&EPA
United States
Environmental Protection
Agency
Environmental Sciences Research EPA 600 2-78-21 2
Laboratory October 1978
Research Triangle Park NC 27711
Research and Development
Investigation of
Particulate Matter
Monitoring Using
Contact Electricity
Final Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-78-212
October 1978
INVESTIGATION OF PARTICULATE MATTER MONITORING
USING CONTACT ELECTRICITY
Final Report
Walter John, Georg Reischl, William Devor
and Jerome J. Wesolowski
Mr and Industrial Hygiene Laboratory Section
California Department of Health Services
Berkeley, California
Grant Number R 803719-01-2
Project Officer
John Nader
Emissions Measurements and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation
lor use.
-------
ABSTRACT
To better understand the contact electrification monitor for particulate
matter, charge transfer by aerosol particles impacting on metal surfaces has
been investigated. Monodisperse, uniformly charged or neutral aerosol
particles (1-5 vim diameter) from a vibrating orifice or fluidized bed
generator were bounced on a metal probe. The transfer of precharge from the
particles was found to be a sensitive indicator of the probe surface condition.
A surface preparation procedure was developed which yielded linear charge
transfer curves.
Measurements were made of methylene blue, potassium biphthalate, sodium
chloride and aluminum particles impacting on stainless steel, Inconel,
titanium and platinum probes. For insulating materials, the transfer of
precharge was independent of particle size while the contact charge was
proportional to the cube of particle diameter and directly proportional to
impact velocity. The magnitude of the contact charge was strongly dependent
on the electrical resistivity of the material. A theoretical model was found
to account semi-quantitatively for all aspects of the data. A major re-
maining difficulty is the lack of knowledge of the contact potential. The
implications of these findings for monitoring applications is discussed in
detail and the advantages and disadvantages of the monitor are assessed.
111
-------
CONTENTS
Abstract . iii
Figures vi
Tables viii
Acknowledgments ix
1. Introduction 1
2. Conclusions and Recommendations 3
3. Tests of the IKOR Monitor It
Experimental Method h
Results , k
U. Charge Transfer Experiments 8
Experimental Procedures 8
Measurements 17
5. Experiments with a Fluidized Bed Aerosol Generator 1*5
The Fluidized Bed 1*5
Charge Transfer Measurements with Aluminum Aerosol 52
6. Theory of Charge Transfer 58
Model of Charge Transfer Process 58
Discussion of the Theory 6U
7. Implications of the Findings for Applications 65
Principal Factors Affecting Monitoring 65
Recommendations for Monitoring Procedures 66
Overall Assessment of the Contact Electrification Monitor 67
References 68
-------
FIGURES
Number Page
1 Experimental Arrangement Used for Testing the IKOR Monitor 5
2 Experimental Arrangement for Charge Transfer Measurements 9
3 Apparatus for Induction Charging of Particles 10
h Detail of the Area Where the Droplets are Formed 10
5 Typical Calibration Curve for the Particle Charger 11
6 Faraday Cup.... 13
7 Impaction Probe for- Charge Transfer Measurements 15
8 Charge Transfer Curve Showing Surface Cleaning Effect 18
9 Nonlinear Charge Transfer Curve 20
10 Nonlinear Charge Transfer Curves .. 21
11 Linear Charge Transfer Curve Obtained with Polished Probe 23
12 Charge Transfer for Potassium Biphthalate Particles on Polished
Stainless Steel 2U
13 Contact Charge vs. Particle Diameter for Methylene Blue on
Stainless Steel 25
lU Contact Charge vs. Particle Diameter for Sodium Chloride 26
15 Contact Charge vs. Particle Diameter for Potassium Biphthalate.... 28
l6 Contact Charge vs. Particle Diameter for Methylene Blue on
Titanium 29
IT Contact Charge vs.. Particle Velocity for Methylene Blue on
Stainless Steel 31
18 Contact Charge/Mass for Various Materials 3^
19. Charge Transfer for Potassium Biphthalate on Titanium 35
Vi
-------
Number
age
20 Charge Transfer for Potassium Biphthalate on Inconel 36
21 Charge Transfer for Sodium Chloride on Titanium 37
22 Charge Transfer for Sodium Chloride on Platinum 38
23 Time Dependence of Charge Transfer from Sodium Chloride on
Titanium 39
2k Same as Figure 23 with Different Prior Probe Treatment.... Uo
25 Time Dependence of Charge Transfer from Sodium Chloride on
Platinum Ul-
26 Same as Figure 25 with Different Prior Probe Treatment \2
27 Fluidized Bed Aerosol Generator ^5
28 Pressure Drop Across Fluidized Bed U6
29 Size Distribution of Aluminum Oxide Particles from the
Fluidized Bed U8
30 Size Distribution of Aluminum Powder Sample 1*9
31 Concentration vs. Time of Aluminum Aerosol. 50
32 Size Distribution of Aluminum Aerosol • 51
33 Contact Charge and Particle Diameter vs. Time for Aluminum
Aerosol 51
3^ Charge Transferred from Aluminum Particles vs. Induction Voltage.. 5^*
35 Contact Charge vs. Time After Restart 55
36 Slope of Charge Transfer Curve vs. Time 55
37 Data from Second Sample of Aluminum 56
38 Contact Charge of Aluminum vs. Velocity 57
39 Contact of Elastic Sphere with Surface During Impact 59
vii
-------
TABLES
Number Page
1 Contact Charge for Sodium Chloride Particles Impacting at
71 m/s on a Stainless Steel Probe 22
2 Charge Transfer Data for Methylene Blue Particles Impacting at
71 m/s on a Stainless Steel Probe 22
3 Contact Charge Dependence on Particle Diameter 30
U Contact Charge in yC/g for Particle Velocity 75 m/s 33
5 Charge Transfer for Particle. Velocity 75 m/s U3
viil
-------
ACKNOWLEDGMENTS
We appreciate the interest, encouragement and suggestions contributed
by John Nader, project officer for this work.
One of us, G. Reischl, on leave from I. Physikalisches Institut,
University of Vienna, was supported in part by International Research
Fellowship 1F05TW022T1, National Institutes of Health.
IX
-------
SECTION 1
INTRODUCTION
The present project was undertaken in order to critically examine the
basis for the indirect measurement of particulate matter mass concentration
by means of contact electrification. The contact electrification monitor (1-6)
offers some important advantages such as measurements in real time with a short
response time. It is a mass monitor in that the electrical charge correlates
accurately with the mass determined gravimetrically. (U,7,8,9) The instrument
is relatively simple and the electrical signal is convenient for monitoring
and data processing. The probe can be inserted directly into a stack to avoid
the complications associated with extraction and external analysis.
In monitoring instruments based on contact electrification, a flow of the
aerosol is directed at a probe. The transfer of charge from particle-probe
collisions results in a current which is continuously monitored with a
sensitive meter. Contact electrification, (10) which refers only to charge
transferred between two bodies as a result of pure contact, will be used here
to describe the charging process for the monitor, although sliding or rubbing
may be present. Little is known about the latter charging mechanisms. We
emphasize that no electrical potential is applied to the probe. The charge
is transferred as a result of the contact.potential between the two materials
involved.
Although it has undergone over a decade of development, the contact
electrification monitor is not widely used. This may be due to uncertainty
arising from a lack of definitive data on the performance of the instrument
and poor understanding of the charge transfer machanism. However, because
of the attractive advantages of this type of instrument, there is continuing
interest in its development. For an account of the development of the theory
and application of contact electrification to monitoring the reader is
referred to the comprehensive review by John, (ll)
j£
In earlier work, John (9) tested the IKOR Air Quality Monitor in the
laboratory with a variety of test dusts. The dynamic response was found to
track well with that of an optical particle counter with the exception of an
initial startup period of some 10 to 20 minutes during which the sensitivity
increased towards the steady-state value. The total charge correlated well
with gravimetric mass, as is required of a mass monitor. Humidity below
saturation had no effect on the instrument. The electrical resistivity of the
particle material was found to have a major effect on the sensitivity. Another
important factor determining the sensitivity was the condition of the probe
surface.
Manufactured by New IKOR, Inc., Gloucester, Mass. 01930.
-------
The earlier vork identified several areas needing further investigation.
These included the influence of the probe surface condition on the sensitivity
and the dependence of the contact charge on particle size. More generally,
it was evident that there was a need for better understanding of the basic
physical processes involved in contact charging so that the inherent limita-
tions of the application to particulate monitoring could be assessed. The
present project was designed to address these questions. First, preliminary
measurements were made with the IKOR Monitor and monodisperse laboratory
aerosol. Then apparatus and techniques were developed to allow investigation
of the transfer of charge between particles and a probe surface under carefully
controlled conditions. A theory was developed for the interpretation of the
experimental results. Finally, the implications of the results for the
application to monitoring were considered.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
This investigation has provided extensive new experimental information on
charge transferred by the impaction of particles on metal surfaces. A
theoretical model has been found to account semi-quantitatively for the data.
Thus a basic understanding of the charge transfer process has been achieved
which can be applied to the development and assessment of the contact
electrification monitor.
Experimental determination was made of the dependence of contact charging
on particle size, impact velocity and electrical resistivity of the particle
material. Contact charging was explored for insulating and conducting particles
impacting on several different metals. The transfer of precharge was also
investigated. The influence of the probe surface condition was explored.
The probe surface preparation procedure and the impaction probe design
can be applied directly to practical monitoring. Implications of the various
experimental and theoretical findings for monitoring have been discussed in
detail. The monitor can produce a continuous relative measure of the mass
flow. However, it must be calibrated for each source and recalibrated'if the
sampled material changes. The main uncertainty in the application of the
monitor concerns the types of materials which can be sampled.
Future progress, both experimental and theoretical, on the basic, charge
transfer process will be difficult owing to the dependence on the unknown
(and normally uncontrollable) properties of the particle and probe surfaces.
The inherent limitations of the monitor restrict its usefulness to certain
sources. Only field studies on specific sources can determine where the
instrument can be applied.
-------
SECTION 3
TESTS OF THE IKOR MONITOR
EXPERIMENTAL METHOD
The purpose of these tests was to investigate factors affecting the
response of the IKOR Monitor to monodisperse laboratory aerosol. The IKOR
Model 206 Portable Particulate Monitor with the standard bullet-shaped Inconel
sensor was tested with monodisperse particles produced by a Berglund-Liu
vibrating orifice aerosol generator. The experimental arrangement is shown
in Figure 1. Particles from the aerosol generator were passed through the
radiation from a Kr-85 source for charge neutralization. The aerosol then
traversed an electrical mobility analyzer having parallel brass plates
12.7 cm x 31.8 cm,spaced 2.5 cm apart. Because the flowrate of the IKOR
Monitor ranges up to 1^ liters per second, clean dilution air was provided
with a HEPA filter. The particle concentration and size distribution in the
manifold were monitored with a Glimet 201 Optical Particle Analyzer.
