).644.4. Summary4.4 SummaryThe poke-and-flow technique combined with simulations of fluid flow pro-vides the advantage of being able to measure viscosities of samples that areboth highly viscous and available only in small sample volumes. The combi-nation of these characteristics provides a challenge that is beyond the reachof current commercially available viscometers. In Chapter 3, only a prelim-inary validation of the poke-and-flow technique combined with simulationsof fluid flow was carried out for viscosities up to 104 Pa s, due to the lackof suitable standards for validation at the time of publication. The currentmanuscript expands on the initial validation experiments in Chapter 3.First, the approach was used to determine the viscosity of sucrose-waterparticles over a wider range of relative humidities than previously done inChapter 3. The lower and upper limits of viscosity at ≈59 % RH were 1.0x 101 Pa s and 1.6 x 104 Pa s, whilst at 37 % RH the corresponding valueswere 7.2 x 104 and 4.7 x 106 Pa s, respectively. The results are in goodagreement with Quintas et al. (2006) who measured a viscosity of 103 Pa sat 54 % RH using a rotational controlled stress rheometer, and Power et al.(2013) who recently reported that mean measured viscosities of ≈5 x 102Pa s to ≈3 x 106 Pa s between 60 % RH and 37 % RH using holographicoptical tweezers.Second, the approach was used to determine the viscosity of two poly-butene standards. The simulated lower and upper limits of viscosity forstandard #1 was 2.0 x 102 and 1.6 x 104 Pa s and for standard #2 1.6 x 102and 2.6 x 104 Pa s. These values are in good agreement with value reportedby Cannon Instrument Company (See Fig. 4.4).The results for both the sucrose-water particles and the polybutene stan-dards shows the poke-and-flow technique combined with simulations of fluidflow is capable of providing both lower and upper limits of viscosity thatare consistent with literature or measured values when the viscosity is inthe range of ≈5 x 102 - ≈3 x 106 Pa s. This covers an important part ofthe range of viscosities of secondary organic material generated in environ-mental chambers. For example, this range of viscosities have been measured654.4. Summaryat atmospherically relevant RHs for both the water-soluble fraction of theSOM produced via the ozonolysis of α-pinene (Chapter 3), as well as thetotal SOM produced via the oxidation of isoprene (Song et al., 2015). Inaddition, this technique has several advantages including being low in costand the experimental setup affording compatibility with cascade impactorsfor particle collection.Whilst the poke-and-flow technique combined with simulations of fluidflow gives good agreement with measured values the upper limit of viscosityfor a given particle is typically a factor of 16-140 larger than the correspond-ing lower limit of viscosity. Thus, the limits of viscosity determined usingthis approach are wide. The largest source of uncertainty in the approach isthe values of surface tension and slip length used in the simulations. Con-straining these values would lead to a reduction in the uncertainty of themeasurements.66Chapter 5Effect of varyingexperimental conditions onthe viscosity of α-pinenederived secondary organicmaterial5.1 IntroductionIn Chapter 3 the viscosity of the water-soluble component of SOM producedfrom the ozonolysis of α-pinene in an environmental chamber was deter-mined. The current manuscript is an extension of the work in Chapter 3.Similar to Chapter 3 the viscosity of SOM particles derived via the ozonol-ysis of α-pinene is studied; however, in contrast, the current study focuseson particles consisting of the whole SOM, meaning both the water-solublefraction and water-insoluble fraction.In the first set of experiments, the effect of relative humidity (RH) onthe viscosity of the whole SOM was investigated. SOM was generated viathe ozonolysis of α-pinene. Reported here are viscosity measurements as afunction of RH between <0.5 % and 50 % RH, using SOM with productionmass concentrations of 520 and 121 µg m−3. The results add to the fewexisting measurements of the effect of RH on the viscosity of SOM producedvia the ozonolysis of α-pinene (Bateman et al., 2015; Kidd et al., 2014; Zhanget al., 2015, and Chapter 3). Understanding the effect of RH on the viscosity675.2. Experimentalof SOM is important as RH in the boundary layer regularly varies betweenroughly 20 % and 100 % RH with varied time and location in the planetaryboundary layer (Hamed et al., 2011).In the second set of experiments, the effect on viscosity of the productionmass concentration of SOM particles (in units of micrograms of SOM perm3 of gas) used when generating SOM was investigated. Experiments haveshown that the composition of SOM particles can change with productionmass concentration (Shilling et al., 2009), possibly affecting the viscosity ofthe SOM particles. The production mass concentrations of the SOM in thecurrent study ranged from 121 to 14,000 µg m−3.5.2 ExperimentalSOM particles were produced either in a flow tube (particle mass concentra-tions of 520 to 14,000 µg m−3) or a chamber (particle mass concentrationsof 121 and 230 µg m−3) at <5 % RH and collected on hydrophobic sub-strates with an impactor (see Sections 5.2.1 and 5.2.2, and Fig. 3.1). Thepoke-and-flow technique in conjunction with simulations of fluid flow wasused to determine the viscosity of the SOM (see Sections 5.2.3 and 5.2.4).5.2.1 Production and collection of SOM generated atproduction mass concentrations from 520 to 14,000µg m−3For SOM generated at production mass concentrations from 520 to 14,000µg m−3, a previously described flow tube was used (Shrestha et al., 2013)to generate the SOM. Alpha-pinene (Sigma-Aldrich, >99.5 % purity, 97 %enantiomeric excess) and 2-butanol (Sigma-Aldrich, >99.5 % purity; usedas an OH scavenger) were introduced into the flow tube at an α-pinene:2-butanol ratio of 1:49, using a dry air flow rate of 0.50 sLpm. Ozone wasproduced prior to the inlet of the flow tube by passing dry air through anozone generator (Jetlight, Model 600) at a rate of 3.0 sLpm, resulting in anozone concentration of 12 ppm at the inlet of the flow tube, as measured by685.2. Experimentalan ozone sensor (Ecosensors, UV-100). Residence time in the flow tube was38± 1 s. The concentration of the α-pinene entering the flow tube was variedto produce samples at a total of five different particle mass concentrations (asmeasured using an SMPS; TSI, model 3934), with the ozone concentrationbeing kept in constant excess. Table 5.1 shows the mass concentrationsand collection times used in the flow tube experiments, as well as the meangeometric size of the particles produced in the flow (Sample names Flowtube #1 - Flow tube #5).After exiting the flow tube the 2 Lpm of dry flow was diluted with an 8Lpm flow of humidified air, giving a total airflow of 10 Lpm with a humidityof 68 ± 2 % RH, as measured using an RH meter (Rotronic, HC2-S). Theairflow then passed through a single stage impactor (MSP Corp.), withinwhich a hydrophobic glass substrate was mounted. Over the course of anexperiment sub-micron sized SOM particles impacted on a hydrophobic glasssubstrate, with the humidified gas serving to reduce the fraction of particlesthat bounced in the impactor. Over time the particles coagulated to formsuper-micron sized particles. The production mass concentration, modediameter, geometric standard deviation, and collection times are detailed inTable 5.1. After collection, the samples were stored at 253 K until use. Allsamples were used within 4 weeks of production. To determine the impactof storing samples at 253 K, the viscosity of one sample (produced using amass concentration of 6,000 µg m−3) was measured first after four days ofstorage and again after 24 further days of storage. The measured lower andupper limits of viscosity differed by <20 % (which is within experimentaluncertainty) when measured at both 30 % and <0.5 % RH.5.2.2 Production and collection of SOM generated atproduction mass concentrations of 121 and 230 µgm−3For production mass concentrations <500 µg m−3, the time required tocollect enough material for the poke-and-flow experiments was >12 hoursusing the flow tube setup described above. As a result, to collect SOM using695.2.ExperimentalTable 5.1: Conditions used for generating and collecting samples of SOM generated via theozonolysis of α-pinene. The whole SOM (both water soluble and water insoluble componentof the SOM) was collected. Originally published in Grayson et al. (2015b).SamplesProduction mass Mode GeometricCollectionSample namestudiedconcentration diameter standardtime (min)(µg m−3) (nm) deviationFlow tube sample #1 3 (1.4 ± 0.1) x 104 265 ± 7 1.43 20Flow tube sample #2 3 (5.9 ± 0.7) x 103 194 ± 5 1.47 90Flow tube sample #3 3 (3.4 ± 0.1) x 103 163 ± 2 1.46 150Flow tube sample #4 3 (1.2 ± 0.2) x 103 121 ± 8 1.46 450Flow tube sample #5 3 (5.2 ± 0.2) x 102 132 ± 2 1.52 800Chamber sample #1 2 (2.3 ± 0.1) x 102 181 ± 12 95Chamber sample #2 2 (1.2 ± 0.1) x 102 169 ± 12 180705.2. Experimentalproduction mass concentrations less than 500 µg m−3 the Leipzig AerosolChamber (LEAK), a cylindrical 19 m3 Teflon bag (Iinuma et al., 2009), wasused. The LEAK chamber could be sampled at higher flow rates than theflow tube (16 Lpm as opposed to 2 Lpm), reducing the required collectiontime.First, ozone was introduced into LEAK, which was operating under dryconditions (<5 % RH). The ozone concentration was held between 64-72 ppb(ozone monitor; 49c Ozone Analyzer, Thermo Scientific, USA). Afterwardsα-pinene (Sigma-Aldrich, >99.5 % purity, 97 % enantiomeric excess) wasinjected into LEAK, and the formation and growth of SOM particles withinLEAK was monitored using an SMPS (TROPOS-type). No OH scavengerwas used during experiments. After 80 minutes of reaction, the submicronsized particles were collected by pumping air from the chamber at a flowrate of 16 Lpm. At the exit of LEAK the air passed through a humidifierunit (FC300-1660-15-LP-01, Perma Pure LLC, NJ, USA), and upon exitingthe humidifier unit the air was determined to be at 91 ± 2.5 % RH, asmeasured using a handheld RH meter (RH85, Omega, USA). The airflowthen passed through a single stage impactor, as described in Section 5.2.1.Particles collected and coagulated on a hydrophobic slide located with theimpactor. After collection, the samples were stored at 253 K until use. Allsamples were used within 10 weeks of production. The production massconcentration, mode diameter, and collection times are detailed in Table 5.1(Samples named Chamber sample #1 and Chamber sample #2)5.2.3 Poke-and-flow techniqueThe viscosities of the SOM collected on the hydrophobic slides were deter-mined with the poke-and-flow technique combined with simulations of fluidflow, which has be described in Section 2.2. As experiments were performedbetween <0.5 % and 50 % RH, the dewpoint monitor was calibrated us-ing the deliquescence dewpoint of potassium carbonate, and found to givereadings within 0.1 K of the expected value at 43 % RH and 293 K.During poke-and-flow experiments the SOM was exposed to a dry or hu-715.2. Experimentalmid gas flow over a period of 3-15 h. During this time semivolatile compo-nents of the SOM may undergo evapouration. If the semivolatile componentswere behaving as plasticizers within the SOM, the viscosity of the SOM maychange. To determine whether this process occurred here and, if so, whetherit had a significant effect on the results, the volumes of particles consistingof whole SOM and produced with a mass concentration of 6,000 µg m−3were determined for up to 45 h while exposed to a dry (<0.5 % RH) flowof nitrogen gas (see Section B.1 for further experimental details). Withinexperimental uncertainty the volume of the particles remained unchanged(Figure B.1).In addition, the viscosity of particles consisting of whole SOM and pro-duced with a mass concentration of 6,000 µg m−3 were determined afterboth 1 h and 45 h of exposure to a dry (<0.5 % RH) flow of nitrogengas. The mean lower and upper limits of viscosity were determined to beroughly double after 45 h of exposure compared to their values after 1 hof exposure (Table B.1). This result suggests it is possible that a smallvolume of semivolatile material may have evapourated during the exposureto dry nitrogen, below the detection limit of the measurements of particlevolume, but enough to result in a small increase in viscosity. Alternatively,oligomerisation or polymerisation could occur within the samples at roomtemperature, with the products of this process being of higher viscosity thantheir precursors. This doubling in viscosity should be considered as a conser-vative upper limit to the effect of evaporation in the rest of the experimentsreported here, which were carried out on a time scale of 3-15 h.Two or three samples were analyzed per set of conditions, and the resultsof the three samples combined to give the values reported here. In total,this study contains the results from experiments on a total of 436 particles.5.2.4 Simulations of fluid flowSimulations of fluid flow were performed as detailed in Section 2.2.3. Shownin Table 5.2 are the estimates of the physical properties (i.e., particle-substrate slip length, surface tension, particle-substrate contact angle, and725.2.ExperimentalTable 5.2: Physical parameters used when simulating particles that exhibited flow withCOMSOL. Originally published in Grayson et al. (2015b).Density Slip length Surface tensionContact angle/ kg m−3 a / m b / mN m−1 cValue used to determine1,300 5 x 10−9 40 See Table B.2 dlower limit of viscosityValue used to determine1,300 1 x 10−5 75 See Table B.2 eupper limit of viscositya Density was varied from 1,000-1,400 kg m−3 based on the work of Chen and Hopke (2009) and deter-mined to have no effect upon viscosities determined via simulation. As such a median value of 1300 kgm−3 was used.b For references and rationale see Chapter 4.c Range of surface tension values based on work on Tuckermann and Cammenga (2004).d The lower value from Table B.2 is used for particles of geometry (R0 - r0) / r0 <2, and the upper valuefrom Table B.2 is used for particles of geometry (R0 - r0) / r0 >2.e The upper value from Table B.2 is used for particles of geometry (R0 - r0) / r0 <2, and the lower valuefrom Table B.2 is used for particles of geometry (R0 - r0) / r0 >2.735.3. Results and discussiondensity) of SOM used during simulations. In addition images acquired dur-ing each experiment were used to determine the dimensions of each particleand its value of τ exp,flow as inputs for simulations. Contact angles weredetermined using 3-D images of the super-micron particles suspended onhydrophobic surfaces using a confocal fluorescence microscope (Leica SP5II) with a 20x objective, a schematic of which is shown in Figure B.2(a).An excitation wavelength of 458 nm was used, causing moieties within theSOM to fluoresce, and a range of emission wavelengths, from 500-700 nm,were used to produce the image. A z-stack series of images with a step sizeof 0.5 µm, was acquired for each particle. Contact angles were subsequentlymeasured from the 2-D cross-sections in the y-z plane using the LB-ADSAplugin for ImageJ (Fig. B.2(b)). Contact angles were determined by mea-suring multiple particles from each sample and are reported in Table B.2.The values used during simulations of a given particle are those determinedfor particles of the corresponding sample.The main source of uncertainty in the viscosity of the SOM arises fromuncertainty in the physical properties of SOM that are used in simulations,including the slip length, the particle-substrate contact angle, and the sur-face tension at the particle-gas interface. The variability in viscosity fromparticle to particle was only a small component of the overall uncertainty(discussed further in Section B.2).5.3 Results and discussion5.3.1 Effect of relative humidity on the viscosity of SOMThe effect of relative humidity on the viscosity of SOM was determined forSOM produced with production mass concentrations of 520 µg m−3 and 121µg m−3. Shown in Figure 5.1 are images of SOM produced in the flow tubewith a production mass concentration of 520 µg m−3 and studied at <0.5% and 50 % RH. Shown in Fig. 5.1(a) (Panels a1-a3) is SOM being studiedat <0.5 % RH. Prior to poking the SOM is in a hemispherical geometry(Fig. 5.1, Panel a1). The act of poking the SOM with the needle led to the745.3. Results and discussionFigure 5.1: Optical images recorded during typical poke-and-flow experi-ments of whole SOM produced at a production mass concentration of 520µg m−3 being poked at (a) <0.5 %, and (b) 50 %, RH. Images a1 andb1 correspond to SOM prior to poking. Images a2 and b2 correspond tothe first frame post-poke (i.e. the first frame after the needle has been re-moved). Images a3 and b3 correspond to images of the experimental flowtime, τ exp,flow, the point at which the diameter of the hole at the centre ofthe torus has decreased to 50 % of its original size. Scale bar in Images a1and b1: 20 µm. Originally published in Grayson et al. (2015b).formation of a half-torus geometry (Fig. 5.1, Panel a2). Upon removal ofthe needle the material flowed and the hole began to close, with a τ exp,flowof 1,074 s (Fig. 5.1, Panel a3). Shown in Fig. 5.1(b) (Panels b1-b3) is SOMbeing studied at 50 % RH. As for the SOM in Fig. 5.1(a), the SOM washemispherical in geometry prior to being poked (Fig. 5.1, Panel b1), and theact of poking the SOM also lead to the formation of a half-torus geometry(Fig. 5.1, Panel b2). However, in this case the flow rate was clearly faster,and the SOM was determined to have a τ exp,flow of 4.3 s (Fig. 5.1, Panela3).Figure 5.2 summarizes the RH dependent studies. For SOM producedat a production mass concentration of 520 µg m−3 the mean τ exp,flow valuewas a factor of 460 lower at 50 % RH than at <0.5 % RH (Fig. 5.2(a)). By755.3. Results and discussionFigure 5.2: Summary of poke-and-flow experiments from <0.5 % to 50 % RHperformed on samples of whole SOM produced at mass concentrations of 520µg m−3 (Panels (a) and (c)) and 121 µg m−3 (Panels (b) and (d)). Panels(a) and (b) show box plots of observed τ exp,flow as a function of RH. Panels(c) and (d) show simulated lower (filled symbols) and upper (open symbols)limits of viscosity. Y-error bars represent 95 % confidence intervals, and x-error bars represent the range of RH at which measurements were made. Theshaded regions are included to guide the eye of the reader. The viscosities ofcommon substances at room temperature have been added to (d) to providepoints of reference, as per Koop et al., 2011. The image of pitch is partof an image from the pitch drop experiment (image courtesy of WikimediaCommons, GNU Free Documentation License, University of Queensland,John Mainstone). Originally published in Grayson et al. (2015b). 