1

Georgia Institute of TechnologySchool of Earth and Atmospheric Sciences

EAS 4641Spring 2007

Lab 5

Particle Number and Mass in an Urban Environment

Purpose of Lab 5:In this lab we will characterize the concentration of particles by number and mass inan urban environment. We will investigate how number and mass vary diurnally andwith respect to each other to gain some insight into particle sources andmeteorological processes that influence concentrations. In addition, we will collect afilter during the period of the number and mass measurements for analysis of aerosolchemical ionic composition in Lab 6.

Introduction. Due to the progression in the understanding of environmental and healthimpacts of air pollutants, air quality regulation of pollutants has focused on gases.However, aerosol particles are the cause of visibility degradation and may be moredamaging to health than gases. Particles are either formed in the atmosphere (calledsecondary particles) by nucleation or directly injected (primary particles). These particleschange as they age. Number concentrations and mass change due to coagulation,condensation of vapors onto particles, chemical reactions on particles, uptake of water,wash out by precipitation, sedimentation, and dilution during transport. Particle size,number, mass, shape, and chemical composition all influence how effectively particlesscatter and absorb light, limiting visibility (haze) and altering the global energy balance.These factors also influence how particles take up water, which in turn affects cloudformation, precipitation, and the hydrologic cycle in general. The toxicity of particles,when inhaled, also is a function of all these properties, although much remains uncertain.

Particle Size Distribution, Number vs Mass. Atmospheric particles span a wide sizerange, from a few nanometers (10-9 m) to tens of microns (we will use the symbol µ, 10-6

m). Figure 1 shows the size of particles of various composition, and compares this size togases and the wavelength of radiation. The way particle number, surface area, mass, orany other parameter varies with particle size is represented as a size distribution function.Figure 2 shows the size distribution function for particle number, surface area, andvolume (or mass). Because particles span a large size range plots of the size distributionfunction are usually made on a semi-log scale, with the log of particle size. To preservethe proportions in this type of plot, the size distribution function is in the form ofDX/DlogD, where, D is the particle diameter and X is the property of interest (e.g.,number, surface area, or mass concentration). The area under the curve of a plot in

2

Figure 2 is equal to the total number, surface area, or volume of all particles in that sizerange). This means that in Figure 2, in terms of number, most particles have sizesbetween 0.001 and 0.1 µm diameter.However, the particles that contributethe most to the total mass or volume,have sizes between 0.1 and 10 µm orso. These plots show that particles arepreferentially found in various sizeranges, or modes. This is a result ofthe various sources of particles, andhow they are processed in theatmosphere.

Sources of Particles. The sources ofparticles have not been completelyestablished. Most of the insight intosources comes from measurements ofparticle chemical composition.However, that is complicated by thefact that the ambient aerosol is verychemically complex, especially theorganic carbon fraction. This happensbecause not only are manycompounds emitted, many chemical

Figure 1. Characteristics of ambient and industrial particles

Figure 2. Number, surface area, and volumedistribution for a typical urban aerosol.

3

compounds are added to the particles by chemical processes that occur in the atmosphere.In a broad sense the main sources of particles are:

1. Sea-spray emissions (primary)2. Soil dust emissions (primary)3. Volcanic emissions (primary and secondary)4. Biomass burning (primary and secondary)5. Fossil-fuel combustion for transportation and energy (primary and secondary)6. Fossil-fuel combustion for industrial processes (primary and secondary)7. Biogenic emissions (primary and secondary)

Various sources can dominate, depending on location. In urban settings in the easternU.S. the main source of particles is 5). Coal-burning power plants lead to high levels ofSO2, which produces sulfate (SO4

2-) aerosol. Burning of fossil fuels in cars and trucksleads to organic (OC) and element carbon particles (EC). Volatilization of organicsolvents and natural hydrocarbons also produce organic particles. The Atlanta summerfine (PM2.5) aerosol is mainly from these sources. In the winter, domestic wood burningcan also be a significant source.

Atmospheric Processing: Coagulation between particles, condensation of vapors, andthe loss mechanisms such as precipitation scavenging and dry deposition combined withthe various sources leads to the multi-modal nature of ambient size distributions. Figure3 provides a summary.Note that in both Figure2 and Figure 3, there is aclear separation betweent w o m o d e s a tapproximately 1 to 3 µmdiameter. This is mostobvious in the volumeor mass distribution.Particles smaller thanthis minimum are calledfine particles, and thoselarger are coarsepar t i c les . Th isdelineation is mainly theresult of fundamentallydifferent sources for thef ine and coarseparticles. As Figure 3shows, fine particles aremainly from condensedgases or particlesemitted directly duringcombustion. Coarseparticles on the otherhand are mainly thrown

Figure 3, Causes for a multi-modal ambient aerosol.

