U.S. Environmental Protection Agency Industrial Environmental Research
Office of Research and Development Laboratory
Research Triangle Park. North Carolina 27711
EPA-600/7-78-009
January 1978
PARTICULATE SAMPLING
SUPPORT: 1977 Annual Report
Interagency
Energy-Environment
Research and Development
Program 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 ancf 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 INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems: and integrated assessments of a wide'range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-009
January 1978
PARTICULATE SAMPLING SUPPORT:
1977 Annual Report
by
K.M. Gushing, William Farthing, LG. Felix,
J.D. McCain, and W.B. Smith
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
Contract No. 68-02-2131
Program Element No. EHE624
EPA Project Officer: D. Bruce Harris
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
TABLE OF CONTENTS
Page
Introduction 1
Executive Summary 3
Technical Summary 9
Task Category - Cascade Impactors
Technical Directive No. 10101 - Cascade Impactor
Computer Data Reduction 9
Technical Directive No. 10201 - HP-65/HP-25
Source Sampling Programs 22
Technical Directive No. 10301 - Non-Ideal Cas-
cade Impactor Behavior 24
Technical Directive No. 10401 - Cascade Impactor
Sampling of Charged Particles 42
Technical Directive No. 10501 - Cascade Impactor
Substrate Media Study 90
Technical Directive No. 20101 - Calibration and
Evaluation of Commercial Impactors 94
Technical Directive No. 20201 - Soviet Impactor-
Cyclone Calibration 115
Task Category - Cyclones
Technical Directive No. 10602 - Develop and Eval-
uate Five Stage Series Cyclone System 124
Technical Directive No. 20302 - Calibration of
the Source Assessment Sampling System Cyclones.. 128
Technical Directive No. 20402 - Recalibration
of the Three SASS Cyclones 132
Technical Directive No. 20502 - High Temperature
Recalibration and Modification of SASS Cyclones. 138
-------
Page
Technical Directive No. 21002 - Calibration of
SASS Cyclones for HERL 140
Task Category - E.S.P. Back-Up
Technical Directive No. 10703 - Develop an E.S.P.
Back-Up for Sampling Systems 151
Task Category - Guidelines, Manuals
Technical Directive No. 10804 - Guideline for
Particulate Sampling and Annotated Bibliography. 153
Technical Directive No. 10904 - Technical Manual
on Particulate Sampling 153
Technical Directive No. 20604 - Procedures Manual
for Electrostatic Precipitator Evaluation 154
Task Category - Review of Documents
Technical Directive No. 20705 - Review Documents
and Reports Furnished by EPA 155
Task Category - Consulting Services
Technical Directive No. 20806 - Participate in
U.S.A. - Soviet Information Exchange Program ... 156
Technical Directive No. 20906 - EPA/IERL/PMB
Booth Exhibit at the 1977 APCA Meeting 156
Task Category - Advanced Concepts - Mass and Size
Technical Directive No. 11007 - Evaluation of a
PILLS IV Particle Sizing Instrument 159
References 160
111
-------
LIST OF FIGURES AND TABLES
Figure No. Page
1 - Cumulative mass loading versus particle
diameter for an individual impactor run 11
2 - Differential mass concentration (dM/dlogD)
versus particle diameter 12
3 - Cumulative mass loading versus particle dia-
meter for eight individual impactor runs made
on the same source 13
4 - Differential mass concentration (dM/dlogD)
versus particle diameter for eight individual
impactor runs made on the same source 14
5 - Differential number concentration (dN/dlogD)
versus particle diameter for eight individual
impactor runs made on the same source 15
6 - Cumulative mass concentration versus particle
diameter averaged for a set of runs 16
7 - Percent of mass concentration (cumulative
percent) versus particle diameter averaged for
a set of runs 17
8 - Differential mass concentration (dM/dlogD)
versus particle diameter averaged for a set
of runs 18
9 - Differential number concentration (dN/dlogD)
versus particle diameter averaged for a set of
runs 19
10 - Penetration-efficiency versus particle diameter
for a number of inlet and outlet impactor runs
at a particular control device 20
11 - Penetration-efficiency versus particle diameter
for a number of inlet and outlet impactor runs
at a particular control device 21
12 - HP-65 and HP-25 Programmable Calculator Source
Measurement Booklets 25
IV
-------
Figure No. Page
13 - Composite of calibration data for the Andersen
impactor — stages 2 through 7 .................... 29
14 - An illustration of the four modeled stage
collection efficiency curves of a typical
stage of the Andersen impactor .................. 3d
15 - Recovered size distributions on a cumulative
percentage basis from the Brink impactor models
for ag = 2.0 and MMD's of 1.5, 4.5, 13.5, and
27 pm ........................................... 32
16 - Recovered size distributions on a cumulative
percentage basis from the Brink impactor models
as shown in Figure 15 with the backup filter
catches omitted from the analysis ............... 33
17 - Recovered size distributions on a cumulative
percentage basis from the Brink impactor models
for ag = 3.0 and MMD's of 1.5, 4.5, 13.5 and
27 \im (backup filter included in the analysis).. 34
18 - Recovered size distributions on a cumulative
percentage basis for the Andersen impactor
for ag = 2.0 and MMD's of 1.5, 4.5, and 13.5 ym. 35
19 - Recovered size distributions on a differential
basis from the Brink impactor models for MMD's
of 4.5 and 27 pm and ag's of 2 and 3 ............ 37
20 - Recovered size distributions on a differential
basis from the Andersen impactor models for
MMD's of 1.5 and 13.5 ym and ag's of 2 and
3 ............................................... 38
21 - Charged particle generator orifice region with
the parallel plates for charge measurement ...... 45
22 - Schematic of the charged particle generator and
sampling arrangement ............................ 46
23 - Particle Charge Versus Charging Voltage as De-
termined by Methods I-IV.
24 - Particle Charge Versus Charging Voltage ......... 54
25 - Control sampling run using MRI-Model 1502 Impac-
tor wi th no charge ......... . .................... 56
-------
Figure No. Page
26 - Sampling charged particles using MRI-Model 1502
Impactor with no grounding wire—high particle
charge 57
27 - Sampling charged particles using MRI-Model 1502
Impactor with grounding wire—high particle
charge 58
28 - Sampling charge-neutralized particles using MRI-
Model 1502 Impactor with neutralizer at Nozzle -
n = 4 x 1011 59
29 - Sampling charge-neutralized particles using MRI-
Model 1502 Impactor with Nozzle Losses Corrected... 60
30 - Sampling charged particles using MRI-Model 1502
Impactor with grounding-wire moderate particle
charge 61
31 - Sampling charge-neutralized particles using MRI-
Model 1502 Impactor with neutralizer at Nozzle -
n = 2.4 x 10' 62
32 - Sampling charged particles using MRI-Model 1502
Impactor with grounding wire - moderate particle
charge 63
33 - Sampling charge-neutralized particles using U. of
W. Mark III Impactor with charge neutralizer at
Nozzle - n = 4 x 101* 64
34 - Sampling charged particles using U. of W. Mark III
Impactor with no grounding wire - high particle
charge 65
35 - Sampling charged particles using U. of W. Mark
III Impactor with grounding wire - high particle
charge 66
36 - Sampling charged particles using U. of W. Mark III
Impactor with grounding wire - moderate particle
charge 67
37 - Sampling charge - neutralized particles using
Andersen Mark III Stack Sampler with neutralizer
at Nozzle - n = 7 x 103 68
P
VI
-------
Figure No. Page
38 - Sampling charged particles using Andersen Mark
III Stack Sampler with grounding wire - moderate
particle charge...«. o 69
39 - Sampling charged particles using Andersen Mark
III Stack Sampler with grounding wire - moderate
particle charge ...... 70
40 - Photographs of deposition patterns of ammonium
fluorescein particles in MRI-Model 1502 Impactor.... 74
41 - Photographs of deposition patterns of ammonium
fluorescein particles in MRI-Model 1502 Impactor.... 75
42 - Photographs of deposition patterns of ammonium
fluorescein particles in Andersen Mark III
Stack Sampler 76
43 - Photographs of deposition patterns of ammonium
fluorescein particles in Andersen Mark III
Stack Sampler 77
44 - Photographs of deposition patterns of ammonium
fluorescein particles in Andersen Mark III Stack
Sampler 78
45 - Sampling charged particles using MRI-Model 1502
Impactor with plexiglass jet plates, 2J, 3J, and
4J 79
46 - Reference curve (solid) giving collection efficiency
as a function of /4> for the Andersen Stack Sampler
and neutral par tides. 81
47 - Efficiency versus /"ijj of stages in Andersen Stack
Sampler for neutral particles (solid curve) and
moderately charged particles (dashed curve) 83
48 - Reference curve (solid) giving collection efficiency
as a function of /4> for the MRI-Model 1502 Impactor
and neutral particles. 84
49 - Efficiency versus /vj; of stages in MRI-Model 1502
Impactor for neutral (solid curve), moderately
charged (dashed curve), and highly charged parti-
cles 85
VII
-------
Figure No. Page
50 - Reference curve (solid) giving collection
efficiency as a function of the /\|; for the
U. of W. Mark III Impactor and neutral par-
ticles ......................................... 86
51 - Efficiency versus /ijJ of stages in the U. of W.
Mark III Impactor for neutral (solid curve),
moderately charged (dashed curve) , and highly
charged particles .............................. 87
52 - Flow chart for acid wash treatment of glass
fiber filter material .......................... 95
53 - Theoretical impactor efficiency curves for rec-
tangular and round impactors showing the effect
of jet-to-plate distance S, Reynolds number Re,
and throat length T ............................ 100
54 - Schematic representation of the Vibrating Ori-
fice Aerosol Generator ......................... 101
55 - PSL calibration system for high and low flow-
rate impactor .................................. 103
56 - Total wall loss vs. Particle diameter .......... 105
57 - Collection efficiency vs. /ijJ . Andersen Mark
III Stack Sampler with glass fiber collection
substrates ..................................... 106
58 - Collection efficiency vs. /4J . Brink Model
BMS-11 Cascade Impactor with glass fiber col-
lection substrates ............................. 107
59 - Collection efficiency vs. /4> . Brink Model
BMS-11 Cascade Impactor with greased collection
plates ......................................... 108
60 - Collection efficiency vs. /ijJ . MRI Model 1502
Inertial Cascade Impactor with greased collec-
tion plates .................................... 109
61 - Collection efficiency vs. /4> . Sierra Model
226 Source Cascade Impactor with glass fiber
collection substrates .......................... 110
62 - Collection efficiency vs. /4> . Sierra Model
226 Source Cascade Impactor with glass fiber
collection substrates .......................... Ill
Vlll
-------
Figure No. Page
63 - Collection efficiency vs. /\p . University
of Washington Mark III Source Test Cascade
Impactor with greased collection plates 112
64 - Soviet 14-Stage Cascade Impactor 116
65 - Soviet 3-Stage Impactor/Cyclone 116
66 - Collection efficiency vs. Particle diameter.
Soviet 3-Stage Impactor/Cyclone 118
67 - Collection efficiency vs. Particle diameter.
Soviet 12-Stage Cascade Impactor. Data shown
for the first nine stages from 2-20 microns... 119
68 - Collection efficiency vs. Particle diameter.
Soviet 12-Stage Cascade Impactor. Data shown
in combined form in the configuration for
field measurements 120
69 - Collection efficiency vs. Particle diameter.
Soviet 14-Stage Cascade Impactor (Small Cut-
points) 122
70 - Collection efficiency vs. Particle diameter.
Soviet 14-Stage Cascade Impactor (Large Cut-
points) 123
71 - Five Stage Series Cyclone System 125
72 - Laboratory calibration for the Five Stage
Series Cyclone System 126
73 - Collection efficiency vs. Particle diameter.
Large SASS Cyclone 129
74 - Collection efficiency vs. Particle diameter.
Middle SASS Cyclone 130
75 - Collection efficiency vs. Particle diameter.
Small SASS Cyclone 131
76 - SASS cyclone cut points 133
77 - Collection efficiency vs. Particle diameter.
Large SASS Cyclone 135
78 - Collection efficiency vs. Particle diameter.
Unmodified Middle SASS Cyclone 136
79 - Collection efficiency vs. Particle diameter.
Modified Middle SASS Cyclone 137
IX
-------
Figure No. Page
80 - Collection Efficiency - Temperature Relation-
ship 142
81 - D50 -Viscosity Relationship 143
82 - Collection Efficiency - Particle Density Re-
lationship 144
83 - Collection Efficiency at 400°F, 4 SCFM 145
84 - Exxon SASS Cyclones 148
85 - A sketch of the electrostatic precipitator
design 152
Table No.
1 - Tentative Project Schedule, Contract No. 68-02-
2131 4
2 - Outputs From Scheduled Tasks, Contract No. 68-02-
2131 8
3 - Simulation Conditions of the Modeled Impactor
Per f ormance 27
4 - Percent Errors in AM/AlogD, Andersen Impactor 39
5 - Percent Errors in AM/AlogD, Brink Impactor 40
6 - Percentage of Trial Cases in Which Recovered Value
of (AM/AlogD) is Within the Indicated Factor of
the True' Value 41
7 - Average Values and Standard Deviations of Charging
Parameters Observed by Reischl, e_t al12 48
8 - Grouping of Surfaces for the Efficiency of Each
Stage 88
9 - Cascade Impactor Calibration Study—Operational
Parameters 97
-------
Table No. Page
10 - Laboratory Calibration of the Five Stage Series
Cyclones - Values of Dso cut points (in micro-
meters) for various conditions of sample flow,
temperature, and particle density 127
11 - Five-Stage Cyclone Calibration Data 141
XI
-------
INTRODUCTION
The scope of the research, development, and support to
be done for the EPA under this contract encompasses every as-
pect of particulate sampling in gaseous process and effluent
streams. Specific objectives which have been identified and
given priority in the technical work plan are:
1. Identify current and future requirements for particu-
late sampling - the nature of the particles (shape,
volatility, concentration, size distribution, charge,
etc.), the sampling conditions (temperature, pres-
sure, entrained fluids, etc.), and the goals of the
sampling programs (control device evaluation, health
effects, etc.).
2. Design prototype high and low flow rate impactors
which incorporate all that we have learned about their
fundamental behavior and operational problems. Con-
tinue research on the non-ideal or unmodelled behavior
of impactors. Investigate problems in using impactors
on nonroutine process streams (wet, hot, high pres-
sure) .
3. Evaluate cyclone systems for particle sizing. Develop
a theory for calculating cyclone D5o's at stack condi-
tions. Consider alternatives to filter backups on
high flow rate systems.
4. Study alternatives to impactors and cyclones for par-
ticle sizing. Possibilities are optical, electrical,
or hybrid systems. Concentrate on devices which offer
the possibility of real time, automatic, sampling and
analysis.
5. Study and evaluate mass monitors. Again the emphasis
should be on new systems with high flow rates or auto-
matic features.
6. Study methods of Quality Assurance in sampling and
calibration programs.
-------
7. Generate and review documents on particulate sampling
and continually update our bibliography and literature
survey.
8. Continue to study and evaluate new ideas, techniques,
and instruments for particulate sampling.
9. Attend and organize meetings and symposia on particu-
late sampling.
10. Provide consulting and support to EPA programs.
Considerable progress has been made in research related
to inertial particle sizing techniques, and less emphasis will
be placed on the development and operation of such devices af-
ter the current tasks have been completed. An exception will
be the special problems of sampling with inertial devices in
high temperature, high pressure process streams.
A second area of special interest is the development of
an advanced technique (perhaps optical) to obtain particle size
information in real time. We are starting now to develop the
necessary background to initiate a research program to develop
or improve instruments for real-time, instack sampling.
There are two main sections in the remainder of this re-
port. The Executive Summary is a review of the work accomplish-
ed to date, and also outlines our projected schedule for the
remainder of the contract. The Technical Summary contains a
description of the Technical Directives issued to date under
this contract. The scope of work for each task is presented,
followed by a summary of the work accomplished to date on each
task.
-------
EXECUTIVE SUMMARY
Under this Term Level of Effort contract, Southern Research
Institute is working toward the completion of several relatively
small tasks, generally less than 2000 man-hours, rather than
one single goal encompassing the total number of contracted
man-hours. The Project Officer at his discretion issues Tech-
nical Directives which specify the scope of the tasks to be
performed and the level of effort. The current tasks are group-
ed under eight Task Categories. These categories are indicated
in Table I, the Tentative Project Schedule, which outlines our
areas of research and the status of each task. Our projected
schedule for some future tasks is also shown.
Through the end of June, twenty Technical Directives had
been issued; twelve tasks have been completed, and eight tasks
are still in progress. Of these twenty tasks, seven come under
the Task Category of Cascade Impactors, five under Cyclones, one
under E.S.P. Back-up, three under Guidelines and Manuals, one
under Review of Documents, two under Consulting Services, and
one under Advanced Concepts - Mass and Size. Tasks numbers
beginning with a 1 (10000) are categorized as research and de-
velopment projects, and those beginning with a 2 (20000) are
categorized as support service projects. At present the 20
Technical Directives are split evenly between these two cate-
gories. A brief Status Report on each task is given below.
Task Category - Cascade Impactors
Cascade Impactor Computer Data Reduction (10101) - This task
is still active. A report including computer program listings
will be issued in several months. It will be applicable to
the Andersen Mark III, Brink, MRI Model 1502, and University
of Washington Mark III cascade impactors.
HP-65, HP-25 Calculator Programs For Source Sampling (10201) -
This task has been completed. Two reports have been issued,
each containing twenty-two programs to facilitate source sam-
pling measurement calculations: "HP-65 Programmable Pocket Cal-
culator Applied To Air Pollution Measurement Studies: Stationary
Sources," EPA-600/8-76-002 (October 1976), and "HP-25 Programmable
Pocket Calculator Applied To Air Pollution Measurement Studies:
Stationary Sources," EPA-600/7-77-058 (June 1977).
Non-Ideal Cascade Impactor Behavior (10301) - This task has
been completed. Non-ideal behavior of both the Andersen Mark
III and Brink impactors was studied. The results were presen-
ted at the 1977 APCA Meeting in Toronto, Canada.
-------
Table 1. Tentative Project Schedule, Contract No. 68-02-2131
TASK CATEGORY
01 CASCADE IMPACTORS
02 CYCLONES
03 E.S.P. BACKUP
04 GUIDELINES. MANUALS
05 REVIEWING DOCUMENTS
AND TEST PLANS
06 CONSULTING SERVICES
4OX OF EFFORT
07 ADVANCED CONCEPTS
MASS AND SIZE
08 MASS MONITORS
1976
DEC JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Co
Imp
—
Ibrot
ic tor
B
1 >
^•65.
Corr
put*
Data
1
Reductior
1 1 1
25 Programs for Air Pollution
r 1 1 1 1 I
on and Evaluation of Commercla PrototvP
Col It
Guid
ration
iinei
of S
•J
L<
viet
<
vol 1
I
^
c
mpac
DC
>—
Docu
» i
III1
Nonideal Bohov or
largod Particle Sampl
Substrata Study
tort i
isign.
— 1
•-
ment
nd C
LJ
yclon
n
Construct, Eval
1 -
-t>
^
SAS
, HP-
no
5
1977
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
f
1 ' 1
jate 5-Series System
Trai
n Col
J
brat)
I
r
ins
IHT
e
HP-2
I
V
HP System
' Construct
^ HERL SX
C
yclon
e Cal
Des gn. Construct, Lab Evaluate
E.S.P. Eval. Proc. Manual!
Gu
T.C
idelir
e ond
Ann
o to tod Bib
llogra
phy
Ca
= rw!f
ibret
3«V«tf
r
on S
ms T
tudln
rst
1
, HTI-
r
i i
•xxor
r
P. S<
Mini
vrot F
' PM,
5m
ss
iratio
)lant
rogra
~
B-IV E
Cti
n
B.g
arged
Poly
dlspo
Dei
Theory
F eld
house
SO At
gn, C
Cy
Evaluation
J
Eval. Pro
Scrub
roso
onstr
clone
bar E
.. Manual
Partlculato Sampling
Test
m, A
n
»CA E
1
«hib
n
. i i
Optical and Au
A
valua
HonT
t, at
n
r1
i
to mated U
r_J
1 1 .1
Select, Assemble
Prototypes
i i
JAN
Sam
uct
for
val. f
trafine
t_
Wine
Com
1978
FEB MAR APR MAY JUN JUL AUG SEP OCT NOV
ling
-uglt
roc.
—
vo SA
Menu
>S
1 i
1 1 .1
•nerd
>l Ins
A
trumanls
1
Eval
jate Is
ew
mpac
tors
—
KEY: •TASK INITIATED. A ESTIMATED COMPLETION DATE. A LATEST ESTIMATE, • TASK COMPLETED
LIKELY CONTINUATION
-------
Cascade Impactor Sampling of Charged Particles (10401) - This
task is still active. A final report is being written which
describes the results of sampling two and five micrometer dia-
meter particles charged to levels near and much greater than
those normally found in electrostatic precipitation control
devices.
Cascade Impactor Substrate Media Study (10501) - This task has
been completed. A final report has been published which pre-
sents the results of a laboratory and field evaluation of over
twenty greases and six types of glass fiber substrate media.
The title of this report is "Inertial Cascade Impactor Substrate
Media For Flue Gas Sampling," EPA-600/7-77-060 (June 1977).
Calibration and Evaluation of Commercial Impactors (20101) -
This task has been completed. A report has been written dis-
cussing the calibration of the Andersen Mark III, Brink, MRI
Model 1502, Sierra Model 226, and University of Washington
Mark III Cascade Impactors. This document is entitled "Particu-
late Sizing Techniques For Control Device Evaluation: Cascade
Impactor Calibrations," EPA-600/2-76-280 (October 1976).
Soviet Impactor-Cyclone Calibrations (20201) - This task has
been completed. Using monodisperse aerosols, two 14-stage,
multiple hole, Soviet impactors, one twelve-stage Soviet im-
pactor, and one three-stage Soviet impactor-cyclone were cali-
brated.
Task Category - Cyclones
Develop and Evaluate Five-Stage Series Cyclone (10602) - This
task is still active. A final report is being written. It
will describe the design and development of this cyclone sys-
tem, as well as the results of calibration at several combina-
tions of flow rate, particle density, and gas stream tempera-
ture.
Calibration of the Aerotherm SASS Cyclones (20302) - This task
has been completed. The three SASS cyclones were calibrated
at 4 ACFM and 3 ACFM using ammonium fluorescein and polystyrene
latex particles under laboratory conditions.
Recalibration of Aerotherm SASS Cyclones (20402) - This task
has been completed. The three SASS cyclones were calibrated
at 400°F and 4 SCFM. The results were not acceptable due to
problems with the physical integrity of the calibration aero-
sol.
High Temperature Recalibration and Modification of the SASS
Cyclones (20502) - This task has been completed. Using a new
calibration aerosol, the "three micrometer" cut point SASS
cyclone was calibrated at 400°F and 4 SCFM. A modification
was made to obtain a Dso closer to 3 micrometers.
-------
Exxon-SASS Cyclone Calibration (21002) - This task has been
completed.The middle("three micrometer") cyclone of the
Source Assessment Sampling System used by Exxon Research
Laboratory at a coal-fired power plant in Kentucky was cali-
brated at 450°F and 13 ACFM.
