EPA-600/2-77-004
January 1977
Environmental Protection Technology Series
               PROCEDURES FOR  CASCADE  IMPACTOR
                      CALIBRATION AND OPERATION IN
                                      PROCESS  STREAMS
                                  Industrial Environmental Research Laboratory
                                        Office of Research and Development
                                       U.S. Environmental Protection Agency
                                 Research Triangle Park, North Carolina  27711

-------
               RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environ
Protection Agency, have been grouped into five series. These five
categories were established to facilitate further development and applies
environmental technology. Elimination of traditional grouping was consi
planned to foster technology transfer and a maximum interface in related
The five series are:

     1.   Environmental Health Effects Research
     2.   Environmental Protection Technology
     3.   Ecological Research
     4.   Environmental Monitoring
     5.   Socioeconomic  Environmental Studies

This report  has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repairer prevent
environmental degradation  from point and non-point sources of pollution. This
work provides the new or  improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.


                    EPA REVIEW NOTICE


This report has been reviewed by the U.S. Environmental
Protection Agency, and  approved for publication.  Approval
does not signify that  the contents necessarily reflect the
views and policy of the Agency, 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-600/2-77-004

                                  January 1977
             PROCEDURES

      FOR CASCADE  IMPACTOR

   CALIBRATION  AND OPERATION

        IN  PROCESS STREAMS
                    by

              D. Bruce Harris

 Industrial Environmental Research Laboratory
   Office of Energy, Minerals, and Industry
      Research Triangle Park, NC  27711
            ROAP No. 21ADM-012
        Program Element No. 1AB012-20
               Prepared for

U.S. ENVIRONMENTAL PROTECTION AGENCY
      Office of Research and Development
           Washington, DC 20460

-------
                                   CONTENTS
Figures   	     v
Tables    	    vi
Acknowledgments 	   vii
     1.0  Introduction  	     1
     2.0  The Presurvey	     3
     3.0  Equipment Selection 	     6
          3.1  Impactor Selection 	     6
          3.2  Sample Trains  	     7
          3.3  Balance Requirements 	    10
     4.0  Substrates	    11
          4.1  Collection Substrates  	    11
          4.2  Back-Up Filters	    13
     5.0  Preparation and Sampling	    14
          5.1  Substrate Preparation  	    14
          5.2  Impactor Orientation 	    14
          5.3  Heating the Impactor	    14
          5.4  Probes   	    15
          5.5  Nozzle and Sampling Rate Selection	    15
          5.6  Pre-Cutter Use	    16
          5.7  Sampling Time  	  .......    16
          5.8  Readying the Impactor	^. .  . .    19
          5.9  Pre-Sample Checks  	    20
          5.10 Taking the Sample  	    20
          5.11 Number of Sample Points  	    20
     6.0  Sample Retrieval and Weighing 	    22
          6.1  Impactor Clean-Up  	    22
          6.2  Drying and Weighing	    22
          6.3  Data Logging	-	    23
     7.0  Quality Assurance 	    25
          7.1  Need for a Quality Assurance Program   	    25
          7.2  Impactor Techniques  	    25
          7.3  Weighing Techniques  	    27
          7.4  General Notes	    28
          7.5  Data Analysis	    28
                                  111

-------
CONTENTS  (cont'd)
     8.0  Data Analysis 	
          8.1  Cascade Impactor Data Analysis   	
          8.2  D5Q Method   	
          8.3  Calculation of Theoretical Stage Den's   .  . .
          8.4  Calculation of Stage DSQ'S from Calibration
               Data   	
          8.5  Differential Particle Size Distributions—Dso
               Method 	
          8.6  Cumulative Particle Size Distributions .  .  . .
     9.0  Reports   	
          9.1  Size Distribution  	
          9.2  Data to be Reported  	
    10.0  Commercial Impactors  	
         10.1  Brink Impactor	 .  .  . .
         10.2  Andersen Impactor  	
         10.3  University of Washington Mark III (Pilat)
               Impactor 	
         10.4  Meteorology Research, Inc. (MRI) Impactor  . .
         10.5  Sierra Model 226 Source Cascade Impactor .  . .
    11.0  Bibliography  	
    12.0  Glossary      	
Appendix A - Impactor Flow Rate Measurement   	
Appendix B - Cascade Impactor Calibration Guidelines,
             EPA-600/2-76-118   	
Appendix C - Design Drawings,  Brink In-Line Cyclone and
             Impactor   	
Appendix D - Cost Estimating Data Sheet   	
Appendix E - Preliminary Survey for Particulate Sizing  .  . .
Appendix F - Safety Checklist    .	
Appendix G - Sample Calculation 	
Appendix H - Metric System Conversion Factors 	
 30
 30
 30
 32

 35

 36
 38
 40
 40
 40
 41
 41
 42

 43
 44
 45
 46
 48
 51

 55

 93
 96
 99
104
106
113
                                    iv

-------
                                   FIGURES

Number                                                           Page
  1       Presurvey sampling using an impactor 	    4
  2       Typical  sample train with heated impactor  	    8
  3       Nomograph for selecting nozzles for isokinetic
          sampling   	17
  4       Nomograph for sampling time selection (50 mg sample)  .   18
  5       Stage collection efficiency for a modified Brink
          impactor (T=72°F, P=2S.SO in. Hg, P=1.35 g/cm3,
          flow = 0.03 acfm)  .	31
B-l       Calibration apparatus  	   59
B-2       Collison atomizer  .	60
B-3       Sampling system arrangement	   64
B-4       Aerodynamic cut diameter vs. impactor flow for
          A.P.T. M-l cascade impactor     	 ....   71
8-5       Aerodynamic diameter vs. diameter for various
          .tensities  . . .	72
B-6       Ffficiency vs. inertial impaction parameter for
          A.P.T. M-l cascade impactor, stage 7   	   79
3-7       Efficiency vs. inertial impaction parameter, data
          from 4 separate U.W, Mark III cascade impactors,
          stage: 5	 ,	80
3-8       Efficiency vs. inert1*c'i Impaction parameter for
          IJ.W. Mark III cascade impactor,  stages 4-7    	81
E-9       Efficiency vs. inertial impaction parameter for
          A.P.T. M-l cascade impactor, stages 4-7   	   82
B-10      Efficiency vs. inertial impaction parameter for
          Andersen non-viable cascade  impactor, stages 4-7   . .   83
B-ll      Efficiency vs. inertial Impaction parameter for
          comparison   ............ 	   85

-------
                                    TABLES
Number                                                             Page
  1       Impactor Decision Making	     5
  2       Sampling Information Required  	    24
B-l       Operating Characteristics of a 3-Jet Collison
          Atomizer	    62
B-2       Available PSL Particle Diameters   	    70
B-3       Length/Jet Diameter (s/de) for Cascade Impactors ...    86
G-l       Data from Brink Impactor Run   	   112
                                    vi

-------
                             ACKNOWLEDGMENTS

     The author gratefully acknowledges the help of the members of the
working group who provided much useful Information from their knowledge
of and experience in the field of particle sizing:  Dr. Seymour Calvert
(APT, Inc.), Dr. Neal Hill (Andersen 2000, Inc.), Gene Kennedy (Monsanto
EnviroChem), Joe McCain (Southern Research Institute), and Dr. Michael
Pilat (University of Washington).
     Special thanks to Ken Cushing and Joe McCain of Southern Research
Institute for their work on the computation section, Seymour Calvert for
the calibration procedure, and Douglas VanOsdell of Research Triangle
Institute for assistance in the compilation of this document.
                                    vii

-------
                               SECTION 1.0
                              INTRODUCTION

     Inertial impactors are commonly used to determine the  particle  size
distributions of particulate matter emitted by industrial  sources.
Impactors have several advantages over competing techniques:   they are
compact, can be inserted directly into the duct (avoiding  the problem of
sample loss in a probe), are fairly accurate, and produce  information
which is widely understood.  The Process Measurements Branch  (PMB) of
EPA's Industrial Environmental Research Laboratory, Research  Triangle
Park, N. C., has been using inertial impactors to determine particle
size distributions for several years, as have a number of  IERL-R7P
contractors.  The PMB has also sponsored an evaluation of  impactors  to
select devices that could be used under normal field conditions.  During
the course of these programs it became evident that no uniform approach
to the field use of impactors was available.
     To develop uniform procedures, a working group of IERL-HTP personnel,
contractors, and independent experts met at Research Triangle Park,  N.  C.
This procedures guide is an outgrowth of the working group's  discussions.
The document has several purposes.  Above all, the PMB wants  to ensure
the comparability of data gathered by different contracted.   That is,
that the contractors use equipment whose characteristics are known,
follow sound sampling procedures, and reduce the data using accepted and
defined techniques.  This document 1s also intended to help impactor
users avoid some of the problems which others have experienced.
     The procedures presented should yield quality data at. most sampling
sites.  Situations will occur where the Information gathered *n *h1:;
document will not be applicable and a suitable procedure will have to be
worked out.  Professional judgment Is still the most Important elamsnt
in successfully determining particle size distributions and fractional
efficiency.

-------
     The scope of this report Includes the preliminary survey, sampling
apparatus, testing procedures, data analysis, reporting requirements,
and Impactor calibration.  The Information 1s applicable to cascade
impactors in general.  Specific commercial impactors are discussed in
the section titled "Commercial Impactors."
     Because the state-of-the-art in Impactor sampling 1s advancing
rapidly, readers of this document are encouraged to seek additional
information.  For IERL-RTP contractors, the Process Measurements Branch
should be contacted for updates on the guidance 1n this document and
perhaps the resolution of some of the questions which have Incomplete
answers at this time.

-------
                                  SECTION 2.0
                                 THE PRESURVEY

     The key to performing a successful  fractional  efficiency evaluation
is thorough planning based on a complete pretest site survey.  The
survey should provide adequate information at as low a cost as possible.
The presurvey form presented as Appendix E is a reasonable guide to the  .
type of information which should be noted.  Some sites will require more
information.  As far as is possible, the information noted during the
presurvey should be measured rather than obtained from plant records or
personnel.
     As the presurvey is generally conducted by one or two men "traveling
lightly," the apparatus used during the presurvey should be as light and
compact as possible.  A presurvey sample train 1s shown in Figure 1.
This system was built into a single, suitcase-sized package, and served
well as a presurvey sample train.  The impactor which is to be used
during the main test program should normally be used during the presurvey
because the suitability of substrates and adhesives must be checked out.
These problems are discussed more fully in later sections.
     In general, the presurvey work should be done using the techniques
described in this document.  Less precision is required, but the accuracy
must be high enough to provide useful information 1n designing the test
program.  The decisions which must be made are summarized  in Table 1.

-------
            DIFFERENTIAL PRESSURE
                  INDICATORS
0-2in.H20                  0-10 in. Hg
                                                                 NEEDLE
                                                                 VALVE
                                            ATMOSPHERIC
                                              PRESSURE
TEMPERATURE   OUTPUT TO
   SENSOR      HEATING
   SIGNAL      TAPE
                                  STACK STATIC
                                   PRESSURE


                     IMPACT PRESSURE
                        COLLAPSIBLE
                        PITOT TUBE
                                                                           HEATING TAPE
                                                                           OR WRAP
                          Figure  1.   Presurvey  sampling using an  impactor.

-------
                                     TABLE  1.   IMPACTOR DECISION MAKING
    Item
      Basis of Decision
               Criteria
Impactor
Loading and size estimate
Sampling rate





Nozzle


Pre-cutter


Sampling time
Collection
substrates
Number of sample
points

Orientation
of Impactor

Heating
Probe
Loading and gas velocity





Gas velocity


Size and loading


Loading and flow rate
Temperature and gas
composition
Velocity distribution
and duct configuration

Duct size, port
configuration, and size

Temperature and presence
of condenslble vapor
Port not accessible using
normal techniques
a.   If concentration of particles smaller
     than 5.0 vm 1s less than 0.46 gin/am3
     (0.2 gra1n/acf), use high flow rate
     Impactor (wO.5 acfm).
b.   If concentration of particles smaller
     than 5.0 gm 1s greater than 0.46 gm/am3
     (0.2 gra1n/acf), use low flow rate
     Impactor (*0.05 acfm).

a.   Fixed, near Isokinetic.
b.   Limit so last jet velocity does not
     exceed:
     -60 m/sec greased
     -35 m/sec without grease.

a.   Near Isokinetic, ±10%.
b.   Sharp edged; minimum 1.4 mm ID.

If pre-cutter loading Is comparable to
first stage loading, use pre-cutter.

a.   Refer to Section 5.5.
b.   No stage loading greater than 10 mg.

a.   Use metallic foil or fiber substrates
     whenever possible.
b.   Use adhesive coatings whenever possible.

a.   At least two points per  station.
b.   At least two samples per point.

Vertical Impactor axis wherever possible.
a.    If  flue  1s above  177°C, sample at
      process  temperature.
b.    If  flue  1s below  177°C, sample at
      11°C  above process temperature at
      Impactor exit external heaters.

a.    Only  1f  absolutely necessary.
b.    Pre-cutter on endin  duct.
c.    Minimum  length and bends  possible.

-------
                                 SECTION 3.0

                             EQUIPMENT SELECTION
3.1  IMPACTOR SELECTION
     The selection of the proper impactor for a particular test situation
is primarily dependent upon the mass loading of the gas stream and its
effect on sampling time.   There are three major criteria to be met to
match an impactor to a particulate stream:
     1)   The sampling period must be long enough to provide a
          reasonable averaging of transient conditions in
          the stack.
     2)   The loading on a given impactor stage must be low
          enough to prevent re-entrainment.
     3)   The sampling rate through the impactor must be low
          enough to prevent scouring of impacted particles by
          high gas velocities.
     For these reasons, an impactor with a comparatively low sample rate
must be used in a gas stream with a high mass loading.  The low sample
rate allows a longer sampling time, although in some situations it will
still be undesirably short.  Conversely, in a low mass loading situation
such as a control device outlet, a high sample rate device must be used
if a significant amount of sample is to be gathered in a reasonable
amount of time.
     A cascade impactor can normally yield useful information over a
range of sample rates differing by a factor of 2 or 3.  As high efficiency
control devices cause the outlet mass loading to differ from the inlet
by a factor of 10, the same impactor can seldom be used on both inlet
and outlet.  Both high and low flow rate impactors are usually required
to determine the efficiency of particulate control devices.

-------
3.2  SAMPLE TRAINS
     Figure 2 is a flow diagram of a typical  impactor sample  train.   As
shown, it is desirable to have the impactor inside the stack  with  a
straight nozzle.  A sampling probe leading to an impactor outside  of the
duct should be used only if absolutely necessary.   The probe  should  be
as short as possible and contain the fewest possible bends.   It is
recommended that a pre-cutter cyclone be mounted at the probe Inlet  to
remove particles larger than approximately 10 micrometers and thus
reduce losses in the probe.
     Heating System ~ The criteria for heating are given in  Table 1.
If heating is required, the entire Impactor must be either wrapped 1n
heating tape or put in a custom-fitted heating mantle.  The temperature
control should be based on the temperature at the outlet end  of the
impactor.  Often the temperature 1s measured between the last stage  and
back-up filter.  The Impactor temperature can be controlled either
manually or automatically.  An automatic controller has been  found to be
worth its cost by releasing the operator for other tasks.
     Flue Gas Conditioning — Often 1t 1s necessary to cool and dry the
flue gas before it reaches the flow measuring section, as condensation
in the orifice would distort the measurement.  Also, drying is useful to
protect the equipment from the condensate which, in S02-conta1n1ng
gases, is likely to be sulfuric acid.  The type of condenser shown 1s
usually satisfactory.  Packed-bed drying columns are commercially available.
The heat exchange coil is used to bring the gas temperature to essentially
ambient so that there will not be a significant-temperature gradient
across the flow measuring devices.
     Flow Measurements — At least two flow measuring devices are used
in series.  Normally, a calibrated orifice is used in conjunction with a
dry gas meter, as shown.   The commonly used diaphragm-type positive
displacement gas meter becomes Increasingly inaccurate at flow rates
less than 5 percent of rated capacity.  For a typical stack sampling

-------
                                  IMPACTOF
          GAS FLOW
      HEATING
      JACKET
              DRYING
             COLUMN
                            CONDENSERS
         HEAT
         EXCHANGER
         TEMPERATURE
         CONTROLLER
                                         ICE BATH
                                               BLEED
                       CALIBRATED
                        ORIFICE
                                       DRY GAS
                                        METER
Q
 VACUUM
  PUMP
              •VENT
                MANOMETERS

          (OR DIFF. PRESSURE GAUGES)
Legend

      -  Pressure Measurement Point

   )  -  Temperature Measurement Point
               Figure 2.  Typical  sample  train with heated impactor.
                                  8

-------
gas meter this would be approximately 1.4 liters/min  (0.05  cfm).  Another
calibrated orifice or a rotameter should then be  used as  the  second  flow
meter.
     Temperature Measurements -- It is necessary  to know  the  temperature
at all .points where flow rate must be known.   Any convenient  device  of
known accuracy can be used to make the measurements.   The in-stack
measurement can easily be made at the probe end with  a thermocouple.
The temperature at the downstream end of the Impactor is  measured
directly behind the final filter and is used to control the heating  tape
if one is used.  If the heat exchanger in the train brings the gas
temperature to about ambient, only one temperature reading will be
necessary at the flow meters.  This is usually most conveniently done at
the dry gas meter, as taps are available on the meter.
     Vacuum Pumps — The vacuum pump should usually be placed at the end
of the sample train.  This is because vacuum pumps tend to leak and all
of the flow measurements must be made upstream of any leak.  The flow
rate can be controlled by using an Inlet side air bleed or with a
recirculating bypass from the pump discharge.  If a leak-free pump  is
available a more convenient train can be obtained by switching the  pump
and orifice.  A standard Method 5 pump box has worked well.
     Pressure Measurements — Most of the pressure measurements are
shown being made with manometers, but calibrated differential pressure
meters are equally acceptable.  The in-stack pressure needed is the
static pressure, which is not exactly the downstream pressure of an S-
type pitot tube.  A true static pressure measurement should be made.   It
is not necessary that this be part of the impactor train,  but  it can be.
     The pressure at the downstream end of the impactor, between the
last stage and the final filter, must be known.  It can be measured, but
this is often inconvenient.  If a flow rate pressure drop  calibration  is
available for the impactor (without final filter), 1t  is normally accept-
able to calculate the pressure drop.  Correction must be made for pressure
and temperature differences between the calibration conditions and  the
actual conditions.  The impactor is treated as an orifice, and the
techniques discussed in Appendix A are used to make the corrections.

