MEMORANDUM
From:
BART
Project
Team,
EPA
Office
of
Air
Quality
Planning
and
Standards
To:
Docket
for
BART
Guidelines,
OAR
2002­
0076
Subject:
Fine
Particles:
Overview
of
Source
Testing
Approaches
Date:
April
1,
2005
A.
Background
Most
stationary
source
test
methods
specified
in
State
rules
do
not
adequately
quantify
either
total
PM
emissions
or
PM2.5
emissions.
Additionally,
some
of
the
current
stationary
source
test
methods
will
not
adequately
provide
a
uniform
indication
of
the
sources'
performance
in
controlling
PM2.5
emissions.
Most
source
test
methods
referenced
in
SIPs
provide
a
measurement
of
the
particulate
matter
that
is
solid
or
liquid
at
a
temperature
specified
in
the
method
or
applicable
standard.
Filtration
temperatures
of
250"
F
and
320"
F
are
typical
although
other
temperatures
may
be
specified
in
a
few
test
methods
or
applicable
standards.
Generally,
these
filterable
particulate
matter
test
methods
are
either
identical
or
very
similar
to
one
of
the
ten
Federal
test
methods
published
in
Appendix
A
of
40
CFR
Part
60
and
used
to
determine
compliance
with
New
Source
Performance
Standards
(
NSPS).
These
test
methods
are
adequate
to
evaluate
the
compliance
status
of
a
source
for
emissions
of
that
component
of
particulate
matter
evaluated
when
the
applicable
rule
was
developed.
However,
these
test
methods
do
not
provide
a
measurement
of
total
particulate
matter
emissions,
or
PM2.5
emissions.
The
test
method
proposed
to
determine
compliance
with
the
first
group
of
NSPS
(
36
FR
15713)
determined
the
sum
of
the
mass
of
material
collected
on
or
prior
to
the
filters
maintained
at
250"
F
and
the
material
collected
in
the
cooled
impingers
that
followed
the
filter.
While
the
material
collected
prior
to
the
filter
provided
a
measure
of
the
filterable
particulate
material,
the
material
collected
in
the
impingers
was
stated
to
measure
vapors
in
the
stack
that
would
become
particulate
matter
at
70"
F
(
36
FR
15495).
When
combined,
the
method
provided
a
measurement
of
the
total
particulate
matter
emissions
from
the
facility
tested.
The
promulgated
test
method
(
36
FR
24888)
did
not
include
the
analysis
of
the
impinger
portion
of
the
sampling
train.
To
accommodate
the
change
in
the
test
method,
EPA
made
adjustments
in
the
promulgated
emission
limits
to
reflect
the
change
in
the
test
method.
The
EPA
made
adjustments
of
up
to
50
percent
in
the
promulgated
emission
limitations
to
reflect
the
measurement
of
only
the
filterable
portion
of
the
emissions.
EPA
recognized
in
setting
several
subsequent
NSPS
that
the
source
test
method
used
to
determine
compliance
with
the
particulate
matter
emissions
limits
measured
only
part
of
the
total
particulate
matter
emitted
by
the
applicable
sources.
This
recognition
was
published
on
October
6,
1975,
in
the
promulgated
Revisions
to
Performance
Testing
Methods
(
40
FR
46250).
Similarly,
EPA
acknowledged
this
in
the
proposal
preamble
to
Subpart
CC­
Standards
of
Performance
for
Glass
Manufacturing
Plants
(
6/
15/
79)
in
the
section
"
Selection
of
Performance
Test
Methods:
The
use
of
EPA
Reference
Method
5
­
Determination
of
Particulate
Emissions
from
Stationary
Sources."
In
developing
the
NSPS
emission
limitations,
it
is
evident
that
only
a
portion
of
the
particulate
matter
emissions
were
considered.
As
a
result,
the
test
methods
that
EPA
selected
for
determining
compliance
with
these
emission
limitations
measured
only
that
same
portion
of
the
particulate
matter
emissions.