Task Category - E.S.P, Back-up
Develop ESP Back-Up To Replace Conventional Filter (10703) -
This task is still active.Under this task an electrostatic
precipitator is being designed to replace the normal glass
fiber back up filters in high volume sampling systems. The
prototype is currently being evaluated.
Task Category - Guidelines and Manuals
Generate a Guideline Report and Annotated Bibliography on
Particulate Sampling (10804) - This task is still active.
This guideline document will be a concise survey of the methods
and instruments used in sampling effluent gaseous process
streams for particulate matter.
Technical Manual on Particulate Sampling Instrumentation and
Techniques (10904) - This task is still active.This compre-
hensive document will be a technical survey of the methods and
instruments used in sampling effluent gaseous process streams
for particulate matter.
Procedures Manual For ESP Evaluation (20604) - This task has
been completed.A report has been published which describes
the methods and instrumentation necessary to completely evalu-
ate the performance of an electrostatic precipitator. Topics
discussed include ESP operation, mass sampling, particle
sizing, resistivity measurements, gas analysis, and test plan
development. The title of this report is "Procedures Manual
For Electrostatic Precipitator Evaluation," EPA-600/2-77-059
(June 1977) .
Task Category - Review of Documents
Review Documents Furnished by EPA (20705) - This task has been
completed.Under this task the IERL/PMB Cascade Impactor Guide-
lines, the TRW Level 1 Assessment Report, and the GCA Sampling
Protocol To Minimize Errors were reviewed and comments sent to
the Project Officer.
Task Category - Consulting Services
USA-Soviet Information Exchange Program (20806) - This task has
been completed.Sampling equipment and personnel were sent to
the Soviet Union to participate in tests at a scrubber on a
metallurgical process plant.
-------
EPA/IERL 1977 APCA Exhibit Booth (20906) - This task has been
completed.Southern Research coordinated the planning for an
exhibit booth at the 1977 APCA meeting in Toronto, Canada.
Four thousand EPA sponsored research publications were distrib-
uted and several pieces of hardware were displayed.
Task Category - Advanced Concepts (Mass and Size)
Evaluation of the PILLS-IV (11007) - This task is still active.
Under this task a complete detailed evaluation of the theory
and operation of the PILLS-IV particle sizing instrument will
be performed.
Several new topics for research and development and support
services are tentatively scheduled during the latter part of 1977
These include (1) Procedures Manual For Fabric Filter Evaluation,
(2) Evaluation of Commercial and Prototype Mass Monitors, (3)
1978 Particulate Sampling Technology Symposium; and (4) Cyclone
For Fugitive Assessment Sampling Train.
Specific outputs from the tasks issued to date are shown
in Table II.
-------
Table II. Outputs From Scheduled Tasks, Contract No. 68-02-2131
01-CASCADE IMPACTORS
Computer Data Reduction
HP-65, 25 Programs
Non-ideal Impactor Behavior
Charged Particle Sampling
Substrate Study
Design, Construct, Evaluate
New Impsctors.
02-CYCLONES
Develop and Evaluate 5-series System.
Develop and Evaluate HTHP System
03-E.S.P. BACKUP
04-GUIDELINES AND MANUALS
OS-CONSULTING SERVICES
A.PCA Exhibit,
Prototype Evaluation
KLD
TRW
HTHP
SYMPOSIUM
07-ADVANCED SYSTEMS-MASS AND SIZE
OPTICAL SYSTEMS
Program Report, November, 1977.
Handbooks, June, 1977.
Paper, APCA, June, 1977
Paper, Env. Sci. & Tech., April, 1978
Special Report, Recommendations, Oct. 1977.
(Also input to Non-ideal Model)
Substrate Report, June, 1977.
Prototypes, Drawings, May, 1978.
Report on Evaluation, June, 1978.
Prototype, Evaluation, June, 1977.
Better Theory, March, 1978.
Prototype, June, 1977.
Continuing Evaluation.
Prototype, Theoretical Analysis,
Calibration, December, 1977.
E.S.P. Evaluation Procedures Manual, June, 1977.
Baghouse Evaluation Procedures Manual, December, 1977,
Guidelines and Annotated Bibliography on
Particle Sampling, December, 1977.
Technical Manual on Particulate Sampling,
December, 1977.
Information Transfer, June, 1977.
Special Reports, Suggestions
Procedures Manuals, Improved Theory.
May, 1978.
Review of State-of-Art, Recommendations, 1977.
Periodic reports of laboratory research.
Evaluation of PILLS--IV, October, 1977.
Automated System and Data Analysis (0.01-2.0 pm) Schedule at Project Officer Discretion.
No sooner than December, 1977.
OBI-MASS MONITORS
Special Report on Evaluation-Comparison
April, 1978.
8
-------
TECHNICAL SUMMARY
In this section a detailed report on work accomplished
under each task is presented. The twenty tasks are organized
according to the seven task categories.
Task Category - Cascade Impactors
Cascade Impactor Computer Data Reduction
(Technical Directive Number 10101)
Description of Task;
The purpose of this task is to improve the accuracy and
efficiency of cascade impactor data reduction. This will be
accomplished in four parts,
1. A computer program will be written to calculate all
of the necessary variables (gas viscosity, particle
mean free path, slip correction factors, etc.) in
order to obtain cascade impactor cut points. Also
this program will be able to incorporate calibra-
tion constants for each impactor stage. Initially,
this program will be written for the Andersen Mark
III, the Brink, the MRI Model 1502, and the Univer-
sity of Washington Mark III impactors.
2. A curve fitting routine will be developed which can
be used to fit the experimental data from individual
impactor runs.
3. A statistical routine will be written to manipulate
the fitted curves and to calculate differential size
distributions and fractional efficiency curves, all
with confidence limits.
4. A final report will be written documenting the entire
program.
Summary of Progress:
Parts 1, 2, and 3 are complete. A series of computer pro-
grams have been written which constitute a cascade impactor data
-------
reduction package. These programs calculate stage cut points,
cumulative mass distributions and differential mass and number
size distributions (dM/dlogD and dN/dlogD). A modified spline
curve fitting procedure is incorporated into this particle-size
package so that individual cumulative mass distributions can be
fitted and groups of impactor run data averaged. Inlet and
outlet distributions for particulate control devices are used
to calculate fractional efficiency and penetration curves with
50 or 90 percent confidence intervals. Graphs which can be
produced by the computer and plotter include cumulative mass vs.
particle size, dM/dlogD, dN/dlogD and fractional efficiency-
penetration curves. The efficiency-penetration curves can be
plotted on log-log or log probability axes.
Figures 1 through 11 illustrate most of the various graph-
ical outputs which the impactor computer data reduction program
provides. All data are plotted with particle diameter in micro-
meters as the abscissa. Data can be reduced and graphed versus
Stokes diameter or aerodynamic particle diameter. All the
following figures show Stokes diameter.
Figures 1 and 2 show data for a single impactor run.
Figure 1 is a cumulative mass loading plot and Figure 2 is the
derivative with respect to log (particle diameter) of the cumu-
lative mass loading curve, the dM/dlogD plot. Data, derived
manually from impactor stage weights, are shown along with
results from the computer generated curve fit. In Figure 1 the
curve fit is drawn as a solid curve and in Figure 2 the deriva-
tive of this fit with respect to log (particle diameter) is
plotted as an & where calculated.
Figures 3 through 5 are computer graphs where the data for
eight individual impactor runs are plotted on each graph. Each
run is identified by a particular plotting symbol. Scatter in
the data can be clearly seen. Cumulative mass loading, dM/dlogD
and dN/dlogD (differential number concentration) data are shown.
No curve fits are drawn.
Figures 6 through 11 show typical results for several im-
pactor runs which have been averaged together with outliers
identified and discarded. Such data as these rely on curve fits
to make such averaging possible. Fifty percent confidence inter-
vals are shown. Figures 6 through 9 show cumulative mass
loading, cumulative percent, dM/dlogD, and dN/dlogD averaged
results, respectively. Figures 10 and 11 are penetration-
efficiency curves (for a fictitious control device) which may
be presented in two formats, a log probability grid or a log-log
grid.
10
-------
CD
CD
a
en
en
I
u
M
<
HI
:>
U
i
£>-.
10
-1.
H 1 I I Hill 1 1 I I III Ij 1 1 I I Mill
10"1 1O° 101 10s
PARTICLE DIAMETER (MICROMETERS)
Figure 1. Cumulative mass loading versus particle diameter for an individual impactor
run. Data generated from stage weights are shown as circles with crosses.
The computer generated curve fit is shown as a solid curve.
11
-------
104"
a
CD
a
LD
a
a
a
10°'
,**
X
X
X
X
X
H 1 I I Hill 1 1 I I I III! 1 1 I I I lll|
icr1 10° lo1 10s
PARTICLE DIAMETER (MICROMETERS)
Figure 2. Differential mass concentrationfdM/dlogD) versus particle diameter. Data
generated from stage weights are shown as circles with crosses. The deriva-
tive with respect to logfparticle diameter) of the curve fit to the cumulative
mass concentration (Figure 1.) is plotted with an X where calculated.
12
-------
a
a
01
LJ
u
TEST 1-S .TEST S-X .TEST 3-D .TEST 4-O .TEST 5-M .
TEST G-Q .TEST 7-0 .TEST S-Q .
3T101
u
CD 103
ID
102::
0
o
o
*0
Q
E
x ::
CD
tn
J-io
"3
i i i mill i i i mill—i i i mill—i i intiH
10"2 1CT1 ±CP 1O1 1O2
PARTICLE DIAMETER (MICROMETERS)
Figure 3. Cumulative mass loading versus particle diameter for eight individual
impactor runs made on the same source. Only data derived from individual
stage weights are shown.
13
-------
TEST 1-e .TEST E-X .TEST 3-D JEST A-O .TEST 5-* .
TEST 6-0 .TEST 7-e .TEST 8-0 .
MO = 2.35 tt/CC
(MG/DNM3)
a
\
a
-Lu -
ICA
103,
101:
^rP
EJ
: 0 E 0 0
: & B ' |
k (D ° * x ^ *
0 rD ^
* 0 ^^
O ° H
: *»s
! 0
X
1. i i i HIM i i i 1 1 im i — i i i inn 1 — i i i HIM
"lO"5 10"1 10° 101
PARTICLE DIAMETER (MICROMETERS)
Figure 4. Differential mass concentrationfdM/dlogD) versus particle diameter for eight
individual impactor runs made on the same source. Only data derived from
individual stage weights are shown.
14
-------
TEST l-6> .TEST E-X .TEST 3-D .TEST 4-0 .TEST 5-X .
TEST 6-0 .TEST 7-0 .TEST B-0 .
IHJ - 5.5 a/n:
m
lO^ir
lO^ir
u
d
a
a
a
lO9^
lo7
10
5
EJ
tf
0
21
0
H 1 I I HIM 1 1 I I IMH 1 1 I I I MH
1C
1
10°
101
PARTICLE DIAMETER (MICROMETERS)
Figure 5. Differential number concentration (dN/dlogD) versus particle diameter for
eight individual impactor runs made on the same source. Only data derived
from individual stage weights are shown.
15
-------
conn wow-4 HJCTBI
MJRWL
•
§ 10i.
5»- ;
CD
3 :
CD
M 10°-
Q •
S ;
-J
01
en
< i
^ 10 -
•
M
h-
j
D lo-€_
3 :
U
.
i rr3-
RHD = W7 QUCC 1
F- •
• •
**'* 1
• f ^
• 5^
; jiji
I1 :
;
: I
: I
I
flU *
rlO"E
flO"3
T 4-in~^
* $
•
• 0
; I :
4 ,
„ •
A
•
•
••
-
_10-5
*
-io~e
1 1 — 1--I 14 1 U 1 1 1 1 1 U i\ 1 1 — U 1 J »I-U
u
u
^
(H
CD
1— J
a
a
_j
en
en
LJ
M
*
i —
<
2D
2
u
icr1 10°
PARTICLE DIAMETER (MICROMETERS)
102
Figure 6. Cumulative mass concentration versus particle diameter averaged for a set
of runs. Fifty percent confidence intervals are shown.
16
-------
BfltET GRCUP - 4 DUCT Bl 5-e-7G NDOUL
P.E7QWT
99.99T
z
w
U
QL
LJ
LL
UJ
M
h-
3
:i
u
99.953
99.9^
99.8:
98 ^
95:
90^
80^
7O4
604
50^
f
»
•»
•
u
•
^ y
1 /
I- J1
40 -| rf
30 4 jl4
e°f i*1*
10$ .51
V
•
E-=
: 5*
- 5
J
i J
0.5? -1
O.ST
O.OBi
0.01-"
ll
: f
=
: T
ili ii«i
-------
CUTLET GROUP - A fflJCT Bi 5-2-7G NORMAL
RHD = B.?7 GMT
lO3-
JC^.
a
\
CD
•*f^
\~s
a
CD
a
_i
a
a
•
IDS
•
•
• 1
f ,
p
lO"1^ r
ic
-1
10
4f
lo
10C
PARTICLE DI/M^ETER (MICROMETERS)
Figure 8. Differential mass concentration(dM/dlogD) versus particle diameter averaged
for a set of runs. Fifty percent confidence intervals are shown.
18
-------
OUREI GSfflP - 4 OKI Bl 5-Z-76 WHJJAL
(Ml = B.27 GM/CC
m
a
&
u
0.
a
M/DLDGD
a
lO11)
10»j
1 ll\
; \
: -I
lOH
1O7"
;!lf
I
I
I
I
1—i i 11 ml 1—i i i mil 1—i i 11 mi
icr1 10° lo1 io2
PARTICLE DIAMETER (MICROMETERS)
Figure 9. Differential number concentrationfdN/dlogD) versus particle diameter averaged
for a set of runs. Fifty percent confidence intervals are shown.
19
-------
PENETRATIONS
4-M/30-7G.5-1-7E DUCT - Bi
RHD= e.
(_J
z
LJ
wfc^
M
U
M
LL
U.
U
1—
LJ
U
LJ
*r*U
Q_
atj » ±J±I -
99.95:
99. 9 '-
99.8-
99- 5:
99 :
98 :
95-
90-
80-
»
•
-
•
_ —
•
•
• i "
M J ~
x X "
I!T »
llil^
" * 1
•
: i "
'~ I II :
• ? ii •
- ^IT! "^ -
11' "..
.
.
.
•
. —
.
i I I I ft I M I 1 fl 1 1 1 1 1 1 1 1 1 1 IMF
- U«U_L
-0.05
-0.1
-O.E
•0.5
•1
- 2
-5
:10
^20
10"1 10° 101 102
a
M
1-
•4^
C£
h-
LU
U
Q_
l_
2:
U
UJ
Q_
PARTICLE DIAMETER (MICROMETERS)
Figure 10. Penetration-efficiency versus particle diameter for a number of inlet
and outlet impactor runs at a particular control device. Log-probability
grid. Fifty percent confidence intervals are sho wn.
20
-------
a
a
u
LJ
a.
u
u
(Y.
LJ
a.
PENETRATIDN-EFFICIENCY
«B/30-7G.5-i-7G DUCT - Bl
lOW
10°-:
10"1-!:
I
^—i i i im| H—i iinn
10"
10°
H—» ' ' "iii 99.99
U
M
U
M
U.
b
z
LJ
u
CK
Q_
1O5
10s
PARTICLE DIAMETER (MICROMETERS)
Figure 11. Penetration-efficiency versus particle diameter for a number of inlet
and outlet impactor runs at a particular control device. Log-log grid.
Fifty percent confidence intervals are shown.
21
-------
Ordinate and abscissa maximum and minimum values can be
fixed or data regulated. If any point lies beyond an axis
limit on a non-data regulated graph, it is plotted with the
correct ordinate or abscissa value slightly beyond the axis
it exceeds.
Part 4 of the program to develop a computer data reduction
package is currently underway and documentation is proceeding.
The program has been placed into a "version 1" status based on
the work of Parts 1, 2, and 3. A future version is anticipated
which will be more automated and will accommodate any impactor.
HP-65/HP-25 Source Sampling Programs
(Technical Directive Number 10201)
Description of Task;
A library of programs to facilitate calculations associated
with source sampling were compiled for the Hewlett-Packard
HP-65 and HP-25 Pocket Programmable Calculators. This library
contains EPA Reference Methods 1-8, cascade impactor programs,
and others which the Project Officer selected. Two documents
were published in two formats—a normal SV'xll" and a special
5"x7" spiral bound version for field use.
Summary of Progress;
Twenty-two programs for air pollution measurement studies
have been written for both the Hewlett-Packard HP-65 and HP-25
Pocket Programmable Calculators. Each program includes a
general description, formulas used in the problem solution, a
numerical example, user instructions, and a program listing.
A list of the calculator programs follows.
APol-01 Method 1 - Sample and Velocity Traverses for Stationary
Sources
APol-02 Method 2 - Determination of Stack Gas Velocity and
Volumetric Flow Rate (Type S Pitot Tube)
APol-03 Method 3 - Gas Analysis for Carbon Dioxide, Excess
Air, and Dry Molecular Weight
APol^04 Method 4 - Determination of Moisture in Stack Gases
22
-------
APol-05 Method 5 - Determination of Particulate Emissions from
Stationary Sources
APol-06 Method 6 - Determination of Sulfur Dioxide Emissions
from Stationary Sources
APol-07 Method 7 - Determination of Nitrogen Oxide Emissions
from Stationary Sources
APol-08 Method 8 - Determination of Sulfuric Acid Mist and
Sulfur Dioxide Emissions from Stationary Sources
APol-09 Cascade Impactor Operation
APol-10 Impactor Flow Rate Given Orifice AH
APol-11 Impactor Flow Rate, Given Gas Velocity and Nozzle
Diameter
APol-12 Impactor Sampling Time to Collect 50 Milligrams
APol-13 Impactor Flow Rate, Sample Volume, Mass Loading
APol-14 Impactor Stage D50
APol-15 /¥ Calculation - Round Jets
APol-16 /¥ Calculation - Rectangular Slots
APol-17 Cumulative Concentration vs D50 and AM/AlogD vs Geo-
metric Mean Diameter
APol-18 Mean, Standard Deviation, 90/95% Confidence Interval,
Mean ± CI
APol-19 Resistivity and Electric Field Strength
APol-20 Channel Concentrations for the KLD Droplet Measuring
Device (1-600 urn) DC-1
APol-21 Aerotherm High Volume Stack Sampler; Stack Velocity,
Nozzle Diameter, Isokinetic AH
APol-22 Flame Photometric Detector Calibration by Permeation
Tube Technique
Both of the program documents were published in two formats,
The first of these was the normal 8*5 "xll" bound version; the
second was a special, 5"x7" spiral bound version with plastic
laminated covers. The HP-65 document was published under the
23
-------
title: HP-65 Programmable Pocket Calculator Applied To Air
Pollution Measurement Studies: Stationary Sources. (EPA-600/
8-76-002)(NTIS-PB 264284/$5.50).The HP-25 document was pub-
lished under the title: HP-25 Programmable Pocket Calculator
Applied To Air Pollution Measurement Studies; Stationary
Sources^(EPA-600/7-77-058). A photograph of the two spiral
bound booklets is shown in Figure 12.
Non-Ideal Cascade Impactor Behavior
(Technical Directive Number 10301)
Description of Task;
The purpose of this task was to theoretically investigate
the effect of non-ideal cascade impactor behavior on field test
data. The feasibility of compensating for non-ideal behavior
during data reduction was investigated. In addition, studies
were made to determine the "ultimate" accuracy of size distri-
butions measured with impactors, given this non-ideal behavior.
Summary of Progress;
Cascade impactors have become commonly used measurement
devices for the determination of size distributions of particu-
late matter emissions from industrial sources. Data obtained
with impactors are used to characterize emissions from sources,
to determine the performance of particulate control devices,
and in the selection and design of control devices for specific
sources.
Data provided by impactors are of relatively low resolution
and do not permit the exact reconstruction of the size distri-
bution of the aerosol being sampled, even over the limited range
of sizes normally covered by most impactors (approximately 0.5
to 10 vim) . However, little has been done to estimate the
magnitude of the uncertainties, or errors, which are inherent
in the method insofar as they relate to industrial source
emission measurements and determinations of fractional collec-
tion efficiencies of control devices. The study described here
was one with the specific goals of estimating the effects of
two non-ideal operating characteristics of impactors on the data
obtained with them. These two non-idealities are (1) the lack
of step function stage collection characteristics and (2) par-
ticle bounce. Several authors1'2'3 have proposed various
deconvolution procedures which, when applied to impactor data,
would to a large degree, correct for the effect of the finite
slopes of the stage collection efficiency curves. However,
24
-------
NJ
U1
^^?M?c.
^Sfe
/2. HP--65 and HP--25 Programmable Calculator Source Measurement Booklets.
-------
little use has been made of these procedures, primarily because
noise in the data frequently results in oscillatory solutions
with large negative values. In any case, little quantitative
information regarding the magnitude of the errors introduced by
the lack of sharp size cuts in impactors commonly used for stack
sampling has been published. The magnitude of errors intro-
duced by particle bounce has not previously been quantified
although the existence of such errors has been described in the
literature."'5'6'7
Technical Procedures
The approach used in this study was the development of a
computer model of cascade impactor performance. The model was
based on actual impactor performance as measured in a calibra-
tion study of commercially available cascade impactors for
stack sampling. A total of four simulation models were used
for both a Brink impactor in a commonly used modified config-
uration for stack sampling and an Andersen Mark III stack
sampler. The use of glass fiber collection substrates was
assumed for both impactors. Both grease and glass fiber sub-
strates are commonly used for sampling at temperatures below
150°C (300°F) but no satisfactory greases have been found for
use at temperatures over 150°C. Therefore, glass fiber sub-
strates must usually be used for collection substrates at
elevated temperatures.
The first model for each impactor was one having ideal col-
lection characteristics, !.•§_• , step functions from 0% to 100%
collection at the stage DSO'S. (The stage DSQ is that particle
diameter at which the stage has a collection efficiency of 50%.
The Dso is generally used as the characteristic cut off diameter
for particles collected by the stage.) This model was used as
a performance standard against which the remaining three models
could be compared and also provided a basis for checking the
program.
The assumed operating conditions and resulting cut sizes
(Dso's) of the two impactors modeled in the study are given in
Table III. The models of the Brink impactor included a cyclone
precollector which was assumed to have the same performance
characteristics in all three of the simulations other than that
of the "IDEAL Brink." The cylone performance was based on cali-
bration data for a cyclone in common use with the Brink impactor
modified for stack sampling. This cyclone has a collection
efficiency of 100% for particles larger than about 20 ym.
26
-------
Table III.