4

into the atmosphere by some mechanical process. This distinction between fine andcoarse is important for a number of reasons, one being how these particles influencehuman health.

Health Effects. It has been established that inhaled aerosol particles are detrimental tohealth. Particles can contain hazardous inorganic and organic substances, such as heavymetals, benzene, polychlorinated biphenyls, and polycyclic aromatic hydrocarbons(PAHs). However, it has also been recognized that particles smaller than 10µm (PM10)are often correlated with asthma and chronic obstructive pulmonary disease. Recentstudies have shown that fine particles (PM2.5) result in more respiratory illness andpremature death than do larger particles. Fine particles are believed to be more toxic dueto their ability to penetrate further into the lung (recall lab 4 on how particles penetratethrough tubes), and are thus much harder to expel. One study (Dockery et al., 1993) ofsix-cities over 16 years found that people living in areas with higher aerosolconcentrations had a lifespan two years less than those living in cleaner areas. Airpollution was correlated with lung cancer and cardiopulmonary disease. Mortality wascorrelated with fine particles, including sulfates. A number of studies suggest that thereis no lower threshold in which exposure to particles becomes safe. These studies focusedon long-term exposure (chronic). For example, long-term exposure to 5 µg/m3 ofparticles smaller than 2.5 µm (PM2.5) above background levels resulted in variouscardiopulmonary health outcomes that included mortality and decreased lung function inadults and children. Short-term (acute) exposure can also be hazardous. One study (Pope2000) found that short term exposure to an increase in PM10 particle mass of 10 µg/m3

resulted in a 0.5 to 1.5% increase in daily mortality, higher hospitalization and health-care visits for respiratory and cardiovascular disease, and enhanced outbreaks of asthmaand coughing. These results usually occur 1-5 days following exposure. Some studiessuggest that exposure to particles plus gaseous pollutants (e.g., ozone) may havesynergistic adverse affects.

To protect human health the U.S. EPA has set standards for PM10 and PM2.5.

Annual Mean 24-Hr AveragePM 10 50 µg/m3 150 µg/m3PM 2.5 15 µg/m3 65 µg/m3

Table 1. National ambient air quality standards for PM. (check out EPA web sitewww.epa.gov, or http://epa.gov/air/criteria.html)

Although air quality standards are based on aerosol mass concentrations, recent researchsuggests that other aerosol characteristics may be more appropriate when attempting toestablish a link with human health. For example, so-called ultrafine particles (diametersless than 0.1 µm) may be especially toxic. Figure 2 shows that ultrafine particlescontribute little to PM2.5 (or PM10) mass, but account for most of the numberconcentration. New studies have indicated that fine particles, especially those frommobile sources (e.g., cars, trucks) are hazardous. If interested in aerosol particle healtheffects see: Kaiser, J., News Focus: Mounting evidence indicts fine-particle pollution,Science, 307, 1858-1861, 2005.

5

Particle Measurement Methods.Number Concentrations. One common method used to measure number concentrationsof all ambient particles is the Condensation Particle Counter (CPC, also calledCondensation Nuclei Counter, CNC). Although a variety of CPC exist, they all work ona c o m m o nprinciple.Particles areexposed to asupersaturatedvapor (typicallyan a lcohol) ,which condenseson the particlescausing them togrow to muchlarger s izes,(typically on theorder of 1µm orso). This is doneso that particlestoo small toscat ter l ightgrow to a sizewhere they canbe de tec t edoptically. Lightscat ter ing ismost intense forparticles near thewavelength oflight 0.5 µm, andparticles smallert h a n a b o u t0.1µm diametercan not bedetectedoptically. In acontinuous flowCPC (the most commonly used, and what we will use in this lab, see Figure 4), the largedroplets and accompanying air pass through a nozzle to make a narrow jet. Individualparticles are counted by an optical device, which consists of a laser beam and optics tocollect scattered light. The jet intercepts a laser beam, and each particle that scatters thelaser light is sensed by a photo-detector that converts electromagnetic radiation to avoltage pulse. Pulses are counter by an electronic counter and summed over some period

Figure 4, Schematic of the TSI 3007 CPC

6

of time. Number concentration can be determined (number of particles/volume of air)from the volumetric rate at which air is sampled by the CPC.