It
was
recognized
that
these
test
methods
were
not
suitable
for
quantifying
the
total
emissions
to
the
atmosphere
and
that
the
impinger
portion
of
the
sampling
train
contained
the
missing
portion
of
the
particulate
matter
emissions.
On
December
17,
1990,
EPA
promulgated
Method
202
in
Appendix
M
of
40
CFR
Part
51
(
56
FR
65433)
to
provide
a
method
for
States
to
use
to
analyze
the
impinger
(
or
"
back
half")
content
of
PM
emissions
and
provide
a
measure
of
the
condensable
particulate
emissions.
The
principal
procedures
in
Method
202
improved
upon
the
original
Method
5
back
half
analysis
proposed
in
1971.
In
developing
this
measurement
method,
EPA
consulted
with
several
State
and
local
agencies
and
incorporated
several
options
to
simplify
or
accommodate
existing
policies
and
source
testing
methodologies
for
condensable
particulate
matter.
We
believe
that
by
excluding
the
optional
components,
the
use
of
EPA
Method
202,
combined
with
EPA
Method
5
or
EPA
Method
17,
provides
a
reasonable
indication
of
total
particulate
matter
emissions
for
the
majority
of
stationary
emission
sources.
However,
the
combination
of
EPA
Method
5
and
Method
202
measures
particulate
matter
that
is
larger
than
2.5
micrometers
in
aerodynamic
diameter,
and
will
not
provide
a
reasonable
measurement
of
the
emissions
of
PM2.5.
Methods
are
available
that
can
separate
particulate
matter
by
aerodynamic
size.
On
April
17,
1990,
EPA
promulgated
EPA
Method
201
and
Method
201A
to
provide
a
source
test
method
that
separated
filterable
particulate
matter
greater
than
10
micrometers
from
filterable
particulate
matter
equal
or
smaller
than
10
micrometers.
The
single
cyclone
used
in
these
methods
replaced
the
nozzle
of
EPA
Method
17
to
separate
the
two
size
classes
of
filterable
particulate.
This
method
allows
sources
to
determine
their
emissions
of
filterable
PM10
when
there
are
size
specific
emission
limits
or
when
there
is
a
need
for
size
specific
emission
inventories.
With
the
addition
of
a
second
smaller
cyclone
following
the
single
cyclone
of
EPA
Method
201A,
the
filterable
particulate
can
be
separated
into
three
size
classifications.
These
classifications
include
filterable
particulate
matter
greater
than
10
micrometers,
filterable
particulate
matter
equal
or
smaller
than
10
micrometers
but
greater
than
2.5
micrometers,
and
filterable
particulate
matter
equal
or
smaller
than
2.5
micrometers.
This
method
is
posted
as
Conditional
Method
40
(
CTM
40)
on
EPA's
Emission
Measurement
Centers
web
page
at
http://
www.
epa.
gov/
ttn/
emc/
ctm.
htm.
Of
the
methods
mentioned
previously,
the
most
reliable
measurement
of
total
direct
PM2.5
emissions
would
combine
the
use
of
Conditional
Method
40
with
EPA
Method
202.
The
acceptability
of
a
source
using
the
existing
SIP
test
methods
for
filterable
particulate
matter
as
an
indication
of
the
source's
relative
performance
in
controlling
PM2.5
emissions
would
depend
on
the
source's
level
of
condensable
particulate
matter
emissions
in
relation
to
filterable
PM
emissions,
the
proportion
of
filterable
particulate
matter
that
is
smaller
than
2.5
micrometers,
the
add­
on
PM
control
device
effectiveness,
and
the
need
to
consider
limiting
the
emissions
of
the
condensable
material.

B.
The
existing
PM
test
methods
and
the
emission
limits
based
upon
these
methods
have
been
acceptable
since
1971,
why
do
they
need
to
be
changed
for
PM2.5?
Several
changes
have
occurred
over
the
last
30
years
that
have
gradually
eroded
the
predictive
capabilities
of
particulate
matter
source
test
methods
used
in
most
SIPs
to
evaluate
the
sources
performance
in
controlling
the
pollutant
measured
by
the
ambient
air
quality
test
method.