765.3. Results and discussioncomparison, SOM produced at a production mass concentration of 121 µgm−3 the mean τ exp,flow value was a factor of 3,600 lower at 50 % RH thanat <0.5 % RH (Fig. 5.2(b)).Based on simulations of the poke-and-flow experiments the viscositiesof SOM produced at a production mass concentration of 520 µg m−3 theviscosity was between 3 x 105 and 2 x 107 Pa s at <0.5 % RH and between4 x 102 and 3 x 104 Pa s at 50 % RH (Fig. 5.2(c)). The viscosity of SOMproduced at a production mass concentration of 121 µg m−3 was determinedto be between 2 x 106 and 6 x 107 Pa s at <0.5 % RH and between 1.8 x102 and 1.4 x 104 Pa s at 50 % RH. The results suggest the viscosity ofboth samples was between that of window putty and tar pitch at <0.5 %RH and that of ketchup and window putty at 50 % RH. The RH-dependentresults are consistent with previous work that has shown that the viscosity ofSOM can depend strongly on RH (Bateman et al., 2015; Saukko et al., 2012;Song et al., 2015; Zhang et al., 2015, and Chapter 3), with the dependenceof the viscosity on RH likely being a combination of water behaving as aplasticizer and the fraction of water present in a particle increasing with RH(Koop et al., 2011).5.3.2 Effect of production mass concentration used whengenerating the SOM on the viscosity of SOMViscosity of SOM as a function of production mass concentration used togenerate SOM was studied at 30 % RH and <0.5 % RH. Figure 5.3 showsexamples of SOM generated at production mass concentrations of 14,000,520, and 121 µg m−3 being poked at <0.5 % RH. In all cases the SOMexhibited flow, and there is a trend of increasing experimental flow timewith decreasing production mass concentration.A summary of the τ exp,flow and viscosity values as a function of produc-tion mass concentration at <0.5 % RH is shown in Figure 5.4. Consideringall the data together, as the production mass concentration decreases from14,000 µg m−3 to 121 µg m−3, the mean τ exp,flow values increase by a factorof 45 (Fig. 5.4(a)). Based on simulations of the poke-and-flow experiments775.3. Results and discussionFigure 5.3: Optical images recorded during typical poke-and-flow experi-ments of particles of the whole SOM produced at production mass concen-trations of (a) 14,000 µg m−3, (b) 520 µg m−3, and (c) 121 µg m−3 beingpoked at <0.5 % RH. Images a1, b1 and c1 correspond to SOM prior topoking. Images a2, b2 and c2 correspond to the first frame post-poke (i.e.the first frame after the needle has been removed). Images a3, b3 and c3correspond to images of the experimental flow time, τ exp,flow, the point atwhich the diameter of the hole at the centre of the torus has decreased to 50% of its original size. Scale bar in Images a1, b1 and c1: 20 µm. Originallypublished in Grayson et al. (2015b).the viscosities of the SOM samples are between 4 x 104 and 1.5 x 106 Pas for SOM produced at a production mass concentration of 14,000 µg m−3and between 6 x 105 and 5 x 107 Pa s for SOM produced at a productionmass concentration of 121 µg m−3 (Fig. 5.4(b)).The inverse relationship between viscosity and production mass concen-tration is consistent with results of Shilling et al. (2009), who observed aninverse relationship between production mass concentration and the oxida-785.3. Results and discussionFigure 5.4: Summary of poke-and-flow experiments performed on samplesof whole SOM at <0.5 % RH. Black symbols represent results from particlesproduced in a flow tube, whilst red symbols represent results from particlesproduced in a chamber. Panel (a) shows box plots of observed τ exp,flowtimes at different production mass concentrations for particles poked <0.5% RH. Boxes represent the 25, 50, and 75 percentiles, open circles representmedian values, and whiskers represent the 5 and 95 percentiles. Panel (b)shows the simulated lower (filled squares) and upper (open squares) limitof viscosity for particles at each production mass concentration poked at<0.5 %. Symbols represent mean values. The y error bars represent 95 %confidence intervals. The shaded regions are included to guide the eye ofthe reader. Also included in (b) are literature viscosities for SOM producedvia the ozonolysis of α-pinene (Renbaum-Wolff et al., 2013a; Zhang et al.,2015). Originally published in Grayson et al. (2015b).795.3. Results and discussiontion level of the resulting SOM. As previously mentioned, higher oxidationlevels are linked to higher glass transition temperatures and an increasedlikelihood that a particle rebounds from an impactor surface.The results for SOM produced in the flow tube (production mass concen-trations of 14,000 to 520 µg m−3) and produced in the chamber (productionmass concentrations of 230 and 121 µg m−3) each exhibit the same trend:τ exp,flow increases as production mass concentration decreases. However,the data are not in perfectly aligned. If the data from the flow tube areextrapolated to lower particle mass concentrations, slightly higher τ exp,flowvalues are predicted compared to observations using samples from the cham-ber (roughly a factor of 2-3 higher). This difference could be due to somedifferences in experimental conditions. For example, the flow tube studieswere carried out in the presence of an OH scavenger, 2-butanol, whereas noOH scavenger was used in the chamber studies. The presence of 2-butanoldecreases the SOM yield from a given amount of precursor (Henry and Don-ahue, 2011; Jonsson et al., 2008). The reaction of OH with both α-pinene,as well as first generation products of α-pinene ozonolysis, can alter thechemical composition of the SOM produced (Vereecken and Peeters, 2012).Another difference in experimental conditions between the flow tube andthe chamber studies was the RH at which the SOM was collected - 68 ± 2 %in flow tube studies and 91 ± 2.5 % in chamber studies. The increased hu-midity while SOM was being collected during the chamber studies may haveresulted in a larger fraction of the more volatile components being presentin the particle phase as material was collected, possibly explaining the lowerthan expected viscosity of the samples collected during chamber studies.Also included in Fig. 5.4 are previous measurements of the viscosity ofα-pinene derived SOM measured under dry conditions. Zhang et al. (2015)studied material produced in the same flow tube as the material used hereusing a production mass concentration of ≈70 µg m−3, and in Chapter 3 theviscosity of the water-soluble component of SOM produced at a productionmass concentration of ≈50 µg m−3 in an environmental chamber was de-termined. The results of Zhang et al. (2015) are consistent with the resultsobtained here. The results from Chapter 3 are not inconsistent with the805.3. Results and discussioncurrent results due to the observed inverse relationship between viscosityand production mass concentration.Other researchers have measured diffusion rates (Abramson et al., 2013;Cappa and Wilson, 2011; Perraud et al., 2012), or mixing times under dryconditions Robinson et al. (2013); Saleh et al. (2013) within SOM producedvia the ozonolysis of α-pinene. In Section B.3 these measurements have beenconverted to viscosities using the Stokes-Einstein relationship. It should bekept in mind that the Stokes-Einstein relationship may break down for smallmolecules (Bones et al., 2012; Price et al., 2015) and for large molecules whenthe viscosity is high and near the glass transition temperature (Championet al., 1997; Corti et al., 2008). Further discussion on the conversion of re-ported diffusion coefficients or mixing times to viscosities for each of thesestudies is given in Section B.3. Figure B.3 shows that most of these pre-vious studies (Cappa and Wilson, 2011; Perraud et al., 2012; Saleh et al.,2013) are not inconsistent with those presented here. Some of the results areoutside of the range reported here (Abramson et al., 2013; Robinson et al.,2013) suggesting factors beyond just a simple relationship between viscosityand production mass concentration are required to explain previous mea-surements. Differences may be due to invalid assumptions made when usingthe Stokes-Einstein relationship or other factors.The effect of production mass concentration on viscosity was also studiedat 30 % RH (Figure 5.5). At this RH, the effect of particle mass concen-tration was not as dramatic. For the samples produced in a flow tube, asthe production mass concentration decreases from 14,000 µg m−3 to 520µg m−3, the mean τ exp,flow values increase by a factor of 5 (Fig. 5.5(a)).For the samples produced in the chamber, as the production mass concen-tration decreased from 230 µg m−3 and 121 µg m−3, the mean τ exp,flowvalues increase by a factor of 1.5. Similar to the experiments at <0.5 %RH, if the results from the flow tube are extrapolated to lower particle massconcentrations, they predict larger τ exp,flow values than observed from thechamber studies. As mentioned above, these differences may be due to smalldifferences in experimental conditions between the flow tube and chamber.Based on simulations the viscosity of the SOM at 30 % RH is between815.3. Results and discussionFigure 5.5: Summary of poke-and-flow experiments performed on particlesof whole SOM at 30 % RH. Black symbols represent results from particlesproduced in a flow tube, whilst red symbols represent results from particlesproduced in a chamber. Panel (a) shows box plots of observed τ exp,flowtimes as a function of SOM mass concentrations for particles studied usingthe poke-and-flow technique at 30 % RH. Panel (b) shows the simulatedlower (filled squares) and upper (open squares) limit of viscosity for particlesat each SOM mass concentration studied using the poke-and-flow techniqueat 30 % RH. Symbols represent mean values, whilst the y error bars represent95 % confidence intervals. The shaded region is included to guide the eyeof the reader. Also included in (b) are literature viscosities from Renbaum-Wolff et al. (2013a) and Zhang et al. (2015), for SOM produced via theozonolysis of α-pinene and studied at 30 % RH. Originally published inGrayson et al. (2015b).825.3. Results and discussion1.0 x 103 and 9 x 104 Pa s at a production mass concentration of 14,000µg m−3 and between 1.2 x 103 and 1.2 x 105 Pa s at a production massconcentration of 121 µg m−3 (Fig. 5.5(b)). The smaller dependence ofviscosity on production mass concentration at 30 % RH compared to <0.5% RH can be explained by the dependence of the viscosity on the watercontent of the SOM. Under dry conditions the measured viscosity is due onlyto the viscosity of the SOM. However, as RH is increased the SOM uptakeswater, and the viscosity of the different SOM samples become increasinglydependent on the viscosity of water and converge, finally approaching theviscosity of water, ≈10−3 Pa s, at 100 % RH.Also included in Fig. 5.5(b) are viscosities of α-pinene-derived SOMmeasured at 30 % RH by Zhang et al. (2015) and reported in Chapter 3. Asmentioned above Zhang et al. (2015) studied material produced in the sameflow tube as the material used here, and in Chapter 3 the water-soluble com-ponent of SOM produced in an environmental chamber was studied. Onepossible explanation of the results shown in Fig. 5.5(b) is a very strongdependence of viscosity on production mass concentration in the range of50 and 121 µg m−3. To determine if a strong dependence of viscosity in therange of 50 and 121 µg m−3 shown in Fig. 5.5(b) exists or due to otherfactors, additional studies are needed. More importantly, additional studiesare needed to determine if the viscosity of the water-soluble component ofSOM is the same as the viscosity of the whole SOM (water-soluble and waterinsoluble components) produced at production mass concentrations around50 µg m−3. In addition, further comparison studies using the techniqueintroduced by Zhang et al. (2015) and the poke-and-flow technique usedhere would be beneficial. Finally, the studies here are carried out at produc-tion mass concentrations greater than those found under ambient conditions(Hallquist et al., 2009; Slowik et al., 2010), and studies carried out using ma-terial produced using ambient concentrations would provide further usefulinformation.835.3. Results and discussion5.3.3 Effect of the water-insoluble component on theviscosity of SOMTo better understand the difference between the viscosity of water-solubleSOM and SOM containing both the water-soluble and water-insoluble com-ponents, additional measurements were carried out using just the water-soluble component of SOM generated by the ozonolysis of α-pinene at aproduction mass concentration of 14,000 µg m−3. Particles were generatedusing the flow tube as discussed in Section 5.2.1, and particles from theoutlet of the flow tube were collected on a Teflon filter. After collection,SOM was extracted from the Teflon filter by placing it in a clean glass jarand immersing the filter in 10 mL of Millipore (18.2 MΩ cm) water. Thejar was shaken for 1.5 h, with the filter being flipped over half way through,after which the filter was removed from the jar, resulting in a solution of thewater-soluble component of the SOM. The solution was then nebulised andsprayed onto a hydrophobic glass substrate, producing super-micron sizedparticles. The particles were then studied using the poke-and-flow techniqueand their viscosities determined using simulations of fluid flow as describedin Section 2.2.Shown in Figure 5.6 are images of a particle comprised of the water-soluble fraction of SOM (Fig. 5.6(a)) and a particle comprised of the wholeSOM, both the water-soluble and water-insoluble fractions (Fig. 5.6(b)).Both were produced at a production mass concentration of 14,000 µg m−3and studied at <0.5 % RH. Although the production of both the water-soluble SOM and the whole SOM took place using equivalent flow tubeconditions, the images of the SOM during the poke-and-flow experimentswere clearly different, with the water-soluble SOM cracking and showing noobservable flow over the course of 14 hours (Fig. 5.6(a), Panels a2 & a3),whilst the whole SOM exhibited flow, with a τ exp,flow of 1074 s (Fig. 5.6(b),Panels b2 & b3).845.3. Results and discussionFigure 5.6: Optical images recorded during poke-and-flow experiments usingparticles consisting of (a) the water-soluble component of the SOM and(b) the whole SOM (i.e., both the water-soluble and the water-insolublecomponents). In both experiments the SOM was produced using a massconcentration of 14,000 µg m−3 and was poked at <0.5 % RH. Images a1 andb1 correspond to the SOM prior to being poked. The brightness in Imagea1 is due to reflection of the source light by the needle positioned just abovethe particle. Images a2 and b2 correspond to the first frame post-poke (i.e.the first frame after the needle has been removed). The particle comprisedof the water-soluble component of SOM exhibited cracking behaviour and,as shown in Image a3, no change in the size or shape of the cracks can beobserved 14 hours after the particle has been poked. The particle comprisedof whole SOM exhibited flow, and Image b3 corresponds to an image ofthe particle at its experimental flow time, τ exp,flow, the point at which thediameter of the hole at the centre of the torus has decreased to 50 % of itsoriginal size. Scale bar in Images a1 and b1: 20 µm. Originally publishedin Grayson et al. (2015b).Table 5.3 summarizes experimental results at <0.5 % RH for both thewater-soluble SOM and the whole SOM produced at a production mass con-centration of 14,000 µg m−3. The τ exp,flow and viscosity of the water-soluble855.3. Results and discussioncomponent were both at least a factor of 300 greater than the τ exp,flow andviscosity of the whole SOM.The difference in viscosity between the whole SOM and the water-solubleSOM may arise from differences in the extent of oxidation of the SOM.Water-soluble SOM is assumed to be composed of the more oxidized com-ponents of the whole SOM and literature suggests that higher oxidation isrelated to a warmer glass transition temperature (Berkemeier et al., 2014;Dette et al., 2014; Koop et al., 2011), implying that viscosity increases withoxidation level.The results in Table 5.3 correspond to a high production mass concen-tration. At lower SOM particle concentrations such as concentrations usedin Chapter 3 the difference between water-soluble SOM and whole SOM islikely smaller, since as the production mass concentration decreases, the ex-tent of oxidation in the particle phase is expected to increase and hence theamount of water insoluble material in the particle phase should decrease.In addition, literature suggests that the SOM formed from the ozonolysis ofα-pinene is largely composed of water-soluble organic compounds (Hall andJohnston (2011) produced using a production mass concentration of <500Table 5.3: Summary of τ exp,flow times and viscosities ofwhole SOM and water-soluble SOM produced in the flowtube at a production mass concentration of 14,000 µg m−3and studied at <0.5 % RH. Originally published in Graysonet al. (2015b).τ exp,flowa Viscosity (Pa s) bWater-soluble SOM >4.3 x 104 >4.8 x 108Whole SOM 90 (57, 144) 3.8 x 104 - 1.5 x 106a For the whole SOM, τexp,flow values represent experimental valuesin the form ”mean (5th percentile, 95th percentile)”. For the water-soluble SOM, the lower limit to τexp,flow represents the shortestexperimental time that the particles were observed.b For whole SOM, the lower limit of viscosity represents the lower 95% confidence interval of the lower limit of viscosity, whilst the upperlimit of viscosity represents the upper 95 % confidence interval of theupper limit of viscosity. For water-soluble SOM the lower limit ofviscosity was calculated for the particles observed over the shortestexperimental time.865.4. Summaryµg m−3. Further, mass spectral analysis has revealed little difference inthe chemical composition of SOM produced via the ozonolysis of α-pineneand extracted using either water or a methanol:water solution (Heaton et al.,2007), and cloud condensation measurements suggest SOM generated via theozonolysis of α-pinene is not limited by solubility of the organic material inwater (King et al., 2009) for SOM produced at production mass concentra-tions of <100 µg m−3. Based on these arguments the results shown in Table5.3 should be considered as an upper limit to the difference between theviscosity of water-soluble SOM and whole SOM produced using productionmass concentrations lower than 14,000 µg m−3.5.4 SummaryThe effect of various experimental parameters on the viscosity of SOM de-rived via the ozonolysis of α-pinene have been studied. First, the effect ofrelative humidity on the viscosity of the whole SOM was studied. For eachsample studied the τ exp,flow values were larger and the simulated viscositieshigher as the RH was decreased from 50 % to <0.5 % (Figs. 5.1 and 5.2).Specifically, for SOM produced at a production mass concentration of 121µg m−3, the τ exp,flow increased by a factor of 3,600 as the relative humidity(RH) decreased from 50 % to <0.5 % RH. Based on simulations, the vis-cosities of the particles were between 3 x 102 and 9 x 103 Pa s at 50 % RHand between 6 x 105 and 5 x 107 Pa s at <0.5 % RH.Second, the effect on viscosity of the production mass concentration usedduring the production of SOM was investigated at 30 % and <0.5 % RH.The measurements provide evidence of an inverse relationship between pro-duction mass concentration in the reaction vessel and viscosity of the SOMmaterial (Figs. 5.3 and 5.4). The effect was most prominent at <0.5 % RHwhere τ exp,flow increased by a factor of 45 as the particle mass concentrationdecreased from 14,000 µg m−3 to 121 µg m−3. From simulations of the poke-and-flow experiments, the viscosity of the SOM produced at a productionmass concentration of 14,000 µg m−3 are between 4 x 104 and 1.5 x 106 Pas and the viscosity of SOM produced at a production mass concentration of875.4. Summary121 µg m−3 are between 6 x 105 and 5 x 107 Pa s at <0.5 % RH (Fig. 5.4).The τ exp,flow and viscosity of the water-soluble component of SOM wasalso observed to be at least a factor of 300 greater than the τ exp,flow andviscosity of the whole SOM when using a production mass concentration of14,000 µg m−3 (Fig. 5.6). This result should be considered as a upper limitto the difference between the viscosity of water-soluble SOM and whole SOMproduced at production mass concentrations lower than 14,000 µg m−3.Overall the results suggest that the RH at which the viscosity was deter-mined and the mass concentration at which the SOM was produced shouldbe considered when laboratory experiments are being compared or whenused to infer viscosities of atmospheric particles.88Chapter 6Viscosity of a tetraol andsaccharide-water mixtures6.1 IntroductionAs outlined in Chapter 1, an improved understanding of the viscosity ofSOM particles over the range of relative humidities found in the atmosphereis required to improve our ability to predict the atmospheric effects of SOM(Adler et al., 2013; Bodsworth et al., 2010; Koop et al., 2011; Kuwata andMartin, 2012; Riipinen et al., 2011; Shiraiwa et al., 2011a; Zelenyuk et al.,2012).The numerous oxidation mechanisms in the atmosphere result in SOMhaving a complex composition, with hundreds or thousands of individualspecies and a range of chemical structures and functionalities (Hallquistet al., 2009). Only ≈10% of the SOM mass has been identified at themolecular level (Hallquist et al., 2009). Molecules that have been identi-fied include alkane, alkene, alcohol, carboxylic acid, aldehyde, ketone, ester,ether, and acid anhydride functional groups, as well as both aromatic andnon-aromatic cyclic molecules (Aschmann and Atkinson, 1998; Chan et al.,2010; Chen et al., 2011b; Christoffersen et al., 1998; Day et al., 2009; Rus-sell et al., 2011; Surratt et al., 2006). A wide range of molar masses havebeen observed for individual species, with some being of molar mass >1000g mol−1, although smaller compounds are thought to account for the ma-jority of the SOM mass fraction (Gao et al., 2004; Praplan et al., 2015;Schobesberger et al., 2013). The average elemental oxygen-to-carbon ra-tio (O:C) of SOM has been determined to range from 0.3-1.1 (Chen et al.,896.1. Introduction2011a; Jimenez et al., 2009; Lambe et al., 2011, 2015). The high values ofO:C can be explained by organic compounds with multiple oxygen contain-ing functional groups (Christoffersen et al., 1998; Hoffmann et al., 1997; Yuet al., 1999). Recent studies have identified the presence of extremely lowvolatility organic compounds (ELVOCs) in SOM, which can be of O:C ≥1and may account for a significant fraction of SOM mass (Ehn et al., 2012,2014; Praplan et al., 2015; Schobesberger et al., 2013).In the following the viscosities of a tetraol (2-methyl-1,2,3,4-butanetetrol)and three saccharides (glucose, raffinose, and maltohexaose) mixed with wa-ter are determined. These systems are used as proxies for highly oxidizedcomponents of ambient SOM. The saccharides and the tetraol studied arelisted in Table 6.1 as well as their relevant physical properties. Althoughthe viscosity of sucrose has been measured over a wide range of relativehumidities (RHs) (e.g. Fo¨rst et al., 2002; Power et al., 2013; Quintas et al.,2006; Swindells et al., 1958; Telis et al., 2007), similar studies for glucose,maltohexaose and raffinose have yet to be carried out. Furthermore, we arenot aware of previous studies of the viscosity of tetraols.The saccharides studied have an O:C of 0.86-1.0, which is similar to thatof the more highly oxidized components of SOM, and molar mass rangingfrom 180-990 g mol−1, which covers the range of molar mass of componentsthat account for the majority of the mass of SOM. Saccharides includinglevoglucosan, glucose, xylose, sucrose, and maltose have been identified insmoke produced during the combustion of wood under both controlled con-Table 6.1: Properties of saccharide and tetraol compounds studied experi-mentally.CompoundChemicalO:CMolar mass RH range of viscosityformula / g mol−1 measurements / %Glucose C6H12O6 1.00 180 28Raffinose C18H32O16 0.89 342 40-85Maltohexaose C36H62O31 0.86 991 50-772-Methyl-1,2,3,C5H12O4 0.80 134 <0.54-butanetetrol906.2. Experimentalditions and in ambient atmospheric samples (Nolte et al., 2001).The tetraol studied, 2-methyl-1,2,3,4-butanetetrol, has an O:C of 0.8,also similar to that of the more highly oxidized components of SOM. 2-methyl-1,2,3,4-butanetetrol has been identified as a product of the oxidationof isoprene, and 2-methyl tetrols as a group are estimated to account forapproximately 1-2 % of the organic carbon mass in PM2.5 aerosols (Claeyset al., 2004; Edney et al., 2005; Hallquist et al., 2009; Surratt et al., 2010).In addition to measuring the viscosity of 2-methyl-1,2,3,4-butanetretrol,a comparison is made between its measured viscosity and the viscosity pre-dicted by two structure activity models (Suzuki et al. 1996; Katritzky et al.2010; Sastri and Rao 1992; Marrero-Morejo´n and Pardillo-Fontdevila 2000).These two models have been derived and validated using organic compoundswith viscosities<100 Pa s. As such, their applicability to organic compoundswith viscosities >100 Pa s, such as the 2-methyl-1,2,3,4-butanetretrol, is un-certain.6.2 Experimental6.2.1 Determination of viscosity of super-micron sizedparticles comprised of a tetraol or a saccharide mixedwith waterThe bead-mobility technique and the poke-and-flow technique combinedwith simulations of fluid flow have been used to measure the viscositiesof particles comprised of 2-methyl-1,2,3,4-butanetetrol or a saccharide (glu-cose, maltohexanose and raffinose) mixed with water.The bead-mobility technique is described in Section 2.1 and the poke-and-flow technique combined with simulations of fluid flow is described inSection 2.2. The physical properties used when simulating the poke-and-flowexperiments are summarised in Tables 6.2 and 6.3.B-D-Glucose (≥99.5 % purity), raffinose (≥98 % purity), and malto-hexaose (≥65 % purity) were obtained from Sigma-Aldrich. 