The Measurement and Characterisation of Aerosol in the Urban Atmosphere (PM 10) and an Evaluation of the Sources of these Particles by Number Andrew Lindsay Dye Abstract The link between human health and the mass of fine particulate matter below 10 tm (PM10) in air is well documented. Current research suggests that the number, size and shape of particles may be of most concern and that in the urban atmosphere combustion sources of PM10, especially diesel engine sources, dominate the fine (< 1µm) and ultra-fine (<0.1 µm) particles. Despite this, the number, size and shape of particles in urban air has not been reported to any great extent or detail, and the percentage contribution to the numbers of particles from different sources is largely unknown. The objectives of this research were to characterise fine particles with respect to their morphology and thus apportion the sources of particles by number. Urban aerosol above 1 µm was initially examined to study the fluctuations in PM10 number and make retrospective analysis of periods of elevated PM10 for source identification in Plymouth, UK. Aerosol was collected via a Burkard spore trap and examined using light microscopy with image analysis between 16 March 1995 and 31 August 1996, at a background site in Plymouth, UK. Two periods, 19 Januamy-4 February and 10-25 March 1996, identified as UK wide PM 10 episodes, were retrospectively studied and compared with PM10 mass measurements. The mean number count for the whole period was 10.5 x 104 ± 7.9 x 104 particles m-3 . The two PM10 episodes had elevated average number concentrations of 13.5 x 104 ± 7.6 x 104 particles m-3 for 19 January - 4 February 1996, and 13.0 x 104 ± 9.7 x 104 particles n13 for 10-25 March 1996. During the periods of elevated PM10 the tapered element oscillating microbalance (TEOM) mass of particles had a low correlation with the particles less than 5 µm and an increased correlation to the particles greater than 5 µm in size. Outside of these peak periods the PM10 TEOM mass was most closely correlated with the number of particles less than 5 µm in size. This work shows the difference in urban aerosol during periods of air quality guideline exceedence. These findings agree with literature that an aged continental aerosol source has a key role in the generation of UK wide PM10 mass exceedances. Further analysis of the fine urban aerosol (< 1 µm) was made using direct sampling of urban aerosol on to porous carbon films (PCF) developed in this research. The efficiency of collection was low (ca 5%) but the samples were representative and enabled transmission electron microscopy (TEM) for sub-micron particle analysis. Measurement was made of the fractal dimensions and diameter of particles. This was used to identif' any ageing and ultimately the sources of aerosol. PCF were used in the simultaneous collection of urban roadside and background aerosol, on seven dates between December 1996 and August 1997 in Plymouth, UK. The average perimeter fractal dimension (PFD) of aerosol was consistently significantly greater at the roadside than the background (+ 0.02), indicative of a smoother, aged aerosol at the background site. The sampling of a variety of combustion engines was made for source identification purposes. The particle morphology produced from the diesel engines showed great uniformity of particle morphology with varying speed and load; no consistent significant differences were found. The morphology results were comparable to other density fractal dimensions and penmneter fractal dimension values found in other studies for diesel. A natural log relationship between the median particle size and the median PFD was found for the diesel engine sources but not in petrol samples. This natural log trend was considered as a tentative 'fingerprint' of diesel engine combustion and was in harmony with literature values of PFD for diesel engine particles. Using the fractal measures, size and particle classification the bulk of aerosol was identified as from hydrocarbon combustion sources; ca. 88-92% of the roadside and ca. 77-86% of background. A component of carbon ceno-spheres were identified contributing ca. 6-12% of both the roadside and background aerosol. Non-combustion particles increased from ca. 1-4% of the roadside to ca. 7-9% of the background, as did the proportion of aged combustion particles, from 0-1% of roadside to 2-3% of the background aerosol. A strong correlation for the median size vs. PFD morphology curve between, the roadside and diesel sources (0.93 - 0.95) and the background and petrol sources was found (0.95). The roadside aerosol was significantly different to the petrol source in the 120-220nm size range (p=O.007) and there was a low correlation of the petrol and the roadside size vs. morphology curve (0.66). This suggests the domination of roadside aerosol by diesel engine particles. The background aerosol was similar to both diesel and petrol engine sources, especially from a dilution tunnel, thus indicative of a mixture of sources and an aged combustion aerosol. Roadside sources thus dominate the fine and ultra fine urban aerosol by number as compared to most other studies which have only apportioned the sources of particles in the air by mass.