In
the
1970'
s
and
early
1980'
s,
the
ambient
air
quality
test
method
quantified
the
total
particulate
matter
suspended
in
the
ambient
air.
At
the
beginning
of
this
period,
particulate
matter
control
measures
were
relatively
poor.
Additionally,
most
of
particulate
matter
control
measures
applied
over
the
last
30
years
have
focused
on
filterable
particulate
matter.
While
some
control
measures
for
other
air
pollutants
also
resulted
in
collateral
reductions
in
condensable
particulate
and
particulate
precursor
emissions,
these
reductions
were
relatively
small.
As
a
result,
the
relative
amount
of
sulfates,
nitrates
and
condensed
organic
matter
in
the
ambient
air
particulate
matter
was
proportionally
greater
in
the
1980'
s
than
it
was
in
the
1970'
s.
The
promulgation
of
the
PM10
NAAQS
in
1987
resulted
in
further
reductions
in
filterable
PM
from
sources,
but
there
were
few
nonattainment
areas
where
control
of
the
condensable
constituents
of
PM10
was
required
in
order
to
achieve
attainment.
As
a
result,
stationary
source
control
measures
that
addressed
only
the
filterable
component
of
particulate
matter
were
generally
adequate
to
achieve
the
PM10
NAAQS.
With
the
promulgation
of
the
PM2.5
NAAQS
in
1997
and
associated
ambient
air
quality
monitoring,
speciation
analyses
of
PM2.5
show
that
a
substantial
portion
of
PM2.5
consists
of
sulfates,
nitrates
and
organic
carbon.
These
constituents
are
also
a
substantial
portion
of
the
condensable
particulate
matter
collected
from
stationary
sources.
With
the
increased
application
of
increasingly
efficient
filterable
particulate
matter
control
measures,
condensable
emissions
have
become
a
larger
percentage
of
overall
PM2.5
emissions
for
several
stationary
source
categories.
Based
upon
the
particle
size
distribution
presented
in
Table
1.1­
6
of
AP­
42,
29
percent
of
the
total
filterable
particulate
matter
is
filterable
PM2.5.
As
a
result,
71
percent
of
the
total
PM2.5
emissions
would
be
condensable
PM.
Since
filterable
particulate
matter
emissions
controls
have
improved
since
1971
and
since
most
sources
achieve
substantially
lower
emissions
than
required
by
State
and
Federal
emissions
limitations,
and
condensable
emissions
have
generally
not
been
significantly
reduced,
the
significance
of
the
condensable
emissions
as
a
proportion
of
direct
PM2.5
emissions
may
be
greater
than
indicated
above.
A
test
method
that
measures
total
filterable
particulate
matter,
commonly
including
mostly
particles
larger
than
PM2.5
and
yet
excluding
condensable
emissions,
is
a
poor
indicator
of
source
performance
at
reducing
PM2.5
emissions.

C.
What
methods
are
available
for
measuring
PM
size
and
condensable
PM
from
stationary
sources?
EPA
has
adopted
one
of
several
methods
that
are
available
for
classifying
particulate
matter
by
aerodynamic
diameter.
The
method
adopted
is
based
upon
the
use
of
centrifugal
forces
created
in
cyclones
to
separate
particulate
matter
into
two
aerodynamic
size
classifications.
The
cyclone
specified
in
EPA
Method
201
and
201A
separates
particulate
matter
with
a
nominal
aerodynamic
diameter
greater
than
10
micrometers
from
the
remaining
particulate
matter.
The
addition
of
a
second
smaller
cyclone
following
the
EPA
Method
201A
cyclone
as
is
specified
in
EPA
Method
CTM
40
separates
the
particulate
matter
that
has
an
aerodynamic
diameter
greater
than
2.5
micrometers
from
the
remaining
particulate
matter.
A
filter
follows
the
final
cyclone
of
these
particle
sizing
methods
to
collect
the
smaller
material.