2-methyl-1,2,3,4-tetraol was prepared in diastereomerically pure form starting from pro-916.2.ExperimentalTable 6.2: Physical parameters used to simulate the flow of material during poke-and-flow experiments where a half-torus geometry was formed and the material subsequentlyobserved to flow. R and r represent dimensions of a half-torus, with R representing theradius of the tube of material, and r representing the radius of the hole at the centre ofthe tube.Slip length Surface tension Density Contact angle/ nm / mN m−1 / g cm−3 / degrees aGlucoseLower limit 5 b 72.0 c 1.0 d66 (if r<2R)73 (if r>2R)Upper limit 10,000 b 95.1 e 1.7 d73 (if r<2R)66 (if r>2R)RaffinoseLower limit 5 b 57.2 f 1.0 d58 (if r<2R)67 (if r>2R)Upper limit 10,000 b 125.8 e 1.9 d67 (if r<2R)58 (if r>2R)MaltohexaoseLower limit 5 b 57.2 f 1.0 d63 (if r<2R)73 (if r>2R)Upper limit 10,000 b 138.2 e 2.0 d73 (if r<2R)63 (if r>2R)a Contact angles determined using photographic images of particles on a hydrophobic substrate.b This range is based on the literature values included in Table 2.1.c Lee and Hildemann (2013).d The density of water.e As predicted by ACD/Labs.f Lower limit of the surface tension of sucrose (Power et al., 2013, and Chapter 4).926.2. ExperimentalTable 6.3: Physical parameters used to simulate a lowerlimit of viscosity for poke-and-flow experiments where par-ticles cracked when impacted by the needle, and no observ-able flow of material was observed over the course of theexperiment.Slip length Surface tension Density Contact angle/ nm / mN m−1 / g cm−3 / ◦0.01 x l a 57.2 1.7 90a l is the grid spacing of the mesh, which ranged from 1-1.7 µmtected cis-2-methylbut-2-ene-1,4-diol (Fontana et al., 2000; Surratt et al.,2010). Protection of the alcohol moieties of cis-2-methylbut-2-ene-1,4-diolwith benzyl groups was followed by dihydroxylation using osmium tetrox-ide and N-methyl morpholine-N-oxide. Deprotection with H2 over Pd/Cresulted in the formation of the desired 2-methyl-1,2,3,4-tetraol. The purityof the tetraol was determined based on 1H and 13C spectra generated usingNMR spectroscopy. In order to test stability, solutions containing 100 mMof the tetraol in 1 M (NH4)2SO4 and D2O were stirred at room temperaturefor one week and monitored by NMR spectroscopy. No changes in compo-sition were observed during this time. Saccharide-water and tetraol-watersolutions for nebulising were prepared using high purity (18.2 MΩ) water.The range of relative humidites at which viscosities have been measuredfor each of the compounds are detailed in Table 6.1. Glucose was studiedonly at 28 % RH as viscosity at higher RH values can be extracted fromliterature data and below 28 % RH the particles stuck to the needle and wereremoved from the substrate during the poke-and-flow experiments, meaningtheir viscosity could not be determined. Raffinose particles were studiedbetween 40 and 85 % RH. At 40 % RH the particles cracked when poked,and did not flow on a laboratory timescale. The same results were expectedat lower RH values, and so experiments at lower RH values were not carriedout. Maltohexaose particles were studied at RH values ranging from 50 to80 %. As the particles cracked and did not flow at both 60 % RH and 50 %RH, experiments were not carried out at lower RH values. 2-Methyl-1,2,3,4-butanetetrol was studied only under dry conditions, though future studies936.2. Experimentalof viscosity as a function of RH for this compound are required as RH in theplanetary boundary layer ranges from approximately 20 to 100 % (Hamedet al., 2011; Martin, 2000).6.2.2 Predictions of viscosity using quantitativestructure-property relationship modelsTwo quantitative structure-property relationship (QSPR) models were usedto estimate the viscosity of 2-methyl-1,2,3,4-butanetetrol. QSPR modelsrelate physical, chemical, or physicochemical properties of compounds totheir structures. The first QSPR model used, which was proposed by Sas-tri and Rao (1992), estimates the viscosity of a compound based on itsvapour pressure, along with the number and type of functional groups inthe molecule. Experimental measurements of the vapour pressure of 2-methyl-1,2,3,4-butanetetrol are not available, and so its vapour pressure wasestimated using the three QSPR models employed by the E-AIM calculator(http://www.aim.env.uea.ac.uk/aim/ddbst/pcalc_main.php). Each ofthe QSPR models used to estimate vapour pressure is based on the boilingpoint of the molecule along with group contributions from the functionalgroups present in its structure. The first model uses the method of Nan-noolal et al. (2004) to predict the boiling point, and the method of Molleret al. (2008) to predict vapour pressure, the second uses the method of Nan-noolal et al. (2004) to predict the boiling point and the method of Nannoolalet al. (2008) to predict vapour pressure, and the third uses the method ofStein and Brown (1994) to predict the boiling point and the method ofMyrdal and Yalkowsky (1997) to predict vapour pressure. The three modelseach gave rise to a different estimation of the vapour pressure of 2-methyl-1,2,3,4-butanetetrol, and the lower and upper estimations of vapour pressurewere used to estimate lower and upper limits of viscosity.The second QSPR model used, which was proposed by Marrero-Morejo´nand Pardillo-Fontdevila (2000), estimates a compounds viscosity based onits molar mass and the type and number of bonds and functional groupswithin the molecule.946.3. Results and discussion6.3 Results and discussion6.3.1 Measured viscosities of saccharides at a range of RHsParticles comprised of saccharides were studied using the bead-mobility andpoke-and-flow techniques across a range of RHs. Shown in Figure 6.1 areexamples of images for poke-and-flow experiments for maltohexaose-waterand raffinose-water particles at RHs of 54 and 50 %, respectively. MarkedlyFigure 6.1: Figure 1: Optical images recorded during poke-and-flow experi-ments using particles of (a) maltohexaose and (b) raffinose. Images a1 andb1 correspond to the particles prior to being poked, with the white haloesbeing an optical effect. Images a2 and b2 correspond to the first frame afterthe needle has been remover. The particle composed of maltohexaose andstudied at 50 % RH exhibited cracking behaviour and, as shown in Imagea3, no change in the size or shape of the cracks can be observed 3 h afterthe particle has been poked. The particle comprised of raffinose and stud-ied at 54 % RH exhibited flow, and Image b3 corresponds to an image ofthe particle at its experimental flow time, τ exp,flow, the point at which thediameter of the hole at the centre of the torus has decreased to 50 % of itsoriginal size. The scale bar in images a1 and b1 corresponds to 20 µm.956.3. Results and discussiondifferent behaviour was observed in the particles. Maltohexaose cracked,and showed no observable flow over the subsequent three hours. In contrast,raffinose flowed upon removal of the needle, with a calculated experimentalflow time, τ exp,flow, of 112 s.The behaviour observed in images like the examples in Fig. 6.1 wassimulated and the viscosity of each saccharide-water particle determined.The simulated viscosities for each of the saccharide particles are grouped byRH and summarised in Table 6.4, and shown in Figure 6.2(a).Also included in Fig. 6.2(a) are literature data for the viscosity ofglucose-water mixtures at ≥75 % RH (Barbosa-Ca´novas et al. 2007; Haynes2015; Achard et al. 1992) and sucrose-water mixtures for RH values ≥25 %RH (Fo¨rst et al., 2002; Power et al., 2013; Quintas et al., 2006; Swindellset al., 1958; Telis et al., 2007). The viscosity of each of the saccharides wasobserved to increase as RH is decreased, with the viscosity at 28 % RH atleast four orders of magnitude greater than at 78 % RH. This inverse re-lationship between viscosity and RH is due to the behaviour of water as aplasticiser (a component that reduces the viscosity of a solution) and thegreater water content in particles at higher relative humidities.Figure 6.2(b) is a plot of viscosity vs. molar mass of the saccharides(primary x -axis) at three RHs for saccharide-water particles. At 28 % RH,the viscosity increased by 3.6-6.0 orders of magnitude as molar mass of thesaccharide increased from 180 to 342 g mol−1. At 77-80 % RH the viscosityincreased by 4.3-6.2 orders of magnitude as the molar mass of the saccharideincreased from 180 to 991 g mol−1. These observations are consistent withprior studies that suggest viscosity and molar mass are related through apower function (Hiemenz and Lodge, 2007; Pachaiyappan et al., 1967).Shown as a secondary x -axis on Fig. 6.2(b) is the number of saccharideunits in a compound’s structure. Molar mass increased as the number ofsaccharide units in a compound’s structure increased, and a relationship wasobserved between viscosity and number of saccharide units. At 28 % RH,glucose, which is comprised of one saccharide unit, was determined to havea viscosity between 2 x 103 and 1 x 105 Pa s, whilst sucrose, comprised oftwo saccharide units, was determined to have a viscosity between 3 x 108966.3. Results and discussionTable 6.4: Summary of experimental viscosity measurements usingthe bead-mobility and poke-and-flow techniques for the tetraol andthe saccharide-water particles studied here, with results from indi-vidual particles grouped by RH. For experiments using the bead-mobility technique the mean is reported along with the 95 % confi-dence intervals. For experiments using the poke-and-flow techniquelower and upper limits of viscosity are reported, taking account ofthe 95 % confidence limits of the simulated lower and upper lim-its of viscosity for the group of particles studied at each RH. N/Ais reported for all experiments performed using the poke-and-flowtechnique, for which no mean viscosity is calculated, and for upperlimits of viscosity for experiments with the poke-and-flow techniquewhere the particle cracked, as only a lower limit of viscosity can becalculated.Compound RH / %Viscosity / Pa sMean Lower limit Upper limitGlucose 28 * N/A 1.4e3 7.7e4Raffinose85 + 1.9e-1 0.8e-1 5.2e-180 + 4.3e0 1.7e0 1.2e161 *N/A5.9e2 4.7e455 * 9.3e3 1.0e648 * 2.7e5 2.4e740 * 4.8e8 N/AMaltohexaose77 * 3.5e3 1.9e574 * 5.3e3 3.7e660 * 3.2e8N/A50 * 2.4e82-Methyl-1,2,3,4-<0.5 + 2.4e2 1.6e2 5.8e2butanetetrol+ Experiment carried out using the bead-mobility technique.* Experiment carried out using the poke-and-flow technique.976.3. Results and discussionFigure 6.2: Plots of log10(viscosity) vs. (a) relative humidity and (b) bothmolar mass and number of saccharide units for glucose, sucrose, raffinose,and maltohexaose. Results determined in the current study using the bead-mobility technique are shown using circle symbols, and those determined us-ing the poke-and-flow technique are shown using squares, with filled squaresrepresenting upper limits of viscosity and open squares representing lowerlimits of viscosity, with y-error bars representing 95 % confidence intervalsfor both techniques, as detailed for Table 3. Also included are literatureviscosity values for sucrose (Fo¨rst et al., 2002; Power et al., 2013; Quintaset al., 2006; Swindells et al., 1958; Telis et al., 2007) and glucose (Achardet al., 1992; Barbosa-Ca´novas et al., 2007; Haynes, 2015), with the viscosityof glucose at 47 % shown in (b) being determined using a polynomial fitof the literature data. The viscosity of water is added to (a), and shadedregions are added to (a) and (b) to guide the readers eye.and 2 x 109 Pa s. At 77-80 % RH, glucose was determined to have a viscosityof 1 x 10−1 Pa s, and maltohexaose, which is comprised of six saccharideunits, was determined to have a viscosity between 3 x 103 and 2 x 105 Pa s.986.3. Results and discussion6.3.2 Measured viscosity of a tetraolThe viscosity of 2-methyl-1,2,3,4-butanetetrol, a compound containing fourhydroxyl functional groups, was measured at <0.5 % RH using the bead-mobility technique. Under these conditions the 2-methyl-1,2,3,4-butanetetrolhad a viscosity of 240 Pa s (95 % confidence intervals of 160-575 Pa s) (Ta-ble 6.4), roughly between that of ketchup and peanut butter (Koop et al.,2011). Recently Song et al. (2015) studied SOM produced via the oxidationof isoprene and determined the SOM to have a viscosity between 2 x 104and 4 x 106 Pa s at <1 % RH. This viscosity is greater than that measuredfor the tetraol compound measured here, suggesting that SOM producedvia the oxidation of isoprene contains some components that are of greaterviscosity than that of the tetraol.Plotted in Figure 6.3 is the viscosity of 2-methyl-1,2,3,4-butanetetrol,along with 2-methyl-butane and 1-hydroxy-2-methyl-butane, as a functionof the number of hydroxyl groups in the structure. Also included on thefigure is a line that passes through the points of 2-methylbutane and 1-hydroxy-2-methyl-butane. The viscosity of 2-methyl-1,2,3,4-butanetetrol is5.9-6.4 orders of magnitude greater than that of 2-methylbutane, and is0.6-1.2 orders of magnitude greater than would be predicted by the extrap-olated line, suggesting that the increase in log(viscosity) may not be a linearfunction of the number of hydroxyl functional groups in its structure.Shown in Figure 6.4 is a plot similar to Fig. 6.3, which shows the viscosityas a function of the number of hydroxyl groups added to n-alkanes with threeto eight carbons. Fig. 6.3 illustrates that the viscosity increases by approxi-mately 1.2 orders of magnitude when a n-alkane with three to eight carbonsincorporates one hydroxyl group, by approximately 2.5 orders of magnitudewhen a C3-C8 alkane incorporates two hydroxyl functional groups, and byapproximately 4 orders of magnitude when a C3 or C6 alkane incorporatesthree hydroxyl functional groups. Based on these values, oxidation reactionsin the atmosphere that lead to the addition of a hydroxyl group to alkanesshould lead to at least one order of magnitude increase in viscosity.996.3. Results and discussionFigure 6.3: A plot of log(viscosity) vs. number of hydroxyl functional groupsadded to 2-methylbutane. The symbols correspond to 2-methylbutane(square), 1-hydroxy-2-methylbutane (upward pointing triangle), and 2-methyl-1,2,3,4-tetraol (leftward pointing triangle). The solid line is a fitof the data for 2-methylbutane and 1-hydroxy-2methylbutane.Figure 6.4: Plot of log(viscosity) vs. number of hydroxyl groups addedto C3-C8 alkanes. Symbols correspond to alkanes (squares), monoalcohols(upward pointing triangles), diols (downward pointing triangles), and triols(diamonds), with lines drawn between compounds with the same number ofcarbons.1006.3. Results and discussion6.3.3 Comparison of the measured and predicted viscosityof a tetraolTwo QSPR models have been used to predict the viscosity of 2-methyl-1,2,3,4-butanetetrol. The first QSPR (Sastri and Rao, 1992) relates viscosityto the molecular structure and vapour pressure of a compound, and predictsa viscosity of 1 x 1010-5 x 1014 Pa s for 2-methyl-1,2,3,4-butanetetrol, 8-12orders of magnitude greater than the experimental value.The second QSPR (Marrero-Morejo´n and Pardillo-Fontdevila, 2000) re-lates viscosity to the molecular structure and molar mass of a compound,and predicts a viscosity of 4 x 101 Pa s for 2-methyl-1,2,3,4-butanetetrol,0.5-1.5 orders of magnitude lower than the experimental value.Figure 6.5 is a plot of experimental vs. predicted viscosity for 2-methyl-1,2,3,4-butanetetrol for the two QSPR models. Also included in this figureare the measured and predicted viscosities of the same organics shown inFigs. 6.3 and 6.4. Literature vapour pressure values (Cai et al., 2015;Cammenga et al., 1977; Perry and Green, 2008; Verevkin, 2004) were used topredict the viscosities of the alkanes, mono, di, and tri alcohols. The QSPRmodel of Sastri and Rao predicts the viscosity of compounds containing zeroFigure 6.5: Plot of experimental vs. predicted log(viscosity) of C3-C8 alka-nes, monoalcohols, and polyols for the models derived by (a) Sastri and Rao(1992) and (b) Marrero-Morejo´n and Pardillo-Fontdevila (2000). Dashed1:1 lines are shown on each plot to guide the readers eye.1016.4. Summaryor one hydroxyl functional groups well, however, it increasingly over-predictsthe viscosity of compounds as the number of hydroxyl functional groups in-creased, over-predicting the viscosity of compounds containing three hy-droxyl functional groups by approximately 6 orders of magnitude (Fig.6.5(a)). These over-estimations may be due to the model over-estimating theeffect of multiple hydroxyl compounds being present in the same molecule.In theory there could also be error in the calculated vapour pressure of 2-methyl-1,2,3,4-butanetetrol, however, this would not account for the largeover-estimation in the viscosity of compounds containing three hydroxylfunctional groups.Shown in Fig. 6.5(b) are predicted viscosities using the QSPR modelof Marrero-Morejo´n and Pardillo-Fontdevila (2000). The model providesmore accurate predictions than that of Sastri and Rao, with the predictedlog(viscosity / Pa s) of all compounds bar 2-methyl-1,2,3,4-butanetetrol be-ing within 0.2 of their experimental values.6.4 SummaryThe viscosity of particles consisting of a tetraol (2-methyl-1,2,3,4-butanetetrol)and particles containing one of three saccharides (glucose, raffinose and mal-tohexaose) mixed with water have been determined. These systems serveas proxies for highly oxidized components of ambient SOM. In addition,tetraols are a component of secondary organic material produced via thephotooxidation of isoprene, and saccharides have been observed in atmo-spheric organic particulate matter.The viscosity of saccharide-water particles was at least four orders ofmagnitude greater at a RH of 28 % than at 78 % RH, indicating that wateris acting as a plasticizer. At 28 % RH, the viscosity of saccharide-waterparticles was observed to increase by approximately 4-6 orders of magnitudeas molar mass of the saccharide is increased from 180 to 342 g mol−1 or thenumber of saccharide units increased from one to two. At 77-80 % RHthe viscosity increased by approximately 4-6 orders of magnitude as themolar mass of the saccharide increases from 180 to 991 g mol−1 or the1026.4. Summarynumber of saccharide units increased from one to six. These results suggestoligomerisation of highly oxidized compounds in atmospheric SOM couldlead to large increases in viscosity, and could be at least partially responsiblefor high viscosities in SOM.The tetraol studied under dry conditions was determined to have aviscosity of 240 Pa s, approximately 6 orders of magnitude greater than2-methylbutane, which shares a carbon backbone with 2-methyl-1,2,3,4-butanetetrol but contains zero hydroxyl functional groups. The measuredviscosity of 2-methyl-1,2,3,4-butanetetrol is lower than the viscosity of SOMproduced via the oxidation of isoprene, suggesting SOM produced via theoxidation of isoprene contains some components of greater