Development of a monitor to quantify lead-based aerosols


Material Information

Development of a monitor to quantify lead-based aerosols
Physical Description:
xiv, 224 leaves : ill. ; 28 cm.
Vanderpool, Robert William, 1955-
Publication Date:


bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1991.
Includes bibliographical references (leaves 223-224).
Statement of Responsibility:
by Robert William Vanderpool.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001722546
notis - AJD5025
oclc - 25764291
System ID:

Full Text







This dissertation is dedicated to the memory of my parents.


I would first like to thank Ken Reed for his assistance in machining

several pieces of the experimental apparatus. His expertise is most


The useful discussions with John Applewhite concerning the capabilities

and limitations of the polarograph are appreciated.

I would also like to acknowledge the assistance of Tom Peters during the

field evaluation of the prototype monitor. The successful demonstration of the

instrument's capabilities would not have been possible without his effort.

I also wish to thank the members of my doctoral committee for their time

and effort required to serve in this capacity. Each member, in his own way, has

served as a source of inspiration during the course of my studies at the

University of Florida. I am particularly appreciative of Dr. Lundgren for his

guidance, encouragement, and confidence. He has provided me with

opportunities for laboratory, teaching, and field experience that contributed

greatly to my professional development.

Finally, I would like to thank my wife, P.J., for her patience and support

during the last three years.



ACKNOWLEDGEMENTS ...........................

LIST OF TABLES .................................

LIST OF FIGURES ................................

KEY TO SYMBOLS ...............................

ABSTRACT .....................................


1 INTRODUCTION .............................

Background .................................
Description of Battery Production Process ...........
Instrument Design Criteria .......................
Summary of Design Approach ....................

2 REVIEW OF UTERATURE ......................


Design Theory ...............................
Prediction of Impactor Performance ................
Calibration of the Particle Collection Section ..........


Introduction .................................
Description of Air Sampling System ................
Prediction of Particle Sampling and Transport
Efficiency ..................................




. .

. .. =..

. o. ,


. ..


Introduction .......................................
Design Criteria ................ ............. ........
Description of Polarographic Analysis ...........................
Analysis of Field Bulk Samples .................. ... .......
Description of Modifications to the Polarograph ...............




Introduction .........................................
Experimental Methods .................................
Experimental Results ...................................

9 SUMMARY AND RECOMMENDATIONS .....................









A IMPACTOR DESIGN DRAWINGS .......................... 155

B IMPACTOR CALIBRATION RESULTS ...................... 160


D LISTING OF PROGRAM ................................ 165


REFERENCE LIST ........................................ 223

BIOGRAPHICAL SKETCH .................................. 225


Table 3-1

Table 4-1

Table 4-2

Table 4-3

Table 4-4

Table 5-1

Table 5-2

Table 5-3

Predicted Operating Characteristics of Teflon
Impactor ...............................

Description of Module 1 Air Flow Components ......

Description of Module 2 Air Flow Components ......

Module 2 Electrical Component Listing and
Description ..............................

Module 2 Electrical Connector Pin Descriptions .....

Polarograph Response Using Copper Internal
Standard ................................

Comparison Tests Between Polarograph and GFAA ..

Component Listing and Description of RUN/STOP
Board ..................................

RUN/STOP Board Connector Pin Functions ........

Component Listing and Description of PCM3 Board ..

Description of Various PCM3 Channels ...........

Component Listing and Description of Decade
Resistance Board ..........................

Decade Resistance Board Connector Pin Functions ..

Component Listing and Description of Model 264A ...

Model 264A Connector Pin Descriptions ..........






























Table 5-11

Table 5-12

Table 6-1

Table 6-2

Table 6-3

Table 6-4

Table 8-1

Table 8-2

Table 8-3

Table 8-4

Table 8-5

Table 8-6

Table 8-7

Component Listing and Description of Digital
Board ......... .........................

Digital Logic Board Connector Pin Descriptions .....

Listing and Description of Liquid Handling
Components .............................

Listing and Description of Nitrogen Handling
Handling Components ......................

Module 1 Electrical Component Listing and
Description ..............................

Module 1 Electrical Connector Pin Descriptions .....

Analysis of Laboratory Lead Standards ...........

Analysis of Field Control Lead Standards ..........

Results of Collocated Sampling Performed at Plate
Offbearing Operation .......................

Results of Unattended Sampling Performed at Plate
Offbearing Operation .......................

Results of Collocated Sampling Performed at Paste
Drying Operation ..........................

Results of Unattended Sampling Performed at Paste
Drying Operation ..........................

Results of Collocated Sampling Performed at Plate
Stacking Operation .........................
















Figure 1-1

Figure 1-2

Figure 3-1

Figure 3-2

Figure 3-3

Figure 3-4

Figure 3-5

Figure 3-6

Figure 4-1

Figure 4-2

Figure 4-3

Figure 4-4

Fow chart of lead-acid battery production
process ................................

Summary of prototype lead-in-air monitor design
approach .. ...........................

Measured particle size distributions at a
battery plant .............................

Schematic of single-stage impactor used for
particle collection ............... ..........

Schematic of generation system used for production
of calibration aerosols greater than 0.5 micrometers
aerodynamic diameter ......................

Scanning electron photomicrograph of ammonium
fluorescein calibration aerosols produced by VOAG

Schematic of generation system used for production
of calibration aerosols less than 0.5 micrometers
aerodynamic diameter ......................

Measured particle size collection efficiency curves
for single stage impactor ....................

Schematic of Module 1 air flow system ...........

Schematic of Module 2 air flow system ...........

Wiring diagram of Module 2 electrical components ...

Measured performance of Module 2 mass flow
controller ................................


Figure 4-5

Figure 4-6

Figure 5-1

Figure 5-2

Figure 5-3

Figure 5-4

Figure 5-5

Figure 5-6

Figure 5-7

Figure 5-8

Figure 5-9

Figure 6-1

Figure 6-2

Figure 6-3

Figure 6-4

Theoretical particle transport efficiency through the
sampling probe's horizontal section .............

Theoretical particle transport efficiency through the
probe's 900 bend ..........................

Polarographic response to 40 ppm concentrations of
lead, cadmium, and copper in 0.1 N HNO3 .........

Typical polarographic response as a function of lead
concentration .............................

Polarographic response to 0-200 ppb lead
concentration using 200 ppb internal copper standard

Polarographic response to 0-8000 ppb lead
concentration using 200 ppb internal copper standard

Schematic of printed circuit board which provides
run/stop functions .........................

Wiring diagram of Keithley PCM3 power control module

Schematic of printed circuit board which provides
decade resistance ranging functions ............

Schematic of Model 264A showing recommended
locations for auxiliary circuit boards ..............

Schematic of printed circuit board which provides
logical switching of stir, purge, and drop
dispense/dislodge functions ...................

Schematic of Module 1 liquid handling system ......

Detail drawing of flow-through cell designed for the
polarograph ..............................

Schematic of Module 1 nitrogen handling system ....

Wiring diagram of Module 1 electrical components ...
















Photograph of lead-in-air monitor during tests
conducted at plate offbearing process ............

Photograph of Module 1 inlet section showing position
of collocated filter holder relative to Module 1
inlet ....................................

Figure 8-1

Figure 8-2




C particle concentration

C, slip correction factor

D particle diffusion coefficient

Dp0 impactor cutpoint

D, tube diameter

g gravitational constant

L tube length

M gas molecular weight

n number of impaction jets

P fractional penetration

P, stack pressure

Q volumetric flowrate

R bubble radius

STK particle Stokes number

T gas temperature

V gas velocity

Vb bubble velocity

V, gas velocity pressure

V, particle settling velocity

W jet width

ad coefficient of diffusional deposition

a, coefficient of inertial deposition

a, coefficient of gravitational deposition

p, gas density

p gas dynamic viscosity

T particle relaxation time


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Robert William Vanderpool

December 1991

Chairperson: Dale A. Lundgren
Major Department: Environmental Engineering Sciences

Manufacture of lead-acid batteries results in the advertent formation of

gas-borne particles containing lead. Since the adverse health effects of

significant lead exposure are well documented, the battery industry is required

to undergo periodic monitoring to demonstrate compliance with existing lead-in-

air regulations.

A prototype instrument has been developed to quantify lead-based

particulates on a near real-time basis. Aerosol samples can be obtained from

high velocity, flowing airstreams as well as from static airstreams. The

instrument thus has potential as an emissions monitor, a recirculation air

monitor, or a monitor of workplace environments. The prototype instrument

contains an air sampling and transport system, a particle collection section, and


a lead analysis section. Proper sequencing of the system's components and

archival of the experimental test results are provided by an IBM-compatible

computer interfaced with a data acquisition and control system. The prototype

monitor may be operated either in attended or automated sampling and

analysis mode. The details of the instrument design will be presented along

with results of its laboratory and field evaluation.




Industrial manufacturing processes often result in the formation of gas-

borne particles (aerosols) containing hazardous metals such as lead, cadmium,

and copper. Since the adverse environmental and occupational health effects

of these substances are well documented, periodic air monitoring of

manufacturing processes involving their use is required to demonstrate

compliance with existing federal and local regulations.

The lead acid battery industry, which accounts for 72% of all domestic

lead usage (Environmental Protection Agency, 1986), is a potential source of

significant lead-based aerosol production. Monitoring of lead emissions in the

battery industry typically requires a separate aerosol sampling and analysis

technique. First, a representative aerosol sample is obtained from the process

using approved sampling techniques. This procedure typically involves aerosol

collection through a high efficiency filter to remove particles from the airstream.

The collected aerosol sample must then be extracted from the collection

substrate and analyzed using an appropriate analytical technique. However, the

time delay between sample collection and analysis is typically on the order of

days to weeks. Moreover, improper handling and extraction of the collected

sample can result in significant errors in estimation of the total lead

concentration. Positive sample bias is a particular problem during

measurement of trace lead levels common during evaluation of occupational or

recirculation air monitoring. Finally, specially approved equipment, sampling

and analysis techniques, and trained personnel are necessary to achieve

accurate results. The need existed, therefore, for an instrument capable of

providing both aerosol sampling and analysis of lead-based particles on a near

real-time basis.

The advantages of an automated air sampling and analysis instrument for

lead measurement tests are obvious. Such a device could be used both for

occupational and emission measurements and would significantly reduce the

cost of each test. Such an instrument would report test results within minutes

of the air sampling. Finally, by proper design and construction of the

instrument, the sources of measurement error inherent to the current

methodology could be significantly reduced. Compliance with existing emission

and occupational regulations could be routinely demonstrated. This report

details the design, construction, calibration, and field evaluation of such an

instrument designed for use by the lead-acid battery industry.

Description of Battery Production Process

The design criteria for the lead-in-air monitor was based on careful review

of the various operations involved in the manufacture of lead-acid batteries. Of

particular importance was consideration of the concentration, size distribution,

and chemical composition of the lead-based aerosols produced at each of the

process points. Figure 1-1 is a flow chart of the various operations involved in

the battery production. Not shown in the figure are various paths involving the

reclaim and reuse of lead scrap which occurs at several of the process points.