-------
     The pressure at the inlet to the metering devices must be known.  In
the system shown in Figure 2, the pressure is metered ahead of the
calibrated orifice and the orifice pressure drop is used to calculate
the pressure going into the dry gas meter.  The dry gas meter pressure
should be measured if there is a reason to think the procedure above was
not adequate.
3.3  BALANCE REQUIREMENTS
     For accurate weighing of collected material a balance with a
sensitivity of at least 0.05 mg is required.  This is especially true
for the lower stages of the low sample rate impactors where collection
of 0.3 mg or less is not uncommon.  The balance must be relatively
insensitive to vibration if it is to be used 1n the field.  It is also
desirable to have a balance with a weighing chamber large enough to hold
the impactor substrates without folding.  These capabilities are available
in several electrobalances marketed in the U.S.
                                  10

-------
                                SECTION 4.0
                                SUBSTRATES
4.1  COLLECTION SUBSTRATES
     For reasons which have been discussed, very accurate determinations
of Impactor stage catch weights are necessary.  Impactor stages are
generally too heavy for the tare capacity of field-usable precision
balances.  For this reason, the particles are captured on substrates
which are lightweight and can be weighed on the balances.  Generally,
these substrates are made of metal foil or glass fiber.
     Glass Fiber Substrates ~ Glass fiber substrates are used on some
commercial impactors.  In addition to providing a lightweight impactlon
surface, glass fiber mats greatly reduce re-entrainment due to particle
bounce.  They are superior to greased metal substrates in very high
temperature applications where the greases tend to evaporate.  These
substrates should be handled carefully to prevent fiber loss after
weighing, and care must also be taken when using glass fiber substrates
in streams containing sulfur dioxide.  Recent experimentation has shown
that glass fiber materials often exhibit anomalous weight gains due to
sulfate uptake on the substrates. Apparently, sulfur dioxide in a gas
stream can react with basic sites on most glass fiber materials and form
sulfates.
     There are two approaches to this problem.  Substrates which will
gain weight from sulfate uptake can be preconditioned 1n the flue gas
before weighing.  Two to 6 hours of exposure to the flue gas will
suffice where mass loadings are high and sample times are short.  In the
situations where sample times are long and the collected amount of
paniculate matter small, it may be necessary to condition the substrates
for as long as 24 hours to eliminate significant sulfate uptake and
                                     11

-------
 weight gains.   Repeated weighings to check weight gains are necessary to
 confirm that the  substrates  can be used.  Another approach is to use a
 fibrous substrate which shows  little weight gain in a sulfur dioxide
 stream, if  one  can  be found.   It should be noted that the particle
 retention characteristics of different fiber materials vary, and the
 impactor calibration could change significantly if the substrate 1s
 changed.
     Greased Substrates — Grease must often be used on metal foil
 substrates  to improve their  particle retention characteristics.  This is
 particularly important with  hard, bouncy partlculate.  Impactor stage
 velocities  of 60-65 m/sec have been used on greased substrates with good
 results, while  particle bounce can become a problem at about half of
 that rate on ungreased substrates.
     Finding a  suitable grease can be difficult.  The grease should not
 flow at operating temperature, and must be essentially non-volatile.
 Gas chromatographic materials  such as polyethylene glycol 600 have
 exhibited more  consistent characteristics than materials such as stopcock
 grease.  Another  class of materials which may be suitable is high vacuum
 greases; Apiezon  L and H in  particular have performed well at temperatures
 up to  120°C (250°F).  The greased substrates must be tested as blanks in
 filtered process  gas before they are used 1n the test program.
     The greases  are normally applied as suspensions or solutions of 10-
 20 percent grease in toluene or benzene.  The mixture 1s placed on the
 substrate with  a  brush or eyedropper, baked at 400°F for 1 to 2 hours,
 and then desiccated for 12 to 24 hours orior to weighing.  It is important
 to avoid an excess of grease.  The desiccated, greased substrate should
 be tacky, but not slippery, with a film thickness about, aqua! to the
diameter of the particles which are to be captured,
     Horizontal operation of the impactors with greased substrates is
not recommended due to possible grease flow.  Care must also be taken to
ensure that grease is not blown off the substrates (which tands to occur
at jet velocities greater than 60 m/sec).   To some degree, grease blow-
off can be avoided by using a light coating of grease on the last stages.
                                      12

-------
This is normally satisfactory from an adhesive standpoint, as the last
stages usually have the lightest loading along with the highest jet
velocity.  Inspection of the stage catches is the best way to check on
this problem.
4.2  BACK-UP FILTERS
     Back-up filters are used on all impactors to collect the material
that passes the last impact!on stage.  Binder!ess glass fiber filter
material is normally used for this purpose in all the Impactors, although
the exact configuration varies.
     Glass fiber back-up filters have the same problems as do glass
fiber substrates.  Their use in process gases containing sulfur oxides
is suspect, and blanks must be run to check out the problems.  Pure
Teflon filters may alleviate this problem if they can be used.
                                    13

-------
                                 SECTION 5.0
                          PREPARATION AND SAMPLING

5.1  SUBSTRATE PREPARATION
     It Is assumed that the substrates have been properly prepared and
that the necessary data quality assurance steps have been taken.  The
substrates should be desiccated, carefully weighed, and kept in a
desiccator until they are to be placed In the Impactor.
5.2  IMPACTOR ORIENTATION
     Whenever possible, the impactor should be oriented vertically to
minimize gravitational effects such as flow of grease or movement of
collected particles.  Sampling situations requiring horizontal placement
will occur, and extra care must be taken on such occasions not to bump
the impactor against the port during entry or removal.
5.3  HEATING THE IMPACTOR
     Unless a condensate is the prime aerosol being measured, all
condensible vapors must be in a gaseous state until they exit from the
impactor.  In gas streams above 350°F, auxiliary heating is not usually
required.  Below 350°F the exit temperature of the impactor should be
maintained at least 2Q°F above the process temperature 1f condensible
vapors are present.  A thermocouple-feedback temperature controller has
proven useful.
     When condensible vapors are present, it is sometimes necessary to
heat the impactor probe to prevent any condensate formed in the probe
from entering the impactor and contaminating the substrates.  Water
vapor is the primary problem.  The probe temperature should be maintained
above the vapor's dewpoint.
                                   14

-------
     Whether the impactor 1s being heated in the duct or externally with
heater tape, an allowance of 45 minutes warm-up time 1s  recommended as a
minimum to ensure that the Impactor has been heated to duct or  operating
temperature.  Thermocouple monitoring of the Impactor temperature  and
gas temperature is recommended.
5.4  PROBES
     Sampling probes leading to an Impactor outside of the duct should
be used only if there is no other way.  They should be as short as
possible and contain the fewest possible bends.  It 1s recommended that
a pre-cutter be mounted at the duct end of the probe to  remove  the large
(>10 ym) particles and thus reduce line losses.
5.5  NOZZLE AND SAMPLING RATE SELECTION
     It is preferable to use as large a nozzle diameter  as possible to
minimize sampling errors resulting from nozzle Inlet geometry.   When
very small nozzles have been used with the Brink Impactor, there have
been some cases in which large amounts of material were retained in the
nozzle or the nozzle was completely blocked.   It is recommended that the
inlet nozzle not be smaller than 1.5 mm, and some types  of particulate
material may require a larger minimum nozzle size.  In some Instances
bent nozzles are necessary due to port location and gas  direction, but
these should be avoided.  Particulate tends to collect in the nozzle,
and it is difficult to determine the size interval In which the deposited
material originated.  Bent nozzles are also difficult to clean.  If they
cannot be avoided, bends should be as smooth as possible and of minimum
angle in order to minimize the losses in the fine particle region.
     For hard, "bouncy" particulate,  the sampling rate must be  such
that the last stage velocity does not exceed 60 m/sec for greased collection
surfaces or 35 m/sec for ungreased plates if no suitable substrate can
be found to limit particle bounce.  The flow rates above should not be
considered the final word on nozzle velocity;  particle bounce has been
observed at nozzle velocities as low as 10 m/sec, while some particulate
                                   15

-------
 materials  are  "sticky"  and will adhere at well above the maximum velocity
 for  hard particles.  The exposed substrate should be visually examined
 for  evidence of  re-entrainment and the rates adjusted accordingly.
 Refer  to the "Quality Assurance" section of this document.
     It is apparent  that sample rate and nozzle size are closely coupled.
 The  requirements for isokinetic or near-isokinetic nozzle flow sometimes
 impose a compromise  on  nozzle selection.  The general order of priorities
 when choosing  the sample rate is nozzle diameter (at least 1.4 mm), last
 stage  jet  velocity,  and flow rate required for isokinetic sampling.
 Selection  of nozzle  diameter and impactor flow rate combinations for
 achieving  near-isokinetic sampling conditions can be made from Figure 3.
 If a choice must be  made between undersized and oversized nozzles,
 undersized nozzles will usually result in lower sampling errors than
 will oversize.
 5.6  PRE-CUTTER USE
     In many instances  the percentage (by weight) of material with sizes
 larger than the first impaction stage cut point is quite high.  In such
 cases  a pre-cutter cyclone is necessary to prevent overloading on the
 upper  impactor stages.  A pre-cutter should always be used for the first
 test.  If  the weight of material collected by the pre-cutter is comparable
 to that on  the first stage, it should be used in all subsequent runs.
 Cyclones can be obtained from the impactor manufacturer or can be shop
 made.  A set of drawings of a cyclone for the Brink 1s included in
 Appendix C.  The basic  design can be adapted for attachment to other
 impactors.   The use of  two first impactor stages in series has also been
 suggested  and appears to be a valid approach; however, no data are
 available.
 5.7  SAMPLING TIME
     The length of the  sampling interval  is dictated by mass loading and
size distribution.   An estimate for initial tests can be obtained from
Figure  4.   Two  conflicting criteria complicate the choice of the sampling
                                   16

-------
                                                     NOZZLE SIZES
•a
K

Ul
3.

-------
00
                    00023
                                         OOZ3
    FUJE GAS MASS LOAOWS, units at right

C123                   2£»
                                                                                                        223
                                                                                                                             22B-9/I
                                                                                      READ DOWN FROM MASS LOADING TO IMWCTOR
                                                                                      SAMPLE RATE,  READ LEFT TO TIME  REQURED
                                                                                      TO COLLECT A SO MG SAMPLE AT THAT SAMPLE
                                                                                      RATE.
                                                                                                                             MOar/acf
      ocfm

       I/coin-
                                                                                                                    o.oe   o.a
                                                                                                      2.05                 OZ83

                                                                                              IMPACTOR SAMPUN6 RATE, vail* at Uft
                               Figure  4.   Nomograph for sampling time selection  (50  mg  sample).

-------
time.  It is desirable from the standpoint of minimizing  weighing  errors
to collect several  milligrams on each stage.   However,  most size distributions
are such that the upper stages are overloaded and are re-entraining
particles by the time the lower stages reach  a few milligrams.  A  rule
of thumb is that no stage should be loaded above 10 mg, but the determining
factor is whether or not re-entrainment occurs.   As is  dicussed later,  a
comparison of the relative distribution determined by a long run with
that from a shorter (about half as long) run  can be used to check  on  re-
entrainment due to stage overloading.
5.8  READYING THE IMPACTOR
     As equipment is not always cleaned up as well as it should be,  the
impactor should be inspected prior to use.  The nozzles must be clean,
gaskets in good shape, and the interior clean.  Nozzles can be cleaned
with fine wire if necessary.
     After inspection, the impactor should be carefully loaded with  the
preweighed stage substrates and assembled.  Teflon-thread sealant  tape
or antiseize compound should be applied to the threads, especially when
high temperatures (>215°C) are encountered.  The thread sealant tape
generally works better and causes fewer problems but probably cannot be
successfully used at temperatures above 290°C.
     If supplemental heating is required, a heating device and temperature
monitor need to be added.  A thermocouple mounted in the gas flow immediately
after the final filter and inside the impactor is best for measuring the
gas temperature within the impactor.
     The supplemental heat can be supplied with either a heating  mantle
which has been made to fit the impactor or by using heating tapes.  If
the tapes are to be used, a heating tape of sufficient wattage is  wrapped
around the impactor.  Glass fiber tape works well for holding the heating
tape.  Insulation such as asbestos tape is then wound around the impactor.
Glass fiber tape is again used to hold the asbestos in place and also
acts as additional insulation.  The impactor can now be mounted on the
                                   19

-------
 appropriate  probe,  taken  to  the sampling position, and Installed in the
 sampling  system.
 5.9   PRE-SAMPLE CHECKS
      Impactors are  prone  to  leak, and they must be checked for leaks at
 operating temperature.  This can be done 1n several ways.  The nozzle
 can  be  plugged and  the  impactor pressure-tested or vacuum-tested.
 Because impactors are basically a series of orifices, they should have a
 constant  flow-to-pressure-drop relationship.  Checking the pressure drop
 on various flows of filtered air will point out deviations from normal
 operations—both leaks  (external or internal) and plugged jets.
 5.10  TAKING THE SAMPLE
      The impactor  should be preheated for at least 45 minutes before
 sampling.  If supplemental heat is being used, the Impactor should be
 brought up to temperature outside the duct and then allowed some time to
 equilibrate after insertion.  The nozzle should not point Into the flow
 field during this phase.  Without supplemental heat, the whole warm-up
 is conducted within the duct, again with the nozzle pointed away from
 the flow  field.  Capping the nozzle during preheat 1n the flue is also
 desirable.
      A predetermined flow rate must be maintained to ensure stable cut
 points.   Any attempt to modulate flow to provide isokinetic sampling
will destroy the utility of the data by changing the cut points of the
 individual stages.  Rapid establishment of the correct flow rate is
especially important for the short sampling times typically found at the
inlets  to control  devices.
5.11  NUMBER OF SAMPLE POINTS
      As  the velocity and partlculate distributions in industrial duct-
work are  unlikely to be ideal, a large number of samples are often
required for accurate particulate measurements.  A velocity traverse
should  be run to check on the velocity distribution.  At least two
                                     20

-------
points within a duct should be sampled in each measurement plane,  and  at
least two samples taken at each of these points.   These are the minimum
sampling efforts and are appropriate only for locations with well  developed
flow profiles in the absence of significant concentration stratification.
If the flow profile at the station is uncertain due to duct configuration
and/or the mass loading is not uniform, the number of samples may  need
to be increased for reliable results.
                                    21

-------
                                SECTION 6.0
                       SAMPLE RETRIEVAL AND WEIGHING
6.1   IMPACTOR CLEAN-UP
      Careful disassembly of the impactor and removal of the collected
participate are essential to the success of the test program.  The
crucial points are to make sure that the collected material stays where
it originally impacted and to remove all the participate.  After the
sampling run, the impactor should be carefully removed from the duct
without jarring it, removed from the probe, and allowed to cool.
Disassembly can be difficult in some cases, particularly if the impactor
was used at elevated temperatures.
      Typically, not all of the dust which collects 1n an Impactor collects
on the substrates.  Some accumulates on the interior surfaces, especially
in the nozzle.  By convention, all of the particulate collected upstream
of a  given impactlon stage is assigned to that stage.
      The collection of this "misdirected" particulate is often troublesome.
If the dust is hard and dry, the particulate can be brushed off into the
weighing container.  A No.7 portrait brush or its equivalent is suggested,
and care must be taken to prevent brush hairs from contaminating the
sample.  If the particulate is sticky or wet, some type of washdown
procedure should be used.  The solvent must be considerably more volatile
than  the particulate.
6.2  DRYING AND WEIGHING
     All of the particulate must be dried to constant weight, with 2
hour checks used to establish the uniformity of the weights.  Hard, non-
volatile particulate is often dried in a convection oven to 212°F,
                                  22

