Você está na página 1de 33

IEST-RP-CC007.

2:
Testing ULPA Filters
IEST Recommended Practices License Agreement for Electronic Documents
You have purchased a non-revisable Adobe Acrobat PDF file that can be stored on your personal computer.
You have purchased a document which is protected internationally by copyright laws and conventions. The
unauthorized reproduction or distribution of this copyrighted work is illegal and may be punishable by criminal law.
Under no circumstances may the electronic file you are licensing be copied, transferred, reproduced in any form, or
placed on a network of any sort without the authorization of the copyright owner. Under no circumstances may it be
resold.
You may print out and retain one-only printed copy of the PDF file. This single-user license permission to print one
copy is valid for each purchased and paid copy. The printed copy is fully protected by US copyright law and
international conventions, and may not be photocopied or reproduced in any form. Under no circumstances may it be
resold.
While all reasonable care is taken in the preparation and review of Recommended Practices, the Institute of
Environmental Sciences and Technology (IEST) does not warrant that the content of the document is accurate or up
to date or that the document will be suitable for your purposes.
To the extent allowed in applicable law, in no event shall IEST be liable for any direct, indirect, punitive, incidental,
special, consequential damages, or any damages whatsoever arising out of or connected with the use or misuse of
this document, with the limit of the refund of the price of the purchased document.
This transaction is governed by and construed in accordance with the laws of Illinois of the United States.

IEST General Copyright Statement


The use of IEST Recommended Practices is subject to the users acceptance of IESTs terms of use and conditions
of copyright for IEST publications, as set out below.
All IEST Publications are protected by copyright. Therefore, and unless otherwise specified, no part of an IEST
Publication may be reproduced or utilized in any form or by any means, electronic or mechanical, including
photocopying, microfilm, and scanning, without permission in writing from the publisher.
While all reasonable care is taken in the preparation and review of Recommended Practices, the Institute of
Environmental Sciences and Technology (IEST) does not warrant that the content of the document is accurate or up
to date or that the document will be suitable for your purposes.
To the extent allowed in applicable law, in no event shall IEST be liable for any direct, indirect, punitive, incidental,
special, consequential damages, or any damages whatsoever arising out of or connected with the use or misuse of
this document, with the limit of the refund of the price of the purchased document.
This transaction is governed by and construed in accordance with the laws of Illinois of the United States.
If you have any difficulties concerning the above terms, any questions regarding copyright, or have a request for
permission, please contact IEST at the following address:

IEST
Arlington Place One
2340 S. Arlington Heights Road, Suite 620
Arlington Heights, IL 60005-4510
Tel (847) 981-0100 Fax (847) 981-4130
E-mail copyright@iest.org
Web www.iest.org

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

INSTITUTE OF
ENVIRONMENTAL
SCIENCES AND
TECHNOLOGY
Contamination Control Division
Recommended Practice 007.2

IEST-RP-CC007.2
Testing ULPA Filters

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY


Arlington Place One
2340 S. Arlington Heights Road, Suite 620
Arlington Heights, IL 60005-4510
Phone: (847) 981-0100 Fax: (847) 981-4130
E-mail: iest@iest.org Web: www.iest.org

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

This Recommended Practice is published by the INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY to
advance the technical and engineering sciences. Its use is entirely voluntary, and determination of its applicability
and suitability for any particular use is solely the responsibility of the user.
This Recommended Practice was prepared by and is under the jurisdiction of Working Group 007 of the IEST Contamination Control Division.
Copyright 2007 by the INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY
Second printing, February 2009
ISBN 978-0-9787868-4-7

PROPOSAL FOR IMPROVEMENT: The Working Groups of the INSTITUTE OF ENVIRONMENTAL SCIENCES
AND T ECHNOLOGY are continually working on improvements to their Recommended Practices and Reference
Documents. Suggestions from those who use these documents are welcome. If you have a suggestion regarding
this document, please use the online Proposal for Improvement form found on the IEST website at
www.iest.org/proposal/form.html.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY


Arlington Place One
2340 S. Arlington Heights Road, Suite 620
Arlington Heights, IL 60005-4510
Phone: (847) 981-0100 Fax: (847) 981-4130
E-mail: iest@iest.org Web: www.iest.org

Copyrighted material

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

Testing ULPA Filters


IEST-RP-CC007.2
CONTENTS
SECTION
1
2
3
4
5
6
7
8

SCOPE AND LIMITATIONS .............................................................................................................................. 5


REFERENCES ....................................................................................................................................................... 5
TERMS AND DEFINITIONS .............................................................................................................................. 6
TEST SYSTEM...................................................................................................................................................... 7
SYSTEM QUALIFICATION AND CALIBRATION ....................................................................................... 11
TEST PROCEDURE ........................................................................................................................................... 15
DATA REDUCTION ......................................................................................................................................... 16
REPORTING AND MARKING ......................................................................................................................... 18

FIGURE
1

TYPICAL TEST SYSTEM ................................................................................................................................... 9

TABLE
A1
B1

UPPER AND LOWER CONFIDENCE LIMITS FOR UPSTREAM AND DOWNSTREAM COUNTS .... 22
MINIMUM FREQUENCY OF CALIBRATION .............................................................................................. 26

APPENDIX
A
B
C
D
E
F

FACTORS INFLUENCING MEASURED PENETRATION ........................................................................... 19


RECOMMENDED CALIBRATION INTERVALS .......................................................................................... 26
CONSIDERATIONS FOR SPECIFIC MEDIA TYPES.................................................................................... 27
OTHER TEST METHODS FOR ULPA FILTERS ........................................................................................... 28
PARTICLE DETECTION AND PARTICLE SIZE DISTRIBUTION ............................................................ 29
BIBLIOGRAPHY ................................................................................................................................................ 30

IEST-RP-CC007.2

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

Copyrighted material

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY


Contamination Control Division
Recommended Practice 007.2

Testing ULPA Filters


IEST-RP-CC007.2
1
1.1

SCOPE AND LIMITATIONS


Scope

This test procedure covers production testing of filters


for particle penetration and pressure drop of ultralowpenetration air (ULPA) filters. The penetration range
of the procedure is 0.0010% to 0.00010%, using particle counters.
This Recommended Practice (RP) describes the
equipment, aerosol properties, processes, and calculations for determining the efficiency of ULPA filters,
using particle counters. The procedure may be applied to production applications. This RP provides
guidelines for constructing a suitable test duct and
sampling system. Also provided are test criteria for
quantifying penetration in the range of 0.0010% to
0.00010%, using test aerosol particles in the size
range of 0.1 to 0.2 m.

1.2

Limitations

Filters tested per IEST-RP-CC007 are typically factory-tested with uniform airflow across the filter. Ducted
filters, fan filter units (FFUs), and poorly designed
inlet housings for in-line filters may result in nonuniform media air velocity that can possibly reduce the
in-situ filter efficiency.
Application of this RP is by mutual agreement between the customer and the supplier. To apply this RP,
the agreement should also include:
a)

acceptance criteria for penetration and pressure


drop

b) the test aerosol


c)

the test volume flow rate

IEST-RP-CC007.2

Prior to testing filters according to this RP, the most


penetrating particle size (MPPS) should be determined.
The determination can be made for the filter medium
in flat sheet form, provided that the test is conducted
with an aerosol as defined in section 4.2.9.
The test is performed at the same velocity as the average velocity through the medium in the assembled
filter at the test volume flow rate.
CAUTION: Testing in accordance with this RP may
involve hazardous materials, operations, and equipment. This RP does not purport to address the safety
problems associated with its use. It is the responsibility
of the user to consult and establish appropriate safety
and health practices and to determine the applicability
of regulatory limitations prior to use of this RP.

1.3

Application of method

The methodology described in this RP may be applied for particle-counter testing of filters outside the
efficiency and particle size range covered in the document.

REFERENCES

The following documents are incorporated into this RP


to the extent specified herein. Users should apply the
most recent editions of the references.

2.1

American Society of Heating,


Refrigerating, and Air-Conditioning Engineers (ASHRAE)

ASHRAE Standard 52.2-2007: Method of Testing


General Ventilation Air-Cleaning Devices for Removal
Efficiency by Particle Size

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

2.2

American Society of Mechanical


Engineers (ASME)

Fluid Meters: Their Theory and Application (ASME)


ASME-N510: Testing of Nuclear Air-Treatment Systems

2.3

Institute of Environmental
Sciences and Technology (IEST)

IEST-RP-CC001: HEPA and ULPA Filters


IEST-RP-CC014: Calibration and Characterization of
Optical Airborne Particle Counters
IEST-RP-CC021: Testing HEPA and ULPA Filter
Media

2.4 Technical Association of the Pulp


and Paper Industry (TAPPI)
TAPPI-T1205: Dealing with Suspect (Outlying) Test
Determinations

2.5

Sources and Addresses

ASHRAE
American Society of Heating, Refrigerating, and
Air-Conditioning Engineers, Inc.
1791 Tullie Circle, NE
Atlanta, Georgia 30329, USA
Phone: 404-636-8400
Fax: 404-321-5478
www.ashrae.org

ASME International
American Society of Mechanical Engineers
Three Park Avenue
New York, NY 10016-5990, USA
Phone: 800-843-2763
Fax: 973-882-1717
www.asme.org
Institute of Environmental Sciences and Technology
Arlington Place One
2340 S. Arlington Heights Road, Suite 620
Arlington Heights, IL 60005-4510
Phone: 847-981-0100
Fax: 847-981-4130
www.iest.org

TAPPI
Technical Association of the Pulp and Paper Industry
15 Technology Parkway South
Norcross, Georgia 30092
Phone: 770-446-1400
Fax: 770-446-6947
www.tappi.org

Copyrighted material

TERMS AND DEFINITIONS

The following terms have special meaning in the context of this RP.
airflow
Airflow refers to volumetric flow rather than mass
flow. Tests are run at the manufacturer-rated airflow
or other airflow as agreed upon by the customer and
supplier.
correlation ratio
The ratio of downstream counts to upstream counts
with no filter in the test system.
For sequential counting systems, the correlation ratio
represents the differences in the upstream and downstream sampling systems, such as:
particle losses in sample lines and test duct
diluter (if used)
sample times
For simultaneous counting systems, the correlation
ratio represents the same factors as for sequential
counting systems, as well as the differences between
the different particle counting instruments used upstream and downstream, such as:
sample flow rate
counting efficiency
DOP
The mineral oil dioctyl phthalate or DEHP (di-2ethylhexyl phthalate) CAS# 117-81-7. With reference
to filter testing, DOP also refers to a polydisperse
aerosol of the above material.
electret filter media
Media made of fibers that carry electrical charge on
their surfaces.