Under
EPA's
source
test
methods
to
separate
PM
based
on
particle
size,
both
of
the
cyclones
and
the
filter
are
maintained
at
the
flue
gas
temperature.
Therefore,
any
material
that
is
in
a
vapor
state
in
the
flue
gas
but
would
be
condensed
as
a
result
of
dilution
and
cooling
when
released
to
the
ambient
air
will
not
be
measured
by
these
particle
sizing
methods.
Vapors
that
would
condense
to
form
particulate
matter
in
the
ambient
air
can
be
quantified
by
EPA
Method
202.
The
EPA
Method
202
is
intended
for
use
in
conjunction
with
a
filterable
particulate
matter
test
method
such
as
Method
201A
or
CTM
40.
Impingers
containing
cold
water
are
used
by
most
methods
to
condense
water
vapor
for
determining
the
flue
gas
moisture
content.
Besides
condensing
water
vapor
in
the
flue
gas,
organic
and
inorganic
chemical
vapors
are
also
condensed
in
these
impingers.
In
EPA
Method
202,
the
organic
and
inorganic
vapors
condensed
in
the
impingers
are
separated
with
an
organic
solvent
and
weighed
after
evaporating
the
water
and
organic
solvent
used
for
separation.
As
recommended
by
the
National
Academy
of
Sciences,
EPA
and
others
are
developing
dilution
based
source
test
methods
for
collecting
and
analyzing
PM2.5.
Rather
than
condensing
vapors
in
chilled
water,
cool
filtered
dilution
air
condenses
the
vapors
prior
to
collection
on
filters.
In
the
new
method
under
development
by
EPA,
particulate
matter
is
sized
using
the
same
cyclones
used
in
CTM
40.
However,
the
in­
stack
filter
used
in
CTM
40
is
removed
so
that
all
of
the
PM2.5
particulate
matter
is
collected
at
near
ambient
temperature
on
the
filters.

D.
Why
is
a
new
dilution­
based
test
method
being
developed
by
EPA?
The
use
of
dilution­
based
particulate
matter
sampling
offers
several
advantages
over
the
combination
of
EPA
Method
CTM
40
and
Method
202.
One
advantage
is
that
the
vapors
are
condensed
and
chemical
reactions
occur
in
a
manner
similar
to
when
stack
gas
is
released
to
the
atmosphere.
As
a
result,
the
potential
for
particulate
matter
formation
that
may
occur
in
water
but
would
not
occur
in
air
is
eliminated.
Another
advantage
is
that
the
potential
for
losing
particulate
matter
during
the
evaporation
of
the
impinger
water
is
eliminated.
With
the
use
of
multiple
filter
types,
the
use
of
dilution
sampling
methods
will
allow
for
the
speciation
of
the
collected
PM2.5
by
the
same
methods
used
for
speciation
of
ambient
air
particulate
matter.
Additionally,
dilution­
based
methods
allow
for
the
measurement
of
the
particle
size
distribution
of
the
particulate
matter
smaller
than
2.5
micrometers.
This
can
be
accomplished
by
modifying
the
hardware
of
the
sampling
equipment
to
extend
the
residence
time
of
the
sampled
particulate
matter.
The
extra
residence
time
allows
the
ultrafine
particulate
matter
initially
formed
during
vapor
condensation
to
grow
toward
its
ultimate
particle
size
distribution.

E.
What
types
of
sources
should
use
the
new
dilution­
based
test
method?
The
new
dilution­
based
test
method
would
be
appropriate
for
most
sources.
Sources
with
very
complex
flue
gas
characteristics
(
e.
g.,
having
several
acidic
and
alkali
gases
with
semi­
volatile
organic
matter)
and
those
sources
that
want
to
generate
a
speciation
profile
specific
to
their
facility
should
use
the
new
dilution
test
method.
Sources
with
very
low
PM2.5
emission
concentrations
and
low
SO2
and
NOX
emission
concentrations
also
may
wish
to
use
the
new
dilution
method.
However,
the
more
complex
operation
and
increased
size
of
the
equipment
associated
with
the
new
method
may
persuade
some
sources
to
use
an
alternative
method.