viscosity thanthat of the tetraol. Finally, two quantitative structure-property relationshipmodels, which were previously developed to predict the viscosity of organicliquids, was used to predict the viscosity of 2-methyl-1,2,3,4-butanetetrol.The viscosity of 2-methyl-1,2,3,4-butanetetrol is vastly over-predicted by 8-12 orders of magnitude by the model of Sastri and Rao (1992), and is under-predicted by 0.6-1.2 orders of magnitude by the model of Marrero-Morejo´nand Pardillo-Fontdevila (2000).103Chapter 7Predicting the viscosity oforganic compounds usingtheir oxygen-to-carbonelemental ratio andsaturation vapourconcentration7.1 IntroductionAerosol particles are ubiquitous throughout the atmosphere and affect theEarth’s climate directly, through the scattering and absorption of incomingsolar radiation, and indirectly, by acting as nuclei for liquid and ice clouds(Stocker et al., 2013). Atmospheric particulate matter is formed in the atmo-sphere, in part through the oxidation of volatile organic compounds (VOCs)emitted from the planet’s surface. These oxidation reactions result in theformation of secondary organic material and SOM is typically estimated toaccount for 30-70 % of the mass fraction of submicron particulate matter inmost regions of the atmosphere (Kanakidou et al., 2005).Due to the complexity of SOM (see Section 1.2), explicit description ofits formation, evolution, and physical properties of SOM cannot currentlybe incorporated into large-scale models (Hallquist et al., 2009). As a re-sult, researchers have used simple methods to describe these processes and1047.1. Introductionproperties. Two properties that are currently being used to describe andmodel the formation, evolution, and hygroscopic properties of SOM are O:Cand saturation vapour concentration (C*) (Donahue et al., 2011, 2006, 2012;Jimenez et al., 2009; Napier et al., 2014; Zhao et al., 2015). C* defines theconcentration above which a species will partition to the particle phase (Don-ahue et al., 2009; Hallquist et al., 2009; Pankow, 1994), and is calculated fora given compound as follows,C∗ =P 0LMwς106RT(7.1), where PL0 is the saturated vapour pressure (units of atm) of the com-pound, Mw is its molar mass (g mol−1), ς is its activity coefficient (typicallyassumed to be 1), R is the ideal gas constant (J K−1 mol−1), and T is thetemperature (K).The use of O:C and C* is a convenient method for predicting the for-mation, evolution, and hygroscopic properties of SOM as they are simpleenough to be incorporated into large-scale models, and they may also bevalidated with field and laboratory experiments.The viscosity of SOM has recently become a topic of interest in thescientific literature as viscosity is important for predicting the environmentalimpact of SOM. For example viscosity can affect the mechanism and rateof growth of particles containing SOM (Riipinen et al., 2011; Shiraiwa andSeinfeld, 2012; Shiraiwa et al., 2013; Zaveri et al., 2014) as well as theirability to uptake water (Bones et al., 2012; Hawkins et al., 2014; Lu et al.,2014; Price et al., 2014; Tong et al., 2011). The importance of the viscosityof SOM is further outlined in Chapter 1.Due to the importance of the viscosity of SOM, researchers have begunto measure the viscosity of specific types of SOM under specific laboratoryconditions (Bateman et al., 2015; Pajunoja et al., 2014; Song et al., 2015;Zhang et al., 2015, and Chapters 3 and 5), however, tools to predict the vis-cosity of SOM particles for a range of environmental conditions are currentlylimited.The relationship between viscosity and the physical property of com-1057.1. Introductionpounds has been studied previously (Katritzky et al., 2010). For example,Suzuki et al. (1996) studied the correlation between 18 different physicalproperties and viscosity for a selection of 237 organic compounds. The physi-cal properties determined to correlate most strongly with viscosity were theenthalpy of vaporisation of a compound at its boiling point, the cohesiveenergy of a compound at 298 K, and the vapour pressure of a compoundat 293 K. Approaches to predict viscosity that combine multiple physicalproperties have also been used - for example, Katritzky et al. (2000) derivedan equation combining five physical properties, with the most importantproperty determined to be a measure of the proportion of the compoundavailable to form hydrogen bonds.The following investigates whether an organic compound’s O:C and C*are good predictors of its viscosity, which would potentially allow viscosityto be incorporated into atmospheric models. Additionally, the predictiveability of a multivariate relationship between viscosity and both O:C andC* is considered.A relationship has previously been established between a molecule’s O:Cand its glass transition temperature (Tg) (Koop et al., 2011), as has a re-lationship between the viscosity and vapour pressure of organic compounds(Sastri and Rao, 1992; Suzuki et al., 1996), however, neither the relationshipbetween viscosity and O:C, nor the relationship between viscosity and C*,have been studied. The common characteristics between O:C and C*, andphysical properties shown to be correlated with viscosity suggest O:C and C*may be good candidates for predicting the viscosity of organic compounds.A greater O:C ratio increases the proportion of a compounds structure thatis available to hydrogen bond, which was determined to be an importantmeasure by Katritzky et al. (2000). Vapour pressure is used to calculatea compounds C*, and has been determined by Suzuki et al. (1996) to bestrongly correlated to the viscosity of organic compounds. Also included inthe calculation of C* is molar mass which, though observed to give rise to alower correlation than vapour pressure, was also observed by Suzuki et al.(1996) to be correlated to viscosity.1067.2. Methods7.2 Methods7.2.1 Selection of compoundsBased on the elemental composition and functional groups of secondaryorganic material outlined in the introduction of this Chapter, and the goalof determining the relationship between viscosity, O: C and C*, the followingcriteria were used when selecting compounds from the literature:i) compounds must contain only carbon and hydrogen or only carbon,hydrogen, and oxygen atoms,ii) compounds must contain only the following functional groups: alka-nes, alkenes, alcohols, carboxylic acids, aldehydes, ketones, esters, ethers,and acid anhydrides, as well as both aromatic and non-aromatic cyclicmolecules, to be consistent with functional groups associated with SOM(see introduction to this Chapter),iii) compounds must have viscosities reported in the literature at 298 K,or viscosities reported at more than three other temperatures, with at leastone measurement being above 298 K and at least one measurement beingbelow 298 K, allowing their viscosity at 298 K to be interpolated using apolynomial fit,iv) compounds must have vapour pressures reported in the literature at298 K, or vapour pressures reported at more than three other temperatures,with at least one measurement being above 298 K and at least one mea-surement being below 298 K, allowing their vapour pressure at 298 K to beinterpolated using a polynomial fit.Application of these criteria resulted in data for 156 suitable compoundsthat range in molar mass from 32-600 g mol−1 and in O:C from 0.0 to2.0, consistent with the properties of components of SOM outlined in theintroduction. The compounds also range in log(C* / µg m−3) from 0 to10, and in log(η / Pa s) from -4 to 0.5. The compounds, along with theirproperties, are detailed in Table C.1, whilst Table 7.1 provides a summaryof the functional groups represented in the compounds.Components of SOM produced during the ozonolysis of α-pinene arepredicted to have log (C* / µg m−3) at least as low as 2.5 (Jimenez et al.,1077.2. MethodsTable 7.1: A summary of the functional groups featured inthe 156 compounds used to determine the relationship be-tween viscosity and either O:C or C*. A compound withmultiple oxygen containing functional groups is counted oncefor each type of functional group it contains, and once in themultiple oxygen containing functional group row. As such,the total sums to >156.Organic compound classNumber of compoundscontaining functional groupAlkane 23Alkene 21Cyclic (non-aromatic) 18Cyclic (aromatic) 19Alcohol 36Acid 10Acid anhydride 2Aldehyde 9Ketone 10Ester 19Ether 16Multiple oxygen containing19functional groups2009), and may be significantly lower (Schobesberger et al., 2013). Theviscosity of individual components has yet to be measured, though α-pinenehas a log(η / Pa s) of -2.89 and experimental studies of the total SOMproduced via the ozonolysis of α-pinene suggest it to be of log(η / Pa s) ≥6(Zhang et al., 2015, and Chapter 4).7.2.2 Parameterisation of dataParameterisations were determined using Origin software. Linear fits wereused to determine the relationship between log(η) and O:C and betweenlog(η) and log(C*), whilst a multiple linear regression was used to determinethe relationship between log(η) and a combination of O:C and log(C*). Theform of each of these parameterisations is detailed in Table 7.2.1087.3. ResultsTable 7.2: Summary of the parameterisations used in this study. In each caselog(η / Pa s) is used as the dependent variable, and O:C and log(C* / µg m-3)are the independent variables. RMSE refers to the root mean square errorbetween the log(η / Pa s) predicted by the parameterisation for each of thecompounds in the study and the experimental viscosities.Equation form, including Values of constantsStatistical valuesindependent variables (standard error)Log10(η) = a + b(O:C)R value = 0.31a = -3.09 (6.74e-2) Adjusted R2 = 0.09b = 0.776 (0.189) p-value <0.001RMSE = 0.49Log10(η) = a + b x Log10(C*)R value = -0.87a = 0.164 (0.141) Adjusted R2 = 0.76b = -0.407 (1.83e-2) p-value <0.001RMSE = 0.23a = -2.49e-4 (0.147)R value = 0.88Log10(η) = a + b(O:C)b = 0.301 (9.72e-2)Adjusted R2 = 0.77+ c x Log10(C*)c = -0.394 (1.83e-2)p-value <0.001RMSE = 0.267.3 Results7.3.1 The relationship between viscosity and O:CFigure 7.1(a) is a plot of log(η) vs O:C for the compounds used for this study(Table C.1). Though a general trend can be observed, with log(η) and O:Cbeing directly related, the relationship has a R value of 0.31, suggesting O:Cexplains only ≈9% of the observed variance in log(η / Pa s). The equationof the best fit to Fig. 7.1(a) is detailed in Table 7.2. Figure 7.1(b) is a plotof experimental viscosities vs. the viscosities predicted using the best fitequation. The predicted viscosities have a root mean square error (RMSE)of 0.49 units of log(η / Pa s), with the predicted log(η / Pa s) of 58 of the156 compounds observed to deviate by >0.5 from their experimental values.1097.3. ResultsFigure 7.1: Plots detailing the relationship between viscosity (log(η / Pas)) and ((a) and (b)) O:C, and ((c) and (d)) log saturation vapour con-centration (log(C*)) for a range of compounds comprised only of carbonand hydrogen or carbon, hydrogen, and oxygen atoms, and containing at-mospherically relevant functional groups. Lines of best fit are included onplots (a) and (c). 1:1 lines are included on plots (b) and (d) to guide theeye of the reader. Statistical values from the plots are reported in Table 7.2.1107.3. Results7.3.2 The relationship between viscosity and C*Shown in Figure 7.1(c) is a plot of log(η) vs log(C*) for the compounds usedin this study (Table C.1). The linear regression between log(η) and log(C*)exhibits a stronger correlation than that between log(η) and O:C, with a Rvalue of -0.87 (Table 7.2) suggesting log(C*) explains ≈76 % of the variationin log(η) for the compounds used in this study, a much greater proportionthan O:C. This correlation occurs over a wide range of values, with viscosi-ties spanning almost 5 orders of magnitude, and C* values spanning morethan 10 orders of magnitude. Figure 7.1(d) is a plot of experimental viscosi-ties and the viscosities predicted using a compounds C*. A much tightergrouping about the 1:1 line is observed than for O:C (Fig. 7.1(b)), and thebest fit equation (Table 7.2) predicts viscosities with a RMSE of 0.23 unitsof log(η), smaller than that predicted using O:C. Of the 156 compoundsstudied, the predicted log(η) of 20 compounds deviated by >0.5 from theirexperimental viscosities.Suzuki et al. (1996) has previously demonstrated a strong correlationbetween viscosity and vapour pressure for a range of organic compounds.Vapour pressure and C* are related as detailed in equation 7.1, and thusthe results here can be considered consistent with those of Suzuki et al.(1996).7.3.3 Multivariate relationship between viscosity and bothO:C and C*A tool often used in modelling is multiple linear regression (MLR), whichsimultaneously optimises the weight given to numerous independent proper-ties to create a single equation to estimate the value of a dependent property.The MLR approach has been applied to the study of the viscosity of organiccompounds previously, using models that contain up to nine properties (Ka-tritzky et al., 2010). As such, a combination of O:C and C* has been usedto determine whether they result in an improved relationship with viscosity.Figure 7.2(a) is a 3-D contour plot with O:C and log(C*) on the x - andy-axes, respectively, and the colour scale on the plot indicating the viscosity1117.3. ResultsFigure 7.2: (a) is a 3-D contour plot with O:C and log10(C* / µg m−3) onthe x - and y-axes, respectively, and the colour scale on the plot indicatingthe viscosity predicted using an MLR with experimental log(η / Pa s) asthe dependent variable and O:C and log(C* / µg m−3) as the independentvariables.(b) is a residual plot showing the difference for each compoundbetween its experimental viscosity and the viscosity predicted by the regres-sion. (c) is a plot of the predicted viscosities for each of the compounds vs.their experimental viscosities, with a 1:1 line to guide the readers eye. 1127.4. The predicted viscosity of products of α-pinene ozonolysispredicted using an MLR with log(η / Pa s) as the dependent variable, andlog(C* / µg m−3) and O:C as independent variables (Table 7.2). Also plottedare the O:C and log(C*) values of the compounds used in this study.Fig. 7.2(b) is a residual plot showing the difference between the ex-perimentally determined viscosities and those predicted by the MLR thatincorporates O:C and log(C*) as predictors of log(η / Pa s). The RMSEof the predicted viscosities (0.26 log units) is similar, though slightly largerthan the RMSE of C* alone (0.23 log units). Though the correlation to vis-cosity between the MLR featuring both O:C and C* is higher (adjusted R2= 0.77) than the parameterisation featuring solely C* (adjusted R2 = 0.76),the small increase suggests O:C plays at best a negligible role in improvingthe ability to predict viscosities using log(C*).The viscosities predicted by the regression model are plotted againsttheir experimental viscosities in Fig. 7.2(c). This data is also, expectedly,centred on the 1:1 line, and a tight grouping is observed, with the predictedviscosities of only 17 of the 156 compounds deviating by >0.5 log units fromtheir experimental viscosities.7.4 The predicted viscosity of products ofα-pinene ozonolysisThe plot of log(η / Pa s) vs log(C* / µg m−3) shown in Fig. 7.1(c) is plottedagain as Figure 7.3, with the pink area representing the area bounded bythe 95 % prediction interval of the linear relationship. The 95 % predictionintervals suggest log(η / Pa s) to range from 0.9 to -0.6 when log(C* / µgm−3) equals 0, and -3.2 to -4.6 when log(C* / µg m−3) equals 10.Shown on the bottom x -axis is the structure of α-pinene, as well asthree products formed during the ozonolysis of α-pinene, cis-pinic acid, cis-pinonic acid, and pinonaldehyde. Shown along the top x -axis are the regionsof the log(C*) scale attributed to semivolatile organic compounds (SVOCs),intermediate volatility organic compounds (IVOCs), and volatile organiccompounds (VOCs), as defined by Donahue et al. (2009, 2012).1137.4. The predicted viscosity of products of α-pinene ozonolysisFigure 7.3: Plot of log(η / Pa s) vs. log(C* / µg m−3) for the compoundsdetailed in Table C.1. The shaded pink region represents the 95 % predic-tion interval of the linear regression between log(C* / µg m−3) vs. log(η /Pa s). Shown on the x -axis are the regions of the log(C*) scale attributedto semivolatile organic compounds (SVOCs), intermediate volatility organiccompounds (IVOCs), and volatile organic compounds (VOCs). Also shownis the position of α-pinene (dark blue circle) (per Donahue et al., 2009),as well as first generation products from α-pinene ozonolysis, cis-pinic acid(green), cis-pinonic acid (red), and pinonaldehyde (cyan) (C* values calcu-lated from Hartonen et al., 2013).Compounds in the SVOC region may be in either the particle or gasphase under ambient atmospheric conditions, whilst IVOCs and VOCs arealmost exclusively in the gas phase. The C* of α-pinene suggests it should1147.5. Summarybe classified as a VOC, whilst those of the three reaction products shown onFig. 7.3 suggest they should be classified as IVOCs. The 95 % predictionlimit of the regression equation suggests the upper limit of log(η / Pa s)for IVOCs to be -0.2, suggesting both IVOCs and VOCs occupy the liquidregime of viscosities (Shiraiwa et al., 2011a, and Chapter 3). These valuesare consistent with the multiple linear regression incorporating both O:Cand C* shown in Fig. 7.2, which predicts the log(η / Pa s) of α-pineneto be -3.0, and the log(η / Pa s) of cis-pinic acid, cis-pinonic acid, andpinonaldehyde to be between -1.4 and -2.3, all of which are in the liquidregime of viscosities.The lower limit of log(C* / µg m-3) for SVOCs is defined as -0.5 (Don-ahue et al., 2012). Though the dataset of compounds used here, and shownin Fig. 7.3, only covers compounds of log(C* / µg m−3) >0, extrapolationof the regression equation and 95 % prediction limits suggests an upperlimit of log(η / Pa s) of 1.1 when log(C* / µg m−3) equals -0.5, suggestingthat SVOCs also occupy the liquid regime of viscosities, although additionalstudies are needed to confirm whether this is the case.7.5 SummaryLiterature data has been used to determine the relationship between viscos-ity and O:C, as well as between viscosity and C*, for 156 organic compounds.The compounds studied have viscosities that span a range of almost five or-ders of magnitude, O:C values that range from 0.0-2.0, and values of C*that span a range of ten orders of magnitude.The relationship between log(viscosity) and O:C was determined to havea R value of 0.31, suggesting O:C explained 9 % of the observed variance inlog(viscosity). The linear relationship between log(viscosity) and O:C wassubsequently used to predict the viscosity of each of the compounds. Thepredicted viscosities had a RMSE of 0.49 for log(viscosity), and of the 154compounds studied, 58 had a predicted log(viscosities / Pa s) that devi-ated by >0.5 from its experimental log(viscosity / Pa s). The relationshipbetween log(viscosity) and C* was determined to be stronger that that be-1157.5. Summarytween log(viscosity) and O:C, having a R value of -0.87, suggesting log(C*)explained 76 % of the observed variance in log(viscosity). The linear rela-tionship between log(viscosity) and log(C*) was subsequently used to pre-dict the viscosity of each of the compounds. The predicted viscosities had aRMSE of 0.23 for log(viscosity), and of the 154 compounds studied, 20 had apredicted log(viscosities / Pa s) that deviated by >0.5 from its experimentallog(viscosity / Pa s).Finally, MLR was used to determine the relationship between viscosityand a combination of O:C and log(C*). The statistical results were similar tothat observed for C* alone, with an R value of 0.88, suggesting the combina-tion of O:C and C* explained 77 % of the observed variance in log(viscosity).The viscosities predicted by the MLR had a RMSE of 0.26 for log(viscosity),and of the 154 compounds studied, 17 had a predicted log(viscosities / Pas) that deviated by >0.5 from its experimental log(viscosity / Pa s).These results suggest, for the compounds studied here, C* is a potentiallyuseful property with which to predict viscosities, whilst O:C was observedto play a negligible role in improving the ability to predict viscosities.Whilst the parameterisations and results presented here provide an ini-tial framework for predicting viscosities of the components observed in SOM,a greater number of experimental values for compounds of higher viscosity,lower C*, and containing multiple oxygen containing functional groups arerequired in order to further evaluate the usefulness of C* for predicting theviscosity of SOM.116Chapter 8Conclusions and future work8.1 Conclusions8.1.1 Development and validation of the poke-and-flowtechniqueChapters 2-4 detail the development and validation of the poke-and-flowtechnique combined with simulations of fluid flow, a novel method for mea-suring small samples of high viscosities. The technique is capable of mea-suring a wide range of viscosities yet requires only micrograms of material,presenting the opportunity to provide the first direct measurements of theviscosity of SOM.In Chapter 3 the technique was used to determine the upper limit