IEST

HEPA (high-efficiency particulate air) filter


An extended-medium, dry-type filter in a rigid frame
when tested at rated airflow having a minimum particle collection efficiency of 99.97% for 0.3-m mass
median diameter particles of DOP when tested in
accordance with MIL-STD-282.
most penetrating particle size (MPPS)
The particle size at which a given filter has its highest
penetration (or lowest efficiency); i.e., the worst-case
particle size with respect to filtration efficiency.
From filtration theory, efficiency is higher for particle sizes smaller or larger than the MPPS. In practice,
the MPPS of a filter is considered to be within a
measured size range typical in commercial instruments rather than at one unique particle size.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

national metrology institute (NMI)


An organization providing national or international
primary test standards used in the unbroken chain of
traceability of test equipment. Traceability to national
metrology institute standards does not always require
the use of the NMI of the country in which a calibration laboratory is located.
particle count
The number of particles detected in a given volume
of air (or time period for a stable instrument sample
flow rate). The number may be determined by counting discrete particles using a particle counter, or it
may be deduced by calibration of a monotonic response (such as the photometric calibration of a condensation nucleus counter) of the instrument to a
known concentration standard.
particle counter
An instrument capable of resolving responses from
individual particles (e.g., an optical particle counter
or condensation particle counter).
particle size
The apparent maximum linear dimension of a particle
in the plane of observation as observed with an optical
microscope, or the equivalent diameter of a particle
detected by automatic instrumentation. The equivalent
diameter is the diameter of a reference sphere having
known properties and producing the same response in
the sensing instrument as the particle being measured.
penetration
The ratio of the number of particles exiting the filter
to the number of particles entering the filter, per unit
time, expressed as a percentage for a stated particle
size range.
observed penetration
The penetration calculated from observed counts.
true mean penetration
The penetration that would be calculated from
the average observed counts, if the tests were repeated indefinitely.
UCL and LCL penetration
The 95% upper confidence limit (UCL) penetration and 95% lower confidence limit (LCL)
penetration are values statistically calculated
from the observed counts. There is a 95% confidence that the true mean penetration is between
the UCL and LCL penetrations.
polystyrene microspheres
Test aerosol made using solid, monodisperse polystyrene particles of a known size. Also known as PSL
(polystyrene latex) spheres. NOTE: Polystyrene microspheres do not contain natural rubber latex, which

IEST-RP-CC007.2

has been shown to potentially cause allergic reactions


on contact with skin.
sequential counting systems
Test systems in which particles are counted by the
same counter or counters both upstream and downstream of the filter. The counters are sequentially
connected to either side of the filter, once or several
times. Sequential counting systems are characterized
by the need for stable aerosol and by reduced problems in correlating counters.
simultaneous counting systems
Test systems in which particles are counted simultaneously upstream of the filter, with a counter or counters
dedicated to upstream counting, and downstream of
the filter, with a counter or counters dedicated to
downstream counting. Simultaneous counting systems
are characterized by reduced need for stable challenge
aerosol concentration and by difficulty in correlating
counters.
super ULPA (ultralow-penetration air) filter
An extended-medium, dry-type filter in a rigid frame,
made with filter medium having a minimum particle
collection efficiency of >99.9999% (i.e., a maximum
particle penetration of <0.00010%) for particles of
the most penetrating particle size (MPPS) when tested at the average medium velocity in the filter in accordance with the methods of IEST-RP-CC021.
ULPA (ultralow-penetration air) filter
A throwaway, extended-medium, dry-type filter in a
rigid frame, having a minimum particle-collection
efficiency of 99.999% (that is, a maximum particle
penetration of 0.0010%) when tested in accordance
with the methods of IEST-RP-CC007. ULPA filters
are defined and described in IEST-RP-CC001.

4
4.1

TEST SYSTEM
General

The basic components of a typical test system are


shown in Figure 1. The configuration of components
shown provides a test system that operates above ambient pressure. This configuration helps prevent system
leaks from affecting the measurement. If the test system is below ambient pressure, small leaks can introduce sufficient particles to interfere with the
measurements. In Figure 1, item 13 is a restriction or
valve used to regulate the system pressure.
In designing systems, it should be recognized that
large volumes, low velocities, and recirculating regions in the flow path lead to long system response
times (see section 5.16).

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

4.2

4.2.7 Aerosol mixing, upstream and


downstream of the filter under test

Components

4.2.1 Inlet filter

(Figure 1, items 8 and 12)

(Figure 1, item 1)

Prior to mixing with test aerosol, the test air should be


cleaned to a level of less than 1% of the challenge aerosol concentration. Filtration with high-efficiency particulate air (HEPA) filters is recommended.

4.2.2 Relative humidity measurement


and control
(Figure 1, items 3 and 10; section A17)

Provisions should be included to maintain the relative


humidity of the test airflow in the range of 30% to
70%.

Upstream of the filter under test, a method should be


used to provide for the uniform challenge of the filter
(see section 5.13). This may be accomplished with mixing vanes, baffles, multiple aerosol injection points, or
other means.
Downstream of the filter under test, a method should
be used to confirm that the downstream sample represents the flow through the entire filter (see section
5.14). This may be accomplished with mixing vanes,
baffles, multiple sampling points, or other means.

4.2.8 Resistance measurement


(Figure 1, item 11)

4.2.3 Temperature measurement and control


(Figure 1, items 3 and 10; section A16)

Temperature control is not required if the temperature


remains in the range of 10 to 38 C (50 to 100 F).
However, if the temperature is other than 21 C
(70 F), appropriate corrections for the flow measurement device should be used to maintain the actual volumetric test flow. In addition, corrections to the
measured pressure drop of the filter should be made as
per section 7.6. Care should be taken to minimize temperature gradients.

4.2.4 Blower
(Figure 1, item 2)

A blower of sufficient capacity to provide steady airflow at the test conditions should be used.
(Figure 1, item 4; sections A3, A16, A17, A18)

Equipment used for flow measurement and control


should be capable of measuring actual volumetric flow
at test conditions within 3%.

4.2.6 Filter test duct


(Figure 1, items 8-13)

All materials for the filter test system that contact the
test airstream, including sampling lines, should be
cleanable, non-particle-shedding, and electrically conductive and grounded. Two preferred materials are
polished stainless steel and aluminum. The test filter
should be mounted in a chucking device that completely encapsulates the filter and frame, so the filter frame
does not become part of the test duct and leaks in the
filter frame are detected.

Copyrighted material

Pressure taps in the test duct should be of the static


pressure type, at right angles to the airstream. To eliminate the possibility of velocity components in any one
reading, several pressure taps appropriately spaced in
the test duct are preferred to only one. An average
reading can be obtained by plumbing all taps together.

4.2.9 Aerosol material

4.2.5 Flow measurement and control

The filter resistance should be a measure of the static


pressure loss of the airstream as the airstream moves
through the test filter at the test flow rate. The airstream should be measured with a differential pressure
device connected to pressure taps in the test duct, upstream and downstream of the test filter. The differential pressure device should have an accuracy of at least
3% of reading or 2.5 Pa (0.01 in. w.c.), whichever
is larger.

The particles of the aerosol used for testing, calibrating, and obtaining correlation ratios should be of
spherical form and have an index of refraction with a
real part between 1.45 and 1.60 and low imaginary
part. The vapor pressure should be low enough to ensure that diameter changes through the test system are
less than 10%. Possible materials include but are not
limited to:
DOP (DEHP)dioctyl phthalate (di-2ethylhexyl phthalate) (CAS# 117-81-7)
PAOpolyalpha olefin (e.g., 4 centistoke,
CAS# 68649-12-7)
DOS (DEHS)dioctyl sebacate (diethyl hexyl
sebacate) (CAS# 122-62-3)
polystyrene microspheres (PSL)
mineral oil

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Figure 1Typical test system.

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

IEST-RP-CC007.2

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

CAUTION: (1) For any material used, the material


safety data sheet (MSDS) should be available, and the
health risks of using the material in aerosol form
should be assessed. As noted in section 1, testing in
accordance with this RP may involve hazardous materials, operations, and equipment. This RP does not
purport to address the safety problems associated with
its use. It is the responsibility of the user to consult and
establish appropriate safety and health practices and to
determine the applicability of regulatory limitations
prior to use of this RP. (2) Filter testing will leave a
miniscule amount of the challenge aerosol as a residue
that may offgas. While all of the aerosols identified in
this section are adequate for filter testing, determining
the effects of offgassing is outside the scope of this RP
and is a matter for agreement between the user and
supplier. Measurement methods for determining offgassing are given in IEST-RP-CC031.
Use of aerosol materials that do not fit the preceding
description may result in penetration measurements
that are different from measurements made using the
listed aerosols. These restrictions do not apply for other sizing methods.

4.2.10 Aerosol source


Although not shown in Figure 1, use of a method of
stopping or bypassing the aerosol flow from the aerosol source is recommended for the purpose of changing the test filter.

4.2.11 Aerosol generator


(Figure 1, item 5; sections A1, A10A14, A19)

The aerosol population may be either polydisperse or a


narrow size distribution having 90% of the aerosol
between 0.1 and 0.2 m (see section 4.2.16).
The aerosol generator should be capable of producing
a stable aerosol concentration, with the number of particles in the test size range sufficient to allow statistically valid counts downstream of the filter in the
desired time period. Count statistics are discussed in
section A5.
Because the slope of the aerosol size distribution over
the particle detector size channels influences the effective size of the channels, the statistical mode of the
aerosol count distribution should be within the test size
range (0.1 to 0.2 m).
For aerosol generators that have output concentration
capacities greater than 105 particles/cm3 (109 particles/ft3) or generators that use a solvent, it may be necessary to add dilution air at the outlet of the generator
to minimize agglomeration or to evaporate a solvent.
Atomizers with integral or subsequent impactors may
be used to generate suitable aerosols from a variety of

10

Copyrighted material

liquids, solutions, and suspensions. Examples of such


atomizers include: Collison style, Laskin nozzle type,
and those described by Scripsick and Soderholm (see
Appendix F). Thermal methods of evaporation and
condensation similar to those used in MIL-STD-282
may be used if adjustments are made to provide the
proper size particles.

