EXPOSURE ANALYSIS -CHAPTER 5 pps

Personal monitors for measuring occupational exposure to volatile organic compounds VOCs often employed small pumps to pull air across an activated charcoal sorbent.. Granger Morgan at C

Lance A Wallace

U.S Environmental Protection Agency (ret.)

CONTENTS

5.1 Synopsis 99

5.2 Introduction 99

5.3 VOC Monitors 100

5.4 Pesticide Monitors 102

5.5 Carbon Monoxide Monitors 103

5.6 Particle Monitors 104

5.7 Newer Monitors for Other Pollutants/Parameters 107

5.8 Conclusion 109

References 110

5.1 SYNOPSIS

This chapter deals with personal air quality monitors, small devices people can carry with them through their daily activities that will sample the air they are breathing Personal monitors are the

“gold standard” for estimating human exposure to air pollutants In no other way can we measure the changing concentrations as persons go from their houses to their cars to their workplaces over the course of a day The development of personal monitors for measuring environmental pollutants began in the 1970s and continues today The chapter provides the seldom-told history of how the U.S Environmental Protection Agency (USEPA) responded to the National Academy of Sciences recommendation to “foster development” of personal monitors Early pioneers in developing mon-itors for volatile organic compounds (VOCs), pesticides, carbon monoxide (CO), and particles are recognized (Current technologies are also described, such as the new “multipollutant personal monitor” developed at Harvard.) Only brief attention is paid to the results of the large-scale studies using these monitors, because these findings are treated at greater length in accompanying chapters

on these pollutants But the monitors themselves are illustrated profusely

5.2 INTRODUCTION

Personal air quality monitors have a long history, perhaps beginning with the radiation badge invented early in the twentieth century to protect workers in radiological industries Until about the late 1960s, personal monitors were used mainly in occupational settings Development of the monitors was encouraged by legislation such as the Occupational Health and Safety Act of 1970 and the Coal Mine Health and Safety Act of 1969

Personal monitors for measuring occupational exposure to volatile organic compounds (VOCs) often employed small pumps to pull air across an activated charcoal sorbent The charcoal was then bathed in a liquid solvent (usually carbon disulfide) and analyzed by gas chromatography with

pollutants, including particles (Total Suspended Particles, or TSP) and several specific gases (CO, ozone, SO2, and NO2) The sixth criteria pollutant was hydrocarbons, or nonmethane hydrocarbons,

a non-specific mix of VOCs not including most chlorinated or oxygenated species The levels of concern for these pollutants were far below the occupational standards and therefore the occupa-tional personal monitors would not be suitable for making environmental measurements (However, the historical record is not clear on whether USEPA scientists gave much thought to personal monitors at all.)

A few years later, some researchers including M Granger Morgan at Carnegie Mellon Institute

of Technology and Leonard Hamilton at the Brookhaven National Laboratory (BNL) realized that technology was rapidly becoming available that would make personal monitors for environmental concentrations feasible They sponsored a workshop on personal monitors at BNL that reviewed recent improvements in the field (Morgan and Morris 1976)

At about the same time, the National Academy of Sciences was asked by Congress to review the scientific approach of the USEPA In 1977, a nine-volume report appeared Volume IV of this effort was entitled “Environmental Monitoring” (NAS 1977), and one of its major recommendations reads as follows

…the Agency is not currently developing personal air quality monitors…By equipping a controlled sample of people with these portable sampling devices, the concentration of the pollutants to which they are exposed could be measured….There is evidence to challenge a common objection that personal monitors are technologically impracticable (… Morgan and Morris 1976, Wallace 1977) Prototypes have been developed to detect several air pollutants at levels near ambient concentrations (Wallace 1977).

• We recommend that EPA coordinate and support a program to foster the development of small, quiet, accurate, and sensitive personal air pollution monitors for use in conjunction with other methods for measuring human exposure to ambient air quality.