Airframes operating in the hypersonic regime are subjected to complex structural and thermal loads. Structural loads are a result of aggressive high G maneuvers, rapid vehicle acceleration and deceleration, and dynamic pressure, while thermal loads are a result of aerodynamic heating. For such airframes, structural members are typically constructed from steel, titanium and nickel alloys. However, with most materials, rapid elevations in temperature lead to undesirable changes in material properties. In particular, reductions in strength and stiffness are observed, along with an increase in thermal conductivity, specific heat and thermal expansion. Thus, hypersonic airframes are typically designed with external insulation, active cooling or a thermal protection system (TPS) added to the structure to protect the underling material from the effects of temperature. Such thermal protection may consist of adhesively bonded, pinned, and bolted thermal protection layers over exterior panels. These types of attachments create abrupt changes in thermal expansion and stiffness that make the structure susceptible to cracking and debonding as well as adding mass to the airframe. One of the promising materials concepts for extreme environments that was introduced in the past is the so-called Spatially Tailored Advanced Thermal Structures (STATS). The concept of STATS is rooted in functionally graded materials (FGMs), in which a directional variation of material properties exists. These materials are essentially composites and consist of two or more phases of distinct materials in which the volume fractions of each phase continuously change in space. Here, the graded material will serve a dual-purpose role as both the structural/skin member and thermal management with the goal of reducing the weight of the structure while maintaining structural soundness. This is achieved through the ability to tailor material properties to create a desired or enhanced thermomechanical response through spatial variation (e.g. grading). The objective of this study is to present a computational framework for modeling and evaluating the thermomechanical response of STATS and FGMs for highly maneuverable hypersonic (Mach > 5) airframes. To meet the objective of this study, four key steps have been defined to study the thermomechanical response of such materials in extreme environments. They involve: (1) modeling of graded microstructures; (2) validation of analytical and numerical modeling techniques for graded microstructures; (3) determination of effective properties of variable composition composites; (4) parametric studies to evaluate the performance of FGMs for use in the hypersonic operating environment; (5) optimization of the material spatial grading in hypersonic panels aiming to improve the thermomechanical performance. Modeling of graded microstructures, representing particulate reinforced FGMs, has been accomplished using power law distribution functions to specify the spatial variation of the constituents. Artificial microstructures consisting of disks and spheres have been generated using developed algorithms. These algorithms allow for the creation of dense packing fractions up to 0.61 and 0.91 for 2D and 3D geometry, respectively. Effective properties of FGMs are obtained using micromechanics models and finite element analysis of representative volume elements (RVEs). Two approaches have been adopted and compared to determine the proper RVE for materials with graded microstructures. In the first approach, RVEs are generated by considering regions that have a uniform to slow variation in material composition (i.e., constant volume fraction), resulting in statistically homogenous piecewise RVEs of the graded microstructure neglecting interactions from neighboring cells. In the second approach, continuous RVEs are generated by considering the entire FGM. Here it is presumed that modeling of the complete variation in a microstructure may influence the surrounding layers due to the interactions of varying material composition, particularly when there is a steep variation in material composition along the grading direction. To determine these effects of interlayer interactions, FGM microstructures were generated using three different types of material grading functions, linear, quadratic and square root, providing uniform, gradual and steep variations, respectively. Two- and three-dimensional finite element analysis was performed to determine the effective temperature-dependent material properties of the composite over a wide temperature range. The outcome of the computational analysis show that the similar effective properties are obtained by each of the modeling approaches. Furthermore, the obtained computational results for effective elastic, thermal, and thermal expansion properties are consistent with the known analytical bounds. Resulting effective temperature-dependent material properties were used to evaluate the time-dependent thermostructural response and effectiveness of FGM structural panels. Structural panels are subjected to time- and spatial-dependent thermal and mechanical loads resulting from hypersonic flight over a representative trajectory. Mechanical loads are the by-product of aggressive maneuvering at high air speeds and angles of attack. Thermal loads as a result of aerodynamic heating are applied to the material systems as laminar, turbulent and transitional heat flux on the outer surface. Laminar and turbulent uniform heat fluxes are used to evaluate the effectiveness of FGM panels graded in the through-thickness direction only. Transitional heat fluxes are used to evaluate the effectiveness of FGMs graded in two principal directions, e.g., through-thickness and the surface parallel to flow. The computational results indicate that when subjected to uniform surface heat flux, the graded material system can eliminate through-thickness temperature gradients that are otherwise present in traditional thermal protection systems. Furthermore, two-dimensional graded material systems can also eliminate through-thickness temperature gradients and significantly reduce in-plane surface temperature gradients when subjected to non-uniform surface aerodynamic heating.