Battery manufacture first begins with the production of lead monoxide

powder by grinding solid lead metal in the presence of air at elevated

temperatures. Lead oxide paste is then produced by combining the lead oxide

powder with water and dilute sulfuric acid along with small amounts of

proprietary materials.

Battery plate grids are cast from molten lead-based alloys which consist

primarily of lead but also contain various trace metals to improve the plate's

mechanical properties. Following a specified curing time, the grids are

transported to the pasting operation where they are pasted and flash dried prior

to being off-loaded from the assembly line. The pasted plates undergo another

curing process before use. Plate stacking then occurs which involves alternate

stacking of plates with paper separators. The stacked plates are then loaded

into plastic cases and the appropriate cells connected together during the cast-

on-strap operation. The case cover is then installed, cells filled with sulfuric

C5 0 | 5 M
tc s


- --



0 o

a t



CB 0

(1) r

acid, and the batteries electrically charged prior to the packaging and shipping


The characteristics of the lead-based aerosol produced at each step of

the production process depends primarily on the composition of the material

handled and the mechanism by which the aerosol is produced. The aerosol

produced during the grid casting operation tends to be relatively small in size

and may contain only 5% to 20% lead by mass. Lead mass concentrations of

aerosol produced at the grid casting operation typically meet allowable emission

limits and are therefore ducted to the environment without the need for any

control measures.

Conversely, production and handling of lead oxide powder, paste and

pasted plates tends to form dispersion-type aerosols relatively large in diameter

and contain 95% to 99% lead by mass. Aerosols from these processes also

tend to be of much higher lead mass concentrations and are therefore ducted

to control equipment (typically baghouses) prior to the airstream's exhaust to

the environment.

With the exception of the acid fill, charging and packaging operations, all

the steps in the battery production process are sources of elevated lead

exposure to the battery plant workers. These processes, therefore, all possess

ventilation systems designed to remove the majority of the lead-based aerosols

and transport them to the particulate control equipment. Since these airstreams

are ultimately exhausted to the environment, an equivalent supply of makeup air

is required to maintain a flow balance. Battery plants in colder climates have

incorporated recirculation air systems to reduce heating costs associated with

heating the makeup air to plant temperature. This recirculation air is typically

obtained from grid pasting, drying, plate offbearing, cast-on-strap, and plate

stacking operations and is cleaned through a system of two high efficiency

particulate air filters (HEPA) in series with one another. This clean recirculation

air is then supplied to various process points as a source of makeup air.

The Occupational Safety and Health Administration has approved use of

this type of recirculation air system but requires continuous monitoring of the

recirculation air to ensure that the airborne lead content not exceed the

regulated 50 micrograms/m3 permissible exposure limit concentration.

Experience has shown that if the double HEPA filter system is installed and

maintained correctly, concentrations much lower than 5 micrograms/m3 can be

routinely obtained.

It is primarily for the purpose of monitoring recirculation air quality that

the lead-in-air monitor was developed. Other potential applications to the

battery industry will be discussed.

Instrument Design Criteria

The development of a lead-in-air monitor must be based on the sampling

and analysis requirements specific to the lead-acid battery industry. First, the

instrument must be capable of obtaining representative aerosol samples from

both flowing airstreams (stack or duct sampling) as well as from static

airstreams (room sampling). Particle size distribution measurements have

shown that efficient particle collection must take place over a fairly wide range

of particle sizes.

Accurate quantitation of the particulate lead content is also essential.

The analytical technique chosen must be specific for elemental lead and

possess minimum physical and chemical interference from other particulate

and gaseous components. Analytical accuracy and repeatability must exist over

the measurement range of interest. The analytical technique must possess low

blank and background response for the lead measurement.

Sampling should be performed over a 10 to 20 minute time period.

Instantaneous measurements are neither technologically feasible nor particularly

desirable because variations in lead handling processes and activities result in

considerable concentration variability with time. Time averaged measurements,

therefore, are more appropriate indicators of average process emissions. In

conjunction with a 15 minute analysis time, a 15 minute air sampling period

would allow a complete sampling and analysis cycle to be repeated

approximately every 30 minutes.

The instrument must also be capable of quantifying lead levels over a

wide range of concentrations. These concentrations range from recirculation

system alarm levels of 10 micrograms/m3 to levels of greater than

1000 micrograms/m3 for emissions from paste mixing facilities and other three-

part operations.

Instrument reliability and ease of use was also a design consideration.

Cleaning, maintenance, and external calibration of the instrument should be

minimized. The working instrument must be at least semi-automated, fairly

portable, and operate on standard 110 VAC, 60 hz power sources.

Finally, it was desired that the instrument's development and construction

costs be kept to a minimum. Once a final instrument design had been

completed and its capability demonstrated, it was desired that duplication costs

be limited to approximately $10,000.

Summary of Design Approach

Development of an automated sampling instrument for the specific

quantitation of lead involved careful consideration of the design criteria outlined

in the previous section. Ideally, the developed instrument would satisfy all of

the listed criteria although certain design and operational tradeoffs could have

been made if necessary. Of primary concern, however, was the instrument's

ability to obtain representative particulate samples from particle-laden airstreams

and accurately quantify lead levels over a reasonable concentration range.

Design of the instrument can be conveniently divided into three separate

sections: the air sampling system, the particulate sampling system, and the lead

analysis system. The sampling system must efficiently capture particles over a

range of sizes and effectively deliver them to the lead analysis section.

Representative sampling in flowing airstreams requires that isokinetic conditions

be achieved (nozzle velocity equal to freestream velocity) through proper

combination of nozzle inlet area and sampling flowrate. It was envisioned that

the sampling system contain several nozzles of varying inlet areas in

conjunction with a 60 to 90 liter per minute (Ipm) capacity sampling pump

whose sampling rate could be controlled and monitored by the user. The range

of operating flowrates was selected to be 10 to 30 Ipm. Following consideration

of available particle collection techniques (filtration, electrostatics, etc.), inertial

impaction was chosen as the mechanism to be employed in the instrument's

particle collection section. Subsequent calibration of the impactor showed that

it would collect 95% to 100% of the lead-based particles present in the in-plant

processes of interest.

Design of the instrument's analytical section required careful

consideration of its specific operating requirements. Of primary concern was

the ability to accurately quantify collected lead in amounts which were expected

at the various in-plant operations. Considering a 15 minute sampling time and

a nominal 15 Ipm sampling flowrate, approximately 0.23 m3 of air would be

sampled. For expected concentrations ranging from 10 micrograms/m3 to

2000 micrograms/m3 at the various in-plant processes, lead deposits of

2.3 micrograms to 460 micrograms would result. If these deposits were placed

in liquid solution (50 ml, for example), lead concentrations of 46 micrograms per

liter (ppb) to 9 milligrams per liter (ppm) by mass would be expected.

A wide variety of analytical techniques was considered during design of

the instrument's analytical section. Most of these techniques were eliminated

from consideration either due to their high cost, level of complexity, lead

detection capability, or failure to lend themselves to convenient automation.

Ultimately, an electrochemical technique was chosen for the quantitation of lead

in liquid solution. A commercially available polarograph was purchased and

adapted for automation.

A schematic of the completed instrument is shown in Figure 1-2. Air flow

through the system's particle collection section is provided by a 90 Ipm capacity

air pump with flow controlled and monitored using a commercially available

mass flow controller. Deposition of the lead-based particles takes place

primarily by inertial impaction into 50 ml of 0.5M nitric acid which is supplied

from a 5 liter capacity reservoir. Following a sufficient sampling time, an aliquot

of the solution is transferred to a flow-through cell in the polarograph for lead

analysis. The lead-in-air concentration can then be calculated from the

sampling time, sampling flowrate, solution volume, and the measured liquid lead


Automated control of the system's components takes place using a data

acquisition and control system (Kiethley Instruments, Inc., Model 575) interfaced

with an IBM-compatible computer. Computer control not only allows for proper

sequencing of the various components but also allows for data reduction and

automatic storage and retrieval of the test results. The Keithley 575 data

acquisition and control system receives necessary power from the computer

and provides both digital and analog input/output capabilities. A Keithley Model

AMM1 analog measurement module provides eight separate analog input




'a -
5 &E
a a.

c ,


0 O
- o2



'5 E
CT~ ""
0 3.

channels with a range of 10V, an input rate of 50 khz, and 12 bit

measurement resolution. The Kiethley 575 contains two analog output channels

with similar capabilities. At power up, the Model 575 self-calibrates the analog

channels with an accuracy of 2%.

Digital control is provided by 32 digital input/output channels with signals

which are transistor-to-transistor (TTL) compatible. A separate 16 channel

module (Kiethley, Model PCM3) provides external AC or DC power control

using TTL activated solid-state relays.

In the 575 system, the user can access the various module channels

using Kiethley KDAC500/I software linked with a GW-BASIC programming

environment. Use of programming logic rather than hard-wired logic allowed

for design flexibility during development of the prototype system.


Design of the proposed lead-in-air monitor was based on a number of

important criteria. The instrument must capture airborne particles in a

representative manner and transfer them effectively to the instrument's lead

analysis section. The analytical technique chosen must be specific for

elemental lead and yield a response that is linearly proportional to mass

concentration. Analytical accuracy and repeatability must exist over the

measurement range of interest and possess minimal interference from the

collection substrate and other particle components.

During the last two decades, considerable progress has been made

towards development of direct reading instruments for aerosol measurement.

These devices include electrical aerosol analyzers, condensation nuclei

counters, aerodynamic particle sizers, optical sizing and counting devices, and

piezoelectric and beta gauge based devices. Note that each of these

techniques measure a different property of airborne particles and thus do not

provide comparable information. Of these devices, only piezoelectric and beta

gauge devices provide direct reading of collected aerosol mass. None of these

techniques provide information in terms of the chemical composition of the

measured aerosol. Typically, devices such as filters, impactors, or electrostatic

collectors are merely used to capture an aerosol for subsequent analysis by a

separate instrument.

Conversely, there exist a wide variety of analytical techniques available

for chemical quantitative analysis. Typically, these techniques require that the

aerosol sample be previously collected in a separate step and manually

transferred to the instrument's analytical section. Some sample pretreatment or

extraction is often required to obtain reliable test results.

Rather limited success has been achieved in integrating these separate

aerosol sampling and analysis techniques in a single, direct reading instrument.

Many of these analytical techniques required for trace level analysis are fairly

sophisticated and do not adapt well to the development of inexpensive, portable

instruments. These sophisticated techniques often require daily calibration and

frequent maintenance to ensure accurate test results. The need for automated

sample transfer and preparation also complicates the instrument design.