Figure 4, shows a schematic of a continuous flow CPC. Continuous flow CPCsconsist of three components, a saturator, condenser, and an optics assembly. Thesaturator and condenser consist of cylindrical tubs. In the saturator, the tube is lined witha wick that soaks up alcohol from a reservoir. The tube is typically drilled out from ablock, which is held at some elevated temperature. The ambient air, containing particles,enters the condenser and becomes saturated with alcohol (saturation ratio S =1). Fromthere, the air moves to the condenser, which is also a tube bored in a block, but in thiscase it is cooled, typically by a thermo-electric device. This cools the alcohol-saturatedair resulting in supersaturated alcohol, S>1. Recall that saturation (or equilibrium) vaporpressure (psat) decreases with decreasing temperature, Clausius-Clapeyron), and that Sequals the ambient vapor partial pressure (p, the amount of vapor in the gas), divided bythe saturation vapor pressure, S= p/psat. As the air travels through the CPC from thecondenser to saturator, p remains fairly constant, but psat decreases, causing S to increaseto values larger than 1.

CPCs detect (count) all particles that are “activated” by the alcohol (i.e., are of sufficientsize to grow by alcohol condensation). Particles not activated, do not increase in size,and are not detected. Activation depends on particle size and is predicted by the Kelvinequation.

where S is the alcohol saturation ratio in the condenser, s, M , and r are the surfacetension, molecular weight, and density of the alcohol, R is the universal gas constant, Tthe temperature, and Dp* the size of particle activated. Thus, as S gets larger (S must belarger than 1) Dp* decreases, and for a given S there is a minimum size that will bedetected (Dp*) for a given CPC geometry, T settings, and working fluid. For ourinstrument (TSI 3007 or 3022), Dp* is roughly 0.01 µm diameter (10 nm), thus theinstrument will count all particles larger than 0.01 µm.

One problem associated with CPCs is that there are limits on how high a concentrationcan be measured. This occurs because as number concentration increases, the likelihoodthat more than one particle will be in the laser beam at a single time increases. When thisoccurs the CPC assumes it was only one particle, since only one pulse is produced. Thiswill result in an under measurement of the concentration. The TSI 3007 has an upperlimit of 105 particles/cm3

Mass Filters and TEOMSConceptually, the measurement of aerosol particle mass concentration is straight forward.One precisely weighs a new (e.g., clean) particle filter, draws ambient air through it forsome period of time (typically 24 hours), than reweighs it. The increase in filter mass isassumed to be due to the collected ambient particles, and the concentration is just the

Dp* =4s M

r RT ln(S)

7

increase in mass divided by the volume of air drawn through the filter. In practice, this isactually a difficult experiment. First, a very sensitive balance is required, LOD of about0.00001 g. Secondly, care must be taken to remove any interference from water absorbedor lost from the filter. Filters are generally placed in a T and RH controlled environmentfor weeks (~3) prior to the initial weight measurement, and then again after loading withambient particles prior to the final weight measurement. The idea is that the amount ofwater on the filter will be the same for both weight measurements.

Filter measurements of mass typically require long sample integration times to obtain asufficient collected mass for an accurate measurement. These long integration timesseverely limit investigating aerosol mass sources, spatial distributions, and establishingrelationships with meteorological parameters. Recently, a Tapered Element OscillatingMicrobalance (TEOM) was developed to measure particle mass on line. This instrumenthas been widely deployed throughout the globe at air quality monitoring networks.

The TEOM operates by drawing sample air through a filter at a constant rate, weighingthe filter and calculating mass in near real-time. In an attempt to limit absorbed waterinterferences, the inlet is heated to 50°C (note this causes other problems, such as loss ofsemi-volatile aerosol compounds), and a hydrophobic filter media is employed (Tefloncoated borosilicate glass). Sample air is drawn into the instrument at 16.7 L/min througha PM2.5 or PM10 inlet. 3 L/min is drawn isokinetically from this flow and conducted tothe filter situated in the mass transducer. The increase in the filter’s weight, which ismeasured automatically by the instrument, is used to calculate the total mass of collectedparticulates. The innovation in this instrument is the mass sensor, which consists of a“tapered element”; a hollow tube with one end fixed to a piezo-electric oscillator, and theother end free to vibrate and to which a small filter is attached. The tapered elementvibrates at its natural frequency. An electronic control circuit senses the vibration andmaintains a constant amplitude (i.e., offsetting losses) through a feed back loop. Thefrequency of oscillation is also precisely measured.

The tapered element can be viewed as a hollow cantilever beam vibrating as a simplemass-spring system in which the natural frequency follows the equation:

F = (K/M)1/2 ,

where F is the frequency in radians/sec, K the spring rate (related to spring constant), andM the mass. For more details see the TEOM manual.

8

Experiment No 5: Particle Number and Mass Concentration

The following experiment will determine the concentrations of ambient particles throughonline measurements of total number and PM2.5 mass. A ~24 hr integrated filter willalso be collected for later IC analysis (in the following this will be referred to asFilter/IC). We will also investigate how sources and meteorology influencesconcentrations.