To design a cyclone abatement system for particulate control, it is necessary to accurately estimate cyclone performance. In this cyclone study, new theoretical methods for computing travel distance, numbers of turns and cyclone pressure drop have been developed. The flow pattern and cyclone dimensions determine the travel distance in a cyclone. The number of turns was calculated based on this travel distance. The new theoretical analysis of cyclone pressure drop was tested against measured data at different inlet velocities and gave excellent agreement. The results show that cyclone pressure drop varies with the inlet velocity, but not with cyclone diameter. Particle motion in the cyclone outer vortex was analyzed to establish a force balance differential equation. Barth??s "static particle" theory, particle (with diameter of d50) collection probability is 50% when the forces acting on it are balanced, combined with the force balance equation was applied in the theoretical analyses for the models of cyclone cut-point and collection probability distribution in the cyclone outer vortex. Cyclone cut-points for different dusts were traced from measured cyclone overall collection efficiencies and the theoretical model for calculating cyclone overall efficiency. The cut-point correction models (K) for 1D3D and 2D2D cyclones were developed through regression fit from traced and theoretical cut-points. The regression results indicate that cut-points are more sensitive to mass median diameter (MMD) than to geometric standard deviation (GSD) of PSD. The theoretical overall efficiency model developed in this research can be used for cyclone total efficiency calculation with the corrected d50 and PSD. 1D3D and 2D2D cyclones were tested at Amarillo, Texas (an altitude of 1128 m / 3700 ft), to evaluate the effect of air density on cyclone performance. Two sets of inlet design velocities determined by the different air densities were used for the tests. Experimental results indicate that optimal cyclone design velocities, which are 16 m/s (3200 ft/min) for 1D3D cyclones and 15 m/s (3000 ft/min) for 2D2D cyclones, should be determined based on standard air density. It is important to consider the air density effect on cyclone performance in the design of cyclone abatement systems.

Background: Fraction of exhaled nitric oxide (FENO) is a promising non-invasive index of airway inflammation that may be used to assess respiratory effects of air pollution. We evaluated FENO as a measure of airway inflammation after controlled exposure to diesel exhaust or ozone. Methods: Healthy volunteers were exposed to either diesel exhaust (particle concentration 300 mu g/m(3)) and filtered air for one hour, or ozone (300 ppb) and filtered air for 75 minutes. FENO was measured in duplicate at expiratory flow rates of 10, 50, 100 and 270 mL/s before, 6 and 24 hours after each exposure. Results: Exposure to diesel exhaust increased FENO at 6 hours compared with air at expiratory flow rates of 10 mL/s (p = 0.01) and at 50 mL/s (p = 0.011), but FENO did not differ significantly at higher flow rates. Increases in FENO following diesel exhaust were attenuated at 24 hours. Ozone did not affect FENO at any flow rate or time point. Conclusions: Exposure to diesel exhaust, but not ozone, increased FENO concentrations in healthy subjects. Differences in the induction of airway inflammation may explain divergent responses to diesel exhaust and ozone, with implications for the use of FENO as an index of exposure to air pollution.

Factors that may govern diesel particulate matter (DPM) oxidation at low temperatures (~200°C) were studied using reactivity and TP-ToFSIMS analysis. Best-case scenarios that give maximum gasification rates were determined for DPM impregnated with KOH and non-catalyzed DPM using temperature programmed oxidation and isothermal experiments. Conditions of intimate catalyst-carbon contact (K/C molar ratio=1/50) and high NO2 concentrations (1%) to improve the reactivity of the carbon reactive sites were unable to meet the steady state gasification rate needed for particulate filter regeneration for a modern diesel engine at 200°C. Oxygen-free thermal annealing (>500°C) caused reactivity losses of a maximum of 40% that correspond to changes to surface morphology and/or concentration of oxygen-containing functional groups. TP-ToFSIMS identified surface functional group changes with temperature on non-dosed and NOX pre-dosed (1.5%NO, 1%NO2, 4.5%O2, balance helium) diesel soot and sucrose char. Detailed analysis of the NOX dosed sucrose char spectra using both inspection and principal component analysis techniques revealed that the 1200 ion fragments created could be reduced to five sets of ions that are chemically and kinetically distinct. These sets presumably represent surface functional groups on the carbon. For example, Set IV may represent carboxylic acid, lactone, or carboxylic anhydride functional groups. Based on these results a mechanism for the surface reaction of NO2 with carbon under vacuum conditions was postulated. At temperatures less than 200°C the ion fragments contain primarily carbon-NO2 type ions. As temperature increases between 200 and 400°C the ion fragments are primarily carbon-NO and carbon-N type fragments. At higher temperatures (>500°C) the surface is enriched with nitrogen containing functional groups. A surface reaction mechanism is proposed where NO2 is bonded to an armchair site and with increasing temperatures and molecular rearrangements the N is incorporated into the carbon ring. The initial surface composition of NOx containing functional groups changes within the area of relevance of low temperature soot regeneration (i.e. between 25° and 200°C). Further studies are needed to understand the effect of N-incorporation on carbon reactivity. No rate processes either in reactor studies or based on surface functional groups met the rate criteria for low temperature DPM oxidation.