Several tape samplers are available which provide combined aerosol

sampling with mass determination. In these automated devices, air flow is

normally directed through a porous tape which acts as the collection medium

for the airborne particles. This technique thus provides an integrated sample

with results time-weighted for the sampling period. In the GCA Model RDM301

sampler, particle matter is deposited by impaction onto a mylar film coated with

an adhesive (Uoy, 1983). At the end of the sampling period, the collected

material is quantified by beta attenuation. For low energy beta radiation, the

attenuation is a function of mass density of the absorber and is largely

independent of chemical composition or the particle's physical properties.

Spagnola (1989) developed an automated sampling instrument based on this

analytical principle. The instrument uses electromechanical devices for

automatic positioning of membrane filters in the sample holder. At the end of

the sampling period, the filter is analyzed and returned to its original position in

the sample holder. Application of beta attenuation has also been applied to a

dichotomous sampler developed by Macius and Husar (1976).

A paper tape sampler was also modified to provide lead analysis by x-ray

fluorescence (Smith et al., 1986). The system samples air at 0 to 10 Ipm and

irradiates the collected aerosol mass with a 10 mCi cadmium 109 radioisotope.

A Nal scintillator in conjunction with a multi-channel analyzer provides counts

versus x-ray energy for the resulting x-rays. The count corresponding to the

x-ray energy characteristic of lead provides a measure of the collected lead

mass. The authors note that the presence of arsenic and zinc positively

interfere with the lead measurement.

Automated mass monitors have also been developed based on the

concept of piezoelectric sensing of changes in aerosol mass. The TSI Inc.

Model 5000 deposits particles onto a piezoelectric crystal by electrostatic

precipitation. The rate of change of the crystal's output frequency is directly

proportional to the deposited mass concentration. Recently, this mass sensing

technique was applied to the development of an instrument which uses a low-

frequency tapered element coupled to a filter collector (Wang, 1985).

Successful measurement of aerosol composition "in-situ" has been limited

and has generally not evolved beyond the laboratory. Arnold and Folan (1987)

devised an instrument which levitates a single airborne particle within an

electrostatic field and irradiates the particle with an infrared beam. The

instrument optically detects changes in particle size caused by evaporation of

the particle due to the heat induced by the infrared absorption. The measured

rate of change is proportional to aerosol mass. The instrument is capable of

detecting picogram quantities of material for both organic and inorganic

functional groups such as sulfate and carbonate (Allen and Palen, 1989).

The majority of analytical techniques which have been adapted for trace

aerosol chemical analysis are available only in an "off-line" mode and do not

adapt well to field use. Martinsson and Hansson (1988) have used particle

induced x-ray emission (PIXE) for the multielement analysis of aerosol

previously deposited onto thin aluminum foils. Deposits were stepwise heated

to several hundred degrees centigrade by applying an electric current through

the sample. An ion beam impinged on the sample during the process. Low

atomic weight element information (such as oxygen and carbon) was obtained

by particle elastic scattering analysis while elements with atomic numbers

greater than 14 were determined by PIXE analysis. The chemical composition

of the sample was inferred from noted vaporization temperatures and

stoichiometry. Laboratory tests found the technique useful for quantitation of

ammonium chloride, ammonium nitrate, and ammonium sulfate.


Xhoffer et al. (1989) used a transmission electron microscope to analyze

laboratory generated aerosols deposited onto carbon coated electron

microscope grids. The analytical technique of electron energy loss

spectroscopy was used to observe the change in kinetic energy as the electron

passes through the sample. This technique is generally useful only for low

atomic weight elements. The authors also noted that high background signals

can be associated with the substrate and that rapid deterioration of

environmental samples can occur. The technique was also found to possess

limited quantitative application since the particle's absorption of the incident

electron beam was a function of particle size.

Electron spectroscopy for chemical analysis (ESCA) was investigated by

Muller et al. (1987) as a means of characterizing sub-micrometer sodium sulfate

aerosol collected on various filter media. This technique used incident x-rays to

excite photoemission of core level electrons from the sample. The measured

energy of the emitted photoelectrons is characteristic of the emitting element's

oxidation state. ESCA provides only surface analysis and is thus of limited

quantitative analysis. ESCA signals can also be difficult to interpret

quantitatively. Incorporating a similar surface analytical technique, Fletcher

(1983) used a laser microprobe to vaporize and ionize a collected sample for

analysis by a time-of-flight mass spectrometer. Reported limitations of this

technique were similar to those experienced by Muller et al. using ESCA


Neutron activation analysis (NAA) has been employed to provide

elemental information of collected aerosol samples. In this technique, the

sample is irradiated with neutrons to produce radionuclides. The activities of

the radionuclide products are then measured and related to the quantities of the

elements present in the sample. Podzimek and Wojnar (1990) used NAA to

quantify masses of laboratory generated titanium oxide particles deposited on

polyester filters. In a related study, Ondov et al. (1990) used NAA to quantify

45 elements in rural sub-micrometer ambient aerosols collected by micro-orifice

impaction. Although this technique was found to be generally useful, the

substrate blank precluded analysis of several elements of interest.

X-ray powder diffraction analysis of collected aerosol deposits was

reported by Tanninen et al. (1985). Powder diffraction allows for direct

determination of crystalline compounds. The authors used the diffraction

technique to analyze fumes generated by metal arc welding of steel. The main

analytical difficulty encountered was low instrument output intensity due to the

small particles analyzed in conjunction with the small quantity of material


Both conventional flame and graphite furnace atomic absorption

spectroscopy have been widely used for the analysis of collected aerosol

samples. Davidson (1979) used graphite furnace techniques to characterize

size fractionated aerosol samples obtained from three national parks. Attempts

to incorporate atomic absorption analysis in an automated monitor have met

with limited success (Smith et al., 1986).


Design Theory

Accurate quantitation of lead-in-air concentrations requires efficient

capture of lead-based particles. If an aerosol was only collected with 10%

efficiency, the resulting lead analysis would obviously greatly underestimate the

actual lead-in-air concentration. This would, in turn, result in an underestimation

of the lead exposure hazard. This section describes the design and calibration

of a device capable of sampling and capturing lead-based particles with the

necessary high efficiency to ensure accurate test results.

There currently exist a variety of aerosol samplers and collectors which

offer a basis for the collector design. These devices include inertial devices

(impactors and cyclones), electrostatic precipitators, thermal precipitators,

diffusion devices, and filtration devices. Each of these devices relies on a

particular property of airborne particles to remove the particles from the

airstream in which they are suspended.

Fundamental to the design of any particulate collection section is

understanding the nature of the aerosol to be measured. In predicting

individual particle properties (settling velocity, diffusional behavior, etc) the most

important parameter is that of particle size. Since industrial aerosols are always

a collection of particles of varying sizes, it is understanding of the aerosol's size

distribution which is of importance.

Figure 3-1 presents particle size distributions measured at a battery plant

in the southeastern United States. The specific sampling locations represent

sources of aerosols which supply recirculation air. It is believed, therefore, this

size distribution can be assumed to be fairly representative of aerosols which

exist in recirculation air systems. Plotted in the figure for the various process

points are cumulative lead mass distributions versus aerodynamic particle

diameter as measured using a University of Washington cascade impactor. The

aerodynamic diameter is defined as the diameter of a spherical particle of unit

density which has the same settling velocity as the particle under consideration.

Use of an equivalent aerodynamic diameter accounts for the particle's size,

shape, and density.

Inspection of Figure 3-1 reveals that these measured size distributions

are fairly large with half of the lead mass associated with particles greater than

10 micrometers aerodynamic diameter. These results agree well with similar

measurements by Hodgkins et al. (1990) from two different battery

manufacturing plants. Note that only 1% to 5% of the lead mass is associated

with particles less than one micrometer aerodynamic diameter. A device

capable of collecting all particles greater than one micrometer, therefore, would

have an overall mass collection efficiency on this aerosol of 95% to 99%.



0 0o
-) C


E cm

0 E

S0 0 0
)) P O

o c C S -s
o o


41 01 I, p i rI ...

JQQuUDiGa UD1j_ ssei SSDMPt pDGl %

Based on a review of these distributions, both thermal and diffusion type

collectors were eliminated as candidates for the collector design. Although

these type collectors can capture submicron aerosols with high efficiency, their

capture efficiencies for particles greater than one micrometer are quite low

(Hinds, 1982).

Filtration probably represents the most common form of aerosol

collection in air sampling. Unlike thermal and diffusion type devices, samplers

which rely on filtration mechanisms possess high collection efficiencies over a

wide range of particle sizes. Aerosols encountered in the battery industry could

be expected to be removed by filtration with virtually 100% efficiency. Despite

these advantages, however, filtration was eliminated from design consideration

primarily for logistic reasons. Due to high particle-substrate attractive forces,

particles collected on high efficiency filters are extremely difficult to remove

efficiently even if wet chemical techniques are employed. Wetting of either

fibrous or porous membrane filters dramatically increases its pressure drop

making the filter difficult to reuse. The alternative to filter reuse is to replace it

after each sampling test. Automated replacement of the filter would have

involved design and cost complications which were considered to be

unacceptably high.

Electrostatic collection was also given consideration during the

instrument's design phase. In these devices, collection relies on the

electrostatic forces which can be exerted on charged particles in an applied

electric field. Similar to filtration samplers, electrostatic samplers possess high

collection efficiencies over a wide range of particle sizes. Use of electrostatic

collection would have involved use of a point-to-plane precipitator for deposition

of airborne particles to a liquid surface similar to the design approach adopted

during the design of the LEAP sampler (Andersen Samplers, Inc.).

This design approach was initially investigated during the instrument

development but was ultimately discarded for two primary reasons. First,

preliminary laboratory tests with a design prototype showed that the flow of

electrical current through the electrolyte resulted in the inadvertent plating out of

previously collected lead ions from solution. Loss of collected lead in solution

would have resulted in underestimation of the total lead-in-air concentration and

thus an underestimation of the lead exposure hazard. Use of electrostatics was

also eliminated due its greater complexity and anticipated maintenance

requirements in the corrosive nitric acid atmosphere of the collection region.

Ultimately, the use of inertial impaction was adopted for the final

instrument design. Inertial impactors have been widely used for aerosol

collection and sizing and can efficiently capture particles greater than one

micrometer diameter. Inertial impactors are relatively simple in construction,

contain no moving parts or electrical components, and require little


Impactors rely on the high inertial forces which can be established on

particles in a high velocity airstream. In a typical impactor stage, particle-laden

gas is accelerated through one or more nozzles and directed toward a solid

surface which forces the airstream to change its flow direction. Due to their

inertia, however, particles will be unable to exactly follow their respective

streamlines and will deviate from the airstream by a distance which is a function

of the particle's aerodynamic diameter and the jet velocity. Particles of sufficient

inertia will deviate a sufficient distance to strike the collection surface and be

removed from the airstream. Conventional impactor stages with high jet

velocities can remove particles of 0.3 micrometers aerodynamic diameter and


Impactor performance is typically described by its Dp, cutpoint which is

the particle size collected with 50% efficiency. Particles larger than the cutpoint

diameter are collected with much higher efficiency. Extensive theoretical and

practical impactor research has shown that the cutpoint can be predicted with

some confidence by

D r=| 9 r I n W3 rS()
Pso Vi 9 i4 pp Q (3-1)

where C, is the Cunningham's slip correction factor determined from the

particle size and the fluid properties, p is the gas dynamic viscosity, n is the

number of jets, W is the jet width, p is the particle density, and Q is the

volumetric flowrate. The Stokes value (STK) can be predicted from knowledge

of the jet Reynolds number, ratio of jet width to jet to plate distance, and ratio of

jet throat length to jet width.