-------
desiccated until cooled to room temperature, weighed, then check-
weighed.  Volatile particulate will require some other technique using
low temperature.  Whatever the technique used, constant weight of the
sample with further drying is the criteria to be met.
6.3  DATA LOGGING
     Permanent records should be kept of all pertinent Information.   It
is generally necessary to keep records 1n three places—In the lab with
the balance (using a bound notebook), and (using either looseleaf data
forms or a bound notebook) at.both the Inlet and outlet of the control
device.  Table 2 presents a fairly complete listing of the Information
required concerning an impactor run.   Notes should be taken on any
abnormalities which occur and on the apparent condition of the collected
particulate.
                                    23

-------
     TABLE 2.  SAMPLING INFORMATION REQUIRED
0    Date
     Time
0    Run Code Number
0    Impactor Type and Identification Number
0    Operator
0    Port Number/Sampling Location
0    Ambient Temperature
0    Ambient Pressure

0    Impactor In-Stack or Out-of-Stack
0    Impactor Orientation
0    Number of Traverse Points
0    Stack Pressure
0    Stack Temperature
0    Nozzle Diameter/Type
0    Probe Depth, if used
     Stack Pi tot Tube Delta P/Stack Gas Velocity
o
     Desired Impactor Flow Rate for Isoklnetic Sampling
     Metering Orifice Identification Number
     Metering Orifice Delta P
     Impactor Temperature
     Scalping Cyclone in Use?  Identification
0    Prefilter Identification
0    Postfliter Identification
0    Substrate Set Identification
0    Pressure Drop Across Impactor

0    Test Start/End Time:  Duration of Test
0    Gas Meter Start/End Readings:  Gas Meter Volume
0    Agreement Between Meter and Orifice
0    Volume of Condensible H20 in Flue Gas
0    Gas Meter Temperature
                     24

-------
                                 SECTION 7.0
                    SOME QUALITY ASSURANCE TECHNIQUES

7.1  NEED FOR A QUALITY ASSURANCE (QA) PROGRAM
     The field use of cascade impactors is a difficult task.  The
accuracy required is more appropriate for a laboratory program than for
a fiold test.  There are many places in the operational sequence where
errors can occur in spite of a conscientious effort to do a good job.  A
quality assurance program attempts to discover Inaccuracies before they
are propagated throughout the test program.  It 1s beyond the scope of
this manual to delineate a complete QA program.  Such a program 1s being
prepared and will be issued separately.  The techniques presented 1n
this section are not the only ways to ensure quality data.  However,
they have been used successfully 1n field testing with Impactors.
7.2  IMPACTOR TECHNIQUES
     Glass Fiber Substrates -- As has been discussed previously, glass
fiber substrates are not without problems.  Two potentially serious
problems are S02 uptake on the substrate and mechanical or manual abrasion
of the filter mat.
     The problem of S02 uptake on the substrate 1s discussed by Smith,
et al.   Two approaches which have been tried are:  to use a substrate
which does not change weight in the flue gas; or to precondition the
substrate in filtered flue gas prior to weighing.  Using a new glass
fiber material, which does not react with the S02, may alter the particle
retention characteristics of the impactor and change the Impactor's
calibration.  This must be checked and the data reported.  The use of
 Smith, W. B., K. M. Gushing, G. E. Lacey, and J. D. McCain, "Particulate
Sizing Techniques for Control Device Evaluation," EPA-650/2-74-102a
(NTIS No. PB 245184/AS), August 1975.
                                    25

-------
 preconditioned filter mats  requires  that  the  glass  fiber  substrates be
 preconditioned long enough  to reduce the  weight change during the expected
 duration of the Impactor runs to  10  percent or less of the minimum stage
 weight.   At the present time, this S02  reaction phenomenon 1s not well
 understood, and only rough  guidelines are available.  For some common
 glass fiber materials tested, the saturation  times  were on the order of
 2 to 6 hours at the temperatures  tested.   Check with  the  IERL-RTP Process
 Measurements Branch for the latest Information 1f glass fiber substrates
 are to be used.
      The applicability of the method chosen to overcome this substrate
 problem must be tested during the presurvey and periodically during the
 test runs by running blanks.
      Glass fiber substrates must  be  handled carefully to  prevent damage
 and possible loss of fibers.   Loose  surface fibers  should be removed by
 shaking  prior to weighing.   After weighing, every precaution must be
 taken to prevent the loss of any  part of  the  substrate.   One approach
 which will  quantify the problem of substrate  abrasion 1s  to prepare a
 substrate set,  load the impactor, then  disassemble  and rewelgh.
      Greased Metal  Substrates  —  The problems which occur with the use
 of greased substrates are usually related  to  the properties of the
 grease.   A grease which has been  applied  too  heavily  or has a low viscosity
 at operating temperature can  be physically blown off  the  Impactor stage.
 The grease could also react chemically  with the flue  gas  or be excessively
 volatile at the  operating temperature.  Again, these  phenomena must be
 checked  during  the presurvey  and  periodically during  the  test program.
      Re-entrainment —  Re-entra1nment is  the  phenomenon of an impacted
 particle being  blown  off the  stage on which 1t was  collected initially
 and  being collected downstream.   This can  be  caused by excessive jet
 velocities  or by overloaded  stages.   The effect of  re-entrainment can be
 serious,  because only a  few large particles on a stage which should
 collect  small particles  can considerably affect the size  distribution.
     One way to  spot  re-entra1nment  is  to  very carefully  examine the
 stage catches.   If, for  example,  a low  velocity through the jets resulted
 in a well-defined pile of partlculate and  a high velocity sample gave a
diffuse deposit,  re-entrainment should  be  suspected at the high sampling
                                    26

-------
 rate.  Microscopic examination of the lower stages and final  filter for
 large particulate (which should have been collected upstream) 1s another
 way to check for re-entra1nment.
     Re-entrainment due to stage overloading can be detected by running
 two otherwise Identical tests for two different test durations.  If the
 two size distributions are not the same, overloading should be suspected
 at the higher stage loadings.
     Impactor Leaks — Two types of leaks can occur with impactors—
 Internal or external.  A flow rate versus pressure drop check or a
 pressure test will pick up most leaks.  An Internal leak, where part of
 the airstream is bypassing the proper flow path, will give results
 similar to re-entrainment.  Leak checks must be made at operating
 temperature.
     General Procedure -- A general procedure for Impactor use, concentrating
 on quality assurance, is outlined below:
     1)   Preparing Impactor
          a.   Wash impactor, using ultrasonic cleaner if available.
          b.   Visually check cleanliness.  Jets must be clear,
               sidewalls clean.  Must be done in good lighting.
          c.   Obtain preweighed substrates and assemble impactor.
     2)   Sampling
          a.   Assemble Impactor train and heat to operating temperature.
          b.   Leak-check the impactor.
          c.   Sample with Impactor.
          d.   Disassemble Impactor, examine stage catches and
               Impactor walls.  Note any anomolies.
     3)   Substrate and Re-entrainment Checks
          a.   Check during presurvey.
          b.   Check substrates if flue gas composition changes
               significantly.
7.3  WEIGHING TECHNIQUES
     Precision and Calibration — The manufacturer's directions should
be followed when operating the balance.  The balance should be calibrated
at least once a day.  The repeatability of measurements should be checked
by repeatedly weighing a substrate and a test weight.
                                    27

-------
      Technique  -- The assembly and disassembly of an Impactor should
 have  no  effect  on the substrate weights.  This should be checked by
 weighing a  set  of substrates, assembling them in an Impactor, then
 disassembling and reweighing.  Any weight losses from this process
 should be within the repeatability of the balance (approximately 0.2
 milligram for an electrobalance).  Dry. weight checks are made by desiccating
 the substrate,  weighing, then desiccating again and reweighing.  When
 the agreement is within the repeatability of the balance, dry weight has
 been  achieved.
 7.4   GENERAL NOTES
      Spare  Parts -- The well-equipped sampling team will travel with an
 adequate supply of spares.  Improvisation due to an equipment failure
 can lead to poor quality data.
      Flow Meters -- At least two flow meters should be used 1n series.
 If they  do  not  agree, the problem should be investigated.
      Pumps  -- Eventually, vacuum pumps 1n sampling trains begin to leak.
 Flow  meters placed upstream of the pump prevent incorrect flow measurements,
 due to leaking  pumps though they complicate the calculations.
 7.5   DATA ANALYSIS
      Final  Filter Data -- The fine particulate Information obtained from
 the final filter can sometimes be misleading.  It is assumed for analysis
 that  a stage captures everything larger than Its DgQ, and captures
 nothing  smaller.  A real stage misses some large particulate.  Under
 some  conditions (including but not limited to re-entrainment), large
 particles will  penetrate to the final filter.  In this case the size
 distribution will be skewed toward the small particles.  Microscopic
examination of  the final filter may provide an Indication of this problem.
 If it occurs, the best choice in data analysis 1s probably to Ignore the
final  filter on runs where this phenomenon was encountered.
     Cumulative Size Data Analysis — If a probe or a pre-cutter cyclone
is used on an impactor,  the probe losses and pre-cutter catches must be
included in cumulative size analysis.  Failure to do so will  lead to an
incorrect cumulative distribution.
                                    28

-------
     Inspection of Data -- After the data have been collected, they
should be examined for any Inconsistencies or weak points.   For example,
the data presented as part of the "Sample Calculation," Appendix G,
display two possible errors:  the mass of sample caught on  Stage 6,
0.008 mg, is inadequate for an accurate weight determination; and the
mass caught on the final filter seems high in comparison to the mass on
the other small particle stages, and re-entrainment should  be suspected.
                                      29

-------
                                 SECTION  8.0
                                DATA ANALYSIS

 8.1   CASCADE  IMPACTOR DATA ANALYSIS
      The information  directly  available  from a  cascade  Impactor  Is the
 weight of particles caught on  a stage.   There are several ways to
 analyze and present this data  as particle  size  distributions and
 fractional efficiencies across  collection  devices.  The method which has
 been  found most  generally applicable  is  the "D50 method," described
 below.
 8.2   D5Q METHOD
      The DQQ method is presently used for  the majority  of cascade
 Impactor data reduction.  The method  is  fairly  straightforward and can
 be hand-calculated, but results  1n  a  somewhat simplified picture of the
 real  distribution.
      The Dg0 of  a stage is the  particle  diameter at which the stage
 achieves 50 percent efficiency;  one half of the particles of that
 diameter are captured and one half  are not.  Figure 5 shows a complete
 set of  theoretical capture efficiency curves for a modified Brink
 Impactor. The DCQ of Stage 4, for example, 1s about 1.2 ym.  The cal-
 culation of stage D5g's 1s discussed  below.
      The D5Q analysis method simplifies  the capture efficiency dis-
 tribution by assuming that a given  stage captures all of the particles
with  a diameter  equal to or greater than the DCQ of that stage and less
than  the D5Q of  the preceding stage.  With this simplification, the mass
collected on a given stage can be assigned to a particular diameter.
      Particle-size distributions may be  presented on a differential or a
cumulative basis.  When using the D5Q method, either type of presentation
may be easily employed.
                                      30

-------
                                                        ^K^CYCUONE 3

                                                     r     ^ ****** Ak**» M
ojot    0.1   as i  2    5  to  20
80  90 95  96 99     9919   99.99
                      COLLECTION EFFICIENCY., percent


   Figure 5.  Stage collection efficiency for a modified Brink Impactor
              {T=72°Ff  P*29.60 in. Hg, p=1.35 g/cn3,  flow = 0.03 acfm).
                                    31

-------
     The  size parameter reported can be aerodynamic diameter, aero-
 dynamic 1mpact1on diameter, or Stokes diameter.  In all cases, the
 particles are assumed to be spherical.  The method of reporting
 diameters depends to a large extent upon the Ultimate use of the size
 distribution Information.  For this reason 1t 1s suggested that the data
 be  reported in three parallel sets:  one set based on aerodynamic
 impaction diameters, one based on aerodynamic diameter, and one based on
 the Stokes diameter.
 8.3 CALCULATION OF THEORETICAL STAQE D^'S
     The  reduction of field data obtained with a cascade Impactor can
 sometimes be troublesome and time consuming because of the computations
 Involved.  The equations below are based Oh the motion of particles for
 which the Reynolds number is less than 1.0.  Although this 1s not always
 true for  impactors, the equations are often a good approximation.  The
 basic equation that defines the theoretical 1mpact1on behavior of a
 given stage of a cascade Impactor is:
                             f
                              18 * u 0.
where:
              1 + —g- I 1.23 * 0.41 exp (  "Utr "'"II           (2)
D = diameter of particle Impacting on the stage, cm
ijj = Stokes impaction parameter, dimensionless
v = viscosity of gas at conditions Immediately downstream of
     Impactor jet(s), poise
Dj= diameter of Impactor jet, cm*, or width of slot Impactor, cm
C = Cunningham Correction Factor
p * density of particle, g/cm
V.= velocity of gas through an Impactor jet, cm/sec
i • mean free path of air molecule at impactor stage conditions* cm.
*The expression for Cunningham Correction Factor 1s empirical; several
slightly different versions'are available in the literature.
                                     32

-------
     As equation (1) 1s written, with the actual  particle density and
the calculated Cunningham Correction Factor, 1t defines the Stokes
diameter. If the particles are treated as if their density, p ,  was 1.0,
equation (1) defines the aerodynamic diameter.   If the Cunningham
Correction Factor 1s also assumed to be equal to 1.0, the aerodynamic
impaction diameter 1s defined by equation (1).
     The collection efficiency of Impactor stages has been both
theoretically and experimentally correlated with the impaction parameter,
ip.   For the purposes of the DSQ method, the value of  at 50 percent
collection efficiency must be known.  These values have been reported by
             2
Ranz and Wong , among others, and are presented below:

          for round jet impactors, i|>50 • 0.145
          for rectangular jet Impactors, I|»CQ = 0.44.

Other Investigators have presented other values for *, and 1f an experimental
value for the Impactor under consideration 1s available, 1t should be
used.
     If the value of ^5Q 1s substituted Into equation  (1), it becomes an
equation with which to calculate the D50 of a given stage of the Impactor.
For a round jet Impactor, equation (1) becomes:
                                           1/2
                               /2.61 y D,
                          D50
     Equation (1) can be put in a form which is more convenient for
field use by substituting the volumetric flow rate Into the Impactor for
the velocity at the jet, obtaining:
ZRanz, W.B. and J.B. Wong, "Impaction of Dust and Smoke Particles on
Surface and Body Collectors," Ind. and Eng. Chem.: 44, No.6,  1371-81
(1952).                                   	
                                    33

-------
                         2.05 y D?
                  D   -'	i
                   50  I  C Pp Qs Ps
where:   X.  =  number of jets on the stage
         P.  =  absolute pressure immediately upstream of jet(s), mm Hg
         Q.  =  volumetric flow rate through impactor at stack
                 conditions, cm3/sec
         PS  = absolute pressure in stack, mm Hg.
     Slmi'larly, the i|»g0 for rectangular jet impactors can be sub-
stituted into equation (1), and the D5Q equation for slot impactors ob-
tained:
                                          1/2
                              3.485 v W. '
                       D50  "
where:  VL = the slot width, cm.

Equation (4) can be converted to a volumetric flow rate basis:
                                                                     (4a)
where:  L. = the slot length, cm.