Prior to each experiment, familiarize yourself with the instruments. Then proceed withthe following steps:

1. Start collecting data with the TEOM and CPC. The first test to ensure that theinstruments are operating correctly, go outside on the platform and place a lowpressure-drop HEPA filter on the inlet to the TEOM and CPC. This will test forleaks etc. If everything is operating properly concentrations will go to zero (Note:never plug the sample line by kinking the hose etc!!). Run the TEOM and CPC inthis way for about two hours. (While waiting, do steps 2, 3, and 4). After thistime remove the outside HEPA filters and let the instruments run, making surethat they are recording data (see step 5). Also make sure to record all proceduresin your lab notebook, including the times when the filters are inline on the TEOMand CPC and then removed. Check the computer clock(s) to make sure the timecorresponding to the measurements is recorded correctly.

2. Turn on the vacuum pump to the Filter/IC system. Although there is no filterinline in the filter holder, make sure the filter holder is screwed tight. Go outsideon the platform and measure the total flow rate for the Filter/IC by connecting the‘gas’ meter to the inlet of the cyclone outside. Do not measure the CPC or TEOMflow rate, it is 0.3 L/min and 16.7 L/min, respectively. These instruments haveautomatic mass flow controllers that are factory calibrated.

3. Estimate the length and internal diameter of the TEOM, Filter/IC, and CPCsample lines. (The TEOM and Filter/IC system use a tube with outside diameterof 0.5 inch, and the CPC a tube with outside diameter of 0.25 inch, insidediameters are slightly smaller. The flow rate and rough estimates of tubing lengthand inside diameter will be needed for estimating particle sample line losses foreach system.

4. Turn the pump to the Filter/IC system off, load a Teflon filter into the filter holderwith clean tweezers. Screw the filter holder shut. Never touch the filter with yourhands. Remove the filter and place in a clean bottle – this will be your blank.Load another filter into the filter holder and reassemble the filter holder.

5. Start sampling: After the zero test of part 1 is completed and make the two filtersoutside from the TEOM and CPC inlets. Start the vacuum pump to the Filter/ICsystem and begin sampling. Record the time you start sampling and the vacuumpressure on the pump (to ensure the flow is critical). The CPC and TEOM shouldbe operating and recording data.

6. After about 24 hours, leave the CPC and TEOM on, but stop the Filter/IC systemvacuum pump, remove the now loaded filter and store in a labeled bottle. Give

9

the bottles with filter blank and loaded filter to the lab instructor. It will be storedin a freezer until analyzed in the next lab.

7. Download the TEOM (PM2.5 mass), CPC (particle number concentration), andmeteorological data files. (The following depends on how the data system wasoperated. It may not be applicable. CPC data acquisition program may onlyrecord time of measurement and number of counts. To convert the counts tonumber concentration divide counts by both flow rate (0.3 l/min) and sampleintegration time of 10 sec – this can also be determined from the sample timecolumn).

Data Analysis1. Check the quality of the data by plotting the time series of the measurements (that

is a graph of Number or Mass Concentration versus time of day – use time labelsthat make sense, not seconds). Check for extraneous spikes in concentrations. Ifjustified, remove spikes.

2. Consider all sampling and transport issues (inlets and tubing losses). Estimate ifthey are significant and result in a lower measurement than the true ambientconcentration.

3. Estimate uncertainties associated with the measurements.4. Plot the 24-hour variation of CN and PM2.5 (ie time series plot). Plot pertinent

meteorological parameters along with your number and mass concentrations.You may wish to plot, wind direction, wind speed, solar intensity (this will atleast tell you when the sun rises and sets), precipitation (if any) etc. Either plotmultiple variables on the same graph (use color) or make all plots with identicaltime-scales.

5. Explain the cause for any diurnal trends, or lack of trends, or any abrupt changesin concentrations. Keep in mind that we are not far from a major highway withtraffic that varies with time of day.

6. Discuss what you think is the major source of CN and PM2.5 mass, why, or whydo they not track each other (give reasons).

7. Compare your data to other measurements (if any) made in Atlanta andsurrounding region. For example, see (http://cure.eas.gatech.edu/ga_air/today-0/Atlanta.html and http://www.air.dnr.state.ga.us/amp/ ). Explain any differences.Compare the PM2.5 mass with the EPA standard, is it every exceeded.

Question:The CPC used in this lab has a lower detection limit of 10 nm. That is, the super-saturation of alcohol in the instrument is sufficient to activate (grow) a 10 nm diameterparticle. Given the properties of the alcohol used in the instrument, determine the alcoholsuper-saturation in this instrument. The temperature in the condenser (where activationoccurs) is 10°C.

Alcohol properties: surface tension = 22.25 dyn/cmDensity = 0.81 g/cm3

Molecular weight = 74.12 g/mole

10