Simulation Conditions of the Modeled
Impactor Performance
Brink
Temperature, °C (°F)
Gas composition
Particle density, gm/cm3
Flowrate, .alpm (acfm)
Barometric Pressure, mm Hg
Stage/Dso
1
2
3
4
5
6
7
8
177 (350)
Std. air
2.27
1.13 (0.040)
749
9.601
6.182
3.39
1.85
1.23
0.905
0.589
0.1982
Andersen
177 (350)
Std. air
2.27
17.0 (0.600)
749
7
7
,84
,40
4.44
2.87
1.58
0.855
0.449
0.200
1. Stage 1 is a cyclone precollector.
2. Stages 2 and 8 are part of the modifications to the impactor,
27
-------
The second model for each impactor used the actual cali-
bration data for each stage. In this model the stage collection
efficiencies increased monotonically with increasing Stokes1
numbers (increasing particle size) to a maximum value of about
90% to 95%. The efficiencies then decreased for larger Stokes1
numbers to a value of 35% to 40% and remained constant there-
after. A composite of the calibration data for stages two (2)
through seven (7) of the Andersen impactor, which illustrates
the behavior described above, is shown in Figure 13. Data for
States 1 and 8 were offset from the tight grouping of the data
for the remaining stages and hence were omitted in Figure 13
for purposes of clarity in illustrating the behavior trends of
the stage efficiency curves. The data for the Brink impactor
exhibited similar trends. This model, Model 2, is called the
"Normal Bounce" model.
The third model was identical to the second except that
the rollover and decline in efficiency for larger Stokes1
numbers was ignored. Instead, the efficiencies were assumed to
smoothly increase to 100% and remain at that value for increas-
ingly larger Stokes1 numbers. This is called the "No Bounce"
model. The fourth model was also identical to Model 2 with the
exception that the collection efficiencies were assumed to drop
rapidly to a value of 2% for Stokes1 numbers larger than that
at which the collection efficiency reached a peak in the cali-
bration data. This model was termed the "Extreme Bounce"
model.
The use of the same basic collection efficiency curves
for the "No Bounce", "Normal Bounce", and "Extreme Bounce"
models for particle sizes smaller than those for which the col-
lection efficiencies were maximal in the calibration data is
probably a realistic representation of the actual performance
of the impactors in collecting various types of particles.
Rao5 found that impactor collection characteristics for dry
solid particles and oil particles were virtually identical when
glass fiber substrates were used for Stokes' numbers smaller
than those at which the peak efficiency was reached for the dry
solid particles. Beyond this point he found that oil particles
were collected with efficiencies which increased to 100% with
increasing Stokes" number while the efficiencies declined for
the dry particles as a result of bounce. Figure 14 shows an
example of the four modeled collection efficiency character-
istics of one stage of the Andersen impactor.
28
-------
98
90h ^
••
> 70
m 30
i
§10 *
8 • • •
ai
0
H
CO
0.5
0.1
*•
0.0ll - 1 - 1
0.04 0.1 1.0 10.0
SQUARE ROOT OF STOKES NUMBER (DIMENSIONLESS)
3630-027
Figure 13. Composite of calibration data for the Andersen impactor-
stages 2 through 7.
29
-------
100 -
5 10 20
PARTICLE DIAMETER, nm
100
3630-028
Figure 14. An illustration of the four modeled stage collection efficiency
curves of a typical stage of the Anderson impactor. Model 1 is
the ideal behavior model, model 2 is the normal bounce model,
model 3 is the no bounce model, and model 4 is the extreme
bounce model.
30
-------
Results and Discussion
The performance of each model of the two impactors was
evaluated for aerosols having log-normal size distributions
with mass median diameters (MMD) of 1.5, 2.6, 4.5, 7.8, 13.5,
and 27 micrometers and geometric standard deviations, ag, of 2,
3 and 4. Representative results are presented in this report.
Figures 15 through 18 show typical results for the two im-
pactors in a cumulative percentage presentation. It is evident
from all four of these figures that particle bounce severely
distorts the size distributions, especially for aerosols having
large mass median diameters. Figures 15 and 16 show the results
of the simulations for the same impactor and size distributions,
the difference between the two being the omission of the back-
up filter catch in presenting the results in Figure 16.
Comparison of Figures 15 and 16 indicate that omitting the
back-up filter in calculating the cumulative percentages greatly
reduces the distortion resulting from bounce. Comparison of
Figures 15 and 16 shows that increasing the width of the input
size distribution (increasing ag reduces the distortion caused
by bounce although the distortion remains appreciable for the
extreme bounce models at large HMD's.
Figure 18 shows the results from the Andersen models cor-
responding to three of the four cases for the Brink Model shown
in Figure 16. Note that the deviations from the input distri-
bution resulting from bounce are more severe in the Andersen
results than in the Brink. This difference in the severity of
the distortions apparently results from the cyclone precollector
on the Brink which removes most of the larger particles that
are responsible for raising the apparent percentages of fines"
in the recovered size distributions.
It should also be noted that the relative errors in mass
median diameters become increasingly large as the MMD of the
input aerosol decreases. The recovered MMD's were found to be
systematically large for test aerosol MMD's below 10 ym. The
values of Og were also systematically high with larger relative
errors at tne lower values of a
-------
99.8
N
I-
V)
Ui
u
oc
111
a.
LU
U
• NO BOUNCE
A NORMAL BOUNCE
• EXTREME BOUNCE
0.01
PARTICLE DIAMETER,
10.0
3630-029
Figure 15. Recovered size distributions on a cumulative percentage basis from
the Brink impactor models for og = 2.0 and MMD's of 1.5, 4.5, 13.5,
and 27 yjn. The bold lines represent the input distributions.
32
-------
99.8
N
co
-------
99.8
N
i/5
Z
tc.
UJ
UJ
Z
UJ
u
DC
UJ
a.
UJ
S
u
• NO BOUNCE
A NORMAL BOUNCE
• EXTREME BOUNCE
0.01
0.1
1.0
PARTICLE DIAMETER,
10.0
3630-031
Figure 17. Recovered size distributions on a cumulative percentage basis from
the Brink impactor models for ag = 3.0 and MMD's of 1.5, 4.5, 13.5
and 27 \im (backup filter Included in the analysis). The bold lines
represent the input distributions.
34
-------
99.8
UJ
N
-------
= mass concentration of particles retained by the
itn stage
and AlogD; = log —7=;—r
(M s o / j_
The particle diameter is then taken to be the geometric mean of
(D5o)i and (Dso)i-l or
Dg = /(Dso)i x (D50)i_i
Figures 19 and 20 illustrate recovered size distributions
presented in such a manner, together with the input distribu-
tions, for representative sets of Brink and Andersen results.
The results for the "extreme bounce" case are not shown in
Figures 19 and 20 but the values in those cases generally fall
between the "no bounce" and "normal bounce" cases except for
the back-up filters, for which the values were much higher in
the "extreme bounce" case than in the other two cases. Except
for the finest size fractions, represented by the back-up filter
catches, and the fine fraction tails of the low 0g distributions,
the agreement between the recovered values of (AM/AlogD)i
generally lie reasonably close to the input distributions.
However, errors of up to ±35% are not infrequent.
Tables IV and V show the errors, expressed as percentages,
in the recovered values of (Am/AlogD)i for several cases for
each of the two impactors. (For the purpose of calculating
log D and D for the filter catches, it was assumed that the
diameter range covered by the filter was (Dso)e down to MDsoJa-
Although no results for Og = 4 have been shown, the agree-
ment between the recovered size distributions and the input
distributions was progressively better as Og increased and was
quite good in all cases for Og = 4 with the exception of back-up
filter catches when bounce was present.
Table VI shows the percentages of cases in which the
recovered values (Am/AlogD)^ lay within factors of 1.2, 1.5, and
2 of the true value. From these results it appears that the
concentrations of fine particles as measured with impactors can
seldom be taken to be known better than to within a factor much
smaller than 1.5 unless the particles are known to be adhesive
or an effective adhesive coating can be applied to the sub-
strates .
36
-------
2
C3
o
NO BOUNCE
NORMAL BOUNCE
ID"4
io-i
10° 101 10'1 10°
GEOMETRIC MEAN DIAMETER (MICROMETERS)
3630-033
Figure 19. Recovered size distributions on a differential basis from the Brink
impactor models for MMD's of 4.5 and 27 nm and og's of 2 and
3. The bold curves represent the input distributions.
37
-------
10°
3 10'1
ID
o
o
10^
10'3
O NO BOUNCE
A NORMAL BOUNCE
I
MMD = 1.5 jum
ag = 3
MMD = 13.5 /urn
ag = 3
10'1
10° 101 10'1 10°
GEOMETRIC MEAN DIAMETER (micrometers)
3630-034
Figure 20. Recovered size distributions on a differential basis from the Andersen
' impactor models for MMD's of 1.5 and 13.5 IJUTI and og's of 2 and
3. The bold curves represent the input distributions.
38
-------
Table IV. Percent Errors in AM/AlogD, Andersen Impactor
ag =
MMD:
Stage/Error
F
8
7
6
5
4
3
2
F
8
7
6
5
4
3
1.5
144
9
-8
-18
-14
6
568
22
-6
-20
-16
5
No Bounce
2
4.5
56
13
-r!9
-26
-19
-23
Normal Bounce
121000
920
111
9
-18
-27
-20
13.5
18
-20
-30
-28
293
39
-17
-30
3
1.5
22
-14
-11
-10
-12
-7
10
9
44
-12
-11
-12
-12
-7
9
4.5
38
-5
-4
-8
-17
-21
-15
-22
173
8
1
-11
-16
-21
-16
13.5
4
-24
-25
0
-24
1720
103
37
4
-11
-24
-25
NOTE: Values are omitted for stages for which the collected
mass would be too small to be detected in field sampling
programs.
39
-------
Table V. Percent Errors in AM/AlogD, Brink Impactor
ag =
MMD:
1.5
No
4.5
Bounce
2
13.5
27
Stage/Error
F
6
5
4
3
2
1
F
6
5
4
3
2
1
220
-6
-43
-19
-15
-1
88
550
-10
-40
-20
-16
-1
86
220
-35
3
-3
-1.0
-14
238
-8
10
-4
-28
-15
84
-8
-15
Bounce
107
7
-18
33
7
100
3
1.5
24
-37
-32
-9
-4
4
34
41
-39
-29
-10
-4
4
33
4.5
53
-21
-39
-12
-9
-20
-3
123
-24
-33
-12
-10
-18
-4
13.5
13
-40
-5
-3
-23
-14
12
-26
0
-3
-20
-16
27
-37
4
8
-22
-14
-10
17
10
-16
-17
NOTE: Values are omitted for stages for which the collected mass
would be too small to be detected in field sampling pro-
grams .
40
-------
Table VI. Percentage of Trial Cases in Which Recovered
Value of (AM/AlogD) is Within the Indicated
Factor of the True Value
Andersen
Brink
Factor:
1.2 1.5
2.0
Stage/Percent
of cases
F
8
7
6
5
4
3
2
0
57
65
65
75
30
40
(30)
31
71
76
80
100
100
90
(90)
38
79
82
90
100
100
90
(90)
Factor:
Stage/Percent
of cases
F
6
5
4
3
2
1
1.2 1.5
0
35
11
80
74
42
71
33
59
47
95
91
96
72
2.0
56
82
100
100
91
100
100
Andersen table covers all cases
with
MMD = 1.5, 2.6, 4.5, 7.8, 13.5 and
0g = 2, 3
for both normal bounce and no
bounce
Brink table covers all cases with
with
MMD = 1.5, 2.6, 4.5, 7.8, 13.5,
27, and
ag = 2, 3
for both normal bounce and no
bounce
41
-------
There is some evidence,8'9 although it is not conclusive,
that the use of adhesive coatings (greases) on the substrates
may become ineffective as the particulate deposits build up
under the impactor jets. This would result in the same type of
errors due to particle bounce resulting in back-up filter con-
tamination by oversize particles with greased substrates as
has been shown to occur with glass fiber substrates.
Conclusions and Recommendations
From the evidence presented here, it is suggested that
back-up filter catches generally should be omitted from data
presentation when dry, non-sticky particulates are sampled.
Exceptions should be made only if the MMD is smaller than about
2.5 ym. In addition it is suggested that cyclone precollectors
having Dso's somewhat larger than the first impaction stage
Dso be used whenever a non-sticky particulate is sampled. The
use of such cyclones tends to greatly reduce errors due to par-
ticle bounce.
The results of this study were reported at the 1977 Air
Pollution Control Association Annual Meeting in Toronto, Ontario,
as Paper No. 77-35.3, entitled "Non-Ideal Behavior in Cascade
Impactors."
Cascade Impactor Sampling of Charged Particles
(Technical Directive Number 10401)
Description of Task:
The purpose of this task is to investigate the errors intro-
duced in particle-size measurements when sampling charged par-
ticles with cascade impactors. One aspect of this study is to
sample charged aerosols with and without charge neutralizers
situated at the impactor nozzle. If it is found necessary to
employ a charge netralizer for in-stack sampling, then a
collaborative effort will be planned with manufacturers to
develop the needed device.
Summary of Progress;
Introduction
The purpose of this study is to identify the effects of
particle charge upon cascade impactor behavior and to evaluate
methods of eliminating the effects, if they are significant.
To accomplish this goal a charged particle generator of
monodisperse, uniformly charged particles was developed and
calibrated. Then cascade impactor sampling of these aerosols
was carried out under a variety of conditions. The parameters
42
-------
varied from run to run include particle charge, particle size
(2 ym and 5 ym diameters), electrical grounding of the impactor,
and the use of a charge neutralizer. The impactors investigated
in this study were
1. Andersen Mark III Stack Sampler (Andersen)
Andersen 2000, Inc.
Atlanta, GA 30320
2. MRI Model 1502 Inertial Cascade Impactor (MRI)
Meteorology Research, Inc.
Altadena, CA 91001
3. University of Washington Mark III Source Test
Cascade Impactor
(U. of W.)
Pollution Control Systems, Inc.
Renton, Washington 98055
The laboratory sampling phase of the project has been
completed and analysis of the data is proceeding. Current
results and conclusions are presented in this report. Fur-
ther analyses to be performed are described in the last sec-
tion.
The main area where the charge on particles is of interest
is for the case of flyash exiting an electrostatic precipitator
(ESP). For large particles (>2 ym in dia.) and common ESP
operating conditions, the upper limit of charge is that of
saturation in field charging. For a typical field strength
(4 x 103 volts/cm), np, the number of charges per particle, is
given by
np = 834 ap2 (1)
where ap is ths particle radius in micrometers.10 This means
that for particles with a 5 ym diameter np is approximately 5600
and for a 2 ym diameter is approximately 900.
In this study a monodisperse aerosol was produced with a
Vibrating Orifice Aerosol Generator (VOAG). Ammonium fluorescein
droplets were charged, dried, and then sampled with a cascade
impactor. The collection efficiencies of the various impactor
surfaces were then found by washing each surface individually
with an ammonium hydroxide solution and measuring the optical
absorbance of the wash to find the total mass on each surface.
This procedure is similar to that of Gushing, et al.11
43
-------
In this discussion, the method of charging particles and
measuring the charge is described first, then impactor data and
conclusions are presented.
Particle Charging
The induction method of charging droplets produced with a
VOAG was introduced by Reischl, et al.12 The region of droplet
formation of a VOAG is depicted in Figure 21, with the entire
aerosol system shown in Figure 22. Ammonium fluorescein solu-
tion is pumped through an orifice mounted on a piezoelectric
ceramic. Due to the applied voltage of frequency f, the crystal
oscillates, producing mechanical vibrations of the liquid jet
formed at the orifice. This causes the jet to become detached
and form droplets at regular intervals. Droplet formation occurs
at a height L above the orifice. The solvent, ammonia water,
evaporates and leaves spherical particles of the solute, ammonium
fluorescein. The particle radius ap is then given by
aP =
3F C- 1/3
sy
4TTf
(2)
where Fsy is the liquid flowrate in ml per second and C' is the
solute concentration. In this study f was always set at 65 kH
and FSy at 0.201 ml per minute. If the liquid conducts, then
charge can be induced upon the droplet during its formation by
applying an electric field. The expression for the induced
charge per droplet np given by Reischl, et al12 for a highly
conducting liquid is
4ireoa,
"P • "P0 - —S^~ TT v<=
where a
-------
DROPLET
STREAM
3630-012
Figure 21. Charged particle generator orifice region with the parallel
plates for charge measurement.
45
-------
IMPACTOR
UNDER TEST
TO FLOWMETER
AND PUMP
GROUNDING CLIP
TO ELECTROMETER
TO ELECTRICAL
MOBILITY ANALYZER
SHIELDED PLEXIGLAS
COLUMN
COLLECTION
ELECTRODE
DRYING AND
LOFTING AIR
TO WATER MANOMETER
PRESSURE ABOVE AMBIENT
CHARGING
VOLTAGE
SIGNAL
GENERATOR
ABSOLUTE SYRINGE
FILTER PUMP
X
.DRY
AIR
3630:013
Figure 22. Schematic of the charged particle generator and sampling
arrangement.
46
-------
a = constant determined from the geometry of the orifice
region = 0.81 for the device shown,
nn = charge from spraying process, and
^o
= capacitance of a sphere.
Equation (3) is based on the model of the spherical droplet im-
mersed in the electric field with a conducting lead (liquid jet)
connected to it. This expression has not been verified on an
absolute basis; however, the linear dependence of np upon Vc was
substantiated by Reischl, e_t al^12 Thus the measurement of charge
was necessary to calibrate the instrument although ad, L, H, a,
and Vc were measured independently. Comparisons of the predic-
tions of Equation (3) and the charge measurements are given. The
most difficult parameter to measure is L. A microscope (125X)
allowed, visual observation of the jet disturbance at the point
of droplet formation and a rough measurement of L. However, this
measurement cannot be made accurately without a strobe and camera
arrangement .
Reischl, et ajL, 1 2 measured the charge on dry particles. No
measurements were made of the droplet charge. During a single
run no variation in charge per particle could be measured, indi-
cating a variation of less than ±2%, From one run to another,
a variation in np and the coefficient of Vc was observed (see
Table VII.
Charge Measurement Methods
In this study four different methods were developed to
measure the charge per particle in order to provide independent
checks as well as to establish a convenient procedure. Initially
two methods were devised to measure charge on the dry particles
of ammonium fluorescein: (I) sampling with a commercial elec-
trical mobility analyzer13 (EAA) and (II) absolute filtering' of
dried particles on an electrometer electrode. Difficulties were
encountered with the EAA due to wall losses and inability to
sample high concentrations of dry particles. Method II was time
consuming. Because of these difficulties in charge measurement
of dry particles, two new methods were then developed and per-
formed which give information about droplet charge: (III) deflec-
tion in a uniform field with a parallel plate mobility analyzer,
and (IV) collection of undried aerosol droplets on an electrometer
electrode. However, there was the question of whether or not
the droplet charge is altered while drying. Zung and Snead1"
observed such an effect. In this study a set of measurements
were carried out using methods II and IV to evaluate that possi-
bility. These methods are described below, and results and
47
-------
Table VII. Average Values and Standard Deviations
of Charging Parameters Observed
by Reischl, et al12
Solute
Methylene
Blue
Potassium
Biphthalate
Sodium
Number
of runs
12
11
5
n.
-1640
± 360
-5540
± 910
-2640
± 160
Coefficient of
Vc in Eg. (2)
-1830
± 670
-2900
± 700
-1140
± 190
48
-------
evaluations are given in the next section. Method IV turned
out to be the most suitable one for this project, while the
others provide verification and have desirable possibilities
for any future work performed in this area.
Method I; Method I employs a Thermo-Systems Model 3030
electrical aerosol analyzer (EAA). This particular device was
designed to sample particles with diameters less than one micron.
Several modifications were made to eliminate wall losses so that
the electrical mobility of larger particles could be measured.
With this device n is given by
np = Zp6irnap/e C
(4)
where C is the slip correction factor,
ap is the particle radius known from the VOAG operating
conditions,
n is the viscosity of air, and
Zp is the electrical mobility in m/volt-sec. The
mobility is determined by
ZD = K/V = 3.98 x 10
-i»
where K is a constant depending upon the flow and geometry
of the analyzer, and
V is the applied voltage in the analyzer at which all
particles are collected; i^.e., removed from the air
flow so that the analyzer current goes to zero.
Method II; With Method II the charge on dry particles
versus charging voltage was measured with a filtering system in
place of the impactor shown in Figure 22. This method employs
an filter membrane made of silver and connected to an electro-
meter. The ammonium fluorescein aerosol is sampled with the
filter at the same flow rate as an impactor. An electrometer
measures the total charge collected by the filter. After
sampling, washing of the filter with an ammonium hydroxide
solution and spectrophotometry of the wash solution determines
the total mass of ammonium fluorescein collected. In this
method np is given by
Qp
np =
AF 3
aP
(5)
49
-------
where
Q
PAF
aP
Mt
charge collected,
density of ammonium fluorescein,
particle radius, and
total mass collected.
Method III; In Method III the droplet stream passes between
parallel plates as depicted in Figure 21. With a high voltage
VHV applied between the plates, the charged droplets could be
deflected as indicated. The equations of motion for the drop-
lets are:
u
y =
yoj
mU,
1 _ e-bt/m
(6)
V
x bd
0 -
_ -b(t-tB)/m
x =
_ V V HV
bd
t-tB -
1 - e-b(t-tB)/m
where Ux and Uy are the horizontal and vertical components of
the droplet velocity,
t = time of flight, tB = elapsed time at the bottom of
the plates,
U
yo = initial droplet velocity = Fsy/Tr (orifice radius)
m = droplet mass = density of liquid x Fsy/f,
g = acceleration due to gravity,
b = 6-na^r], where r\ is the viscosity of air, and a^
is the droplet diameter, and
d = separation between the plates.
50
-------
With Vuv = 0 (no deflection) and UyQ = 793 m/sec, it was found
that the .droplets rise much higher (y = 15 cm) than the equations
predict (y = 5 cm) using the viscosity of air (182.7 x 10~6 poise)
at standard pressure and temperature. The droplets rise higher
because of the movement of air along with the droplet stream.
Therefore, the droplets experience a lower viscous drag than they
would in still air. An effective viscosity of 57 x 10~6 poise
was determined which gives the correct height of the flight with
vtJV = 0* This value was then used in the determination of np
with YT, YB, VHV/ and d known. The stream of charged droplets
does not spread due to the applied field VHV, thus demonstrating
a uniformity of the charge per droplet within a few percent.
Method IV; Another method (IV) employed to measure droplet
charge collects the droplets onto an aluminum foil electrometer
electrode as they pass through the charging plate. The electrode
is a hollow cylinder with a hole at the bottom for the droplets
to enter. The top of the cylinder is covered with fine mesh
metal netting (pore size, 0.16 mm, and 36% open) to allow air
to flow through. With this arrangement np is given by
nP = fe"
where I = electrometer current,
f = frequency of the oscillator driving the VOAG, and
e = elementrary unit of charge.
The electrode can be inserted and charge measured before and
after an impactor sampling run.
Two variations of this method were attempted in search of
a measurement method which can be used during impactor runs.
An electrode with the same cylindrical shape was formed entirely
of fine mesh metal netting. In one variation, a high voltage
wire was placed along the cylinder axis. This wire, having the
same polarity as the droplets, enhances the collection
efficiency of the netting. With the high voltage wire at zero
volts a large portion of the charged aerosol passes through the
net electrode and is sampled by the impactor. Then, to inter-
mittently measure the particle charge, the voltage on the wire
can be increased to collect the aerosol on the electrode. In
the other variation of this method, the cylinder, made of metal
netting, is used without the high voltage wire. The dispersion
and drying air can be varied to give a high collection efficiency
for the electrode during charge measurement and then changed to
values which reduce the droplet collection of the electrode so
51
-------
that impactor sampling can proceed. Although both of these
variations showed promise, time did not permit sufficient re-
finement to justify their use in the final analysis of impactor
behavior.