Based on these accepted design criteria, an inertial collector was

designed and constructed for use as the instrument collection section.

Figure 3-2 is a schematic of this device showing key components. Complete

detail drawings of each of the unit's components are presented in Appendix A.

In most inertial impactors, the impaction jet is directed onto a preweighed

stationary surface previously coated with a viscous, nonvolatile grease to

minimize particle bounce and losses due to blowoff. Following the sampling

run, the impaction surface is reweighed to quantify the mass of the collected

aerosol deposit. If chemical analysis is to be performed, the aerosol must be

extracted from the grease in an efficient and representative manner.

The collector developed for the lead-in-air monitor does not impact

particles onto a solid surface but into 50 ml of 0.5 M HNO3. Use of the liquid

impaction substrate eliminates the need for greased substrates and provides for

direct transfer of the particles from the gaseous phase to the liquid phase. High

velocity impaction also provides for the liquid turbulence necessary for efficient

particle dissolution so that aqueous lead ions are formed. High liquid

turbulence also tends to homogenize the liquid sample so that only an aliquot of

the sample need be analyzed for accurate lead determination.

Aerosol Inlet

Liquid Inlet

^; ^
11 olo

o ??
/f y Yle
/ f /
/ /,

/ / / /
^ / /
01 0 0
' //

/ //
/ //

Liquid Outlet

Figure 3-2. Schematic of single-stage impactor used for particle collection.

Air Outlet


During the sampling run, particle laden air enters the impactor inlet at a

flowrate of 10 to 30 Ipm (depending on the sampling situation) and is

accelerated through a single stage impactor. As shown in Appendix A, the

impaction stage contains 20 circular jets of 0.026 inch diameter. Following its

impingement into the liquid, the airstream then enters an expansion section

which reduces the upward air velocity and thus reduces inadvertent liquid carry

over. Use of a small plug of glass fiber filter at the air outlet has shown to

further minimize the liquid carry over. At the conclusion of the air sampling

period, a liquid outlet provides for transfer of the solution to the instrument's

lead analysis section.

Prediction of Impactor Performance

Inspection of Equation 3-1 allows for the generalized prediction of

impactor performance as a function of sampling air flowrate and the critical

impactor dimensions. Table 3-1 presents a summary of these calculations

assuming standard temperature and pressure conditions for air. Also presented

are actual measured stage pressure drops as a function of the sampling

flowrate. Calculated cutpoints at flowrates of 10, 20, and 30 Ipm are 0.92, 0.63,

and 0.42 micrometers, respectively.

It should be emphasized that these calculations are based primarily on

theoretical and experimental evaluations of impaction onto a solid surface.

However, immersion of an impaction stage and impingement of an airstream

o '3 (
1r CJ 3

0. I CM
QE 00) 0
2 co5

T-v d
o oo
0 ,- COJ

00 0
It 9 c N
CMl lqt I-

co E






CL 0



> ^


into a liquid represents an entirely different flow situation. While impaction is

certainly the most dominant collection mechanism involved, the entire process

may more correctly be described as a bubbling phenomenon.

Unlike the extensive studies related to conventional inertial impaction,

there has been relatively limited research devoted to the absorption of aerosols

by bubbling. As a simplified model, some researchers have considered particle

deposition in air bubbles which freely rise through a stationary liquid. As a

further simplification, these studies only consider the motion of bubbles less

than 0.2 cm diameter which have been found to be approximately spherical in

shape. Larger bubbles become flattened as they rise through the liquid and

pulsations in the bubble shape can occur.

Deposition of particles with a buoyantly rising bubble results primarily due

to the relative motion of the bubble with respect to the liquid. Due to the shear

force which exists at the liquid-gas interface, a gas circulation within the bubble

is established and maintained. As a result of this circulation pattern, large

particles may be removed by inertial impaction to the liquid-gas interface.

Gravitational settling of large particles can also occur as well as removal of

submicron particles by molecular diffusion.

The first analytical and experimental treatment of this particle absorption

phenomenon within a freely rising bubble was performed by Davies and Taylor

(1950). This theory was extended by Fuchs (1964), who developed coefficients

of absorption (deposition efficiency per unit cm of bubble rise) for the separate

absorption mechanisms of inertial impaction, sedimentation, and diffusion.

Later, related experimental studies by Melandrie and Prodi (1971) confirmed the

validity of this approach.

The coefficient of absorption due to inertial deposition, a,, is related to

the bubbles upward velocity, Vb, the particle's relaxation time, T, and the radius

of the bubble, R, by

9 Vt
a 9 = 2 (3-2)

Because particle relaxation time increases with the square of the particle

diameter, this inertial mechanism is dominant for large particles.

Similarly, the absorption coefficient due to particle sedimentation, a,, is

expressed as

3 9gT
a 4R Vb (3-3)

For most bubble sizes of interest, inertial deposition is an order of magnitude

greater than that of sedimentation.

Finally, diffusional deposition efficiency per unit path length can be

calculated by accounting for the particle's diffusion coefficient, D

a = 1.8 VD
Vd-b VR (3-4)

Large bubbles (R > 1 cm) rising in a liquid tend to flatten and pulsate

such that their rate of rise is difficult to predict. However, smaller bubbles can

be assumed to be spherical and their upward velocity can be predicted by a

empirical relationship developed by Melandrie and Prodi (1971)

Vb = 3.05 v (3-5)

where the dimensions of Vb and R are m/sec and m, respectively.

For a known bubble size and particle diameter, the overall change in

particle concentration (C) per unit length of bubble rise can be calculated by

= ( as + a, + ad ) C (3-6)

For example, Jonas and Schulz (1991) used this relationship to calculate an

overall collection efficiency of 98.8% of one micrometer radius iron oxide

particles within 0.5 cm radius bubbles buoyantly rising through one meter of


Inspection of the absorption equation reveals that for a given particle

size, decreasing the bubble radius results in an increase in absorption

efficiency. The equation also indicates that for a given bubble size, the

dominant collection mechanism depends strongly on particle size. For large

particles (greater than 0.3 micrometers radius) which possess large relaxation

times, inertial and sedimentation forces dominate while diffusive absorption is

negligible. Conversely, particles less than 0.05 micrometers radius are removed

almost entirely by molecular diffusion. Thus, regardless of bubble diameter,

there exists an absorption minimum which exists for particles of approximately

0.2 to 0.4 micrometer diameter. The actual magnitude of the minimum depends

primarily on bubble radius.

Use of these theoretical considerations to exactly predict the

performance of the lead-in-air monitor's impactor proved to be difficult. This

simplified approach assumes bubbles of known size rising through a stable

liquid of constant height. Experimental verification of the theory by other

researchers has been limited and performed at relatively low flow rates using

uniform sized bubbles. However, the dynamics within the teflon impactor

represent a vastly different flow situation. Temporary replacement of the

impactor's opaque housing with a glass enclosure of similar dimensions allowed

for the direct observation of the bubbling phenomenon. This observation

revealed the limitations of the simplified theory to this particular collector. First,

it was evident that the bubbles produced were not uniform in size but existed

over a wide range of sizes. To apply the absorption theory to a distribution of

bubble sizes, the distribution itself must be known and the collection efficiency

equation integrated over the bubble size range. At these high flowrates, it was

also evident that the bubbles themselves experienced considerable oscillation

and occasional breakup due to the high liquid turbulence. Finally, it was

observed that the liquid itself underwent considerable turbulence which changes

the nature of the liquid-bubble interface and results in uncertainties in predicting

the path rise of the individual bubbles.

Due to this lack of reliable design theory, no exact predictions of the

impactor's actual performance were made. Preliminary review of conventional

impactor design theory and bubbling theory were used only as a basis for the

impactor design. For accurate evaluation of the impactor's actual performance,

extensive experimental efficiency tests were conducted with the impactor in the


Calibration of the Particle Collection Section

In evaluating the performance of the particle collector, it was understood

that the collection efficiency would depend on the unit's critical dimensions,

liquid solution volume, particle diameter, and the sampling flowrate. Since the

impactor's dimensions and the solution volume were fixed by design, the

particle size and sampling flowrate were normally the only two parameters

varied during the collection efficiency tests.

The particle collection characteristics of the impactor were evaluated in

the laboratory using monodisperse calibration aerosols of solid, spherical

ammonium fluorescein. These calibration aerosols were produced by two

separate generation systems depending upon the size of the particle required.

For the generation of calibration aerosols 0.5 micrometers aerodynamic

diameter and larger, a Thermal Systems Inc. Model 3050 vibrating orifice

aerosol generator (VOAG) was used. A schematic of the complete generation

and sampling system is presented in Figure 3-3. The VOAG operates on the

principle of the controlled mechanical breakup of a liquid jet into uniform

droplets of predictable size. Periodic breakup of the liquid jet was achieved by

application of an AC signal at a known frequency to a piezoelectric crystal

which housed the liquid orifice. Since the liquid consisted of a known quantity

of a non-volatile solute dissolved in a volatile solvent, the generated droplets

dried to form uniform particles of predictable size. For the calibration of the

impactor, the liquid solution was prepared by dissolving a fluorescein powder in

aqueous ammonia. After their production, the droplets dried to form spherical

particles of solid ammonium fluorescein. The particles were then mixed with

clean dilution air and charge neutralized by exposure to a "Kr radioactive

source. The particles were then mixed with additional dilution air and sampled

by the impactor for the calibration tests. An optical microscope was used to

verify the size and quality of the generated particles collected on a fibrous or

membrane filter before each test series. Figure 3-4 is a scanning electron

photomicrograph of a monodisperse aerosol generated by this technique.

For calculating the particle's aerodynamic diameter from its physical

diameter, an ammonium fluorescein particle density of 1.35 g/cm3 was

assumed. This value was based on density measurements by Stober and

Flachsbart (1973), John and Wall (1983), Vanderpool and Rubow (1988), and

Chen and Crow (1986).










0 E








Figure 3-4. Scanning electron photomicrograph of ammonium fluorescein
calibration aerosols produced by VOAG.

Calibration aerosols less than 0.5 micrometers aerodynamic diameter

were generated using a Thermal Systems Inc. Model 3071 electrostatic

classifier. Figure 3-5 is a schematic of this generation setup. Polydisperse

droplets were first generated by air nebulization of fluorescein solutions through

a Collison nebulizer at an operating pressure of 20 psig. The polydisperse

droplets were then dried in a diffusion drier to form a polydisperse aerosol of

solid ammonium fluorescein particles. This aerosol was then passed through a

single stage impactor to reduce the large particle concentration. The pressure

drop across the impactor stage was monitored to verify the magnitude and

stability of the aerosol flowrate. Monodisperse particles of the desired size were

then produced by electrical classification within the classifier. Lastly, the

monodisperse aerosol was diluted to the necessary sampling flowrate using

clean, dry dilution air and sampled by the impactor.