Soon after this recommendation was published, USEPA sponsored a symposium on personal monitors (Mage and Wallace 1979) and set aside a small amount of resources ($250,000) to begin developing personal air quality monitors

5.3 VOC MONITORS

In the meantime, progress was being made on several other fronts Small quiet pumps capable of pumping air at 2–4 Lpm were developed Batteries capable of supplying adequate current to power the pumps for 24 hours or more were developed However, in the case of the VOCs, the single most important development was the serendipitous discovery of a new sorbent as part of an industrial process in Switzerland The sorbent was called Tenax, and had several advantages over activated

charcoal First, it could be thermally desorbed This had the advantage of not diluting the analytes first with a solvent before reconcentrating them Second, it gave up its sorbed agents more easily and reproducibly than activated charcoal Third, it was hydrophobic, which meant that water molecules could not take up all the active sites when monitoring under humid conditions And fourth, following thermal desorption, it could be reused, an important consideration, since it was quite expensive

Early exploiters of the new sorbent included Edo Pellizzari of the Research Triangle Institute, and Boguslaw Krotoszynski of IIT Research Institute in Chicago Pellizzari studied ambient air concentrations of VOCs in petrochemical production areas such as the Louisiana-Texas Gulf of Mexico area (Pellizzari 1977) Krotoszynski studied exhaled breath concentrations of nurses in Chicago (Krotoszynski, Gabriel, and O’Neill 1977; Krotoszynski, Bruneau, and O’Neill 1979) Thus, almost from the first uses of Tenax, the double usage for environmental (air) and body burden (breath) sampling was established

In 1979, a USEPA initiative committed the Agency to personal monitoring, beginning a program called the Total Exposure Assessment Methodology (TEAM) Studies The first TEAM study investigated the possibility of measuring personal exposure to four groups of pollutants: VOCs, particles, pesticides, and PAHs A full-scale 1-year pilot study on nine persons in New Jersey and three in the Research Triangle Park area of North Carolina ensued (Wallace et al 1982) At the end of this effort, however, only the VOCs were considered feasible targets with the methodology then available USEPA researchers reviewed progress and research needs in 1981–1982 (Wallace 1981; Wallace and Ott 1982)

The first TEAM study equipped 355 persons in New Jersey with Tenax monitors (Figure 5.1), measuring not only their daytime and overnight exposures to 25 target pollutants, but also their exhaled breath at the end of the 24-hour monitoring period (Wallace et al 1985) Repeat visits were made in different seasons over a period of 3 years (1981–1983) Smaller comparison studies

of 25 persons each were carried out in North Carolina and North Dakota Another large study took place in Los Angeles, Antioch, and Pittsburg, California in 1984, with 120 persons visited over two seasons (Pellizzari et al 1987a,b; Wallace 1987; Wallace et al 1988.) Several new target pollutants were added to bring the total up to 32; the single most prevalent VOC was d-limonene,

a natural constituent of citrus fruits and a popular fragrance (lemon scent) in cleaning products The TEAM study is now considered a watershed in environmental studies Not only was it the first large-scale study to focus on personal monitors for environmental measurements, it was also the first to use a probability design Such a design is the only accepted way to use a small group

to extrapolate to a larger population The 1979–1984 TEAM studies sampled about 500 persons but the total population represented was nearly 500,000 residents of New Jersey and California Finally, the additional focus on body burden (exhaled breath) measurements, allowed the discovery

of the overwhelming importance of smoking in exposing smokers to benzene and other aromatic compounds such as xylenes, ethylbenzene, and styrene (Wallace et al 1987)

The most surprising finding of the TEAM studies was undoubtedly the importance of indoor sources of VOCs, outweighing the outdoor contribution for essentially all of the compounds studied

In some cases, these indoor sources were easily identified: dry-cleaned clothes as a source of tetrachloroethylene; moth crystals and air fresheners as sources of paradichlorobenzene; volatiliza-tion of chloroform from chlorinated water in homes Persons visiting dry cleaners or gas stavolatiliza-tions

in the past week had significantly higher levels of tetrachloroethylene and benzene in their breath The basic findings of the TEAM studies have been replicated repeatedly in studies in Europe Large-scale studies of 300 homes in the Netherlands (Lebret et al 1986) and 500 homes in West Germany (Krause et al 1987) took place shortly after the TEAM studies Since then, the German national studies have included the entire country and more than 1,000 participants, with VOCs, pesticides, and metals as major targets (Seifert et al 2000)