     One approach that can be used to further simplify the computations
is to develop curves for the impactor stage cut points at one set of
conditions; e.g., air at standard conditions and a particle density of
1.0.  Then a suitable correction factor can be applied to these curves
for the actual sampling conditions.  Unfortunately, further simplifications
                                    34

-------
are involved in making the correction factor simple enough  to  be  of
value.  Therefore, the use of this type of approach suffers from  some
restrictions.  Figure 5, presented earlier, shows a calibration for dry
air at laboratory conditions with an assumed particle density  of  1.35
g/cm .
     All of the assumptions and calculations involved in going from
equations (1) and (2) to the calibration curve can be quite awkward,
particularly in cases where different types of sources are  being  sampled.
Perhaps the best approach is to write or obtain a computer  program based
on the rigorous equations given initially.  The program can both  calculate
impactor stage cut points, and compute concentrations of particles in
each size range, as well as differential and cumulative size distributions.
It permits more sophisticated data reduction methods to be  used  than
would be possible by hand.  Manual calculation of cut points typically
takes several hours; using a computer program, several sets of data  can
be calculated in a few minutes.  Not only is the computer faster but the
possibility of computational errors 1s greatly reduced.  Programs for
both set-up calculations and data reduction are available for  use on
small, hand-held, programmable calculators.3
8.4  CALCULATION OF STAGE D50'S FROM CALIBRATION DATA
     Some very recent research has pointed out that the theoretical
relationships presented in Equations (1) and (2) are not always  accurate
in commercial impactors.  Work is in progress to define the full  extent
of the errors.  Equation (3a) has been modified by the Inclusion of an
empirical constant, and the constant is calculated from calibration
data.  The new form of Equation (3a) is:
                        K
                                    3
                            2.05 u D? X,P
where:  K = empirical constant, dlmenslonless, varying with  Impactor
            and stage.
3Cushing, K., J. McCain, J. W. Ragland, and W. Smith,  "HP-65  Programmable
 Pocket Calculator Applied to Air Pollution Measurement Studies-Stationary
 Sources," EPA-600/8-76-002, October 1976.
                                     35

-------
      D5o's from Call*brat1°n ^ta should be used whenever available.   If
 calibration data are not available, the theoretical  equations can be
 used.  As before, this equation is best used as part of a computer
 program for the calculation of D50's.   Similar calibration constants can
 be determined for slot impactors.
 8.5  DIFFERENTIAL PARTICLE SIZE DISTRIBUTIONS -  DCft METHOD
                                                   bU
      The true particle size distribution of almost any particulate-
 containing stream outside of a laboratory is a smooth and continuous
 curve.  As impactors have a finite  number of stages,  they break this
 continuous particle size distribution  into a series  of discrete piles  of
 particulate of different size intervals.   The challenge of impactor  data
 analysis is to transform the discrete  data into a good approximation of
 the real, continuous distribution.   The D5Q method described  below is
 commonly used.
      It is assumed for the purpose  of  analysis that  all  of the mass
 caught  upon an impaction stage consists of material  having aerodynamic
 diameters equal  to,  or greater than, the  DSQ for  that stage,  and  less
 than the D5Q for the next higher stage.   For the  first stage  (or  cyclone),
 it  is assumed that all  of the material  caught has aerodynamic diameters
 greater than,  or equal  to,  the D5Q  for  that stage  (or cyclone), but  less
 than the maximum particle size.  If the maximum particle  size is  not
 known,  some arbitrary large value,  say  100  ym,  is used.
      As  the true particle size distribution is  continuous, the  amount of
 material  having  diameters between D and D+dD can be  represented by dM.
 Then the integral
                             D2/*
                               I  dM
                              J  dD
D',
  ml  JM
        dD
     QU
'1
yields the total mass having diameters between D, and D2
                                     36

-------
            Many  cascade  impactors are designed so that the relationship between
       successive stage D50's  1s logarithmic.  For this reason, and to minimize
       graphical  scaling  problems, the differential particle size distributions
       are  plotted on  log-log  or semi-log paper with dM/d(log D) as the ordlnate
       and  log D  as the abscissa.  The mass on stage "n" Is designated by AM
       and  is, In approximation, the mass of particulate with diameter between
       ^D50^n and (D50^n+r  The A(lo9 D) associated with AMn is log
       (D50)n+i - log  (D5Q)n.  Using these approximations, the derivative term
       associated with stage "n" is:
                                 AM
                                   n	    _      Mass on Stage "n'
^ nvi       _ n        _       ass
ogD)|n  '   Ado, D50)|n   -  1og(D50)
                                                    50n+l
           Plotting this approximation of dM/d (log D) versus log D results 1n
      a histogram.  From such a histogram, the total mass of particles with
      diameters between (D).- and (D)- can be calculated as the sum:
                                          AM,
                                            *
Z
                                    A(log D  )
      where "fe" takes on values corresponding to the discrete increments of
      the histogram.
           If an impactor with an infinite number of stages were available,
      the histogram would approach a continuous function, the A(log D5Q) terms
      would approach d(log D), and the mass between Dm and Dn could be calculated
      as:
                                     D
                                            HM
                                     Dm
                                          37

-------
 Such  an  impactor does  not exist, but the histogram can be plotted as a
 smooth curve  by assigning some average of  (D50)n+1 and (D50)n to the
 AM/Alog  D^glp term.  The geometric mean of the 050*5 1s often used,
 This  curve  is then a continuous function approximating the actual particle
 size  distribution.  Such a curve 1s needed to calculate fractional
 efficiencies  of control devices if the D^Q's differ for inlet and outlet
 measurements.  The accuracy of the approximation is limited by the number
 of points,  and by the  basic Inaccuracy of neglecting the non-ideal
 behavior of the impactors, especially overlapping collection efficiencies
 for adjacent  stages.
      To  normalize the  differences in mass of sample collected by various
 instruments,  the mass  on each stage is usually divided by the standard
 volume of the sample,  yielding concentration units; I.e., dC/d(log D).
 A sample calculation using impactor data 1s presented as appendix G
 of this  document.
 8.6   CUMULATIVE PARTICLE SIZE DISTRIBUTIONS
      The data  may be presented on a cumulative basis by summing the mass
 on all the  collection  stages and back-up filter, and plotting the fraction
 of the mass below a given size versus size.  This is frequently done on
 special  log-probability paper.  Semi-log paper may be preferable for
 distributions  that are not log-normal.
      Cumulative distributions are very easy to understand and present
 the data with  clarity.  For this reason, they should be presented as
 part  of  each  particle  sizing report.  Cumulative distributions do have a
 couple of disadvantages when compared to differential distributions.  An
error in stage weight will  be propagated throughout a cumulative analysis,
 but will  be isolated by the differential approach.  Also the differential
method does not involve the use of total mass concentration or total
 size distribution from diameters of zero to infinity, and so 1s useful
 in comparing instruments with overlapping but different size fractionation
ranges and different stage cut points.
                                   38

-------
       When cumulative plots are used,  the abscissa  Is normally  the logarithm
  of the particle diameter and the ordlnate  Is  the weight percent smaller
  than this size.  The value of the ordlnate at a given  (D50). would be
Weight percent smaller than
In } s
(05Q'k
fe-1
X~"*
7 AMj
^ 1
K
                                                              * 100%
                                              Z
                                               1=0
 where:
 1  = o corresponds to the filter,
 1  = fe corresponds to the stage under study,  and
 i  = K corresponds to the coarsest jet or cyclone

      This  equation requires  that the stages  be  counted  from  the  final
 filter up.  There 1s no (D50)Q,  as  the "o" corresponds  to the filter.
 (°5g)-| 1s  the  cut point of the  last stage, which collects mass,  AM,.
     An analytical  curve can  be  fitted to the cumulative distribution
obtained above, and  values of dM/d(log D) obtained by differentiation
of the analytical  expression.   In general this  requires some a priori
assumptions 1n determining the form of the expression to be used 1n the
curve fitting process,  but several  Independent groups have used this
technique to good advantage.
                                     39

-------
                                 SECTION 9.0

                                    REPORTS

9.1  SIZE DISTRIBUTION
     As a minimum requirement, size distribution information should be
reported in cumulative size distribution plots.
     The plots should be presented on log-probability paper or on semi-
log paper.  Plots are to be prepared based on aerodynamic diameter,
physical diameter, and aerodynamic impaction diameter.  The total sample
                                        o
weight in mg and the sample volume in Nm  must be included on the plot.
             o
(The term "Nm " means "normal cubic meters," with normal being defined
as a temperature of 20°C and a pressure of 760 mg Hg.)
9.2  DATA TO BE REPORTED
     The data from particulate size measurements must often be recast at
a later date for a purpose other than that for which they were gathered.
This can be an unsatisfactory situation unless all of the data which was
collected has been reported.  For this reason, all of the data which was
used to calculate the final distributions should be presented in the
final report, probably in the Appendix.  The data presented in Table 6-1
of Appendix G is almost complete.  The only additional Information
required would be port location, type of source, appearance of particulate,
and special  circumstances.
                                    40

-------
                                SECTION  10.0
                            COMMERCIAL  IMPACTORS
10.1 BRINK IMPACTOR
     The Brink Impactor 1s a six-stage,  low sample  rate, cascade  Impactor,
suitable for measurements In high mass loading  situations.  Appendix C
contains detailed drawings of another stage and a cyclone  for  the Brink.
The Brink uses a single round jet on each of its stages.
     Sampling Rate — The usual  sampling rates  for  the Brink are  in the
range of 0.05 -0.2  1/min (0.02  to 0.07 acfm).   The sampling rate must
be low enough to prevent re-entrainment of particles from  the  lower
stages.  With hard, bouncy particulate, the last stage nozzle  velocity
must be less than 30-35 m/sec with ungreased substrates, and less than
65 m/sec with greased substrates.
     Collection Substrates and Adhesives -- The Brink Impactor collection
stage is too heavy to use without a substrate.   Foil cups  are  commonly
preformed and fit into the collection cups of the  Brink stages.  If
grease is to be used, the top stages require about 5 or 6  drops of
solution while the bottom stages normally require  only about 1 drop in
the center of the cup.  Glass fiber substrates cut to fit  the  collection
cups have also been found satisfactory in many situations.
     Back-up Filter — The Brink back-up filter is normally made of
binderless glass fiber filter material.  Two l-1nch diameter disks  of
filter material are placed under the spring in the last stage  of the
impactor.  The filter is protected by a Teflon 0-r1ng and  the  second
filter disk acts as a support.
     Pre-cutter Cyclone — A pre-cutter cyclone for the Brink  is not
presently commercially available.  A drawing for shop construction  of a
cyclone is presented as Appendix C, along with details for further
modifications to the impactor which have been found useful.
                                    41

-------
      Sampling Train  —  The  Brink  uses  the  usual type of sampling train.
 Orifices  on the order of  0.03,  0.06, and 0.09  inches in diameter allow
 full  coverage of its range  of sampling rates at reasonable pressure
 drops.
      Brink Clean-up  j—  Careful  disassembly of  a Brink  impactor is
 necessary for obtaining good stage weights.  If a pre-cutter cyclone has
 been  used* all  material from the  nozzle to the outlet  of the cyclone is
 included  with the cyclone catch.  All  of this  material should be brushed
 onto  a  small, tared, 2.5  x  2.5  cm aluminum foil square to be saved for
 weighing.   Cleaning  the nozzle  1s also Important, especially if 1t is a
 small bore nozzle.   All material  between the cyclone outlet and the
 second  stage  nozzle  is  included with material  collected on the first
 collection substrate.   All  appropriate walls should be brushed off, as
 well  as around  the underside of the nozzle, where as much as 30 percent
 of  the  sample has been  found.
 10.2  ANDERSEN IMPACTOR
      The  Andersen impactor  is a relatively high sample rate impactor.
 Normal  sample rates  are about 15  1/min (0.5 acfm).   The Andersen is a
 multiple  jet, round  hole  Impactor.
      Sampling Rate -- The Andersen sampling rate is around 15 1/min (0.5
 acfm).  As  with  other cascade impactors, the flow rate  must be low
 enough  to  prevent re-entralnment  of impacted dust.
      Collection Substrates  and  Adhesives -- Andersen substrates are
 obtained  precut from the manufacturer.   The substrates are glass fiber
 and of  two types—one cut for the odd  numbered stages, one for the even.
 As discussed earlier, normal Andersen  substrates have  a tendency to
 absorb  S02 on basic  sites in the  substrate and, therefore, gain weight.
     The Andersen requires careful assembly, as overtightenlng will
cut the substrates or cause them  to stick to the metal  separator rings.
                                      42

-------
      Back-up Filter — The Andersen uses a 2-1/2 Inch diameter disk
 placed above the final F-stage.  (This F-stage 1s an option not normally
 included with the standard stack head.)  The filter should be cut from
 binderless glass fiber filter material such as Reeve-Angel 934AH filter
 paper or a similar material.
      Pre-cutter Cyclone — A pre-cutter cyclone for the Andersen is
 available from the manufacturer.  It is necessary to have a 6-inch or
 larger sampling port when using the pre-cutter cyclone with its nozzle.
      Andersen Sampling Train ~ The Andersen requires the usual type of
 sample train.  The pumping and metering systems of the commercial
 Method 5, EPA mass sampling train are appropriately sized for use with
 the Andersen.
      Care should be exercised never to allow a gas flow reversal  to
 occur through the impactor.   Material  could be blown off the collection
 substrate onto the underside of the jet plate or the collection substrates
 could be  disturbed.   A check valve  or maintenance of a very low flow
 while removing the Impactor from the duct avoids this problem.
      Andersen Clean-up -- Cleaning  an  Andersen impactor is difficult.
 Foils should  be  cut to hold  the substrates,  and each foil  and substrate
 weighed together before and  after the  run.   For disassembly,  the  foil  to
 hold the  stage  1  substrate should be laid out.   Next the nozzle and
 entrance  cone  should  be brushed out and onto  the foil.   Then  the material
 on  stage  0 should  be  brushed  onto the  foil.   The stage  1  filter substrate
 material  should  then  be placed  on the  foil and,  lastly,  the top of  the
 stage 1 plate 0-ring and  cross  piece should be  brushed  off.   Depending
 on  how tightly the  Impactor was  screwed shut,  some filter  material may
 stick to  the 0-ring edge  contacting  the substrate.   This  should be
 carefully brushed onto  the appropriate  foil.  This process 1s continued
 through the lower stages.  Finally, the filter  1s carefully removed.
 10.3  UNIVERSITY OF WASHINGTON MARK  III  (PILAT)  IMPACTOR
     The Mark III impactor 1s a  seven-stage, high flow rate device with
generally the same characteristics as the Andersen.  The Mark III 1s a
round hole, multiple jet impactor.
                                     43

-------
      Sampling  Rate  -- The  Mark  III sampling rate is on the order of 0.5
 acfm (15  Vmin).  The flow rate must be low enough to keep scouring of
 impacted  particles  to a minimum.
      Collection Substrates and  Adhesives — The Mark III has often been
 used with supplementary foil  (aluminum or stainless steel) substrates.
 These substrates  require the  use of grease for easily re-entrained
 particles.   Enough of the grease solution is placed evenly on the
 substrate to adequately cover the area under the jets.  The normal
 cautions  on the use of greased  substrates apply as discussed in the
 text.
      Pre-cutter Cyclone — A  BCURA (British Coal Utilization Research
 Association) designed pre-cutter cyclone is available from the manufacturer.
      Mark III  Sampling Train  -- As the Mark III is a high flow rate
 device, its sampling train is similar to that of the Andersen.
      Mark III  Clean-up —  Mark  III impactor clean-up is similar to that
 for  the Brink.  Some problems have been noted with 0-rings sticking
 rather tenaciously  and care must be exercised not to dislodge the sample
 while trying to separate the  stages.
 10.4 METEOROLOGY  RESEARCH,  INC. (MRI) IMPACTOR
      The  MRI impactor is a high flow rate sampler.  The body of the
 instrument is  constructed  with  quick-disconnect rings which allow flexibility
 in configuration  of the impactor and a positive gas seal between stages.
 The  impactor uses multiple round jets in its stages.
      Sampling  Rate  -- The  sampling rate is nominally 0.5 acfm in the
 seven-stage configuration.  Higher flow rates have been used by removing
 the  last  stage.
      Collection Substrates and  Adhesives — The MRI collection disk 1s a
 self-supporting foil (316  stainless steel) which is functionally similar
 to the collection cup or tray and inserts used 1n other Impactors.  The
 collection disks  are mass  produced and normally are used only once and
 discarded.
     Grease,  applied as  described earlier,  is recommended for most
applications.
                                      44

-------
     Back-up Fi1ter — The MRI Impactor has a built-in filter holder for
47 mm diameter filters.  Normally, blnderless glass fiber filters  are
used. Filter losses can be prevented by placing tared Teflon washers on
both sides of the filter during the test.
     MRI Sampling Train — The MRI sampling train 1s similar to that of
the Andersen.
     MRI Clean-up -- The cleannip of the MRI impactor 1s similar to the
Brink.  The device 1s clamped in a vise and all of the sections and
nozzles are loosened with wrenches.  The wall losses are carefully
brushed onto the appropriate collection disk.  Care 1s taken not to
brush contamination from the threads Into the sample.  A tared foil dish
Is used to collect the back-up filter.  Any worn 0-r1ngs should be
replaced and the whole unit carefully cleaned before the next test.
10.5 SIERRA MODEL 226 SOURCE CASCADE IMPACTOR
     The Sierra Impactor Is a six-stage, high sample rate cascade  Impactor.
The Sierra Instrument uses a radial-slot design.
     Sampling Rate — The Sierra Impactor has a nominal sampling rate  of
0.25 acfm.  The flow rate must be low-enough to prevent re-entralnment of
particles.
     Collection Substrates and Adhesives — Substrates for the Sierra
are obtained precut from the manufacturer.  These are glass fiber substrates
and should be checked for weight gain.  Stainless steel substrates are
also available and these should normally be coated with grease as  described
earlier.
     Back-up Filter — The back-up filter uses a 47 mm glass fiber
filter mat.  It 1s supported by a screen from below.
     Pre-cutter Cyclone — A pre-cutter cyclone 1s available from the
manufacturer.
     Sampling Train — The sampling train for the Sierra is similar to that
of the Andersen.
     Clean-up — Clean-up of the Sierra Is fairly similar to Andersen  clean-
up.  Care should be taken to be sure the glass fiber substrates are
removed intact.
                                    45

-------
                                SECTION 11.0

                                BIBLIOGRAPHY
1.   Andersen, A.A., "New Sampler for the Collection, Sizing, and
     Enumeration of Visible Airborne Particles," J. Bacterlol.:  76,
     471 (1958).                                                ~~

2.   Brink, J. A., Jr., "Cascade Impactor for Adiabatic Measurements,"
     Ind. and Eng. Chetn.: 50., 645 (1958).

3.   Cohen, J. J. and D. M. Montan, "Theoretical Considerations, Design,
     and Evaluation of a Cascade Impactor," Amer. Ind. Hyg.  Assoc.  J.;
     28, 95, (March 1967).