The air exiting the impactor was passed through a filter holder housing a

47 mm diameter glass fiber filter which was used to collect any airborne

particles not collected by the impactor. Flowrate through the sampling system

was provided by a 90 Ipm capacity vacuum pump fitted with a primary valve

and a recirculation air valve. The pressure drop across a calibrated orifice was

used to adjust and monitor the sampling flowrate. A calibrated dry gas meter

measured the total volume of the gas sampled during each sampling run.

The collection efficiency of the impactor was defined as that mass

fraction of particles exiting the impaction jet which was collected in the liquid



I -


1 i-

li .



0 *-

0 0
0 Zi
11, *5

C.) z











solution. The total mass exiting the.jet was estimated by summing the mass

measured in the solution with that mass measured on the 47 mm after-filter.

Collected aerosol mass deposits were quantified fluorometrically using a

Turner Model 112 fluorometer calibrated for fluorescein concentrations ranging

from one part per billion to 14,000 parts per billion. For the collection efficiency

tests, the impactor's liquid solution consisted of 50 ml of 0.1N NH4OH. At the

conclusion of the sampling run, this solution was analyzed directly. To quantify

the mass of aerosol collected on the after-filter, the filter was placed in a 100 ml

polypropylene container with 50 ml of 0.1N NH4OH and the container placed in

an ultrasonic bath for approximately 20 minutes. Previous analytical tests have

shown that virtually 100% extraction efficiency can be achieved using this


Run times for the calibration tests varied depending upon the amount of

aerosol mass collected. For the large particle calibration tests, only 10 minute

run times were necessary. For the smallest particles generated, however, three

hour run times were necessary to achieve sufficient mass deposits for reliable


The performance of the impactor was measured as a function of both

particle size and sampling flowrate. Thirteen discrete particle sizes were used

ranging from 0.05 micrometers to 20 micrometers aerodynamic diameter. For

each particle size generated, collection efficiency tests were performed at

flowrates of 10, 20, and 30 Ipm. Each test series combination of particle size

and flowrate consisted of three separate collection efficiency tests performed

under identical test conditions.

Results of the collection efficiency tests are plotted in Figure 3-6.

Individual test results are tabulated in Appendix B. As expected, the impactor's

collection efficiency varies as a function of both aerodynamic particle size and

sampling flowrate. The right hand side of the collection efficiency curves is of

particular importance. Independent of flowrate, the shape of the curves

resemble that of conventional inertial impactors with impaction efficiency

increasing monotonically with increasing particle size. Clearly, inertial impaction

is the predominant collection mechanism for particles in this size range. In

accordance with impaction theory, the collection efficiency also increased with

increasing flowrate. Measured cutpoints in this region at 10, 20, and 30 Ipm

were 0.90, 0.53, and 0.46 micrometers aerodynamic diameter, respectively.

These measured values compare somewhat favorably with those predicted by

standard impactor theory (0.92, 0.63, and 0.42, micrometers).

The overall shape of the complete curves is quite similar to that predicted

by the simplified bubbling theory. Impaction is clearly dominant for large

particles while molecular diffusion accounts for deposition of small submicron

particles. Note also that an absorption minimum occurs in the range of 0.2 to

0.36 micrometers aerodynamic diameter which is approximately the size region

predicted by bubbling theory. The position and magnitude of the absorption

minimum was also noted to be a function of sampling flowrate. Below


S -
o 000

OI C 00
n 0



4 0 0


< 0



oo C
O 4

I I I I I I I I o o

O O o o o o ( '0 N









(%) AOua!90013 uo!01alo00


0.2 micrometers diameter, molecular diffusion dominates the particle absorption

and efficiency actually decreases with increasing flowrate. This effect is

believed to result as a consequence of the gas' lower residence time within the

liquid column at higher flowrates.

In terms of quantitative chemical analysis, it is the overall mass efficiency

of the collection device on the aerosol which is of importance. Recall from the

measured size distributions in Figure 3-1, that 95% to 99% of the lead-based

mass is associated with particles greater than one micrometer aerodynamic

diameter. Since the impactor's collection efficiency on these particles is high,

integration of the collection efficiency curve over the measured size distribution

predicts that the impactor should collect this aerosol with over 95% mass


Particle losses through the impactor's inlet section were also quantified

during the calibration tests. Following each test series, the impactor was

disassembled and the inlet section carefully rinsed with a known volume of

0.1N NH4OH. Particle mass deposits were quantified by fluorometric analysis of

the collected rinse solution. In general, test results showed that inlet losses

represent only a small fraction of the sampled aerosol mass. Submicron

aerosol deposits were normally below the limit of quantitation of the fluorometric

technique. Large particle losses were found to be more significant with losses

of 2.2%, 3.2%, and 4.0% measured for particles of 6, 10, and 20 micrometers

aerodynamic diameter, respectively. Subsequent evaluation of the impactor

during its actual field use supported these test results.

Based on the calibration tests, it was concluded that the impactor's

design met the criteria outlined for the overall instrument design. For the

aerosols of interest in the battery industry, the impactor is capable of capturing

the aerosols with high collection efficiency. Use of a liquid solution as the

impaction substrate eliminates particle bounce and substrate overloading which

can exist for solid impaction substrates. Turbulence within the fluid during

collection provides for effective dissolution of the lead-based particles as well as

providing for a well mixed, homogeneous solution. Unlike solid impaction

substrates, the liquid is easily replaced after each sampling test and is

compatible with the chosen lead analysis technique. Lastly, the impactor can

be easily constructed, is physically compatible with the corrosive nitric acid, and

requires little periodic maintenance.



Accurate quantitation of lead-in-air aerosols requires representative

collection and efficient transport of the lead-based particles to the instrument's

lead analysis section. Since it is infeasible to capture and analyze an entire

aerosol, a fraction of the aerosol must be extracted which reflects the aerosol's

particle size distribution, concentration, and chemical composition. Because

sampling flowrates may vary from one test to another, the designed

instrument's sampling flowrate must be controllable and accurately measured.

Description of Air Sampling System

Figure 4-1 is a schematic of the air sampling components associated

with Module 1. A description of these components is given in Table 4-1. In

addition to the components shown, Module 1 contains the system's nitric acid

reservoir, liquid pumps, and lead analysis system. Referring to Figure 4-1, the

vertical orientation of the impactor is compatible with its liquid contents and

minimizes particle transport losses from the Module 1 inlet to the impaction

stage. The Module 1 sampling system consists of a set of sharp-edged



To Module 2
Vacuum Source

Figure 4-1. Schematic of Module 1 air flow system. Component descriptions
are provided in Table 4-1.

1 2

Table 4-1
Description of Module 1 Air Flow Components

Figure 4-1 Code Description

1 3/16, 1/4, or 3/8 in. Diameter SS Nozzle

2 SS Female Connector, 1 in. Tube 1 in. NPT(F)

3 SS Rexible Hose, Teflon Lined
Cajon, Inc. Model SS-16BHT-36

4 SS Bulkhead Union, 1 in. Tube 1 in. Tube

5 SS 3/8 in. O.D. Tube, 4.25 in.

6 SS Male Connector, 1 in. Tube 1 in. NPT(M)

7 SS Male Elbow, 3/8 in. Tube 1/2 NPT(M)

8 SS 3/8 in. O.D. Tube, 3 7/8 in.

9 SS Union Elbow, 3/8 in. Tube

10 SS 3/8 in. O.D. Tube, 7 7/8 in.

11 SS Bulkhead Union, 3/8 in. Tube

12 3/8 in. Polypropylene Tubing, 6 ft.

nozzles, sampling probe, and impactor in addition to their associated tubing

and compression fittings.

Sampling of indoor air (non-ducted airstreams) is used to evaluate

occupational lead exposures. For indoor air measurements, the sampling

nozzles and sampling probe are not needed. Instead, a sharp-edged nozzle

has been fabricated from a one inch diameter stainless steel tube and attaches

to the top of Module 1. Its vertical orientation allows for direct collection and

transfer of airborne particles to the collector with minimal transport losses.

Unlike indoor air sampling, the sampling of ducted airstreams such as

recirculation airstreams requires sampling from a horizontal duct with high

associated air velocities. As will be discussed, accurate particle sampling from

a flowing airstream requires that a thin-walled nozzle be oriented directly into

the airstream. In addition, the sampling flowrate must be controlled such that

the nozzle inlet velocity equals that of the freestream velocity at the point of

sampling. Depending upon the measured freestream velocity, a nozzle of 3/16,

1/4, or 3/8 inch diameter can be used.

Sampling from horizontal ducts requires that the sampled aerosol must

undergo a 90 degree change of direction prior to entering the instrument's

vertical collection section. The sampling system contains a 39 inch long flexible

probe which is used for this purpose. The probe consists of a teflon inner

lining surrounding by a stainless steel braid. The teflon lining is compatible with

elevated temperatures up to 400 OF and is resistant to corrosion. The probe's

0.88 inch inner diameter matches that of the impactor inlet tube thus minimizing

particle transition losses due to air expansion or contraction.

Air flow through Module 1 is provided by connection to the vacuum

source located in Module 2. As shown in Figure 4-2, Module 2 contains an air

filter, mass flow controller, and cooling fan. Component descriptions are

presented in Table 4-2. The cartridge filter contains both activated charcoal and

fibrous filters and serves to protect the mass flow controller and sampling pump

from any particles remaining in the airstream. A high capacity vacuum pump

with teflon coated wetted surfaces provides the driving force for air flow through

the sampling system. As shown in Figure 4-2, this pump has been fitted with a

recirculation valve to reduce the load to the pump.

The sampling flowrate in the lead-in-air monitor is controlled using a

surplus Tylan Model FC-262 mass flow controller which has been adapted for

use in the system. The FC-262 is a 0-50 slpm capacity unit with flowrate

controlled by an internal solenoid valve. Selection of the flowrate is normally

provided by manual adjustment of a potentiometer to supply a 0 to 5 VDC

command signal to the flow controller. Feedback circuitry within the flow

controller continually adjusts the valve setting so that the actual air flow equals

the desired flowrate selected by the user. A measure of the actual flowrate is

provided by a corresponding 0 to 5 VDC output signal which is normally read

by the user from a digital panel meter. For adaptation to the prototype monitor,

the potentiometer input circuit and output display were removed and the

From Module 1

1 19 //
1 ,18

2 7 13


Valve Vacuum

Filter 4 15 P Pump


Outlet Il
5 7 13 1

6 Mass
Flow 10

Module 2
(Top View)

Figure 4-2. Schematic of Module 2 air flow system. Component descriptions
are provided in Table 4-2.