Tenax, like all sorbents, collects only a fraction of the VOCs that are present in ambient or indoor

air It cannot measure very volatile organics (VVOCs) because they have a small breakthrough

volume, the volume of air that can be passed through the monitor before the target chemical begins

to emerge from the back end of the sampling cartridge For example, Tenax cannot measure vinyl chloride and methylene chloride if less volatile organics are also targets, and chloroform is at the edge of measurability Also, Tenax cannot measure many of the more reactive chemicals such as 1,3-butadiene and those containing oxygen

To counter these deficiencies, multisorbent sampling cartridges can be used (Hodgson, Girman, and Binenboym 1986) These commonly employ Tenax for the bulk of the target organics, but also include some form of activated carbon to pick up the very volatile and oxygenated chemicals that Tenax cannot collect Although these were used in the 100-building BASE (Building Assessment Survey and Evaluation) study as fixed indoor monitors, they have not been widely used as personal monitors (USEPA 2005)

5.4 PESTICIDE MONITORS

The next personal monitor to be developed employed polyurethane foam to collect vapor-phase semivolatile organics (SVOCs) including many common pesticides (Lewis and Macleod 1982) A new TEAM study of exposure to pesticides was carried out for 250 residents of Springfield, Massachusetts, and Jacksonville, Florida, (Immerman and Schaum 1990a,b) The same procedures including probability-based selection of participants and use of personal samplers were employed Once again the findings were unexpected, with indoor exposures even more dominant than in the case of the VOCs (Lewis et al 1988) Although five pesticides had been banned as of the time

of the study (DDT, aldrin, dieldrin, chlordane, heptachlor) they were all found in quantities such that they presented the highest risk of all the target pesticides Since DDT had been banned more than a decade before, this was evidence that these pesticides were stable for long periods of time

in the environment For an illustration of the pesticide monitor, see Chapter 15

FIGURE 5.1 Tenax cartridge and vest used in the TEAM VOC studies The vest has a Velcro flap to protect

the glass cartridge The pump is in the pocket of the vest The flow is 30 cc/min for a nominal 12-hour sample.

5.5 CARBON MONOXIDE MONITORS

A third personal monitor came on line in the mid-1980s: an electrochemical monitor for carbon monoxide (CO) Electrochemical monitors oxidize CO to CO2 and collect the free electrons created

— the current is proportional to the CO concentration For optimal use in a field study, it was important to develop a way to monitor not only the concentrations but also the locations where the person was This problem was solved by attaching a data logger and allowing the participant to punch in code numbers indicating the times he or she entered or left each different microenvironment encountered during the day (Ott et al 1988) (Figure 5.2) The program then calculated the average concentration associated with each microenvironment Also, an average was calculated for each hour At the end of the 24-hour exposure period, a breath sample was collected in a Tedlar® bag and analyzed by a single CO monitor in a central laboratory This third TEAM study was carried out in Washington, DC, and Denver, Colorado, on more than 1,200 persons

The study resulted in a rich database with scores of microenvironments ranked according to the CO exposures in each (Akland et al 1985) CO exposures were shown to depend on car travel,

as expected The breath samples were instrumental in identifying a problem with the CO monitors that had gone undetected; the weakening of the battery toward the end of the 24-hour period resulted

in lowering the measured exposures at that time Thus, the breath samples, together with the model relating breath levels to air levels, were used to adjust the readings of the monitors during the latter part of the monitoring period

Several electrochemical monitors are available for CO, but one, the Langan monitor, has certain special features of interest Many of the monitors only supply readouts to the nearest ppm, but CO levels have decreased so much in recent years that monitors capable of reading sub-ppm values are desirable The Langan has a feature allowing selection of a 10-times scale allowing readings down to 0.1 ppm Drawbacks include a sensitivity to temperature that must be corrected by a user-specified algorithm

FIGURE 5.2 CO monitor/data logger designed by Ott for use in a USEPA study of 1,200 residents of

Washington, DC, and Denver, Colorado The display contains an “activity” button that was pressed by the participant each time he changed microenvironments This caused the data logger to average the measurements since the last time the activity button was pressed By reading the time on the display, the participant could write down the location and the time so that each microenvironment was matched to the proper average.