.For some of the tests, the IKOR Monitor probe was inserted directly into
the manifold. For other tests, the IKOR sensor and blower were separated as
shown in Figure 1, to allow .the placement of a Faraday cup between them. The
Faraday cup consists of a 130 mm diameter high efficiency glass-fiber filter
mounted in a metal housing. The housing is insulated by Teflon bushings and
surrounded by a grounded metal shield. Charges stopping on the filter may
flow directly to the metal housing or cause an equal charge to flow according
to Gauss' Law. The current was monitored with a Keithley Model 6l6 Digital
Electrometer. The arrangement of the IKOR components shown in Figure 1 is
similar to the normal instrument configuration when the in-line filter is
used for gravimetric determinations except that the filter is electrically
insulated to permit the total current to be monitored.
RESULTS
Static Charging in the Teflon Hose
The IKOR Monitor is equipped with a flexible hose which connects the
sampling probe to the sensing unit. The tube is lined with Teflon. A time-
varying charging effect was traced to the flexible hose. With a steady
aerosol flow, the charging effect caused a slow drift of the currents
observed with the IKOR electrometer or the Faraday cup. Since this effect
interferes with measurements, the flexible hose was replaced with a metal
tube. The experience with the Teflon tube indicates the desirability of
using grounded metal tubing to sample aerosol for charge measurements.
Response to Gaseous Ions
Using particle-free air from the aerosol generator and a total flowrate
of lU.l«s~l, a current of 2.10~12A was observed with IKOR Monitor and
2.7'10~11A on the Faraday cup. When the Kr-85 source was removed, the IKOR
-------
Compressed
Air
ELECTRICAL MOBILITY
ANALYZER
Over-
flow
Room Air
HEPA
FILTER
Exhaust <
Figure 1. Experimental arrangement used for testing the IKOR monitor.
-------
Monitor current was undetectable and the Faraday cup read 1-10 12A. Thus, the
IKOR Monitor responds 'to the gaseous ions created by the Kr-85 radiation. A
net charge on the air can be produced by unequal losses of ions of opposite
sign to the walls caused by differing diffusion constants.
The possibility that gaseous ions could be removed from the air stream
by a transverse electric field was investigated. Particle-free air from the
aerosol generator at a flowrate of 100 1-nT1 was passed through the Kr-85
source radiation. A d.c. voltage was applied in steps to the plates of the
electrical mobility analyzer and the current monitored on the Faraday cup
which was placed directly after the mobility analyzer.
Surprisingly, the current was observed to increase from 0.6'10~1:1A at 0 V
to a maximum of 2.6-10~11A at 20 V. Thereafter, the current decreased to
zero at 63 V.. At higher voltage, a negative current appeared. These obser-
vations indicate that the ions consist of a complex mixture of molecular
species having a range of electrical mobilities. Removing ions of one sign
causes an increase in the apparent current of the opposite sign.
Although the observed complications indicate need for caution in the
setting of electric field intensity, a suitable field can greatly reduce the
ion current. Since the monitor has been shown to respond to ions, this pre-
caution is necessary.
Response to Precharge
Some tests were conducted with the Kr-85 source removed. Methylene blue
particles were observed to cause negative currents in the IKOR Monitor and
the Faraday cup. As a result of observations taken with and without the Kr-85
neutralizer, it was believed that the IKOR Monitor was responding to particle
precharge (charge on the particles prior to contact with the probe). However,
because of a number of uncertainties, the tests were not conclusive.
As a result of the data obtained in later experiments (described below)
it is now known that the precharge on the methylene blue particles from the
spraying process in the aerosol generator is considerably greater than the
contact.charge. Therefore, the IKOR Monitor was probably responding to the
precharge.
IKOR Sensor Efficiency
Some measurements were made to ascertain the approximate efficiency of
the IKOR sensor. One factor determining the efficiency is the probability
that particles will impact on the probe. This was investigated by coating
the probe with Dow Corning silicone vacuum grease to make the particles
stick after impaction. Methylene blue particles were used so that the mass
could be quantitated by dissolving the dye in alcohol-water solution and
quantitating on a spectrophotometer. The total particle mass in the aerosol
was measured by the use of an in-line glass fiber filter and quantitating. by
the same procedure. For H.5 pm diameter particles, 5$ of the mass was found
on the sensor. For 1.5 ym particles, 9% was on the sensor.
-------
Other data were provided by a comparison of the currents from the IKOR
Monitor and the Faraday cup operated in series. A typical result is the
following: for 3-9 um methylene "blue particles generated with the Kr-85 source
removed and at a flowrate of lU 1-s"1, the IKOR current was 1.2-10~13A; the
Faraday cup current was 8.0'10~13A.
These tests can be summarized by stating that the -IKOR sensor efficiency
for 1-5 ym particles is approximately 10%. This is not surprising since the
streamlined shape of the sensor facilitates the flow of aerosol around the
sensor.
-------
SECTION U
CHARGE TRANSFER EXPERIMENTS
EXPERIMENTAL PROCEDURES
The preceding experiments with the IKOR Monitor revealed the need for the
development of apparatus to allow observation of various charge transfer
effects under controlled conditions. It is desirable to control the charge
on the aerosol test particles. This includes charging or neutralizing the
particles without producing gaseous ions. Further, it is desirable to observe
the charge transfer between particles and probe with each particle colliding
with the probe surface once and only once. An additional requirement is a
probe configuration allowing changes of probe material and surface preparation.
The experimental arrangement developed for these measurements is shown
as a block diagram in Figure 2. Particles from a Berglund-Liu vibrating
orifice aerosol generator (12) could be charged or neutralized with a particle
charger developed for this work. The aerosol then entered a chamber via the
mobility analyzer (the voltage was normally off for this work); the particle
concentration and size distribution were continuously monitored with a Climet
201 Optical Particle Analyzer. The aerosol was sampled from the plenum with
an impaction probe especially designed for this work. The current from
particle-probe charge transfer was monitored with a Keithley Model 6l6 Digital
Electrometer. The aerosol was also sampled with the Faraday cup to determine
the total charge on the aerosol. The flows through the impaction probe and
the Faraday cup were maintained constant by Sierra* mass flow controllers.
Each of the principal components of the apparatus are discussed in more
detail below.
Particle Charger
A technique was developed for the induction charging of droplets from
the vibrating orifice aerosol generator. The induction electrode is shown
in Figure 3 and Figure k. By applying a d.c. voltage to the electrode,
highly uniform charges of either sign can be placed on the droplets. After
droplet drying, the charges are left on the residual particles. A typical
calibration curve is shown in Figure 5 where the charge on 1.5 ym diameter
methylene blue particles is shown to vary linearly with the induction voltage.
The intercept on the y-axis of the graph in Figure 5 represents the charge
placed on the particles in the absence of applied voltage, which we call
spray charge. The figure also illustrates the neutralization of the spray
charge by cancellation with an opposing induction charge. In this case,
approximately -2 V results in zero charge (intercept on x-axis).
Sierra Instruments, Carmel Valley, CA
-------
Compressed *
Air
ELECTRICAL MOBILITY
ANALYZER
Figure 2. Experimental arrangement for charge transfer measurements
-------
Orifice Area
(detail shown\^
in Fig.
To Power
Supply
s
o
t--
Stainless
Steel
•Epoxy
Figure 3. !' Apparatus for induction charging of particles from the
vibrating orifice generator. The dotted lines outline the
vibrating orifice assembly.
V —=
Stainless n .,/-__
Steel 1.26mm
Platinum Orifice Plate
Figure U.
Detail of the area vhere the droplets are formed in
the vibrating orifice generator.
10
-------
Particle charge,
elementary units
•lOxlO3
--5
Induction voltage,
volts
Figure 5.; Typical calibration curve for the particle charger. These
data were taken with 1.5 ym methylene blue particles.
11
-------
The induction technique places a uniform charge on the particles with
almost instantaneous control of the magnitude of the charge. This includes
neutralization of the charge. Moreover, no gaseous ions are produced.
Additional details of the particle charger and some charging effects beyond
the scope of the present work have "been published (13).
Particle Counting and Sizing
The particle sizes quoted here are the physical geometric diameters of
the spherical particles. Particle sizing was "based primarily on calculation
from the known vibration- frequency applied to the orifice, the liquid flow-
rate and solute concentration. This was checked by sizing in a Zeiss micro-
scope equipped with Epi (vertical) illumination. The optical particle
analyzer served as a secondary standard and real time monitor of the mean
size and size distribution. The particles were highly monodisperse, the
geometric standard deviation being of the order of 1.02.
The optical particle counter sampled from the plenum. It was necessary
to calibrate this particle count to the particle concentration sampled by
the impaction probe and the Faraday cup. This was done by counting while
simultaneously collecting particles on a high-efficiency filter in the
Faraday cup. The methylene blue was then dissolved in alcohol-water solution
and quantitated in a spectrophotometer. When a glass fiber filter was used
it was found necessary to separate the fibers in a blender in order to extract
all of the methylene blue. Ultrasonification was also useful. The slurry
was then vacuum filtered through a glass frit to separate the solution from
the glass fibers. When a Mijlipore membrane filter was used, both the filter
and deposit were dissolved in acetone for quantitation. A similar calibration
was carried out with the impaction probe. The particles deposited on the
walls of the impactor were included in the analysis by washing it out with
solvent.
Faraday Cup
A Faraday cup was constructed to measure the charge on the particles.
It consists of a filter surrounded by an aluminum cylinder connected to an
electrometer (see Figure 6). The cup is insulated by Teflon and surrounded
by a grounded aluminum shield. It should be noted that it is immaterial
whether the charge collected oh the filter is conducted to ground. The
instant a charged particle enters the cup, an equal induced charge appears
on the outside of the cup as required by Gauss' Law. This induced charge
is removed and measured by the electrometer.
The Faraday cup shown in Figure 6 is our second model which incorporates
several .improvements over our first model (Figure l). The principal features
are listed below:
1. The parts are rigidly constructed to avoid the induction of currents by
mechanical vibrations.
2. Teflon is used as the insulator. The surface exposed to the aerosol, flow
is kept small to avoid charging effects.
12
-------
AEROSOL IN
TO
ELECTROMETER
SCALE, CM.
-5
ALUMINUM
TEFLON
47mm FILTER
TO PUMP
Figure 6. Faraday cup used to measure the electrical charge on
particles.