Sources
where
the
flue
gas
is
near
ambient
temperature
or
where
the
sampled
gas
can
be
cooled
to
near
ambient
temperature
could
use
CTM
40
or
its
equivalent
to
quantify
PM2.5
emissions.
Sources
with
less
complex
flue
gas
characteristics
may
want
to
use
CTM
40
combined
with
EPA
Method
202.

F.
What
are
the
main
features
of
the
new
test
method?
The
main
features
of
the
new
test
method
are
in
the
areas
of
sample
extraction,
particle
sizing,
sample
flow
rate
measurement,
dilution
air
conditioning,
dilution
air
flow
rate
measurement,
sample
mixing
with
dilution
air
and
sample
filtration.
An
additional
major
feature,
where
particulate
speciation
is
desired,
is
the
method
of
extracting
an
aliquot
of
the
diluted
sample.
Flue
gas
is
extracted
isokinetically
at
a
flow
rate
that
produces
particulate
matter
sizing
at
10
and
2.5
micrometers
by
the
two
in­
stack
cyclones.
The
sampled
flue
gas
and
the
PM2.5
particulate
matter
is
extracted
from
the
stack
prior
to
dilution
and
cooling
with
ambient
air
that
has
been
conditioned
by
removing
excess
moisture
and
ambient
particulate
matter
with
a
HEPA
filter.
The
objective
of
all
the
methods
is
to
achieve
complete
mixing
prior
to
filtration
and
to
minimize
sample
losses
on
the
internal
surfaces
of
the
hardware.
The
PM2.5
is
removed
from
the
diluted
sample
gas
by
a
Teflon
filter.
The
PM2.5
deposited
on
the
internal
surfaces
of
the
hardware
is
quantitatively
recovered
with
acetone.
Both
the
Teflon
filter
and
the
PM2.5
recovered
from
the
internal
surfaces
of
the
sampler
are
weighed.
When
speciation
of
the
PM2.5
is
desired,
aliquots
of
the
diluted
sample
gas
are
extracted
for
collection
on
filters.
The
ambient
air
speciation
criteria
are
followed
with
respect
to
the
filter
media
used
and
analytical
finish
of
the
three
filters.

G.
What
is
the
schedule
for
finalization
of
the
new
test
method?
We
have
posted
the
dilution­
based
PM2.5
source
test
method
on
the
TTN
web
as
"
Conditional
Test
Method
39"
and
expect
that
this
method
will
provide
the
basis
for
a
40
CFR
Part
51,
Appendix
M
method
to
be
proposed
at
a
later
date.
Beyond
proposing
the
EPAdeveloped
dilution
test
method,
we
may
identify
the
use
of
a
source
test
method
developed
by
a
national
voluntary
consensus
standard
setting
organization.
Public
Law
104­
113,
also
known
as
the
National
Technology
Transfer
and
Advancement
Act
(
NTTAA),
requires
that
we
use
technical
standards
that
are
developed
or
adopted
by
voluntary
consensus
standards
bodies
as
a
means
to
carry
out
policy
objectives
where
appropriate.
The
law
also
requires
us
to
consult
with
such
bodies
when
it
is
in
the
public
interest
to
participate
with
them
in
the
development
of
technical
standards.
On
April
23,
2003
the
ASTM
Source
and
Ambient
Atmospheres
Committee
started
Work
Item
WK752
to
develop
a
PM2.5
source
test
method
similar
to
the
method
we
are
developing.
EPA
is
an
active
participant
in
this
committees
activities
for
this
work
item.
We
believe
that
it
is
in
the
interest
of
the
public
and
the
Agency
to
participate
in
the
ASTM
process
of
developing
a
PM2.5
source
test
method.
While
we
cannot
predict
when
an
ASTM
standard
will
be
available
and
whether
it
will
be
a
suitable
test
method
for
EPA
to
specify
for
use
in
SIPs,
we
expect
to
make
a
decision
on
the
final
test
method
by
December
2004.
