
C:\
dmautop\
temp\
dwcgi­
23438­
1077572953­
553683000.
wpd
MEMORANDUM
January
7,
2004
To:
Docket
File
From:
Rebecca
L.
Nicholson
Subject:
Telephone
Conference
between
Rebecca
L.
Nicholson
(
RTI)
and
David
Word
(
NCASI)
regarding
Comments
on
Proposed
Plywood
and
Composite
Wood
Products
MACT
 
NCASI
Method
CI/
WP­
98.01
Dr.
Word
was
contacted
on
July
17,
2003
to
find
out
more
about
NCASI
Method
CI/
WP­
98.01
in
an
effort
to
respond
to
a
comment
about
this
test
method
submitted
by
the
State
of
Minnesota.
Representatives
from
the
State
of
Minnesota's
Pollution
Control
Agency
stated
that
their
experience
with
the
NCASI
Method
CI/
WP­
98.01
(
proposed
for
measurement
of
methanol
and
formaldehyde)
showed
that
sampling
stack
emissions
with
a
high
moisture
content
would
require
a
large
impinger,
isokinetic
version
of
this
method
to
ensure
accuracy
(
See
Docket
No.
A­
98­
44,
item
No.
IV­
D­
04).
Therefore,
they
requested
that
EPA
add
guidance
in
the
final
rule
to
allow
the
use
of
the
isokinetic
version
of
the
NCASI
method
when
needed.
Dr.
Word
provided
the
following
information
about
isokinetic
sampling
to
provide
some
context
for
the
comment
submitted
by
the
State
of
Minnesota.

Isokinetic
sampling
is
designed
to
obtain
a
representative
particulate
matter
sample
within
a
stack
(
i.
e.,
collect
all
of
the
size
fractions
present
in
the
emission
stream).
To
achieve
isokinetic
sampling,
laminar
flow
must
continue
into
the
sampling
probe
such
that
the
velocity
in
the
sampling
probe
is
the
same
as
the
gas
velocity
in
the
stack.
If
the
velocity
in
the
sampling
probe
is
lower
than
the
stack
velocity,
then
some
of
the
smaller
particles
are
forced
around
the
probe
(
i.
e.,
are
not
collected
in
the
sample)
whereas
larger
particles
exhibit
some
inertia
and
are
more
inclined
to
enter
the
sampling
probe.
Under
this
scenario,
the
sample
collected
will
be
biased
towards
the
larger
particles
and
thus
would
not
be
a
representative
sample.
The
reverse
would
also
be
true,
that
is,
samples
would
be
biased
towards
the
smaller
particles
if
the
sampling
probe's
velocity
was
higher
than
the
velocity
of
the
stack
gas.
Because
stack
gas
velocities
are
relatively
high,
isokinetic
sampling
systems
have
fairly
high
flows
(
e.
g.,
20
ft3/
hr)
and
require
larger
impingers
than
non­
isokinetic
sampling
systems.

Isokinetic
sampling
is
generally
not
required
in
test
methods
designed
for
gas­
phase
pollutants
because
the
assumption
is
made
that
the
gases
are
equally
distributed
in
the
emissions
stream/
stack
(
i.
e.
do
not
have
the
momentum
issues
associated
with
particles
in
an
air
stream).
Therefore,
the
sampling
probe
velocity
can
be
lower
when
sampling
for
gaseous
pollutants,
allowing
for
smaller
air
flows
and
impingers.
EPA
Method
M0011,
used
to
measure
formaldehyde,
is
one
exception
to
this
general
rule
regarding
test
methods
for
gaseous
pollutants.
EPA
Method
0011
is
an
isokinetic
method
and
it
does
not
include
a
filter.
Therefore,
any
2
C:\
dmautop\
temp\
dwcgi­
23438­
1077572953­
553683000.
wpd
formaldehyde
that
is
bound
up
in
a
particle
would
be
collected
along
with
the
gas­
phase
formaldehyde.
The
other
test
methods
in
the
proposed
PCWP
rule
(
including
the
NCASI
method
referenced
by
the
commenter
[
State
of
Minnesota])
are
non­
isokinetic.

One
of
the
concerns
with
the
non­
isokinetic
methods
occurs
when
the
gas
stream
being
sampled
is
saturated
with
water
(
e.
g.,
100%
relative
humidity).
In
such
cases,
when
the
temperature
decreases,
the
water
condenses
and
forms
droplets
in
the
air
stream.
These
water
droplets
do
not
exhibit
laminar
flow
(
more
of
a
Brownian
motion
instead),
and
they
can
absorb
gaseous
methanol
and
formaldehyde
in
the
air
stream
such
that
they
become
tiny
"
scrubbers."
Therefore,
in
these
conditions,
the
use
of
a
non­
isokinetic
sampling
system
(
which
won't
capture
a
representative
portion
of
these
droplets)
can
result
in
emission
measurements
that
are
biased
low.

Emissions
streams
associated
with
plywood
and
composite
wood
products
manufacturing
can
sometimes
be
saturated,
particularly
if
the
sampling
location
is
after
a
wet
scrubber
or
wet
ESP.
Therefore,
the
issue
of
water
droplet
formation
can
be
an
issue
for
all
of
the
non­
isokinetic
methods
included
in
the
proposed
PCWP
rule.
For
the
most
part,
the
water
droplet
phenomenon
is
ignored
by
test
contractors
and
State
agencies.
Dr.
Word
noted
that
there
is
no
"
large
impinger,
isokinetic
version"
of
NCASI
Method
CI/
WP­
98.01
and
that
all
of
the
HAP
data
collected
by
NCASI
to
characterize
emissions
from
PCWP
process
units
was
collected
using
NCASI
Method
IM/
CAN/
WP­
99.01,
a
non­
isokinetic
variation
of
Method
CI/
WP­
98.01
which
has
a
similar
sample
train
configuration
and
operation
but
which
includes
a
canister
for
collection
of
non­
water
soluble
gases.
However,
he
also
noted
that
some
test
contractors
NCASI
has
worked
with
have
used
EPA
Method
5
sampling
equipment
with
NCASI
Method
CI/
WP­
98.01
when
they
were
testing
for
methanol
and
formaldehyde
(
only)
and
already
had
the
Method
5
sampling
trains
set
up,
such
that
it
was
more
convenient
for
them
to
use
these
large
impinger,
isokinetic
sampling
systems.
In
these
situations,
spiking
tests
were
run
to
ensure
that
the
larger
impinger
systems
provided
acceptable
results.
However,
NCASI
has
not
attempted
to
develop
and
validate
an
alternative
large
impinger,
isokinetic
version
of
NCASI
Method
CI/
WP­
98.01.