13
-------
3. The geometry is chosen to minimize leakage of electrical flux from the
particles on the filter to the outside, and to minimize the pickup of
flux from the outside. Backing the filter with a metal grid also helps
in this regard. A coarse grid could be placed on top of the filter, but
was not found necessary.
h. The taper on the cone leading to the filter was made with a 15° half
angle, resulting in negligible particle loss to the walls. This
simplifies the calibration procedure.
5. The top of the inner cylinder is removable to permit the filter to be
changed. An 0-ring seals to the metal and to the edge of the filter.
Both top and bottom of the inner cylinder are directly connected to the
electrometer.
6. The electrical resistance of the cup to ground was measured to be > 10llt
ohms. Since the input resistance to the electrometer was only 106 ohms
there was negligible leakage from cup to ground.
Impaction Probe
The impaction probe shown in Figure 7, is essentially a one stage inertial
impactor with a rectangular slit. The air flow is maintained well above the
particle cut off so that nearly all of the particles impact on the insulated
probe surface. The top of the probe is beveled to eliminate multiple bouncing
of particles.
The impaction probe shown in Figure 7 was our second, improved model.
The first model was a modified Multi-day* impactor. The slit width is 0.28 mm
and the slit-to-probe distance 0.6 mm. Some important features of the design
are:
1. The impaction efficiency is nearly 100% for particles with diameters
greater than 1 urn for the particle velocities used.
2. Multiple bouncing of particles is eliminated by the beveled probe
surface and the sharpened slit edges.
3. The long taper on the inlet minimized particles losses.
U. The slit and probe assemblies are self-aligned to high precision but can
be easily disassembled for cleaning.
5. The after filter permits a quantitative mass analysis and also micro-
scopic examination of particles after impaction.
6. The interior of the apparatus has been shaped and polished to reduce the
deposition of particles everywhere except on the after filter in order
to facilitate the quantitative mass analysis of particles. This was only
partially successful.
*
Sierra Instruments, Carmel Valley, CA
-------
AEROSOL IN
47mm DIA.
FILTER'
Figure 7-
Impaction probe for charge transfer measurements.
-------
7. The probe was designed to meet several requirements:
a. The beveled upper surface should be highly polished and smoothly
joined to the adjacent insulating and shielding materials.
b. The probe should be electrically shielded except for the upper
surface.
c. The resistance to ground should be greater than 108 ohms.
d. The electrical capacity should be minimized. ">•
e. The probe should be mechanically rigid to avoid the induction of
currents.
f. The area of exposed insulator should be minimized to avoid charging
effects.
g. The probe should be easily removed in order to facilitate cleaning
and the substitution of other probe materials.
As shown in the detail on Figure 7, the probe has a rectangular exposed
surface 0.051 cm wide by 2.22 cm long. It was inserted into a slot in the
brass housing, shimmed with Teflon strips and the remaining space filled with
a high-resistivity epoxy. A wire soldered, to the bottom led to the electro-
meter connection. During the machining process, particles become imbedded in
the surface of the epoxy, lowering the probe resistance to ground. However,
it was found that the imbedded particles could be eliminated by a careful
polishing procedure. Aluminum oxide powders were used with successively
smaller grain size down to 0.05 vm followed by rinsing with high purity
water and baking at 110°C under vacuum.
Probe Surface Preparation—
For this work it is desirable to have a polished probe surface free of
chemical contaminants and mechanical strains. These requirements correspond
to those for the preparation of specimens for metallography. Therefore a
similar procedure was adopted: V
1. The surface is hand polished with Buehler* Micropolish C (l.O micron
alpha alumina) and distilled water using a Buehler polishing cloth
(Metcloth No. U07156). .The surface is rinsed with distilled water.
2. Step 1 is repeated, but substituting Buehler Micropolish A (0.3 micron
alpha alumina).
3. Repeat polishing but with Buehler Micropolish B (0.05 micron gamma
alumina). After rinsing, the surface is checked under a microscope at
6UOX. If scratches are visible, steps 2 and 3 are repeated.
*
Buehler, Ltd., Evanston, Illinois
16
-------
U. The surface is rinsed thoroughly with distilled water to remove all the
alumina powder, and then scrubbed vigorously with a wet tissue paper
(Kimwipe). It is then rinsed with water followed by isopropyl alcohol
(reagent grade, filtered through 0.2 ym Millipore membrane). The probe
is then baked for 15 minutes at 70°C under vacuum to remove the alcohol.
5. After the probe is replaced in the apparatus, it is flushed with clean
dry air for at least 20 minutes. This step appears to remove static
charge effects from the insulator material surrounding the probe which
has been affected by the cleaning procedure. The flushing also restores
the probe to room temperature.
6. During subsequent use, if it was determined that the probe needed
cleaning, it usually sufficed to repeat steps k and 5.
Probes were prepared of stainless steel, Inconel, titanium and platinum.
The leakage resistance to ground, measured with a Keithley electrometer,
ranged from 1011 to > 1013 ohms. The background current with clean air was
5 to 10-10~15A. When tested with methylene blue particles it was found that
most of the particles bounce from the surface. Microscopic examination of
the probe surface after exposure showed that only a few percent of the
particles stick to the surface. For particles less than H ym in diameter,
the deposition on the walls of the impactor was less than 15% of the total,
the remainder being deposited on the after filter.
Properties of the Particles Generated
Three types of particles were used in the present work, methylene blue,
sodium chloride and potassium biphthalate. Methylene blue is an organic dye
substance with both polar and nonpolar properties. Zinc-free methylene blue
was dissolved in 50-50 isopropyl alcohol-water solution for atomization. The
dried particles have a density of 1.3^3 and are smooth, spherical and bouncy.
The dilution air in the aerosol generator was maintained at low humidity,
producing methylene blue particles having a bright golden appearance under
the microscope.
Sodium chloride was atomized from the aqueous solution. The morphology
of the dried particles varied from fairly angular to nearly spherical depend-
ing on the size of the microcrystals in the aggregate particle. Examination
of the deposit on the after filter of the impaction probe revealed the
presence of crystalline fragments of particles indicating that a significant
fraction of the sodium chloride particles break up during the impaction.
Potassium biphthalate was atomized from aqueous solution, producing
smooth, white, spherical particles. When sampled with an inertial impactor
these particles prove to be quite bouncy, i.e., they show very little tendency
to stick to the impaction surface.
MEASUREMENTS
Effect of Probe Surface Preparation on Charge Transfer
The probe surface preparation procedure described above is the final
version developed as a result of preliminary tests. It is instructive to
review the data obtained with various surface treatments. Figure 8 shows
IT
-------
l.Sjjm NaCl
Probe surface:
A untreated
• cleaned
Transfer charge,
elementary units
- -20xlOJ
-.15
Precharge,
elementary units
-5
Figure 8. Charge transferred to the stainless steel probe
as a function of the particle precharge. The
dashed line corresponds to complete charge
transfer. These preliminary data were taken to
investigate surface cleaning effects.
18
-------
charge transfer curves for 1.5 ym diameter sodium chloride particles and the
stainless steel probe. The charge per particle transferred to the probe is
plotted as a function of the precharge on the particles. These charge
transfer curves are nonlinear. Interestingly, the curve for the untreated
probe surface shows that positive charge is more readily transferred than
negative, i.e., there is a rectifying action. To investigate the possible
role of surface contaminants, the probe was cleaned with isopropyl alcohol
and Kimwipes. The charge transfer curve for the "cleaned" probe is symmetric
with respect to charge sign, although charge transfer in the "forward"
direction (positive charge) was lowered by the cleaning. The contact charge
is simply given by the y-intercept of the charge transfer curve which
corresponds to zero precharge. We note that the contact charge for Nad on
stainless steel is positive; further, the contact charge was relatively
unaffected by the surface treatment in this case.
These observations establish the importance of the surface condition of
the probe for charge transfer. Even the mild cleaning had a large effect on
the charge transfer. The cleaning removed some surface contaminants (the
rectifying action suggests an oxide layer) but also apparently left a film
inhibiting some charge transfer. Thus, the investigation of charge transfer
cannot be separated from the question of probe surface condition. The
transfer of precharge, however, provides a sensitive tool for the assessment
of the surface condition. Additional data were taken to investigate various
probe treatments. Some results are shown in Figure 9 and Figure 10 for
methylene blue particles. The curves are nonlinear and vary considerably.
We note that the contact char.ge for methylene blue is affected by the surface
treatment. Data related to surface treatments are summarized in Table 1
and Table 2.
Finally, a cleaning procedure was devised which gave reproducible results.
This procedure, which involves polishing the probe as described above, produces
linear charge transfer curves. (See Figure 11 and Figure 12) After a polishing,
the probe yields good data for a period of time. The duration of this period
was not determined by the present work, but exceeds several hours. As pre-
viously noted, the transfer of precharge is more sensitive to surface condition
than the contact charge. Thus, some contact charge data taken with only an
alcohol, water, and tissue cleaning were consistent with data taken with a
freshly polished surface. (See Figures 13 and lU).
Currents from Gaseous Ions
The contact charge is defined as the charge transferred to the probe by
a particle which is initially uncharged. In the initial phases of this work,
the particles were neutralized (Boltzmann equilibrium charge distribution) by
passing the aerosol through the radiation field of a Kr-85 source. However,
a current was observed from the impaction probe and Faraday cup even in the
absence of particles. This current was traced to residual ions from the Kr
source. The current varied as a function of voltage applied to the mobility
analyzer (Figure 2), in sign as well as magnitude, indicating the presence of
a complex bipolar mixture of ions having various mobilities. At any given
voltage on the mobility analyzer, the net charge was observed downstream.
It was found that an electric field of 60 V/cm was sufficient to remove
essentially all of the ions. Such a field is too small to remove an appre-
ciable number of aerosol particles. These observations definitely established
that the impaction probe responds to gaseous ions.
19
-------
Transfer charge,
elementary units
- -5xl03
Methylene Blue
3.5um dia.
unpolished probe ..
Precharge,
elementary units
5
Figure 9. Charge transfer curve for methylene blue particles
impacting on an unpolished stainless steel probe show-
ing nonlinear charge transfer and anomalous contact
charge.
20
-------
Transfer charge,
elementary units
Methylene Blue
3.5jim dia.
• probe cleaned
A probe baked
-10
X
X
X
X
lOxlO4
Precharge,
elementary units
Figure 10. Nonlinear charge transfer curves obtained with two different
surface treatments on a stainless steel probe.