4.2.12 Aerosol neutralizer


(Figure 1, item 6; sections A10 and A14)

Many aerosol generators produce highly charged aerosols as a result of mechanical shearing of the liquid in
the process of making drops. These charges should be
neutralized to minimize particle deposition as a result
of electric fields. Commercially available radioactive
or electrostatic neutralizers produce high numbers of
bipolar ions that effectively neutralize the aerosol.
These neutralizers should be used immediately after
the aerosol is generated and before the aerosol is introduced into the filter test duct.
Electrostatic neutralizers may produce large numbers
of ultrafine particles (<0.05 m) that will contribute to
the challenge aerosol. This is not important for particle
counters that are not sensitive to particles smaller than
0.05 m. Radioactive neutralizers do not produce such
ultrafine particles.
Ionizers of the pulse type, if used, require thorough
mixing with the aerosol for effective neutralization.
Otherwise, ionizers can charge rather than neutralize
the particles and thus be detrimental.

4.2.13 Aerosol source output rate adjuster


(Figure 1, item 7)

a) When testing filters the aerosol challenge upstream of the test filter should be:
1) sufficiently concentrated to complete the test
in a reasonable period of time (see section 6.6);
2) sufficiently concentrated that the background
count rate of the downstream counter does not exceed 5% of the average downstream count;
3) less than 105 particles/cm3 (109 particles/ft3)
to avoid agglomeration.
During calibration of the test system, it is necessary to
have a substantially lower concentration of the same
aerosol to correlate the upstream and downstream
sampling systems and counting system (see section
5.9).
b) The aerosol used for correlation should:
1) be the same material and size distribution as
the test aerosol;

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

2) have a concentration low enough to limit the


error rate resulting from high counting rate in any
particle counter to a maximum of 5%;
3) be sufficiently concentrated to accomplish the
correlation in a reasonable period of time.
Therefore, a method should be provided to reduce the
concentration of aerosol. This reduction can be accomplished by dumping excess aerosol or by filtering
the aerosol out of a substantial portion of the aerosol
source flow using a dilution system.

4.2.14 Sampling, sample lines


(Sections A9 and A15)

The sampling system should be designed to obtain a


representative sample. The sampling tubing should be:
a) electrically conductive (preferably stainless steel,
aluminum, other metal, or other conductive flexible
tubing);
b) electrically grounded;
c)

as short as possible.

Turbulence should be avoided. Use of valves, restrictions, and bends should be minimized.

4.2.15 Diluter
(Figure 1, item 14; section A6)

High upstream concentration can result in particlecounting errors. When the high-count-rate error exceeds 5%, a diluter should be used.
Design and operation of diluters for low flow rates
may present difficulties such as very small capillary
sizes. Hence, for low-flow-rate particle counters, it
may be desirable to sample from the test duct at a
higher flow rate than the particle counter and then subsample for the particle counter, as shown in Figure 1,
item 15.

4.2.16 Particle counters


(Figure 1, items 17, 18, and 20; sections A1,
A6-A8, A11, A19)

The particle-sensing equipment should be one or


more particle counters capable of counting particles
in the size range of 0.1 to 0.2 m. Polydisperse particles may be sized with equipment using light scattering, electrical mobility classification, diffusional
separation, or other method demonstrated to be of
equal or better accuracy.
CAUTION: Particle-sensing instruments should be
capable of responses independent of particle size in the
0.1 m to 0.3 m size range; that is, their count efficiency should be constant in this size range. In general,
it is good practice to select particle counters that are

IEST-RP-CC007.2

capable of detecting and counting particles with at


least half the diameter of the size of interest. For example, counters that can measure particles as small as
0.1 m may be suitable for measurement of filter media performance at 0.2 m. On the other hand, if the
MPPS of the media is 0.2 m or smaller, it is generally
not recommended to use optical particle counters and
polydisperse aerosol, since many commercial optical
particle counters are not suitable for measuring 0.1 m
or smaller particles.
When using a narrow size distribution of particles,
such as polystyrene latex spheres, having 90% of the
aerosol between 0.1 and 0.2 m, the size should be
measured and verified prior to filter testing. It is not
necessary that the particle counters measure the size
during the filter testing. When using particle counters
that have sensitivity to a broad range of particle sizes
but no particle sizing capability (such as a condensation nucleus counter), it is important to verify that the
aerosol generation and neutralization mechanism does
not create additional particles of other sizes that may
be detected by the particle counters.
The equipment should be able to repeatably and accurately measure the concentration both upstream and
downstream of the filter. For high concentrations resulting in greater than 5% error as a result of high
counting rate, a diluter should be used. For low concentrations, the zero count background level of the
counter should be less than 5% of the concentration
being measured.

4.2.17 System control, data acquisition, and


data reduction equipment
To reduce test time and operator errors, computer control of the test data is recommended. Appendix A, sections A6 and A7, discuss reasons to automate the data
acquisition and reduction.

SYSTEM QUALIFICATION
AND CALIBRATION

Qualification tests and calibrations should be performed prior to using the filter test system. In addition,
periodic testing and calibration should be performed to
provide continued accurate results. Table B1 summarizes the requirements. Certain tests or calibrations
should precede others. The tests and calibrations described in sections 5.1 through 5.4 may be conducted
in any order; the remaining procedures should be performed in the order in which they are listed.
Qualification tests and calibrations should be performed prior to initial startup and prior to startup after
any system modifications.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

11

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

5.1

NMI and have an accuracy of 5% relative humidity


or better.

Flow

a) Calibration of the measurement equipment required to obtain the volume flow rate should be traceable to a national metrology institute (NMI), primary
standards, or recognized equivalent. If a nozzle as described in ASHRAE-52 is used, the following devices
should be calibrated:
1) linear measuring devices for nozzle and duct
diameter,
2) the differential pressure measuring device,
3) absolute pressure measuring devices (such as
a barometer and a differential pressure measuring
device),
4) the temperature measuring device.
b) The standard for duct lengths, pressure tap location, and so forth should also be followed.
Possible standards to follow include, but are not limited to:
1) Fluid Meters: Their Theory and Application
(ASME)
2) ASHRAE 52.2-2007, Section 4.5 Device
Flow Measurement
3) Orifices calibrated with secondary standard
orifice plates that are themselves calibrated as described by Fain and Selby (see Appendix F).
The flow measurement device should be calibrated prior
to system use and a minimum of once a year thereafter.
The calibration should be accurate within 3%.

5.2

Filter pressure drop

The differential pressure measurement device used to


measure the filter pressure drop should be calibrated
prior to system startup and at least yearly thereafter.
The calibration should be traceable to an NMI with an
accuracy of 3% of reading or 2.5 Pa (0.01 in. w.c.),
whichever is larger.

5.3

Temperature

The flow temperature measurement device should be


calibrated prior to system startup and a minimum of
once a year thereafter. The calibration should be traceable to an NMI and have an accuracy of 1.7 C
(3 F) or better.

5.4

Relative humidity

The relative humidity measurement device should be


calibrated prior to system startup and at least yearly
thereafter. The calibration should be traceable to an

12

Copyrighted material

5.5

Particle counters

Particle light-scattering spectrometers (if used) should


be calibrated using the procedures defined by IESTRP-CC-014. Other sizing equipment calibration should
be NMI traceable. All particle counters and particle
sizing equipment should be calibrated prior to system
startup and a minimum of once a year thereafter.
The particle counter background count rate should be
determined daily by using appropriate high-efficiency
(e.g., HEPA or ULPA) filters on the counter inlet.
There are no NMI-traceable particle-concentration
calibration methods. The correlation methods used in
this RP eliminate the need for such calibration in cases
where the counting efficiency is not dependent on concentration, generally true for low concentrations. It is
necessary to know the upper limit of the count rate for
the counters being used. The aerosol concentrations in
the test system should be kept sufficiently low so the
counting error resulting from a high count rate does
not exceed 5% (see section 5.6).
When aerosols other than those defined by section
4.2.9 are used, a procedure for particle counter response correction should be included, as agreed by the
customer and the supplier.

5.6

Particle counter
high-count-rate error

Particle counter high-count-rate errors cause the test


method outlined in this RP to be conservative, so outof-specification filters will not test as though they are
within specification. That is, filters that truly do not
meet the minimum specified efficiency will be rejected. However, filters that would otherwise meet the
minimum specified efficiency may also be rejected;
therefore it is desirable to minimize this source of error. Such a source of error can be minimized by setting
the aerosol source output rate and the aerosol dilution
ratio to produce a count rate below the count-rate limit
of the particle counter.
One test is effective for counters having a cumulative mode (in which the channels have only lower
size limits): each channel should have more counts
than the next (larger) channel. This test checks all
channels except the smallest, which can then be tested
for a count reasonably agreeing with the size distribution suggested by the larger channels.
A second test is to uniformly reduce particle concentration (by a known or unknown factor) entering the
counter and observe the shape of the distribution. If the

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

shape of the distribution changes by shifting toward


smaller sizes, the count was too high. Other tests include (1) verifying that counts vary in proportion to the
sample flow rate, (2) careful dilution, or (3) correctly
measuring a known filtration efficiency.
Simply observing that stated counts are below manufacturer-specified limits is insufficient, as counters
typically err by wrongly stating counts within the limits, while true counts would be outside the limits.

5.7

Aerosol neutralization

Aerosol neutralization systems should be used according to manufacturer instructions. Operation should be
verified as directed in manufacturer instructions at system startup and a minimum of once a year thereafter.
Radioactive aerosol neutralizers should be replaced after
one half-life has passed since the time of manufacture.