Table 4-2
Description of Module 2 Air Flow Components

Figure 4-2 Code Description

1 SS Bulkhead Union, 1/2 in. Tube 1/2 in.

2 SS 1/2 in. Tube, 2 3/16 in.

3 SS Female Connector, 1/2 Tube 3/8 NPT(F)

4 Disposable Air Filter
Koby Inc., Model "Junior King"

5 SS Female Connector, 3/8 in. Tube 3/8 NPT(F)

6 3/8 in. Polypropylene Tube, 6 5/16 in.

7 SS Elbow, 3/8 in. Tube 3/8 in. Tube

8 SS 3/8 in. Tube, 1 3/4 in.

9 Mass Flow Controller, 0-50 slpm
Tylan Corp., Model FC262

10 SS 3/8 in. Tube, 2 in.

11 3/8 in. Polypropylene Tube, 5 3/8 in.

12 SS Male Tee Run, 3/8 in. Tube 3/8 NPT(M)

13 SS 3/8 in. Tube, 2 5/8 in.

Table 4-2--continued

Figure 4-2 Code Description

14 SS 3/8 in. Tube, 2 9/16 in.

15 Valve, Swagelok Corp., Model B1RS6

16 Vacuum/Pressure Pump, 1/4 hp, teflon lined
3.2 cfm, 4A at 0 psi
Thomas Industries, Model 727CM39-TFE

17 3/8 in. Polypropylene Tube, 2 1/4 in.

18 SS Bulkhead Female Connector, 3/8 in. Tube
to 3/8 in. NPT(F)

19 Muffler, Koby Inc., Model M-1-N

command signal is now supplied by an analog output channel from the data

acquisition and control system. The flow controller's output signal is read by a

corresponding analog input channel on the data acquisition and control system.

Figure 4-3 is a complete wiring diagram of the Module 2 circuitry.

Component descriptions, functions, and commercial sources are listed in Table

4-3. Table 4-4 provides connector pin descriptions. A standard 120 VAC, 60

hz source supplies power to the mass flow controller power supply, vacuum

pump, and cooling fan. Power to the flow controller is supplied by a 15 VDC,

200 mA power supply. The power supply is protected using a 0.5 amp fuse.

The sampling pump is internally protected against current overload and thus

requires no additional fusing. On/off control of the air pump and the cooling

fan is provided by use of a solid state relay capable of handling load currents

up to 25 amps. The relay itself is activated through application of a 3-32 VDC

signal which is compatible with standard TTL digital command signals. All

Module 2 input and output signals are communicated by a standard 9 pin

ribbon cable to the Kiethley 575 data acquisition and control system.

Prior to its use, the Tylan mass flow controller had been repaired and

recalibrated by the manufacturer. Its condition and calibration were then

verified in the laboratory. At a variety of flowrates and upstream and

downstream pressures, the actual mass flowrate was measured as a function of

selected mass flowrate. For these tests, a National Bureau of Standards

traceable Gillian Model D-80028 bubble flow meter was used as a flow








8. g







Table 4-3
Module 2 Electrical Comoonent Ustina and Descriotion


Barrier Terminal

Solid State Relay

Appliance Plug

Cooling Fan

Fan Filter Screen(2)

Vacuum/Pressure Pump

Power Supply


Fuse Block

Connector KK

Connector MM


3 Contact, Double Row, 20A
Cinch, Inc. type 3-141

SSR Series, 240 VAC/25A
3-32 VDC input
Potter and Brumfield
Type SSR-240D25

250/115 VAC, 10/15A
Feller, Inc. type 6100-31

115 VAC, 0.15A
34 cfm at 0 in H20 S.P.
Comair/Rotron type SV2A1

Anodized, for 3.5 in fan
Fan-S/Qualtek type 06325-B

115 VAC, 1/4 hp, teflon lined
3.2 cfm, 4A at 0 psi
Thomas Industries type 727CM39-TFE

Dual.Voltage (-15/+ 15 VDC), 200 mA
Sola, Inc. type 85-15-2120

1 1/4 X 1/4 in, 0.5A slow-blow
SPC Technology type F-500SS-0.5

Screw termination
SPC Technology type FB1PS

9 pin D-Sub, crimp style plug
SPC Technology type DEC-9P

22 position, card-edge connector

Table 4-4
Module 2 Electrical Connector Pin Descriptions





























Pump, fan on/off

Ground Ground

LL-A MFC Command Signal
0-5 VDC

LL-3 MFC Output Signal
"High", 0-5 VDC

LL-2 MFC Output Signal
"Low", 0-5 VDC

KK-5 MFC Output Signal
"Low", 0-5 VDC

KK-4 MFC Output Signal
"High", 0-5 VDC

Power +15 VDC Supply

KK-3 MFC Command Signal
0-5 VDC




-15 VDC Supply





calibration standard. For each operating pressure and flowrate range, six

separate measurements were made. As shown in Figure 4-4, the correlation

between the actual mass flowrate and the desired flowrate was generally good.

In the 10-30 slpm flowrate range of interest, the agreement was within 2%.

Note that conversion from actual to standard flowrate must account for the

fluid's temperature and pressure at the point of sampling. Conversion from

standard temperature as defined by the mass flow controller's manufacturer

(0 C) to standard temperature as defined by occupational regulations (25 OC)

must also be taken into account. In the prototype lead-in-air monitor, these

conversions take place automatically through the instrument's software.

Prediction of Particle Sampling and Transport Efficiency

Accurate measurement of lead-in-air concentrations requires that lead-

based particles be sampled, transported, collected, and analyzed efficiently.

Design of an efficient particle sampling and transport system depends primarily

on the particle size distribution expected, the fluid properties at the point of

sampling (such as temperature and velocity), and the air sampling flowrate

required by the instrument's particle collection section. Efficient design,

calibration, and field testing of a separate sampling system was beyond the

scope of this project. As a preliminary approach, a particle sampling and

transport system was adapted from EPA Reference Method 12, "Determination

of Inorganic Lead Emissions from Stationary Sources." In the forthcoming





5 10

15 20 25 30 35

Set Flowrate


Figure 4-4. Measured performance of Module 2 mass flow controller.

sections, the sampling and transport efficiencies of this design will be predicted

based on published theoretical and empirical studies. Alternatives to this

sampling and transport system will be discussed.

Sampling from Still Air

Evaluation of occupational exposures to lead-in-air concentrations

represents sampling from an airstream at zero or low velocity. Because

particles possess higher stopping distances at higher velocities, sampling and

transport losses during still air sampling can be negligible when compared to

sampling from ducted, high velocity airstreams. There exist, however, two

types of still air sampling biases which can result particularly with regard to the

large particle distributions characteristic of the battery industry.

Efficient still air sampling requires use of a vertically oriented sampling

tube. When the inlet air velocity is low, a sampling bias can result due to high

settling velocities characteristic of particles with large aerodynamic diameters.

At low sampling velocities, this results in an overestimation of particle

concentrations. An additional bias can result due to the effect of particle inertia

on particle trajectory. For sampling in still air, curved streamlines result as air is

accelerated towards the inlet nozzle. Due to their inertia, particles will be unable

to follow their respective streamlines and will be displaced a distance which may

be large with respect to the dimensions of the inlet. Inertial effects of this type

can result in an underestimation of actual particle mass concentrations.

Based on these possible biases, various criteria have been developed for

the design of air samplers. Davies (1968) developed inlet criteria based on

parameters of nozzle inlet diameter, sampling flowrate, and particle

aerodynamic diameter. Agarwal and Lu (1980) predicted less than 10%

sampling losses if the probe diameter (Dt) meets the following criterion

De > 20 r2 g (4-1)

where r is the particle relaxation time and g is the gravitational constant.

Equation 4-1 was used to predict inlet losses of the 0.88 inch nozzle

used for still air sampling of the lead-in-air monitor. Results showed that

particles up to 60 micrometers aerodynamic diameter would be collected

without significant error. Particles greater than 60 micrometers would be

collected with less than 90% efficiency. The equal diameter of the nozzle the

inlet section ensures that particle transport losses should be minimized.

Sampling from Ducted Air Streams

Unlike sampling from still air, sampling from ducted airstreams can result

in significant sampling biases particularly with regard to the high velocity

airstreams and large particle distributions characteristic of the battery industry.

Moreover, the horizontal arrangement of ventilation ducts requires that a

captured aerosol must be transported a short distance to the instrument which

also involves a change in the aerosol's flow direction. As a result, particle

losses within the aerosol transport system may significantly reduce the mass

concentration detected by the lead analysis section. The factors affecting

particle sampling and transport biases will be discussed.

The purpose of the aerosol sampling system is to obtain a representative

aerosol sample from the airstream of interest. In ducted airstreams, this

involves use of a sampling nozzle and probe to extract a representative aerosol

sample and transport it to the instrument's collection section. Sampling bias

can occur, however, due to particles' inertia in flowing airstreams. Similar to the

sampling from still air, sampling criteria have been developed for efficient

sampling from ducted airstreams. These criteria are generally of three types.

First, sharp-edges inlet nozzles must be used to obtain representative air

samples. Blunt-edged orifices can disrupt the airflow entering the nozzle

resulting in nonrepresentative particle collection and high inlet losses (Vincent

and Gibson (1981), Vincent et al. (1982), Vincent et al. (1985), Stevens (1986),

Vincent (1987), Vincent and Gibson (1988)). The second criterion is that the

nozzle must be oriented directly into the oncoming airstream. Changes in the

sampled air direction due to nozzle misalignment are a function of particle

inertia, the misalignment angle, and the probe diameter (Durham and Lundgren,

1980). Lastly, the sampling should be performed under isokinetic conditions

(nozzle inlet velocity equal to freestream velocity). Anisokinetic sampling can

result in significant large particle sampling biases whose magnitude and

direction depend on the velocity ratio and the particle Stokes number. Note

that isokinetic sampling requires prior measurement of the airstream's velocity

at each of the desired sampling points. When sampling, achieving desired

nozzle inlet velocities requires proper combination of nozzle diameter and

sampling flowrates.

Achieving representative aerosol sampling does not ensure efficient

transport of the aerosol from the inlet nozzle to the particle collector. Particle

losses can occur both in the inlet nozzle as well as in the transport section.

Deposition can take place due to molecular diffusion, thermophoretic effects,

electrostatic effects, gravitational effects, and inertial effects. For the sampling

of aerosols of interest in the battery industry, only gravitational and inertial

mechanisms will result in significant transport losses.

Numerous theoretical and experimental studies have been conducted in

relation to understanding particle transport in sampling tubes. In general, these

studies agree on the recognized parameters of importance (such as particle

Stokes number, sampler geometry, and fluid conditions) but often vary widely in

terms of their conclusions and recommendations. Theoretical predictions of

transport losses requires extensive mathematical modeling of both fluid flow

fields and particle trajectories within the flow fields. The validity of these models

is often not confirmed. Differences in sampler geometries and flowrates can

result in significantly different predicted losses. Experimental loss

measurements are generally more practical but differences in sampler

geometries, operating flowrates, and particle properties make comparison of

separate studies difficult.