ofviscosity for particles that formed a half-torus geometry on being pokedby the needle, and subsequently exhibited flow to re-attain a hemisphericalmorphology. An initial validation showed the simulated viscosities to be inagreement with literature viscosities for sucrose-water mixtures with viscosi-ties of approximately 104 Pa s. The technique was also used to determinethe lower limit of viscosity for particles that cracked upon impaction by theneedle, that exhibit no observable flow over subsequent hours. The simu-lated lower limit of viscosity for raffinose-water particles at low RHs was 5x 108 Pa s, not inconsistent with literature data suggesting them to be ofviscosity >1012 Pa s.In Chapter 4 simulations of particles that took on a half-torus geometryand subsequently exhibited flow were extended to also include an estimatedlower limit of viscosity. The technique was also validated across a widerrange of viscosities, with simulated viscosities of sucrose-water particles ob-1178.1. Conclusionsserved to be in agreement with literature values at least across the range of≈5 x 102 - ≈ 3 x 106 Pa s. Simulated viscosities obtained using the poke-and-flow technique combined with simulations of fluid flow were also shownto be in agreement with measured viscosities for two commercially availablepolybutene high viscosity standards.8.1.2 Measuring the viscosity of SOM and atmosphericallyrelevant compoundsThe viscosity of SOM produced via the oznolysis of α-pinene was measuredin Chapters 3 and 5, and the viscosity of atmospherically relevant com-pounds, a tetraol and numerous saccharide-water mixtures was measured inChapter 6. These results add to the limited values in literature for SOM aswell as highly oxidized atmospherically relevant compounds.In Chapter 3, the poke-and-flow technique combined with simulations offluid flow was used along with a second novel technique, the bead-mobilitytechnique, which was also detailed in Chapter 2. The viscosity of the water-soluble component of the SOM was determined to be strongly dependenton RH, and range in viscosity from approximately 100 Pa s at 90 % RH,roughly equivalent to that of honey, to >108 Pa s at ≤30 % RH, greaterthan that of tar pitch.In Chapter 5, the viscosity of the whole SOM, comprised of both thewater-soluble and water-insoluble components, was collected and measuredusing the poke-and-flow technique in combination with simulations of fluidflow. The effect on viscosity of varying experimental conditions was alsostudied. The viscosity of the total SOM was determined to be dependent onRH, with the viscosity of the sample produced under the most atmospher-ically relevant conditions observed to be three to four orders of magnitudegreater at <0.5 % RH than at 50 % RH. The viscosity of the total SOMwas determined to be dependent on the SOM particle mass concentrationat which the SOM was produced, with viscosities measured at <0.5 % RHranging by greater than an order of magnitude for the range of productionmass concentrations studied. Finally, the viscosity of the whole SOM was1188.1. Conclusionsdetermined to be lower than the viscosity of the water-soluble componentof the SOM.In Chapter 6, the bead-mobility and poke-and-flow technique in com-bination with simulations of fluid flow were used to measure the viscos-ity of a tetraol and some saccharide-water mixtures. These measurementsadd to the limited measurements that exist in literature for the viscosity ofhighly oxidized atmospherically relevant compounds. The viscosities of thesaccharide-water mixtures were determined to be dependent upon RH, withviscosities determined to be at least four orders of magnitude greater at 28% RH than at 78 % RH. The viscosities of the saccharide-water mixtureswas also determined to be dependent upon the molar mass of the saccharide.The viscosity measurement of the tetraol studied were compared to litera-ture data to demonstrate viscosity is strongly dependent upon the numberof hydroxyl functional group in a molecule, suggesting oxidation reactionsin the atmosphere that lead to the addition of a hydroxyl group to alkanesshould lead to at least one order of magnitude increase in viscosity.8.1.3 Predicting the viscosity of SOMChapter 7 details the application of literature data in order to determine therelationship between viscosity and elemental oxygen-to-carbon ratio (O:C),as well as between viscosity and saturation vapour concentration (C*), themass based equivalent of saturation vapour pressure. The relationship be-tween log(viscosity) and O:C had a R value of 0.31, compared with a Rvalue of -0.87 between log(viscosity) and log(C*). A multivariate linear re-gression comprising both O:C and log(C*) was similar to log(C*) in terms ofits correlation to, and ability to predict, viscosities, suggesting O:C playeda negligible role in improving the ability to predict viscosities. Finally, therelationship between log(viscosity) and log(C*) suggests that volatile or-ganic compounds (VOCs), and intermediate volatility organic compounds(IVOCs) will be liquids (viscosities <102 Pa s). An extrapolation of thisrelationship a short distance beyond the compounds included in the studyof Chapter 7 suggests the same to be true for at least some semivolatile1198.2. Directions for future workorganic compounds (SVOCs).8.2 Directions for future workThe samples and relationships studied here provide a starting point for pre-dicting the viscosity of SOM in the atmosphere, however, further experi-ments would be beneficial.Studies of the physical properties of SOM typically report the values forSOM produced and studied under a limited set of conditions, for exampleproduced at dry RH and studied at room temperature (see Chapter 3, aswell as Abramson et al., 2013; Cappa and Wilson, 2011; Perraud et al.,2012; Zhang et al., 2015), and caution must be taken when applying theseresults to the atmosphere, where conditions such as RH and temperatureexhibit significant spatial and temporal variation. The results presented inChapter 5 outline the effect on the viscosity of SOM of varying experimentalconditions, and Kidd et al. (2014) have reported changing the RH at whichSOM is produced also plays a role in dictating the chemical compositionof the SOM. Further, viscosity and temperature are known to be inverselyrelated (Viswanath et al., 2007). Whilst the majority of laboratory studiesare performed at room temperature, the temperature in the tropospherecan be as low as -56 0C (Seinfeld and Pandis, 2006), suggesting SOM inthe atmosphere would be of higher viscosity than at room temperature.Initial observations supporting this theory have recently been provided byJa¨rvinen et al. (2015), who observed the transition of SOM particles fromnon-spherical to spherical as RH was increased, and determined an increasein the RH at which the transition occurred as temperature was decreased.This relationship was also in agreement with those prior estimations (Koopet al., 2011; Wang et al., 2015). Due to the effect on viscosity of varyingexperimental parameters, the effect on viscosity of varying production RHshould be studied in more detail, as well as the effect on viscosity of thetemperature at which SOM is both produced and studied.In order to better understand and predict the viscosity of the individ-ual components of SOM, additional measurements of the viscosity of at-1208.2. Directions for future workmospherically relevant compounds are required. For example, the recentexperimental discovery and identification of extremely low volatility organiccompounds in SOM, which may be of O:C ≥1.0 and account for a signif-icant fraction of SOM mass (Ehn et al., 2012, 2014; Praplan et al., 2015;Schobesberger et al., 2013), has highlighted the lack of experimental viscositymeasurements of compounds that contain certain types of oxygen containingfunctional groups, compounds containing multiple oxygen containing func-tional groups, or compounds that have low volatilities. For example, ELVOCcompounds are known to contain multiple hydroperoxide functional groups,whilst organonitrate and organosulfate functional groups are also known tobe present in components of SOM (e.g. Russell et al., 2011), however, fewliterature measurements of viscosity are available for compounds containingthese functional groups. Furthermore, due to a lack of available literaturedata no compounds considered to be of low volatility (per Donahue et al.,2009, 2012) could be included in the study in Chapter 7. Literature datafor a greater number of compounds containing multiple oxygen containingfunctional groups or of low volatility would allow the relationship betweenviscosity and saturation vapour concentration (C*) observed in Chapter 7 tobe established over a wider range of values. The results presented in Chap-ters 4 and 6 here, as well as the viscosities of sucrose reported by Power et al.(2013) and the vapour pressures of atmospherically relevant compounds re-ported by Bilde et al. (2015), are some of the examples of early steps inproviding some of these measurements.121BibliographyAbramson, E., Imre, D., Bera´nek, J., Wilson, J., and Zelenyuk, A.: Ex-perimental determination of chemical diffusion within secondary organicaerosol particles, Phys. Chem. Chem. Phys., 15, 2983–2991, 2013.Achard, C., Gros, J. B., and Dussap, C. G.: Pre´diction de l’activite´ de l’eau,des tempe´ratures d’e´bullition et de conge´lation de solutions aqueuses desucres par un mode`le UNIFAC, Ind. Agric. Ailment., 109, 93–101, 1992.Adler, G., Koop, T., Haspel, C., Taraniuk, I., Moise, T., Koren, I., Heiblum,R. H., and Rudich, Y.: Formation of highly porous aerosol particles byatmospheric freeze-drying in ice clouds, Proc Natl Acad Sci USA, 110,20 414–20 419, 2013.Aiken, A. C., Decarlo, P. F., Kroll, J. H., Worsnop, D. R., Huffman, J. A.,Docherty, K. S., Ulbrich, I. M., Mohr, C., Kimmel, J. R., Sueper, D., Sun,Y., Zhang, Q., Trimborn, A., Northway, M., Ziemann, P. J., Canagaratna,M. R., Onasch, T. B., Alfarra, M. R., Pre´voˆt, A. S. H., Dommen, J., Du-plissy, J., Metzger, A., Baltensperger, U., and Jimenez, J. L.: O/C andOM/OC Ratios of Primary, Secondary, and Ambient Organic Aerosolswith High-Resolution Time-of-Flight Aerosol Mass Spectrometry, Envi-ron. Sci .Technol., 42, 4478–4485, 2008.Aldrich Chemical Company: Technical bulletin AL-128, URLhttps://www.sigmaaldrich.com/content/dam/sigma-aldrich/docs/Aldrich/Bulletin/al_techbull_al128.pdf, 1996.Angell, C. A.: Formation of glasses from liquids and biopolymers, Science,267, 1924–1935, 1995.122BibliographyArneth, A., Monson, R. K., Schurgers, G., Niinemets, U¨., and Palmer, P. I.:Why are estimates of global terrestrial isoprene emissions so similar (andwhy is this not so for monoterpenes)?, Atmos. Chem. Phys., 8, 4605–4620,2008.Aschenbrenner, O., Supasitmongkol, S., Taylor, M., and Styring, P.: Mea-surement of vapour pressures of ionic liquids and other low vapour pressuresolvents, Green Chemistry, 11, 1217, 2009.Aschmann, S. M. and Atkinson, R.: Kinetics of the gas-phase reactions ofthe OH radical with selected glycol ethers, glycols, and alcohols, Int. J.Chem. Kinet., 30, 533–540, 1998.Baltensperger, U., Dommen, J., Alfarra, M. R., Duplissy, J., Gaeggeler, K.,Metzger, A., Facchini, M. C., Decesari, S., Finessi, E., Reinnig, C., Schott,M., Warnke, J., Hoffmann, T., Klatzer, B., Puxbaum, H., Geiser, M., Savi,M., Lang, D., Kalberer, M., and Geiser, T.: Combined determinationof the chemical composition and of health effects of secondary organicaerosols: The POLYSOA project, J. Aerosol. Med. Pulm. Drug Deliv.,21, 145–154, 2008.Barbosa-Ca´novas, G. V., Fontana, A. J., Schmidt, S. J., and Labuza, T. P.,eds.: Water activity in foods: fundamentals and applications, Wiley-Blackwell, Hoboken, NJ, USA, 1 edn., 2007.Bateman, A. P., Nizkorodov, S. A., Laskin, J., and Laskin, A.: High-resolution electrospray ionization mass spectrometry analysis of water-soluble organic aerosols collected with a particle into liquid sampler, Anal.Chem., 82, 8010–8016, 2010.Bateman, A. P., Bertram, A. K., and Martin, S. T.: Hygroscopic influenceon the semisolid-to-liquid transition of secondary organic materials, J.Phys. Chem. A, 119, 4386–4395, 2015.Baudry, J., Charlaix, E., Tonck, A., and Mazuyer, D.: Experimental evi-dence for a large slip effect at a nonwetting fluid-solid interface, Langmuir,17, 5232–5236, 2001.123BibliographyBaustian, K. J., Wise, M. E., Jensen, E. J., Schill, G. P., Freedman, M. A.,and Tolbert, M. A.: State transformations and ice nucleation in amor-phous (semi-)solid organic aerosol, Atmos. Chem. Phys., 13, 5615–5628,2013.Berkemeier, T., Shiraiwa, M., Po¨schl, U., and Koop, T.: Competition be-tween water uptake and ice nucleation by glassy organic aerosol particles,Atmos. Chem. Phys., 14, 12 513–12 531, 2014.Bilde, M., Barsanti, K., Booth, M., Cappa, C. D., Donahue, N. M.,Emanuelsson, E. U., McFiggans, G., Krieger, U. K., Marcolli, C., Top-ping, D., Ziemann, P., Barley, M., Clegg, S., Dennis-Smither, B., Hal-lquist, M., Hallquist, A˚. M., Khlystov, A., Kulmala, M., Mogensen, D.,Percival, C. J., Pope, F., Reid, J. P., Ribeiro da Silva, M. A. V., Roseno-ern, T., Salo, K., Soonsin, V. P., Yli-Juuti, T., Prisle, N. L., Pagels, J.,Rarey, J., Zardini, A. A., and Riipinen, I.: Saturation vapor pressures andtransition enthalpies of low-volatility organic molecules of atmospheric rel-evance: from dicarboxylic acids to complex mixtures, Chem. Rev., 115,4115–4156, 2015.Blunk, R. and Wilkes, J. O.: Surface-tension-driven flows of coatings: Bond-line readout formation, J. Coating. Tech., 73, 63–71, 2001.Bodsworth, A., Zobrist, B., and Bertram, A. K.: Inhibition of efflorescencein mixed organic–inorganic particles at temperatures less than 250 K,Phys. Chem. Chem. Phys., 12, 12 259, 2010.Bond, T. C., Doherty, S. J., Fahey, D. W., Forster, P. M., Bernsten, T.,DeAngelo, B. J., Flanner, M. G., Ghan, S., Ka¨rcher, B., Koch, D., Kinne,S., Kondo, Y., Quinn, P. K., Sarofim, M. C., Schultz, M., Venkataraman,C., Zhang, H., Zhang, S., Bellouin, N., Guttikunda, S. K., Hopke, P. K.,Jacobson, M. Z., Kaiser, J. W., Klimont, Z., Lohmann, U., Schwarz, J. P.,Shindell, D., Storelvmo, T., Warren, S. G., and Zender, C. S.: Boundingthe role of black carbon in the climate system: A scientific assessment, J.Geophys. Res. Atmos., 118, 5380–5552, 2013.124BibliographyBones, D. L., Reid, J. P., Lienhard, D. M., and Krieger, U. K.: Comparingthe mechanism of water condensation and evaporation in glassy aerosol,Proc. Nat. Acad. Sci. U.S.A., 109, 11 613–11 618, 2012.Booth, A. M., Murphy, B., Riipinen, I., Percival, C. J., and Topping, D. O.:Connecting bulk viscosity measurements to kinetic limitations on attain-ing equilibrium for a model aerosol composition, Environ. Sci .Technol.,48, 9298–9305, 2014.Cai, C., Stewart, D. J., Reid, J. P., Zhang, Y.-h., Ohm, P., Dutcher, C. S.,and Clegg, S. L.: Organic Component Vapor Pressures and Hygroscopic-ities of Aqueous Aerosol Measured by Optical Tweezers, J. Phys. Chem.A, 119, 704–718, 2015.Cammenga, H. K., Schulze, F. W., and Theuerl, W.: Vapor pressure andevaporation coefficient of glycerol, J. Chem. Eng. Data, 22, 131–134, 1977.Cappa, C. D. and Wilson, K. R.: Evolution of organic aerosol mass spec-tra upon heating: implications for OA phase and partitioning behavior,Atmos. Chem. Phys., 11, 1895–1911, 2011.Cappa, C. D., Lovejoy, E. R., and Ravishankara, A. R.: Evidence for liquid-like and nonideal behavior of a mixture of organic aerosol components,Proc. Nat. Acad. Sci. U.S.A., 105, 18 687–18 691, 2008.Cavalli, F., Facchini, M. C., Decesari, S., Emblico, L., Mircea, M., Jensen,N. R., and Fuzzi, S.: Size-segregated aerosol chemical composition at aboreal site in southern Finland, during the QUEST project, Atmos. Chem.Phys., 6, 993–1002, 2006.Champion, D., Hervet, H., Blond, G., Le Meste, M., and Simatos, D.: Trans-lational diffusion in sucrose solutions in the vicinity of their glass transitiontemperature, J. Phys. Chem. B, 101, 10 674–10 679, 1997.Chan, A., Chan, M. N., and Surratt, J. D.: Role of aldehyde chemistry andNOx concentrations in secondary organic aerosol formation, AtmosphericChemistry & Physics, 10, 7169–7188, 2010.125BibliographyChen, Q., Liu, Y., Donahue, N. M., Shilling, J. E., and Martin, S. T.:Particle-phase chemistry of secondary organic material: modeled com-pared to measured O:C and H:C elemental ratios provide constraints,Environ. Sci .Technol., 45, 4763–4770, 2011a.Chen, X. and Hopke, P. K.: Secondary organic aerosol from α-pinene ozonol-ysis in dynamic chamber system, Indoor Air, 19, 335–345, 2009.Chen, Y., Randerson, J. T., Morton, D. C., DeFries, R. S., Collatz, G. J.,Kasibhatla, P. S., Giglio, L., Jin, Y., and Marlier, M. E.: Forecasting fireseason severity in South America using sea surface temperature anomalies,Science, 334, 787–791, 2011b.Cheng, J. T. and Giordano, N.: Fluid flow through nanometer-scale chan-nels, Phys. Rev. E, 65, 031 206, 2002.Choi, C.-H. and Kim, C.-J.: Large slip of aqueous liquid flow over a nano-engineered superhydrophobic surface, Phys. Rev. Lett., 96, 066 001, 2006.Christoffersen, T. S., Hjorth, J., Horie, O., Jensen, N. R., Kotzias, D.,Molander, L. L., Neeb, P., Ruppert, L., Winterhalter, R., Virkkula, A.,Wirtz, K., and Larsen, B. R.: cis-Pinic acid, a possible precursor fororganic aerosol formation from ozonolysis of α-pinene, Atmos. Environ.,32, 1657–1661, 1998.Churaev, N. V., Sobolev, V. D., and Somov, A. N.: Slippage of liquids overlyophobic solid surfaces, J. Colloid. Interface. Sci., 97, 574–581, 1984.Claeys, M., Graham, B., Vas, G., Wang, W., Vermeylen, R., Pashynska,V., Cafmeyer, J., Guyon, P., Andreae, M. O., Artaxo, P., and Maenhaut,W.: Formation of secondary organic aerosols through photooxidation ofisoprene, Science, 303, 1173–1176, 2004.Cocker IIII, D. R., Clegg, S. L., Flagan, R. C., and Seinfeld, J. H.: Theeffect of water on gas–particle partitioning of secondary organic aerosol.Part I: α-pinene/ozone system, Atmos. Environ., 35, 6049–6072, 2001.126BibliographyCorti, H. R., Frank, G. A., and Marconi, M. C.: Diffusion-viscosity decou-pling in supercooled aqueous trehalose solutions., J. Phys. Chem. B, 112,12 899–12 906, 2008.Craig, V. S., Neto, C., and Williams, D. R.: Shear-dependent boundary slipin an aqueous Newtonian liquid, Phys. Rev. Lett., 87, 054 504, 2001.Czechowski, G. and Jadzyn, J.: The viscous properties of diols.II. 1,2- and1,5-pentanediol in water and 1-pentanol solutions, Z. Naturforsch., 58a,321–324, 2003.Darke, C. S., Knowelden, J., Lacey, J., and Milford Ward, A.: Respiratorydisease of workers harvesting grain, Thorax, 31, 294–303, 1976.Daubert, T. E. and Danner, R. P.: Physical and thermodynamic propertiesof pure chemicals data compilation, Taylor & Francis, Washington, DC,1989.Day, D. A., Takahama, S., Gilardoni, S., and Russell, L. M.: Organic com-position of single and submicron particles in different regions of westernNorth America and the eastern Pacific during INTEX-B 2006, Atmos.Chem. Phys., 9, 5433–5446, 2009.Dette, H. P., Qi, M., Schro¨der, D. C., Godt, A., and Koop, T.: Glass-forming properties of 3-methylbutane-1,2,3-tricarboxylic acid and its mix-tures with water and pinonic acid, J. Phys. Chem. A, 118, 7024–7033,2014.Djojoputro, H. and Ismadji, S.: Density and viscosity of several aldehydesfragrance compounds in their binary mixtures with ethanol at (298.15 K,308.15 K, and 318.15 K), J. Chem. Eng. Data, 50, 2003–2007, 2005.Donahue, N., Epstein, S., Pandis, S., and Robinson, A. L.: A two-dimensional volatility basis set: 1. organic-aerosol mixing thermodynam-ics, Atmos. Chem. Phys., 11, 3303–3318, 2011.127BibliographyDonahue, N. M., Robinson, A. L., Stanier, C. O., and Pandis, S. N.: Cou-pled partitioning, dilution, and chemical aging of semivolatile organics,Environ. Sci .Technol., 40, 2635–2643, 2006.Donahue, N. M., Robinson, A. L., and Pandis, S. N.: Atmospheric organicparticulate matter: From smoke to secondary organic aerosol, Atmos.Environ., 43, 94–106, 2009.Donahue, N. M., Kroll, J. H., and Pandis, S. N.: A two-dimensional volatilitybasis set–Part 2: Diagnostics of organic-aerosol evolution, Atmos. Chem.Phys., 12, 615–634, 2012.Donaldson, K., Tran, L., Jimenez, L., Duffin, R., Newby, D. E., Mills, N.,MacNee, W., and Stone, V.: Combustion-derived nanoparticles: A reviewof their toxicology following inhalation exposure, Part. Fibre Toxicol., 2,10, 2005.Du, J., Cheng, T., Zhang, M., Chen, J., He, Q., and Wang, X.: Aerosolsize spectra and particle formation events at urban Shanghai in EasternChina, Aerosol Air Qual. Res., 12, 1362–1372, 2012.Edgeworth, R., Dalton, B. J., and Parnell, T.: The pitch drop experiment,Eur. J. Phys., 5, 198–200, 2001.Edney, E. O., Kleindienst, T. E., and Jaoui, M.: Formation of 2-methyltetrols and 2-methylglyceric acid in secondary organic aerosol from lab-oratory irradiated isoprene/NOX/SO2/air mixtures and their detectionin ambient PM2.5 samples collected in the eastern United States, Atmo-spheric Chemistry & Physics, 2005.Ehn, M., Kleist, E., Junninen, H., Peta¨ja¨, T., Lo¨nn, G., Schobesberger,S., Dal Maso, M., Trimborn, A., Kulmala, M., Worsnop, D. R., Wah-ner, A., Wildt, J., and Mentel, T. F.: Gas phase formation of extremelyoxidized pinene reaction products in chamber and ambient air, Atmos.Chem. Phys., 12, 5113–5127, 2012.128BibliographyEhn, M., Thornton, J. A., Kleist, E., Sipila¨, M., Junninen, H., Pullinen, I.,Springer, M., Rubach, F., Tillmann, R., Lee, B., Lopez-Hilfiker, F., An-dres, S., Acir, I.