Factors that may govern diesel particulate matter (DPM) oxidation at low temperatures (~200°C) were studied using reactivity and TP-ToFSIMS analysis. Best-case scenarios that give maximum gasification rates were determined for DPM impregnated with KOH and non-catalyzed DPM using temperature programmed oxidation and isothermal experiments. Conditions of intimate catalyst-carbon contact (K/C molar ratio=1/50) and high NO2 concentrations (1%) to improve the reactivity of the carbon reactive sites were unable to meet the steady state gasification rate needed for particulate filter regeneration for a modern diesel engine at 200°C. Oxygen-free thermal annealing (>500°C) caused reactivity losses of a maximum of 40% that correspond to changes to surface morphology and/or concentration of oxygen-containing functional groups. TP-ToFSIMS identified surface functional group changes with temperature on non-dosed and NOX pre-dosed (1.5%NO, 1%NO2, 4.5%O2, balance helium) diesel soot and sucrose char. Detailed analysis of the NOX dosed sucrose char spectra using both inspection and principal component analysis techniques revealed that the 1200 ion fragments created could be reduced to five sets of ions that are chemically and kinetically distinct. These sets presumably represent surface functional groups on the carbon. For example, Set IV may represent carboxylic acid, lactone, or carboxylic anhydride functional groups. Based on these results a mechanism for the surface reaction of NO2 with carbon under vacuum conditions was postulated. At temperatures less than 200°C the ion fragments contain primarily carbon-NO2 type ions. As temperature increases between 200 and 400°C the ion fragments are primarily carbon-NO and carbon-N type fragments. At higher temperatures (>500°C) the surface is enriched with nitrogen containing functional groups. A surface reaction mechanism is proposed where NO2 is bonded to an armchair site and with increasing temperatures and molecular rearrangements the N is incorporated into the carbon ring. The initial surface composition of NOx containing functional groups changes within the area of relevance of low temperature soot regeneration (i.e. between 25° and 200°C). Further studies are needed to understand the effect of N-incorporation on carbon reactivity. No rate processes either in reactor studies or based on surface functional groups met the rate criteria for low temperature DPM oxidation.

Starting in March of 1999, the ASACA study has measured PM2.5 mass and composition using 24-hr integrated and continuous measurement techniques. The ASACA network has one rural (Fort Yargo) and three urban (Fort McPherson, South Dekalb, and Tucker) monitoring sites located in the metropolitan Atlanta area. Supplementary data from the SEARCH and STN monitoring networks is also used where applicable. Yearly-averaged TEOM measurements recorded violations of the annual PM2.5 NAAQS (>15 μg/m3) every year of the study, and the daily NAAQS (>65 μg/m3) was exceeded on five separate occasions. Seven-year PM2.5 averages for the sites ranged from 18.8 – 19.8 μg/m3. PCMs were employed to collect PM2.5 composition data, detect spatial variations of PM species, and compare results with the continuous mass measurements. From 2004 – 2005, approximately 28% of the mass was OC, 24% was sulfate, 10% was ammonium, 6% was nitrate, and 3% was EC. Lesser ions contribute less than 3% to the total PM2.5 mass. Spatial variation of the major species was minimal, especially for species formed from secondary processes. South Dekalb did exhibit elevated levels of EC compared to the other sites, most likely due to its proximity to an interstate heavily used by diesel vehicles. PCM averages were found to be less than the averaged TEOM data due to the presence of unidentified matter (UM). Depending on the season, UM can contribute as little as 5% and as much as 50+% of the total mass. Secondary organic aerosol (SOA) concentrations from 2004 – 2005 were predicted using the EC-tracer method. Peak SOA occurs in mid-summer, and winter concentrations are significant due to biomass burning increasing the estimated OC/EC ratios. PCM, TEOM, and aethalometer data was also subjected to seasonal, day-of-the-week, and diurnal temporal variations. Active photochemistry plays an important role, as most species exhibit higher concentrations during summer months. The lone exception was nitrate, whose peak occurs in winter. Daily-averaged PM2.5 concentrations tend to peak late in the work-week and reach their low point on Sundays. Morning and afternoon rush-hour spikes in one-hour averaged PM2.5 are visible most days.