Evaluation of the Charged Particle Generator
The behavior of the charged particle generator has been
characterized with the four methods of charge measurement de-
scribed above. The results are depicted in Figures 23 and 24
where np is plotted as a function of charging voltage Vc.
Figure 23 shows the full range of Vc employed, while Figure 24
gives a better view for small np.
The open triangle in Figure 23 depicts the charge measured
on droplets using Method III. The single point giver) was
determined from the average of three Vc-values taken on differ-
ent occasions with VHV = 2000V. The procedure in this measure-
ment was to vary the charging voltage Vc to deflect the stream
to the top of one plate as shown in Figure 21. This value
varied by ±6%. The most likely source of error in this calcu-
lation is the estimation of the effective viscosity in
Equation (6).
The measurements using Method IVr cylindrical electrode
made of foil, are depicted by the solid line in Figure 23. This
line is a least squares fit to Eq. (3) from 33 measurements
obtained in four different runs over a period of two weeks. The
fluctuation of measured np values from this line were low as
judged from the coefficient of determination (= 0.997 where 1.000
is a perfect fit). The value of npo is -3050 ±580 and the co-
efficient of Vc (slope of the line) is 1280 ±20 volt"1. These
results are comparable to those of Reischl, e_t al12 for sodium
chloride (see Table VII).
The measurements on dry particles, denoted by open circles,
were obtained with the modified EAA. These results are more
easily viewed in Figure 24 where the lower quarter of Figure 23
is reproduced. These measurements with the EAA are consistent
with the results of the other methods. Although the preliminary
results shown in Figures 23 and 24 were obtained by this method,
attempts to repeat this measurement were unsuccessful. The
problem was that no stable current of any significance (>10~11*
amps) was produced by charged particles passing through the
device. Apparently, the wall losses were too great and the
number of particles sampled too small.
The data obtained by Method II for dry particles is depicted
by closed circles in Figure 23, Although time consuming, these
measurements are important because Zung and Snead11* observed a
52
-------
3 —
DASHED LINE - THEORY (EQ. I)
A DROPLETS, PAR. PLATES (METHOD III)
SOLID LINE - DROPLETS ELECTRODE (METHOD IV)
O PARTICLES, EAA (METHOD I)
• PARTICLES, FILTER (METHOD II)
T
T
T
I
-60 -90 -120
CHARGING VOLTAGE
-150
-180
Figure 23. Particle Charge Versus Charging Voltage as Determined by
Methods I - IV. (ap = 2.6 nm and ad =23
53
-------
50
40
30
in
o
a
c
20
I I I I
O PARTICLES, EAA (METHOD I)
• PARTICLES, FILTER (METHOD II)
DROPLETS, ELECTRODE (METHOD IV)
FOIL ELECTRODE. 100% COLLECTION
SOLID LINE-ALL DATA
V SAME RUN AS METHOD IV DATA
0 SEPARATE RUN
METAL MESH ELECTRODE
AIR FLOW FOR IMPACTOR SAMPLING
B AIR FLOW TO MAXIMIZE COLLECTION
I I I I
-10' >,
•15 -20 -25
CHARGING VOLTAGE
Figure 24. Particle Charge Versus Charging Voltage
54
-------
change in charge of droplets while drying. The agreement between
measurements using Methods II, III, and IV eliminate this possi-
bility and thus justify the use of Method II which determines
particle charge from the droplet charge measurement.
The dashed line in Figure 23 gives the predictions of Eq.
(3) where the parameters L, H, and a were measured independently
of np. Since npo, the charge given to particles in the spraying
process cannot be predicted, its value in Figure 23 was set as
that obtained by Method IV. The length of the jet [L in Eq. (3)]
at droplet formation was measured to be 6 x 10~*cm ±10%. How-
ever, the end of the jet is difficult to locate without high-
speed photography. This difficulty causes additional uncertainty
in the measurements.
As indicated, Figure 24 gives a more detailed view of the
lower part of Figure 23, In addition some data points obtained
with Method IV are shown. For the foil electrode one set is
shown which was measured in the same run as the Method II data.
Another set, obtained in a separate run, is plotted to show the
variation one may expect from one run to another.
The data depicted by squares in Figure 23 are presented to
illustrate data which may be obtained with the two described
variations of Method IV. The data shown as open squares are a
result of the metal-mesh electrode with a high voltage wire
along the electrode axis. The several values given for np at
constant Vc are for different levels of this high voltage: 0,
1000, and 2000 volts. The higher np values correspond to
greater collection efficiencies. Although a 100% collection
efficiency was not achieved in this series of measurements, it
can be, as shown by the data point depicted by a solid square.
This datum was the result of adjusting the dispersion and drying
air of the VOAG to maximize the electrometer current with the
high voltage wire at zero volts. Data concerning the variations
of Method IV is given only to illustrate the possibility for
use in further work if the ability to monitor particle charge
is important.
The solid line in Figure 24 is being used as the calibra-
tion curve for the charged particle generator. Its agreement
with Methods I and II and the results of Reisch, ejt al12
justify its use. In addition to the calibration, further
measurements of particle charge with Method IV were performed
before and after most impactor sampling runs.
Impactor Data
Impactor sampling data are given in Figures 25 through 39
in the form of histograms. Each figure shows data from one
55
-------
0.7
0.6
0.5
IU
™l
O 0.4
u
I
u.
0 0.3
z
O
§
cc
u,
0.2
0.1
MRI
5.2 /An AEROSOL
vc-ov
CHARGE NEUTRALIZER - ORIFICE _
J • JST
S-SUBSTRATE
H - HOUSING
F - FINAL
N • NOZZLE
18 x 10s PARTICLES
111
JSH
222
JSH
333
JSH
444
JSH
555
JSH
666
JSH
777
JSH
SH
IMPACTOR SURFACE
3630-014 a
Figure 25. Control sampling run using MPi-Model 1502 Impactor
with no charge.
56
-------
ff
U.
0.7
0.6
0.5
o
iu
O 0.4
o
UL
0 0.3
0.2
0.1 -
n
-n
MR)
5.2 jun AEROSOL
V...33V
np - 4 * 104
J-JET
S- SUBSTRATE
H - HOUSING
F • FINAL
N • NOZZLE
5.9 x 10s PARTICLES
Jirf
N 111 222 333 444 655 666 777 FSSH
JSH JSH JSH JSH JSH J S H JSH
IMPACTOR SURFACE
3630-0153
Figure 26. Sampling charged particles using MRI-Model 1502 Impactor
with no grounding wire— high particle charge. (See Figure 25
for comparison.)
57
-------
0.7
0.6
0.5
o
tu
8
o
o
o
<
cc
0.3
MRI
5.2 ym AEROSOL
Vc= -33V
n - 4 x 10*
P
J-SET
S•SUBSTRATE
H • HOUSING
F - FINAL
N • NOZZLE
3.6 x 10 5 PARTICLES
IMPACTOR SURFACE
FS SH
3630-0163
Figure 27. Sampling charged particles using MRI-Model 1502 Impactor
with grounding wire • high particle charge (See Figure 25 and 26
for comparison.)
58
-------
0.7
0.6
0.5
O
5
O 0.4
g
u.
O
O
O
<
C
0.3
0.2
0.1 • -
MRI
fkJ
AEROSOL
Vc = - 33V
CHARGE NEUTRALIZER - NOZZLE
J - JET
S- SUBSTRATE
H • HOUSING
F • FINAL
N - NOZZLE
25 x 10s PARTICLES
_rTL
N 111 222 333444 555 666 777 FSSH
JSH JSH JSHJSH JSH JSHJSH
IMPACTOR SURFACE
3630-017(9
Figure 28. Sampling charge-neutralized particles using MRI-Model 1502
Impactor with neutralizer at nozzle - n - 4 x 10*.
59
-------
0.7
0.6
0.5
O
O
m
8
u.
O
z
O
g
c
0.3
0.2
0.1
i-Tk
Ikd
MRI
5.2 pm AEROSOL
Vc = -33V
CHARGE NEUTRALIZER - NOZZLE
J-JET
S- SUBSTRATE
H • HOUSING
F • FINAL
18 x 10s PARTICLES
J S H
222
J S H
333
J S H
444
J S H
S 6 S
J S H
666
J S H
IMPACTOR SURFACE
777 FS SH
J S H
3630-018
Figure 29. Sampling charge-neutralized particles using MRI-Model 1502
Impactor with nozzle losses corrected. (Run data identical
to Figure 28.)
60
-------
0.7|
0.61
0.5*
O
u
O 0.4
o
I
0.3*
oc
u.
0.2\
MR)
5.2 fjm AEROSOL
VC = -8V
nn - 7 x 103
N - NOZZLE
J-JET
S- SUBSTRATE
H • HOUSING
F - FINAL
6.3 x 10s PARTICLES
IMPACTOR SURFACE
7 7 7 FS SH
J S H
3630-0193
Figure 30. Sampling charged particles using MRI-Model 1502
Impactor with grounding-wire moderate particle charge.
(See Figures 25 and 27 for comparison.)
61
-------
.5
O
Ul
_l
C^ .4
O
I
u.
O
O
o
<
cc
.3
.2
.Ob
nJ
MR I
2.1 fJm AEROSOL
CHARGE NEUTRALIZER • NOZZLE
J-JET
S- SUBSTRATE
H - HOUSING
F • FINAL_
N • NOZZLE^
16 x 106 PARTICLES
zfl
N
1 1 1
J S H
222
J S H
333
J S H
444
J S H
555
J S H
666
J S H
777
J S H
FS SH
IMPACTOR SURFACE
Figure 31. Sampling charge-neutralized particles using MRI-Model 1502
Impactor with neutralizer at nozzle - n * 2.4 x 10?. See
Figure 29 for comparison with 5.2 pm aerosol data.
62
-------
.7
.5
_l .4
O
U
O .3
O
.2
MR)
2.1 /An AEROSOL
VC = -4V
np = 2.6x 103
J-JET
S- SUBSTRATE
H - HOUSING
F-FINAL
N • NOZZLE
20 x 106 PARTICLES
N
1 1 1
J S H
222
J S H
333
J S H
444
J S H
555
J S H
666
JS H
777
J S H
FS SH
IMPACTOR SURFACE
Figure 32. Sampling charged particles using MRI-Model 1502
Impactor with grounding wire - moderate particle
charge. (See Figure 30 for comparison.)
63
-------
0.7
0.6
0.5
2
O
8
o
o
cc
u.
0.3
0.2
0.1
rTLr
U. of W.
5.2pm AEROSOL
Vc = -33V
CHARGE NEUTRALIZER - NOZZLE
N- NOZZLE
J -JET
S- SUBSTRATE
H - HOUSING
F - FINAL
15 x 10s PARTICLES
f~Tl nr-ii—i
N NH 1
S
2 2
J S
3 3
J S
4 4
J S
5 5
J S
6 6
J S
FS SH.
IMPACTOR SURFACE
Figure 33. Sampling charge-neutralized particles using U. of W. Mark III
Impactor with charge neutralizer at nozzle - nQ = 4 x 1(T.
64
-------
0.7
(X6
0.5
0.4
0.3
0.2
0.1
rffl
U. of W.
5.2 pn AEROSOL
Vc = -33V
np = 4 x 104
N- NOZZLE
J-JET
S- SUBSTRATE
H - HOUSING
F - FINAL
7.2 x 10s PARTICLES
f—4 \ t—(~~] I~~L
N NH 1
S
2 2
J S
3 3
J S
4 4
J S
5 5
J S
6 6
J S
7
J
FS SH
IMPACTOR SURFACE
Figure 34. Sampling charged particles using U. of W. Mark Ml.
Impactor with no grounding wire - high particle charge.
(See Figure 33 for comparison.)
65
-------
0.7
0.6
0.5
O
H
O
LU
o o.a.
o
I
o
z
o
o
<
oc
0.3
0.2
0.1
U. of W.
5.2 fJm AEROSOL
Vc = -33V
nQ = 4 x 104
N - NOZZLE
J-JET
S- SUBSTRATE
H - HOUSING
.F - FINAL
13 x 10s PARTICLES
r-n
N NH I 22
S J S
3344
J S J S
5566
J S J S
77 SH FS
J S
IMPACTOR SURFACE
Figure 35. Sampling charged particles using U. of W. Mark III
Impactor with grounding wire - high particle charge.
(See Figures 33 and 34 for comparison.)
66
-------
0.71
0.6
0.5
U
LLJ
O 0.4
I
U.
z 0.3
O
0.2
O.I
r-Thf
U. of W.
5.2 /Jm AEROSOL
VC = "8V
n = 7 x 103
JP-JET
S - SUBSTRATE
H - HOUSING
F - FINAL
N - NOZZLE
19 x 106 PARTICLES
N H 1
S
2 2
J S
3 3
J S
4 4
J S
5 5
J S
6 6
J S
7 7
J S
FS FH
IMPACTOR SURFACE
Figure 36. Sampling charged particles using U. of W. Mark III
Impactor with grounding wire - moderate particle
charge. (See Figures 33 and 35 for comparison.)
67
-------
0.7 i
0.6
0.5
O
O
O
O
u.
O
z.
O
(-
u
0.4
0.3
0.2
0.1
JL
N NH 1 1
J S
2 2
J S
3 3
.1 S
4 4
J S
ANDERSEN
5.2 /Jm AEROSOL
Vc - -8V
CHARGE NEUTRALIZER -
J-JET
S—SUBSTRATE
H - HOUSING
F - FINAL
N - NOZZLE
42 x 10* PARTICLES
NOZZLE —
5 5
J S
6 6
J S
7 7
J S
8 8
J S
FS SH H
IMPACTOR SURFACE
Figure 37. Sampling charge - neutralized particles using Andersen
Mark III Stack Sampler with neutralizer at nozzle -
np = 7 x 70s.
68
-------
.7
.6
Z
o
o
o
o
u.
O
Z
o
.4
.3
.2
rTI
Id
ANDERSEN
AEROSOL
VC=-8V
np = 7 x 103
J- JET
S - SUBSTRATE
H - HOUSING
F - FINAL
N - NOZZLE
16.2 x 106 PARTICLES
N. NH 1 1
J S
2 2
J S
3 3
J S
4 4
J S
5 5
J S
6 6
J S
7 7
J S
8 8
J S
FS SH H
IMPACTORSURFACE
Figure 38. Sampling charged particles using Andersen Mark III
Stack Sampler with grounding wire -moderate
particle charge. (See Figure 37 for comparison.)
69
-------
.7
.6
.5
O
O
O
U
u.
O
O
.4
.3
.2
m
ANDERSEN
2.1 /An AEROSOL
n
•4V
' 600
P
J-JET
S - SUBSTRATE
H - HOUSING
F - FINAL
N - NOZZLE
20 x 106 PARTICLES
n
N NH 1 1
J S
2233
J S J S
4455
J S J S
6677
J S J S
8 8 FS SH H
J S
IMPACTOR SURFACE
Figure 39. Sampling charged particles using Andersen Mark III
Stack Sampler with grounding wire - moderate
particle charge.
70
-------
sampling test; the fraction of total particles collected is
given for each surface in the impactor starting with the nozzle
and ending with the housing of the absolute filter at the exit.
The surfaces are distinguished according to nozzle, jet plate,
substrate, and housing if separable. Each of these surfaces
was washed separately and the mass of fluorescein in the wash
water determined with a spectrophotometer.
The parameters varied in these tests were particle size
(5.2 ym and 2.1 ym) , particle charge, electrical grounding, and
the use of a charge neutralizer. The sampling times, varied
between one to one-and-a-half hours. However, the total number
of particles sampled was smaller for highly charged particles,
a result of greater wall losses in the drying chamber.
All sampling tests were performed at ambient conditions
with the same flow rate, 14 LPM. The impactor nozzles were
1.27 cm (0.5 in.) in diameter giving a gas velocity of 1.84
m/sec. The stream velocity where sampling occurred was approxi-
mately one m/sec. Glass fiber substrates were employed in each
test.
MRI Impactor: Figures 25 through 32 show MRI Model 1502
Cascade Impactor data (Figures 25-30 for 5.2 ym diameter and
31-32 for 2.1 ym diameter particles).
Figure 25 shows data for a "reference" run with no particle
charge. The data shown in Figure 26 is for particles with about
4 x 10** charges and an electrically isolated impactor. The in-
crease in the number of particles collected on the nozzle and
metal jet plates is quite significant. Also there appears to
be a slight increase in the number reaching the stages below
number three. Figure 27 repeats the data of Figure 26 with the
impactor electrically grounded. The collection efficiency of
the nozzle and stage one jet plate (1J) apparently increased.
However, only the nozzle and 1J of the MRI can be grounded from
the outside due to an anodized coating. In these measurements
a grounded wire was wrapped around the threads between U and
its substrate housing. The resistance between ground and U and
2J was measured to be of the order of 109 ohms with the grounding
wire and about 15% higher without the grounding wire.
Figure 28 presents a sampling run with identical operating
conditions to that of Figure 27 except that a charge neutralizer
was mounted upstream of the impactor nozzle. The neutralizer
is a polonium 210 strip made by Nuclear Products Company. For
use with an impactor it is bent to form a ring and placed at the
entrance to the nozzle. In Figure 28 the high collection on the
nozzle is thought to be a result of turbulence induced by the
charge neutralizer. Collection efficiencies were calculated
71
-------
excluding the nozzle losses. These values are given in Figure
29. Comparing, this figure to Figure 25 shows that indeed the
effect of particle charge is eliminated by the ion source
neutralizer. After this run the mount was changed to produce
less disturbance to the flow.
Figures 26 and 27 show an effect of high particle charge
upon impactor behavior. Here, the charge level was higher than
encountered in most effluent streams. In the run depicted in
Figure 30 the charge per particle is approximately 7 x 103
elementary charges, approximately that expected in a precipi-
tator. Collection by the stage 3 substrate (3S), is nearly
twice that with 4 x 101* charges/particle (Figure 27) but is
still 15% less than with no charge (Figure 25). These data
indicate that in sampling 5.2 ym diameter particles exiting a
precipitator, approximately 25 percent of the particles collect
on surfaces due to their charge and not size. This conclusion
is based on the assumption that all particles exiting a pre-
cipitator have the level of charge predicted by Eq. (1).
Figures 31 and 32 show the effect of moderate charge on
2.1 ym diameter particles. 3.T and 4J again show an increased
collection when charged particles are sampled. However, the
effect for 2.1 ym particles charged to saturation is much
smaller than with the larger particles at saturation.
University of Washington; Figures 33 through 35 show data
obtained from the University of Washington Mark III Impactor
in sampling 5.2 ym diameter particles with 4 x 10u charges per
particle. Figure 31 shows results with the charge neutralizer
and Figures 32 and 33 without the neutralizer. The impactor
was grounded electrically in the run depicted in Figure 32 and
was not in Figure 33. It is seen that particle charge effects
the collection efficiency of the various surfaces. The behavior
of the U. of W, impactor is similar to that of the MRI impactor
except that the use of a grounding lead appears to produce a great-
er change in the deposition of charged particles for the U. of W.
impactor.
The results obtained with the moderate charge level,
7 x 103 elementary charges, on 5.2 ym particles are shown in
Figure 36. Comparison of these data with those in Figure 33
shows no significant effect due to this charge level on surface
collection efficiency.
Andersen: Figures 37 and 38 show data obtained with the
Andersen Mark III Stack Sampler when sampling 5.2 ym diameter
particles with the moderate charge level. Comparison of these
two figures shows a significant difference in the fraction of
particles collected upstream of stages three and four. As with
72
-------
the MRI impactor about 25 percent of the 5.2 ym particles with
moderate charge collect on wall and jet surfaces because of
their charge rather than size.
Figure 39 shows sampling data for particles with 2.1 ym
diameters and moderate charge, 600 elementrary charges. The
control run for this data; that is, the sampling of neutral
particles with 2.1 ym diameters, had to be discarded because of
an apparent syringe pump malfunction. However the agreement of
the data in Figure 39 with impactor theory for neutral particles
and a previous sampling study1l shows that no significant effect
of charge was present in sampling the 2.1 ym particles.
In the following section the data shown in Figures 25
through 39 are presented in terms of collection efficiency per
stage versus Stokes number for different levels of charge.
Particle Deposition Patterns
The deposition patterns of fluorescein particles on im-
pactor surfaces were observed with ultraviolet light, and in
the presence of a water mist to enhance the fluorescence.
Photographs in Figures 40 and 41 show representative depo-
sition patterns of particles on the top and bottom of jet plates
and on substrates in the MRI impactor. Most particles, lost to
jet plates, are on the downstream side of the plate as shown in
these photographs. This pattern is visible with both charged
and uncharged particles, but more particles are deposited on
the plates in the charged case. Apparently, particles pass in
the vicinity of these surfaces whether charged or not. As ex-
pected, the U. of W. impactor, with a similar geometry, had
deposition patterns similar to these on the MRI.
Collection efficiencies of jet plates in the Andersen im-
pactor were also slightly higher for charged particles than for
neutral ones. Also, substantial numbers of charged particles
were observed to be deposited on the upper surface around jets
as well as the back surface. The photographs in Figures 42-44
illustrate the Andersen deposition data.
Nonconducting Jet Plates
In an effort to study the effect that jet plates made from
an insulated material would have on deposition due to charge on
particles, plates 2J, 3J, and 4J were fabricated of plexiglass
for the MRI impactor in an attempt to eliminate the effect. The
data for this experiment are shown in Figure 45. Comparison with
Figures 25 and 30 show two significant differences. First, the
stage cut points seem to be shifted to larger sizes. The origin
73
-------
FRONT OF SECOND JET PLATE
BACK OF SECOND JET PLATE
SECOND SUBSTRATE
3630-024
Figure 40.
Photographs of deposition patterns of ammonium fluorescein
particles in MRI-Model 1502 Impactor. Particle diameter was
5.2 nm. Samp/ing test data is in Figure 27.
74
-------
FRONT OF THIRD JET PLATE
BACK OF THIRD JET PLATE
THIRD SUBSTRATE 3630023
Figure 41. Photographs of deposition patterns of ammonium fluorescein
particles in MRI-Model 1502 Impactor. Particle diameter was
5.2 urn. Sampling test data is in Figure 27.
75
-------
FRONT OF FIRST JET PLATE
BACK OF FIRST JET PLATE
FIRST SUBSTRATE
Figure 42. Photographs of deposition patterns of ammonium fluorescein
particles in Andersen Mark III Stack Sampler. Particle diameter
was 5.2 urn. Sampling test data is in Figure 38.
76
-------
FRONT OF SECOND JET PLATE
BACK OF SECOND JET PLATE
SECOND SUBSTRATE
Figure 43. Photographs of deposition patterns of ammonium fluorescein
particles in Andersen Mark III Stack Sampler. Particle diameter
was 5.2 urn. Sampling test data is in Figure 38.
77
-------
FRONT OF THIRD JET PLATE
BACK OF THIRD JET PLATE
THIRD SUBSTRATE
Figure 44. Photographs of deposition patterns of ammonium fluorescein
particles in Andersen Mark III Stack Sampler. Particle diameter
was 5.2 [itn. Sampling test data is in Figure 38.