21
-------
TABLE 1. CONTACT CHARGE FOR SODIUM CHLORIDE PARTICLES IMPACTING AT 71 m/s
ON A STAINLESS STEEL PROBE
Particle Diameter
ym
Probe Surface
Treatment
Contact Charge
el. chg./particle
1.5
2.5
3.5
5.0
1.5
3.5
Untreated
Untreated
Untreated
Untreated
Cleaned*
Cleaned*
380
1025
2850
5800 .
300
3250
*Cleaned with isopropyl alcohol and Kimwipes
TABLE 2. CHARGE TRANSFER DATA FOR METHYLENE BLUE PARTICLES IMPACTING AT
. 71 m/s ON A STAINLESS STEEL PROBE
Particle
Probe Surface
Contact Charge
Transfer Charge
Diameter , \m
2.5
3.0
3.0
3.5
3.5
2.5
3.5
3.5
3.5
3.5
Treatment el. chg. /particle Precharge
Polished* - ITU 0.0375"
Polished* - 280 0.0300
Polished* - 260 O.OU60
Polished* - 500 0.0108
Polished* - U'TO 0.0300
Unpolished ^ 0
Unpolished | Q
»
Nonlinear
< 100 Nonlinear
av.
0.031
1 0.013
Baked 36 h. - 200 Nonlinear
Chem. cleaned - 150 Nonlinear
Chem; cleaned |Q
< 2UO Nonlinear
*Polishing procedure described in text.
tCleaned with surface active agent, RBS 25, Pierce Co. Rockford, 111.
^Cleaned with RBS 25, isopropyl alcohol, Kimwipes.
22
-------
Methylene Blue
3.5um dia.
polished probe
Transfer charge,
elementary units
--2xl03
- -1
-2
Precharge,
elementary units
-—1
2
Figure 11. Linear charge transfer curve obtained with the stainless
steel probe polished by the procedure described in the
text.
23
-------
Potassium Biphthalate ^
Transfer charge,
elementary units
5xl02
AS.O^m
polished probe
-5
f—I 1-
5xl04
Precharge,
elementary units'
Figure 12. Charge transfer data obtained with potassium "biphthalate
. particles and a polished stainless steel probe.
-------
-103
C
3
><
s
C
d>
0)
I
o
§
o
-10
Methylene Blue Aerosol
Q = 8.4 D3'31
Vp = 85 m s-1
-1
• Cleaned
A Polished
1.0
I
I I I I
10
Particle diameter,
Figure 13. Contact charge vs. particle diameter for methylene blue
particles impacting on a stainless steel probe at tvo
different particle velocities. The lines are fitted to
the data "by least squares. This plot indicates that the
contact charge is less sensitive to probe surface condition
than is precharge.(Fig. 10).
25
-------
104
•H
C
>»
'3
I
0)
» 103
0)
bo
J3
o
§
o
10'
NaCl Aerosol
Q = 124 D2'45
Probe condition
• Untreated
A Cleaned
\
I
I I I I II
1.0 10
Particle diameter, urn
Figure 1)4.Contact charge vs. particle diameter for sodium chloride
particles impacting on a stainless steel probe.
26
-------
Contact Charge Dependence on Particle Size
Data on the effect of particle size on the contact charge are valuable
for understanding of the physical mechanism of the charging process and have
important implications for the application to monitoring. The particle size
dependence was investigated by measuring the contact charge for monodisperse
particles of methylene blue, sodium chloride and potassium biphthalate im-
pacting on the stainless steel probe and for methylene blue impacting on the
titanium probe.
The initial measurements were made with methylene blue particles
neutralized by the Kr-85 source. The current from the impaction probe was.
integrated with the electrometer to obtain the total charge which was then
divided by the number of particles. The latter was determined by dissolving
the deposits in the impactor, quantitating on the spectrophotometer and
dividing by the average volume of a particle. Results are shown in Figure 13
for two different particle velocities and for two different probe surface
treatments. Most of the other measurements were made by neutralizing the
particles with the induction charger or by taking the intercept of the charge
transfer curve. Particle number was based on the optical particle count with
.the calibration determined by spectrophotometry of the deposit. Results are
shown in Figures I*t-l6. The lines were fitted to the data by the method of
least squares. :
Each set of data points is well fitted by a straight line on the log-log
plots indicating that the contact charge depends on the particle diameter
raised to an exponent approximately equal to three. The data are summarized
in Table 3. Judging from the methylene blue data, the exponent does not vary
significantly with particle velocity or probe material. Relative to methylene
blue the exponent is slightly lower for sodium chloride and slightly higher
for potassium biphthalate. However, the accuracy of the data is not sufficient
to conclude that the differences are significant. The measurements are
difficult because of the strong dependence of the contact charge on particle
diameter.
The weighted average value of the exponent for all the measurements except
sodium chloride is 3.0^ +_ 0.08. Therefore the contact charge is essentially
proportional to the cube of the particle diameter or equivalently, to the
particle volume or mass. This result will be further discussed interms of a
theoretical model below. We note that the cube law implies that the contact
charge from sampling a number of particles of the same material will be
proportional to the total mass of the particles, independent of the particle
size distribution. Thus we have established that on a microscopic (particle
by particle) basis the contact electrification monitor is a mass monitor.
Contact Charge Dependence on Particle Velocity
The particle impact velocity was varied by changing the air flow rate in
the impaction probe. The resulting data for methylene blue particles impact-
ing on the stainless steel probe is shown in Figure IT. The data are well
fitted by straight lines through the origin indicating that the contact charge
is proportional to the velocity. Combined with the preceding result that the
contact charge is proportional to the mass, this implies that the contact
charge is proportional to the momentum of the particle.
27
-------
in -1
a
3
0)
bo
I
w
-p
o
a
-in2
1.0
Potassium Biphthalate Aerosol
Polished probe
Q = 2.6 D3-83
10
Particle diameter, jim
Figure 15. Contact charge vs. particle diameter for potassium
biphthalate particles impacting on a stainless
steel probe.
28
-------
CO
EH
10"
a
w
•l
w
10'
Q=15.9D2-89.
1.0
10
PARTICLE DIAMETER,/(m
Figure 16. Contact charge vs. particle diameter for methylene
blue particles impacting on a titanium probe.
29
-------
TABLE 3. CONTACT CHARGE DEPENDENCE W PARTICLE DIAMETER
Particle
Methylene blue
Methylene blue
Methylene blue
Sodium chloride
Potassium bi-
Probe
Stainless Steel
Stainless Steel
Stainless Steel
Stainless Steel
Stainless Steel
Particle
Velocity
85 m/s
75 m/s
33 m/s
50 m/s
85 m/s
Exponent ,
Std. Dev.
3.31+0.16
2.90 +_ 0.15
2.80 +_ 0.16
2.U5 + O.lU
3.83 + 0.27
phthalate
Methylene blue
Titanium
75 m/s
2.89 +0.16
Weighted average 3.0U +_ 0.08
(Excluding sodium chloride) '
30
-------
xlO2
-12
§-10
-M
1-8
0>
rH
O>
JS
o
-w _4
c
o
o
-2
Methylene blue
• 4.0>im
A 3.5jun
100
Particle velocity, m/sec
n i i i i r i
50
Figure 17. Contact charge vs. particle velocity for methylene
blue particles impacting .on stainless steel.
31
-------
Contact Charge for Several Particle and Probe'Materials
The contact charge was measured for methylene "blue, potassium Mphthalate
and sodium chloride particles impacting on stainless steel, Inconel, titanium
and platinum probes. By using the cube law and the particle density the
results were expressed in terms of charge per mass and listed in Table k.
The stainless steel probe data were taken at 85 m/s and corrected to 75 m/s,
assuming the proportionality to velocity established above.
The data are also shown graphically in Figure 18. In both Table H and
Figure 18 the data are arranged to display the trends. Considering probe
material, the absolute magnitude of the contact charge generally increases
in order from stainless steel to Inconel, titanium and pfatinum. The
difference between Inconel and titanium was small, however. The absolute value
of the contact charge increased generally in the order methylene blue., potas-
sium biphthalate and sodium chloride. The contact charge was always negative
for methylene blue and potassium biphthalate and positive in the case of
sodium chloride except for the anomalous negative value obtained with the
platinum probe. The considerable variation of the contact charge/mass with
materials is evident.
Transfer of Precharge
The transfer of precharge from particle to probe was discussed above in
connection with the probe surface preparation and plots were shown in Figures
11 and 12. Additional examples of charge transfer for various particle and
probe materials are shown in-Figures 19, 20, 21 and 22. All of the data show
linear charge transfer, however, Figures 20, 21 and 22 require lines with
different slopes depending on the sign of the precharge. Additional cleaning
of the probe did not result in equal slopes (single straight line plot as in
Figure 19). Other examples of the type of Figures 20-22 were encountered as
the particle-probe combinations were varied, in fact, it appears that this
type of plot is more common than the single straight line. Figures 20-22
show a systematic effect which was observed in all the measurements which
were not fitted by a single straight line. When the contact charge is posi-
tive (intercept on positive y-axis) the slope is greater for positive
precharge (positive x-axis). Similarly when the contact charge is negative,
the slope is greater for negative precharge. As will be discussed later,
this effect is very similar to the action of a biased p-n junction.
The linear charge transfer curves were obtained after the probe was
polished and cleaned according to the established procedure. However, during
the initial exposure to a given aerosol, the charge transfer varied before
approaching an asymptotic value over a period of 20 to 60 minutes. Some
illustrative data is shown in Figures 23-26. It can be seen that prior
exposure to a different aerosol can drastically affect the early time depen-
dence of the transfer charge. The procedure adopted for taking charge .
transfer data with a given aerosol was to first expose the probe to the aerosol
until the asymptotic region was reached.
For each combination of particle and probe materials the ratio of.
transfer charge to precharge was determined from the slope of the charge
transfer line. The data are summarized in Table 5- Unlike the contact
charge, the fraction of precharge transferred does not appear to depend on
32
-------
TABLE U. CONTACT CHARGE* IN yC/g FOR PARTICLE VELOCITY 75 m/£
Probe
Stainless Steel
Inconel
Titanium
Platinum
Particle Material
Methylene blue
- 2.7 ± 0.1
- 1.1 +_ 0.2
- 3.6 +_ 0.5
- 3.U + 1.6
Potassium biphthalate
- 1.9 +_ 0.1
- 2.7 + 0.6
- 6.9 1 0.3
-20+3
Sodium chloride
+ 8.0 +_ 0.6
+ 25 + U
+ 2k +_ 2
- 75 +5
*Charge transferred to the probe.