On/Off Activity

Charger

Cap

Inlet

Flow Meter Scrubber

CO Cell

5.6 PARTICLE MONITORS

Meanwhile, the late 1970s and early 1980s saw the development of the Harvard Impactor (HI), and its use in another watershed study, the Harvard 6-City Study This study, which went on for

20 years, has provided invaluable data on indoor concentrations of particles and associated lung function declines in children and adults The HI is not a personal monitor, but has had a long useful life as an indoor and outdoor monitor (Figure 5.3) An impactor uses a collector plate to intercept particles too large to follow the sharply curved streamlines within the monitor The smaller particles deposit on a filter that has been preweighed, and is then weighed again to determine the particle concentration By contrast, the cyclone monitors widely used in occupational sampling use the walls of the monitor to collect particles that fail to follow helical streamlines within the monitor

An impactor has a sharper size cut (the dividing diameter separating the two particle groupings)

than a cyclone The size of particles is very important in determining the depth of penetration into the lung Therefore the USEPA standards have been set for particles of a specified size Over the years, the standards have evolved from total suspended particles (TSP) to inhalable particles (particles smaller than 10 µm, or PM10) to fine particles (those smaller than 2.5 µm, or PM2.5)

In the early days of the 6-City Study, a personal cyclone monitor, with a cut point at 3.5 µm was employed (Dockery and Spengler 1981) Later a new personal impactor was developed in conjunction with Virgil Marple of the University of Minnesota (Marple et al 1987) The Marple personal exposure monitor (PEM) employs a design that can provide a cut point of any desired magnitude by adjusting the airflow rate and by blocking off a specified number of the 10 holes in the inlet (The PEM is pictured in Figure 5.4.)

The PEM adjusted to sample PM10 was employed in the fourth and last TEAM study, the Particle TEAM (PTEAM) study in Riverside, California (Clayton et al 1993; Thomas et al 1993) One hundred seventy-eight Riverside residents wore the monitor for two 12-hour periods (day and night) in the fall of 1990 To ensure comparability between the personal and fixed samples, the PEM was pressed into service as a stationary indoor monitor (SIM) and stationary outdoor monitor (SAM) in each home

Results of the PTEAM study included discovery of the “personal cloud,” a widespread phe-nomenon in which personal exposures exceed the time-weighted average of indoor and outdoor concentrations (Özkaynak et al 1996a,b) Another important finding was the importance of cooking, especially frying, in contributing to indoor exposures Cooking appeared to be second only to smoking as a particle source

FIGURE 5.3 Harvard impactor (HI), also known as the Black Box Shown is a PM2.5 nozzle The HI is programmable using the timer to collect data automatically for specified periods up to 2 weeks.

The Ethyl Corporation sponsored a year-long study of personal exposure to manganese (Mn)

in Toronto, Canada, and also in Indianapolis, Indiana (Pellizzari et al 1999, 2001) Mn is produced

in auto exhaust when methylcyclopentadienyl manganese tricarbonyl (MMT) is used in gasoline

to improve engine performance MMT has been used in Canada for many years, and is also legal for use in the United States, although little market penetration has occurred due to the opposition

of environmental groups and also automobile manufacturers, who fear the “poisoning” of their onboard computer systems by Mn The Toronto study was made in part to meet USEPA desires for a full-scale study of personal exposure to Mn For the Toronto and Indianapolis studies, which were carried out by the Research Triangle Institute, a new personal monitor was developed with several interesting features It had a timer to turn the pump on and off in a 1 minute on, 2 minutes off sequence, allowing measurements to extend over 3 days but requiring battery power to be applied for only 24 hours There was also a device that kept track of physical activity by measuring electrical capacitance of a transducer kept on the person This was useful in establishing whether the monitor was actually being worn at the times and places the person reported