5.8

Aerosol source

The aerosol source characteristics that should be verified on startup include the particle size distribution,
output rate, or concentration and stability.
When polydisperse aerosols are used with particlesize-detecting counters, the statistical mode of the
count distribution should be in the size range of 0.1 to
0.2 m. When monodisperse aerosols are used, it
should be shown that other potentially interfering particles do not interfere with the system operation.
The aerosol generator output rate should be compared
to the particle-counter saturation limits, as described in
section 5.6, for the following two conditions:
a) When determining the correlation ratio (with no
test filter in place) with the aerosol source adjusted for
low output, it should be determined that the downstream count rate is sufficiently low to keep the highcount-rate error less than 5%.
b) When testing a filter with the aerosol source in its
high output, it should be determined that the upstream
count rate including the diluter, if necessary, is low
enough to keep the high-count-rate error less than 5%.
These tests should be conducted at the minimum test
flow rate to be used.
The aerosol source stability should be demonstrated.
This is particularly important for a sequential counting
system where the challenge aerosol concentration
should not vary by more than 5% between subsequent
upstream-downstream cycles.
The aerosol source characteristics should be verified at
least yearly, or each time a new aerosol is used.

IEST-RP-CC007.2

5.9

Dilution

Dilution systems, if employed, should have demonstrated stability and accuracy. As a minimum, the dilution ratio at the size range used for penetration
measurement should be measured at the sampling
flows to be used during testing and with the challenge
aerosol to be used. The aerosol concentration should
allow measurement upstream and downstream of the
diluter with the same instrument. The diluter should be
capable of reducing the upstream challenge concentration to an error rate (resulting from high concentration)
below 5% for the particle counter being used.
Several measurements of diluter ratio should be made.
Averages and standard deviations of these measured
values should be calculated to determine the confidence level in the final penetration calculation.
The actual dilution ratio should be measured by comparing particle concentrations upstream and downstream of the diluter, measured with the same sampling
system and particle counter. By using the correlation
ratio described in section 5.17, the measured dilution
ratio is automatically included as part of the correlation
ratio.
Dilution system measurements as described in this
section should be performed at startup, after modification, and yearly thereafter.

5.10 Duct leak test


The leak integrity of the test duct should be verified by
testing in accordance with section 6 of ASME-N510, at
system startup and after modification. Leaks should not
exceed 0.06 m3/min per square meter of duct at 2500 Pa
(0.2 ft3/min per square foot of duct at 10 in. w.c.).

5.11 Cleanliness of test air


The cleanliness of the test air with the aerosol source
turned off should be checked prior to system startup
and daily thereafter. The background concentration
should be less than 1% of the challenge aerosol concentration. (If the background concentration is
checked with an ULPA filter of known efficiency in
the test chuck, then the background count rate of the
downstream counting system can be checked at the
same time).

5.12 Influent airflow distribution


The airflow distribution should be measured upstream
of the test filter, with the filter in place. The airflow distribution should be verified for the largest and the smallest flow capacities of the system and for the largest and
smallest filters to be tested. The minimum number of
velocity measurements should be 10, equally spaced

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

13

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

across the entire face, with a reading at the center of the


test filter. All readings should be within 10% of the average. The airflow distribution test should be done at
system startup and after modification.

5.13 Uniformity of challenge aerosol


distribution
The verification of proper mixing of the aerosol with
the test air upstream of the test filter is a prerequisite to
testing. This verification need be made only upon
completion of initial system installation or modification and should be conducted at both minimum and
maximum design airflows.
An air-aerosol mixing uniformity test should be performed on the system in accordance with ASMEN510. DOP or other acceptable aerosols may be used.
Aerosol concentration readings should be taken with
the upstream sampler located at its normal test position
(as determined by test system design), and at eight or
more other test positions equally spaced over the filter
face area and in the same plane as the upstream sampler test position. Maximum and minimum readings
should be within 20% of the average of all readings.

5.14 Leak detection ability (downstream aerosol distribution)


The downstream mixing and sampling system should
be challenged with a simulated leak. A leak may be
simulated by the injection of aerosol through a small
tube just downstream of the filter. The injection points
should be on a 127 mm 127 mm (5 in. 5 in.) grid
covering the entire face of the filter. The injection
points that lie along the edge or at the corners of the
filter should be in contact with the inside wall of the
test duct or filter frame. All injection points should be
in a plane within 1 in. of the downstream face of the
filter. The simulated leak should increase the penetration through an otherwise intact filter by a factor of at
least 10, but the simulation leak should not saturate the
downstream counter. The challenge should be performed with the largest and smallest filter sizes to be
tested.
The downstream mixing and sampling system should
be adequate to detect the leak regardless of the location. The maximum and minimum penetrations measured with the leak should be within 20% of the
average.
This challenge with a simulated leak should be done at
system startup and after modification.

14

Copyrighted material

5.15 Particle losses


Excessive particle losses in the test system make it
difficult to obtain reliable penetration measurements.
When correction factors are calculated and used as
explained in section 7.1, the particle losses are automatically compensated. However, in the system qualification testing, when the system is first constructed or
following system changes, the particle losses in the test
duct and in the sampling systems should be measured.
The particle losses in the test duct between the upstream and downstream sampling locations can be
determined by measuring the aerosol concentration at
both sampling locations with the same sampling system and particle counter, with a blank or open frame
filter in position. The downstream concentration
should not exceed the upstream concentration by more
than 2% and should not be more than 10% less than
the upstream concentration.

5.16 System response times


(Section A8)
A system response time is the time period during
which a measured variable within the test system reequilibrates in response to a change in status of part of
the system as the test progresses. For example, if, during testing, a particle counter sampling valve switches
the counter from challenge aerosol to filtered downstream air, then, for some response time period, the
counter purges itself and counts more particles than is
correct for the filtered air. Other measured variables
within the system are typically characterized by different response times.
Any change in status and subsequent measurement
should include a waiting period longer than the relevant system response time. System startup should include time-resolved studies to demonstrate the system
response times at minimum and maximum flow rates.
While in principle a re-equilibration is never truly
complete, any errors due to incomplete equilibrations
should be kept smaller than other, uncontrollable
sources of instrumental error.
System response times to be determined include:
a) the time for particles introduced into the downstream ductwork during loading of the test filter to be
cleared from the system.
b) the time for the challenge aerosol concentration to
stabilize after being turned on or diverted into the test
duct. Measurement of the system response times
should be made for both the downstream sample system and the upstream sample system.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

c) the time for the challenge aerosol to be cleared


from the test system after being turned off or diverted.

accumulated and correlation ratios determined for other size ranges also.

d) for sequential systems, the time for the counting


system to stabilize after switching from one sample
port to the other in both directions.

5.18 Pressure drop of filter mount

e) the time for the airflow rate to stabilize after the


test filter is changed.
f) the time for the system air temperature and relative humidity to stabilize after system startup.

5.17 Correlation ratio


If separate upstream and downstream counters are
used, the two counting systems should be correlated
with one another. When a single particle counter is
used sequentially to sample upstream and downstream,
it is necessary to correlate the sampling systems. The
daily correlation ratio check should be stable and repeatable within 10% of the correlation ratio calibration
when sampling the same aerosol.

5.17.1 Sequential counting systems


With a blank or open frame filter in position, the upstream and downstream counters should measure the
particle counts in air that passes over both sampling
systems, in series. The test aerosol should be introduced at a sufficiently low concentration so that the
error rate as a result of high counting rate does not
exceed 5% when sampling either upstream or downstream. Counts should be accumulated by sequentially
sampling upstream and then downstream, with appropriate purge times between counts. The upstreamdownstream cycle should be performed at least three
times, for a total of three upstream and three downstream samples.
These data are used to calculate the correlation ratio as
defined in section 7.1 for particle sizes of 0.1 to 0.2
m. Counts may be accumulated and correlation ratios
calculated for other size ranges also. If unequal upstream and downstream sampling times are used in
correlation or testing, the times should be included in
correlation.

5.17.2 Simultaneous counting systems


With a blank or an open frame filter in position, both
upstream and downstream particle counters and diluters, if used, should sample the same airstream. The test
aerosol should be introduced at a low enough concentration to insure that the count rate is below the upper
limit of the particle counter. These counts should be
accumulated and the correlation ratio should be calculated for particle sizes of 0.1 to 0.2 m. Counts may be

IEST-RP-CC007.2

The resistance measured by the differential pressure


device with no test filter in place should be considered
a tare resistance. This measurement should be taken
using a blank filter with the mounting chuck closed
and the airflow equal to the test airflow. The tare resistance should be subtracted from the gross measurement indicated by the differential pressure device for
the static pressure loss across the test filter. The reported filter resistance should not include any static pressure losses that are attributable to test duct
configuration or adapters or blank-offs required to
mount the filter.
The pressure drop tare measurement should be repeated each time the filter mount is changed to accommodate different size or style filters.

5.19 Reference filters


While no traceable standard of filter efficiency exists,
it is nonetheless useful to establish reference filters for
individual use. Immediately after thorough and careful
qualification testing and calibration of a test system as
previously described, several filters should be carefully
tested to measure their characteristic penetration and
pressure drop.
The length of the test should be sufficient to provide
reliable counting statistics (e.g., several hundred counts
of particles in the 0.1-to-0.2-m range, if possible). As a
result, repeated tests of one of the filters will provide
information about the overall repeatability of the test
system from test to test in short time intervals, and from
day to day, week to week, and so forth. The other filters
should be carefully stored for future reference.
If the measured penetration or pressure drop of the
reference filter changes (gradually, through filter loading or test-system drift; or rapidly, through damage to
the reference filter or fault in the test system), a backup
reference filter may be used to help pinpoint the problem. A reference filter should be measured daily. Prior
to using the last of the reference filters, a new set of
reference filters should be measured.

TEST PROCEDURE

Test system qualification and calibration in accordance


with section 5 should be completed prior to performing
the following test procedures. Test procedure should
be documented and should include such items as the
sequence of operations, the daily startup checks (see

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

15

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

Appendix B), and routine filter tests. The response


times required for each operation should be included.

6.1

Mounting of filters under test


(Section A20)

The test filter should be placed in the chuck of the test


duct in the manner which best simulates its actual use,
as noted in section 4.2.6. The test airflow direction and
filter sealing face should be recorded.