33
-------
+20 |_
ss
0 I
-20
-40
-60
-80 L-
—NaCl
Inc
_KBiPh
MB
Ti
MB
KBiPh
-NaCl
-MB
-KBiPh
r-MB
-KBiPh
Pt
I-NaCl
Figure 18. Contact charge/mass for various materials. Particle
materials are: NaCl-sodium chloride, KBiPh-potassium
biphthalate, MB-methylene blue. Probe materials are:
SS-stainless steel, Inc-Inconel, Ti-titanium, Pt-
platinum.
-------
-4xl04 -3
TRANSFER CHARGE,
ELEMENTARY UNITS
PRECHARGE,
ELEMENTARY UNITS
3xl04
Figure 19- Charge transfer for 2.78.urn potassium "biphthalate particles impacting
on a titanium probe.
35
-------
TRANSFER CHARGE,
ELEMENTARY UNITS
- -1x103
-4x104
-3
-2
- -3xlOJ
3xl04
PRECHARGE,
ELEMENTARY UNITS
Figure 20. Charge transfer for 2.78 pm potassium biphthalate
particles impacting on an Inconel probe.
36
-------
TRANSFER CHARGE,
ELEMENTARY UNITS
- - 3x104
— -. O
PRECHARGE,
ELEMENTARY UNITS
-6x10
-2
4x10"
IxlO4
Figure 21. Charge transfer for 3.5 um sodium chloride particles
impacting on a titanium probe.
37
-------
TRANSFER CHARGE,
ELEMENTARY UNITS
PRECHARGE,
ELEMENTARY UNITS
Figure 22. Charge transfer for 3.5/tni NaCl particles
impacting on a platinum probe.
38
-------
20
o
rH
X
& 15
H
W
10
o 5
I
CO
H
I
10
20 30
TIME, MINUTES
40
50
Figure 23. Time•dependence of transfer charge from 3.5 pm NaCl particles with a.
precharge of -56HO elementary charges to a freshly polished titanium
probe. The aerosol generator vas interrupted from 32-37 minutes.
39
-------
20
40
TIME, MINUTES
60
80
Figure 2k. Same as Figure 23 except for the probe exposure history.
Following the Nad run shown in Figure 23, the probe was
exposed to potassium biphthalate particles for about one
hour and then to Nad particles for 0.5 hour. It was then
cleaned with alcohol and water before starting the above
NaCl run.
-------
I—I
X
w
EH
as;
-10
w
J
w"
c_>
K
-20
-O—
-30
10
20
TIME, MINUTES
30
40
Figure 25. Time dependence of transfer charge from 3-5 ym NaCl particles with
a precharge of -6060.elementary charges to a platinum probe cleaned
with alcohol-water.
-------
TIME, MINUTES
Figure 26, Same as Figure 25 except that the precharge was -61*00 elementary
charges and the probe vas first exposed to potassium biphthalate
aerosol.
-------
TABLE 5. CHARGE TRANSFER FOR PARTICLE VELOCITY 75 m/s
Particle
Methylene Blue
n it
n n
n n
n n
Potassium Bi-
phthalate
ti n
ti n
n it
" "
n n
Sodium Chloride
it n
n n
n ii
n n
Probe-
Stainless steel
Stainless steel
Inconel
Titanium
Platinum
Stainless steel
Inconel
Inconel
Titanium
Titanium
Platinum
Inconel
Titanium
Titanium
Platinum
Platinum
Transfer charge /Precharge
Precharge > 0
0.026
0.031
O.OUl
0.030
0.039
0 . 022
0.023
0.067
0.02k
0.023
0.016
0.56
O.U7
0.60
0.32
0.32
Precharge < 0
0.068
0.031
o.oin
0.053
0.072
0.022
0.055
0.063
0.031
0.023
0.060
0.20
0.2l*
0.30
0.55
0.62
Average
0.01*7
0.031
o.oi*i
0.0^2
0.056
0.022
0.039
O.Q65
0.028
0.023
0.038
0.38
0.36
0.1*5
O.kk
0.1*7
-------
the probe material. The fraction of precharge transferred by methylene blue
averaged over all probe materials is 0.0^3 +_ 0.009; the corresponding value
for potassium biphthalate is 0.036 +_ 0.016. The errors quoted are the
standard deviations computed from the data.. Since the values for the two
particle materials are not significantly different, we form the overall
average for methylene blue, potassium biphthalate and all probe materials
obtaining 0.039 ±. 0.013. On the other hand the fraction of precharge trans-
ferred by sodium chloride averaged over all probe materials is 0.^2 +^0.05.
This is an order-of-magnitude larger and in fact represents a large fraction
of the total precharge. The implication is that the sodium chloride is more
conducting than the methylene blue and potassium biphthalate, probably due
to residual moisture in spite of the low relative humidity (< 5$) of the
dilution air used to dry the particles.
-------
SECTION 5
EXPERIMENTS WITH A FLUIDIZED BED. AEROSOL GENERATOR
THE FLUIDIZED BED
A separate series of measurements were performed with aerosol generated
by a fluidized bed. The object of this work was to produce aerosol particles
for investigations of contact charging with insoluble materials, including
metals, which cannot be used in the vibrating orifice aerosol generator.
The fluidized bed affords a method of generating aerosols from a powder
or dust sample, (l^) The principal problem encountered in-producing such an
aerosol is agglomeration of the particles which adhere to each other with
very strong forces. The fluidized bed aerosol generator consists of a bed of
large particles suspended in a fluidized state by an upward flow of gas. A
powder sample of smaller particles is added to the bed. The smaller particles
become attached to the bed particles and are dispersed throughout the bed.
During bed particle collisions, individual powder particles may-'be dislodged
and become entrained in the gas flow, emerging from the bed to form an aerosol.
Figure 27 is a drawing of the fluidized bed which was machined from
stainless steel and chromium plated for abrasion resistance. The powder is
fed into the bed by a Teflon screw driven by a stepping motor. The rotation
rate is controlled by the pulse frequency from the electronic motor drive
circuit.
AEROSOL OUT
BAFFLE
AIR
'PRESSURE
POWDER
1S-MOTOR
OVERFLOW
Figure 27.
POROUS PLATE
Fluidized Bed Aerosol Generator
-------
Several versions of the feed screw were tried before one was found which
functioned. The problem is sticking of the feed screw caused by the powder.
The best results, were obtained with a loose fitting feed screw and by applying
air pressure on the powder container.
Clean air enters the bed via the plenum at the bottom. A porous plate of
sintered stainless steel is clamped between the bottom plenum and the upper
cylinder. The sintered plate supports the 150 ym nickel bed particles which
initially filled the space up to the overflow tube. The conical baffle at the
top is designed to prevent bed particles from leaving the cylinder.
As the air flow rate was increased, the pressure drop across the bed as
measured with a manometer increased linearly until a break point was reached
at about 2^.5 LPM, indicating the onset of fluidization (Figure 28). An
operating flow rate of 30 LPM was chosen. At this flow rate, the bed has
expanded approximately 10$ vertically and the top -surface resembles that of
a gently boiling liquid with bubbles rising to the surface.
40
30
E
O
2 20
Q
CO
w
os
10
24.5 LPM
20 40
FLOW RATE, LPM
60
Figure 28. Pressure drop across the fluidized bed
1 as a function of air flow rate. The
break in the curve at 2U.5 £/min indi-
cates the onset of fluidization.
-------
When the fluidized bed is initially operated with a fresh charge of nickel
particles only, a high concentration of small particles (more than 101* particles
per liter with a peak size less than 0.3 ym) is observed in the output air.
Several days of continuous operation are required to reduce the concentration
to a tolerable background level. Examination of filter samples showed that the
particles are fragments of the nickel bed particles. Considerable variation
in the nickel particle shape has been found in two different batches of powder
from the manufacturer (Sherritt-Gordon Mines Limited, Alberta, Canada). One
batch contained particles which were relatively smooth and spherical, while.a
second batch had particles with a more crystalline angular appearance. The
second batch produced much more background aerosol.
Aluminum Oxide Aerosol
The initial tests of the fluidized bed were conducted with aluminum oxide
powder labelled 3.0 micron, 99-98$ pure alumina, agglomerate-free by the
manufacturer (Adolf Meller Co., Providence, Rl). The size distribution observed
with the Climet optical counter is shown in Figure 29- The peak is slightly
below 0.5 ym and most of the particles are smaller than 1 ym in diameter
(optical size). According to the manufacturer, the aluminum oxide particles
are grown from seed crystals in the alpha form using a high temperature vapor
process. The particle size with which the sample is labelled is said to
conform to an industry convention for this product. The actual mean particle
size is claimed to be about 1.5 ym- Examined under a microscope, the powder
appears to consist of aggregates of small crystals which are individually a
few tenths of a micrometer in diameter. It is not possible to decide whether
the particles are loosely aggregated or strongly bonded on the basis of
microscopic examination. However, since aluminum oxide is an extremely hard
substance, individual crystals will probably not be broken by nickel collisions
in the fluidized bed. The observations can be explained by assuming that the
powder is deagglomerated in the fluidized bed and that the actual mean particle
size is considerably smaller than determined by the manufacturer.
Unfortunately, the aluminum oxide particles produced were too small for
use in the contact charging experiment since the impaction probe has a 50$
particle size cutoff at 0.6 ym (aerodynamic diameter). The fluidized bed
appears to be capable of generating a usable aluminum oxide aerosol if a
properly sized powder sample could be obtained. The scarcity of suitable
powder samples appears to be one of the principal difficulties in using the
fluidized bed as an aerosol generator.
Aluminum Aerosol
Samples of aluminum particles were obtained through the courtesy of the
Alcan Aluminum Corporation, Albany, CA. Produced by atomizing from the melt,
the particles are spherical. The aluminum is 99-5$ pure, with Si (0.2%) and
Fe (0.2%) the principal impurities. The material contains traces of Ni, Cu
and Mn with some A12C>3 also present. Most of our measurements were made with
sample X-T1 which has a narrow size distribution peaking at 3 ym, according
to the manufacturer's measurement with a Coulter Counter. However, our own
measurements made with a microscope (Figure 30) show a broad size distri-
bution peaking below 1 ym.
-------
5
xlO3
OS
W
cx
CO
EH
0.3
J I
0.5 0.7
DIAMETER, fjm
0.9
Figure 29. Size distribution of aluminum oxide particles from
the fluidized bed aerosol generator.
-------
100
s
g 80
W
I
t—(
I
o.
oi
w
60
40
20
468
PARTICLE DIAMETER./m
10
12
Figure 30. Size distribution of the aluminum particles in the
powder sample before it was placed in the fluidized
bed.
1*9
-------
A few grams of aluminum powder were added to the .fluidized bed and the
concentration monitored with the optical counter as a function of time after
startup. As shown in Figure 31, the concentration falls off rapidly at first
and more slowly as time progresses. The initial concentration is so high
that a single charge of fresh powder can "be used for several hours without
running the feed screw.