4.   Davies, C.N., and M. Alward, "The Trajectories of Heavy Solid
     Particles 1n a Two-Dimensional Jet of Ideal Fluid Impinging
     Normally Upon a Plate," Proc.  Phys. Soc.; B64, 889 (1951).

5.   May, K.R., "The Cascade Impactor:  An Instrument for Sampling
     Coarse Aerosols," J. Sc1. Instr.:, 22, 187, (Oct. 1945).

6.   Mercer, T.T., "The Interpretation of Cascade Impactor Data,"
     Amer. Ind. Hyg. Assoc. J.: 26_, 236-41 (1965).

7.   Parkes» 6.J., "Some Factors Governing the Design of Probes  for
     Sampling in Particle and Drop  Laden Streams,"  Atmospheric Environment;
     2, 477-490, (1968).

8.   Pilat, M. J., D. S. Ensor, and J. C. Busch, "Cascade Impactor for
     Sizing Partlculates in Emission Sources," Amer. Ind. Hygiene Assoc.  J.:
     32, No.8, 508-11, (August 1971).

9.   Smith, W. 8., K. M. Gushing, and J. D. McCain, "Particulate Sizing
     Techniques for Control Device  Evaluation," Environmental Protection
     Technology Series, EPA-650/2-74-102 (NTIS No.  PB 240670/AS),
     October 1974.

10.  Smith, W. B., K. M. Cushing, and G. E. Lacey,  "Andersen Filter
     Substrate Weight Loss," Environmental Protection Technology Series,
     EPA-650/2-75-022 (NTIS No.  PB  240720/AS), February 1975.

11.  Smith, W. B., K. M. Cushing, G. E. Lacey, and  J. D. McCain,
     "Particulate Sizing Techniques for Control Device Evaluation,"
     Environmental Protection Protection Series, EPA-650/2-74-102a
     (NTIS No. PB 245184/AS), August 1975.
                                     46

-------
12.  Strom, L., "Transmission Efficiency of Aerosol  Sampling Lines,
     Atmos. Envir.;  6, 133, (1972).

13.  Ranz, W. B., and J. B. Wong:   "Jet Impactors for Determining  the
     Particle Size Distribution of Aerosols," AMA Arch.  Ind.  Hyq.  Occup.  Med.:
     5: 464-77 (1952).                        	"	*"	

14.  Ranz, W. B., and J. B. Wong:   "Impactlon of Dust and Smoke Particles
     on Surface and  Body Collectors," Ind.  and Eng.  Chem.: 44, No,6,
     1371-81 (1952).                  	
                                   47

-------
                                SECTION 12.0
                                  GLOSSARY
Aerodynamic Diameter -- The aerodynamic diameter,  D.,  of  a  particle
     is the diameter of a sphere of unit density which would  behave
     in an impactor the same as does the actual  particle.
Aerodynamic Impaction Diameter. D«» — The  D^»  is  an  indicator of the
     way a particle behaves in an inertial  Impactor.   The DAI is
     numerically equal to:

                          DAI = "W^iF"

     where:    DS - Stokes  diameter of particle
               P  - particle density, g/cm3
               C  = Cunningham Correction Factor,  dimensionless.

Blank -- A blank run is one in which filtered process  gas is  run
     through the impactor to make sure that stage  weight  changes
     are due only to participate collection.  Volatile greases or S02
     uptake on glass fiber  substrates can be identified with  blanks.
Bounce -- Bounce in this document refers to inadequate attractive forces
     between a particle and the impaction surface.   If the  particle does
     not adhere, it is said to bounce.
Condensation — Condensation in an impactor refers  to  the coalescence
     of vapors as liquid particulate either in the  gas stream or on
     the impactor walls.
                                   48

-------
Control -- A control is a quality assurance step similar 1n purpose  to
     a blank.  Usually controls are used to check on substrate
     preparation and weighing rather than the effects of flue gases.
     An extra substrate is prepared and taken out to the field,  but
     not used.  After the run is over, the control substrate 1s
     desiccated and reweighed as are the used substrates.  Weight
     change by the control substrate indicates the magnitude of  handling
     and weighing errors.
Cut-point— The cut-point of an impactor stage 1s the particle diameter
     for which all particles of equal or greater diameter are captured
     and all particles with smaller diameters are not captured.   No
     real impactor actually has a sharp cut-point, but the D^Q of a
     stage is often called its cut-point.
D50 "" The D50 of an imPactor stage is the particle diameter at  which
     the device is 50 percent efficient.  One half of the particles  of
     that diameter are captured and one half are not.  The Dgg is
     normally calculated from the equation that describes the
     theoretical impaction behavior of Impactors at 50 percent
     collection efficiency:
where:
     i|;50 = Stokes inertial impaction parameter at 50 percent efficiency,
           dimensionless
      y  = gas viscosity, poise
     D.  = impactor jet diameter (for slot Impactors, the slot width), cm
      j
       C  * Cunningham Correction Factor, dimensionless
     p^  = particle density, g/cm3
      V. = gas velocity through impactor jet, cm/sec
       J
                                     49

-------
      This  equation can be used to calculate different diameters (Stokes,
      aerodynamic, or aerodynamic impaction) as described below:
      D50  s  is ca''cu''ated using the actual particle density and
         '   Cunningham Correction Factor,
      Drn  .  is calculated using the actual Cunningham Correction
      ou'rt  Factor, but with p  = 1.0, and
      Den  nr is calculated with both C and p>n = 1.0.
      oO,AI                                p

Grease -- In impactor terminology, grease is a substance which is
      placed on an impactor stage or substrate to serve as an adhesive.
Isokinetic  Sampling -- This is sampling with the bulk fluid velocity
      through the impactor nozzle equal to the velocity in the duct.
      This is necessary to prevent sample bias.
Physical Diameter -- See Stokes Diameter.
Pre-cutter or Pre-collector -- A collection device, often a cyclone,
      which  is put ahead of the impactor in order to reduce the first
      stage  loading.  This is necessary in some streams because the high
      loading of large particulate would overload the first stage before
      an acceptable sample had been gathered on the last stages.
Re-entrainment -- Re-entrainment in an impactor is the phenomenon of
      particles which impacted on a given stage being picked up by the
      gas stream and moving downstream to another stage.
Stage — A stage of an impactor is usually considered to be the
     accelerating jet (or plate containing multiple jets) and the surface
     on which the accelerated particles impact.
Stokes Diameter. D  — The Stokes diameter of a particle is the diameter
     of a sphere (with the same density as the actual particle) which
     would behave in an impactor the same as does the actual particle.
Substrate -- The removable, often disposable, surface on which impacted
     particles are collected.  Substrates are characteristically light
     and can be weighed on a microbalance.
Wall Losses — Wall  losses are the portion of the particles 1n the
     gas stream which impact with and adhere to surfaces 1n the
     impactor other than the substrates.   They must be collected and
     assigned to a stage.
                                    50

-------
                                APPENDIX A
                      IMPACTOR FLOW RATE MEASUREMENT


     The flow rate through an Impactor must be accurately measured In
order to set the isoklnetic sampling rate and to determine the correct
impactor stage cut-points.  Unfortunately, 1t 1s usually very Inconvenient,
and sometimes Impossible, to measure the Impactor flow rate at the
conditions present in the Impactor.  The gas 1s normally drier, cooler,
and at a lower pressure by the time the flow rate 1s measured; and the
flow must be corrected to Impactor conditions.   The use of calibrated
orifices and dry gas meters is discussed below.
Units
     The equations presented 1n the Appendix are valid only if the units
of the various terms are consistent.  For instance, the pressure drop
terms could be in units of Inches HgO or cm H20 or something else* but
all pressure drop terms must have the same units.  The same 1s true for
the other properties.  Note that pressure and temperature are both
absolute measurements.
Orifice Meters
     The gas flow rate through a particular orifice meter 1s related to
the pressure drop across that orifice by an equation of the form:

                                Q2 = C-f-                         (A-l)
where:
     Q  = volumetric flow rate at upstream conditions
     C.  = dimensional constant, (}ength)5(mass)(t1me)"2(force)"1
                                     51

-------
AP = pressure drop across orifice
 p = density of gas at upstream conditions.
    Solving for the constant, C, in equation (A- I), one obtains
                          C = -pfi.                       (A-2)

     As C is a constant at all conditions, its value can be obtained
at a convenient set of conditions with a known flow rate and used
later to calculate the flow rate.  Equation (A-2) can be rewritten to:
                                2
                          r    °c "c                      (A-2a)
                               iPc
     The subscript "c" indicates that these parameters were determined
during a calibration.  Density and flow rate are at upstream conditions.
     Substituting equation (A-2a) into equation (A-l) yields an equation
suitable for obtaining flow rates from a calibrated orifice:
                                                          (A-3)
                                            m
     The subscript "m" denotes the parameters of the gas as 1t Is
being "measured."  All are at conditions Immediately upstream of the
orifice.  For use with impactors, the measured flow rate, Q^, must be
converted to a flow rate at stack conditions, Q$.  Assuming that the
stack gas was dried as well as altered in temperature and pressure, the
stack flow rate is related to the measured flow rate by:
                  V1 -
                                     52

-------
where:
     F =  water removed from flue gas, expressed as a volumetric fraction
     P =  absolute pressure
     T =  absolute temperature
     The subscript "s" refers to stack conditions
     At the usual conditions of relatively high temperature and low
pressure which occur during stack sampling, the flue gas behaves very
much like an ideal gas.  The density of an ideal gas can be approximated
as:
                                                                 (A-5)
                           "     RT

where:

     MW = the molecular weight of the gas
      R = the universal gas constant

     Equations (A-3), (A-4), and (A-5) can be combined and rearranged into a
form which gives the pressure drop which must exist across the calibrated
orifice, APm> to obtain the required  impactor flow rate, Qg.
               Q2
                S   I*   c    \e. i    9   \ i  in  v.   i  /	in   \   /. e\
                    I' - Fu n)   -rr-5—  	?—  —no—    (A-6)
0-        HoO
wc
where:
      MW   = molecular weight of  the  stack gas at  the  orifice;
               normally the dry molecular weight.
      MW   • molecular weight of  the  calibration gas.
        w
                                     53

-------
Dry Gas Meter
     The dry gas meter, like the orifice, can only directly measure the
flow rate of the gas which passes through It.  This measured flow rate
can be converted to the flow rate through the Impactor (which is at
stack condition) using the equation:
                                        0 - FH20>                 
                                   54

-------
                                          EPA-600/2-76-118

                                          April  1976
                   APPENDIX B


     CASCADE IMPACTOR CALIBRATION GUIDELINES
                       by

Seymour Calvert, Charles Lake, and Richard Parker

                  A.P.T., Inc.
         4901 Morena Boulevard, Suite 402
           San Diego, California 92117
              Contract No.  68-02-1869
               ROAP No. 21ADJ-037
            Program Element No.  1AB012
      EPA Project Officer: Leslie E. Sparks

   Industrial Environmental Research Laboratory
     Office of Energy, Minerals, and.Industry
        Research Triangle Park, NC  27711
                  Prepared for

      U.S. ENVIRONMENTAL PROTECTION AGENCY
       OFFICE OF RESEARCH AND DEVELOPMENT
              WASHINGTON, DC  20460
                      55

-------
                                  ABSTRACT

     This report contains guidelines for routine calibration of cascade
impactors.  The basic calibration technique discussed in the report
involves generating uniformly sized particles, testing Individual stages,
determining particle number concentrations by light scattering, and
calculating efficiencies for given test parameters.  Each component of
the technique 1s discussed.  The results of calibrations of three cascade
Impactors and comparisons with published studies are presented.
                                  56

-------
                                  SECTION B-l
                                  INTRODUCTION

     The following report has been prepared at  the request of  the
Particulate Technology Branch of the Environmental Protection  Agency
(EPA) Utilities and Industrial Power Division.   The contents are  intended
for use as a guideline and experimental  calibration of cascade impactors
(C.I.) and for Air Pollution Technology  (APT)  internal purposes.
     The calibration process is based on approximately four years  of  APT
experience in connection with work for the EPA and other clients.   In
our experience, we have used University  of Washington Mark III, Anderson
non-viable (not the in-stack model), and APT M-l* cascade impactors and
have found that calibration of these devices is necessary for  accurate
sizing of particles in gas streams.  Variation of hole diameter and
shape due to production machinery, corrosive chemical action,  and particle
deposition can contribute to measurement inaccuracy when using impactors.
There are also discrepancies among the published studies of Inertial
impaction devices.  Many C.I. manufacturers have not thoroughly calibrated
their instruments experimentally but use the experimental results of
Ranz and Wong (1952) even though Mercer and Stafford (1969) and Stern,
et al. (1962) report much different results. Consequently, substantial
uncertainties exist in the prediction of C.I. performance from past
studies and manufacturer's specifications.
     The basic calibration technique used by APT  includes: generating
uniformly sized particles, testing individual stages, determining particle
number concentrations by light scattering, and calculating efficiencies
for given test parameters.
     The scope of this report covers the experimental technique, results
of calibrations of three C.I.s, and comparisons with published studies.
*Not available  commercially.

                                    57

-------
                                  SECTION B-2
                              AEROSOL GENERATION


 B-2.1  CALIBRATION SYSTEM
       The calibration system (Figure B-l) consists of an air-liquid
 atomizer, aerosol drier, charge neutralizer, particle counter, dilution
 air  lines and appropriate metering and flow equipment.  Further details
 on the components are included in succeeding sections.
 B-2.2  PARTICLES
       Monodisperse aerosols can be produced using suspensions of
 polystyrene latex (PSL) spheres available from Dow Chemical Corporation.
 Original suspensions as received are concentrated, 10 percent by weight,
 and  can be obtained with diameters ranging from 0.087 to 2.0 microns (urn)
 with a standard deviation of less than 0.01 urn, and a particle density
 of 1.05 g/cm3. Suspensions of PSL spheres larger than 2 pm are available;
 however, the standard deviations are much larger and the suspensions are
 more polydisperse.
     Particles of 0.5 ym to 2.0 ym diameter have provided a sufficient
 size range for calibration of the lower four stages of both U.W. Mark
 III and APT C.I.'s.  This is also the size range of most Importance in
 fine particle control device evaluation.   Consequently, APT has been
most concerned with calibrations in this size range.
     Useful  suspensions of PSL can be made by diluting small quantities
of the original  suspension with deionized water.   Any settling that
occurs over a period of days can be resuspended by gentle agitation.
The PSL is diluted to a concentration sufficient to minimize the occurrence
of agglomerates.   Dilutions of stock 10 percent solutions of PSL can be
estimated from a paper by Raabe (1969); however,  the amount of dilution
necessary actually depends on the specific device used for spraying the
                                    58

-------
                                II A II
       To Counter  -4-
                       «—x-
                      Dilution
        To Counter
              Filter
           Pressure-
            Gauge
Air

Pressure
Regulator  Filter!
                              t
                                    )C"B
                                    y ^
                                          t> To Manometer
                                         '^
                                       C.I. Case

                                       Spacer
                                       S\	Impaction  Plate
                                           """Met'Plate
                                           ^Impaction  Plate
                              Dilution
                               Mixer
                                            Rotameter


                                 Mixer  prying Tube

                                      -fr
                                            To Manometer and
                                           • Pressure Relief
                          Atomizer Fl Filter

                                        Aerosol Dilution
                                        Air Rotaraeter
                                    > f
                                    / v
                                Valve
                                                        o
                                                        o
                                                        c*.
                                                        o
                                                        «J
                                                        e
                                                        01
                                                        o
                                                        H
                                                        CO
o
o
<
                  Figure 6-1.  Calibration apparatus
                                59

-------
  Compressed Air
                     -,
                Liquid
                Drainag
                  0.159
                  cm ID
          Liquid
Figure B-2.  Collison atomizer
                              Aero s o1
                              Out let
0343s
: ID
0.1
"cm

Si
1]

i
)
                               Baffle
                            Mason  Jar
          60