78
-------
0.7
0.6
0.5
2
g
I-
o
o
u
u.
o
o
u
<
DC
0.4
0.3
0.2
0.1
I" ^r- __
MRI
5.2 Jim AEROSOL
Vc = - 8V
n = 7 x 103
N - NOZZLE
J - JET
S - SUBSTRATE
H - HOUSING
F - FINAL
38 x 10s PARTICLES
111 222
J S H J S H
333444555
JSHJSHJSH
6 6 6 7 7 7 FSSH
J S H J S H
IMPACTOR SURFACE
Figure 45. Sampling charged particles using MRI-Model 1502
Impactor with plexiglass jet plates, 2J, 3J, and 4J.
Other conditions are the same as the test depicted
in Figure 30.)
79
-------
of this behavior is not known. One cause would be changes in
the jet hole sizes. However, measurements ruled out this source
of discrepancy. Secondly, the collection efficiencies of 2J, 3J,
and 4J were reduced.
Effect of Charge on Efficiency Versus /if
Figures 25 through 39 depict impactor behavior under various
circumstances involving particle charge. As discussed above, a
high charge effects the deposition of particles. However, it is
not clear from those figures to what extent charge alters the
calculated size distribution in a collected sample.
The size distribution inferred from impactor data is based
upon stage collection efficiency E^ as a function of the square
root of the inertial impaction parameter, /vp. The true collec-
tion efficiency is defined as the amount of material of a given
size and density collected by a stage divided by the amount in-
cident upon it.
* = DpCpPVo/18nDj
where Dp = particle diameter (cm),
C = Cunningham slip factor,
Dj = jet diameter (cm),
Pp = particle density (gm/cm3),
n = gas viscosity (poise), and
V = jet velocity (cm/sec).
A representative example of efficiency versus /if is given as a
solid line in Figure 46. The use of E-^ versus /if gives the
desired stage efficiency as a function of particle size for a
range of sampling conditions. In practice a stage is calibrated
by experimentally determining efficiency versus /if for a par-
ticular substrate material to obtain /i|T5 0; that is the value of
/if at 50% collection efficiency. Then, for particular sampling
conditions the effective stage cut diameter D50 is obtained from
/ifsT. It is assumed that all particles caught by an impactor
stage are those particles having diameters equal to or greater
than the DSO of that stage, but less than the cut point of
the preceding stage. Therefore, the effect of particle charge
upon the calculated size distribution depends on its effect
upon /vpso. A more sophisticated deconvolution of impactor
data which uses the entire efficiency curve has been pro-
posed,1'2'3 but use of these techniques is limited to low-
noise sampling data.
80
-------
00
*
100
90
80
70
If 60
u
50
40
3 30
20
10
0
i i i i
i i i i i i i i i i r
i i i
I I I I I I J_ l_l_L _L i
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Figure 46. Reference curve (solid) giving collection efficiency as a function
of y/W for the Andersen Stack Sampler and neutral particles. The
measurements ("O") of this study for neutral particles are plotted
by the procedure given in the text. The dashed lines depict
the envelope of all stage efficiency curves superimposed.
-------
In order to completely characterize the effect of particle
charge, the function Ei (/ij>,np) ; that is, efficiency expressed
as a function of the square root of the inertial impaction
parameter and charge (np) for stage i is required. For neutral
aerosols, E;(/\Ji,O) has been determined in an extensive calibra-
tion studyir for each stage of the Andersen, MRI, and U. of W.
impactors at the sampling conditions used here, except that
grease substrates were used with the latter two. To a first ap-
proximation, Ei(/ij;,0) has the same form for all stages with the
same substrate material and impactor. The reference curve which
we denote by E(/$,0) in Figure 46 was drawn by superimposing the
stage calibration curves of the Andersen from the previous study.11
Since the major differences in Ej (/ip,O) from one stage to another
is not in the shape of the curve, but /i|>50, for the purpose of
this analysis the calibration curves are normalized so that
/^so of each coincided at 0.4. The dashed lines in Figure 46
define the envelope of those curves.
The collection efficiency of each stage was calculated for
each run from the data given in Figures 25 through 39. These
efficiencies were "corrected for wall losses" by combining the
particles collected on the various surfaces according to
Table VIII. These groupings are based upon the procedure used
in field testing and the laboratory observations discussed
above concerning deposition on jet plates. In the field, par-
ticulate mass on the top of jet plates is combined with the
preceding stage, and that on the bottom is combined with the
stage below the jet plate. The results for sampling neutral
particles with the Andersen are given in Figure 46 as points.
For comparison of these data to our reference curve, each
measured efficiency is normalized and plotted as described above.
The deviations of the data points in Figure 46 from the refer-
ence curve is thus a measure of the degree to which impactor
calibration data for neutral particles may vary. In Figure 47
the same procedure was followed to relate the measured effi-
ciencies of the Andersen stages with charged particles to its
generalized efficiency curve for neutral particles. The dashed
curve is the resulting E(/4J,np) for moderate np. The effect of
charge is to reduce the sharpness (slope) of the curve at /^
values below the peak while not changing /i^so significantly.
At /ty values above the peak the effect of charge reduces the
efficiency in a manner similar to particle bounce.
Figures 48 and 49 and Figures 50 and 51 present information
about the MRI and U. of W. impactors analogous to that in
Figures 46 and 47 for the Andersen. It must be noted that stage
efficiency curves of the previous study11 for neutral particles
were obtained using grease rather than glass fiber substrates.
Therefore some adjustments in the reference curves had to be
made based upon the efficiency data of the present study and
82
-------
00
U)
100
90
80
o
a.
u.
* 50
O
0 40
UJ
8 30
20
10
I I I I I I I I 1 I I I I I I I I I
i i i i i i i i i i i i i i
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Figure 47. Efficiency versus \/* of stages in Andersen Stack Sampler for
neutral particles (solid curve) and moderately charged particles
(dashed curve). The measurements of this study for moderately
charged particles ("$" - 5.2 urn and "9" - 2.1 nm diameter} are
plotted by the procedure given in the text.
-------
00
100
90
80
* 70
of 60
u
uJ 50
O
6 40
UJ
20
10
0
I I I I I I I I I I I I I I
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Figure 48. Reference curve (solid) giving collection efficiency as a function
of\/y? for the MRI-Model 1502 Impactor and neutral particles.
The measurements ("&') of this study for neutral particles are
plotted by the procedure given in the text. The dashed lines
depict the envelope of all stage efficiency curves superimposed.
-------
00
U1
100
90
80
' 70
HI
o 60
50
40
S 30
20
10
9
I I I I I I I I < I 1 I I I I I 1 I
i i i i i i i i i r~i~.j 4 i
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Figure 49. Efficiency versus \/^F of stages in MRI-Model 1502 Impactor for
neutral (solid curve), moderately charged (dashed curve), and highly
charged ( ) particles. The measurements of this study for
moderately charged ("&' - 5.2 \im and "A" - 2.1 nm diameters)
and highly charged 5.2 pm diameter ("&" - grounded and "^/' -
not grounded) particles are plotted by the procedure given in the
text.
-------
co
-------
00
100
90
80
s 70
1 I I I |,n-K I I I I I I I I I I \ T
o
o
u.
50
O
u 40
UJ
8 30
20
10
I I I I I I I I I I I I I I I
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
Figure 51. Efficiency versus \ffy of stages in the U. of W. Mark /// Impactor
for neutral (solid curve), moderately charged (dashed curve), and
highly charged (—•) particles. The measurements of this study
for.5.2 \jjrn diameter panicles, moderately charged ("W) and
highly charged ("O" - grounded and "{$' not grounded) are
plotted by the procedure given in the text.
-------
Table VIII. Grouping of Surfaces for the
Efficiency of Each Stage
Stage
1
2
3
4
5
6
7
8
Andersen
surface
N
NH
U
IS
2J
2S
3J
2S
4J
4S
5J
5S
6J
6S
7J
7S
8J
8S
FS
SH
H
MRI
surface
N
1J
IS
1H
2J
2S
2H
3J
3S
3H
4J
4S
4H
5J
5S
5H
6J
6S
6H
7J
7S
7H
FS
SH
U. Of W.
surface
N
NH
IS
2J
2S
3J
3S
4J
4S
5J
5S
6J
6S
7J
7S
FS
SH
88
-------
comparisons of available efficiencies of another impactor with
these two substrates to determine the differences. The result-
ing reference curve of the MRI impactor is shown in Figure 48.
For /ij> values up to 0.5, following the same procedure as in
Figure 46, the reference curve of Figure 48 was obtained using
MRI efficiency curves for grease substrates. The use of these
is justified because in this region the efficiencies for neutral
particles measured in this study, do not substantially deviate
from this reference curve.
In the region of /if values greater than 0.5 glass fiber
substrates produce substantial bounce, meaning that efficiency
drops to much less than 100%. This is indicated by the low
measured efficiencies shown in Figure 48. Therefore in this
region a reference curve was determined from experimental data
for neutral particles taken in the course of this study.
Figure 49 shows the effect of high and moderate particle
charge upon the MRI stage efficiency versus /if. For moderate
charge the effects are similar to those with the Andersen. At
the higher charge level however the changes in E(/if,np) are
quite large at all /if values.
The effect of particle charge upon U. of W. stage efficiency,
shown in Figure 51, is small for moderate charge and substantial
with high charge. It is interesting that in the "bounce"
region, moderate charge appears to increase the efficiency sub-
stantially, reducing the effects of bounce. However, data
obtained for Sty values to the right of the peak in efficiency
are inherently subject to fluctuation because small numbers of
particles are involved. Therefore exceptions to the overall
trends in this region are questionable.
Figures 47, 49, and 50 show two obvious effects of particle
charge. At low values of /if the stage efficiency is increased
due to the collection of particles on metal jet plates. At
high values of /if, past the peak, the efficiency is reduced as
if repulsive forces of charge are significant. Charge produced
very small changes in Dso.
The foregoing discussion presents the results from an ex-
perimental study done with monodisperse aerosols over a limited
range of test parameters. A second set of experiments are now
being done with polydisperse aerosols. The results of both
studies will be analyzed and combined into a single final report.
89
-------
Cascade Impactor Substrate Media Study
(Technical Directive Number 10501)
Description of Task;
The purpose of this task was to conduct an experimental
program to investigate the magnitude of weight changes of cas-
cade impactor substrates when exposed to industrial flue gases.
This study determined the causes of these weight changes and
investigated all candidate materials to find out which were
most stable. Also preconditioning techniques were investigated
to stabilize weight changes in those substrates. At the con-
clusion of this study, a detailed report was written giving the
test results and suggesting a protocol for minimizing the im-
pactor substrate weight changes.
Summary of Progress;
Cascade impactors are widely used to determine particle size
distributions in air pollution control device research programs.
In these research programs a large variety of flue streams are
encountered with temperatures ranging from ambient to around
370°C (700°F). Gas analyses show that many of these sources
contain some SOX components, particularly those associated with
fossil fuel fired boilers.
Most impactors have collection stages which are too heavy to
allow accurate measurements of the mass of the particles collec-
ted in each size fraction. Weighing accuracy can be improved by
covering the stage with a lightweight collection substrate made
of aluminum foil, teflon, glass fiber filter material, or other
suitable lightweight materials, depending upon the particular
application. Some manufacturers now furnish lightweight inserts
to be placed over the collection stages. With such arrangements
it is possible to collect enough material on each stage to make
an accurate determination of the mass collected and avoid over-
loading the stage. If the stage is overloaded, some deposited
particulate matter can be reentrained and deposited on another
stage or the back-up filter and lead to erroneous results.
Substrate materials may also serve the purpose of changing
the surface characteristics from those of a bare metal or plastic
to something better suited to holding particles with impact.
Thus, various greases are often used, either on bare impactor
plates, or, more frequently, on metal foil substrates.
90
-------
Presented in the published final report are the results of
investigations concerning the use of two classes of impactor
substrates—greased metal foils, and fiber filter materials.
Tests were made under both laboratory and field conditions to
evaluate each of several greases and filter materials. The
general purpose of this study was to identify specific materials
and handling techniques which may be used to improve the accuracy
of weight measurements in impactors by reducing uncertainties
arising from changes in substrate weights.
Although normal substrate preparation includes baking and
desication before the initial weighing, it is frequently found
that weight losses can occur when sampling clean air. Previous
tests were conducted to investigate this phenomenon in detail.15
It was found that with careful handling, weight loss per glass
fiber substrate for Andersen impactors can be kept below 0.1 mg.
This loss is attributed to loss of fibers which stick to seals
within the impactor and to "superdrying" when sampling hot, dry
air. Weight losses of 0.1 mg are small compared to most stage
catches when sampling particulate matter, and this are within an
acceptable range for sampling errors.
A more significant problem is excessive weight gain of the
glass fiber material itself due to gas phase reactions. These
reactions appear to be caused by the SOX component in flue gases.
A series of studies was thus directed toward developing procedures
to passivate glass fiber materials against the effect of SOX com-
ponents in flue gases.
Although greases offer good impedance to particle bounce on
substrates, they are subject to temperature limitations. Sampling
clean, hot air while using greased substrates may result in
severe weight losses. These losses appear to result from one or
more of several mechanisms which may include continued loss of
volatile components, erosion of grease by the action of the gas
jet in the impactor, and occasional flow of grease from the sub-
strate to other surfaces within the impactor. In addition,
chemical reactions may play role in some cases. Occasionally,
some of the weight lost on upper impactor stages has been found
to reappear on a back-up filter, which is an indication that the
grease has been blown off the collection surface or has chemically
reacted to form a fine "smoke" which was then collected by the
back-up filter.
Summary of Results of Evaluation of Greases
Upon preliminary screening by static heating tests in the
laboratory, six of the nineteen greases tested were found to have
acceptable characteristics at elevated temperatures. Among those
greases eliminated by these tests, large changes in mass or in
consistency had occurred.
91
-------
In the field tests greases were applied to metal foil im-
pactor substrates and were subjected to a flue gas sampling
procedure. Particulate matter was removed by a prefilter so
that the effects of the flue gas alone on the greased sub-
strates could be observed.
As a result of the field studies it was concluded that
Apiezon H grease performed best of the greases tested. Other
greases studied displayed changes in consistency or a tendency
to flow under the influence of the gas stream.
Further tests on Apiezon H have demonstrated that this grease
is a suitable substrate material for applications where the
temperature does not exceed approximately 177°C (350°F).
Summary of Results of Evaluation of Filter Media
Untreated glass fiber filter materials used as impactor sub-
trates will almost invariably increase in mass when subjected to
the hot flue gases normally encountered in field applications.
Conversion of SOa to various sulfates appears to be the cause of
mass gains. The various filter materials tested vary widely in
the amount of mass change which occurs under a particular set of
flue gas conditions.
Preconditioning techniques can be used to force the pro-
duction of sulfates in a filter medium, leaving a minimal number
of sites available for chemical reaction in the flue gas, and
hence, providing substrate material for which minimum mass gains
occur during use in an impactor. The best results were achieved
when substrates were washed in sulfuric acid, baked, and con-
ditioned in situ.
Of the filter materials studied, Reeve Angel 934AH was
found to be most suitable in all respects for use as cascade im-
pactor substrates.
Conclusions and Recommendations
Collection stages of most types of cascade impactors are
very heavy in comparison with the amounts of particulate material
normally collected. It is therefore the usual practice to aug-
ment each collection stage with a lightweight substrate to
improve weighing accuracy. Generally, two classes of substrates
are used—greased metal foils, and glass fiber filter material.
Greased foils provide resistance to particle bounce and
scouring effects, but greases tend to be unstable at elevated
temperatures. Some tend to harden, and in others the viscosity
may become reduced so that they may flow or be blown off the
92
-------
surface by the high velocity gas flowing through the impactor
jets. Of the greases tested, Apiezon H was found to perform
most satisfactorily. This grease may be used at temperatures
up to approximately 177°C (350°F). No greases were found to be
useable at higher temperatures.
Mass gains exhibited by glass fiber filter materials when
they are exposed to the SOX components in flue gas streams pose
a complicated problem. Experiments show that these mass gains
are caused by formation of sulfates due to a gas phase reaction
with SOX. Laboratory and field experiments indicate that the
only glass fiber filter material suitable for use as a cascade
impactor substrate is Reeve Angel 934AH. When this material is
acid treated, according to a procedure given below mass gains
caused by flue gas reactions can be kept quite small.
It is recommended that acid washing, baking and in situ
conditioning be used whenever large blank mass gains with large
standard deviations are expected. In this context, "large"
refers to substrate mass gains greater than several tenths of a
milligram.
Further research may provide a technique for passivating
glass fiber materials to all mass gains. It has been suggested
that a high temperature polymer or silicon compound might be
developed to coat the glass fibers in much the same way that the
Gelman Spectro-Grade material is prepared for use at low temper-
atures.
The Final Report for this task was published in a document
entitled "Inertial Cascade Impactor Substrate Media for Flue
Gas Sampling", EPA-600/7-73-060, June 1977.
Procedure for Acid Washing of Substrates
1. Submerge the substrates to be conditioned in a 50-50
mixture (by volume) of distilled water and reagent grade concen-
trated sulfuric acid at 100°-115°C (230-239°F) for 2 hours. This
operation should be carried out in a hood with clean glassware.
Any controllable laboratory hotplate is suitable.
The substrates may need to be weighted down to keep them
from floating. For this purpose, place a teflon disc on the top
and bottom of the substrate stack. The top disc can be held down
with a suitable glass or teflon weight.
2. When the substrates are removed from the acid bath they
should be allowed to cool to room temperature. They are next
placed in a distilled water bath and rinsed continuously with a
water flow of 10-20/cm3/min. The substrates should be rinsed
93
-------
until the pH of the rinse water, on standing with the substrates,
is nearly the same as that of the distilled water. The importance
of thorough washing cannot be over-emphasized.
3. After rinsing in distilled water the substrates are
rinsed in reagent grade isopropanol (isopropyl alcohol). They
should be submerged and allowed to stand for several minutes.
This step should be repeated four to five times, each time using
fresh isopropanol.
4. Allow the substrates to drain and dry. They can be
spread out in a clean dry place after they have partially dried
(dry enough to handle).
5. When the filters are quite dry to the touch they should
be baked in a laboratory oven to drive off any residual moisture
or isopropanol. Bake the substrates at 50°C (122°F) for about
two hours, at 200°C (392°F) for about two hours, and finally at
370°C (700°F) for about three hours. The substrates are now
ready for iri situ conditioning.
As a final check, place two substrates in about 50 ml of
distilled water, and check the pH. The substrates to be checked
for pH should be torn into small pieces, placed in the water, and
stirred for about 10 minutes before the pH is measured. If the
pH is significantly lower than that of the distilled water, then
the filters should be baked out at 370°C (700°F) for several
hours more to remove any residual sulfuric acid. The boiling
point of sulfuric acid is 338°C (640°F), so high temperatures
must be used.
Figure 52 is a flow chart representing the acid wash pro-
cedure described in the foregoing paragraphs.
Calibration and Evaluation of Commercial Impactors
(Technical Directive Number 20101)
Description of Task:
Under EPA Contract Number 68-02-0273 the calibration and
evaluation of five commercial impactors using ammonium fluorescein
aerosols was begun. At the conclusion of this contract the upper
stages had been calibrated. Under Contract Number 68-02-2131
the lower impaction stages were calibrated using a Pressurized
Collison Nebulizer System and polystyrene latex aerosols. The
final report for Contract 68-02-0273 was modified to include
these results.
94
-------
WASH SUBSTRATES IN
50% H2S04 SOLUTION
RINSE IN WATER
(ROOM TEMPERATURE)
RINSE IN ISOPROPANOL
(ROOM TEMPERATURE)
DRY SUBSTRATES
IN AMBIENT AIR
BAKE OUT
RESIDUAL MOISTURE
pH TOO LOW
TEST pH OF
SUBSTRATES
STORE IN
OESSICATOR FOR
ULTIMATE USE
Figure 52. Flow chart for acid wash treatment of glass fiber filter material.
95
-------
Summary of Progress;
This summary presents a brief description of the methods
and results of an evaluation and calibration of five commercially
available cascade impactors. These cascade impactors and their
manufacturers are listed below:
1. Andersen Mark III Stack Sampler (Andersen)
Andersen 2000, Inc.
Atlanta, Georgia 30320
2. Brink Model BMS-11 Cascade Impactor (Brink)
Monsanto Enviro-Chem Systems, Inc.
St. Louis, Missouri 63166
3. MRI Model 1502 Inertial Cascade Impactor (MRI)
Meteorology Research, Inc.
Altadena, California 91001
4. Sierra Model 226 Source Cascade Impactor (Sierra)
Sierra Instruments, Inc.
Carmel Valley, California 93924
5. University of Washington Mark III Source Test
Cascade Impactor (U. of W.)
Pollution Control Systems, Inc.
Renton, Washington 98055
The normal 5-stage Brink impactor was modified to include
an inline cyclone pre-collector, a "0" stage, and a "6" stage.
Table IX presents the operational parameters of the five cas-
cade impactors. The Andersen impactor was used with glass fiber
substrates supplied by the manufacturer. For the Brink impactor
a small disc of glass fiber material was tested as well as a
thin grease layer (Vaseline). The MRI and U. of W. impactors
were tested with thin films of grease (Vaseline) on the collec-
tion plates. The Sierra was supplied with pre-cut glass fiber
mats.
To the user of an inertial cascade impactor the most impor-
tant consideration is the degree to which the data that is
obtained will duplicate the actual particulate size distribution
which is sampled. In order to transform the mass collected by
several impaction stages into a size distribution, an accurate
knowledge of the relationship between collection efficiency and
particle size for each stage is essential.
Theoretically, cascade impactor operation can be described
by the theory of impaction from a jet. The end result of such
a calculation is impaction efficiency versus particle size.
96
-------
Table IX. Cascade Impactor Calibration Study
Operational Parameters
Laboratory Conditions - 73°F/22°C
Aerosol Particles - Ammonium Fluorescein
- Polystyrene Latex
No. of
29.5" Hg/750 mm Hg
Density - 1.35 gm/cm3
Density - 1.00 gm/cm3
Nominal Sampling
Impact or
Andersen
Brink (Modified)
Brink (Modified)
MRI
Sierra
Sierra
U of Washington
Stages
8
7
7
7
6
6
7
Substrate Material
Pre-cut glass fiber
filter mats
Glass fiber filter
inserts
Greased collection
plates
Greased collection
plates
Pre-cut glass fiber
filter mats
Pre-cut glass fiber
filter mats
Greased collection
plates
Flow Rate
0.5 ACFM/
14.16 LPM
0.03 ACFM/
0.85 LPM
0.03 ACFM/
0.85 LPM
0.5 ACFM/
14.16 LPM
0.5 ACFM/
14.16 LPM
0.25 ACFM/
7.08 LPM
0.5 ACFM/
14.16 LPM
97
-------
Impaction efficiency is defined as the fraction of particles of a
certain size in the jet which impact on a collection plate. This
value can be obtained theoretically or experimentally under ideal
conditions. The collection efficiency of an impactor stage,
however, is the ratio of the mass (or number) of particles of a
certain size collected on an impaction surface to the total mass
(or number) of particles of the same size in a jet impinging on
that surface. The collection efficiency is the product of the
theoretical impaction efficiency and the adhesion efficiency.16
The adhesion efficiency is the fraction of the number of particles
which adhere to the surface after touching it by the impaction
process. This depends in a large part on the surface character-
istics of the particle and collection surface. Thus, there will
be disagreement between the theoretical impaction efficiency and
the experimentally determined collection efficiency in the cases
where particle bounce, reentrainment , electrostatic effects, wall
losses, and non-ideal geometry have an effect. For this reason,
the theory of impaction may not be sufficiently accurate in pre-
dicting impactor performance.