The size distribution of the Al aerosol measured with the optical counter
is shown in Figure 32. This result was surprising since it has no resemblance
to the initial size distribution of the powder [Figure 30). The size distri-
bution of the aerosol has a fairly narrow peak. Moreover, the peak position
was relatively constant over a period of 5 hours (Figure 33). In the first
30 minutes, the mean particle size decreased by 20% to U0$. Thereafter, the
mean size was observed to increase or decrease only slightly, depending on the
run.
1.5
xlO5
g
OH
OS
ffl 0.5
•z-
60
120
180
TIME, MIN.
Figure 31
Figure 31. Concentration of aluminum aerosol from the
fluidized bed as a function of time after the
aluminum was added to the bed.
50
-------
10
X103
33
o
a
CO
D = 3.5ym
AD= 1.3pm
3 456789 10
DIAMETER, /im
Figure 32. Size distribution of aluminum particles generated by fluidized
bed, measured with the optical particle counter.
a -e
o w
-4
OE-
el "2
o
8 I
ce
2 g
«
0.
100
200
300
TIME, MIN.
Figure 33. Variation of contact charge and mean particle diameter with
time for aluminum aerosol from the fluidized bed. The air
flow through the bed was interrupted at 76 minutes and 2UO
minutes. The increase of contact charge with a long time
constant is attributed to increased conductivity of the
surfaces of the aluminum particles due to grinding in the
fluidized bed.
51
-------
A sample of the aluminum aerosol was taken on a Nuclepore filter. Examina-
tion of the sample under an optical microscope revealed a surprise—the aluminum
particles were disk-shaped. The particles were subsequently studied'with the
higher resolution afforded by a scanning electron microscope. The elemental
composition of the particles was checked by x-ray fluorescence. The aluminum
powder used to generate the aerosol was found to be initially spherical with
a wide range of particle size.
The aluminum particles on the Nuclepore filter were found to be thin disks
with diameters in the^H-6 ym range. The disk shapes were remarkably regular;
some of them had scalloped edges. Some large irregular particles were seen
which may have been deposited early in the run before stable .conditions were
established.
The flat shape of the aluminum particles obviously implies that the
aluminum is deformed during collisions between the large, hard nickel bed
particles. We postulate that the remarkably uniform disk shape and uniform
size can be explained by assuming that the circular area of the disks corresponds
to the contact area between the nickel particles. We can test this idea
quantitatively by applying the Hertz theory of elastic deformation to a
collision of two spherical nickel particles with a small aluminum particle
between them. We assume that the velocities after impact are zero. (The
conclusions are unchanged even if the particles rebounded with nearly their
initial velocities.) Taking the diameter of the contact area of the nickel
particles to be U ym, the initial velocities of the nickel particle are
calculated to be H5 cm/sec or 8$ of the terminal falling velocity. This
seems fairly reasonable. The pressure on the contact area is 1.1 X 1010
dynes/cm2. The yield strength of annealed aluminum ranges from 1.22 X 108
dynes/cm2(99.996$ Al) to 3.^5 X 108 dynes/cm2 (99-0$ Al). Therefore, the
calculated pressure exceeds the yield strength by a factor 3.0 or more. We
conclude that the postulated model of the disk formation process is quanti-
tatively consistent with the observations. Evidently, the aluminum is
worked on during the collisions, small particles are amalgamated and large
particles broken up so that the emergent aerosol particles have a size
distribution determined by the fluidized bed and not by the aluminum feed
material.
CHARGE TRANSFER MEASUREMENTS WITH ALUMINUM AEROSOL
Preliminary experiments showed that the aluminum aerosol from the fluid-
ized bed is uncharged. This was a somewhat surprising finding. The background
nickel aerosol observed during the clean-up phase of the fluidized bed gives
a positive contact current to the stainless steel probe. As a precaution
for subsequent work the Kr-85 neutralizer was placed just before the impaction
probe. .
The contact charge observed from the aluminum aerosol (Sample X-71) and a
stainless steel probe is shown in Figure 33. The decrease during the first
30 minutes is probably related to the decrease in mean particle size during
that time and possibly to changes in particle shape. Subsequently, the contact
charge increases, approaching an asymptotic value. The air flow in the bed
was interrupted briefly at 76 minutes and 21*0 minutes. The contact charge
dropped sharply, and then increased rapidly following start-up. These
observations can be explained by assuming that the aluminum particles are
52
-------
coated with an oxide layer. In operation, the working of the aluminum brings
fresh metal to the surface. The higher conductivity increases the observed
contact charge. Evidently the oxidation is a rapid process, affecting the
contact charge after only a few minutes of interruption. The magnitude of
the asymptotic contact charge is about 10,000 elementary charges, much
larger than that obtained for insulating materials. These results agree
with the dependence of the contact charge on resistivity observed pre-
viously by John. (9)
The transfer of induced charge was also observed. Because the particles
are initially uncharged, a new method was devised to charge the particles
by induction, namely, by placing a positive potential on the slit which is
opposite the grounded impaction probe. The particles then contact the probe
in the presence of an electric field which superimposes an induced charge
on the contact charge. When the induction voltage is varied, a charge
transfer curve is observed as shown in Figure 3^. The straight line shows
that the oxide layer is not thick enough to cause a rectifying action.
Although the data for Figure 3^ were taken after the bed had operated for
about 200 minutes, data taken for earlier times also yielded linear charge
transfer curves.
Data on the contact charge from another run is shown in Figure 35• The
bed had been previously run, therefore the time shown is after restart, and
the charge increases immediately in agreement with the previous data. For
comparison, g, the slope of the charge transfer curve, is plotted in Figure 36.
This shows that g also increases with running time but it does not approach
the asymptotic value as rapidly as Qc- In our previous work with insulators
(particles from the vibrating orifice aerosol generator) we observed that Qc
was less sensitive to the surface condition of the probe than was g. The
present observations are similar, except that the oxide layer is on the particles
rather than the probe.
Figure 37 summarizes measurements made with a different sample of aluminum
powder. From an unknown supplier, the particles were large and irregular in
shape but smooth as though derived from a molten state. Although this aluminum
feed material was quite different, the particles produced by the fluidized bed
were closely similar to those previously observed, and the charge measurements
showed qualitatively similar behavior.
Finally, the contact charge was measured as a function of the flow rate
of the air in the impactor (Figure 38), using the X-1T aluminum. Because Qc
increases with time, the measurements were cycled from low flow rates to
high and back down. Within the errors the data are again in agreement with
a linear dependence of Q,, on particle impact velocity.
53
-------
-6
i
as
s
-7
-8
-9
£ -io
CO
u
-11
-12
Qc = -9075 [e]
/3 = 6.36[e]/V
I
-600 -400 -200 0 200 400
SLIT VOLTAGE, VOLTS
600
Figure 3k. ; Charge transferred from aluminum particles to
a stainless steel probe as a function of
induction voltage.
-------
-15
, xlO'
O
I
O
O
-5
j I
I I
I
I I
60
120
TIME, MIN.
180
240
Figure 35 • Contact charge for aluminum particles on stainless steel
vs. time after' restart of the fluidized "bed shoving a
rapid initial rise.
240
Figure 36. Slope of the charge transfer curve, B, vs. time for the
same run as Figure 35•
55
-------
« 6
w
§ 4
20
Diameter
-15
h-"
O
-5
40
60
TIME, MIN.
80
100
120
w
o
t-H
§
a
«
S'
u
H
§•
O
Figure 37. Particle size, contact charge and slope of the charge
transfer curve vs. time for a different aluminum feed
material than shown in the previous figures.
56
-------
1
-3
-2
I I
0 10 20 30
FLOW RATE, LPM
Figure 38. Contact charge of aluminum vs. air flow rate in the impaction
probe.
57
-------
SECTION 6
THEORY OF CHARGE TRANSFER
MODEL OF CHARGE TRANSFER PROCESS
The origins of the model to be considered here are in the work by Cheng
and Soo (15) who pointed out that the charge transfer is a dynamic process.
An elastic sphere impacting on a surface will be deformed, a circular contact
area being established during the impact, as illustrated in Figure 39- Cheng
and Soo calculated the charging of the capacity of the sphere by the contact
potential. Later, Cole, et al., (l6) considered that the relevant capacity
is formed by the circular contact area and the effective separation of the
surfaces. Masuda, et al., (IT) employed this model with an additional factor
to account for the incomplete charging of the capacity owing to the resistance
of the material. Both Cole, et al., and Masuda, et al., included, additional
effects which we do not retain here. It will be shown that one.of the effects
is negligible. The other effects are complications from multiple bounces or
space charge which are not present here. We wil.1 selectively develop here
those, parts of the theory which we believe to be applicable. In fact, the
unique data presented here allows the first direct test of the theory for
single impacts of monodisperse particles of known uniform charge.
Theory of Elastic Impact
When a sphere impacts on a plane surface, it can be shown from the
elasticity theory of Hertz that the contact area, A, and the duration of the
impact, At, will be given by (18):
(l) A = ira2a
(2) At = ^-^ aa
5 —
where a = [j- ir2pv2(k + k )]5
" .S
a = particle radius
v = particle velocity
p = particle density
k = elasticity parameter for particle material
k = elasticity parameter for surface material
5
In obtaining the above expression for a we have taken the rebound coefficient
to be one, the impact angle 9 = i , and the surface radius and mass to be
infinite. 2
58
-------
CONTACT ARfiA
Figure 39- Illustration of the contact area and separation
distance Z during the impaction of an elastic
sphere vith a surface.
59
-------
The elasticity parameter k is defined by
TTE
where v = Poisson's Ratio ( =0.3 for most materials)
E = Young ' s Modulus
Transfer of Precharge from Insulating .Particles
The duration of the impact is quite short, as can be seen by estimating
At for methylene blue or potassium biphthalate particles impacting on a metal
probe. We take a = 1.5 ym, v = 75 m/s, k = 9.10"11™. , ks = 1.5 . 10~12^.2.
Then equation (2) gives At = 10 8 sec. Since At is small and the
particle is an insulator, only the precharge within the Hertz contact area
is transferred. We .further assume that the precharge is on the surface of the
particle and the charge within the contact area is completely transferred.
The ratio of the charge transferred, Q^ , to the precharge on the particle, Qo,
is given by:
5t _ A
Q^" ~ Uira*
a
-IT
7T =F t £
^*o
Equation (3) predicts that the fraction of precharge transferred is independent
of particle radius, as we observe experimentally.