6.2

Airflow

The volume flow rate of the test air should be set,


measured, and recorded.

6.3

Pressure drop

The filter pressure drop should be measured, the tare


pressure subtracted, and the result recorded.

6.4

Temperature

The test airflow temperature should be measured and


recorded.

6.5

Challenge aerosol

The challenge aerosol should be activated, and the


required response time should be allowed to elapse.

6.6

6.7

Acceptance or rejection
of the filter

If the UCL of penetration is below the specified


penetration, pass the filter. If the LCL of penetration
is above the specified penetration, fail the filter. If the
specified penetration is between the UCL and LCL of
penetration, the tester has a choice of two actions:
1) fail the filter, or 2) continue the test to collect more
data.
If filter testing is continued, all upstream and downstream counts measured on that filter should be
summed, and the total upstream and total downstream
counts should be used to calculate the UCL and LCL
as before. If the test is continued for a length of time
sufficient to permit three or more sets of particle
counts to be measured, the procedures of TAPPIT1205 should be used to identify outlying counts that
might have resulted from some nonrandom cause.
Those outlying counts identified in this way should be
excluded from the sum of all counts.

DATA REDUCTION
(Section A5)

Symbols

Sequential counting should be performed by measuring upstream and downstream particle counts, making
appropriate allowance for system response times (see
section 5.16).
A single upstream-downstream sampling cycle is sufficient for generated aerosol sources producing upstream counts that vary by less than 5% from one
sample period to the next in the particle size range
from 0.1 to 0.2 m or appropriate range. For aerosols
that vary by more than 5% from one sample period to
the next in the particle size range from 0.1 to 0.2 m or
appropriate range, the upstream-downstream sampling
cycle should be repeated twice more for a total of three
upstream and three downstream samples.
The upstream and downstream counts and sampling
periods should be recorded, if the number and duration
of the sampling periods are not equal.

6.6.2 Simultaneous counting systems


The particles for both upstream and downstream
should be counted and recorded.

Copyrighted material

The upper confidence limit of penetration (UCL) and


the lower confidence limit (LCL) of penetration should
be calculated, as explained in section 7.

Particle counting

6.6.1 Sequential counting systems

16

6.6.3 Calculation

U
D
R
P
T

upstream counts
downstream counts
correlation ratio
penetration
sampling time

Subscripts
o
ucl
lcl
spec
c
t
uc
dc
ut
dt

observed
upper confidence limit
lower confidence limit
specified performance
correlation (i.e., no filter installed)
testing a filter
upstream during correlation
downstream during correlation
upstream during test
downstream during test

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

7.1

Correlation ratio

7.1.1 Observed valuecorrelation ratio


Using counts obtained with no filter installed, the observed correlation ratio should be calculated as:

Ro =

Do,c
U o,c

7.1.2 UCL and LCL valuescorrelation


ratio
For numbers n 50, Table A1 should be used to determine the upper and lower confidence limits for the
upstream and downstream counts with no filter installed.
For numbers n > 50, then
n2 n
should be used. The correlation ratios should be calculated as:
Ducl ,c
Rucl =
U lcl ,c

Rlcl =

7.2

7.3

Retest
(Section 6.7)

If Pucl Pspec, the filter has passed.


If Plcl > Pspec, the filter has failed.
If Plcl < Pspec < Pucl, a retest is permitted as described in
section 6.7, or, if no retest is done, the filter has failed.

7.4

Calculations for unequal


sample times

If

Tuc Tut
=
Tdc Tdt
then no adjustments for sampling time need to be
made.
If this condition is not met, then the equation for the
observed penetration is:

Do,t
T ut

Po =

D T T dt
U o,t o,c uc
U o,c T dc

Dlcl ,c
U ucl ,c

Penetration

7.2.1 Observed valuepenetration


With the test filter installed, upstream and downstream
counts should be obtained to calculate the observed
penetration:
Do,t
Po =
U o,t Ro

The equations for the UCL and LCL values of the penetration are:

Ducl,t
T ut T dc

Pucl =

D
T uc T dt
lcl,c
U lcl,t

U ucl,c

Dlcl,t
T ut T dc

Plcl =

D
T uc T dt
U ucl,t ucl,c
U lcl,c

7.2.2 UCL and LCL valuespenetration


The UCL and LCL values should be calculated for the
upstream and downstream counts, using Table A1 for
numbers n 50, or using
n2 n

for n > 50. The UCL and LCL of the penetration


should be calculated as:

Pucl =
Plcl =

IEST-RP-CC007.2

Ducl ,t
U lcl ,t Rlcl
Dlcl ,t
U ucl ,t Rucl

7.5

Significant digits and rounding

Calculated values of penetration should be rounded,


using standard conventions, to two significant digits.
When comparing the upper confidence limit on the
filter penetration to the acceptance specification, it is
not appropriate to round to fewer than two significant
digits.
Such rounding of digits should not be understood to
mean that the measurement is precise to two significant digits. The spread between the observed penetration and the 95% confidence levels on the penetration

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

17

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

for repeated tests on reference filters provides estimates of the precision of the penetration measurement.

7.6

(Appendix A, sections A4 and A16)


Resistance of the filter should be corrected for air viscosity as a result of test air temperature that is different
from the standard air temperature of 21 C (70 F).
The suggested correction factor formula is:

Pm
C1 + [(C 2 )(TT )]

1) manufacturer name or symbol


2) filter serial number, model number, and date
3) reference to this RP (e.g., Tested per IESTRP-CC007)
4) test flow rate: actual volume per unit time
5) 95% UCL penetration for at least one particle
size range in the span of 0.1 to 0.2 m

where:
Pc = filter resistance in Pa (in. w.c.), corrected to 21
C (70 F)
Pm = filter resistance in Pa (in. w.c.), measured
C1 = 1st constant, 0.9468 for C (0.902 for F)
C2 = 2nd constant, 0.00252 for C (0.0014 for F)
TT = test air temperature, C (F)

REPORTING AND MARKING

a) Unless otherwise specified by contract, each filter


unit should be marked with the following information:

Resistance

Pc =

6) observed penetration for same size range


7) particle size range for which the 95% UCL
and observed penetrations are reported
8) identification of MPPS in m
9) airflow resistance corrected as described in
section 7.6
10) arrow showing direction of test airflow and
identification of mounting surface during test
(e.g., Filter sealed to upstream face during test)
11) test aerosol material
In addition, penetration data for other particle sizes that
span the most penetrating size should be provided.
b) Marking should be centrally located on the top of
the filter frame with the pleats vertical.

18

Copyrighted material

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

APPENDIX AFACTORS INFLUENCING MEASURED PENETRATION


A1
A2
A3
A4
A5
A6
A7
A8
A9
A10
A11
A12
A13
A14
A15
A16
A17
A18
A19
A20

Particle size sphere equivalence


Mixing of aerosols
Airflow measurement
Viscosity and pressure drop
Poisson statistics and counting
Particle counter saturation
Particle counter background count rates
System response time
Anisokinetic sampling
Charge distribution of aerosols
Particle size distribution slope across size
channels
Solid aerosol cake buildup
Liquid aerosol and fiber diameter buildup
Air ion density at the filter medium
Charge distribution inside aerosol transport lines
Air temperature
Air relative humidity
Barometric pressure
Particle specific gravity
Mechanical vibration of the filter

A1 PARTICLE SIZE SPHERE


EQUIVALENCE
Particle measuring instruments can provide only sufficient information about a particle to state that the particle is equivalent to a sphere of stated size and has other
specified properties, for example, density and refractive index. The equivalent sphere concept should be
used because particle composition (which may determine density and refractive index) and particle shape
are unknown. These properties will affect the response
of the instrument. Further, a population of particles
may have a wide range of shapes and physical properties. The equivalency measurement may involve the
amount of light scattered, the electrical mobility, or
another variable. No instrument can provide an equivalent particle size directly suitable for the determination
of filter performance in the size range of interest. Instead, aerosols capable of producing equivalent responses in optical sizing instruments should be used. A
list is provided in section 4.2.9.
Instruments using different designs, or different implementations of the same principle (e.g., different
scattering angles for laser particle counters), may have
a systematic bias greater than the precision of the in-

IEST-RP-CC007.2

struments in particle sizing. Even if instruments have


been adjusted to provide an identical response with a
calibration aerosol, measurements of aerosols having
different compositions may yield different results.

A2 MIXING OF AEROSOLS
In testing for the efficiency of a filter, the challenge
aerosol should be homogeneous over the entire upstream area, and the downstream sample should represent filtered air from all points equally.
The measured penetration of a filter may vary by as
much as several orders of magnitude when the filter is
rotated about its axis between tests in a test duct that
does not mix upstream and downstream aerosol. To
ensure uniformity of the challenge aerosol, the aerosol
should pass through static or motionless mixers between the aerosol generator and the upstream sampling
probe. Downstream air should pass through similar
mixers between the test filter and the downstream
sampling probe.

A3 AIRFLOW MEASUREMENT
The pressure drop across a filter medium as a function
of airflow is a primary measure of filter performance.
Particle collection efficiency is the only other key
measurable property of the filter. Both measurements
depend on how airflow is controlled within the test
duct.
Increasing airflow, inadvertently or deliberately, will
lead to the following:
a) The pressure drop will increase, and the observer
may incorrectly ascribe this increase to a change in the
filter.
b) Particle capture efficiency by impaction (if any)
will be enhanced.
c) However, particle capture efficiency by diffusion
will be reduced. Because particle capture by interception is unchanged, the MPPS (where the sum of both
interception and diffusion is at a minimum) will become smaller, and the efficiency at the new size will
be lower than it was at the old size.
d) If the output of the generator remains constant,
even the challenge aerosol concentration will change.
There are two possible measurements of flow: mass
flow and volume flow. Volume flow is expressed in
units of volume per time, for example, cubic feet per
minute. Mass flow is expressed as mass per time, for
example, pounds (mass) per hour. Mass flow is im-

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

19

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

portant in chemical reactions such as combustion or in


heat transfer such as steam heating. Volume flow is the
controlling parameter in valves, aerosol concentrations, sheet permeabilities, and filtration. Therefore, it
is volume flow and not mass flow that is relevant to
filtration and is measured.

the standard air temperature of 21 C (70 F). A correction factor formula is:

Volumetric flow may be directly measured by volume


flow devices. Alternatively, mass flow may be measured by mass flow sensors and the volumetric flow
determined by computation. Commercial mass flow
sensors measure flow rate from heat transfer from a
heated wire in the airflow (e.g., hot-wire anemometers)
and typically provide this conversion automatically.

where:

In practice, airflow should be measured with a device


of desired accuracy and with a stable output. If a mass
flow instrument is used, the air temperature and barometric pressure should be measured and corrections
should be applied. The manufacturer recommendations
should be followed carefully. Periodic checks with a
second, different type of flow instrument are desirable
for routine work. ASME-defined flow nozzles, used
with liquid-filled manometers, are ideal. For rigorous
work, an alternate type of in-line flow measurement is
recommended.