Time series studies in the 1990s established the importance of air pollution in apparently causing morbidity and mortality among sensitive subpopulations (Dockery, Schwartz, and Spengler 1992) However, although much evidence points toward fine particles, a fair amount of evidence suggests other agents, including CO, SO2, NO2, ozone, sulfates, nitrates, and coarse particles Clearly it would be desirable to collect information on personal exposure to as many of these “co-pollutants”

or possible confounders as possible Harvard University School of Public Health accepted this challenge and developed a multipollutant personal monitor (Figure 5.5a and Figure 5.5b) The monitor has a 5.2 Lpm airflow split four ways: 1.8 Lpm each through a PM10 and PM2.5 nozzle with a Teflon filter; 0.8 Lpm to collect nitrates or sulfates (depending on which is more prevalent

in the region of interest) through a denuder onto a sodium carbonate-coated glass fiber filter; and 0.8 Lpm through a PM2.5 nozzle onto a quartz filter that can be analyzed for organic carbon and elemental carbon (Demokritou et al 2001) A clever addition to the monitor’s range of pollutants was made by attaching two passive Ogawa badges to the elutriator portion of the monitor, assuring

a steady passage of air across the surface of the badges, which can be selected to measure ozone and either SO2 or NO2 (again, depending on the pollutant of interest in the region) The monitor was used in a series of panel studies in Atlanta, Boston, and Los Angeles sponsored by the USEPA and also (Los Angeles only) by the California Air Resources Board (Wallace, Williams, and Suggs 2004) It is now commercially available

FIGURE 5.4 MIE optical scattering monitor and Marple PEM The MIE uses a laser at 880 nm wavelength

to estimate real-time particle numbers in the range of about 0.1–10 µ m The PEM employs a Teflon filter that can be analyzed for a dozen or so elements using x-ray fluorescence after weighing.

In the 1990s, personal monitors employing optical scattering to detect and count particles became available The most widely used of these is the MIE personal DataRAM™, or pDR This passive monitor uses a laser tuned to 880 nm wavelength It is most sensitive to particles of about 0.6 µm diameter, but can detect particles as small as about 0.1 µm and as large as 10 µm Since the MIE is calibrated to a road dust aerosol of specific gravity 2.6, it overestimates the concentration

of ambient particles, which typically have a specific gravity of about 1.5–1.7 This means it is an uncertain indicator of the exact mass of the particles of interest; however, the great advantage of having nearly continuous measurements is the ability to determine short-term peaks and identify sources more easily than with a 24-hour integrated sample The MIE has been used in many studies (Wallace et al 2003), including the EPA-supported University of Washington study of four Seattle cohorts (Liu et al 2002, 2003; Allen et al 2003, 2004) and another EPA-sponsored study in Research Triangle Park, North Carolina (Wallace et al 2006) In the latter study, participants carried both the gravimetric Marple PEM and the optical-based MIE pDR, allowing a comparison

of their respective responses to PM (Both monitors are pictured in Figure 5.4.) All these studies used the MIE to identify short-term peaks due to indoor sources When these peaks were then subtracted from the record, the remaining concentrations were assumed to be due to outdoor particles penetrating the home This separation of indoor particles into those from outdoors and those from indoors is desired by epidemiologists so that they can determine the correlation between

(a)

(b)

FIGURE 5.5 Harvard multipollutant monitor, as worn (a) and with parts identified (b) The pump (not shown)

has a 5.2 Lpm capacity, with 1.8 Lpm flowing through each of the impactors at the top (PM2.5 and PM10) 0.8 Lpm flows through the quartz filter for later analysis of elemental carbon (EC) and organic carbon (OC) The final 0.8 Lpm flows through a sodium carbonate–coated tube to a filter for collection of nitrates or sulfates Two passive Ogawa badges can be mounted on the vertical elutriator to measure ozone, NO2, or SO2.