In order to test equation (3) quantitatively the elastic constant k is
needed. Since k is unknown for methylene blue and potassium biphthalate,
we estimate that k is in the range spanned by that for ionic crystals like
NaCl and plastics like polystyrene, i.e., from 3-10"11 to 9-10"11 2_ .
N
Then taking a = 3 ym, v = 75 m/s, k- = 1.5'10~125L2 (steel)
N
Ck
•2- = 0.02U to 0.037
where the spread corresponds to the range in estimated k. This is to be
compared to the experimental average value, 0.039 +. 0.013, for methylene
blue and potassium biphthalate particles impacting on stainless steel,
Inconel, titanium and platinum. The agreement is very good for the larger
value of k. If k is actually near the lower value, the additional charge
transfer observed experimentally could be explained by surface conduction
from a small distance outside the contact area. It would be necessary to
assume conduction from a distance only 25$ larger than the radius of the
contact area.
60
-------
In the above discussion we have assumed complete transfer of the charge in
the contact area. Hovever, the experimental data shows that for some materials
the fraction of precharge transferred is somewhat greater for positive pre- .
charge than for negative precharge. In these cases, the contact charge is
always positive. When the contact charge is negative the dependence on the
sign of the precharge is reversed. This effect can be interpreted in terms
of a weakly rectifying p-n junction (19) formed at the particle-probe contact.
For positive contact charge, the particle surface material is p-type, the
probe surface material n-type. When neutral particles contact the probe,
electrons flow from the probe to the particle and positive charge (holes) flow
from the particle to the probe until the contact potential is established.
If the particle carries a positive precharge, this produces a positive bias
voltage across the junction, increasing the charge transfer rate. On the
other hand negative precharge will impede the charge transfer. For negative
contact charge the effects are reversed. The rectifying action is considered
to be weak because the slopes of the transfer curves are not greatly different
with change of sign of precharge and the curves are linear.
Figures 23-26 show that the charge transferred varies with time. The
effect was observed with all particle and probe combinations. We can show
that the time dependence is consistent with the assumption that the charge
transfer characteristics of the probe surface are altered by contact with
particles. We must calculate the rate at which the probe surface is covered .
by Hertz contact circles whose centers are randomly located. The result, is
t_
(U) b = 1-e-T
.where b = fraction of probe surface contacted
t = time of exposure to aerosol
T = time constant
The time constant is given by
nA
with S = total area of probe surface
n = number of particles incident per unit time
A = Hertz contact area of a particle
Taking parameter values from our .experiments, n = lO1* particles/s,
a = 1.5 pm, A = 0.03 (^ira2), S = 0.0*1 cm'2, we obtain T = 8 minutes.
This is a quite reasonable estimate. Since microscopic examination
revealed that very few particles stick to the surface, we must assume that the
changes produced by contact originate in transfer of minute amounts of
material and, possibly, some mechanical alterations from the impact. The high
sensitivity to surface impurities again suggests semiconductor effects.
61
-------
For NaCl, ^ ranged from 0.2 to 0.6. This large fractional charge
QQ
transfer probably results from surface conduction produced by adsorbed moisture.
Contact Charge
A parallel plate capacitor is formed by the Hertz contact area and the
separation of the particle and probe surfaces. The separation of the surfaces
will be somewhat larger than the range of repulsive molecular forces owing to
surface irregularities. (l6) These irregularities also mean that the entire
Hertz area is not in molecular contact. However, the unknown surface topography
can be included in an equivalent "effective" separation z. In the following,
z, will be set equal to 10~9m (10 A) which is probably correct within an order
of magnitude. (20)
The contact charge, Qc, will be given as a function of time by:
-At
(5) Qc = CVc (1-e T )
where C is the electrical capacity and T is the charge relaxation time of the
particle material. Vc is the contact potential between the particle and probe
materials. (The probe material will always be taken to be a metal). The
capacity C = e°A , and A is given by equation (l).
z
Therefore
v T -At
(6) Qc = ffedve [ £ TT2pv2(k + k )]• a2(l-e ~T) (for all mate-
~z 4 rials)
Contact Charge for Conductors—
When the particle material is-a metal, At»T, and the charging will be
complete:
(7) Q = ire°Vc [ £ TT2p(k + k )]Va2 (for conductors)
C Z 4 S .
This equation predicts that for conductors, Q^ is proportional to v5 and a2.
No data is available on the dependence of Q on particle radius in the case
of conductors. A comparison to the theory is partially afforded by the NaCl
data. As previously noted, the large fraction of precharge transferred by
NaCl particles indicates the presence of surface conductivity. In addition, •
Qc for NaCl was proportional to a^ ^ — O-14/. The exponent was lower than
the average value of 3-04 +_ 0.08 for the insulating particles, (it will be
shown below that the theory predicts an exponent of 3 for insulators). Thus
the NaCl particles, which are partially conducting, also yield a value of the
exponent intermediate between the prediction for conductors and insulators.
n fl '
The predicted dependence of Qc on v is to be compared to the measured
dependence on v-^-0 for aluminum particles. This is already fair agreement.
It is probable that z, the effective particle-surface spacing will decrease
as the velocity increases. This would increase the velocity exponent, bringing'
the theory closer to experiment. However, the change in z cannot be.predicted
theoretically.
62
-------
We evaluate equation (7) for aluminum particles impacting on steel and the
following parameters: a = 1.5 ym, Vc = 1 V, v = 75 m/s , z = 10~9m. The result is
QC = 2-101+ elementary charges. This compares favorably to the measured value
of I'lO1* elementary charges for (non-spherical) aluminum particles. The agree-
ment is well within the uncertainty in the values to be used for z and Vc which
require a detailed knowledge of the surface for accurate evaluation.
It is necessary to discuss why the contact charge does not depend on the
magnitude of the precharge. This is an experimental fact implied by the
observed linearity of the charge transfer curves. The precharge on the
surface of the particle produces a potential VQ, at 'the contact point which
tends to transfer charge. In computing VQ, the effect of the image charge in
the metal surface of the probe must be taken into account. The result is (l6)
(8) v =
Evaluating this expression for z = 10~9m, a = 1.5 ym and Qo = 103 elementary
charges ,
V = l-10~3volt
«i
Since V is of the order of 1 volt, V0 can be neglected.
c Q
Contact Charge for Insulators —
For particles of insulating materials, T»At, and
Q = CV •
C C T
At
V • -
z c T
The relaxation time T is given by (21)
T = Ke0pp
with K = dielectric constant
p = resistivity
Finally,
(9) Qr = 2;9UirVc [ I Tr2p(k + k ) ]5 v 5 a3 (for insulators)
<- zKp 4 s
P
This equation predicts that for insulators, QC is proportional to a3 as was
experimentally observed. Q^ is also predicted to be proportional to v*-1-"
whereas experimentally Qc is proportional to v^ . As discussed above for
conductors, the agreement with theory would be better if the change of z
with velocity were taken into account. On the other hand we note that the
velocity dependences for conductors and insulators are predicted to be v^'"
and v^'" respectively, whereas both are measured to be v-*-'^.
63
-------
To calculate Qc from equation (9) the resistivity p must be known. In
practice p_j is not known arid the possible range of resistivities is enormous.
Therefore we will reverse the procedure, calculating pp from the observed Qc.
For methylene blue and potassium biphthalate, the average Q for particles
with 1.5 Vim radius was approximately TOO elementary charges. We further
take Vc = 1 volt, z = 10~9m, K = 3, v = 75 m/s. The result is pp = 6-101*
ohm-m. This is an acceptable value for the resistivity since the resulting
relaxation time is T = 2-10~6s. Then we obtain
f = 5-10-3
Therefore the contact time is much smaller than the charge relaxation time
of the material, satisfying the definition of an insulator.
DISCUSSION OF THE THEORY
It has been shown that the theoretical model accounts quantitatively for
the fraction of precharge transferred and for the dependence of contact charge
on particle size. The theoretical dependence of contact charge on particle
velocity and resistivity are consistent with experiment and the calculated
contact charging is of the right order of magnitude. Thus the model can be
used as a guide for future investigations and applications of charge transfer
between particles and surfaces.
The principal limitation to the theory is that the contact potential is
unknown because the surfaces of the particles and probe are insufficiently
characterized even with considerable care in their preparation. The probe
polishing procedure developed during this work is a recipe for the preparation
of a reproducible surface, but the detailed physical and chemical characteris-
tics of the surface are unknown. Similarly the details of the particle
surfaces are unknown, although it may be worth noting that -the generation of
aerosol particles does produce fresh surfaces. The asymptotic increase of
the probe current when the probe is first exposed to aerosol also shows the
extreme sensitivity of the charge transfer to contaminants on the surface.
Thus future progress in understanding the charge transfer process will require
elucidation of the properties of the surfaces involved.
-------
SECTION 7
IMPLICATIONS OF THE FINDINGS FOR APPLICATIONS
PRINCIPAL FACTORS AFFECTING MONITORING
The experimental and theoretical findings presented above can be used for
a critical analysis of the application of contact electrification to monitoring
of particulate matter. The principal factors vill "be discussed separately.
Probe Design Consideration
The charge transferred to the probe also depends on the probability that
the particle impacts on the probe. The impaction probe design developed for
this work ensures that' all particles above the impaction cutoff size will impact
with essentially 100$ probability. It is well known that it is difficult to
employ inertial impaction below about 0.5 urn particle diameter. Possibly
turbulence could be used to cause small particles to impact; this has not been
explored here.
The streamlined probe used in some of the IKOR instruments would appear
to be an ineffective design in terms of achieving high sensitivity and de-
tection of small particles. (IKOR has more recently offered a probe designed
to cause turbulence.) The tubular pipe design of the Konitest causes multiple
particle contacts with the probe. Even for insulators the current will depend
on the particle size distribution. Additional uncertainty is introduced by
the possibility of sliding particle-probe contacts.
It has been shown that the contact charge is proportional to the particle
impact velocity. Thus high velocity is desirable in terms of increased
sensitivity as well as the lowered particle cutoff size for impaction. Instack
probe designs such as that offered by IKOR raise concern that the stack velocity
be sufficiently high and constant.
Dependence of Sensitivity on the Physical Properties of the Particulate Material
The finding that the contact charge varies as the cube of the particle
diameter in the case of insulators means that the probe current will be
proportional to the particle mass flow, independent of particle size. However,
for metallic conductors, the contact charge is predicted to vary as the square
of the particle diameter. The probe current per unit mass will then vary
inversely as the particle diameter. Partial conductors, such as an ionic salt
which is not completely dry will exhibit an intermediate behavior. Therefore,
in general, the probe current will be independent of the particle size dis-
tribution only if all the particles are composed of insulating materials.