A4 VISCOSITY AND
PRESSURE DROP
The pressure drop required to force air through a filter
medium is a function of the viscosity of the air. In fact,
in contrast to ducts and valves, filter media usually
owe nearly all their resistance to viscosity and almost
none to inertial effects. Therefore, if the viscosity of
the test air changes, the pressure drop across the filter
medium changes proportionally.

A4.2 Practice
The viscosity of air varies by 5% between 10 C and
29.4 C (50 F and 85 F). Thus, an uncompensated
tester operating in this range will wrongly indicate a
5% variation in inherent media resistanceprobably
the largest non-reproducibility that can be suffered in
an otherwise functional, volumetric flow system.

A4.3 Recommendation
The best approach is to regulate test temperature, but
this method can be very expensive.
When it is not possible to test filters at 21 C (70 F),
the resistance of the filter should be corrected for air
viscosity due to variations in test air temperature from

Copyrighted material

Pm
+ c(TT ST

Pc = filter resistance, corrected to 21 C (70 F)


Pm = filter resistance, measured

= viscosity of air at 21 C (70 F)


(0.0178 centipoise)

= viscosity change with temperature,


centipoise/C (1.4 10-5)
[centipoise/F (7.8 10-6)]

TT = test air temperature, C


ST = standard air temperature, (21 C or 70 F)
A shortened form of suitable accuracy is given in section 7.6. The range of applicability of these corrections
is 10 to 38 C (50 to 100 F).

A5 POISSON STATISTICS
AND COUNTING
A5.1 Theory

A4.1 Theory

20

Pc =

When a well-mixed, stable aerosol penetrates a filter,


penetrating particles will appear downstream of the
filter (or in a small downstream air sample) randomly,
but at some average population density. A particle
counter will detect these particles randomly in time,
but at an average rate. For the purpose of calculating
penetration, the average rate (particles per unit time or
per unit volume) is obtained from the cumulative count
measured over the time period of the test or over the
volume sampled.
The statistics of particle counting become increasingly
important as the filter penetration, and hence downstream counts, decrease. These variations are described
by Poisson statistics. Of primary importance to this
type of testing is the relationship between the results of
a single test and the results that would be obtained
from a test of infinite duration, the true mean result.
This relationship between an observed result and the
implied confidence limits on the true mean result is
described by Box, et al (Appendix F).

A5.2 Practice
When top-performance, noise-free particle counters
are used in a good duct according to this RP, count
statistics become the largest error on highly efficient

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

filter tests. When testing ULPA filters according to this


RP, low downstream particle counts are the largest
source of uncertainty.

A5.3 Recommendation
A5.3.1 Determination of confidence limits
on a count
This procedure uses particle count data to establish the
confidence limits on penetration. Section A5.3.5 gives
the 95% confidence limits on a single observed particle count from 0 to 50. For a single observed particle
count n, there is a 95% confidence that the true mean
count is between the upper and lower limits given in
the table below. The true mean count is the average
count that would be obtained if the tests were repeated
indefinitely.
For larger values of counts (n), the Poisson distribution
tends toward normality with mean n and variance n. In
this case, the 95% confidence levels for a count of n
can be expressed
n2 n

See Abramowitz and Stegun (Appendix F).


Example:
Observed count (n)
0
10
100
10,000

95% Confidence Limits


Lower
Upper
0
3.7
4.7
18.4
80
120
9,800
10,200

Once the confidence limits on a particle count are established, it is necessary to establish the confidence
limits on the correlation ratio and penetration.
In most cases, one number in the ratio is substantially
smaller than the other, and it is the uncertainty in the
smaller value that dominates the uncertainty in the
ratio. In this case, it is reasonable to calculate the confidence limits of the ratio by calculating the ratio with
the limits of the smaller value and the observed level
of the larger value.

A5.3.2 Correlation ratio


Statistical uncertainty exists in the ratio of downstream
to upstream counts, with no filter in the system. This
uncertainty should be established before the penetration is calculated.

IEST-RP-CC007.2

Example:
Observed Value
10,000 counts upstream
1,000,000 counts
downstream

95% Confidence Limits


Lower
Upper
9,800
10,200
998,000
1,002,000

NOTE: In this example, the upstream counts are lower


than the downstream counts because the upstream
sample has been taken through a diluter.
Thus, the confidence limits on the correlation ratio are:
Lower: 1,000,000 10,200 = 98
Upper: 1,000,000 9,800 = 102
If the uncertainty in the correlation ratio is significantly less than in the filter under test, it is reasonable to
use the observed value of the correlation ratio, in this
case 100. Otherwise, the 95% confidence limits should
be used.

A5.3.3 Penetration
This correlation example is used to calculate the penetration of a filter test:
Observed Value
10,000 counts upstream
10 counts downstream

95% Confidence Limits


Lower
Upper
9,800
10,200
4.7
18.4

Using the 95% confidence limits on the correlation


ratio results in 95% confidence limits on the penetration of:
Lower: 4.7 (10,200 102) = 4.5 10-6
Upper: 18.4 (9,800 98) = 1.9 10-5
It should be observed that, in this case where the uncertainty in the correlation ratio was small, the same
results would have been obtained by using the observed value of the correlation ratio.
In this example, it can be stated that, with 95% confidence, the filter penetration is less than 1.9 10-5,
that is, 0.0019%; or, the efficiency is greater than
99.9981%.
NOTE:

The particle size range in which the counts were


obtained should also be given.

This confidence level is based on counting statistics only.

Other error sources may contribute to the uncertainty of the penetration measurement.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

21

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

For higher penetration filters, this analysis may


not be applicable.
The statistical procedures described here apply
only to raw count data. These methods should
not be applied to data that have been scaled,
multiplied by correlation ratios, converted to
rates or concentrations, and so forth; to do so
will yield erroneous results.

A5.3.4 Test procedures


In a typical testing situation, where the filter should
prove to be greater than a specified efficiency, either of
two methods may be used.
a) The internal test specification written to ensure
performance in a pass/fail manner is suitable for manual systems.
b) Or, the confidence levels may be calculated continuously as the counts accumulate, until the filter
passes or fails.
Method (b) is faster but requires an automated test.
The pass/fail testing method can be illustrated by the
previous examples. To prove a filter is more than
99.998% efficient in a particle size range, it is sufficient to accumulate 1,000,000 counts upstream and
require 10 or fewer downstream. In the case of a system with two counters and a correlation ratio of 1002,
it is sufficient to accumulate 10,000 counts upstream
and require 10 or fewer downstream.
For the method in which confidence limits are constantly calculated as counts are accumulated, the test is simply continued until the filter is proved to pass or fail the
specification. This method requires less time and will
not reject good filters as a result of test variability.

A5.3.5 95% confidence limits for the mean


value of a Poisson variable
Table A1. Upper and lower confidence limits
for upstream and downstream counts.
Observed
Count (n)
Lower
0
0.0
1
0.1
2
0.2
3
0.6
4
1.0
5
1.6
6
2.2
7
2.8

22

Copyrighted material

Upper
3.7
5.6
7.2
8.8
10.2
11.7
13.1
14.4

Observed
Count (n)
Lower
8
3.4
9
4.0
10
4.7
11
5.4
12
6.2
13
6.9
14
7.7
15
8.4
16
9.2
17
9.9
18
10.7
19
11.5
20
12.2
21
13.0
22
13.8
23
14.6
24
15.4
25
16.2
26
17.0
27
17.8
28
18.6
29
19.4
30
20.2
31
21.0
32
21.8
33
22.7
34
23.5
35
24.3
36
25.1
37
26.0
38
26.8
39
27.7
40
28.6
41
29.4
42
30.3
43
31.1
44
32.0
45
32.8
46
33.6
47
34.5
48
35.3
49
36.1
50
37.0

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Upper
15.8
17.1
18.4
19.7
21.0
22.3
23.5
24.8
26.0
27.2
28.4
29.6
30.8
32.0
33.2
34.4
35.6
36.8
38.0
39.2
40.4
41.6
42.8
44.0
45.1
46.3
47.5
48.7
49.8
51.0
52.2
53.3
54.5
55.6
56.8
57.9
59.0
60.2
61.3
62.5
63.6
64.8
65.9

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

A6 PARTICLE COUNTER
SATURATION AND
COINCIDENCE
Particle counters produce erroneous data when sampling high-concentration aerosols. Particle counters
exhibit coincidence error when two or more particles
occupy the sensing zone at the same time because overly rapid pulses cause the electronics to saturate. Reported counts are erroneously low in these situations.
Low counts due to coincidence errors are not obvious.
If the errors are not caught, the result is usually a filter
efficiency curve that has an apparent low efficiency for
small particle sizes because the challenge count was
saturated in these sizes.
For routine work, the test system should be known
always to operate below saturation levels by design
and subsequent measurement, as recommended in section 5.6.

The speed at which a test system returns to a new


steady-state (i.e., the response time) may be measured
by inserting a filter into the test system and observing
the decay of the downstream counts. An unusually
long response time may point to other problems with
the system. In practice, when purge times between
tests are not longer than the response time measured,
one may observe erroneously high counts downstream.
During design of a test system, the response time should
be considered on the basis of flow rates and interior
volumes. System volume should be minimized, without
causing undue pressure losses or extreme aerosol velocitiesfor example, by maintaining air velocities at levels
from approximately 100 up to 1000 m/min (a few hundred to as high as 2000 ft/min).