In the lead-in-air monitor's sampling system, particle transport losses

during duct sampling can occur either in the inlet nozzle or in the flexible

sampling probe. To predict nozzle inlet losses, a brief review of pertinent

literature was made. Although the actual loss will depend on nozzle diameter,

sampling flowrate, and particle size, the literature suggests that large particle

losses may exceed 5% by mass. In the lead-in-air monitor, additional nozzle

losses might also be expected due to the airflow's slight expansion from the

nozzle outlet diameter of 0.75 inches to the 0.88 inch diameter of the sampling

probe. The magnitude of the losses in the expansion section will depend on

sampling flowrate and particle size.

Losses in the sampling probe can more accurately be predicted than

those associated with the inlet nozzles. Considering the flexible probe oriented

as in Figure 4-1, the probe can be divided into three main sections for analysis:

an 8 inch long horizontal section, a 23 inch long section with a 14.5 inch radius

of curvature, and a 8 inch long vertical section. The particle losses in these

sections were considered separately. These calculations were simplified by the

fact that laminar flow conditions exist in the 0.88 inch diameter probe even at a

flow of 30 Ipm. Turbulent flow conditions alter the nature of the fluid flow field

and particle trajectories making accurate loss predictions more complex.

Particle deposition in the 8 inch long horizontal section occurs primarily

due to the force of gravity. During their transit time within the tube, particles

may settle a distance sufficient to be removed from the airstream. The

magnitude of these losses was predicted based on a study by Thomas (1958)

on sedimentation losses in horizontal cylindrical tubes. For laminar flow

conditions, the transport efficiency (penetration) can be predicted

P = 1 [ 2 [ 2 1- 4 2/3 + arcsin0l/3 4 1/3 /1- 2/3 ] (4-2)

3x L Vs R

where P is the fractional penetration, L is the tube length, V, is the particle

settling velocity, R is the tube radius, and Q is the volumetric flowrate through

the tube.

This relationship was evaluated for the 8 inch long section as a function

of particle size and sampling flowrate and is presented graphically in Figure 4-5.

Independent of flowrate, particles less than 10 micrometers aerodynamic

diameter can be transported through this section with fairly high efficiency.

Particle penetration above 10 micrometers, however, is predicted to decrease

fairly rapidly as a function of particle size. Note also that the lower sampling

flowrate results in lower transport efficiency due to the gas' longer residence

time within the horizontal tube. Efficient transport through this section,

o 0
0 0




0 U


(0 0-


O ~



I Nc o m M
.O .) .


'o 0

therefore, would imply the use of high sampling flowrates and tubes of smaller


Unlike losses in horizontal sampling tubes, particle losses in 900 bends

are due primarily to inertia and centrifugal forces on the airborne particles. Due

to their inertia, particles will be unable to follow their respective streamlines

directly and may be displaced a sufficient distance to impact on the duct walls.

Air flow around the bend can also establish a rotational flow pattern normal to

the mean direction of flow. This rotational profile can result in additional

particle losses around the bend. Cheng and Wang (1981) developed a

relationship for transport efficiency in a 900 bend as a function of particle Stokes


P = 1 STK (4-3)

=1- pp DP2
18 l x R3

This relationship holds for laminar flow conditions and is plotted in Figure 4-6 as

a function of particle size and flowrate. Similar to losses in the horizontal

section, losses around the bend are minimal for particles less than

10 micrometers aerodynamic diameter. Particles greater than 20 micrometers

aerodynamic diameter are difficult to transport efficiently. Note also the effect of

sampling flowrate on transport efficiency. Higher flowrates correspond to larger


So .

/ 2

/ G )

co c
r / o 3

/ .
*/ .O 0-
l'*' c-L e
I/ o O L


I1 I I 0 I
S o, ,. o. 0 o .-
o o

0.0 N 0L

0 O C cOt N0

S* 6
-o 00 0 0

" 0 0 0 0o 00 0 0 0-- 0



stopping distances which would result in greater particle deposition. Inspection

of equation 4-3 would dictate use of low sampling flowrates and large diameter


Note that the effects of flowrate and tube diameter on bend losses are

contradictory to those required for efficient horizontal transport. For a given

tube diameter, higher flowrates result in higher horizontal transport efficiency but

lower transport efficiency around the bend. Conversely, lower flowrates

minimize bend losses but maximize losses in the probe's horizontal section.

Combined losses within these two sections are thus relatively independent of

sampling flowrate. The magnitude of these losses depends primarily on

aerodynamic particle size.

Particle losses within the probe's final 8 inch long vertical section are

predicted to be minimal. Since the tube is oriented vertically, no horizontal

deposition is possible. Operation of the probe in the laminar flow regime also

eliminates the possibility of turbulent deposition which can be significant for

particles with high Stokes numbers in turbulent flow fields. This prediction of

low particle losses in this section is supported by the low impactor inlet losses

measured during the impactor's laboratory calibration.

It can be concluded, therefore, that use of the proposed sampling nozzle

and probe may not allow efficient transport of large particles to the lead-in-air

monitor's collection and analysis section. Combined losses within the nozzle,

nozzle outlet, and sampling probe may be prohibitively high. It should be

noted, however, that these loss calculations may significantly overestimate

particle losses. Theoretical loss calculations normally assume that when a

particle contacts a surface, it is removed from the airstream. Use of liquid

aerosols during experimental loss tests would support these calculations. In

reality, solid particles contacting solid surfaces at high velocity may experience

only fractional sticking efficiencies. Reentrainment of deposited particles also

results in higher particle transport efficiency.

It is apparent that special sampling considerations may need to be given

to the on-line measurement of ducted airstreams in the battery industry.

Conventional stack sampling techniques and equipment may not enable

representative sampling and transport of these large aerosols. As possible

alternatives, three sampling approaches could be investigated. First, the

ductwork at the monitoring point of interest could be modified to include a short

vertical section so that the entire flow is directed downward. A sharp-edged

sampling nozzle attached vertically to the lead-in-air monitor could then

isokinetically extract a representative sample with minimal inlet and transport

losses. Alternatively, a small flexible test section attached to a high volume

vacuum source could isokinetically extract a portion of the major duct flow and

direct this secondary airstream downward over the monitor's sampling nozzle.

Aerosol not sampled by the monitor could then be returned to the major

airstream or filtered prior to its exhaust to the atmosphere. This concept is

similar is design by Baron (1988) and Ball and Mitchell (1990) for efficient

transport of coarse aerosols to particle sizing instrumentation. As a final

alternative, the conventional sampling nozzle could be replaced with a

precollector designed to remove all particles greater than 10 micrometers

aerodynamic diameter. The remaining fraction could then be transported

efficiently to the monitor's collection section. In the human respiratory tract,

particles less than 10 micrometers represent the size fraction capable of

penetrating to the alveolar region and thus represent the best measure of the

actual lead exposure hazard. Particles larger than 10 micrometers are

deposited in the upper regions of the respiratory tract and are absorbed in the

bloodstream with much lower efficiency. Recent developments in

10 micrometer size selective inlets indicate that this design approach is feasible.

Use of a precollector to evaluate worker exposure, however, must first meet

with approval by the Occupational Health and Safety Administration.



As discussed, accurate evaluation of lead-in-air concentrations requires

efficient aerosol sampling, transport, collection, and accurate analysis of the

collected aerosol's lead content. Previous sections have outlined the

techniques required to ensure representative aerosol sampling, transport, and

collection. This chapter will outline design criteria for the lead analysis

section, briefly review available analytical techniques, and describe in detail the

analytical technique chosen for incorporation into the lead-in-air monitor.

Design Criteria

Selection of the instrument's lead analysis technique was based on

careful consideration of several design criteria. First, the technique must be

specific for elemental lead and capable of accurately quantifying lead

concentrations ranging from 50 ppb to 9000 ppb. It was desired that the

technique's response be fairly linear over a wide range of concentrations and

the technique display adequate repeatability. Because the collected aerosol

sample's composition is not entirely lead, the analytical technique should

possess minimal physical and chemical interference from these other

components. The technique must also be physically compatible with the

chosen particle collection substrate. It was important that the selected

analytical technique require little periodic maintenance or recalibration. Finally,

the selected technique must meet cost requirements and be adaptable to

convenient automation.

Based on these criteria, a thorough review of pertinent literature was

made with regard to selection of the instrument's analytical technique. There

exist a wide variety of analytical techniques available for the quantitation of lead

either in solid or liquid phase. These techniques include both conventional and

graphite furnace atomic absorption spectroscopy, ion chromatography,

inductively coupled argon plasma spectroscopy, x-ray and electron techniques,

colorimetry, neutron activation analysis, and electrochemical techniques. For an

automated device, most of these techniques were eliminated from consideration

either due to their high cost, level of complexity, inadequate lead detection

capabilities, or failure to adapt themselves to convenient automation. For

example, conventional atomic absorption is incapable of detecting sub-ppm

levels of lead in solution. Although graphite furnace atomic absorption

possesses the necessary detection capabilities, its use would have greatly

increased the complexity and cost requirements of the monitor. In addition, the

response of both conventional and graphite furnace atomic absorption displays

a fairly limited range of linearity. Other techniques such as colorimetry involve a

number of chemical handling steps which would have been inconvenient to


Ultimately, an electrochemical method of lead analysis was regarded as

most appropriate for the lead analysis. Because metals display high

electroactivity, their presence is solution can be detected over a wide range of

concentrations. Electrochemical techniques display high linearity and

repeatability. In particular, the electrochemical technique of polarography was

selected as that which best met the outlined design criteria. In forthcoming

sections, the principle of polarography will be reviewed and its application to

lead-in-air monitor described.

Description of Polarographic Analysis

Electrochemical analysis is based on the observed behavior of

electroactive species in solution when subjected to an electric potential gradient.

For the detection of a particular species of interest (such as metals), a varying

potential is applied to the solution through a conductive electrode. If at a

particular potential a species is oxidized or reduced, an electric current will flow

at the electrode. Since the rate of electron transfer is a function of the number

of redox reactions occurring at a given potential, the measured current

produced is proportional to the concentration of that component in solution.

Because electroactive species display characteristic redox potentials, the

observed potential at which the redox reaction occurs can be used to identify

the component of interest. By applying a range of potentials to a sample and

noting the potential and resulting current of several redox reactions,

simultaneous quantitation of several solution components is possible. This

analytical technique is commonly referred to as voltametry.

Although the study of voltametry is not new, recent advances have led to

the development of reliable, commercially available instruments suitable for trace

metal detection. Current designs now incorporate mercury drops as the active

working electrode surface. Since a fresh mercury drop is produced for each

separate sample analysis, carry over contamination of the working electrode is

virtually eliminated. Recent advances in microprocessor technology have also

led to increasing measurement sensitivity and instrument capabilities.

Commercially available analyzers allow the user to select from various applied

potential waveforms which optimize detection of the component of interest. For

many metals (including lead) part per billion detection capability is routine.