PM 2.5 and PM 10 PEMs

EC/OC Mini-Sampler

exposure to particles of outdoor origin and the measurements of those particles at central sites Drawbacks of the MIE include a sensitivity to high humidity, which is not important for use as a personal monitor, but requires adjustments when used as an outdoor monitor Another drawback

is a tendency for a zero drift, which is correctable if the drift is positive, but because the manufacturer suppresses negative readings, is not completely correctable if the drift is negative Other optical scattering devices are also available, including the Radiance Research instrument and the Grimm monitor (Figure 5.6) The Grimm has the advantage of including a filter along with the optical scattering optics to collect particles for later weighing to compare with the optical scattering estimates It also measures eight size categories simultaneously

Although not strictly a personal monitor, the Piezobalance can be mentioned here due to its importance in early studies of environmental tobacco smoke (ETS) (Repace and Lowrey 1980) The Piezobalance employs a piezoelectric crystal electrically driven to oscillate at a certain fre-quency (Figure 5.7) Particles depositing on the crystal change the frequency The change is detected

by comparing it to another crystal isolated from the airflow The particles are induced to deposit

on the crystal by an electrostatic precipitator The instrument weighs about 20 pounds, so it is too large to qualify as a personal monitor, but it is portable and quiet and has been employed in many publicly accessible places such as restaurants, bars, and even churches Because the response of the crystal is not linear beyond a certain threshold, the Piezobalance requires frequent cleaning, up

to every half hour in a high-concentration aerosol, a requirement that limits its usefulness

5.7 NEWER MONITORS FOR OTHER POLLUTANTS/PARAMETERS

In recent years, due to intense interest in particle exposures related to the finding of increased morbidity and mortality, a number of new approaches to particle measurement have been instituted

It remains to be seen which of these will find widespread use in field studies Following is a brief survey of these newer monitors

A diffusion charger, which responds to the active surface area rather than the mass or the number of particles, is commercially available (Figure 5.8) (The active surface area is an important

FIGURE 5.6 The Grimm 1.108 laser particle counter provides 1-minute particle counts in 14 size ranges In

this photograph, the display shows a count of 42,360 particles/L in the sizes above 0.3 micrometers and 40 particles/L above 2 micrometers The instrument stores its readings in a memory card located underneath the display window, and the time series of 1-minute counts for 14 sizes can be downloaded by cable into a personal computer.

feature to measure, since it is surface chemistry that may determine the toxicity of many particles.) The diffusion charger uses a corona discharge to charge all the particles in the sensing volume A sensitive electrometer measures the resultant current Since the number of charges attaching to the particle depends on surface area, it gives a nearly direct measure of surface area Diffusion chargers have been used to measure exposures, in conjunction with health measures, of children with asthma Polyaromatic hydrocarbons (PAHs) include a number of carcinogenic agents such as benzo-a-pyrene Since PAHs have a wide range of volatilities, some exist as gases and others are attached

to particle surfaces at normal temperatures A personal monitor that responds to these

particle-bound PAHs is the EcoChem PAS CE2000 (Figure 5.9) The PAS 2000 series employs UV emission from a krypton element at an energy of 5.6 eV to ionize PAH residing on particle surfaces (Since some non-PAH compounds, most importantly black carbon, are also ionizable at this energy, there can be a problem with interferences.) The electrons emitted by the ionized PAHs form a current detectable by the instrument The analyzer signal is a combined measure of total PAH adsorbed

on carbon particles and the underlying black carbon itself A second major drawback of this instrument is the inability to calibrate it in the field Thus it must be considered semiquantitative

FIGURE 5.7 Kanomax 8510 piezoelectric microbalance (Piezobalance) for measuring particle mass

concen-tration at 1-minute or 2-minute averaging times In this photograph, the display reads “2997” momentarily to show the user the initial starting frequency of the piezoelectric crystal An internal size impactor is installed inside the instrument to allow it to measure either PM2.5 or PM3.5 respirable suspended particulates (RSP) mass concentrations.

FIGURE 5.8 Diffusion charging (DC) monitor The DC monitor provides a measure of the surface area of

particles less than 1 micrometer in diameter.