The theory of the impaction process has been developed by
several researchers17'18 to a state where the efficiency, E, of
impaction can be determined as a function of the particle size
(D ) , Reynolds Number (Re) , jet diameter or width (Dj ,W) , the jet
topplate distance (S) , and the jet throat length (T) .
E = E (D , Re, S/D.., T/Dj)
It is common practice to relate the particle size Dp to the
square root of the inertial impaction parameter, /iJT. fy is the
ratio of the particle stopping distance, H , (the distance a
particle will travel in air when given an initial velocity, VQ)
to the jet diameter or width (D. or W) .
Cp V
« * - V IS
where C = Cunningham Slip Factor,
D . = Jet Diameter (cm) ,
p = Particle Density (gm/cm3),
y = Gas Viscosity (poise), and
V = Jet Velocity (cm/sec).
98
-------
The square root of the inertial impaction parameter, /ijT,
is used in impaction theories as a dimensionless quantity pro-
portional to particle size.
CPPV0
18 y D.
The inertial impaction parameter is useful in graphing
impactor calibration data because information from all stages
of an impactor can be placed on a single graph, and under many
circumstances would in theory lie along a common curve. The
value of Sty at 50% collection efficiency, *^)50 , defines the
impaction stage Dso, the particle size at which half the
particles of that size are collected and half are passed to
the next stage. Thus, the DSO is used as the effective stage
cut diameter.
Recently Marple18 has been able to construct theoretical
impaction efficiency curves for several values of the jet to
plate distance, jet Reynolds Number, and jet throat length.
Figure 53 shows the results of these calculations for both
round and rectangular jet impactors. The value of the square
root of the Stokes number used by Marple, /STK, differs from
Sty by a factor of SI. It can be seen from Figure 53 that for
certain ranges of Re, S/Dj; S/W, or T/S; T/W, the magnitude
of /ij>5 Q is sensitive to these parameters.
Description of Experimental Procedures
Laboratory evaluation of the cascade impactors involved the
use of two types of particle generation systems: A Vibrating
Orifice Aerosol Generator (VOAG) and a Pressurized Collison Neb-
ulizer System (PCNS). The VOAG was used to generate monodisperse
ammonium fluorescein particles with diameters from 18 micrometers
to 1 micrometer. The PCNS was used to disperse three sizes of
monodisperse Dow Corning Polystyrene Latex (PSL) spheres: 2.02
micrometers, 0.82 micrometer, and 0.46 micrometer diameter.
The VOAG used in this study was designed and built at
Southern Research Institute, although similar devices have been
reported by several authors previously,19'20'21 and a commercial
unit is available from Thermo Systems, Inc., St. Paul, MN 55113.
Figure 54 is a schematic diagram showing the operating principle
of the VOAG. A solution of known concentration (in our case, a
solution of fluorescein (CzoHiaOs) in 0.1 N NH^OH) is forced
through a small orifice (5, 10, 15, or 20 ym diameter). The
orifice is attached to a piezoelectric ceramic which, under
99
-------
100
B 80
111
> 60
u
u.
tti 20
T/DjIS/Dj
Round
i
| f — — —• Rectangular
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
VSTK"
(a) EFFECT OF JET TO PLATE DISTANCE (Re=3,000)
0.1
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
VSTK"
Ib) EFFECT OF JET REYNOLDS NUMBER (T/W-1)
100
5 80
LU
> 60
o
Lit
5
IL
S 20
0
Round
_J
— — — - Rectangular
l I i I
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
(c) EFFECT OF THROAT LENGTH IRe=3,000)
Figure 53. Theoretical impactor efficiency curves for rectangular and round
impactors showing the effect of jet-to-plate distance S, Reynolds
number Re, and throat length T. Note that ^/STK = Dp(CppV0/9 nDj)1/2,
whereas v^ = Dp(CppV0/18 nDj)1^. After Marplef3}
100
-------
PlexigassDrying
Chamber
Vibrating
Orifice
Flow
Meters
Control
Valves
Charge Neutralizer
Signal Generator
Membrane
Filter
Syring
Pump
Dry Air
Figure 54. Schematic representation of the Vibrating Orifice Aerosol Generator.
101
-------
electrical stimulation, will vibrate at a known frequency. This
vibration imposes periodic perturbations on the liquid jet causing
it to break up into uniformly sized droplets of ammonium fluores-
cein. Knowing the liquid flow rate and the perturbation frequency,
the droplet size can be readily calculated. The solvent evaporates
from the droplets leaving the non-volatile solute as a spherical
residue. The final dry particle size can be calculated from the
droplet size through the known concentration of the liquid solution,
During each test, when it had been determined that particles
of the correct size were being generated, each cascade impactor
was allowed to operate for the required length of time to collect
a suitable sample. Except for one impactor, non-isokinetic
sampling was performed; however, it was determined by a series of
tests that this did not affect the collection efficiency as com-
pared to isokinetic sampling results. It is likely that the
nozzle wall losses were influenced, however. After pulling the
sample, the quantity of particulate matter on each surface was
determined by absorption spectroscopy. With the mass on each
plate and surface known, the wall losses and stage collection
efficiencies could be calculated.
A Pressurized Collison Nebulizer System similar to that
reported by Calvert— was assembled as shown in Figure 55. A
stream of dry dilution air was mixed with the nebulized suspen-
sion to reduce the aerosol concentration and to aid in drying.
Valves placed upstream of the impactor allowed variations in the
flow rate. Three sizes of PSL particles were used (2.02 pm,
0.82 ym, and 0.46 micrometer diameter). Each impactor stage was
tested at three flow rates near the nominal or designed impactor
flow rate. Thus nine calibration points were obtained for each
stage of each impactor. A Climet Instruments Model 208A Particle
Analyzer was used to monitor the impactor inlet and outlet con-
centrations. Stages of each impactor were tested individually.
This system was designed to allow two different air flow
strategies depending on whether the impactor flow rate was
higher or lower than the Climet inlet flow rate. Because of the
limited availability of large PSL spheres, and because these data
were supplementary to the ammonium fluorescein data, only the
impactor stages for which information could be obtained with
sizes of 2 micrometers diameter and smaller were tested.
Generally these were the lower 3 or 4 impactor stages. For sim-
plicity and convenience, mass flowmeters were used to measure
the critical gas flow rates.
102
-------
PRESSURE GAUGE
O
OJ
I MASS FLOWMETER
THREE-WAY VALVE
DIFFUSIONAL DRYER
BLEED VALVE
ABSOLUTE
FILTER
MASS FLOWMETER
| ,nPRESSURE
\J/ GAUGE
CLIMET PARTICLE
ANALYZER
COLLISON
ATOMIZER
DILUTION .
AIR ROTAMETER ^^ COLLISON ROTAMETER
VALVE
DRYER
VALVE
AUXILIARY PUMP
COMPRESSED AIR LINE
REGULATOR ^~~* ABSOLUTE FILTER
DRYER
REGULATOR ABSOLUTE FILTER
Figure 55. PSL calibration system for high and low flowrate impactor;
-------
Results of the Calibration Study
Wall Losses;
During the portion of the calibration procedure using
ammonium fluorescein aerosols, data on wall losses were tabulated.
By washing each cascade impactor surface after sampling a test
aerosol, it was possible to obtain information on particle losses
occurring in nozzles, inlet cones, jet plates, and other internal
surfaces. The total wall loss was calculated as a percentage of
the total amount of aerosol entering the impactor.
The results are shown in Figure 56. All wall loss data are
based on the results of non-isokinetic sampling, except for the
Sierra impactor operated at 7 LPM. The degree to which non-iso-
kinetic sampling influenced the nozzle wall loss contribution is
unknown; however, on the average the nozzle wall loss was about
40% of the total wall loss.
In general, wall losses tend to decrease with particle size
and are negligible for particles smaller than about 1-2 micro-
meters in diameter. The majority of the losses occurred in the
nozzles and inlet cones.
Wall losses can be attributed to particle settling, diffusion,
electrostatic attraction, bounce, and reentrainment. Visual in-
spection of the nozzles and inlet cones indicate that the losses
were predominately due to settling. All impactors except the
Brink were run in a horizontal position.
Calibration Data - Efficiency vs. /ij7;
Recall from the Introduction that if two or more impactor
stages have the same Reynolds Number, jet to plate spacing, and
jet throat length, then according to Marple's theory™ they should
have the same particle collection efficiency for the same magni-
tude of the square root of the inertial impaction parameter. If
these physical parameters differ, or if different collection sub-
strate media are used, then these curves may not be identical.
These differences are evident in Figures 57 through 63, which
present in alphabetical order the Stage Collection Efficiency
Versus /i|7 for each cascade impactor configuration. These effi-
ciencies are based solely on the particles actually collected by
an impaction surface and those found downstream of that surface.
Particles on other surfaces, such as the jet plate surface above
the impaction surface, were not included in the efficiency shown
in these figures.
The differences in /ijJso for the different impactor stages
shown in the foregoing figures require that empirical calibrations
104
-------
_i
<
o
/o
60
50
40
30
20
1 1 1 1 1 1 1 1
—
—
V
"~ O
A
0 o
o
1 1
O
y
A
O
D
8
A
O
A
—
—
— •
—
—
10
*
o
§ »a
9
V
2
1
0.5
0.2
0.1
L. v _
a
— ~—~
— —
^_ __
I I I I I I I I I I I
1.5
3 4 56789 10
PARTICLE DIAMETER, micrometers
15
20
O ANDERSEN MARK III STACK SAMPLER. NONISOKINETIC SAMPLING.
O MODIFIED BRINK MODEL BMS-II CASCADE IMPACTOR. GLASS FIBER SUBSTRATES. NONISOKINETIC SAMPLING.
A MODIFIED BRINK MODEL BMS-II CASCADE IMPACTOR. GREASED COLLECTION PLATES. NONISOKINETIC SAMPLING.
V MRI MODEL 1502 INERTIAL CASCADE IMPACTOR. NONISOKINETIC SAMPLING.
O SIERRA MODEL 226 SOURCE CASCADE IMPACTOR. 14LPM. NONISOKINETIC SAMPLING.
• SIERRA MODEL 226 SOURCE CASCADE IMPACTOR. 7LPM. ISOKINETIC SAMPLING.
A U. of W. MARK III SOURCE TEST CASCADE IMPACTOR. NONISOKINETIC SAMPLING.
Figure 56. Total wall loss vs. Particle diameter.
105
-------
o
z
UJ
o
111
o
UJ
8
99.8
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5
0.2
TT
I IT
0.03
STAGE SYMBOL
I I »
0.05 0.08 0.1
0.2
0.4 0.6 0.8 1.0
2.0
Figure 57. Collection efficiency vs. ,JW . Andersen Mark III
Stack Sampler with glass fiber collection substrates.
106
-------
a?
LU
O
LL
UL
LU
LU
O
O
99.8
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5i
0.2
0.03
I I
STAGE SYMBOL
I TT
I
I I
I
i I I
0.05 0.08 0.1
0.2
0.4 0.6 0.8 1.0
2.0
Figure 58. Collection efficiency vs. y^" • Brink Model BMS-11 Cascade
Impactor with glass fiber collection substrates.
107
-------
99.8
0.2
0.03 0.05 0.080.1
0.2
Figure 59. Collection efficiency vs. \f$. Brink Model BMS--11 Cascade
Impactor with greased collection plates.
108
-------
99.8
*.!
o
z
LU
O
LL
LL.
Ill
LU
Oi
o1
0.03
0.05 0.08 0.1
Figure 60. Collection efficiency vs. -^F Mfif Model 1502 Inertia!
Cascade Impactor with greased collection plates.
109
-------
99.8
99
98
95
90
O
80
70
iZ 60
LL
m 50
O 40
§ 30
_i
20
O
O
10
2
1
0.5
0.2
I I
STAGE SYMBOL
I TT
i
2
3
4
5
6
I
I I
I
I I I
0.03 0.05 0.08:0.1
0.2
0.4 0.6 0.8 1.0
2.0
Figure 61. Collection efficiency vs. -y/~* . Sierra Model 226 Source
Cascade Impactor with glass fiber collection substrates.
Sampling flow rate is 14 LPM.
110
-------
u
—
o
u.
LL
111
O
o
LU
O
U
99.8
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5 U-
0.2
II
STAGE SYMBOL
1 I I I
1
2
3
4
5
6
_L
1
0.03 0.05 0.080.1
0.2
0.4 0.6 0.8 1.0
2.0
Figure 62. Collection efficiency vs. -J~&. Sierra Model 226 Source
Cascade Impactor with glass fiber collection substrates.
Sampling flow rate is 7 LPM.
Ill
-------
o
z
UJ
o
LU
z
o
o
99.8
99
98
95
90
80
70
60
50
40
30
20
10
5
1
0.5
0.2
0.03
Tl
STAGE SYMBOL
I I
I
I I I
0.05
0.08 0.1
0.2
0.4 0.6 0.8 1.0
2.0
Figure 63. Collection efficiency vs. ^/^. University of Washington
Mark III Source Test Cascade Impactor with greased
collection plates.
112
-------
be done. No existing theory is comprehensive enough to compensate
for the variations in geometry and application techniques for
single devices, among different devices, or among different users.
A difference can also be seen between the shapes of the
empirical calibration curves and the curves predicted by Marple's
theory. The empirical curves show a smooth tail which approaches
zero for small values of Sty. Marple's theory, however, predicts
a sharp intersection between the efficiency curves and the
abscissa. The "tails" are possibly due, at least in part, to the
fact that the calibration aerosols cannot be made perfectly mono-
disperse, and always contain multiplets.23 Such "tails" are
probably unavoidable when using the VOAG, because these devices,
at best, produce about 4% doublets at all times.
The graphs also indicate a more severe problem, common to all
cascade impactors. On a majority of the stages the collection
efficiency does not reach 100% at any value as would be expected
theoretically. Also, after reaching a maximum point, the curves
fall off to lower efficiencies at higher Sty values. This means
that some large particles will not be collected on upper stages
and will be passed through the impactor to lower stages or even
the back up filter. An accurate knowledge of this type of
behavior is essential to the proper design and application of
cascade impactors.
The theoretical data of Marplea 8 plotted in Figure .53 show
shifts in the square root of the Stokes Number, Sty, which depend
upon the jet to plate spacing ratio (S/Dj) and the Reynolds
Number (Re). The values of Sty corresponding to a given particle
collection efficiency increase for increasing jet to plate spacing
ratio and decrease for increasing Reynolds Number.
For the Andersen Mark III the values of the jet to plate
spacing ratio change from about 1.5 to 10 between Stages 1 and 8.
Theoretically this should cause the /Stk curve to shift to larger
values and become much steeper. The Reynolds Number changes from
about 45 to 500 between Stages 1 and_8. Theoretically this should
cause a shift to the left in the /S^tk values as well as causing
them to be steeper.
Although it is difficult to pinpoint the exact cause for the
shifts in the Andersen data shown in Figure 57, the general shift
to smaller Sty values and the corresponding increase in steepness
could be due, in part, to a combination of these effects due to
jet to plate spacing ratio and Reynolds Number,
113
-------
The values of these parameters for the other impactors
indicate that the effects would be much more difficult to
separate and characterize because of the smaller changes in
these values.
Conclusions
Five commercially available cascade impactors have been
calibrated during this study. The method of calibration and the
presentation of the results should make this data useful to both
the field operator of these devices as well as those interested
in the theory of impactor design.
Based on this work, several conclusions can be drawn.
1. The value of /ijTTo" for each stage of a multiple stage
impactor may be different. Prior to the current awareness of
the importance of impactor calibration, it was the practice of
many cascade impactor manufacturers and users to assume that the
value of /4>50 for every stage was identical. In many cases the
experimental value determined by Ranz and Wong17 was used.
Attempts to perform calibrations were not comprehensive. The
theories of cascade impactor operation at this time do not des-
cribe the behavior of cascade impactors accurately enough to
make it unnecessary to calibrate each device empirically.
2. The stage collection efficiencies are sensitive to the
type of impactor collection substrate which is used. This is
evident in the comparison of the Brink Cascade Impactor data
using glass fiber collection substrates and greased collection
plates. This strong dependence of collection efficiency on
stage collection substrate material has also been explicitly
illustrated and discussed by Willeke2"* and Rao.15
3. In the majority of cases the stage collection efficiency
never reaches 100% for any particle size but reaches a maximum
value that usually falls between 80% and 95%. This implies that
some greatly oversized particles will reach every stage beyond
the first stage. Unless suitable compensation can be made for
the presence of these oversized particles, their presence will
tend to bias the apparent particle size distribution toward
higher than actual concentrations of fine particles and reduced
concentrations of large particles. These errors probably tend
to be more significant for the fine particle end of the distri-
bution.
114
-------
4. Ideally, an impactor stage should reach 100% collection
efficiency for some particle size and stay at that value for all
larger particle sizes. In practice, however, this is not the
case as demonstrated by this study. In general, the stage col-
lection efficiency reaches a maximum less than 100% and then rolls
off and decreases for particles larger than a certain size. This
is attributed to the fact that these larger particles strike the
plate with appreciable momentum, bounce, and are thus carried to
a lower stage. The use of grease on the collection plates as well
as a reduction in the impactor flow rate tends to decrease the
magnitude of this problem. The Sierra impactor data illustrate
the increase in collection efficiency which resulted from a
decrease in sampling flow rate and concomitant reduction in bounce
and cascading of large particles to lower stages. This study
shows that in some cases the maximum efficiency was almost doubled
by lowering the flow rate. A discussion of this phenomenon has
also been presented by Rao.16
A more detailed description of this calibration study and
its results has been published in a document entitled "Particle
Sizing Techniques for Control Device Evaluation: Cascade Impactor
Calibration." EPA 600/2-76-280. NTIS PB-262-849 ($5.00).
Soviet Impactor-Cyclone Calibration
(Technical Directive Number 20201)
Description of Task;
Under a joint U.S.-Soviet Technological Information Exchange
Program four Soviet built cascade impactors were sent to Southern
Research Institute to be calibrated. The upper stages were
calibrated using aerosols generated with a Vibrating Orifice
Aerosol Generator. The lower stages were calibrated using poly-
styrene latex spheres dispersed with a Pressurized Collison
Nebulizer System.
Summary of Progress;
Under this task three Soviet cascade impactors and one Soviet
impactor/cyclone were calibrated. The three cascade impactors in-
cluded one twelve stage device and two fourteen stage impactors.
The three stage impactor/eyelone had a single impaction stage
followed by two cyclonic stages. All four devices had an integral
back-up filter holder which uses a plug of glass wool fibers.
A photograph of one of the fourteen stage Soviet impactors is
shown in Figure 64. Figure 65 shows the Soviet impactor/cyclone.
115
-------
. Soviet 14-Stage Cascade Impactor.
Figure 65. Soviet 3-Stage Impactor-/Cyclone.
116
-------
The upper stages of these sizing devices were calibrated
using ammonium fluorescein aerosols (20 ym to 2 ym) dispersed
by a Vibrating Orifice Aerosol Generator. The lower stages were
calibrated using monodisperse polystyrene latex spheres (2.0 ym
to 0.46 ym) dispersed by a Pressurized Collison Nebulizer System.
Soviet Three Stage Impactor/Cyclone
As mentioned above, this instrument consists of a single
impaction stage followed by two cyclonic stages. A back-up
filter collected all particles passing the last stage. This
device was constructed from titanium. A set of nozzles was
supplied with the impactor/eyelone. All three stages were
calibrated using monodisperse ammonium fluorescein particles
with sizes ranging from 18 micrometers diameter to 2.3 micro-
meters diameter. The ambient pressure was 29.5" Hg, the
temperature was 22°C, the particle density was 1.35 gm/cm3, and
the sampling flow rate was 10 liters per minute. At these
conditions the cut points of the three stages were determined
to be 13.5, 6.4, and 2.6 micrometers. The results are presented
graphically in Figure 66.
Soviet 12-Stage Cascade Impactor
This impactor is uniquely designed in that several stages
are designed to have identical cut points. Because of this
feature, the impactor has seven effective stages. The twelve
stages are paired as follows: 1 and 2, 3 and 4,5,6 and 7, 8,
9 and 10, 11 and 12. In field use the mass collected by single
stages 1 and 2 is combined as the catch for effective stage 1.
In a similar manner the mass collected by single stages 11 and
12 becomes the combined catch for effective stage 7. The cali-
bration conditions were an ambient pressure of 29.5" Hg, a
temperature of 22°C; the particle density was 1.35 gm/cm3 and
the sampling rate was ten liters per minute. Because of the
type of construction of this impactor, it was only possible to
calibrate it using ammonium fluorescein aerosols between 2 and
20 micrometers diameter. The calibration data are shown by
stage both on an individual basis (Figure 67) and on an effective
stage basis (Figure 68).
Soviet 14-Stage Cascade Impactor (Small DSQ's)
Soviet 14-Stage Cascade Impactor (Large DSO'S)
These cascade impactors are similar in design to the Soviet
12 Stage impactor in that their 14 single stages are divided into
seven effective stages with a pair of single stages per effective
stage. Each member of a pair is theoretically designed to have
the same cut point. These seven effective stages are numbered
117
-------
100
90
.5 .6.7.8.91.0
56789 10
20
Particle Diameter, Micrometers
Figure 66. Co/lection efficiency vs. Particle diameter. Soviet 12~Stage
Impactor -/Cyclone.
1—lmpaction Stage 2— 1st Cyclone 3— 2nd Cyclone
(29.5 in. Hg, 2FC, 1.35 gm/cm3, 10 LPM)
118
-------
<*>
o
0)
•H
O
w
o
•H
4J
0
O
o
100
90
80
70
60
50
40
30
20
10
I I I I I
I T
l l I I I
L
\
i i
.5 .6.7.8.91.0 2 3 4 5 6 7 89 10
Particle Diameter, Micrometers
Figure 67. Collection efficiency vs. Particle diameter. Soviet 12-Stage
Cascade Impactor.
Data shown for the first nine stages from 2 to 20 microns.
(29.5 in. Hg, 22°C, 1.35 gm/cm*, 10 LPM)
20
119
-------
100
90
80
o 70
Q>
•H
•H 60
W
c 50
o
••H
+J
O
0) 40
o
u
30
20
10
lilt
x-9&10
.5 .6.7.8.91.0 2 3 4 5 6 7 89 10
Particle Diameter, Micrometers
Figure 68. Collection efficiency vs. Particle Diameter. Soviet 3--Stage
Cascade Impactor.
Data shown in combined form in the configuration for
field measurements.
(29.5 in. Hg, 22°C, 1.35 gm/cm3, 10 LPM)
20
120
-------
1.1/1.2, 2.1/2.2, 3.1/3.2, 4.1/4.2, 5.1/5.2, 6.1/6.2, and 7.1/7.2.