Based on a review of available electrochemical detectors, an EG&G

Model 303A static mercury drop electrode was purchased along with an

accompanying Model 264A control unit. The Model 303A contains a mercury-

based working electrode, a silver/silver chloride reference electrode, and a

platinum counter electrode which are all immersed in a static cell containing the

solution of interest. The Model 303A also contains a 5 Ib capacity mercury

reservoir and a delivery system for dispensing individual mercury drops of

repeatable size.

Control of the Model 303A electrode system is provided through

communication with the Model 264A control unit. The Model 264A sets the

Model 303A operating parameters, provides for automatic sequencing of the

analysis steps, receives and conditions the resulting output signals, and

communicates the current versus voltage scan results to an x-y recorder. The

user is able to select from a variety of a scan types including normal direct

current, normal pulse, and differential pulse polarography as well as direct

current and differential pulse stripping analysis. Although these techniques

possess varying degrees of capabilities and applications, differential pulse

stripping analysis provides for the most sensitive detection of lead in solution.

This mode, therefore, is used exclusively for lead quantitation in the lead-in-air


Quantitation of lead in solution by stripping analysis typically involves the

following sequence of steps. First, high purity nitrogen is bubbled through the

liquid sample at a flowrate of approximately 40 cc per minute for approximately

3 minutes. This deaeration step removes the electroactive interference of any

dissolved oxygen in the solution. The Model 303A then dispenses a single

mercury drop of selectable size which hangs on the tip of a glass capillary

immersed in the solution. A constant electric potential is applied to the mercury

drop whose magnitude is more negative than the half-wave potential of the

analyte of interest. This results in a electro-reduction of the analyte onto the

surface of the mercury drop. To ensure repeatable mass transfer rates during

this step, the solution is stirred and deposition is performed for a fixed time

period (such as 3 minutes). The preconcentration step is responsible for the

technique's low detection limit for many electroactive species.

At the conclusion of the deposition step, the stirrer is turned off and the

electroactive species is stripped from the electrode by making the electrode

more positive and oxidizing the species back into solution. In this stripping

step, varying the potential at a constant rate (such as 10 mV/sec) allow

different species to oxidize at different times. In differential pulse stripping

analysis, the current measured during each species' oxidation reaches a

maximum value that is proportional to the concentration of each species that

was deposited.

As an example, Figure 5-1 presents the results of a typical analysis of

40 ppm lead, cadmium, and copper standards in 0.1N nitric acid. The position

of the peaks (with potentials referenced to the Ag/AgCI electrode) identifies the

species present in solution and the peak height for each species is proportional

to its concentration in solution.

To ensure reliable test results of trace metal analysis, special

experimental techniques were observed during all polarographic analyses. To

minimize sample contamination and species absorption losses, all polarographic

cells were made of teflon and were acid washed with 6M nitric acid prior to use.

Metal standards were prepared from 1000 ppm stock metal standards certified

by the manufacturer to within 1% accuracy. Serial dilutions were performed


00 d

co '


o o

o *0

I n I

> S

o C .
o c a
0 0 > I
Q I- I-- I cI S


(V 90-30 L) -ua-0in


using positive displacement pipettes with disposable tips rated to within 2%

accuracy. Dilute nitric acid solutions were prepared using trace metal grade

concentrated nitric acid diluted with distilled, deionized water with a measured

resistivity of 17 x 106 ohm-cm. Only certified, triple distilled mercury and ultra-

high purity nitrogen were used in the polarograph.

If polarographic stripping analyses are performed under carefully

controlled conditions, repeatable test results can be obtained over a wide range

of concentrations. Figure 5-2 presents the instrument's response as a function

of lead concentration in 0.1N nitric acid. Each data point represents the

average of 6 separate tests performed under identical test conditions. The

technique's precision (as expressed by the relative standard deviation) was

measured to be approximately 2%.

Considerable attention was devoted to identification and subsequent

optimization of the parameters which affect the polarographic's response to

trace lead concentrations in solution. Factors affecting the deposition step were

identified to be initial deposition potential, mercury drop size, deposition time,

and solution stirring dynamics. The instrument's response during the stripping

phase was found to vary with the potential's rate of change (scan rate) and the

height of the differential wave applied to achieve oxidation. The overall

response was also found to vary with the condition of the Ag/AgCL reference

electrode, solution temperature, and the concentration of the nitric acid solution




0 c
\o -.

00 0


\.o c

\ 0

u CC
o c c

co c

\o 2

\ a ,1

00 .2
\ ^--

C, 2o o
o o c a,

Within limits, these variables can be controlled by proper selection of the

polarograph's operating parameters. However, complete control of all of these

variables in an automated monitor is impractical to achieve. In stripping

analysis, the instrument's response to a given lead standard was noted to vary

as much as 50% during successive test series. Ultimately, this variability was

virtually eliminated by including an internal standard (200 ppb copper) in the

nitric acid electrolyte.

Use of the internal copper standard was developed based on numerous

observations of the polarograph's response to several species in solution.

During the analysis of equal mass concentrations of lead, copper, and

cadmium, it was noted that the ratios of the resulting peak heights were

essentially constant regardless of the operating parameters of the instrument.

For example, although reducing the deposition time during stripping analysis

reduced the magnitude of each element's peak, the respective peak ratios

remained essentially constant. Similar behavior was noted by changing the

mercury drop size, nitric acid concentration, analysis scan rate, solution

temperature, and solution stirring dynamics. Thus, if the instrument response to

a known concentration of an analyte is measured, the concentration of a

different species can be calculated from its observed peak height and the

known peak ratio. During the course of the project, the instrument's response

to equal lead and copper concentrations was measured as a function of

concentration and instrument's operating parameters. For 156 separate

measurements, the ratio of the copper to lead peak height was measured to

average 2.33 with a relative standard deviation of 1.8%. Assuming the use of a

200 ppb copper internal standard, the concentration of lead in solution can thus

be calculated using the following relationship

Lead Cone. (ppb) = Lead Peak Height X 2.33 X 200 ppb
Copper Peak Height (5-1)

Use of an internal standard in the lead-in-air monitor offers several

advantages. First, the need for periodic calibration of the analytical section is

eliminated since the instrument essentially calibrates itself during each sample

analysis. Analytical accuracy is no longer sensitive to variables which would be

difficult to control within required limits. Lastly, use of the internal copper

standard provides an internal quality control check during each analysis. If the

instrument's response to the copper standard is outside the expected range, it

serves as an indication that either the electrolyte was prepared incorrectly or

that a possible instrument malfunction may exist.

To determine the value of this design approach, lead standards of

concentrations ranging from 0 to 8000 ppb were prepared in 0.5M nitric acid

with a 200 ppb copper internal standard. The instrument's response to these

solutions was then measured and Equation 5-1 used to predict the actual lead

concentration. For each concentration, six separate measurements were

performed under identical test conditions. Figure 5-3 presents the results of the

0 to 200 ppb lead measurements in comparison to the instrument's ideal

50 100 150 200

Actual Lead Conc.


Figure 5-3. Polarograph response to 0-200 ppb lead concentration using
200 ppb internal copper standard.












response. It can be seen that agreement between the measured and ideal lead

concentration is quite good for lead concentrations greater than or equal to

approximately 20 ppb. The observed response at 0 ppb is due to the presence

of hydronium ions in solution which provide a residual current overlapping the

position of the lead peak. At a 0 ppb lead concentration this interference

typically provides a calculated lead concentration of 5 to 15 ppb. For lead

concentration above 20 ppb, however, the contribution of the hydronium ions is

insignificant. A minimum quantitation limit of 20 ppb lead is thus reported for

the monitor's analytical section.

Figure 5-4 presents the results of the 0 to 8000 ppb lead response tests.

Note that use of the internal copper standard is accurate even over this wide

range of concentrations. A summary of these test results is tabulated in

Table 5-1 including the technique's precision as expressed by the relative

standard deviation. For lead concentrations ranging from 20 ppb to 8000 ppb,

the precision was calculated to be approximately 2%.

Analysis of Field Bulk Samples

In any quantitative chemical analysis, it is important to ensure that the

analytical technique chosen be compatible with the sample to be analyzed.

Physical and chemical interference associated with the sample's measurement

should be noted and effort made to minimize them in order to achieve reliable

test results. In the case of the lead-in-air monitor, it was important to ensure


7000 Measure






0 1000 2000


3000 4000 5000 6000 7000 8000

Lead Conc. (ppb)

Figure 5-4. Polarograph response to 0-8000 ppb lead concentration using
200 ppb internal copper standard.






Table 5-1
Polaroaraph Response Using Copper Internal Standard

Actual Lead Conc. (DDb)













Polarograph Response
Lead Conc. (ppb) RSD(o%)













that the expected aerosol's composition be compatible with polarographic


During battery manufacturing, airborne particles are produced primarily

by mechanical breakup (dispersion) of lead oxide powder, lead oxide paste,

and the lead-based plate material. It can thus be expected that the chemical

composition of the aerosol produced from these materials is virtually identical to

that of its parent material. Bulk samples of lead oxide powder, paste, and plate

material were obtained at the battery plant and returned to the laboratory for

analysis in order to identify any potential analytical interference.

During these tests, a known mass of each material was dissolved in nitric

acid solution and analyzed using the polarograph. For comparison purposes,

these samples were also analyzed using a graphite furnace atomic absorption

(GFAA) spectrophotometer manufactured by Perkin-Elmer (Model 5100). The

GFAA is more widely used for trace metal analysis than the polarograph and

thus represents an accepted criterion upon which to base the polarograph's

accuracy. For each test series, the GFAA was calibrated for lead

concentrations in its linear operating range of 0 to 40 ppb by autodilution of a

100 ppb lead stock standard solution. For comparison purposes, the diluent

consisted of 0.5M nitric acid with a 200 ppb copper internal standard. To

minimize matrix effects associated with lead measurement by GFAA analysis, a

fixed quantity of a matrix modifier (0.2 mg NH4H20PO4 + 0.01 mg Mg(NO3)

was injected with each 5 pl sample.

The results of these comparison tests are presented in Table 5-2.

Analysis was first performed on certified laboratory samples of lead nitrate, lead

monoxide, and lead acetate. Based on the known formula weights of these

materials, each sample was prepared to obtain an expected lead concentration

of 100 ppb. As shown in Table 5-2, both the polarograph and the GFAA

measured these solutions with reasonable accuracy. Based on the three

replicate analyses for each material, the polarograph displayed somewhat better

repeatability than did the GFAA.

Similar solutions were prepared from the bulk materials collected at the

battery plant. Although the lead content of the lead monoxide powder can be

assumed with some confidence, its actual percentage in the lead paste can only

be estimated based on discussions with the battery plant personnel. Results

shown in Table 5-2 generally verified these estimates. Similar to the analysis of

the certified laboratory samples, these results indicated fairly good agreement

between the polarograph and the GFAA. More important, no interference was

observed during the polarographic analysis of these battery plant samples. It

was concluded, therefore, that no interference can be expected during analysis

of any airborne particles produced from these bulk materials.