-H., Rissanen, M., Jokinen, T., Schobesberger, S., Kan-gasluoma, J., Kontkanen, J., Nieminen, T., Kurte´n, T., Nielsen, L. B.,Jørgensen, S., Kjaergaard, H. G., Canagaratna, M., Maso, M. D., Berndt,T., Peta¨ja¨, T., Wahner, A., Kerminen, V.-M., Kulmala, M., Worsnop,D. R., Wildt, J., and Mentel, T. F.: A large source of low-volatility sec-ondary organic aerosol, Nature, 506, 476–479, 2014.Evans, E. B.: Parts VII and VIII, J. Inst. Petrol. Tech., 24, 537–553, 1938.Farmer, D. K., Matsunaga, A., Docherty, K. S., Surratt, J. D., Seinfeld,J. H., Ziemann, P. J., and Jimenez, J. L.: Response of an aerosol massspectrometer to organonitrates and organosulfates and implications foratmospheric chemistry, Proc. Nat. Acad. Sci. U.S.A., 107, 6670–6675,2010.Fontana, A., Messina, R., Spindella, A., and Cimino, G.: Simple and ver-satile synthesis of branched polyols: (+)-2-C-methylerythritol and (+)-2-C-methylthreitol, Tetrahedron Lett., 41, 7559–7562, 2000.Fo¨rst, P., Werner, F., and Delgado, A.: On the pressure dependence of theviscosity of aqueous sugar solutions, Rheologica Acta, 41, 369–374, 2002.Gao, S., Keywood, M., Ng, N. L., Surratt, J., Varutbangkul, V., Bahreini,R., Flagan, R. C., and Seinfeld, J. H.: Low-molecular-weight andoligomeric components in secondary organic aerosol from the ozonoly-sis of cycloalkenes and α-pinene, J. Phys. Chem. A, 108, 10 147–10 164,2004.Gao, Y., Hall, IV, W. A., and Johnston, M. V.: Molecular composition ofmonoterpene secondary organic aerosol at low mass loading, Environ. Sci.Technol., 44, 7897–7902, 2010.Gaschen, A., Lang, D., Kalberer, M., Savi, M., Geiser, T., Gazdhar, A.,Lehr, C.-M., Bur, M., Dommen, J., Baltensperger, U., and Geiser, M.:129BibliographyCellular responses after exposure of lung cell cultures to secondary organicaerosol particles, Environ. Sci .Technol., 44, 1424–1430, 2010.Grayson, J. W., Song, M., Sellier, M., and Bertram, A. K.: Validationof the poke-flow technique combined with simulations of fluid flow fordetermining viscosities in samples with small volumes and high viscosities,Atmos. Meas. Tech., 8, 2463–2472, 2015a.Grayson, J. W., Zhang, Y., Mutzel, A., Renbaum-Wolff, L., Bo¨ge, O., Ka-mal, S., Herrmann, H., Martin, S. T., and Bertram, A. K.: Effect ofvarying experimental conditions on the viscosity of α-pinene derived sec-ondary organic material, Atmos. Chem. Phys. Disc., 15, 32 967–33 002,2015b.Grunberg, L. and Nissan, A. H.: Rheological measurements and adhesion,Nature, 4175, 799–800, 1949.Guenther, A., Geron, C., Pierce, T., and Lamb, B.: Natural emissions ofnon-methane volatile organic compounds, carbon monoxide, and oxidesof nitrogen from North America, Atmos. Environ., 34, 2205–2230, 2000.Hall, IV, W. A. and Johnston, M. V.: Oligomer content of α-pinene sec-ondary organic aerosol, Aerosol Sci. Technol., 45, 37–45, 2011.Hallquist, M., Wenger, J., Baltensperger, U., Rudich, Y., Simpson, D.,Claeys, M., Dommen, J., Donahue, D., George, C., Goldstein, A. H.,Hamilton, J., Herrmann, H., Hoffmann, T., Iinuma, Y., Jang, M., Jenkin,M. E., Jimenez, J. L., Kiendler-Scharr, A., Maenhaut, W., McFiggans,G., Mentel, T. F., Monod, A., Pre´voˆt, A. S. H., Seinfeld, J. H., Surratt,J., Szmigielski, R., and Wildt, J.: The formation, properties and impactof secondary organic aerosol: current and emerging issues, Atmos. Chem.Phys., 9, 5155–5236, 2009.Hamed, A., Korhonen, H., Sihto, S.-L., Joutsensaari, J., Ja¨rvinen, H.,Peta¨ja¨, T., Arnold, F., Nieminen, T., Kulmala, M., Smith, J. N., Lehtinen,K. E. J., and Laaksonen, A.: The role of relative humidity in continentalnew particle formation, J. Geophys. Res., 116, D03 202, 2011.130BibliographyHan, Z., Tang, X., and Zheng, B.: A PDMS viscometer for microliter New-tonian fluid, J. Micromech. Microeng., 17, 1828–1834, 2007.Hartonen, K., Parshintsev, J., Vilja, V.-P., Tiala, H., Knuuti, S., Lai, C. K.,and Riekkola, M.-L.: Gas chromatographic vapor pressure determinationof atmospherically relevant oxidation products of β-caryophyllene and α-pinene, Atmos. Environ., 81, 330–338, 2013.Hashim, E. T., Ghalib, L., and Adell, H.: A generalized formula for esti-mating pure alkenes viscosity, Pet. Sci. Tech., 30, 2341–2347, 2012.Hawkins, L. N., Baril, M. J., Sedehi, N., M, G. M., De Haan, D. O., Schill,G. P., and Tolbert, M. A.: Formation of semisolid, oligomerized aqueousSOA: lab simulations of cloud processing, Environ. Sci .Technol., 48, 2273–2280, 2014.Haynes, W. M.: CRC handbook of chemistry and physics: A ready-referencebook of chemical and physical data, CRC Press, Boca Raton, 96 edn.,2015.Haywood, J. and Boucher, O.: Estimates of the direct and indirect radiativeforcing due to tropospheric aerosols: A review, Rev. Geophys., 38, 513–543, 2000.Heaton, K. J., Dreyfus, M. A., Wang, S., and Johnston, M. V.: Oligomersin the early stage of biogenic secondary organic aerosol formation andgrowth, Environ. Sci .Technol., 41, 6129–6136, 2007.Held, I. M. and Soden, B. J.: Water vapor feedback and global warming,Ann. Rev. Energy Environ., 25, 441–475, 2000.Henry, K. M. and Donahue, N. M.: Effect of the OH radical scavengerhydrogen peroxide on secondary organic aerosol formation from α-pineneozonolysis, Aerosol Sci. Technol., 45, 696–700, 2011.Hiemenz, P. C. and Lodge, T. P.: Polymer chemistry, CRC Press, BocaRaton, USA, 2 edn., 2007.131BibliographyHodzic, A., Jimenez, J. L., Madronich, S., Aiken, A. C., Bessagnet, B.,Curci, G., Fast, J., Lamarque, J.-F., Onasch, T. B., Roux, G., Schauer,J. J., Stone, E. A., and Ulbrich, I. M.: Modeling organic aerosols duringMILAGRO: importance of biogenic secondary organic aerosols, Atmos.Chem. Phys., 9, 6949–6981, 2009.Hoffmann, T., Odum, J. R., Bowman, F., Collins, D. R., Klockow, D.,Flagan, R. C., and Seinfeld, J. H.: Formation of organic aerosols from theoxidation of biogenic hydrocarbons, J. Atmos. Chem., 26, 189–222, 1997.Hosny, N. A., Fitzgerald, C., Tong, C., Kalberer, M., Kuimova, M. K., andPope, F. D.: Fluorescent lifetime imaging of atmospheric aerosols: a directprobe of aerosol viscosity, Farad. Discuss., 165, 343–356, 2013.Houle, F. A., Hinsberg, W. D., and Wilson, K. R.: Oxidation of a modelalkane aerosol by OH radical: the emergent nature of reactive uptake,Phys. Chem. Chem. Phys., 17, 4412–4423, 2015.Huff-Hartz, K. E.: Cloud condensation nuclei activation of monoterpene andsesquiterpene secondary organic aerosol, J. Geophys. Res., 110, D14 208,2005.Huff-Hartz, K. E., Tischuk, J., Chan, M., Chan, C. K., Donahue, N. M.,and Pandis, S.: Cloud condensation nuclei activation of limited solubilityorganic aerosol, Atmos. Environ., 40, 605–617, 2006.Hussein, T., Hameri, K., Aalto, P., Paatero, P., and Kulmala, M.: Modalstructure and spatial-temporal variations of urban and suburban aerosolsin Helsinki-Finland, Atmos. Environ., 39, 1655–1668, 2005.Hyva¨rinen, A., Lihavainen, H., Gaman, A., Vairila, L., Ojala, H., Kulmala,M., and Viisanen, Y.: Surface tensions and densities of oxalic, malonic,succinic, maleic, malic, and cis-pinonic acids, J. Chem. Eng. Data, 51,255–260, 2006.132BibliographyIinuma, Y., Bo¨ge, O., Kahnt, A., and Herrmann, H.: Laboratory cham-ber studies on the formation of organosulfates from reactive uptake ofmonoterpene oxides, Phys. Chem. Chem. Phys., 11, 7985–7997, 2009.Jang, M., Ghio, A. J., and Cao, G.: Exposure of BEAS-2B cells to secondaryorganic aerosol coated on magnetic nanoparticles, Chem. Res. Toxicol., 19,1044–1050, 2006.Jarosiewicz, P. and Czechowski, G.: The viscous properties of diols. V. 1, 2-hexanediol in water and butanol solutions, Z. Naturforsch., 59a, 559–562,2004.Ja¨rvinen, E., Ignatius, K., Nichman, L., Kristensen, T. B., Fuchs, C.,Ho¨ppel, N., Corbin, J. C., Craven, J., Duplissy, J., Ehrhart, S., El Had-dad, I., Frege, C., Gates, S. J., Gordon, H., Hoyle, C. R., Jokinen, T.,Kallinger, P., Kirkby, J., Kiselev, A., Naumann, K.-H., Peta¨ja¨, T., Pin-terich, T., Prevot, A. S. H., Saathoff, H., Schiebel, T., Sengupta, K., Si-mon, M., Tro¨stl, J., Virtanen, A., Vochezer, P., Vogt, S., Wagner, A. C.,Wagner, R., Williamson, C., Winkler, P. M., Yan, C., Baltensperger, U.,Donahue, N. M., Flagan, R. C., Gallagher, M., Hansel, A., Kulmala, M.,Stratmann, F., Worsnop, D. R., Mo¨hler, O., Leisner, T., and Schnaiter,M.: Observation of viscosity transition in α-pinene secondary organicaerosol, Atmos. Chem. Phys. Disc., 15, 28 575–28 617, 2015.Jeong, J. T. and Moffatt, H. K.: Free-surface cusps associated with flow atlow Reynolds number, J. Fluid Mech., 241, 1–22, 1992.Jimenez, J. L., Canagaratna, M. R., Donahue, N. M., Prevot, A. S. H.,Zhang, Q., Kroll, J. H., DeCarlo, P. F., Allan, J. D., Coe, H., Ng, N. L.,Aiken, A. C., Docherty, K. S., Ulbrich, I. M., Grieshop, A. P., Robinson,A. L., Duplissy, J., Smith, J. D., Wilson, K. R., Lanz, V. A., Hueglin,C., Sun, Y. L., Tian, J., Laaksonen, A., Raatikainen, T., Rautiainen, J.,Vaattovaara, P., Ehn, M., Kulmala, M., Tomlinson, J. M., Collins, D. R.,Cubison, M. J., Dunlea, E. J., Huffman, J. A., on, c. T. B., Alfarra,M. R., Williams, P. I., Bower, K., Kondo, Y., Schneider, J., Drewnick,133BibliographyF., Borrmann, S., Weimer, S., Demerjian, K., Salcedo, D., Cottrell, L.,Griffin, R., Takami, A., Miyoshi, T., Hatakeyama, S., Shimono, A., Sun,J. Y., Zhang, Y. M., Dzepina, K., Kimmel, J. R., Sueper, D., Jayne,J. T., Herndon, S. C., Trimborn, A. M., Williams, L. R., Wood, E. C.,Middlebrook, A. M., Kolb, C. E., Baltensperger, U., and Worsnop, D. R.:Evolution of organic aerosols in the atmosphere, Science, 326, 1525–1529,2009.Jin, S., Huang, P., Park, J., Yoo, J. Y., and Breuer, K. S.: Near-surfacevelocimetry using evanescent wave illumination, Exp. Fluids, 37, 825–833,2004.Joly, L., Ybert, C., and Bocquet, L.: Probing the nanohydrodynamics atliquid-solid interfaces using thermal motion, Phys. Rev. Lett., 96, 046 101,2006.Jonsson, A˚. M., Hallquist, M., and Ljungstro¨m, E.: Influence of OH scav-enger on the water effect on secondary organic aerosol formation fromozonolysis of limonene, ∆ 3-carene, and α-pinene, Environ. Sci .Technol.,42, 5938–5944, 2008.Joseph, P. and Tabeling, P.: Direct measurement of the apparent slip length,Phys. Rev. E, 71, 035 303(R), 2005.Kamens, R. M., Jang, M., Chien, C.-J., and Leach, K.: Aerosol formationfrom the reaction of α-pinene and ozone using a gas-phase kinetics-aerosolpartitioning model , Environ. Sci .Technol., 33, 1430–1438, 1999.Kanakidou, M., Seinfeld, J. H., Pandis, S. N., Barnes, I., Dentener, F. J.,Facchini, M. C., Van Dingenen, R., Ervens, B., Nenes, A., Nielsen, C. J.,Swietlicki, E., Putaud, J. P., Balkanski, Y., Fuzzi, S., Horth, J., Moortgat,G. K., Winterhalter, R., Myhre, C. E. L., Tsigaridis, K., Vignati, E.,Stephanou, E. G., and Wilson, J.: Organic aerosol and global climatemodelling: A review, Atmos. Chem. Phys., 5, 1053–1123, 2005.Katritzky, A. R., Chen, K., Wang, Y., Karelson, M., Lucic, B., Trinajstic,N., Suzuki, T., and Schuurmann, G.: Prediction of liquid viscosity for134Bibliographyorganic compounds by a quantitative structure-property relationship, J.Phys. Org. Chem., 13, 80–86, 2000.Katritzky, A. R., Kuanar, M., Slavov, S., Hall, C. D., Karelson, M., Kahn,I., and Dobchev, D. A.: Quantitative correlation of physical and chemicalproperties with chemical structure: utility for prediction, Chem. Rev.,110, 5714–5789, 2010.Kestin, J., Sokolov, M., and Wakeham, W. A.: Viscosity of liquid water inthe range 8 ◦C to 150 ◦C, J. Phys. Chem. Ref. Data, 7, 941–948, 1978.Kidd, C., Perraud, V., Wingen, L. M., and Finlayson-Pitts, B. J.: Integrat-ing phase and composition of secondary organic aerosol from the ozonol-ysis of α-pinene, Proc. Nat. Acad. Sci. U.S.A., 111, 7552–7557, 2014.King, S. M., Rosenoern, T., Shilling, J. E., Chen, Q., and Martin, S. T.:Increased cloud activation potential of secondary organic aerosol for at-mospheric mass loadings, Atmos. Chem. Phys., 9, 2959–2972, 2009.Kirchstetter, T. W., Corrigan, C. E., and Novakov, T.: Laboratory and fieldinvestigation of the adsorption of gaseous organic compounds onto quartzfilters, Atmos. Environ., 35, 1663–1671, 2001.Knopf, D. A. and Rigg, Y. J.: Homogeneous ice nucleation from aqueousinorganic/organic particles representative of biomass burning: water ac-tivity, freezing temperatures, nucleation rates, J. Phys. Chem. A, 115,762–773, 2011.Koop, T., Kapilashrami, A., Molina, L. T., and Molina, M. J.: Phase tran-sitions of sea-salt/water mixtures at low temperatures: Implications forozone chemistry in the polar marine boundary layer, J. Geophys. Res.,105, 26 393, 2000.Koop, T., Bookhold, J., Shiraiwa, M., and Po¨schl, U.: Glass transition andphase state of organic compounds: dependency on molecular propertiesand implications for secondary organic aerosols in the atmosphere, Phys.Chem. Chem. Phys., 13, 19 238–19 255, 2011.135BibliographyKostenidou, E., Lee, B.-H., Engelhart, G. J., Pierce, J. R., and Pandis, S. N.:Mass spectra deconvolution of low, medium, and high volatility biogenicsecondary organic aerosol, Environ. Sci .Technol., 43, 4884–4889, 2009.Kuwata, M. and Martin, S. T.: Phase of atmospheric secondary organicmaterial affects its reactivity, Proc. Nat. Acad. Sci. U.S.A., 109, 17 354–17 359, 2012.Ladino, L. A., Zhou, S., Yakobi Hancock, J. D., Aljawhary, D., and Abbatt,J. P. D.: Factors controlling the ice nucleating abilities of αpinene SOAparticles, J. Geophys. Res. Atmos., 119, 9041–9051, 2014.Lambe, A. T., on, c. T. B., Massoli, P., Croasdale, D. R., Wright, J. P.,Ahern, A. T., Williams, L. R., Worsnop, D. R., Brune, W. H., and Davi-dovits, P.: Laboratory studies of the chemical composition and cloudcondensation nuclei (CCN) activity of secondary organic aerosol (SOA)and oxidized primary organic aerosol (OPOA), Atmos. Chem. Phys., 11,8913–8928, 2011.Lambe, A. T., Chhabra, P. S., on, c. T. B., Brune, W. H., Hunter, J. F.,Kroll, J. H., Cummings, M. J., Brogan, J. F., Parmar, Y., Worsnop,D. R., Kolb, C. E., and Davidovits, P.: Effect of oxidant concentration,exposure time, and seed particles on secondary organic aerosol chemicalcomposition and yield, Atmos. Chem. Phys., 15, 3063–3075, 2015.Lee, J. Y. and Hildemann, L. M.: Surface tension of solutions containingdicarboxylic acids with ammonium sulfate, D-glucose, or humic acid, J.Aerosol Sci., 64, 94–102, 2013.Lewandowski, F. Y. and Dupuis, D.: Dynamic measurements of surfacetension of solutions of polyisobutylene in mixtures of polybutene oil andDecalin, J. Non-Newton. Fluid, 52, 233–248, 1994.Li, C., Zhang, J., Zhang, T., Wei, X., Zhang, E., Yang, N., Zhao, N.,Su, M., and Zhou, H.: Density, viscosity, and excess properties for 1,2-diaminoethane + 1,2-ethanediol at (298.15, 303.15, and 308.15) K, J.Chem. Eng. Data, 55, 4104–4107, 2010.136BibliographyLi, D. Y., Jing, D. L., Pan, Y. L., and Ahmad, K.: Slip length measurementof water flow on graphite surface using atomic force microscope, Adv.Mater. Res., 941-944, 1581–1584, 2014.Lignell, H., Hinks, M. L., and Nizkorodov, S. A.: Exploring matrix effectson photochemistry of organic aerosols, Proc. Nat. Acad. Sci. U.S.A., 111,13 780–13 785, 2014.Lin, Y.-Y., Lin, C.-W., Yang, L.-J., and Wang, A.-B.: Micro-viscometerbased on electrowetting on dielectric, Electrochimica Acta, 52, 2876–2883,2007.Longinotti, M. P. and Corti, H. R.: Viscosity of concentrated sucrose andtrehalose aqueous solutions including the supercooled regime, J. Phys.Chem. Ref. Data, 37, 1503, 2008.Loza, C. L., Coggon, M. M., Nguyen, T. B., Zuend, A., Flagan, R. C.,and Seinfeld, J. H.: On the mixing and evaporation of secondary organicaerosol components, Environ. Sci .Technol., 47, 6173–6180, 2013.Lu, J. W., Rickards, A. M. J., Walker, J. S., Knox, K. J., Miles, R. E. H.,Reid, J. P., and Signorell, R.: Timescales of water transport in viscousaerosol: measurements on sub-micron particles and dependence on condi-tioning history, Phys. Chem. Chem. Phys., 16, 9819–9830, 2014.MacDonald, G. A., Lanier, T. C., Swaisgood, H. E., and Hamann, D. D.:Mechanism for stabilization of fish actomyosin by sodium lactate, J. Agric.Food Chem., 44, 106–112, 1996.Marczak, W., Adamczyk, N., and Lezniak, M.: Viscosity of associated mix-tures approximated by the Grunberg-Nissan model, Int. J. Thermophys.,33, 680–691, 2012.Marrero-Morejo´n, J. and Pardillo-Fontdevila, E.: Estimation of liquid vis-cosity at ambient temperature of pure organic compounds by using group-interaction contributions, Chem. Eng. J., 79, 69–72, 2000.137BibliographyMartin, S. T.: Phase transitions of aqueous atmospheric particles, Chem.Rev., 100, 3403–3454, 2000.Mewis, J. and Metzner, A. B.: The rheological properties of suspensions offibres in Newtonian fluids subjected to extensional deformations, J. FluidMech., 62, 593, 2006.Moller, B., Rarey, J., and Ramjugernath, D.: Estimation of the vapourpressure of non-electrolyte organic compounds via group contributionsand group interactions, J. Mol. Liq., 143, 52–63, 2008.Murray, B. J. and Bertram, A. K.: Inhibition of solute crystallisation inaqueous H+–NH4+–SO42–H2O droplets, Phys. Chem. Chem. Phys., 10,3287, 2008.Murray, B. J., Wilson, T. W., Dobbie, S., Cui, Z., Al-Jumur, S. M. R. K.,Mo¨hler, O., Schnaiter, M., Wagner, R., Benz, S., Niemand, M., Saathoff,H., Ebert, V., Wagner, S., and Ka¨rcher, B.: Heterogeneous nucleation ofice particles on glassy aerosols under cirrus conditions, Nat. Geosci., 3,233–237, 2010.Murray, B. J., Haddrell, A., Peppe, S., Davies, J. F., Reid, J. P., O’Sullivan,D., Price, H. C., Kumar, R., Saunders, S. M., Plane, J., Umo, N. S., andWilson, T.: Glass formation and unusual hygroscopic growth of iodic acidsolution droplets with relevance for iodic mediated particle formation inthe marine boundary layer, Atmos. Chem. Phys., 12, 8575–8587, 2012.Myrdal, P. B. and Yalkowsky, S. H.: Estimating pure component vaporpressures of complex organic molecules, Ind. Eng. Chem. Res., 36, 2494–2499, 1997.Nannoolal, Y., Rarey, J., Ramjugernath, D., and Cordes, W.: Estimationof pure component properties. Part 1. Estimation of the normal boilingpoint of non-electrolyte organic compounds via group contributions andgroup interactions, Fluid Phase Equilibria, 226, 45–63, 2004.138BibliographyNannoolal, Y., Rarey, J., and Ramjugernath, D.: Estimation of pure compo-nent properties. Part 3. Estimation of the vapor pressure of non-electrolyteorganic compounds via group contributions and group interactions, FluidPhase Equilibria, 269, 117–133, 2008.Napier, W. J., Ensberg, J. J., and Seinfeld, J. H.: Insight into the numeri-cal challenges of implementing 2-dimensional SOA models in atmosphericchemical transport models , Atmos. Environ., 96, 331–344, 2014.Nga, N. L., Kroll, J. H., Keywood, M. D., Bahreini, R., Varutbangkul, V.,Flagan, R. C., Seinfeld, J. H., Lee, A., and Goldstein, A. H.: Contributionof first- versus second-generation products to secondary organic aerosolsformed in the oxidation of biogenic hydrocarbons, Environ. Sci .Technol.,40, 2283–2297, 2006.Nolte, C. G., Schauer, J. J., Cass, G. R., and Simoneit, B. R. T.: Highlypolar organic compounds present in wood smoke and in the ambient at-mosphere, Environ. Sci .Technol., 35, 1912–1919, 2001.Noureddini, H., Teoh, B. C., and Davis Clements, L.: Viscosities of vegetableoils and fatty acids, J. Am. Oil Chem. Soc., 69, 1189–1191, 1992.O’Brien, R. E., Neu, A., Epstein, S. A., MacMillan, A. C., Wang, B., Kelly,S. T., Nizkorodov, S. A., Laskin, A., Moffet, R. C., and Gilles, M. K.:Physical properties of ambient and laboratory-generated secondary or-ganic aerosol, Geophys. Res. Lett., 41, 4347–4353, 2014.Odum, J. R., Yu, J., and Kamens, R. M.: Modeling the mass transfer ofsemivolatile organics in combustion aerosols, Environ. Sci .Technol., 28,2278–2285, 1994.Ottar, B.: Arctic air pollution: A Norwegian perspective, Atmos. Environ.,23, 2349–2356, 1989.Pachaiyappan, V., Ibrahim, S. H., and Kuloor, N. R.: Simple correlation fordetrmining the viscosity of organic liquids, Chem. Eng. J., 74, 193–196,1967.139BibliographyPajunoja, A., Malila, J., Hao, L., Joutsensaari, J., Lehtinen, K. E. J., andVirtanen, A.: Estimating the viscosity range of SOA particles based ontheir coalescence time, Aerosol Sci. Technol., 48, i–iv, 2014.Pankow, J. F.: An absorption model of the gas/aerosol partitioning involvedin the formation of secondary organic aerosol, Atmos. Environ., 28, 189–193, 1994.Pare´, P. W. and Tumlinson, J. H.: Plant volatiles as a defense against insectherbivores, Plant Physiol., 121, 325–331, 1999.Perera, F. P., Li, Z., Whyatt, R., Hoepner, L., Wang, S., Camann, D., andRauh, V.: Prenatal airborne polycyclic aromatic hydrocarbon exposureand child IQ at age 5 years, Pediatrics, 124, 195–202, 2009.Perera, F. P., Tang, D., Wang, S., Vishnevetsky, J., Zhang, B., Diaz, D.,Camann, D., and Rauh, V.: Prenatal polycyclic aromatic hydrocarbon(PAH) exposure and child behavior at age 6–7 years, Environ. HealthPerspect., 120, 921–926, 2012.Perraud, V., Bruns, E. A., Ezell, M. J., Johnson, S. N., Yu, Y., Alexan-der, M. L., Zelenyuk, A., Imre, D., Chang, W. L., Dabdub, D., Pankow,J. F., and Finlayson-Pitts, B. J.: Nonequilibrium atmospheric secondaryorganic aerosol formation and growth, Proc. Nat. Acad. Sci. U.S.A., 109,2836–2841, 2012.Perry, R. H. and Green, D. W.: Perry’s Chemical Engineers’ Handbook,McGraw Hill, 2008.Pfrang, C., Shiraiwa, M., and Po¨schl, U.: Chemical ageing and transforma-tion of diffusivity in semi-solid multi-component organic aerosol particles,Atmos. Chem. Phys., 11, 7343–7354, 2011.Pipe, C. J. and McKinley, G. H.: Microfluidic rheometry, Mech. Res.Comm., 36, 110–120, 2009.140BibliographyPower, R. M. and Reid, J. P.: Probing the micro-rheological properties ofaerosol particles using optical tweezers, Rep. Prog. Phys., 77, 074 601,2014.Power, R. M., Simpson, S. H., Reid, J. P., and Hudson, A. J.: The transitionfrom liquid to solid-like behaviour in ultrahigh viscosity aerosol particles,Chem. Sci., 4, 2597–2604, 2013.Praplan, A. P., Schobesberger, S., Bianchi, F., Rissanen, M. P., Ehn, M.,Jokinen, T., Junninen, H., Adamov, A., Amorim, A., Dommen, J., Du-plissy, J., Hakala, J., Hansel, A., Heinritzi, M., Kangasluoma, J., Kirkby,J., Krapf, M., Ku¨rten, A., Lehtipalo, K., Riccobono, F., Rondo, L., Sar-nela, N., Simon, M., Tome´, A., Tro¨stl, J., Winkler, P. M., Williamson,C., Ye, P., Curtius, J., Baltensperger, U., Donahue, N. M., Kulmala, M.,and Worsnop, D. R.: Elemental composition and clustering behaviourof α-pinene oxidation products for different oxidation conditions, Atmos.Chem. Phys., 15, 4145–4159, 2015.Price, H. C., Murray, B. J., and Mattsson, J.: Quantifying water diffusion inhigh-viscosity and glassy aqueous solutions using a Raman isotope tracermethod, Atmos. Chem. Phys., 14, 3817–3830, 2014.Price, H. C., Mattsson, J., Zhang, Y., Bertram, A. K., Davies, J. F.,Grayson, J. W., Martin, S. T., O’Sullivan, D., Reid, J. P., Rickards,A. M. J., and Murray, B. J.: Water diffusion in atmospherically relevantα-pinene secondary organic material, Chem. Sci., 6, 4876–4883, 2015.Pruppacher, H. R. and Klett, J. D.: Microphysics and clouds and precipita-tion, Springer, Netherlands, 2 edn., 2010.Putman, A. L., Offenberg, J. H., Fisseha, R., and Kundu, S.: Ultrahigh-resolution FT-ICR mass spectrometry characterization of α-pineneozonolysis SOA, Atmos. Environ., 46, 164–172, 2012.Quintas, M., Branda˜o, T. R. S., Silva, C. L. M., and Cunha, R. L.: Rheologyof supersaturated sucrose solutions, J. Food Eng., 77, 844–852, 2006.141BibliographyRastogi, N., Singh, A., Sarin, M. M., and Singh, D.: Temporal variability ofprimary and secondary aerosols over northern India: Impact of biomassburning emissions, Atmos. Environ., 125, 396–403, 2016.Reist, P.: Aerosol Science and Technology, McGraw-Hill Professional, NewYork, NY, USA, 2 edn., 1992.Renbaum-Wolff, L., Grayson, J. W., Bateman, A. P., Kuwata, M., Sell-ier, M., Murray, B. J., Shilling, J. E., Martin, S. T., and Bertram, A. K.:Viscosity of α-pinene secondary organic material and implications for par-ticle growth and reactivity, Proc. Nat. Acad. Sci. U.S.A., 110, 8014–8019,2013a.Renbaum-Wolff, L., Grayson, J. W., and Bertram, A. K.: Technical Note:New methodology for measuring viscosities in small volumes characteristicof environmental chamber particle samples, Atmos. Chem. Phys., 13, 791–802, 2013b.Riipinen, I., Pierce, J. R., Yli-Juuti, T., Nieminen, T., Ha¨kkinen, S., Ehn,M., Junninen, H., Lehtipalo, K., Peta¨ja¨, T., Slowik, J., Chang, R., Shantz,N. C., Abbatt, J., Leaitch, W. R., Kerminen, V.