-------
hydrosol.  Generally, the necessary dilution must be determined ex-
perimentally by observing collected samples of dried aerosol  with a
microscope.  If doubles run more than 2 to 5 percent, the suspension
should be further diluted.
     Concentrations of 0.004 percent to 0.03 percent (by weight) for
particles of 0.5 to 2.0 ym were found to be compatible with the spray
device in Figure B-2.
B-2.3  PARTICLE GENERATOR
       Drops containing PSL particles are produced from suspensions  with
a Collison type atomizer as shown in Figure B-2.  The operating range of
this device is normally 170 to 350 kPa (10 to 50 psig) and operating
characteristics vary with each device.  May (1973) has tabulated some of
the characteristics of a 3 jet model Collison atomizer (Table B-l).
     Number concentrations and drop size distributions produced by the
atomizer are consistent for a given operating pressure, provided that
the liquid level and concentration  are maintained.  Evaporation losses
will cause an increase in the particle concentration and the number of
agglomerates. This has not been a problem for test periods up to 3 hours
in duration. For longer periods it would be necessary to periodically
sample and test for agglomerates, and further dilute the PSL solution as
necessary.
     Drops leaving the atomizer are mixed with dry, filtered air
(approximately 45 1/min).  This minimizes agglomeration of wet particles,
dilutes the aerosol to a given number concentration, and aids in drying
out the wet particles.
     The aerosol is dried by passing it through a 1.2 m (4 ft) section
of a 3.8 cm (1.5 in) diameter glass tube mounted horizontally with a
layer of silica gel ("1.5 cm deep) spread evenly along the bottom.
Three "Staticmaster 2U500" 500 yc, Po210 alpha emitters (available from
Nuclear Products Company) are situated end to end at the "dry" end of
the glass tube to reduce the excess charge on particles to the minimum
level described by Boltzman's law.  No license is required to use these
units.
                                    61

-------
                            TABLE B-1.  OPERATING CHARACTERISTICS OF A 3  JET  COLLISON  ATOMIZER

Air Pressure
Free air consumption
Water loss, drop + vapor
Approx. water vapor output
Approx. drop output
Total water cone, in outlet port
Droplet cone, in outlet port
kPa
1/min
ml/hr
ml/hr
ml/hr
g/m3
g/m3
103
6.1
7.8
4.6
3.2
21.3
8.7
138
7.1
8.7
5.4
3.3
20.4
7.7
172
8.2
9.5
6.2
3.3
19.3
6.7
207
9.4
10.4
7.1
3.3
18.4
5.9
276
11.4
12.0
8.6
3.4
17.5
5.0
345
13.6
14.0
10.2
3.8
17.2
4.7
INJ

-------
                                   SECTION B-3
                       PARTICLE CONCENTRATION MEASUREMENT

B-3.1  PARTICLE COUNTER
       Particle number concentrations are determined using a Climet Cl  205
Particle Analyzer.  Other commercially available instruments utilizing
similar "electro-optical" techniques are also satisfactory.  The Climet
device has the capability of counting all particles with diameters greater
than a preset value (0.3, 0.5, 1, 3, 5, or 10 urn).  Further discrimination
can be achieved by using a potentiometer to provide a continuous selection
over the range from 0.3 to 10 ym.
B-3.2  COUNTING PROCEDURE
       The particle counter is used within a selected band of particle
diameters, centered about the known PSL diameter.  This reduces the effect
of spurious counts resulting from fine impurities and agglomerates.  The
particle count for the larger diameter setting may be subtracted from that
for the smaller diameter setting to determine the number concentration of
particles within a desired size interval.  It has been our experience that
spurious counts may still be a problem within the size ranges available on
the Cl 205.  Therefore, it is recommended that a potentiometer be used to
zero in as closely as possible to the actual PSL particle size.
     The maximum count allowed for the Cl 205 is 3.5 x 107/m3 (106/ft3).
The sample must be taken from a stream at or very near ambient pressure.
The Cl 205 requires a flow rate of 7 a/min.
     The sampling inlet arrangement is Illustrated in Figure B-3.  A 4 mm
OD tube is used for the inlet to the particle counter.  It is inserted a
few millimeters into the sampling tube (10 to 15 mm ID).  Thus,a sampling
                                      63

-------
             Sample

     Counter
     inlet tube
                        Excess
                        to vent
             Particle

             Counter
Figure B-3.  Sampling System Arrangement
                 64

-------
flow rate larger than 7 l/m1n is handled by letting the excess flow
exit through the annular space between the tubes.  This arrangement also
ensures that the Inlet flow to the particle counter 1s at atmospheric
pressure.  Tygon tubing and variable pinch clamps have been used
satisfactorily as throttles to control the sampling flow rate and the
flow rate through the impactor.  They are shown as "A" and "B" 1n Figure
B-l.
                                       65

-------
                                   SECTION B-4
                              CALIBRATION PROCEDURE

B-4.1  GENERAL CONSIDERATIONS
       There are alternative approaches to calibration which offer
advantages and disadvantages in terms of the amount of information gained
versus the time, effort, convenience, and simplicity.  Before going into
the details of the calibration procedure, a few important principles will
be pointed out.
     Determining the stage cut diameter Is the primary objective of the
calibration.  The computation of particle size distribution can be based
on stage cut diameters with good accuracy, so long as particle bouncing
on the upper (larger cut diameter) stages is prevented.  Therefore, the
calibration should concentrate on the particle size range in the vicinity
of the stage cut diameter.  For all stages, the inertia! impaction para-
meter (defined in equation B-l) at the stage cut point has a value of
approximately 0.2.  Therefore, it is suggested that the calibration of
any stage should cover an Impaction parameter range of 0.1 to 0.3.  These
values are for round jet impactors.
     The present procedure calibrates one Impaction stage at a time because
it is simpler, in determining collection efficiencies from inlet and
outlet particle concentrations, not to have to account for the contributions
of two or more stages in series.  Interference from the upstream plate,
noted by Willeke and McFeters (1975), has been observed mainly for large
particles in rectangular jets.  Very little effect has been noted for
small  cut diameter stages with round jets.  In calibrating single stages,
however, it is necessary to ensure that the flow pattern is very nearly
the same as in actual operation.  Therefore, the present procedure requires
an impingement plate to be placed upstream of the jet plate being calibrated.
                                     66

-------
 This arrangement 1s shown in Figure B-l.   Further experiments are
 presently underway to determine the Influence  of one or more upstream
 Impactor stages.
      Some question has arisen as to expansion  effects  on  hole diameter
 under hot and cold flow conditions.  The  most  practical method for
 resolving this problem is to calibrate the Impactor at various temperatures.
 B-4.2  DEFINITIONS
        The Inertial Impaction parameter 1s defined as  follows:

                      D2 C P_ V.     D2  V.       R
                v  =	P_J.	fll  J_ x 10~B                (B-l)
                "P     9 y Dj        9 y Dj      1U

 where K  = inertial impaction parameter, dlmenslonless*
       C  = Cunningham slip correction factor «

            1 + 2fa  [1.257 + 0.40 exp (-1.10 D/2i)]

      i.    = mean free path of gas molecules, cm
      D.,  = aerodynamic particle diameter,  ymA
      p    = particle density, g/cm3
      V.   = gas (particle) velocity through jet,  cm/sec
      y    = gas viscosity, poise, g/cm-sec
      D.   = jet diameter, cm
      ymA  = ym (g/cm3)-/2

 Aerodynamic diameter is defined as:

                  DAI = Ds (C pp)1/2 ]°4' ""*                      (B"2)
*The inertial  impaction parameter as defined 1n this appendix, K_, 1s
defined such that  it has twice the value of the t|» defined 1n thepbody of
the report.  The difference arises from an arbitrary decision 1n the
derivation of  equation (B-l); the 9 in the denominator of equation (B-l)
rather than an 18  is another consequence.
                                     67

-------
     For the case where the stage 1s 50 percent efficient (I.e.,  the
 cut point)  following  parameters  are substituted into equation (B-l).


      P5Q = inertial  1mpact1on cut parameter; K  at 50% efficiency
      D5Q = cut diameter or diameter (D) at which stage 1s 50% efficient

     DAI 50 = aerodynamic cut diameter
Thus,
                     D50  C pp V  =   DAI.5Q V
B-4.3  PARTICLE SIZE SELECTION

B-4.3.1  Obtain Theoretical Impactor Curves

         Theoretical curves of cut diameter versus flow rate through the
impactor are usually provided by the manufacturers of commercially available
cascade impactors.  These curves are generally based on some value of the

inertial impaction parameter at the cut point (I.e., K    in equation
(B-3)).  If these curves are not available, they may be approximated using

equation (B-3) written in the form:


                          0.135 TT y D? X. K
                DAI 50 = -  J  J  ?™                   (B-4)
                 AI*50         QS00'8)


where Q  = total flow rate, 1/min
      X. = number of jets or holes in an impaction stage
       J

A good approximation may be obtained using K_   =0.2.
                                            P50
                                     68

-------
     Curves of the aerodynamic cut diameter versus flow rate  for  the
seven-stage APT M-l cascade impactor are presented in Figure  B-4  and
are based on K_  = 0.2.
              P50
B-4.3.2  Determine PSL Particle Diameters and Flow Rates for  the  Calibration
         The PSL particle diameters available commercially are listed  in
Table B-2.  The standard deviation is very large for particles greater
than about 2 ym.  Therefore, 2 ym is the largest size PSL particle suitable
for use as a standard aerosol for calibration.
     The aerodynamic diameter corresponding to a 2.02 ym diameter PSL
particle can be calculated from equation (B-2) and is equal to 2.13 ymA,
a minimum flow rate of 20 1/min is required for Stage 4.  Therefore,  20
1/min is a convenient flow rate to use.  From Figure B-4 it can be seen
that a flow rate of 20 1/min corresponds to aerodynamic cut diameters  of
about 1.1, 0.6, and 0.4 ymA, for Stages 5, 6, and 7 respectively.  The
most suitable PSL particle diameters for each stage can be obtained from
Table B-2.  Figure B-5 is a convenient plot (from Calvert et al.  (1972))
for the conversion between aerodynamic and physical diameters.
     For example, assume that 0.5 ym diameter PSL particles are being
used to calibrate Stage 7 of the APT M-l impactor.  From Figure B-5, the
corresponding aerodynamic diameter for a particle density of 1.05 g/cm3
is about 0.58 ymA.  The flow rate required for a cut diameter of 0.58
MmA is obtained from Figure B-4 and equals about 8.5 1/min.  This cut
diameter 1s for K    = 0.2.   Using equation (B-2), the flow rates required
for Kn   » 0.1 and K0   = 0.3 can be calculated to be 4.3 1/min and 12.7
     "50            PBO
1/min respectively.
                                     69

-------
TABLE B-2.  AVAILABLE PSL PARTICLE DIAMETERS
       (Source:  Dow Chemical Company)

Avg. Dlam.
In Microns
0.087
0.091
0.109
0.176
0.234
0.255
0.312
0.357
0.364
0.460
0.481
0.500
0.527
0.600
0.721
0.760
0.794
0.801
0.804
0.807
0.822
1.011
1.099
1.101
2.020
5.7
15.8
One Std. Dev.
In Microns
0.0046
0.0058
0.0027
0.0023
0.0026
0.0022
0.0022
0.0056
0.0024
0.0048
0.0018
0.0027
0.0125
0.0030
0.0057
0.0046
0.0044
0.0035
0.0048
0.0056
0.0043
0.0054
0.0059
0.0055
0.0135
1.5
5.8
Density
Material g/ml
Styrene-Butadiene =1.05 unless other-
wise noted
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Styrene-Butadiene
Polystyrene
Polystyrene
Styrene-Butadiene
Polystyrene
Polystyrene
Polystyrene
Styrene-Butadiene 0.99
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polystyrene
Polyvinyl toluene 1.027
Styrene Divinyl benzene
Polystyrene
                    70

-------
!
u
t-l
I
u
                  Convenient particle ^
                  diameter selection
                  for calibration
   1.0 |bSii±:
   0.1
                   ACTUAL FLOW,  l/min
     Figure B-4.   Aerodynamic cut diameter vs.  impactor
                 flow for A.P.T. M-l cascade impactor.
                          71

-------
10.0
 0.1
    0.1
            1.0

PARTICLE  DIAMETER,  d.
10.0
 Figure B-5.  Aerodynamic diameter vs. diameter  for
             various densities.
                           72

-------
B-4.4   IMPACTOR PREPARATION
B-4.4.1  Inspect and Clean Impacti'on Plates
         Inspect the jet orifices with a microscope to ensure that they
are clean and round.  If necessary,  clear the orifices with  a wire.
Clean the jet and impaction plates in an ultrasonic bath with a detergent
solution.  Rinse first with distilled water, then with acetone.  Re-
inspect  the jet orifices and repeat cleaning If necessary.  Some jets
may be  irregularly  shaped because of poor manufacturing.  Such jets may
be calibrated as they are, or returned to the manufacturer.
B-4.4.2  Grease Impaction Plates
         Apply grease to the impaction plates or foil substrates in the
same manner as for  normal laboratory or field use of the cascade impactor.
APT uses Dow Corning high vacuum silicone grease or equivalent.  If fibrous
substrates are being evaluated, they should not be greased.
     Because the particulate concentration will be counted upstream and
downstream of the impaction plate, 1t 1s not necessary to perform a
gravimetric analysis of the substrate and collected particulate.  However,
a gravimetric analysis is suggested as a check on the particulate mass
balance  of the system.  If a mass balance check is being conducted, it 1s
necessary to record the flow rate and time for the duration of the test.
This will enable the prediction of the total particulate mass entering and
leaving  the system.  The difference can then be compared to the mass
collected on the impaction plate.
B-4.4.3  Assemble Impaction Stage
         Place the  Impaction stage 1n the calibration tube as shown 1n
Figure  B-l.  To provide the proper flow pattern, place an Impaction plate
upstream and downstream of the jet plate.  The regular cascade Impactor
casing  can be used  as the calibration tube if tubular spacers are made to
provide  proper alignment and seals between the components.  Alternatively
a special calibration tube could be constructed.
                                       73

-------
B-4.5   MEASURE STAGE PRESSURE DROP
B-4.5.1  Measure Pressure Drop as a Function of Flow Rate
         This provides data which enable the use of the stage as an orifice
flow meter during the calibration.  It also provides a means for checking
the jet orifice size by comparing the pressure drop/flow rate data against
data for a known jet plate.  This is a useful check both in the laboratory
and in the field.
     Measure the flow rate with a calibrated rotameter.  The rotameter
should be calibrated against a wet test meter every six months or whenever
it is in disagreement with secondary flow measurements (for example, the
pressure drop measurements).
     Measure the pressure drop across the impaction stage with an open end
manometer attached to the upstream side of the impactor casing and downstream
of throttle valve "B" in Figure B-l.
B-4.5.2  Plot Pressure Drop Against Flow Rate
         Pressure drop is conveniently plotted against flow rate on log-log
paper to give a straight line relationship.  A typical impaction plate will
have a pressure drop directly proportional to the square of the flow
rate.  For example, Stage 5 of a U.W. M-III Impactor has a flow resistance
which follows the relationship:

                        AP (cm W.C.) = 0.2 (Qs)2              (B-5)

B-4.6  PARTICLE PENETRATION
B-4.6.1  Select Dilution Air Flow Rate
         Select the dilution air flow rate necessary to dry the aerosol
and also to bring 1t to the desired particle concentration range for
counting.  The general flow rate range is given in Section B-2 above.
                                      74

-------
B-4.6.2  Start Dilution Flow to the Dryer
         After checking the system to be sure that valves are open and
closed as required to allow the flow to pass through the impactor stage,
start the dilution air flow to the aerosol dryer.  Use the pressure
regulator, throttle valves, and rotameter to adjust the flow rate.
B-4.6.3  Load and Start Atomizer
         Load the atomizer with about 500 ml of PSL suspension diluted
from roughly 0.5 ml of concentrated latex.  Start the atomizer, controlling
the flow rate by adjusting the pressure regulator and the pressure gage.
The atomizer flow resistance should be checked periodically by passing
air from a dry atomizer (no liquid present) through a flow meter to
detect any nozzle plugging.
B-4.6.4  Adjust Flow Rate
         Adjust the upstream sampling throttle "A" so that the pressure
drop across the impactor is proper for the desired impactor flow  rate,
"Q ".  Keep throttle "B" open as much as possible.  Set the outlet dilution
air flow rate to provide more than 7 1/min total flow into the particle
counter.
     When flow or concentration changes are made, approximately five
system volume change time  intervals should elapse before any data are taken
to allow steady state conditions to be reached.  This can take several
minutes for some low flow  rate impactors  (e.g., Brink impactors).
B-4.6.5  Measure Particle  Concentration
               ^~™^"^      _ «^—•—™^^™^^»
         Warm up the circuits of the particle counter for several minutes
as recommended by the manufacturer.  Particle counting should be  done on
the potentiometer setting  as close as possible to the PSL particle diameter.
To check the extent of agglomeration, it  Is helpful to use the next highest
channel also.  In general, record counts  in all  three channels, below,
above, and at the PSL diameter.
                                     75