In practice the mass collected by individual stages 1.1 and 1.2,
for example, is combined to give a total mass collected by the
first effective stage. This procedure is repeated for the rest
of the impaction stages. These two impactors have different cut
points but share identical stages on five of the seven effective
stages as illustrated in the following chart.
14-Stage Cascade Impactor
Small DSQS Large Dsos
1.1/1.2
2.1/2.2
3.1/3.2
4.1/4.2
5.1/5.2
6.1/6.2
7.1/7.2
Effective stages 1, 2, 3, 4, and 5 of the small Dso impactors
are identical with stages 1, 4, 5, 6, and 7 of the large Dso im-
pactor, respectively. It was possible to calibrate both impactors
with ammonium fluorescein and polystyrene latex particles; how-
ever, it was not possible to complete the calibration on stages
7.1/7.2 of the small DSQ impactor since the pressure drop across
these stages was greater than our calibration apparatus could
maintain. The results of the calibration are depicted graphi-
cally in Figures 69 and 70.
121
-------
100
90
80
o 70
c
0)
•l-l
•H 60
HI
§ 50
•H
•P
O
0) 40
o
u
30
20
10
.5 .6.7.8.91.0 2 3 4 56789 10
Particle Diameter, Micrometers
20
Figure 69. Collection efficiency vs. Particle diameter. Soviet 14-Stage
Cascade Impactor. (Small Cutpoints)
Data shown in combined for for the first six of the seven
stage pairs. 0
(29.5 in. Hg, 22 C, 1.35 gm/cmj, 10 LPM)
122
-------
100
90
10
.5 .6.7.8.91.0 2 3 4 5 6 7 89 10
Particle Diameter, Micrometers
Figure 70. Collection efficiency vs. Particle diameter. Soviet 14~Stage
Cascade Impactpr. (Large Outpoints)
(29.5 in. Hg, 22°C, 1.35 gm/cm3, 10 LPM)
20
123
-------
Task Category - Cyclones
Develop and Evaluate Five Stage Series Cyclone System
(Technical Directive Number 10602)
Description of Task;
Based on the work previously done under Contract No. 68-02-
0273, a series cyclone system was designed and constructed. The
design goals were 5 cut points and compactness to fit into a
four inch port.
1. The cyclones were calibrated using laboratory aerosols.
This was done over the range of temperatures expected for
field sampling.
2. Minor modifications were made to the prototype and
recalibrated.
3. Tests will be performed at a suitable site to compare the
performance of the cyclone with cascade impactors.
4. A special report on the design, calibration, evaluation,
and operation of the system is being written.
Progress Summary;
A Five Stage Series Cyclone has been designed and a
schematic is shown in Figure 71. It has a nominal sampling
rate of 472 cm3/sec (1.0 ft3/min). Both a black anodized
aluminum and a titanium prototype have been constructed. Pre-
liminary calibration results for the five cyclones are shown in
Figure 72.
A more extensive calibration of these cyclones has also
been completed. The effect on cyclone behavior due to changes
in flow rate, gas temperature, and particle density was studied.
The preliminary results are shown in Table X. Work is also in
progress on developing a theory of cyclone behavior based on
these experimental data.
A complete field test of these prototypes is yet to be
performed.
124
-------
CYCLONE 1
CYCLONE 4
CYCLONE 5
CYCLONE 2
CYCLONE 3
OUTLET
INLET NOZZLE
Figure 71. Five Stage Series Cyclone System.
125
-------
a?
o
z
UJ
u.
u.
UJ
u
UJ
O
U
100
90
80
70
60
50
40
30
20
10
i rpM f
I I I I
I I I
0.2 0.3 0.40.50.60.8 1.0 2 3 4 5 6
PARTICLE DIAMETER, micrometers
• FIRST STAGE CYCLONE
• SECOND STAGE CYCLONE
£ THIRD STAGE CYCLONE
V FOURTH STAGE CYCLONE
O FIFTH STAGE CYCLONE
8 10
20
Figure 72. Laboratory calibration for the Five Stage Series Cyclone System.
(472 cnrr/sec, particle density— 1.0 gm/cm3)
126
-------
Table X
Laboratory Calibration of the Five Stage Series Cyclones -
Values of Dso cut points (in micrometers) for various conditions
of sample flow, temperature, and particle density.
v-y«j J.UJIB
Particle Density
(gm/cm3)
Flow Temperature
(LPM) (°F)
7.0 77
14.2 77
28.3 77
28.3 200
28.3 400
i
2.04/1.00
5.9/8.4
3.8/5.4
4.4/-
6.4/-
JLJ.
2.04/1.00
2.4/3.5
1.5/2.1
2.3/-
2.9/-
J.J.J.
2.04/1.00
0.96/1.4
1.2/-
1.9/-
XV
2.04/1.00
-/2.5
1.0/1.5
0.44/0.63
V
2.04/1.00
-/I. 5
0.59/0.85
0.22/0.32
127
-------
Calibration of the Source Assessment
Sampling System Cyclones
(Technical Directive Number 20302)
Description of Task;
The purpose of this task was to calibrate the SASS cyclones
using both ammonium fluorescein aerosols (20 ym to 2.0 ym) and
polystyrene latex spheres (2.0 ym to 0.46 ym). The calibration
accuracy was to be sufficient to determine the cut point of the
Middle Cyclone (nominally 3.0 ym) to within ±0.5 micrometers
diameter.
Summary of Progress ;
The three (3) cyclones of the Source Assessment Sampling
System (SASS) were calibrated as described below. In the follow-
ing discussion the terms Large, Middle and Small Cyclone refer to
cyclones designed for room temperature DSQ values of 10 ym, 3 ym,
and 1 ym, respectively.
The calibrations of the Large and Middle Cyclones were per-
formed using ammonium fluorescein aerosols generated with Southern
Research Institute's Vibrating Orifice Aerosol Generator (VOAG).
Monodisperse ammonium fluorescein aerosols with diameters of 2, 3,
4, 5, 7.5, 10.5, and 14.5 micrometers were sampled at flow rates
of 4 ACFM and 3 ACFM. These particles have a density of 1.35
gm/cm3. After sampling for a sufficient length of time, each
cyclone was washed with a known amount of NH^OH to dissolve the
ammonium fluorescein aerosol. The total mass of the collected
aerosol was determined using absorption spectroscopy. The
collection efficiency of each cyclone was calculated and plotted
versus particle diameter. Figures 73 and 74 present these
Collection Efficiency Versus Particle Diameter data for the Large
and Middle Cyclones, respectively.
The Small Cyclone was calibrated using Dow Corning PSL
spheres dispersed with the Institute's Pressurized Collison
Nebulizer System. Using an auxiliary pump, aerosols were pulled
through the Small Cyclone at two flow rates, 3.1 ACFM and 1.8 ACFM.
A Climet Instruments Model 208A Particle Analyzer was used to
measure the number concentration of unit density 0.82 ym, 1.1 ym,
and 2.2 ym diameter PSL spheres upstream and downstream of the
Small Cyclone. The collection efficiencies were calculated and
plotted versus the particle diameter as shown in Figure 75.
128
-------
to
VO
u
ui
O
U
UI
O
U
100
90
80
70
60
50
40
30
20
10
.3
1 I I I I I I
"II TTT
I I Mill
4 ACFM
r
/
3 ACFM
I Kl I I I I
.4 .5 .6 .7 .8 .9 1.0
6789 10
20
PARTICLE DIAMETER, MICROMETERS
Figure 73. Collection efficiency vs. Particl&diameter.
Particle Density-—J.35gm/cm3
Large SASS Cyclone.
-------
100
90
80
70
o
2, 60
o
01
O
01
50
40
O 30
20
10
0
I I , I I/I I I
.3 .4 .5 .6 .7 .8 .9 1.0 2 3 456789 10
PARTICLE DIAMETER, MICROMETERS
20
Figure 74. Collection efficiency vs. Particle.diameter. Middle SASS Cyclone.
Particle Density ..... 7.35 gm/cmr
-------
u
UJ
O
u.
uu
O
O
ui
O
U
100
90
80
70
60
50
40
30
20
10
1 I I I I I I
3.1 ACFM
I 1 I
J/~~l I I I I I I
1.8 ACFM
I I I I I I I I
.3 .4 .5 .6 .7 .8 .9 1.0 2 3 456789 10
PARTICLE DIAMETER, MICROMETERS
Figure 75. Co/lection efficiency vs. Particle diameter. Small SASS Cyclone.
Particle Density—1.35 gm/cmr
20
-------
Based on the results of this limited calibration study, it
appeared that the cyclone Dso cut point varied approximately
inversely with the square root of the particle density as pre-
dicted by theory.
The specific data obtained in this study were extrapolated
to obtain cyclone Dso's for four combinations of flow rate and
particle density (4 ACFM and 3 ACFM flow rates and 1.00 gm/cm3
and 2.3 gm/cm3 particle densities). The resulting graphs shown
in Figure 76 indicate the approximate range of DSQ cut points
which can be expected at temperatures near ambient.
Recalibration of the Three SASS Cyclones
(Technical Directive Number 20402)
Description of Task;
The three Source Assessment Sampling System cyclones were
recalibrated at a sampling rate of 4 SCFM and a gas temperature
of 400°F. There were two objectives of this task: (1) to
establish the cut points of the three SASS cyclones as they
presently exist, and (2) to determine what modifications and
design parameters were required to achieve cut points of 10, 3,
and 1 micrometers at a sampling rate of 4 SCFM and a gas tempera-
ture of 400°F.
Summary of Progress;
A recalibration of the three (3) cyclones of the Source
Assessment Sampling System (SASS) at 400°F and 4 SCFM was per-
formed. All data presented here have been corrected to unit
density aerosols. In the following discussion the terms Large,
Middle, and Small Cyclone are used to indicate the nominal 10 ym,
3 ym, and 1 ym Dso cyclones, respectively. As reported under
Technical Directive Number 20302, the previous calibration at
75°F and 4 ACFM gave approximately 0.86 ym, 3.5 ym, and 11.0 ym
D50 for the Small, Middle, and Large Cyclones, respectively.
The calibration of the Large and Middle Cyclones was per-
formed using ammonium fluorescein aerosols generated with
Southern Research Institute's Vibrating Orifice Aerosol Generator
(VOAG). With the cyclones placed in a heated oven and using a
heated inlet line, the temperature of the gas stream at the inlet
to the Large Cyclone was maintained at 400°F. Particle integrity
of the ammonium fluorescein at high temperature was a major
problem. It appears that rapid heating of aerosol particles
132
-------
2.8 -TTT
u>
(jj
••H
(0
c
(1)
Q
0)
rH
O
•H
4*
M
m
.5
2 345
Micrometers
10
20
30 40
F/jyt/re 7& S>4SS cyclone cut points. (At 23 C.)
-------
which had not dried sufficiently after generation caused these
particles to explode creating a large concentration of contami-
nating small particulate matter. This problem was alleviated by
allowing the aerosol to come up to temperature more slowly. We
also observed a color change in the ammonium fluorescein before
and after heating. Microscopic observation also indicated a
possible crystalline change, causing us to question the integrity
of the dry ammonium fluorescein particles.
The Large Cyclone was modified to try to obtain a DSO closer
to the desired 10 micrometers. This modification involved the
removal of the vortex buster from the Large Cyclone outlet.
Unfortunately, there was no apparent effect on the performance
of this cyclone. The data for the Large Cyclone is shown in
Figure 77.
The Middle Cyclone was also modified in an attempt to obtain
a Dso closer to the desired 3 micrometers. This was done by
reducing the Middle Cyclone inlet diameter from 0.62 inches to
0.53 inches. The effect of this change was minimal as can be
seen in Figure 78 and Figure 79 for the Unmodified and Modified
Middle Cyclone, respectively.
The approximate Dso"s for the Large and Middle Cyclones at
400°F and 4 SCFM are 15 micrometers and 4.4 micrometers, respec-
tively.
After considering the data obtained during this recalibration,
it was felt that the desired DSO'S might possibly be obtained by
removal of the vortex busters in the collection cups of the Large
and Middle Cyclones. The resulting large tangential velocities
near the walls of these cups should remove a sufficient number of
particles to cause a significant and measurable change in the cut
points. However, it was felt that a definitive characterization
of the behavior of these cyclones could only be performed if the
cyclones were kept for approximately two months. This would
allow time for careful consideration of solutions to problems
which might arise during the calibration procedure.
On September 14, 1976, a meeting was held at Research Triangle
Park to discuss the results of the recalibration of the SASS train
cyclones at 4 SCRM and 400°F. It was decided that the results of
this study were not sufficiently conclusive to recommend changes
in the cyclone construction to obtain the desired cut points of
10, 3, and 1 micrometer at operating conditions of 400°F and 4
SCFM. Data were presented indicating possible physical changes
in the ammonium fluorescein aerosol at high temperature. Also the
shift in the cyclone calibration curves at these conditions was
not expected based on current cyclone operation theories.
134
-------
O
z
UJ
U
uT
u.
ui
u
UJ
O
U
100
90
80
70
60
50
40
30
20
10
1 I Mill
1 I I I I I I I
I I I I I I I
1
.3 .4 .5 .6 .7 .8 .9 1.0 2 3 456789 10
PARTICLE DIAMETER, MICROMETERS
Figure 77. Collection efficiency vs. Particle diameter. Large SASS Cyclone.
(4 SCFM, 40(fF, LOOgm/crt3)
20
-------
a*
>
z
ui
O
u.
u.
UI
g
u
UI
o
o
100
90
80
70
60
50
40
30
20
10
0
IT
.3
TT
1 1 l I i I I
1
O
1 1
.5 .6 .7 .8 .9 1.0 2 3 456789 10
PARTICLE DIAMETER. MICROMETERS
20
Figure 78 Collection efficiency vs. Particle diameter. Unmodified Middle SASS Cyclone.
(4 SCFM, 400*F, 1.00gm/cmJ)
-------
100
1 I I I I I I
1 \ I I I I I I
90
80
u>
U
ui
o
Z
u.
ui
§
g
Ul
_l
_l
O
U
70
60
50
40
30
20
10
I I I I I I I
I
I
I I I I I II
.3 .4 .5 .6 .7 .8 .9 1.0 2 3 456789 10
PARTICLE DIAMETER, MICROMETERS
20
Figure 79 Collection efficiency vs. Particle diameter. Modified Middle SASS Cyclone.
(4 SCFM, 400°F, 1.00 gm/cmj)
-------
It was concluded that calibration with aerosols which could
withstand these high temperatures should take place. Technical
Directive Number 20502 was issued to complete this assignment.
High Temperature Recalibration and Modification
of SASS Cyclones
(Technical Directive Number 20502)
Description of Task;
The purpose of this task was to determine a material which
would be suitable as a high temperature aerosol (400°F). After
finding such a material, it was to be used to generate aerosols
to be sampled by the SASS cyclones in a calibration procedure.
Approximately ten to twelve monodisperse aerosols were to be
generated using a Vibrating Orifice Aerosol Generator. After an
initial calibration, any necessary modifications would be per-
formed and the cyclones would be r.etested. The major emphasis
was placed on calibration of the Middle Cyclone (3 ym DSO). If
necessary, testing at 75°F, 200°F, and 350°F and extrapolating
data to 400°F would be acceptable.
Summary of Progress:
The goal of this task was to calibrate and modify, if
necessary, the Source Assessment Sampling System Middle Cyclone
so that it would have a 3 ± 0.5 ym D5Q cut point for 1.00 gm/cm3
particles when operated at 400°F and 4 SCFM. Figure 83 shows
the collection efficiency of the SASS Middle cyclone when
sampling a turquoise dye (particle density = 2.4 gm/cm3) at the
above conditions with and without the vortex buster in the cup.
The DSQ of the cyclone with the vortex buster in place is 3.4 ym
for a particle of density 2.04 gm/cm3, or 4.0 ym for a particle
of density 1.00 gm/cm3. The DSO °f the cyclone with the vortex
buster removed is 2.5 ym for a particle of density 2.04 gm/cm3,
or 3.5 ym for a particle of density 1.00 gm/cm3. Thus, with the
vortex buster removed, the cyclone satisfactorily attained the
original goal.
Experimental Procedure
A decision was made to calibrate the Source Assessment
Sampling System Middle Cyclone at 400°F and 4 SCFM. Because
previous tests have indicated ammonium fluorescein was unstable
at 400°F, a search was initiated for an aerosol with some or all
of the following characteristics:
138
-------
Non-toxic
Stable at temperatures up to 500°F or above
Soluble in water or other non-toxic, non-residue forming
solvent
Amorphous - dries to form solid, homogeneous spheres when
dispersed in solution from a VOAG.
Known or easily measured density
Has a definite, distinct absorption spectrum peak for
absorption spectroscopy measurement between 400 NM
and 900 NM.
The initial search for such an aerosol was unsuccessful, so,
ammonium fluorescein was used to determine the DSQ cut points of
the Middle Cyclone at 70°F, 200°F, and 350°F and 5.45 ACFM and
the data obtained from these tests was used to extrapolate to
determine the DSQ cut point at 400°F. This method proved to be
difficult when it was found that ammonium fluorescein particles
smaller than 4 ym in diameter were unstable at 350°F. Attempts
to alleviate this problem were largely unsuccessful. Meanwhile
the search for an acceptable aerosol was continued using
commercially available dyes.
Of several samples from three chemical companies, du Font's
"Pontamine" Fast Turquoise 8 GLP dye was the first found to
satisfactorily meet the requirements listed above. A spectral
analysis performed on a dilute water solution of this dye indi-
cated a distinct absorption peak at 622 nanometers. Measurements
with a Helium-Air pycnometer gave a density of 2.04 gm/cm . The
sample seemed pure and its stability at 400°F was excellent. The
expansion problems encountered with small diameter ammonium
fluorescein particles were absent. Aerosol particles made from a
solution of the dye in distilled water were very nearly, if not
perfectly, round.
The calibration of the SASS train cyclones was performed
using the Institute's Vibrating Orifice Aerosol Generator (VOAG).
The VOAG generated monodisperse ammonium fluorescein particles and
turquoise dye particles with diameters from 2 micrometers to 7
micrometers.
Throughout the testing, close watch was kept on the temperature
and flow rate of the aerosol stream. Any discrepancies were quickly
corrected, and readings of all temperatures were recorded periodi-
cally. Therefore repeatability of the tests and test results was
insured.
139
-------
After each test, the cyclone and filter substrate were washed
to dissolve and rinse off all the aerosol particles. The wash
solutions used were 0.1 N NHitOH for ammonium fluorescein and a
sodium bicarbonate solution for turquoise dye.
A Bausch and Lomb Spectronic 88 Spectrophotometer, calibrated
with solutions of known concentration of the aerosol solute (tur-
quoise dye or ammonium fluorescein) was used to measure the
absorbance of the wash from the cyclone and the filter. From
knowledge of the amount of wash solution, the dilution factor, if
any, and the absolute concentration, the mass of particles in the
cyclone and on the filter was calculated. With these two masses
known, the collection efficiency of the cyclone for that particular
particle size was calculated.
Results
Table XI lists the DSQ cut points of the cyclone at various
conditions. Figure 80 shows the collection efficiency curves of
the cyclone when calibrated with ammonium fluorescein particles
at 70°F, 200°F, and 350°F. The Dso's obtained from these curves
and plotted in Figure 81 indicate that the Dso-gas viscosity
relationship is linear. This does not correspond to Lapple's25
prediction that Dso would vary directly with the change in the
square root of the gas viscosity.
Figure 82 shows the collection efficiency of the cyclone for
ammonium fluorescein particles and turquoise dye particles when
collected under similar conditions. The relative differences
between the two aerodynamic Dso cut points derived from the two
curves differ from the prediction of Lapple's equation by only
7%.
Figure 83 shows the collection efficiency curves of the
cyclone when calibrated with turquoise dye particles with and
without the vortex buster in place. It was determined using
Lapple's equation that the Dso cut point of the cyclone with the
vortex buster removed is 3.5 ym for 1.00 gm/cm3 particles, which
is within the acceptable range.
Calibration of SASS Cyclones for HERL
(Technical Directive Number 21002)
Description of Task;
Exxon Research and Engineering Corporation has used a Source
Assessment Sampling System at a coal-fuel power boiler in Kentucky
to size the particulate effluent. This work was performed for the
Health Effects Research Laboratory. Southern Research Institute
140
-------
Table XI
Five-Stage Cyclone Calibration Data
Material
Vortex
Buster Temperature
Flow Rate
ft.3/min
Actual/Standard
DSD Physical D5o Aerodynamic
Micrometers Micrometers
Ammonium
Fluorescein
Ammonium
Fluorescein
Ammonium
Fluorescein
Turquoise
Dye
Turquoise
Dye
Turquoise
Dye
IN
IN
IN
IN
OUT
IN
Ambient
200°F
350°F
Ambient
400°F
400°F
5.37/
541 /
5.46/
5.42/
6.50/4.00
6.50/4.00
2.8
3.5
4.2
2.2
2.5
3.4
3.3
4.0
4.9
3.1
3.5
4.9
-------
100
o
O
u
o
o
75
50
25
I I
AMMONIUM FLUORESCEIN
070°F, 5.37ACFM
£200°F. 5.41 ACFM
D350°F. 5.46ACFM
2 3 4 5 6789 10
PARTICLE DIAMETER, micrometers
3630-007
Figure 80. Collection Efficiency - Temperature Relationship
SASS Middle Cyclone
Ammonium Fluorescein
Particle Density = 1.35 gm/cm^
142
-------
I
o
o
in
Q
350°F
200°F,
140
I I
180 220
VISCOSITY, poise x 10'6
260
3630-011
Figure 81. DQQ • Viscosity Relationship
SASS Middle Cyclone
Ammonium Fluorescein
Particle Density = 1.35 gm/cm^
143
-------
100
80
o
o
60
LU 40
O
,0
20
I I I I II
DENSITY COMPARISON
O AMMONIUM FLUORESCEIN'
70°F. 5.45 CFM
° TURQUOISE DYE
70°F, 5.42 CFM
1 2 3 456789 10
PARTICLE DIAMETER, micrometers
3630-008
Figure 82. Collection Efficiency - Particle Density Relationship
SASS Middle Cyclone
Ammonium Fluorescein Particle Density = 1.35 gm/cm^
Turquoise Dye Particle Density = 2.04 gm/cm^
144
-------
100
o
•z
UJ
5
o
8
TURQUOISE DYE
400°F, 4 SCFM
w OWITHOUT VORTEX BUSTER
6WITH VORTEX BUSTER
J I I I I I I I
6 7 8 910
PARTICLE DIAMETER, micrometers
3630-009
Figure 83. Collection Efficiency at 400°F, 4 SCFM
SASS Middle Cyclone
Turquoise Dye
Particle Density = 2.04 gm/cm3
145
-------
calibrated the Middle SASS cyclone used by Exxon. The cyclone
was to be calibrated at the actual run conditions as operated
by Exxon test personnel.
Summary of Progress;
This report contains the results of the Exxon-SASS Cyclone
system calibration. Exxon used the following sampling conditions.