The contact charge is strongly dependent on the electrical resistivity
of the particle material. Theoretically, the range of contact charge per unit
mass corresponding to the range of resistivities of all materials would cover
a factor of 101L*. In practice, we have observed a range somewhat less than 10 ,
65
-------
probably because the particle surfaces are never completely dry and clean.
This will also be expected for particles emitted by sources, and the response
of the detector will vary considerably, depending on the particle resistivity.
The contact charge depends weakly on the elastic parameters of the
materials. The theory predicts a maximum variation of the contact charge by
a factor of about 5 from hard materials to plastics.
Complications are introduced by friable materials. It is not possible to
predict with confidence what the contact charging will be in the presence of
particle breakup. Some breakup was present in our Nad experiments which did
not appear to greatly affect,the dependence of the charge transfer on particle
size. The large fractional charge transfer has been attributed to surface
moisture.
Probe Surface Condition
Ample evidence has been presented of the strong dependence of the charge
transfer to the preparation of the probe surface and its history of exposure
to aerosol. The probe preparation procedure which has been presented here
yields reproducible charge transfer. This procedure or one which produces.
equivalent results should be used to obtain reliable monitoring performance.
It is clear, however, that the detailed properties of the.surface are unknown.
Similarly, the properties of the particle surfaces are unknown, so that the
contact potential cannot be predicted. Contact charge can be of either sign.
For a mixture the net charge transferred is detected.. If the composition varies,
the net sensitivity can vary. It is desirable that the monitoring instrument
indicate the sign of the current as an additional useful datum.
RECOMMENDATIONS FOR MONITORING PROCEDURES
The following recommendations are .based on the findings of the present
investigation:
1. Prior to monitoring, the probe surface preparation procedure described
above should be used.
2. The probe should be exposed to aerosol from the source until the current
reaches the asymptotic value.
3. Sticky material should not be monitored since it will coat the probe and
affect the charge transfer.
U. The monitor should not be used to sample material whose composition varies
with time since the sensitivity and calibration of the instrument varies
with the chemical and physical properties of the material.
5. The monitor must be calibrated for each source. Any change in source
conditions will require recalibration.
6. The aerosol should be neutralized because the monitor responds to pre-
charge on the particles. The aerosol is likely to become charged as a
result of mechanical and electrical processes existing within the source.
66
-------
T. Gaseous ions should be removed from the aerosol to avoid collecting
spurious charge.
8. All parts of the monitoring system contacted by the aerosol should be
constructed of metal and grounded to avoid charging effects from particles
and gas exposed to surfaces of insulators.
OVERALL ASSESSMENT OF THE CONTACT ELECTRIFICATION MONITOR
Contact electrification can be used for monitoring in real time with a
short time constant. The current is proportional to the mass flow, but the
instrument must be calibrated for each source. Any change in source conditions
will require recalibration. Because .the response depends on unknown detailed
properties of surfaces the calibration cannot be predicted quantitatively in
terms of fundamental properties of the material sampled. The instrument
furnishes a relative measure of the mass flow. The components of a mixture
will contribute to the total current with varying sensitivity, depending
mainly on the electrical resistivity of the materials. Because the contact
potential can be positive or negative, an increase in the proportion of one
constituent of a mixture can cause the magnitude of the total current .to
increase or decrease. The overall sensitivity will depend on the choice of
probe material. .
For insulating particles the monitor's response is independent of particle
size distribution, provided that the probability of impaction on the probe is
also independent of particle size. Partial or good conductors are detected
with a sensitivity which increases with decreasing particle size for a given
mass emission rate. In practice then, change in particle size distribution
will usually require recalibration. There will be a lower particle size
cutoff in sensitivity determined by the size dependence of the particle
impaction probability.
The principal disadvantages of the monitor derive from the fact that the
charge transfer is a surface phenomenon. Care is required in preparing and
equilibrating the probe. Loading of the probe will require cleaning at
intervals depending on the particular source; it will not be practical to .
sample some sticky materials. The monitor's performance is vulnerable to
subtle changes in the properties of the material sampled.
Within the inherent limitations it would appear .that the monitor can be
useful for certain types of sources. The instrument is relatively simple and
low in cost. The electrical signal is convenient for data processing. Trans-
ient events can be monitored which may be useful in detecting malfunctions of
emission control devices, etc. The total emission can be determined by a
simple integration of the current signal. The output is proportional to
gravimetric mass, assuming constant source emission conditions.
67
-------
REFERENCES
1. Schutz, A. Eine Anordnung zur Registrierenden Kontactelektrischen
Staubmessung. Staub 2k: No. 9, 359-363, 196U.
2. Schutz, A. A Recording Dust-Measuring Instrument Based on Electric
Contact, with Logarithmic Indication. Staub 26: No. 5, 18-22, 1966.
3. Prochazka, R. Neueste Entwicklung des auf Kontaktelektrischer Basis
Beruhenden Staubgehaltsmessgerates Konitest. Staub 2h: No. 9, 353-359,
196U.
U. Prochazka, R. Recording Dust Measurement with the Konitest. Staub 26:
No. 5, 22-28, 1966.
5. Schutz, A. Technical Dust Control Principles and Practice. Staub 26:
No. 10, 1-8, 1966.
6. Cheng, L. and S. L. Soo. Charging of Dust Particles by Impact. J. Appl.
Phys. Ul: 585-591, 1970.
7. IKOR, Inc. , unpublished reports, and A. H. Gruber, private communication.
8. Sennitzler, H. Messtand fur die Prufund und Kalibrierung von Registrie-
renden Staub-und Gasmessgeraten in einem Steinkohlengefeuerten Kraftwerk.
SchrReihe Ver. Wass-Boden Lufthyg, Berlin-Dahlem, V. 33, Stuttgart, 1970.
9- John, W. Investigation of Particulate Matter Monitoring Using Contact
Electrification. Environmental Protection Agency, Research Triangle Park,
NC, Technology Series Report Number EPA-650/2-75-0^3, February, 1975, ^5 pp.
10. Harper, W. R. Contact and Frictional Electrification. Oxford U. Press,
Oxford, 1967, 369 pp.
11. John W. Contact Electrification Applied to Particulate Matter-Monitoring.
In: Fine Particles, Aerosol Generation? Measurement, Sampling and Analysis,
B.Y.H. Liu, ed., Academic Press, NY, 1976, pp 6^9-667.
12. Berglund, R. N. and B.Y.H. Liu. Generation of Monodisperse Aerosol
Standards. Envir. Sci. and Tech. 7:1^7-153, 1973.
13- Reischl, G. , W. John and W. Devor. Uniform Electrical Charging of
Monodisperse Aerosols. J. Aerosol Sci. 8:55-65, 1977.
lU. Guichard, J. C. Aerosol Generation Using Fluidized Beds. In: Fine
Particles, Aerosol Generation, Measurement. Sampling and Analysis,
B.Y.H. Liu, ed., Academic Press, NY, 1976, pp. 173-193-
68
-------
15. Cheng, L. and S. L. Soo. J. Appl. Phys. hi: 585-591, 1970.
16. Cole, B. N., M. R. Baum and F. R. Motts. An Investigation of Electrostatic
Charging Effects in High-Speed Gas-Solids Pipe Flows. Proc. Instn. Mech.
Engrs. 18U, Pt 3C: 77-83, 1969-70.
17. Masuda, H., T. Komatsu and K. linoya. The Static Electrification of
Particles in Gas-Solids Pipe Flow. Unpublished manuscript, Kyoto Univ.,
Kyoto, Japan.
18. Soo, S. L. Dynamics of Charged Suspensions. In: Topics in Current Aerosol
Research, Vol. 2, International Reviews in Aerosol Physics and Chemistry.
Pergamon Press Ltd., Oxford, 1971, p. 71.
19- Malmstadt, H. V., C. G. Enke, S. R. Crouch and G. Horlick. Electronic
Measurements for Scientists. W. A. Benjamin, Inc., Menlo Park, CA, 197^-,
pp. 207-208.
20. Dahneke, B. The Influence of Flattening on the Adhesion of Particles.
J. Colloid and Interface Sci. 1;0: 1-13, '1972.
21. Hendricks, C. D. Introduction to Electrostatics. In: "Electrostatics
and Its Applications, A. D. Moore, ed. John Wiley and Sons, NY, 1973,
p. 26. •
69
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-78-212
4. TITLE AND SUBTITLE
INVESTIGATION OF PARTICIPATE
CONTACT ELECTRICITY
Final Report
7. AUTHOR(S)
Walter John, Georg Reischl,
Jerome J. Weslowski
2.
MATTER MONITOR I
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
Nr, LKiNr, October 1978
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
William Devor, and
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Air and Industrial Hygiene Laboratory
California Department of Health
2151 Berkeley Way
Berkeley, California 94704
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory -
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
15. SUPPLEMENTARY NOTES
Previous Report: EPA-650/2-75-043, February
10. PROGRAM ELEMENT NO.
1AD712 BA-04 (FY-77)
11. CONTRACT/GRANT NO.
R 803719-01-2
13. TYPE OF REPORT AND PERIOD COVERED .
RTP3 I\|P Final
14. SPONSORING AGENCY CODE
EPA/600/09
1975
16. ABSTRACT
To better understand the contact electrification monitor for particulate matter, charge
transfer by aerosol particles impacting on metal surfaces has been investigated. Mono-
disperse, uniformly charged or neutral aerosol particles (1-5 pm diameter) from a
vibrating orifice or fluidized bed generator were bounced on a metal probe. The trans-
fer of precharge from the particles was found to be a sensitive indicator of the probe
surface condition. A surface preparation procedure was developed which yielded linear
charge transfer curves.
Measurements were made of methylene blue, potassium biphthalate, sodium chloride and
aluminum particles impacting on stainless steel, Inconel, titanium, and platinum
probes. For insulating materials, the transfer of precharge was independent of par-
ticle size while the contact charge was proportional to the cube of particle diameter
and directly proportional to impact velocity. The magnitude of the contact charge was
strongly dependent on the electrical resistivity of the material. A theoretical model
was found to account semi-quantitatively for all aspects of the data. A major remain-
ing difficulty is the lack of knowledge of the contact potential. The implications of
these findings for monitoring applications are discussed in detail and the advantages
and disadvantages of the monitor are assessed.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
*Air pollution
*Aerosols
*Particles
*Monitors
*Contact potentials
*Evaluation
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b. IDENTIFIERS/OPEN ENDED TERMS
_
19. SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
c. COSATi Held/Group
13B
07D
21. NO. OF PAGES
80
22. PRICE
EPA Form 2220-V (Rev. 4-77)
PREVl'OUS EDITION IS OBSOLETE
70
------- |