A9 ANISOKINETIC SAMPLING

A7 PARTICLE COUNTER
BACKGROUND COUNT
RATES

Large particles behave somewhat like projectiles. If


the air velocity in a probe is identical to the bulk flow
and the probe is aligned with the flow direction, all
particles are collected without sampling bias due to
inertia effects. In contrast, the fraction of large particles will be enriched if the probe velocity is lower
than the bulk flow.

Particle counters will register some counts when there


are no particles. This phenomenon is particularly true
of laser particle counters in their most sensitive channel, because the lower size limit is generally set as
close to the noise threshold limit as the manufacturer
considers possible.

In industrial ventilation, particles of 100 m and larger


are carried in ducts at several hundred cm/sec. Collecting a representative sample requires isokinetic sampling. In practice this is achieved by using a sharpedged probe that has a forward-facing inlet with the
same air velocity as the stream being sampled.

For routine production work, a filter of much higher


efficiency than those routinely tested should be periodically tested. This higher-efficiency filter is considered
a reference filter. This periodic testing verifies that
background count rate does not significantly change
routine results. Although damage or change in the reference filter can cause the check not to proceed as intended, this test is conservative in that it cannot cause a
false sense of security.

For routine production work involving particles


smaller than 1 m, the anisokinetic effect can be ignored since sampling biases are small. For particles
10 m and larger at velocities of more than
1000 m/min (thousands of ft/min), corrections should
be made for the sample bias introduced, or the sample bias should be avoided through the use of isokinetic sampling conditions.

A8 SYSTEM RESPONSE TIME

A10 CHARGE DISTRIBUTION OF


AEROSOLS

When a test system has been disturbed (such as by


inserting a filter sample or by switching a sampling
location valve), aerosol concentrations inside the system are affected temporally and spatially. The rate of
return to a new steady-state depends on interior volumes, flow rates, and rates of mixing for the various
regions inside the tester, as well as on how these regions follow one another in the system as a whole. In
the case of transition from high aerosol concentration
to much lower, concentration within a given region
decays exponentially and may occur over a significant
length of time.

In gases containing bipolar ions and aerosol particles,


an equilibrium charge distribution will be reached on
the aerosol as a result of the random thermal motion of
the ions and frequent collisions between the aerosols
and the ions. Such equilibrium charge distributions are
found in atmospheric particles due to ionization by
cosmic rays and natural and man-made radioactive
materials. In the laboratory, aerosols exposed to bipolar ion sources, such as Kr85 (a radioactive isotope of
krypton), will also carry such an equilibrium charge
distribution. That is, the aerosol itself is neutral while

IEST-RP-CC007.2

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

23

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

individual particles may carry one or more charges of


either polarity. Such a laboratory aerosol is also termed
a neutralized aerosol.
The bipolar equilibrium charge distribution on aerosols
has been extensively investigated, and the charge distribution was found to obey the Gaussian distribution
predicted by Boltzmanns Law. More detailed discussions with further references on equilibrium charge
distribution on aerosols are given by Liu and Pui, 1974
(see Appendix F).
Because particles used in filter testing may be charged
due to their generation and transport, they may be
readily collected in the sampling and measurement
systems. Therefore it is recommended that aerosols,
especially those of unknown origin, be neutralized
before use. A charge neutralizer (aerosol neutralizer)
should be installed considerably upstream of the filter
so that most air ions are scavenged by tube and duct
walls before reaching the filter medium itself, thus
minimizing the effect of air ions on the results.

A11 PARTICLE SIZE


DISTRIBUTION SLOPE
ACROSS SIZE CHANNELS
The ideal particle counter records the number of particles that lie within each of several channels, each
channel having an upper and a lower size limit. It often
happens that a filter test aerosol size distribution slope
is steeper than normal distribution, resulting in many
more particles in one channel than in the adjacent
channel. Thus, the median diameter of the particles
within a channel may be nearer to the upper or lower
size limit of the channel than to the midpoint (geometric or arithmetic) of the channel. If penetration data are
plotted using the midpoints of the channels, the curve
may be shifted.
By using a challenge aerosol having a mean diameter
equal to the MPPS of the filter under test, and by
measuring the penetration for that MPPS, errors due to
particle size distribution slope for the MPPS can be
reduced or eliminated. This method also helps minimize errors in measuring penetration that result from
particle-sizing errors.

A12 SOLID AEROSOL CAKE


BUILDUP

Copyrighted material

Solid aerosol buildup can have a more dramatic effect


on electret media. In these situations, ultrafine solid
aerosol has been shown to mask the electrostatic field,
causing a noticeable drop in filter efficiency.

A13 LIQUID AEROSOL AND FIBER


DIAMETER BUILDUP
As in the case of buildup of a cake of solid aerosol
particles, filter performance might be expected to
change as a liquid aerosol is collected. However,
there are additional reasons for altered filter performance related to the use of liquid aerosols, including
the following:

The diameters of fibers within a filter medium


are known to be very important in particle capture. Liquid particles can wet a fiber, adding to
the thickness of the filter.

The same quantity of liquid causes a significantly larger fractional increase in the diameter
of a small fiber than in a larger fiber.

Filter media differ in their tendency to saturate


and seal themselves with a liquid.

Electret filter media may become neutralized by


a thin film of liquid around the fibers.

Oily liquids are known to enhance the adhesion


of large particles to air filter fibers and are
sometimes added to filters to increase particle
adhesion.

A14 AIR ION DENSITY AT THE


FILTER MEDIUM
Air ions are 1-nm clumps of 30 to 40 gas molecules
that have a net charge of 1 charge unit. Air ions are
created spontaneously by ionizing radiation (hence, the
name) or by strong electric fields. The net effect of air
ions in air is to make the air functionally slightly conductive. A preponderance of positive or negative ions
can enable air to impart a net charge to any object.
All of the electrical effects in filtration and aerosols
can be changed by the presence of ions.

If enough solid aerosol particles are deposited in or on


a filter, the accumulated mass of particles acts as another filter in series with the first. The penetration
through a mechanical filter medium has been observed
to decay exponentially with time under a constant aerosol challenge. In testing low penetration filters, the

24

aerosol cake prefiltration can combine with poor count


statistics to make an efficiency immeasurable, even
though there might be infinite sampling time and zero
counter noise.

Ions diffuse rapidly because of their small size. For


this reason, ions are stripped from air when transported
in conductive plumbing. When an ion is able to return
to charge state zero, it rapidly degenerates into molecules of the gases that compose air.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

A15 CHARGE DISTRIBUTION


INSIDE AEROSOL
TRANSPORT LINES
Moving air can impart chargestreaming potential
to solid objects, including the inside of a tube carrying
a test aerosol. Insulating materials can allow the charge
to accumulate, or the charge can slowly migrate
around the surface. Such behavior is erratic and depends heavily on humidity and surface contamination.
In the presence of an electric field, particles with net
charge (usually most particles) are affected by
Coulombic attraction or repulsion. This behavior can
cause particle deposition loss resulting in measurement
errors. These errors can be avoided by using grounded
transport lines of conductive materials such as stainless
steel. Stainless steel or aluminum transport lines, mixers, and perhaps ductwork should be used throughout.
These precautions help prevent charge buildup.
If sampling lines are nonconductive, electrostatic attraction will dominate other particle removal mechanisms for particles in the range of 0.1 to 0.2 m. It is of
utmost importance to have a sampling system that is
completely conductive, including fittings and tubing.
A length as short as 25 to 50 mm (1 to 2 in.) of highstatic
tubing
such
as
polyethylene
or
polytetrafluoroethylene can effectively precipitate out
all charged particles in the size range 0.025 to 0.5 m.

A16 AIR TEMPERATURE


(Section A4)
Variation in air temperature and barometric pressure
causes variations in several phenomena related to filtration. In extreme cases, such variation may also affect diffusion effects of particles.
The thermal effects can be minimized by maintaining
constant laboratory temperatures, typically 21 oC
(70 oF), and by using volumetric flow rather than absolute flows. It is also good practice to correct for barometric pressure when using mass flow measurements
to determine volumetric (actual) flow rates for the test.

A17 AIR RELATIVE HUMIDITY


Relative humidity profoundly affects the conductivities
of surfaces and high-surface-area matrices. These conductivities, in turn, control filtration mechanisms that
may occur, including: Coulombic and dipolar electrophoresis, streaming potential, and triboelectrification.
In fact, humidity is used as a diagnostic tool in the
study of these phenomena.

IEST-RP-CC007.2

In addition, humidity also affects adhesive forces for


large particles. Further, particle size or other properties
of some aerosols change, depending on humidity.
Large changes in humidity are known to change test
results for electret filter media and to change size distributions for NaCl aerosols. Humidity is sometimes
applied as a test of electrostatic filtration mechanisms.
HEPA filters are often used in controlled-relativehumidity environments.
To minimize the effects of humidity variations, a RH
between 30% and 70% should be maintained.

A18 BAROMETRIC PRESSURE


Barometric pressure influences many of the same variables as does temperature.
Filters for compressed air or vacuum line service are
observed to behave differently when pressures deviate
from one atmosphere.
At the minimum, flow should be recorded and corrected to actual volume flow. For rigorous work, flow and
pressure drop could be accurately calculated back to
1013.2 mBar (29.92 in. Hg) as a standard. Pressure
could also conceivably be controlled in a test duct.

A19 PARTICLE SPECIFIC


GRAVITY
When particles of identical size and shape but different
specific gravity are collected in a filter by impaction or
diffusion, the particles will be collected in different
quantities. (Also, some instruments will indicate particles of identical size and shape but different specific
gravity to be of different size.)
Different penetration results have been obtained in
studies of aerosols having different specific gravities.
The type of test aerosol should be stated. If the collection mechanism is either impaction or diffusion, the
effect of density should be predicted and, for rigorous
work, confirmed experimentally.