-M., Worsnop, D. R.,Pandis, S. N., Donahue, N. M., and Kulmala, M.: Organic condensation:a vital link connecting aerosol formation to cloud condensation nuclei(CCN) concentrations, Atmos. Chem. Phys., 11, 3865–3878, 2011.Robinson, C. B., Schill, G. P., and Tolbert, M. A.: Optical growth of highlyviscous organic/sulfate particles, J. Atmos. Chem., 71, 145–156, 2014.Robinson, E. S., Saleh, R., and Donahue, N. M.: Organic aerosol mixingobserved by single-particle mass spectrometry, J. Phys. Chem. A, 117,13 935–13 945, 2013.Roe, R.-J.: Surface tension of polymer liquids, J. Phys. Chem., 72, 2013–2017, 1968.Rounds, S. A. and Pankow, J. F.: Application of a radial diffusion model to142Bibliographydescribe gas/particle sorption kinetics, Environ. Sci .Technol., 24, 1378–1386, 1990.Roy, M. N., Das, R. K., and Bhattacharjee, A.: Density and viscosity ofacrylonitrile + cinnamaldehyde, + anisaldehyde, and + benzaldehyde at(298.15, 308.15, and 318.15) K, J. Chem. Eng. Data, 53, 1431–1435, 2008.Russell, L. M., Bahadur, R., and Ziemann, P. J.: Identifying organic aerosolsources by comparing functinoal group composition in chamber and at-mospheric particles, Proc. Nat. Acad. Sci. U.S.A., 108, 3516–3521, 2011.Saathoff, H., Naumann, K.-H., Schnaiter, M., and Scho¨ck, W.: Coatingof soot and (NH4)2SO4 particles by ozonolysis products of α-pinene, JAerosol Sci., 34, 1297–1321, 2003.Saggin, R. and Coupland, J. N.: Rheology of xanthan/sucrose mixtures atultrasonic frequencies, J. Food Eng., 65, 49–53, 2004.Saleh, R., Donahue, N. M., and Robinson, A. L.: Time scales for gas-particlepartitioning equilibration of secondary organic aerosol formed from alpha-pinene ozonolysis, Environ. Sci .Technol., 47, 5588–5594, 2013.Sastri, S. R. S. and Rao, K. K.: A new group contribution method forpredicting viscosity of organic liquids, Chem. Eng. J., 50, 9–25, 1992.Saukko, E., Lambe, A. T., Massoli, P., and Koop, T.: Humidity-dependentphase state of SOA particles from biogenic and anthropogenic precursors,Atmos. Chem. Phys., 12, 7517–7529, 2012.Schill, G. P., De Haan, D. O., and Tolbert, M. A.: Heterogeneous ice nucle-ation on simulated secondary organic aerosol, Environ. Sci .Technol., 48,1675–1682, 2014.Schnell, E.: Slippage of water over nonwettable surfaces, J. Appl. Phys., 10,1149–1152, 1956.Schobesberger, S., Junninen, H., Bianchi, F., Lonn, G., Ehn, M., Lehtipalo,K., Dommen, J., Ehrhart, S., Ortega, I. K., Franchin, A., Nieminen, T.,143BibliographyRiccobono, F., Hutterli, M., Duplissy, J., Almeida, J., Amorim, A., Bre-itenlechner, M., Downard, A. J., Dunne, E. M., Flagan, R. C., Kajos, M.,Keskinen, H., Kirkby, J., Kupc, A., Kuerten, A., Kurte´n, T., Laaksonen,A., Mathot, S., Onnela, A., Praplan, A. P., Rondo, L., Santos, F. D.,Schallhart, S., Schnitzhofer, R., Sipila¨, M., Tome, A., Tsagkogeorgas, G.,Vehkamaki, H., Wimmer, D., Baltensperger, U., Carslaw, K. S., Curtius,J., Hansel, A., Peta¨ja¨, T., Kulmala, M., Donahue, N. M., and Worsnop,D. R.: Molecular understanding of atmospheric particle formation fromsulfuric acid and large oxidized organic molecules, Proc. Nat. Acad. Sci.U.S.A., 110, 17 223–17 228, 2013.Seinfeld, J. and Pandis, S.: Atmospheric chemistry and physics: from airpollution to climate change, Wiley Interscience, New York, NY, USA, 2edn., 2006.Seinfeld, J. H. and Pankow, J. F.: Organic atmospheric particulate material,Annu. Rev. Phys. Chem., 54, 121–140, 2003.Shilling, J. E., Chen, Q., King, S. M., Rosenoern, T., Kroll, J. H., Worsnop,D. R., McKinney, K., and Martin, S. T.: Particle mass yield in secondaryorganic aerosol formed by the dark ozonolysis of α-pinene, Atmos. Chem.Phys., 8, 2073–2208, 2008.Shilling, J. E., Chen, Q., King, S., Rosenoern, T., Kroll, J. H., Worsnop,D. R., DeCarlo, P. F., Aiken, A. C., Sueper, D., Jimenez, J. L., andMartin, S. T.: Loading-dependent elemental composition of alpha-pineneSOA particles, Atmos. Chem. Phys., 9, 771–782, 2009.Shiraiwa, M. and Seinfeld, J. H.: Equilibration timescale of atmosphericsecondary organic aerosol partitioning, Geophys. Res. Lett., 39, L24 801,2012.Shiraiwa, M., Ammann, M., Koop, T., and Po¨schl, U.: Gas uptake andchemical aging of semisolid organic aerosol particles, Proc. Nat. Acad.Sci. U.S.A., 108, 11 003–11 008, 2011a.144BibliographyShiraiwa, M., Sosedova, Y., Rouvie`re, A., Yang, H., Zhang, Y., Abbatt,J. P. D., Ammann, M., and Po¨schl, U.: The role of long-lived reactiveoxygenintermediates in the reaction of ozone withaerosol particles, Nat.Chem., 3, 291–295, 2011b.Shiraiwa, M., Zuend, A., Bertram, A. K., and Seinfeld, J. H.: Gas-particlepartitioning of atmospheric aerosols: interplay of physical state, non-idealmixing and morphology, Phys. Chem. Chem. Phys., 15, 11 441–11 453,2013.Shrestha, M., Zhang, Y., Ebben, C. J., Martin, S. T., and Geiger, F. M.:Vibrational sum frequency generation spectroscopy of secondary organicmaterial produced by condensational growth from α-pinene ozonolysis, J.Phys. Chem. A, 117, 8427–8436, 2013.Silber-Li, Z. H., Tan, Y. P., and Weng, P. F.: A microtube viscometer witha thermostat, Exp. Fluids, 36, 586–592, 2004.Slowik, J. G., Stroud, C., and Bottenheim, J. W.: Characterization of a largebiogenic secondary organic aerosol event from eastern Canadian forests,Atmos. Chem. Phys., 10, 2825–2845, 2010.Smith, G. D., Woods, E., DeForest, C. L., Baer, T., and Miller, R. E.:Reactive uptake of ozone by oleic acid aerosol particles: Application ofsingle-particle mass spectrometry to heterogeneous reaction kinetics, J.Phys. Chem. A, 106, 8085–8095, 2002.Solomon, S., Qin, D., Manning, M., Alley, R. B., Bernsten, T., Bindoff,N. L., Chen, Z., Chidthaisong, A., Gregory, J. M., GC, H., Heimann, M.,Hewiston, B., Hoskins, B. J., Joos, F., Jouzel, J., Kattsov, V., Lohmann,U., Matsuno, T., Molina, M., Nicholls, N., Overpeck, J., Raga, G., Ra-maswamy, V., Ren, J., Rusticucci, M., Somerville, R., Stocker, T. F.,Whetton, P., Wood, R. A., and Wratt, D.: Climate Change 2007: ThePhysical Science Basis. Contribution of Working Group I to the FourthAssessment Report of the Intergovernmental Panel on Climate Change,145BibliographyCambridge University Press, Cambridge, United Kingdom and New York,NY, USA, 2007.Song, M., Marcolli, C., Krieger, U. K., Zuend, A., and Peter, T.: Liquid-liquid phase separation and morphology of internally mixed dicarboxylicacids/ammonium sulfate/water particles, Atmos. Chem. Phys., 12, 2691–2712, 2012.Song, M., Liu, P. F., Hanna, S. J., Li, Y. J., Martin, S. T., andBertram, A. K.: Relative humidity-dependent viscosities of isoprene-derived secondary organic material and atmospheric implications forisoprene-dominant forests, Atmos. Chem. Phys., 15, 5145–5159, 2015.Srivastava, N. and Burns, M. A.: Analysis of non-Newtonian liquids usinga microfluidic capillary viscometer, Anal. Chem., 78, 1690–1696, 2006.Srivastava, N., Davenport, R. D., and Burns, M. A.: Nanoliter viscometerfor analyzing blood plasma and other liquid samples, Anal. Chem., 77,383–392, 2005.Stein, S. E. and Brown, R. L.: Estimation of normal boiling points fromgroup contributions, J. Chem. Inf. Comput. Sci., 34, 581–587, 1994.Stillinger, F. H. and Debenedetti, P. G.: Alternative view of self-diffusionand shear viscosity, J. Phys. Chem. B, 109, 6604–6609, 2005.Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung,J., Nauels, A., Xia, Y., Bex, V., and Midgley, P. M., eds.: IPCC 2013: Cli-mate Change 2013: The Physical Science Basis. Contribution of WorkingGroup I to the Fifth Assessment Report of the Intergovernmental Panel onClimate Change, Cambridge University Press, Cambridge, United King-dom And New York, NY, USA, 2013.Surratt, J. D., Murphy, S. M., Kroll, J. H., Ng, N. L., Hildebrandt, L.,Sorooshian, A., Szmigielski, R., Vermeylen, R., Maenhaut, W., Claeys,146BibliographyM., Flagan, R. C., and Seinfeld, J. H.: Chemical Composition of Sec-ondary Organic Aerosol Formed from the Photooxidation of Isoprene, J.Phys. Chem. A, 110, 9665–9690, 2006.Surratt, J. D., Chan, A. W. H., Eddingsaas, N. C., Chan, M., Loza, C. L.,Kwan, A. J., Hersey, S. P., Flagan, R. C., Wennberg, P. O., and Sein-feld, J. H.: Reactive intermediates revealed in secondary organic aerosolformation from isoprene, Proc. Nat. Acad. Sci. U.S.A., 107, 6640–6645,2010.Suzuki, T., Ohtaguchi, K., and Koide, K.: Computer-assisted approach todevelop a new prediction method of liquid viscosity of organic compounds,Computers Chem. Eng., 20, 161–173, 1996.Swindells, J. F., Snyder, C. F., Hardy, R. C., and Golden, P. E.: Viscositiesof sucrose solutions at various temperatures: Tables of recalculated values,United States Department of Commerce, 1958.Telis, V. R. N., Telis-Romero, J., Mazzotti, H. B., and Gabas, A. L.: Vis-cosity of aqueous carbohydrate solutions at different temperatures andconcentrations, Int. J. Food. Prop., 10, 185–195, 2007.Tillmann, R., Hallquist, M., Jonsson, A˚. M., Kiendler-Scharr, A., Saathoff,H., Iinuma, Y., and Mentel, T. F.: Influence of relative humidity andtemperature on the production of pinonaldehyde and OH radicals fromthe ozonolysis of α-pinene, Atmos. Chem. Phys., 10, 7057–7072, 2010.Tong, H. J., Reid, J. P., Bones, D. L., Luo, B. P., and Krieger, U. K.:Measurements of the timescales for the mass transfer of water in glassyaerosol at low relative humidity and ambient temperature, Atmos. Chem.Phys., 11, 4739–4754, 2011.Tretheway, D. C. and Meinhart, C. D.: Apparent fluid slip at hydrophobicmicrochannel walls, Phys. Fluids, 14, L9–L12, 2002.Tuckermann, R. and Cammenga, H. K.: The surface tension of aqueous147Bibliographysolutions of some atmospheric water-soluble organic compounds, Atmos.Environ., 38, 6135–6138, 2004.Twomey, S.: Aerosols, clouds and radiation, Atmos. Environ., 25A, 2435–2442, 1991.Urban, R. C., Alves, C. A., Allen, A. G., Cardoso, A. A., and Campos, M.L. A. M.: Organic aerosols in a Brazilian agro-industrial area: Speciationand impact of biomass burning , Atmos. Res., 169, 271–279, 2016.US Environmental Protection Agency: US Clean Air Act (and subsequentamendments/rulings), Tech. rep., 2008.Vaden, T. D., Song, C., Zaveri, R. A., Imre, D., and Zelenyuk, A.: Morphol-ogy of mixed primary and secondary organic particles and the adsorptionof spectator organic gases during aerosol formation, Proc. Nat. Acad. Sci.U.S.A., 107, 6658–6663, 2010.Vaden, T. D., Imre, D., Bera´nek, J., Shrivastava, M., and Zelenyuk, A.:Evaporation kinetics and phase of laboratory and ambient secondary or-ganic aerosol, Proc. Nat. Acad. Sci. U.S.A., 108, 2190–2195, 2011.Varutbangkul, V., Brechtel, F. J., Bahreini, R., Ng, N. L., Keywood, M. D.,Kroll, J. H., Flagan, R. C., Seinfeld, J. H., Lee, A., and Goldstein, A. H.:Hygroscopicity of secondary organic aerosols formed by oxidation of cy-cloalkenes, monoterpenes, sesquiterpenes, and related compounds, Atmos.Chem. Phys., 6, 2367–2388, 2006.Vereecken, L. and Peeters, J.: A theoretical study of the OH-initiated gas-phase oxidation mechanism of β-pinene (C10H16): first generation prod-ucts, Phys. Chem. Chem. Phys., 14, 3802–3815, 2012.Verevkin, S. P.: Determination of vapor pressures and enthalpies of vapor-ization of 1,2-alkanediols, Fluid Phase Equilibria, 224, 23–29, 2004.Verevkin, S. P., Krasnykh, E. L., Vasiltsova, T. V., and Heintz, A.: De-termination of ambient temperature vapor pressures and vaporization en-thalpies of branched ethers, J. Chem. Eng. Data, 48, 591–599, 2003.148BibliographyVinson, C. G. and Martin, J. J.: Heat of vaporization and vapor pressure of1,4-dioxane, J. Chem. Eng. Data, 8, 74–75, 1963.Virkkula, A., Van Dingenen, R., Raes, F., and Hjorth, J.: Hygroscopicproperties of aerosol fromed by oxidation of limonene, α-pinene, and β-pinene, J. Geophys. Res., 104, 3569–3579, 1999.Virtanen, A., Joutsensaari, J., Koop, T., Kannosto, J., Yli-Pirila¨, P., Le-skinen, J., Makela¨, J. M., Holopainen, J. K., Po¨schl, U., Kulmala, M.,Worsnop, D. R., and Laaksonen, A.: An amorphous solid state of bio-genic secondary organic aerosol particles, Nature, 467, 824–827, 2010.Virtanen, A., Kannosto, J., Kuuluvainen, H., Arffman, A., Joutsensaari, J.,Saukko, E., Hao, L., Yli-Pirila¨, P., Tiitta, P., Holopainen, J. K., Keskinen,J., Worsnop, D. R., Smith, J. N., and Laaksonen, A.: Bounce behaviorof freshly nucleated biogenic secondary organic aerosol particles, Atmos.Chem. Phys., 11, 8759–8766, 2011.Viswanath, D. S., Ghosh, T. K., Prasad, D. H. L., Dutt, N. V. K., andRani, K. Y.: Viscosity of liquids: theory, estimation, experiment, anddata, Springer, Dordrecht, The Netherlands, 2007.Wang, B., Laskin, A., Roedel, T., Gilles, M. K., Moffet, R. C., Tivanski,A. V., and Knopf, D. A.: Heterogeneous ice nucleation and water uptakeby field-collected atmospheric particles below 273 K, J. Geophys. Res.,117, D00V19, 2012.Wang, B., O’Brien, R. E., Kelly, S. T., Shilling, J. E., Moffet, R. C., Gilles,M. K., and Laskin, A.: Reactivity of liquid and semisolid secondary or-ganic carbon with chloride and nitrate in atmospheric aerosols, J. Phys.Chem. A, 119, 4498–4508, 2015.Watanabe, K. and Udagawa, Y.: Drag reduction of Newtonian fluid in acircular pipe with a highly water-repellent wall, J. Fluid Mech., 381, 225–238, 1999.149BibliographyWehner, B., Wiedensohler, A., and Tuch, T. M.: Variability of the aerosolnumber size distribution in Beijing, China: New particle formation,dust storms, and high continental background, Geophys. Res. Lett., 31,L22 108, 2004.Wilson, J., Gobble, C., and Chickos, J.: Vaporization, sublimation, andfusion enthalpies of some saturated and unsaturated long chain fatty acidsby correlation gas chromatography, J. Chem. Eng. Data, 60, 202–212,2015.Yatavelli, R. L. N., Stark, H., Thompson, S. L., Kimmel, J. R., Cubison,M. J., Day, D. A., Campuzano-Jost, P., Palm, B. B., Hodzic, A., Thorn-ton, J. A., Jayne, J. T., Worsnop, D. R., and Jimenez, J. L.: Semicon-tinuous measurements of gas–particle partitioning of organic acids in aponderosa pine forest using a MOVI-HRToF-CIMS, Atmos. Chem. Phys.,14, 1527–1546, 2014.You, Y., Renbaum-Wolff, L., Carreras-Sospedra, M., Hanna, S. J., Hi-ranuma, N., Kamal, S., Smith, M. L., Zhang, X., Weber, R. J., Shilling,J. E., Dabdub, D., Martin, S. T., and Bertram, A. K.: Images reveal thatatmospheric particles can undergo liquid-liquid phase separations, Proc.Nat. Acad. Sci. U.S.A., 109, 13 188–13 193, 2012.Yu, J., Cocker, III, D. R., Griffin, R. J., Flagan, R. C., and Seinfeld,J. H.: Gas-phase ozone oxidation of monoterpenes: gaseous and particu-late products , J. Atmos. Chem., 34, 207–258, 1999.Zahardis, J. and Petrucci, G. A.: The oleic acid-ozone hetergeneous reactionsystem: products, kinetics, secondary chemistry, and atmoshperic impli-cations of a model system - a review, Atmos. Chem. Phys., 7, 1237–1274,2007.Zaveri, R. A., Easter, R. C., Shilling, J. E., and Seinfeld, J. H.: Modelingkinetic partitioning of secondary organic aerosol and size distribution dy-namics: representing effects of volatility, phase state, and particle-phasereaction, Atmos. Chem. Phys., 14, 5153–5181, 2014.150Zelenyuk, A., Imre, D., Bera´nek, J., Abramson, E., Wilson, J., and Shri-vastava, M.: Synergy between secondary organic aerosols and long-rangetransport of polycyclic aromatic hydrocarbons, Environ. Sci .Technol., 46,12 459–12 466, 2012.Zhang, K., Yang, J., Yu, X., Zhang, J., and Wei, X.: Densities and viscositiesfor binary mixtures of poly(ethylene glycol) 400 + dimethyl sulfoxide andpoly(ethylene glycol) 600 + water at different temperatures, J. Chem.Eng. Data, 56, 3083–3088, 2011.Zhang, Y., Sanchez, M. S., Douet, C., Wang, Y., Bateman, A. P., Gong,Z., Kuwata, M., Renbaum-Wolff, L., Sato, B. B., Liu, P. F., Bertram,A. K., Geiger, F. M., and Martin, S. T.: Changing shapes and impliedviscosities of suspended submicron particles, Atmos. Chem. Phys., 15,7819–7829, 2015.Zhao, B., Wang, S., Donahue, N. M., Chuang, W., Hildebrandt Ruiz, L.,Ng, N. L., Wang, Y., and Hao, J.: Evaluation of one-dimensional andtwo-dimensional volatility basis sets in simulating the aging of secondaryorganic aerosol with smog-chamber experiments, Environ. Sci .Technol.,49, 2245–2254, 2015.Zhou, S., Lee, A. K. Y., McWhinney, R. D., and Abbatt, J. P. D.: Burialeffects of organic coatings on the heterogeneous reactivity of particle-bornebenzo[a]pyrene (BaP) toward ozone, J. Phys. Chem. A, 116, 7050–7056,2012.Zhou, S., Shiraiwa, M., McWhinney, R. D., Po¨schl, U., and Abbatt, J. P. D.:Kinetic limitations in gas-particle reactions arising from slow diffusion insecondary organic aerosol, Farad. Discuss., 165, 391–406, 2013.Zhu, L., Attard, P., and Neto, C.: Reconciling slip measurements in sym-metric and asymmetric systems, Langmuir, 28, 7768–7774, 2012.Zobrist, B., Marcolli, C., Pedernera, D. A., and Koop, T.: Do atmosphericaerosols form glasses?, Atmos. Chem. Phys., 8, 5221–5244, 2008.151Appendix AAppendix to chapter 3A.1 Composition of SOM from ozonolysis ofα-pineneSOM formed from the ozonolysis of α-pinene is largely composed of water-soluble organic compounds. Hall and Johnston (2011) determined that 63(±27) % of the total particle mass is extractable from filters with water.Heaton et al. (2007) compared the mass spectra from extracts using 100% water and 50:50 methanol:water and showed that there was little differ-ence in the chemical composition of the two extracts. Cloud condensationmeasurements suggest that secondary organic material generated from theozonolysis of α-pinene is not limited by solubility of the organic materialin water under similar conditions as used in our experiments (King et al.,2009). Finally, SOM produced by limonene ozonolysis, a species structurallyrelated to α-pinene, has been shown to be nearly completely water-solubleby comparing water and acetonitrile extracts (Bateman et al., 2010).A.2 Viscosity prediction using mixing rulesA common approach to estimating the viscosity of mixtures (ηmix) is toapply mixing rules to the viscosities of the pure components of the mixturebased on experimental data and/or theory. Among the simplest of mixingrules for complex mixtures was first proposed by Arrhenius (1885) and con-sidered the mole fraction of each component (xi) and the viscosity of eachpure component (ηi) (Equation A.1) (Grunberg and Nissan, 1949):152A.2. Viscosity prediction using mixing