-------
     The data recorded should Include the following:
     a.   Stage Identification
     b.   Particle Identification parameters
     c.   Particle suspension (as used) specifications
     d.   Barometric pressure
     e.   All rotameter readings
     f.   All pressure readings
     g.   Air temperature at impactor inlet
     h.   Particle counts on 2 or 3 channels for inlet and outlet
          of impactor.
     Make counts over the complete range of air flow rate going both up
and down.  Plot the data obtained in the simplest meaningful form so
they can be checked for consistency.  Computation methods, as discussed
in Section B-5, are used to obtain particle penetration.  A plot of
penetration versus impactor pressure drop requires the least computation
and serves the purpose.  Inspect the plot for scatter of data and compare
with the anticipated curve.  If it 1s unsatisfactory, make any worthwhile
modifications and repeat the run.
     At this point it is advisable to inspect the impaction plate by eye
and with a microscope.  A visual examination can show whether the plate
is overloaded.  Microscopic examination of light deposits enables the
detection of spurious particles.
     Continue taking and plotting data until satisfied that reproducible
data have been obtained.
                                     76

-------
                                   SECTION B-5
                                     RESULTS


B-5.1  DATA REDUCTION
B-5.1.1  Compute Inlet and Outlet Particle Concentrations
         Compute inlet and outlet particle concentrations by subtracting
the concentration measured on the high diameter channel from that
measured on the lower channel.  For example, if counting 0.5 ym diameter
particles, subtract the concentration counted on the 1.0 ym channel from
that counted on the 0.3 ym channel (see sample calculation, Section B-8.4),
The net concentration should be equal to the concentration measured at
the potentiometer setting closest to the PSL diameter.  If this is not
the case, spurious counts may be a problem, and the concentration measured
closest to the PSL diameter should be used.  If possible, use the
potentiometer to narrow the band around the PSL diameter until no spurious
counts are detected.
B-5.1.2  Adjust Outlet Concentrations
         Adjust outlet concentrations to the same (undiluted) basis as
the inlet concentration in cases where dilution of the outlet sample has
been necessary.
B-5.1.3  Compute Particle Penetration
         Compute particle penetration, Pt, as the ratio of outlet to Inlet
particle concentrations (undiluted basis).   Particle collection efficiency
•jn fractional form is (1-Pt), and in percentage It is 100 times the
fractional efficiency.
                                      77

-------
 B-5.1.4  Obtain  the  Impactor Gas Flow Rate
          Obtain  the  Impactor gas flow rate from the previously determined
 plot of pressure drop  versus flow rate, using the measured pressure drop
 for each penetration data pair.
 B-5.1.5  Compute Inertial Impaction Parameter
          Compute inertial impaction parameter from the measured air flow
 rate and properties, particle properties, jet hole size and number of holes
 by  means of  equation B-l.  Jet velocity may be computed from the hole
 diameter, number of  holes, and air flow rate.
 B-5.1.6  Results
          Results can be plotted in several ways, depending on their desired
                                                          i
 use.   For checking the data quickly during the calibration procedure,
 a plot of penetration  versus impactor pressure drop is convenient.  A plot
 of  efficiency versus impaction parameter is more useful for interpretative
 purposes.
 B-5.2   TYPICAL DATA
 B-5.2.1   Particle Collection Efficiency Data Points
          Particle collection efficiency data points for a single calibration
 run  on State 7 from an APT M-l impactor are plotted against impaction
 parameter in Figure B-6.  The curve 1n Figure B-6 represents the composite
 of data  points for several Stage 7 runs.
 B-5.2.2  Data for Stage 5 Plates
         Data for four separate U.W. M-III Stage 5 plates are plotted in
 Figure  B-7.
B-5.2.3  APT M-l, U.W. M-III, and Andersen Calibrations
         Figures B-8 through B-10 show the results of calibrations of
the APT M-l, U.W. M-III, and Andersen non-viable cascade Impactors In terms
of collection efficiency versus impaction parameter for each stage.  The
curves represent the composites of several runs on different plates.  Data
points are omitted for clarity.
                                     78

-------
                                      Composite for
                                      several  runs
        0        0.10       0.20        0.30       0.40      0.50


                   INERTIAL IMPACTION PARAMETER, K
                                                     P

Figure B-6.   Efficiency vs. inertial impaction parameter for A.P.T.
            M-l cascade impactor,  stage 7.
                               79

-------
        100
            0        0.1        0.2         0.3        0.4

                        INERTIAL  IMPACTION PARAMETER, K


Figure B-7.  Efficiency vs.  inertial impaction  parameter,  data from 4
            separate U.W. Mark  III  cascade impactors , stage 5.
0.5
                                 80

-------
                         INERTIAL IMPAC7ION PARAMETER,  K
                                                            P

Figure  B-8.  Efficiency vs.  inertia] impaction parameter for U.W. Mark

            III cascade impactor, stages  4-7.
                                81

-------



                                         - f  **


                                      •Jm-=t- ^Nt-^-feiH^^
                                       ft£


       ** 70 L

         60



        20
        10


   £|g
                                    / - -
                                  -ht-
   -  ^L.^
       |EF*
                                 • •/ '» "i^-—   -.-^—-•- -rr'-r—- -•:










                                                 rrtr

          -,-+. -n—'--vn-f--H-|-
          -.3^5= '^ctTtri*




                                                          T:HHE




                                                           .
.-" T>-



                                                          "fr^tj

                                0.2         0.3        0,4

                                     ,   [ON  PARAMETER,  K

figure  i     Efficiency vs.  ii        i.ipdCuiori  parameter fo^ A.r
            M-l cascade impactor, stages 4-7.

-------
                                           7 S
            >- • f • M i t • i j. 1.1 i 1+4 •>!>>>—• • i.i. • -_i. r-j_i.     : * t^   Trr ! r r T
Figure B-10.
    0.1        C.2        0.3         0.4


       INITIAL JMPACTION PARAMETER,  K



            Inertia!  'mnact^on paramster for Andersen
-viable cascciic  impart.or, staqes 4-7.
                                                                     0.5

-------
                                  SECTION B-6
B-6.1  REPRODUCIBILITY
       The data obtained with the. procecure described above have been
quite reproducible, as illustrated by Figure B-6.  Generally the data
for a single stage fall within a oar.j Width of inertial impactor parameter
values ranging about 0.02 or 0.03.  Thus, the scatter about a cut
parameter is about ±8 percent.  The -corresponding scatter of cut diameter
values would be ±4 percent.  Similarly, c; variation of ±10 percent in
impaction parameter corresponds to a variation of about ±5 percent in
particle cut diameter.
B-6. 2  ACCURACY
       One way to estimate the accuracy of the impactor calibration is to
compare the results with published theory and experimental data.  Figure
B-ll is an efficiency plot which compares the averaged results for all
four stages of each impactor with a few published experimental and
theoretical results.  It can be seen that the overall average cut parameters
are within a spread of about 0.3 (signifying about 8 percent spread of
cut diameters) for everything but the Ranz :ind Wong experimental data.
     Curves "C" and "D" show the effect of jet length to diameter ratio,
s/dc> ranging from 3 for "C" to 10 for "D", from Mercer and Stafford's
experimental data.  The jet length/diameter ratios for the impactor
stages whose calibrations are reported here are presented in Table B-3.
It can be seen that the ratios range between 2.4 and 12.5, with the APT
M-l having the smallest variation.
     Another factor whose effect is related to that of (s/dc) is the shape
of the orifice in the jet plate.  Of the three impactors calibrated, only
                                     84

-------

                           -tgE-jt-A. Ranz  £  Wong, theoretical,


                         «B««          > 1'°
                    £2B±t:G.  U.W. Mark  III
                    •rv-rmrt- *..
                           K.  Andersen  non-viable

                                                 rrmurctmrnttmtttttfl titS
    0
0.1        0.2       0.3        0.4



     INITIAL IMPACTION PARAMETER, K
0.5
                                                                   0.6
Figure B-ll.   Efficiency vs. inertial impaction parameter for comparison.

-------
TABLE B-3.   LENGTH/JET DIAMETER(s/dJ
        FOR CASCADE IMPACTORS

Stage - s/dc
Impactor
APT M-l
U.W. Mark III (New)
Andersen (non-viable)
4
3.1
4.0
4.7
5
2.4
6.2
7.3
6
3.5
9.2
9.8
7
4.6
12.5
9.8
                 86

-------
the APT M-l has converging orifices in the jet plates.   For the impactors
with cylindrical orifices the jet diameter and velocity depend on jet
plate thickness as well as (s/dj.  While detailed discussion of Impactor
                               C
design and theory is beyond the scope of this report, it is important  to
note that such factors do influence the performance of the impactor and
should be recognized when comparing experimental and theoretical results.
B-6.3  STAGE VARIATIONS
       Turning back to Figures B-8, B-9, and B-10, it can be seen that
the cut characteristics are generally "sharp" (i.e., efficiency rises
steeply over a small impaction parameter range).  However, the fourth
stages for both the APT and the UW impactors show a pronounced decrease in
curve slope above the cut point.  This suggests that particle bounce may
occur at higher velocities for these stages.
     The variation of cut parameter values is least for the APT stages
and greatest for the Andersen.  This is believed to be due in part to
the close control of (s/dc) and the use of converging orifices  in the APT
M-l.  It was also noted that some of the jet holes 1n the Andersen plates
tested were not round, but roughly triangular.  Such non-uniformity of
the jet holes could be responsible for variations in cut parameter between
impactor stages.
B-6.4  CONCLUSIONS
       The procedure outlined in this report provides a simple technique
by which inertial impaction devices may be calibrated.  Such factors as
loading, particle bounce, wall losses, electrostatic and condensation
effects, which may occur during source testing, are evaluated by this
technique to some degree.  The calibration method, therefore, provides a
reliable, intrinsic efficiency and is applicable to laboratory and field
data obtained by careful C.I. operation.
                                     87

-------
                                   SECTION B-7

                                   REFERENCES
May, K.R.  The ColHson Nebulizer:   Description,  Performance  and
Application.  Aerosol Science, Vol.  4.  P.  235-243.   1973.

Calvert, S., Goldshmid, J., Lelth,  D.,  and Mehta, D.   Scrubber  Handbook*
EPA-R2-72-118a (NTIS No. PB 213016), P. 4-148.   1972.

Ranz, W.E., Wong, J.B.  Impact!on of Dust  and Smoke Particles.   In-
dustrial Engineering Chemistry.  P.44,  1371.   1952.

Stern, S.C., Zeller, H.W., and Schekman, A.I.  Collection  Efficiency  of
Jet Impactors at Reduced Pressures.   Industrial  and Engineering Fundamentals.
Vol. 1, No. 4.  P. 273.  1962.

Mercer, T.T., Stafford, R.G.  Impactlon from  Round  Jets.   Ann.   Occupational
Hygiene.  Vol. 12.  P. 41-48.   1969.

Willeke, K. and McFeters, J.J., The Influence of Flow  Entry and Collecting
Surface on the Impaction Efficiency of  Inertia!  Impactors, J. Colloid
and Interface Sci., 53, 121 (1975).

Raabe, O.G., Generation and Characterization  of Aerosols,  from  Inhalation
Carcinogenesis, Proc. of the Biology Division,  Oak  Ridge Nat. Laboratory
Conf., Gatlinburg, Tennessee,  October 8-11, 1969.   (reference from
personal communication from J. McCain,  Southern Research Institute).

-------
                                   SECTION B-8
                              SAMPLE  CALCULATIONS
B-8.1  AERODYNAMIC DIAMETER
       PSL Diameter - 0.5 ym
       Particle density, p  = 1.05 g/cm3
       Cunningham slip correction factor  corresponding to a diameter
          of 0.5 ym, 1s 1.33
                    /    \          /           \
            DAI = D (C pp )   = 0.5  ((1.33)(1.05))  = 0.59 ymA

     (The use of Figure B-4 yields a value  of about 0.58 ymA)

B-8. 2  GAS VELOCITY THROUGH JET

       Number of jets, X, - 110
       Diameter of jet, Dj  = 0.0343 cm
       Sample flow rate at conditions of operation, Qs  3 20  /m1n
         V. = -	4(20)0000).	3.28 x  103  cm/sec
          J   XJ VJ    (110)(iO(0.0343)Z(60)
                                     89

-------
B-8.3  INERTIAL IMPACTION PARAMETER

       Gas velocity at conditions  of jet, V, = 3.28 x lo3 cm/sec
       Viscosity of Gas at conditions  of jet, y = 1.8 x lo"4 poise
       Jet diameter, D. » 0.0343 cm
                      J
       Aerodynamic diameter =0.58 ymA


         K  =    °AI Vj    =  (0.58)2(3.28 x 1Q3)(1Q"8) = Q 2Q
          p      9 M Dj        9(1.8 x 10"4)(0.0343)

B-8.4  CONCENTRATION MEASUREMENT

       Consider PSL particle diameter  = 0.5 ym
       Impactor Inlet
          Concentration on 1.0 ym  channel = 0.01 x 1Q6 cm"3
          Concentration on 0.3 ym  channel = 1.01 x io6 cm"3
          Net concentration measured = (1.01 - 0.01) x io6 cm"3
                                    = 1.00 x io6 cm"3
          Concentration at 0.5 ym  potentiometer setting = 1.00 x io6 cm"3
          Sample flow rate before  dilution = 3.5 1/min
          Sample flow rate after dilution = 7.0 l/min
          Actual concentration entering impactor,


         v   3.1 iff!!  i1-00 x  lfl6  ra"3' * 2-°°x 106  cm~3
                                    90

-------
     Impactor Outlet
        Concentration on 1.0 ym channel  •  0.001 x  106  cm"3
        Concentration on 0.3 ym channel  =  0.50 x 106 cm"3
        Net concentration measured =  0.50  x  106 cm"3
        Concentration at 0.5 ym potentiometer setting =  0.50 x 10 cm"3
        Sample flow rate before dilution =3.5  i/min
        Sample flow rate after dilution  •   7.0  1/min
        Actual concentration exiting  impactor,
         n  =  7.0 1/min (0>5Q x 1Q6 cn)-3)  = I.QO x io6  cm"3
          o    3.5 l/mm
B-8.5  EFFICIENCY AND PENETRATION
       Inlet concentration, n^ = 2 x io6 particles/cm3
       Outlet concentration, nQ = 1 x 106 particles/cm3
              ni      2 x io6
        Pt = _. = — 1 x 1Q6  = 0.5, fraction
         n = 1 - Pt = 0.5

        Percent efficiency « 0.5 x 100 = 50 percent
                                     91

-------
B-8.6  CUT DIAMETER FOR A GIVEN STAGE
       Inertial impaction parameter at 50  percent efficiency, K_  a 0.20
Jet gas velocity, V. = 3.28 x io3  cm/sec
Viscosity of gas at conditions of  jet,  y  =  1.8 x
Jet diameter, D. = 0.0343 cm
                                                            poise
0
 AI,50
"
(0.

20K9H1
(3.28
.8 x K
x 103)
3~**)(
(iQ-

8)
0343)

*"

                                                      = 0.58 yMA
                                     92

-------
-co
                                        on  1?ot~r
                                                                                                                                O/A*i£Te#  x
                                                                                                                            TU Be , 3/L  S.S. ,  .is-*)
                                                                                                                                  ttjTvmat  suiffAc.cz  TO
                                Fouf.  Pt-*<*S  lo*
                                                                                                             iff  AI-SO
                                                                                                                   -C-2.
                                                                                                                                SOUTHEW  USiAfiCH  DBTITOTE
                                                                                                                                  IIIMIMOHAM. AIA»AMA

-------
fitnsfi MWHC-

 POLISH  UTXXIOK. Jcj*(***e J  TO t f*itecf*c.n K* i
               SOUTHEM KSUIQI MSnTUIl
                 tKUlUNOHAM. AUMAMA 35IM
                    J-  "2-77-3-C-Z.

-------
     >••«.
V.
2.
- 3l(o Sro't..
             «-l
              o/^
           J4 secTio(0
                                                   r
                                                      /.^-
                                                               •"
                                                             iL
                                                             4-1—i-.Joo
                                                       >v» uuri.fr
                                95

-------
                                   APPENDIX D

                          COST ESTIMATING DATA SHEET
 Quote  No.