To design a cyclone abatement system for particulate control, it is necessary to accurately estimate cyclone performance. In this cyclone study, new theoretical methods for computing travel distance, numbers of turns and cyclone pressure drop have been developed. The flow pattern and cyclone dimensions determine the travel distance in a cyclone. The number of turns was calculated based on this travel distance. The new theoretical analysis of cyclone pressure drop was tested against measured data at different inlet velocities and gave excellent agreement. The results show that cyclone pressure drop varies with the inlet velocity, but not with cyclone diameter. Particle motion in the cyclone outer vortex was analyzed to establish a force balance differential equation. Barth??s "static particle" theory, particle (with diameter of d50) collection probability is 50% when the forces acting on it are balanced, combined with the force balance equation was applied in the theoretical analyses for the models of cyclone cut-point and collection probability distribution in the cyclone outer vortex. Cyclone cut-points for different dusts were traced from measured cyclone overall collection efficiencies and the theoretical model for calculating cyclone overall efficiency. The cut-point correction models (K) for 1D3D and 2D2D cyclones were developed through regression fit from traced and theoretical cut-points. The regression results indicate that cut-points are more sensitive to mass median diameter (MMD) than to geometric standard deviation (GSD) of PSD. The theoretical overall efficiency model developed in this research can be used for cyclone total efficiency calculation with the corrected d50 and PSD. 1D3D and 2D2D cyclones were tested at Amarillo, Texas (an altitude of 1128 m / 3700 ft), to evaluate the effect of air density on cyclone performance. Two sets of inlet design velocities determined by the different air densities were used for the tests. Experimental results indicate that optimal cyclone design velocities, which are 16 m/s (3200 ft/min) for 1D3D cyclones and 15 m/s (3000 ft/min) for 2D2D cyclones, should be determined based on standard air density. It is important to consider the air density effect on cyclone performance in the design of cyclone abatement systems.

Cyclone separators are widely used in agricultural processing industries as air pollution abatement devices. The performance of cyclones is a function of the geometry of the cyclone, operating parameters, and the particle size distribution (PSD) of the entrained aerosol. Multiple models have been proposed to predict the performance of cyclones given different geometric proportions, but many of these models do not quantify changes in performance with changes in inlet velocity or cyclone diameter given fixed geometric proportions. The Texas A&M Cyclone Design (TCD) method is a simple method for designing cyclones based on an inlet design velocity. The TCD method specifies “ideal” inlet velocities of 975 ± 120 m/min (3200 ± 400 fpm) and 914 ± 120 m/min (3000 ± 400 fpm) for 1D3D and 2D2D cyclones, respectively. However, there is evidence that higher dust collection efficiencies may be obtained from cyclones using different inlet velocities than those specified as the “ideal” velocity. Furthermore, the TCD method assumes that cyclone performance is independent of cyclone diameter. The present research demonstrates that, for large particles, the collection efficiency of 15.24 cm (six inch) diameter 1D3D and 2D2D cyclones is similar for inlet velocities from 10.16 standard m/s (2000 fpm) up to the design velocity, with significantly lower pressure drop at lower inlet velocities, resulting in lower energy requirements. However, the performance of cyclones is a function of cyclone diameter. Using similar operating parameters, the collection efficiency of a 60.96 cm (24 inch) diameter 1D3D cyclone was significantly lower (&#945; = 0.05) than that of a 15.24 and a 30.48 cm (6 and 12 inch) diameter cyclone, and the collection efficiency of a 91.44 cm (36 inch) cyclone was significantly lower (&#945; = 0.05) than that of a 60.96 cm (24 inch) diameter cyclone. The results of this research suggests the need for a new mathematical model to predict the performance of cyclones.

A methodology was developed to dissect diesel particulate filters to study the time effect of loading for two different fuels: ULSD and a biodiesel blend. Filters loaded with soot from a diesel engine for exposure times of 1, 2, 5 and 10 hours were fractured and samples of filter substrates were analyzed using Raman spectroscopy and scanning electron microscopy. Observations revealed the sharp rise in pressure drop to be attributable to the clogging of the pores in the channel wall, leading to the formation of a pore-bridge. Cross sectional imaging of wall sections revealed this pore-bridge to be shallow, with significant particulate depositions limited to the first quarter of the depth of the filter walls. Images revealed increasingly dense deposits and the formation of coarse particles and soot cakes. Raman spectroscopy revealed no significant graphitization of the soot cake. The dissection methodology exhibits significant potential for future studies on DPFs.