ruleslog(ηmix) =n∑i=1xilog(ηi) (A.1)Equation A.1 was later modified by (Grunberg and Nissan, 1949) toinclude a group interaction parameter, G, accounting for non-ideality of thecomplex mixture, the result of which was a closer fit to the experimentaldata (Equation A.2):log(ηmix) =n∑i=1xilog(ηi) +n∑i=1n∑j=1xixjGi,j (A.2)If the interaction parameter, Gi,j , between the SOM components is small,and the mole fraction of water is a linear function of RH, Equation A.2predicts a linear dependence of log(ηSOM−H2O) vs. RH, roughly consistentwith what is observed in Fig. 3.3, where log(ηSOM−H2O) is the viscosity ofthe SOM-water mixture. Any deviation from linearity, for instance between30 and 40 % RH, may be due to non-ideality and/or a gel or glass transitionsin that RH regime. Accurate predictions of viscosities in aqueous mixturesmay require an additional interaction term in Equation A.2, the result ofwhich is further non-linearity in the plots of log(ηmix) (or log(ηSOM−H2O))(Marczak et al., 2012).It should be noted that the use of equations such as Equations A.1 andA.2 for mixtures containing components with very large differences in theviscosities of the pure components is not well established and are used hereonly for a first-order approximation of the trend in the viscosity expectedwith changing RH. However, other formulations modelling the viscosity ofspecific polar mixtures such as sucrose-water mixtures predict a linear cor-relation between log(ηmix) and concentration even with a large difference inpure component viscosities Longinotti and Corti (2008).153Appendix BAppendix to chapter 5B.1 Effect of carrier gas flow on SOM particlepropertiesTo determine whether particles evapourate whilst being exposed to a flowof N2 gas, a sample (generated with a mass concentration of 6,000 µg m−3)was mounted in the flow cell and exposed to dry (<0.5 % RH) N2 gasfor a period of 45 h. A series of nine images were taken of each of fiveparticles over a time period of 45 h. The area at the particle-substrateinterface of each particle was measured using Leica software, and equilibriumcontact angle measurements of particles in the sample were made after thecompletion of the experiment. From this information, the volume of each ofthe particles was determined at each time point, with the equilibrium contactangle assumed to remain constant during the duration of the experiment. Toremove the potential of photo-induced changes to the sample the light sourcewas only turned on when acquiring images. poke-and-flow experiments werealso performed after 1 h and 45 h of exposure to the dry N2 gas flow todetermine whether the viscosity of the particles changed due to the extendedexposure. The results of these studies are shown in Figure B.1 and TableB.1.154B.2. Particle-to-particle and sample-to-sample variabilityB.2 Variability between particles in the samesample, and between all particles producedunder equivalent conditionsTo determine the variability in τ exp,flow and the variability in the simulatedlower and upper limits of viscosity for particles produced using equivalentconditions, the percent relative standard deviations (% RSD) have beencalculated for poke-and-flow experiments performed at <0.5 % RH, and arereported in Table B.3.Table B.3 shows the % RSD for all particles produced under equivalentconditions ranged from 26-117 %. For samples on the same substrate, the %RSD ranged from 14-83 %. In general there appears to be a reasonable levelof reproducibility in results both between particles on the same substrate,and between particles on separate substrates produced under equivalent con-ditions. The majority of the uncertainty in the reported viscosities is due touncertainty in the physical parameters used during simulations, rather thanexperimental variability or error, as can be seen in Figs. 5.2(b), 5.4(b), andB.3(b).B.3 Calculation of viscosity for prior studies ofα-pinene derived SOMSaleh et al. (2013) and Robinson et al. (2013) reported mixing times for par-ticles of a given size, which have been used to calculate diffusion coefficientsthrough the relationship,D =d2p4pi2τmixing(B.1)where dp is the diameter of the particle (m), and τmixing is the mixingtime in s (Shiraiwa et al., 2011a). The calculated diffusion coefficients forSaleh et al. (2013) and Robinson et al. (2013) should be considered as lowerlimits, as the mixing times used were the upper limit of those reported,155B.3. Calculated viscosity for prior studies of α-pinene derived SOMand other processes besides molecular diffusion within the particles mayhave been the rate determining step for mixing in their experiments. Thesecalculated upper limits to diffusion coefficients were then converted to lowerlimits of viscosities through the Stokes-Einstein relationship,µ =kBTxpirD(B.2), where kB is the Boltzmann constant (J K−1), T is the temperature (K),x is a coefficient ranging from 4-6 dependent upon the assumption of slipor no-slip at the surface of the diffusing species, and r is the hydrodynamicradius of the diffusing molecule (m). A summary of the values used tocalculate viscosities is included in Table B.4.As the size of the diffusing molecules were not known in Saleh et al.(2013) and Robinson et al. (2013), a hydrodynamic radius of 0.38 nm wasassumed, which corresponding to the radii of a symmetrically sphericalmolecule of molar mass 175 g mol−1 (Huff-Hartz, 2005) and density 1.3g cm−3 (Chen and Hopke, 2009). Further, x has been given a value of 4, togive conservative upper limits to viscosity.Abramson et al. (2013) determined the diffusion coefficient for pyrenemolecules in particles of SOM generated via the ozonolysis of α-pinene. Thehydrodynamic radius of pyrene is 0.4 nm, whilst, as suggested by Abramsonet al. (2013) pyrene may form clusters consisting of up to 1,000 molecules,with a 1000 molecule cluster having a radius of ≈4 nm. Hence, when cal-culating viscosity using Equation B.2, for the studies of Abramson et al.,values of r = 0.4 and 4 nm were used. Further, values of x = 4-6 were usedin Equation B.2 to calculate conservative lower and upper limits of viscosity.Cappa and Wilson (2011) observed the change in chemical composition ofSOM particles produced via the ozonolysis of α-pinene as the particles wereheated, and conservatively estimated an upper limit of diffusion coefficientfor the particles. A value of x = 6 has been used in Equation B.2, alongwith a value of r = 4 nm, in order to determine a lower limit of viscosity.Perraud et al. (2012) studied the particulate nitrate concentration inSOM particles generated via the ozonolysis of α-pinene and determined an156B.4. Simulations of fluid flow for particles that exhibit crackingupper limit for the diffusion coefficient of the particles. A value of x = 6has been used in Equation B.2, along with a value for r of 4 nm, in order todetermine a lower limit of viscosity.B.4 Simulations of fluid flow for particles thatexhibit crackingIn some cases (for the water-soluble SOM at low RH) the needle did notpenetrate the particle. Instead, the needle caused the particle to ’crack’,resulting in sharp, defined edges in the SOM (see Fig. 5.6, panel b2 for anexample). In cases where cracking occurred, the material was observed foran extended period of time (at least 12 h). If, over that time, the sharp,defined, edges exhibited no detectable movement, the observation time wastaken as a lower limit of τ exp,flow.Particles that exhibited cracking behaviour and no detectable flow overthe course of an experiment were simulated using a particle of quarter-spheregeometry, with one flat surface in contact with a solid substrate (see Fig.2.10, and Movie S5 in Renbaum-Wolff et al., 2013a). The bottom surface,which represented the material-substrate interface, was allowed to undergofree deformation in the horizontal plane. All other surfaces were allowed toundergo free deformation in all directions. In these simulations the viscosityof the particle was varied until the sharp edge at the top of the particlemoved by 0.5 µm over the experimental time. A value of 0.5 µm was chosenbecause this amount of movement is a clearly detectable threshold for themicroscopy. Thus, for experiments for which no detectable movement wasobserved in microscope images, the viscosity determined via this method isa lower limit. The values of density, particle-substrate slip length, surfacetension, and contact angle used when simulating the lower limit of viscosityfor these particles are detailed in Table B.5.157B.5. Tables and figuresB.5 Tables and figuresTable B.1: Summary of τ exp,flow times and viscosities of sample analysedafter both 1 hour and 45 hours of exposure to a dry (<0.5 % RH) flow ofNitrogen gas. Originally published in Grayson et al. (2015b).Exposure τ exp,flowMean simulated viscosity ± 95 %time (sec) aconfidence intervals (Pa s) bLower limit Upper limit1 hour104.31.5 x 104 ± 5.0 x 103 6.4 x 105 ± 9.0 x 104(80.3, 120.5)45 hours107.83.0 x 104 ± 9.3 x 103 1.0 x 106 ± 2.3 x 105(80.6, 141.1)a τexp,flow values represent experimental values in the form ”mean (5th percentile, 95thpercentile)”.Table B.2: Experimentally determined contact angles for each of the samplesstudied. The range of values represent the 95 % confidence intervals of thevalues measured for multiple particles. Originally published in Grayson et al.(2015b).Sample name (production mass conc- Particle-substrate contact angle (0)entration during SOM production) Sample 1 Sample 2 Sample 3Water-soluble SOM (14,000 µg m−3) 43.5 - 50.0 47.2 - 53.3 50.5 - 72.0Flow tube sample #1 (14,000 µg m−3) 57.5 - 63.7 49.6 - 55.4 18.3 - 21.4Flow tube sample #2 (6,000 µg m−3) 59.8 - 61.3 60.5 - 66.9 63.3 - 67.2Flow tube sample #3 (3,200 µg m−3) 41.2 - 52.2 38.5 - 45.3 49.2 - 51.1Flow tube sample #4 (1,100 µg m−3) 31.4 - 35.6 61.5 - 65.5 44.3 - 47.8Flow tube sample #5 (520 µg m−3) 55.5 - 61.8 56.2 - 60.6 36.1 - 47.0Chamber sample #1 (230 µg m−3) 64.5 - 69.0 64.1 - 66.5Chamber sample #2 (121 µg m−3) 60.2 - 65.1 60.7 - 80.1158B.5.TablesandfiguresTable B.3: Summary of the percent relative standard deviation (% RSD) in τ exp,flow, lowerlimits of viscosity, and upper limits of viscosity for particles produced using equivalentconditions and studied via the poke-and-flow technique in combination with simulations offluid flow at <0.5 % RH. Three samples were studied per production mass concentrationin the flow tube. Values prior to parentheses represent the relative standard deviationbetween all particles studied that were produced at a given mass concentration, whilst thevalues inside each parenthesis represent the average relative standard deviation betweenparticles on the same substrate. Originally published in Grayson et al. (2015b).SOM mass particle % RSD of % RSD of simulated % RSD of simulatedconcentration (µg m−3) τ exp,flow lower limit of viscosity upper limit of viscosity520 54 (30) 68 (67) 56 (39)1,100 89 (24) 117 (44) 96 (27)3,200 27 (22) 67 (42) 45 (30)6,000 26 (20) 103 (47) 58 (29)14,000 31 (27) 57 (27) 40 (28)159B.5. Tables and figuresTable B.4: Summary of parameters used to estimate viscosity from liter-ature studies of SOM produced via the ozonolysis of α-pinene. Originallypublished in Grayson et al. (2015b).Reference dp (nm) τmixing (sec) x r (nm)Cappa and Wilson (2011) N/A 6 4.0Perraud et al. (2012) N/A 6 4.0Saleh et al. (2013) a 112 b 3,600 4 0.38Saleh et al. (2013) c 38 b 3,600 4 0.38Abramson et al. (2013) N/A 4 - 6 0.4 - 4.0Robinson et al. (2013) >158 b 60 4 0.38a Values for experiments conducted with an SOM mass concentration of 350 µgm−3.b An aerodynamic diameter was reported, and has been converted to a geometricdiameter here.c Values for experiments conducted with an SOM mass concentration of 1-12 µgm−3.160B.5.TablesandfiguresTable B.5: Physical parameters when simulating particles that don’t exhibit flow inCOMSOL. Originally published in Grayson et al. (2015b).Density Slip length Surface tension Contact(kg m−3) a (m) b (mN m−1) c angle (0)Value used to determine1,300 1 - 1.7 x 10−8 40 90lower limit of viscositya Density was varied from 1,000 - 1,400 kg m−3 based on the work of Chen and Hopke (2009) anddetermined to have no effect upon simulations values. As such a median value of 1300 kg m−3 wasused.b For references and rationale see Chapter 4.c Surface tension value based on work on Tuckermann and Cammenga (2004).161B.5. Tables and figuresFigure B.1: A plot of particle volume vs. time for five particles exposed toa dry (<0.5 % RH) N2 gas flow. Dotted lines represent the measured meansize of a particle. Error bars on the y-axis represent the uncertainty inmeasuring both the area of the particle, and the equilibrium contact angle,at the particle-substrate interface. Originally published in Grayson et al.(2015b).162B.5. Tables and figuresFigure B.2: (a) Schematic representation of instrumental setup for contactangle images. (b) Fluorescence image obtained of an SOM particle. Thegreen overlay is used to determine the contact angle of the particle, in thiscase 60 0, and was produced using the LB-ADSA plugin for ImageJ. Origi-nally published in Grayson et al. (2015b).163B.5. Tables and figuresFigure B.3: Plot of production mass concentration vs. viscosity for wholeSOM produced via the ozonolysis of α-pinene and studied at <5 % RH.Shown are the results determined here along with those previously reportedin literature (Abramson et al., 2013; Cappa and Wilson, 2011; Perraud et al.,2012; Renbaum-Wolff et al., 2013a; Robinson et al., 2013; Saleh et al., 2013;Zhang et al., 2015). Originally published in Grayson et al. (2015b).164Appendix CAppendix to chapter 7C.1 TableTable C.1: A list of compounds included in the study, alongwith selected chemical and physical properties.CompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 ηButane 1 58 0 5.38 -3.78 a b2-Methylbutane 1 72 0 4.96 -3.67 a bIsopentane 1 72 0 4.95 -3.67 c cPentane 1 72 0 4.57 -3.65 a b2-Methyl pentane 1 86 0 4.18 -3.55 a bHexane 1 86 0 4.3 -3.53 a b3-Methyl pentane 1 86 0 4.15 -3.51 a bContinued on next page165C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 η3-Methyl hexane 1 100 0 4.02 -3.46 c c2,2-Dimethyl butane 1 86 0 4.33 -3.45 a b2,4-Dimethyl pentane 1 100 0 3.9 -3.44 d d2,3-Dimethyl butane 1 86 0 4.29 -3.42 d b2-Methyl hexane 1 100 0 3.72 -3.42 d dHeptane 1 100 0 3.73 -3.41 a b2,2,4-Trimethyl pentane 1 114 0 3.47 -3.32 a bOctane 1 114 0 3.24 -3.3 a b2,2,3-Trimethyl pentane 1 114 0 3.4 -3.28 d d2,2,3-Trimethyl butane 1 100 0 3.92 -3.24 d dNonane 1 128 0 2.73 -3.18 c c2,7-Dimethyl octane 1 142 0 2.65 -3.08 a bDecane 1 142 0 2.25 -3.07 a bUndecane 1 156 0 1.74 -2.97 a bDodecane 1 170 0 1.25 -2.86 a bTetradecane 1 198 0 0.26 -2.68 a bTridecane 1 184 0 0.75 -2.65 a bPentadecane 1 212 0 2.8 -2.6 a bContinued on next page166C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 ηPropylene 2 42 0 6.06 -4.09 a e1,3-Butadiene 2 54 0 5.41 -3.85 a b1-Butene 2 56 0 5.47 -3.85 a e1-Pentene 2 70 0 4.73 -3.71 d d2-Methyl-2-butene 2 70 0 4.77 -3.69 c c2-Methyl-1,3-butadiene 2 68 0 4.86 -3.68 a b2-Methyl-1-butene 2 70 0 4.71 -3.65 d d1-Hexene 2 84 0 4.39 -3.6 a e1-Heptene 2 98 0 3.87 -3.48 a eFuran 2, 3, 11 68 0.25 4.7 -3.42 d d1-Octene 2 112 0 3.36 -3.35 a e1-Nonene 2 126 0 2.85 -3.24 a eCyclohexene 2 82 0 4.07 -3.2 a b1-Decene 2 140 0 2.35 -3.12 a eAllyl alcohol 2, 5 58 0.33 3.45 -2.91 a bCyclopentene 3 68 0 4.5 -3.46 d dCyclopentane 3 70 0 4.64 -3.38 a bMethyl cyclopentane 3 84 0 4.06 -3.32 a bContinued on next page167C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 ηMethyl cyclohexane 3 98 0 3.61 -3.16 a bEthyl cyclohexane 3 112 0 3.32 -3.11 a bLimonene 2, 3 136 0 3.01 -3.07 a bCyclohexane 3 84 0 4.09 -3.05 a balpha-pinene 3 136 0 3.3 -2.89 a bIndane 3, 4 134 0.11 2.2 -2.87 c cFurfural 2, 3, 8, 11 96 0.4 2.34 -2.8 c cbeta-pinene 2, 3 136 0 3.26 -2.8 a btrans-Decahydronaphthalene 3 138 0 2.15 -2.71 c cCyclohexanone 3, 9 98 0.17 3.25 -2.69 a bIsophorone 3, 9 138 0.11 1.73 -2.63 c ccis-Decahydronaphthalene 3 138 0 1.92 -2.52 c cToluene 4 92 0 3.43 -3.25 a bm-Xylene 4 106 0 2.96 -3.24 a bBenzene 4 78 0 4.08 -3.22 a bp-Xylene 4 106 0 3.09 -3.21 a bEthyl benzene 4 106 0 3.29 -3.2 a bStyrene 2, 4 104 0 3.03 -3.16 a bContinued on next page168C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 ηCumene 4 120 0 2.75 -3.13 c co-Xylene 4 106 0 3.06 -3.12 a bCymene 4 134 0 2.82 -3.1 a bButyl benzene 4 134 0 2.42 -3.02 a bAnisole 4, 11 108 0.14 2.6 -2.98 c cPhenetole 4, 11 122 0.13 2.2 -2.92 c cNaphthalene 4 128 0 1.01 -2.65 c fAnisaldehyde 4, 8, 11 136 0.25 0.62 -2.43 c gEugenol 2, 4, 5, 11 164 0.2 -0.19 -2.14 a cDiethyl phthalate 4, 6 222 0.33 0.35 -2 c bm-Cresol 4, 5 108 0.14 0.36 -1.89 c cMethanol 5 32 1 4 -3.26 a bTetrahydrofuran 3, 5 72 0.25 4.12 -3.26 d dEthanol 5 46 0.5 3.8 -2.98 a b4-Methyl-3-heptanol 5 130 0.13 2.63 -2.96 c c5-Methyl-3-heptanol 5 130 0.13 2.51 -2.93 c c1-Propanol 5 60 0.33 3.36 -2.72 a b2-Propanol 5 60 0.33 3.59 -2.68 a bContinued on next page169C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 η1-Butanol 5 74 0.25 2.96 -2.6 a bDiacetone alcohol 5, 9 116 0.33 2.31 -2.55 c c2-Butanol 5 74 0.25 3.38 -2.51 a b1-Pentanol 5 88 0.2 2.51 -2.47 a b3-Methyl-1-butanol 5 88 0.2 2.63 -2.43 a bt-Amyl alcohol 5 88 0.2 3.38 -2.41 a b2-Heptanol 5 116 0.14 1.79 -2.4 c c2-Hexanol 5 102 0.17 3.1 -2.39 a b4-Heptanol 5 116 0.14 2.01 -2.38 c c2-Methyl-2-propanol 5 74 0.25 3.75 -2.36 a b2-Methyl-1-butanol 5 88 0.2 2.61 -2.35 a b2-Methyl-1-pentanol 5 102 0.17 2.12 -2.28 a b1-Octanol 5 130 0.13 1.02 -2.12 h b1,2-Ethanediol 5 62 1 1.04 -1.77 i jDiethylene glycol 5, 11 106 0.75 -0.34 -1.52 c c1,2-Propanediol 5 76 0.67 1.24 -1.39 i c1,2-Butanediol 5 90 0.5 0.95 -1.3 i kCyclohexanol 3, 5 100 0.17 2.66 -1.24 c cContinued on next page170C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 η1,2-Hexanediol 5 118 0.33 0.16 -1.21 i l1,2-Pentanediol 5 104 0.4 0.53 -1.2 i k1,5-Pentanediol 5 104 0.4 0.05 -1.01 c kGlycerol 5 92 1 -2.02 -0.03 m cHexanetriol 5 134 0.5 -3.94 0.42 n oPropionic acid 6 74 0.67 2.62 -2.99 a b2-Methylpropanoic acid 6 88 0.5 2.23 -2.91 c cAcetic acid 6 60 1 3.06 -2.91 d dButanoic acid 6 88 0.5 2.32 -2.85 c cFormic acid 6 46 2 3.71 -2.8 a bHexanoic acid 6 116 0.33 0.74 -2.49 c pOctanoic acid 6 144 0.25 -0.72 -2.3 q bPhenethyl alcohol 4, 6 122 0.13 0.95 -1.95 c bOleic acid 2, 6 282 0.11 -5 -1.54 r sAcetic anhydride 7 102 0.75 3.3 -3.07 a bPropionic anhydride 7 130 0.5 2.22 -2.98 a bAcetaldehyde 8 44 0.5 5.08 -3.67 a bPropanal 8 58 0.33 4.42 -3.39 d dContinued on next page171C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 ηHexanal 8 100 0.17 3.1 -3.09 t uParaldehyde 8 132 0.5 3.19 -2.97 c cOctanal 8 128 0.13 2.18 -2.92 t uNonanal 8 142 0.11 1.67 -2.86 t uDecanal 8 156 0.1 1.26 -2.81 t uAcetone 9 42 0 4.48 -3.5 a bMethyl ethyl ketone 9 72 0.25 4.09 -3.4 a b2-Butanone 9 72 0.25 3.88 -3.37 d d3-Pentanone 9 86 0.2 3.69 -3.35 a b2-Pentanone 9 86 0.2 3.67 -3.33 a b4-Methyl-2-pentanone 9 100 0.17 3.44 -3.23 d d2-Heptanone 9 114 0.14 2.65 -3.15 c cMethyl formate 10 60 1 4.61 -3.48 a bEthyl formate 10 74 0.67 4.53 -3.42 a bMethyl acetate 10 74 0.67 2.69 -3.42 a bEthyl acetate 10 88 0.5 4.06 -3.37 a bMethyl propionate 10 88 0.5 3.87 -3.35 a bPropyl formate 10 88 0.5 4.04 -3.31 a bContinued on next page172C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 ηMethyl isobutanoate 10 102 0.4 3.87 -3.31 c cEthyl propionate 10 102 0.4 3.62 -3.29 a bMethyl butyrate 10 102 0.4 3.47 -3.28 a bPropyl acetate 10 102 0.4 3.73 -3.26 a bEthyl butyrate 10 116 0.33 3.39 -3.2 a bButyl formate 10 102 0.4 3.52 -3.19 a bIsobutyl acetate 10 116 0.33 3.43 -3.17 c cButyl acetate 10 116 0.33 3.42 -3.16 a b2-Methylpropyl acetate 10 116 0.33 3.11 -3.16 d dButyl propionate 10 130 0.29 3.31 -3.11 a bButyl butyrate 10 144 0.25 3.3 -3.02 a b1,4-Dioxane 10 88 0.5 3.7 -2.93 a vMethyl benzoate 10 136 0.25 2.69 -2.73 a bDiethyl ether 11 74 0.25 4.65 -3.61 d dEthyl propyl ether 11 88 0.2 4.16 -3.49 d dDipropyl ether 11 102 0.17 4.03 -3.4 c cn-Propyl ether 11 100 0.17 3.85 -3.4 a bEthylene glycol dimethyl ether 11 90 0.5 3.97 -3.38 a bContinued on next page173C.1.TableTable C.1 – continued from previous pageCompoundFunctional MwO:CLog10(PL0) Log10(η) Referencesgroup(s) / g mol −1 / Pa / Pa s PL0 ηButyl ethyl ether 11 102 0.17 3.64 -3.38 d dEthylene glycol diethyl ether 11 118 0.33 3.6 -3.22 a bDibutyl ether 11 130 0.13 2.96 -3.2 c cEthylene glycol monomethyl ether 11 76 0.67 3.33 -2.81 a bPEG 600 11 600 0.54 -0.7 -0.77 w xFunctional groups: 1: alkane, 2: alkene, 3: cyclic (non-aromatic), 4: cyclic (aromatic), 5: alcohol, 6: acid, 7: acidanhydride, 8: aldehyde, 9: ketone, 10: ester, 11: etherReferences: a: Perry and Green (2008), b: Viswanath et al. (2007), c: Haynes (2015), d: Suzuki et al. (1996),e: Hashim et al. (2012), f: Evans (1938), g: Roy et al. (2008), h: Daubert and Danner (1989), i: Verevkin (2004),j: Li et al. (2010), k: Czechowski and Jadzyn (2003), l: Jarosiewicz and Czechowski (2004), m: Cammenga et al.(1977), n: Cai et al. (2015), o: Aldrich Chemical Company (1996), p: Katritzky et al. (2000), q: Cappa et al.(2008), r: Noureddini et al. (1992), s: Wilson et al. (2015), t: Verevkin et al. (2003), u: Djojoputro and Ismadji(2005), v: Vinson and Martin (1963), w: Aschenbrenner et al. (2009), x: Zhang et al. (2011)174

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0300661/manifest

Comment

Related Items