• 600°F inlet gas temperature to the oven
• Oven temperature 375°F
• No filter element in the filter housing in the heated
oven
• Pump flow wide open with filter on pump inlet and
muffler on pump outlet
• Vacuum measured ahead of pump filter - 14 inches
vacuum under sampling conditions
• Stack moisture 8%
• Gas temperature in middle cyclone unknown
• Vortex busters in the collection cups of the Large and
Middle cyclones.
These conditions and those of the calibration system were not
completely compatible. The inlet gas temperature was 450°F, the
constraint being the temperature limit of the calibration aerosol.
The oven temperature was 375°F. The humidity of the air was not
measured.
No probe was supplied with the cyclones; however, a telephone
conversation revealed that Exxon used an Aerotherm probe which
was twelve feet long. The sampling line of this probe is one-half
inch O.D. Stainless Steel tube. The heat loss through this probe
in our lab was unacceptable, and the particle loss would probably
be quite high. Also a filter had to be used in the filter holder
so that all the mass would be caught for collection efficiency
determinations. Therefore, the flow rate at the inlet of the
large cyclone had to be determined. The field set-up was dupli-
cated as closely as possible and the flow was measured to be 13
ACFM at an oven inlet temperature of 399°F. Then the probe was
removed and a glass fiber filter was added to the filter holder.
The flow was adjusted so that the inlet flow to the cyclones was
13 ACFM. There was a definite but unknown amount of water in the
air.
146
-------
A note should be made as to the conditions of the cyclones
when they arrived. There were a few dents in the cup and the
inlet of the Middle cyclone was bent such that the air stream
did not enter the cyclone strictly tangentially. The only modi-
fication made to the cyclones by S.R.I, was replacement of some
teflon gaskets which were badly deformed.
The calibration system used was a vibrating orifice aerosol
generator with du Pont "Pontamine" Fast Turquoise 8 GLP dye as an
aerosol. The collected particles were washed off the cyclones and
filter with a sodium bicarbonate solution and their mass deter-
mined with a spectrophotometer.
For a temperature of 450°F, a flow of 13 cfm, and a particle
density of 2.04 gm/cm3, the Dso cut points of the large and
middle cyclones were 7.6 ym and 2.13 ym respectively (see Figure
84). For a particle density of 1.00 gm/cm3 and the same condi-
tions, the Dso cut point of the large and middle cyclones would
be 11 ym and 3.0 ym respectively.
Estimation of Dso of Middle Cyclone at 600°F
The DSO cut point for the Middle cyclone is 3.0 ym for a
temperature of 450°F, a flow rate of 13 CFM, and a particle
density of 1.00 gm/cm3, with the vortex buster in place. Experi-
mental work with the middle cyclone has indicated the following
may be obtained by extrapolating the DSO vs. viscosity curve.
Flow Rate Dso Temperature Particle Density
5.4 CFM 5.0 ym 375°F 1.00 gm/cm3
5.4 CFM 5.8 ym 450°F 1.00 gm/cm3
5.4 CFM 6.2 ym 600°F 1.00 gm/cm3
Assuming that the D50 cut point will increase at the same rate
with increasing temperature at flow rates of 5.4 and 13.0 CFM,
the following relationships are obtained.
Dso (600°F; 13 CFM) _ Dso (6QO°F; 5.4 CFM)
Dso (450°F; 13 CFM) DSO (450°F; 5.4 CFM)
and
Dso (375°F; 13 CFM) = Dso (375°F; 5.4 CFM)
Dso (450°F; 13 CFM) DSO (450°F; 5.4 CFM)
147
-------
100
I
ONOZZLE
A LARGE CYCLONE
DMIDDLE CYCLONE
80
O
z
UJ
O
LL
H
o
UJ
O
O
60
50
40
20
1.0
Temperature = 450°F
Particle Density = 2.04 gm/cm3
Flow = 13 ACFM
5.0 10.0
PARTICAL DIAMETER, micrometers
20.0
Figure 84. Exxon SASS Cyclones
148
-------
Thus,
Dso (600°F, 13 CFM) = 3.0 x JU| = 3.2 ym.
and
Dso (375°F; 13 CFM) = 3.0 x f-41 = 2.6 ym.
Therefore, the Dso cut points of the middle cyclone for a flow
rate of 13 ACFM are estimated to be 2.6 ym for a temperature of
375°F, and 3.2 ym for a temperature of 600°F.
Estimation of Dso of Small Cyclone at 600°F
The small cyclone has previously been calibrated at two flows
Flow Dso Temperature Particle Density
1.8 ACFM 1.88 77°F 1.00 gm/cm3
3.1 ACFM 1.11 77°F 1.00 gm/cm3
Using this relationship Dso = KQ from Chan and Lippmann26
Dso (1.88) _ K (1.8)N
050 <1- K .
N = -0.97
So
Dso = KQ-°-97
or
1.88 = K (1.8)~°*97
K = 3.32
Assuming the Dso vs. Q relationship above holds at Q = 13 ACFM,
D50 = 3.32Q~°'97
= 3.32(13)~°'97
= 0.277 ym.
149
-------
Recent experimental work with the Five Stage Series Cyclone
II (which has identical dimensions as the SASS small cyclone
except for the width and depth of the cup) has indicated the
following
Flow DSQ Temperature Particle Density
1 CFM 2.30 urn 77°F 1.00 gm/cm3
1 CFM 5.21 ym 600°F 1.00 gm/cm3
1 CFM 4.11 ym 375°F 1.00 gm/cm3
Assuming that the Dso cut point will increase at the same
rate with increasing temperature at flows of 1 and 13 CFM, we
obtain the following relationships:
Dso (600°F; 13 CFM) _ Dso (600°F, 1 CFM)
Dso (77°F; 13 CFM) Dso (77°F, 1 CFM)
and
Dso (375°F; 13 CFM) _ DSQ (375°F; 1 CFM)
Dso (77°F; 13 CFM) DSO (77°F; 1 CFM)
Thus, the Dso cut points of the small cyclone for a flow
of 13 ACFM are estimated to be 0.49 ym for a temperature of 375°F
and 0.63 ym for a temperature of 600°F.
150
-------
Task Category - E.S.P. Back-Up
Develop an E.S.P. Back-Up for Sampling Systems
(Technical Directive Number 10703)
Description of Task;
Southern Research Institute was to develop and test an
electrostatic precipitator to be used in lieu of a conventional
filter for back up to high volume sampling systems. This was to
be a high efficiency, low pressure drop system.
Summary of Progress;
The design of an electrostatic precipitator to be used in
lieu of a conventional back-up filter for high volume particulate
mass sampling systems has been completed and a prototype has been
constructed. A modified disc-cylinder geometry was selected as
the design after an evaluation of wire-cylinder and disc-cylinder
configurations indicated that the disc-cylinder would provide a
higher collection efficiency per unit length of cylinder. In
addition, the disc electrode would have greater endurance than a
wire electrode.
The fact that a laminar flow is obtained in the disc-cylinder
geometry allowed a straightforward calculation of the length of
cylinder required for 100% collection of 0.5 ym diameter particles.
In order to maximize the efficiency of collection with a minimal
overall precipitator length, it was necessary to increase the area
of the collection surfaces. This was done by using a concentric
cylinder configuration with alternate high voltage and grounded
cylinders (see Figure 85). Such a modification would not be
possible in conventional wire-cylinder precipitators.
At this time the prototype precipitator is being evaluated
using submicron aerosols.
151
-------
my. I nan mi i itm
Urai-C.
Zoi SMSLJ-
3/i-s.t
7o> a*se /<->
rr /» 702 3o KM /
i~rf»j
j7€7»-i e>A
/m/uf, / ruoe,
SOUTHERN RESEARCH INSTITUTE
BIRMINGHAM, ALABAMA 33203
ESP BACKUP FILTER. AS5V.
Figure 85. A sketch of the electrostatic precipitator design.
-------
Task Category - Guidelines, Manuals
Guideline for Particulate Sampling and
Annotated Bibliography
(Technical Directive Number 10804)
Description of Task:
The purpose of this task is to generate a guideline document
which will be a concise survey of the methods and instruments used
in sampling gaseous process streams for particulate matter.
Briefly describe the considerations involved in setting up and
carrying out a particulate sampling program. Also describe the
devices currently in use for making sizing and non-sizing measure-
ments. Include a condensed "where to learn more" bibliography
keyed to each topic discussed to provide more extensive coverage.
This document will essentially be an executive summary of
the document generated under Technical Directive Number 10904.
It will be a complete, concise discussion aimed at Program
Managers.
Summary of Progress:
Work has been initiated on this task with a completion
date set for early 1978.
Technical Manual on Particulate Sampling
(Technical Directive Number 10904)
Description of Task;
A comprehensive document is to be written which will be a
technical survey of the methods and instruments used in sampling
effluent gaseous process streams for particulate matter. This
document should include information on how to set up and carry
out a particulate sampling program on industrial and energy
process streams and on control devices such as electrostatic
precipitators, scrubbers, and fabric filters. Information on
the devices currently in use for making sizing and non-sizing
particulate measurements, along with descriptions of their
theory of operation and use should be included. To provide
further reading, a comprehensive bibliography keyed to each
topic is to be included.
Summary of Progress:
Requests for information were sent to 110 manufacturers of
sizing and non-sizing particulate measurement instruments. To
date responses have been received from approximately eighty
companies. This material is currently being reviewed and the
text of the manual is being drafted. The final draft is ex-
pected by the end of December, 1977.
153
-------
Procedures Manual for Electrostatic Precipitator Evaluation
(Technical Directive Number 20604)
Description of Task:
A procedures manual for the evaluation of electrostatic pre-
cipitators was written. This document gave specific methods for
collecting data on precipitator design and operating parameters,
flue gas composition, particulate mass sampling, and methods for
particle sizing and fractional efficiency determination.
Summary of Progress;
A procedures manual for evaluating electrostatic precipitators
has been published under the title "Procedures Manual for Electro-
static Precipitator Evaluation", EPA-600/7-77-059, June 1977. The
purpose of this procedures manual was to describe methods to be
used in characterizing the performance of electrostatic precipi-
tators for pollution control. A detailed description of the
mechanical and electrical characteristics of precipitators is
given. Procedures are described for measuring the particle size
distribution, the mass concentration of particulate matter, and
the concentrations of major gaseous components of the flue gas-
aerosol mixture. Procedures are also given for measuring the
electrical resistivity of the dust. A concise discussion and
outline is presented which describes the development of a test
plan for the evaluation of a precipitator. Several appendices
contain detailed information on testing methods as well as a
listing of the Federal Stationary Source Performance Standards
and Federal Source Testing Reference Methods.
154
-------
Task Category - Review of Documents
Review Documents and Reports Furnished by EPA
(Technical Directive Number 20705)
Description of Task;
The following three documents furnished by the Project
Officer were critically reviewed. A list of corrections and
suggestions was submitted to the Project Officer.
1. Cascade Impactor Operation and Calibration Guideline
2. Sampling Protocol to Minimize Errors Due to Source
Fluctuations
3. Level 1 Assessment - EPA/IERL
Summary of Progress;
All items in this task have been completed.
155
-------
Task Category - Consulting Services
Participate in U.S.A. - Soviet Information Exchange Program
(Technical Directive Number 20806)
Description of Task;
For this task, personnel at Southern Research Institute
were to participate as consultants on particle sizing in the
Soviet-U.S.A. Information Exchange Program. Equipment was to
be prepared and shipped to the U.S.S.R. for field tests. One
of the Institute staff members was to go to the Soviet Union to
supervise the test program, and reports as required were to be
written.
Summary of Progress;
Field testing equipment was shipped to the Soviet Union
during July, 1976. Field testing on a scrubber at a metallur-
gical plant in Russia took place during August, 1976. No
results have been published to date.
EPA/IERL/PMB Exhibit Booth at the 1977 APCA Meeting
(Technical Directive Number 20906)
Description of Task;
The purpose of this task was to coordinate the arrangements
for the EPA/IERL/PMB exhibit at the 1977 Air Pollution Control
Association in Toronto, Canada, June, 1977. This exhibit dis-
played the research and development efforts of the IERL Task
Level of Effort contractors. Certain pieces of hardware and
software from those contractors were displayed.
Summary of Progress;
On June 21, 22, 23, 1977 the Process Measurements Branch of
IERL/RTP supported an exhibit booth at the 70th Annual Air
Pollution Control Association Meeting in Toronto, Ontario,
Canada. This 10' x 20' booth used a color scheme of dark blue
booth back wall and side walls, light blue carpet, and green
draped tables. Three tables along the back wall were used for
document display. Two tables, one on each side, were used to
display hardware. On the white I1 x 20' header board in black
letters was printed the following title:
156
-------
United States Environmental Protection Agency
Industrial Environmental Research Laboratory - RTF
Process Measurements Branch
On either side of the title was an EPA LOGO in color.
On the back wall were hung six 3' diameter discs which
briefly described the research and development efforts of the
six Task Level of Effort contractors for the PMB. These six
contractors are Acurex/Aerotherm, Arthur D. Little, Inc.,
Research Triangle Institute, Southern Research Institute, The
Research Corporation, and TRW, Inc.
The hardware on display included a complete Source Assess-
ment Sampling System, a KLD Droplet Analyser, and the Five Stage
Series Cyclone and Advanced Sampling System.
Approximately 200 copies of twenty-one documents were dis-
tributed on a first come first serve basis to the 4200
registrants at the meeting.
The following is a list of the documents which were available
for distribution:
HP-25 Programmable Pocket Calculator Applied to Air Pollution
Measurement Studies: Stationary Sources. EPA-600/7-77-058,
June 1977
Procedures Manual for Electrostatic Precipitator Evaluation
EPA-600/7-77-059, June 1977
Industrial Environmental Research Laboratory - RTP Annual
Report 1976
Flow and Gas Sampling Manual. EPA-600/2-76-203, July 1976
IERL-RTP Procedures Manual: Level 1 Environmental Assessment.
EPA-600/2-76-160a, June 1976
Selection and Evaluation of Sorbent Resins for the Collection
of Organic Compounds. EPA-600/7-77-044, April 1977
Technical Manual for Measurement of Fugitive Emissions:
Upwind/Downwind Sampling Method for Industrial Emissions.
EPA-600/2-76-089a, April 1976
157
-------
Technical Manual for the Measurement of Fugitive Emissions:
Roof Monitor Sampling Method for Industrial Fugitive
Emissions. EPA-600/2-76-089b, May 1976
Technical Manual for Measurement of Fugitive Emissions:
Quasi-Stack Sampling Method for Industrial Fugitive
Emissions. EPA-600/2-76-089c, May 1976
Technical Manual for Process Sampling Strategies for
Organic Materials. EPA-600/2-76-122, April 1976
Technical Manual for Analysis of Organic Materials in
Process Streams. EPA-600/2-76-072, March 1976
Particulate Sizing Techniques for Control Device
Evaluation: Cascade Impactor Calibrations. EPA-600/2-76-
280, October 1976
Inertial Cascade Impactor Substrate Media for Flue Gas
Sampling. EPA-600/7-77-060, June 1977
Operating and Service Manual: Source Assessment Sampling
System. Acurex/Aerotherm Report UM-77-81, March 1977
Environmental Assessment Sampling and Analysis: Phased
Approach and Techniques for Level 1. EPA-600/2-77-115,
June 1977
Procedures for Cascade Impactor Calibration and Operating
in Process Streams. EPA-600/2-77-004, January 1977
Technical Manual for Inorganic Sampling and Analysis.
EPA/2-77-024, January 1977
Development and Trial Field Application of a Quality
Assurance Program for Demonstration Projects. EPA-600/2-
76-083, March 1976
*
HP-65 Programmable Pocket Calculator Applied to Air
Pollution Measurement Studies: Stationary Sources. EPA-
600/8-76-002, October 1976.
Process Measurements Branch: Report Listing, June 1, 1977
Pollution Control Technology and Environmental Assessment:
Prochure Describing the Research and Development Efforts of
the Six PMB Contractors.
158
-------
At the conclusion of the exhibit all of the reports had
been given out, except for a few HP-65 booklets. Reaction to
the exhibit by the attendees at the meeting was quite favorable.
Task Category - Advanced Concepts - Mass and Size
Evaluation of a PILLS IV Particle Sizing Instrument
(Technical Directive Number 11007)
Description of Task;
The contractors shall perform a detailed evaluation of the
theory and operation of the PILLS IV particle sizing instrument,
This evaluation shall consider the following:
1. Verify the validity of the theoretical principle.
2. Evaluate or interpret existing experimental data.
3. Investigate problems of obtaining representative
samples. Generate a test aerosol and look at
response vs. dilution, sample velocity, particle
sizing, and purge effects.
4. Investigate the possibility of multiple scattering
as an interference.
5. Attempt to measure the beam dimensions and view
volume.
6. Investigate the differences in response to scattering
of broad or narrow plane waves.
7. Investigate the cumulative effects of 4 and 6.
Summary of Progress;
Laboratory evaluation of the PILLS-IV is complete and
a final report is being written.
159
-------
References
1. Cooper, Douglas W. and John W. Davis, "Cascade impactors
for aerosols: Improved data analysis," Amer. Ind. Hyg.
Assoc., p. 79, 1972.
2. Cooper, Douglas W. and Lloyd A. Spielman, "Data inversion
using nonlinear programming with physical constraints:
Aerosol size distribution measurement by impactors,"
Atmospheric Environment, Vol. 10, pp. 723-729, 1976.
3. Picknett, R. G., "A new method of determining aerosol size
distributions from ministage sampler data," Aerosol Science,
1972.
4. McCain, J. D., K. M. Gushing, and W. B. Smith, 'Methods
for determining particulate mass and size properties:
Laboratory and field measurements," J. APCA 24(12);
1172-1176, December 1974.
Rao, A. K., "An experimental study of inertial impactors,"
Ph.D. Dissertation, University of Minnesota, Minneapolis,
5.
Minnesota, 1975.
6. Dzubay, T. G., L. E. Hines and R. K. Stevens, "Particulate
bounce errors in cascade impactors," Atmospheric Environ-
ment, Vol. 10, pp. 229-234, 1976.
7. Natusch, D. F. S. and J. R. Wallace, "Determination of air-
borne particle size distributions: Calculation of cross-
sensitivity and discreteness effects in cascade impaction,"
Atmospheric Environment, Vol. 10, pp. 314-324, 1976.
8. Gushing, K. M., J. D. McCain, and W. B. Smith, "Experimental
determination of sizing parameters and wall losses of five
commercially available cascade impactors," in Proceedings
of the 69th Annual Meeting of the Air Pollution Control
Association, Portland, Oregon, 1976, paper No. 76-374.
9. Lundgren, D. A. (1967). "An aerosol sampler for determina-
tion of particle concentration as a function of size and
time," J. APCA 3.7, pp. 225-229.
10. Smith, W. B., and J. R. McDonald. Development of a Theory
for the Charging of Particles by Unipolar Ions. J. Aerosol
Sci., 7(2): 151-166.
160
-------
References (Cont'd.)
11. Gushing, K. M., G. E. Lacey, J. D. McCain, and W. B. Smith.
Particle Sizing Techniques for Control Device Evaluation:
Cascade Impactor Calibration. EPA-600/2-76-280, U.S.
Environmental Protection Agency, Research Triangle Park,
North Carolina, 1976. 79 pages.
12. Reischl, G. P., W. John, and W. Devor. Uniform Electrical
Charging of Monodisperse Aerosols. J. Aerosol Sci., sub-
mitted for publication May 17, 1976.
13. Particle Technology Laboratory. In the Performance of the
Electrical Aerosol Analyzer. No. 237, Mech. Eng. Dept.
University of Minnesota, Minneapolis, MN 55455. 15 pages.
14. Zung, J. T., and C. C. Snead. Evaporation Kinetics of
Liquid Droplets. U.S. Dept. of Army, Edgewood Arsenal-
DAAA-15-67-C-0151, Field Evaluation Division, Edgewood
Arsenal, Maryland 21010, 1968. 116 pages.
15. Smith, W. B., K. M. Gushing, and G. E. Lacey. Andersen
Filter Substrate Weight Loss Study. Special Report under
EPA Contract No. 68-02-0273, Research Triangle Park, NC
27711.
16. Rao, A. K. Sampling and Analysis of Atmospheric Aerosols.
Particle Tech. Lab. Publ. No. 269, Department of Mechanics
Engineering; University of Minnesota, Minneapolis, Minnesota
55455, June 1975.
17. Ranz, W. .D., and J. B. Wong. Impaction of Dust and Smoke
Particles, Ind. and Eng. Chem., 50, No. 4 (April, 1958).
18. Marple, V. A. A Fundamental Study of Inertial Impactors.
Ph.D. Thesis, Mechanical Engineering Department, Univer-
sity of Minnesota, Minneapolis, Minnesota 55455, 1970.
19. Berglund, R. N., and B. Y. H. Liu. Generation of Monodis-
perse Aerosol Standards. Environmental Science and Tech-
nology, Vol. 6, No. 2, 1973.
20. Lindblad, N. R., and J. M. Schneider. Production of Uni-
form-Sized Liquid Droplets. J. Sci. Instru., Vol. 42,
1965.
161
-------
References (Cont'd.)
21. Strom, L. The Generation of Monodisperse Aerosols by
Means of a Disintegrated Jet of Liquid. Rev. Sci. Instr.,
Vol. 40, No. 6, 1969.
22. Calvert, S. Cascade Impactor Calibration Guidelines.
EPA-600/2-76-118, U.S. Environmental Protection Agency,
Research Triangle Park, NC, 1976.
23. Jaericke, R., and I. H. Blifford. The Influence of Aerosol
Characteristics on the Calibration of Impactors. Journal
of Aerosol Science, 5(5): 457-464, 1974.
24. Willeke, K. Performance of the Slotted Impactor. Am. Ind.
Hygiene Assoc. J., 683-691, September 1975.
25. Lapple, C. E. Process Use Many Collector Types, Chem. Eng.,
58, pp. 144-151 (1951).
26. Chan, Tai and Morton Lippmann. Particle Collection Effi-
ciencies of Air Sampling Cyclones: An Empirical Theory,
Envir. Sci. and Tech., Vol. 11, No. 4, pp. 377-382 (377),
1977.
162
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing)
1. REPORT NO.
EPA-600/7-78-009
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Particulate Sampling Support: 1977 Annual Report
5. REPORT DATE
January 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
K.M. Gushing, William Farthing, L.G. Felix,
J.D. McCain. andW.B. Smith
8. PERFORMING ORGANIZATION REPORT NO.
SORI-EAS-77-661
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-2131
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Annual; 11/76-10/77
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES jERL-RTP project officer is D.
541-2557.
Bruce Harris, Mail Drop 62, 919/
16. ABSTRACT
The report describes the activities supporting the particulate sampling
efforts of EPA/IERL-RTP during FY 1977. Twenty technical directives were issued
in seven categories: cascade impactors (7), cyclones (5), sampling electrostatic
precipitators (1), guidelines and manuals (3), document review (1), consulting
services (2), and advanced concepts (1). Significant results of the 12 completed tasks
are presented.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Dust
Sampling
Impactors
Cyclone Separators
Electrostatic Precipitators
Air Pollution Control
Stationary Sources
Particulate
Cascade Impactors
13 B
11G
14B
07A
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS /This Report/
Unclassified
21. NO. OF PAGES
174
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
163
------- |