A20 MECHANICAL VIBRATION


OF THE FILTER
The ability of vibration to dislodge particles collected
by filters has been observed in practice, especially for
large particles.
To minimize this effect, the filter housing should be
isolated from mechanical vibration of the fan mounting, pumps, and so forth. Fan noise and other acoustical energy should be muffled before it reaches the
test filter.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

25

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

APPENDIX BRECOMMENDED CALIBRATION INTERVALS


Table B1. Minimum frequency of calibration.

Calibration Procedure
Flow
Pressure drop
Temperature
Relative humidity
Particle counters
Zero count
Sizing
High-count-rate error
Neutrality of aerosol
Aerosol source
Size distribution
Stability
Dilution
Duct integrity (leaks)
Cleanliness of test air
Airflow distribution
Upstream uniformity
Downstream leak
System particle losses
System response time
Correlation
Pressure drop tare
Reference filter test

26

Copyrighted material

Daily

Per Filter
Model

Yearly
X
X
X
X

Initial system
startup and
after system
changes
X
X
X
X

X
X
X

X
X
X
X

X
X
X
X

X
X
X
X

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

X
X
X
X
X
X
X
X
X
X
X
X
X

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

APPENDIX CCONSIDERATIONS FOR SPECIFIC MEDIA TYPES


C1 TESTING MEMBRANE
FILTERS

C2 ELECTROSTATICALLY
CHARGED MEDIA

In certain filters with efficiencies greater than


99.9999%, expanded polytetrafluoroethylene (ePTFE)
is an alternative to the traditional micro-fiberglass media used in the construction of these filters. Although
this type of media is a membrane, the media has a fibrous structure and thus properties similar to fibrous
media for particle penetration. Two distinct features
may affect testing of this type of filter:

Synthetic media are available with nominal efficiency


as high as 99.999%. The high efficiency is typically
achieved by electrostatically charging the fiber or media to enhance the filtration property. Several commercial and patented processes for charging are available
with different performance claims. Some electrostatic
media now available in commercial ULPA filters are
expected to be considered as alternatives to traditional
ULPA filters with uncharged media.

a) The mean size of the fibrous structure is much


smaller than micro-fiberglass, resulting in a MPPS
significantly less than 0.1 m for this media. By comparison, the MPPS for a typical micro-fiberglass media
is between 0.1 and 0.25 m. Therefore, testing this
variety of filters at their MPPS requires the ability to
detect particles as small as 0.05 m, which is well outside the useful range of particle counters. As a result,
ePTFE filters may require the use of condensation
nucleus counters (CNCs) or other methods sensitive to
the small particle sizes.
b) Unlike traditional micro-fiberglass media, the ePTFE
membrane is a thin layer of fibrous structure. This
membrane may be too delicate to handle alone and
therefore may be layered onto other, easier-to-handle
media, which may affect filtration. Further, in practice,
many manufacturers layer the ePTFE to compensate for
spatial non-uniformity in each layer of the ePTFE that
could cause filter failure due to leaks.

IEST-RP-CC007.2

Unlike actively charged electrostatic precipitators that


use external power to maintain a charge, electrostatic
charge in these media dissipates as particles are attracted to the media and collected, or with time.
Charge dissipation is particularly evident when liquid
or oily aerosols are collected. Consequently, the performance of these charged filters will vary depending
on the type of aerosol material used to test them. Further, performance deteriorates as the filters collect particles while in use. In some cases, performance
deteriorates several orders of magnitude when all the
effects of charge are dissipated. Since ULPA filters are
usually in continuous service for months at a time and
in critical applications, testing filters with charged media requires special considerations.
Finally, because of the effects of charge dissipation,
caution should be used when considering electrostatically charged filters in critical applications.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

27

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

APPENDIX DOTHER TEST METHODS FOR ULPA FILTERS


Various recommended practices and standards are
used throughout the world to test ULPA filters. Appendix C (Other Filter Test Standards and Methods) of
IEST-RP-CC001 contains an overview of some of
these test methods.

28

Copyrighted material

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

APPENDIX EPARTICLE DETECTION AND PARTICLE


SIZE DISTRIBUTION
E1 THEORY
Accurate measurement of the upstream and downstream
particle size and concentration is critical to the quantification of filter performance. A complication is that the
challenge aerosol may have a distribution of sizes
(polydisperse), each with an unique filter collection
efficiency. Therefore, if a polydisperse aerosol is used as
a challenge an instrument capable of size discrimination
should be used. If an instrument incapable of size discrimination is used, ideally a monodisperse (single particle size) aerosol should be used as a challenge.
The particle size and concentration in real-time instruments are not measured directly but are deduced from
light scattering or electrical current. Sometimes the
aerosol particles are separated prior to detection by
particle diffusion, electrical mobility, or inertia. In optical particle counters, the light scattered from individual
particles is counted and the magnitude of scattered light
is related to particle size. Another complication is that
the dominant collection mechanisms for highefficiency filters are interception and diffusion depending on particle diameter. Both mechanisms depend on
physical particle size. Therefore, it is desirable to
measure or deduce physical particle size of the upstream and downstream aerosol.
All aerosols have a distribution of particle size; the
critical question is the range of the distribution. Even
calibration polystyrene microspheres have a very small
variation in size. Polystyrene microsphere calibration
particles are directly determined by microscopy for
certification as a standard (these data are usually provided by the supplier). Particle size distribution data
are characterized by two statistics: a measure of the
central tendency and measure of the distribution. The
central tendency may be determined by the mean,
mode, median, or geometric mean. The distribution
may be measured by the standard deviation or geometric standard deviation. In addition, the central tendency
statistic depends on the property measured. For example, separation of the aerosol by impaction and detection with piezoelectric elements will yield a
measurement of particle size distribution by mass.
Measurement by optical particle counting will yield a
particle size distribution by number. The particle size
distributions measured by the two techniques, in particular the mean particle diameter, will be quite different. The relative magnitudes are given by:

IEST-RP-CC007.2

Particle size by number < Particle size by surface <


Particle size by mass
More details on particle size distributions are found in
Hinds (1982) in Appendix F.

E2 PRACTICE
In practice, particle size distributions are assumed to
follow a lognormal relationship. Lognormal distributions are a normal probability function with logarithm
of the particle size. The properties of the size distribution are the geometric mean and the geometric standard deviation. This is usually a valid assumption if the
aerosol is generated by one mechanism, such as nebulization. In lognormal size distributions, the geometric
standard deviation does not depend on the property of
the size distribution measured, e.g. number, surface or
mass. The various means can be related mathematically. For example (Hinds, 1982, Appendix F):

MMD = CMD exp(3 ln 2 g )


where:
MMD is the mass mean diameter
CMD is the count mean diameter

g is the geometric standard deviation.


For careful work, in particular if the particle size distribution deviates from lognormal, various means and
standard deviation are computed directly from the distribution data.

E3 RECOMMENDATIONS
If the challenge aerosol particle size distribution characteristics need to be reported, the geometric number
mean and geometric standard deviation should be used.
If means based on surface or mass are used, this should
be explicitly identified. If test data from two test facilities are compared, care should be taken that the statistics
used to report the size distribution are identical and instrumentation differences are understood.
The method of measurement and the specific instrument including manufacturer and model number
should be explicitly reported.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

Copyrighted material

29

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

APPENDIX FBIBLIOGRAPHY
Abramowitz, M. and I.A. Stegun. 1972. Handbook of
Mathematical Functions. National Bureau of Standards (available from National Institute of Standards
and Technology (NIST), 100 Bureau Drive, Stop 1070,
Gaithersburg, MD 20899-1070, USA; 301-975-6478;
www.nist.gov).
ASHRAE-51: Laboratory Methods of Testing Fans for
Rating.
ASTM F1471-93: Standard Test Method For Air
Cleaning Performance Of A High-Efficiency Particulate Air Filter System.
Box, G.E.P., W.G. Hunter, and J.S. Hunter. 1978. Statistics for Experimenters. John Wiley & Sons.
Fain, D.E. and T.W. Selby. 1984. Calibration and Use
of Filter Test Facility Orifice Plates. Oak Ridge Gaseous Diffusion Plant, Martin Marietta Energy Systems
Corp., reported page 1168, 18th Department of Energy
Nuclear Airborne Waste Management Air Cleaning
Conference, Baltimore, Maryland.
Hinds, W.C. 1982. Aerosol technology, properties,
behavior, and measurement of airborne particles. New
York: John Wiley and Sons.
IEST-RD-CC011: A Glossary of Terms and Definitions Relating to Contamination Control.

30

Copyrighted material

Lee, K.W. and B.Y.H. Liu. 1980. On the Minimum


Efficiency and the Most Penetrating Particle Size for
Fibrous Filters. Air Pollution Control Association
Journal 30 (4): 377-381.
Liu, B.Y.H. and D.Y.H. Pui. 1974. Equilibrium Bipolar ChargeDistribution of Aerosols. Journal of Colloid and Interface Science 49 (2).
Liu, B.Y.H., D.Y.H. Pui, K.L. Rubow, and W.W.
Szymanski. 1985. Electrostatic Effects in Aerosol
Sampling and Filtration. Annals of Occupational Hygiene 29 (2): 251-269.
MIL-STD-282: Filter Units, Protective Clothing, GasMask Components, and Related Products: Performance-Test Methods.
Scripsick, R.C. and S.C. Soderholm. Final Report:
Evaluation of Methods, Instrumentation and Materials
Pertinent to Quality Assurance Filter Penetration
Testing. Los Alamos National Laboratories Report No.
LA-10748 (available from NIST, 100 Bureau Drive,
Stop 1070, Gaithersburg, MD 20899-1070, USA; 301975-6478; www.nist.gov).
U.S. DOE-STD-3025: Quality Assurance, Inspection,
and Testing of HEPA Filters.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY

IEST-RP-CC007.2

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

Licensed by IEST to Gerardo Gutierrez - Laboratorios Grossman SA. Order # 13218/Downloaded: 9/3/2014
Single-user license only, copying and networking prohibited.

INSTITUTE OF ENVIRONMENTAL SCIENCES AND TECHNOLOGY


Arlington Place One
2340 S. Arlington Heights Road, Suite 620
Arlington Heights, IL 60005-4510
Phone: (847) 981-0100 Fax: (847) 981-4130
E-mail: iest@iest.org Web: www.iest.org