 Job  No.
    Client

     Date
 Location
 Purpose  of Test 	

 Type  of  Test  — Particulate

           Gases
            Particulate Sizing
Auxiliary Analysis 	
PERSONNEL
       COST ESTIMATE-PROFESSIONAL SERVICES
Chemist/Engineers          Project Director
Meteorologist	Superv. Engr.	
$         /Hr\             $         /W.
                           Mrs.
              Cost
Mrs.
Cost    Total Cost
1.0  Pre-Survey
1.1  Travel Time
     No. of Men
2.0  Consulting
3.0  Preparation of
     Test Protocol
4.0  Construction
     Special Test Equip.
5.0  Project Planning
6.0  Field Test
     No. of Men
7.0  Lab Analysis
8.0  Calculations &
     Report Writing
                                       96

-------
Appendix D (continued)
Page 2

PERSONNEL                     COST ESTIMATING-PROFESSIONAL SERVICES
                            Chemist/EngineersProject Director
                            Meteorologist	Superv. Engr.
                            $         /Hr"             $         /Hr:

                              Mrs.     Cost      Mrs.    Cost    Total Cost
9.0    Consulting &
       Design

10.0	

11.0   Total
          Out-of-Pocket Cost
          (see next page)
          TOTAL COST
                                      97

-------
Appendix D (.continued)
Page 3
Quote No.
Job No.
Client
   Date
Location
                     COST ESTIMATING OUT-OF-POCKET EXPENSES
1.
?.
••>
,-» ,
4.
5.
0.
7.
3.
9.
10.
11.
Air Fare Round Trips
Air Fare Round Trips
Air Fare Round Trips
Vehicle Rentals
Shipping (Air Freight, Excess
Per Diem men x $
Per iJiern men x $
Per Diem men x $
Per Diem men x $
Limousine Service, Parking
Miscellaneous
uOiiiputer Time
Car Expenses it/mile x
"roller Expenses  Equipment
General Supplies
@ $ x men
@ $ x men
@ $ x men

Baqqaqe)
/day x days
/day x days
/day x days
/day x days



miles + tolls
miles + tolls


No Charge
                                       TOTAL:
                                      98

-------
                                 APPENDIX E
                  PRELIMINARY SURVEY FOR PARTICIPATE SIZING
PLANT DATA
Company Name:
     Address:
Name of Contacts:
Telephone Number: 	
Process Description:
  Date:
City:
State:
      Title:
      Title:
      Title:
(Operating Schedule):
(Batch or Continuous);
(Rates/Variability);
AIR POLLUTION CONTROL EQUIPMENT
Description:   	
(Operating Schedule):
(Rate/Variability);
                                     99

-------
Appendix E (continued)
Page 2
Preliminary Survey for Particulate Sizing
Sketch of Sampling Sites (with approximate dimensions, ports located,
  upstream and downstream equipment if important).
                                   100

-------
     Appendix E (continued)
     Page 3
     Preliminary Survey for Particulate Sizing
                                          CONDITIONS AT SAMPLING SITES
      Pressure
     Temperature
      Gas  Rate
     Gas Composition
     Particulate Loading
o    Pre-cutter Required?
     Approx. Size Dist.
     Weight Gain/Loss
     by Substrates, Filter
     Weight Gain/Loss
     by Grease
     Particulate Condition-
     hard, sticky, etc.
     Wet or Dry
     Port Size/Fitting Type
     Condensation?
     Notes

-------
 Appendix E (continued)
 Page 4
 Preliminary Survey for Particulate Stzing
 1)    Electricity Source
      a.    Amperage per circuit
      b.    Location of fuse  box
      c.    Extension cord  lengths	 Quantity
      d.    Adapters  Needed	
      e.    Electrician
 2)    Safety Equipment  Needed
      a.    Hard  hats    	   d.   Safety shoes
      b.    Safety  glasses	   e.   Alarms
      c.    Goggles       	   f.   Other
 3)    Ice
      a.    Vendor	
      b.    Location	
 4)    Solvents
      a.    Vendor
     b,   Location 	
5)   Sampling Ports
     a.   Who will provide	Welder:
     b.   Size opening 	
6)   Scaffolding
a.
b.
c.
Height
Length
Vendor
                    Address
                    Telephone
7)   Distilled Water
     a.   Vendor
     b.   Location
                                     102

-------
Appendix E (continued)
Page 5
Preliminary Survey for  Participate Survey
8)   Test Site Facilities
     a.   Parking  	
     b.   Restroom
     c.   Laboratory Facilities	
     d.   Clean-up Area	;	.
9)   Motels:
     a.	 Phone	Rate
     b.	 Phone	Rate
     c.	Phone     .	Rate
10)  Restaurants:
     a.   New Plant           	                         	
     b.   Near Motel
11)  Airport Convenient to Plant	Distance
12)  Comments: 	
                                     Survey By:
                                    103

-------
                                  APPENDIX F

                               SAFETY CHECKLIST
                                        Date
Company 	
Location
     A.    MEDICAL:
          1)    Plant first aid  available  (yes/no)           .
               If available give location  of unit  and  telephone  number
          2)    Phone  number  for  ambulance

          3)    Phone  number  for  hospital

          4)    Comments:    	
     B.    TEST  SITE  CHECKLIST:   Check  if OK.

          1)    Ladders:
               General conditions	, rest  stops 	,  cage
               Comments:
          2)   Scaffolds/Platforms:
              General conditions	, guardrails
              toeboards	, "screening 	
              Comments:
    C.   PERSONNEL PROTECTION EQUIPMENT:  Check if needed.
         1)   Safety glasses 	_, side shields	.
              face shields 	, goggles            . hard hat	,
              safety shoes	electrical  hazard shoes 	,
              life belt and safety block
              hearing protective devices      , ladder climbing devices
          )   Respiratory equipment:
              Air purifying	, air supplied	,
              self-contained	,
              Other
                                     104

-------
Appendix F (continued)
Page 2
Safety Checklist
          3)   Body protection:
               Chemical protective garments
               Heat protective garments	
               Chemical gloves
               Heat resistant gloves
               Other
     D.   ARE FIRE EXTINGUISHERS AVAILABLE AT SITE	       '

     E.   SPECIAL OR UNUSUAL TEST PROCEDURES AND SAFETY PRECAUTIONS
          NECESSARY:
                                     105

-------
                                  APPENDIX G
                              SAMPLE CALCULATION

     This sample calculation is based on data taken with a Brink, impactor ^
The impactor was run in a wind tunnel with redispersed fly ash in an
essentially ambient air stream.  No SOg was present.  The data are
presented in Table 6-1.
     As stated in the section on reporting, it is necessary to determine
an.i report, cumulative distributions based on DA, DS> and D^j.  The
calculation based on DS will be presented in this appendix, with comments
on DA and OAI when appropriate.  Stage 3 will be completely worked
t1* round.
ST'IP I:  CALCULATE THE  PRESSURE ON  STAGE  3
     Thfe fraction of the pressure drop which occurs by stage 3 of the
impactor 1114,5. f, be measured for a given impactor.  Usually most of the
dmp 1s across the last stage or so.  For the Brink impactor, it has
bs;?n Jatennlned that only 1.4 percent of the total Impactor pressure
drop occurs ahead of stage 3.  The total pressure drop across the impactor
ivas measured at 17.8 m He,
     f,jp - fraction of Apj occurring ahead of stage  [0.014]
     A.PJ = total pressure drop across the impactor, mm Hg [17.8 mm Hg]
     Pc  - stack pressure, mm Hg [764.8 mm Hg]

               P. = 764.8 mm Hg - 0.014 (17.8 mm Hg)

                   P. = 764.6 mm Hg at stage 3.
                                     106

-------
STEP 2:  CALCULATE THE GAS MEAN FREE PATH AT STAGE  3 CONDITIONS
     The equation for mean free  path presented below ts based on the
Chapman-Enskog equation.
                           8CTO   f   I 	_1
                          .

where:
     T. = temperature at stage, °K [300°K]
     MW = molecular weight of gas stream [29
     P. = pressure on stage, mm Hg [764.8 mm Hg]
      n = viscosity at stage, poise [1.82 x 10"4  poise]
        = mean free path,  cm
                              = 6.6 x 10   cm
STEP 3:  ITERATIVE SOLUTION OF EQUATION (G-3) AND EQUATION  (Q-4)


                      /2.05 v D? P,.xV/2                          ,r ^
              n     - [ _   3  J J\                             (G-3)
              Ds,50   1   CPpQsPs   '
Where:  Df Kn  = Stokes diameter cut-point of stage, cm
         5,OU
           D.  = jet diameter on stage, cm [0.1396 cm]
           x.  = number of jets on stage [1]
           P   = density of particles, g/cm3 [2.5 g/cm3]
           Q   = volumetric flow rate through Impactor at stack
                  conditions, cmVsec [20.3 cm3/sec]
                                     107

-------
                   »                        -0.44 Dc
       C = i +     *    [ (1.23 + 0.41 exp  	p- )]           (6-4)
                 s,50
where:

   C = Cunningham Correction Factor, dimensionless.
                   K3j- x 10	((1.23 + 0.41 exp (-6.67 x 104 Dc ,n)J
                      us,50    v                               S'b0/
     For DS 5Q in the range of 1 ym (10   cm), the exponential term is
<0.001 and can be neglected.
                                   Ds, 50
     From experience, choose DS 5Q = 1.5 x 10"4 cm for the initial
trial:

     Trial 1:  Ds>50 = 1.5 x lo"4cm

              C1  = 1.108 •*• Ds 5Q|  = 1.33 x 10"4 cm
     First trial  DS 5Q was too high.

     Trial 2:   Dc ,n = 1.2 x 10"4 cm
        1        S ,OU
               C2 * 1.134 -  DSj5Q|2 * 1.31  x 10~4 cm
     Second trial was too low.
                                    108

-------
     Trial 3;   D$ 5Q =• 1.32 x 10~4 cm
          C3 = 1.122   Ds 5Q|3 * 1.32 x 10"4 cm
     AGREEMENT:  Dg 5Q on stage 3 = 1.32 x lo"4cm = 1.32 ym

     This series of calculations must be repeated for each stage of the
impactor.  The results are presented in Table Grl,
STEP 4:  CALCULATION OF DA 5Q FOR IMPACTOR
     DA,50 is calculated ^ the same way DS 5Q was, except that the
density Is taken as 1.0 g/cm3.
STEP 5:  CALCULATION OF Dftl 5Q FOR IMPACTOR
     In theory, D^j ^ is calculated In the same way as is D  5Q,
except that the density is taken as 1.0 g/cm3 and the Cunningham Correction
Factor is set at 1.0.  It can also be calculated as:
                  DAI,50 ' Ds,50
STEP 6:  CALCULATE'THE STANDARD VOLUME OF THE SAMPLE*
     The known flow rate, Qs, 1s volumetric flow rate
It must be converted to standard conditions of 273°K and 760 mm Hg.
The known flow rate, Qs, 1s volumetric flow rate at stack conditions.
*This step 1s not necessary for a cumulative distribution, but will be
illustrated.  In steps 7 and 8, the mass actually caught on the stage,
M., could be used rather than the AC- which is Illustrated.  The ACj term
ii usually used in place of mass for differential distributions.  The
calculation 1s valid as shown.
                                     109

-------
                                            -    9                        ts-6)
where:
          V,. = impactor sample volume at standard conditions, Ncm3

          Qs = sample rate at stack conditions, cm3/sec [20.3 cm3/sec]

        FH 0 = volume fraction of water in gas [0.019]

           e = sample time, sec [1800 sec]


          VN = 3.28 x 104Ncm3 = 0.0328Nm3


STEP 7:  CALCULATE THE CONCENTRATION OF PARTICULATE ON A STAGE


                               AM,                                        (G_7)
Y^ (0.001)
 VN
where;
          ACp = concentration of particulate which impacted on the
                stage, g/Nm3

          AMp = mass of particles caught on the stage, mg [0.576 mg]
              AC3  = 0.0175 g/Nm3 on stage 3.


STEP 8:  CALCULATION OF CUMULATIVE CONCENTRATION FRACTION SMALLER THAN
         STAGE D5Q
Weight percent smaller than (D5Q)fe = - ^ - x 100%  (G-8)
                                                         ACi
where:                                            £^t      '
                                                  i=0
     i=o corresponds to the filter
     i = k corresponds to the stage under study
     i - K corresponds to the coarsest jet or cyclone.
                                      110

-------
     Note that the stages are numbered differently for use In this
equation than they have been previously; one must count from the bottom
up.  However the numbering is done, the mass percent caught on the
final filter is the mass percent smaller than the DSQ of the smallest
stage.  The sum of the mass percents caught on the smallest stage and on
the final filter is the cumulative mass percent smaller than the D^Q
of the next to smallest stage, and so on up the impactor.

   „ . . .        .     ,,      0.0085+0.0002+0.0017+0.0090 0.0175  -,QQ%
   Weight percent smaller - 	1.86144	
   than  D50 of Stage 3
     Percent less than D5Q of stage 3, 1.32 wm = 1.98%


STEP 9:   PLOT DATA ON CUMULATIVE MASS PLOT

          Percent less than stage Dc Cft versus Dc cn.
                                   s,bU         s,ou
STEP 10:  CALCULATE THE CUMULATIVE DISTRIBUTIONS FOR Dfl en AND FOR
          DAI,50 AND PLOT
                                       111

-------
   TABLE G-l.  DATA FROM BRINK IMPACTOR RUN

Stage
Arbitrary
Maximum
Cyclone
0
1
2
3
4
5
6
Filter
Totals
Other Data:
M, mg ACn g/Nm3 % Smaller
n specified
100%
47.840 1.454 21.89%
6.336 0.193 11.52%
3.848 0.117 5.23%
1.992 0.0605 1.98%
0.576 0.0175 1.04%
0.296 0.0090 0.56%
0.056 0.0017 0.47%
0.008 0.00024 0.45%
0.280 0.0085
1.86144
APj = 17.8 mm Hg
Ps = 764.8 mm Hg
fDp for stage 3 = 0.014
y = 0.000182 poise
MW of air = 29 g/g-mole
Tj = 300° K
D. = 3rd stage jet diameter = 0.1396 cm
x. = 1 jet per stage
p = particle density "2.5 g/cm3
Q = sample data at stack conditions = 20
RS = stack temperature = 300° K
FM n = volume fraction water in gas » 0.019
than De Kn» WR
size **w
100
9.39
5.73
3.25 .
1.93
1.32
0.71
0.46
0.28
~
--
.3 cm3/sec
e  = sample time * 1800 sec
                    112

-------
                                  APPENDIX H
                       METRIC SYSTEM CONVERSION  FACTORS

Non-metric                     Multiplied by                Yields Metric
acfm                           28.317                      Hters/min
°F                             5/9 (QF-32)                  °C
in.                             2.54                       cm
gr/acf                          0.0023                     g/liter
ft                             30.48                       cm
gr/acf                          2.3                        mg/liter
                                       113

-------
                                 TECHNICAL REPORT DATA
                          (Please read Instructions on the reverse before completing)
 1. REPORT NO.
 EPA-600/2-77-004
                            2.
                                                       3. RECIPIENT'S ACCESSION-NO.
 4. TITLE AND SUBTITLE
 Procedures for Cascade Impactor Calibration and
  Operation in Process Streams
                             5. REPORT DATE
                             January 1977
                             6. PERFORMING ORGANIZATION COOt
 . AUTHOH(S)

 D. Bruce Harris
                             B. PERFORMING ORGANIZATION REPORT NO.

                             EERL-RTP-236
9. PERFORMING ORGANIZATION NAME AND ADDRESS


See Block 12.
                                                       10. PROGRAM ELEMENT NO.
                             1AB012; ROAP 21ADM-012
                             11. CONTRACT/GRANT NO.

                             NA  (Inhouse report)
 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
                             Final; 12/75-9/76
                             14. SPONSORING AGENCY CODE
                              EPA-ORD
 is.SUPPLEMENTARY NOTES  Author's phone is 919/549-8411 Ext 2557-Mail Drop is 62.  Tenta-
 procedures are covered in EPA-600/2-76-023,  February 1976.
 ie. ABSTRACT The report ]Q an outgrowth of discussions by members of a working group
 of EPA/IERL-RTP personnel,  contractors, and independent experts who met to
 develop uniform procedures for the field use of intertial impactors to determine par-
 ticle size distributions from industrial particulate sources.  It is intended to pro-
 mote individual tests of similar quality so that valid comparisons can be made.
 Procedures for measuring particle size which have  yielded valid data in stationary
 sources are presented based on laboratory and field experience.  Following these
 methods should help the users of cascade impactors to obtain the  information desired.
 The report discusses the preliminary survey, the sampling apparatus, testing
 procedures, data analysis,  calibration procedures,  quality assurance, and reporting
 requirements.  The information applies to cascade impactors  in general.  Specific
 commercial impactors are discussed.
 7.
                             KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                 b.lDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
Air Pollution
 [mpactors
 Dust
 Measurement
Size Determination
Industrial Processes
Calibrating
Sampling
Quality Assu-
 rance
Air Pollution Control
Stationary Sources
Inertial Impactors
Cascade Impactors
Particulate
Fractional Efficiency
13B

11G
14B

13H
14D
 8. DISTRIBUTION STATEMENT
 Unlimited
                                           19. SECURITY CLASS (ThisReport)
                                           Unclassified
                                                                    21. NO. OF PAGES
                                               121
                 20. SECURITY CLASS (Thispage)
                 Unclassified
                                          22. PRICE
EPA Form 2220-1 (9-73)
               114

-------