MEMORANDUM
Date:
August
28,
2003
To:
Kevin
Cavender,
EPA
From:
Jeff
Coburn,
RTI
Subject:
Evaluation
of
Carbon
Adsorption
Systems
to
Control
Organic
HAP
Emissions
from
Cooling
and
Shakeout
Operations
at
Iron
and
Steel
Foundries
An
engineering
evaluation
was
conducted
to
determine
whether
or
not
carbon
adsorption
systems
are
technically
capable
of
meeting
a
20
ppmv
VOC
(
as
hexane)
emission
limit
for
cooling
and
shakeout
operations.
Specifically,
although
two
steel
foundries
use
carbon
adsorption
systems
to
control
emissions
from
cooling
and
shakeout
operations,
no
carbon
adsorption
systems
are
currently
used
at
iron
foundries.
Comments
were
received
noting
that
steel
foundries
do
not
use
seacoal
and,
therefore,
cooling
and
shakeout
emissions
from
iron
foundries
are
different
than
from
steel
foundries.
The
commenters
subsequently
question
the
ability
of
carbon
adsorption
systems
to
meet
a
20
ppmv
VOC
emission
limit
for
cooling
and
shakeout
operations
at
iron
foundries.

Data
from
the
CERP
Mexico
study
(
CERP,
1998)
were
used
to
assess
the
organic
HAPs
emitted
from
green
sand
iron
foundries.
These
data
are
summarized
in
Table
1.
This
foundry
is
an
automotive
foundry
that
uses
green
sand
molds
and
chemically­
bonded
cores.
The
data
from
this
foundry
generally
showed
higher
acetaldehyde
emissions
than
other
available
data
(
an
important
point),
but
generally
agrees
well
with
other
data
of
organic
emissions
from
green
sand
foundries.
Using
these
emission
factors
and
using
a
lower
range
of
exhaust
flow
rates
per
ton
of
metal
poured,
a
reasonable
upper
estimate
of
the
concentration
from
the
combined
cooling
and
shakeout
exhaust
is
50
ppmv
as
hexane.
The
concentration
of
each
constituent
(
in
ppmv
as
hexane)
in
inlet
to
the
carbon
adsorber
is
alo
provided
in
Table
1.

Tables
contained
in
"
Internal
Instruction
Manual
for
ESD
Regulation
Development
­
Carbon
Adsorption
Control
for
Organic
Emissions"
(
U.
S.
EPA,
1993)
indicate
that
acetaldehyde
and
formaldehyde
are
generally
not
well
controlled
by
carbon
adsorption
systems,
but
all
of
the
other
organic
HAP
are
amendable
to
control
with
carbon
adsorption
systems.
Especially
when
other
more
easily
adsorbed
compounds
are
present,
these
other
compounds
will
compete
with
and
displace
acetaldehyde
and
formaldehyde
from
the
adsorption
sites.
Therefore,
it
can
be
expected
that
these
compounds
will
not
be
effectively
controlled
by
a
carbon
adsorption
system.
Other
organics
present
were
assumed
to
be
controlled.
Two
different
efficiencies
were
assessed:
90
percent
and
a
95
percent
control
efficiency.
This
is
the
range
of
expected
removal
efficiencies
for
these
other
chemicals
in
this
concentration
range.
Based
on
these
values,
the
exhaust
concentrations
from
the
carbon
adsorber
were
calculated
(
see
Table
1).
2
Table
1.
Summary
of
Projected
Cooling
and
Shakeout
Emissions
from
Green
Sand
Iron
Foundry
Emission
Factor
lb/
ton
metal
Percent
of
total
VOC
emissions
Projected
Exhaust
Steam
VOC
Concentration
(
ppmv
as
hexane)
for:

Carbon
Adsorber
Inlet
Carbon
Adsorber
Outlet
at
90%
Control
Efficiency
Carbon
Adsorber
Outlet
at
95%
Control
Efficiency
Benzene
0.064
21.26%
10.63
1.06
0.53
Acetaldehyde
0.061
20.39%
10.20
10.20
10.20
Toluene
0.042
14.01%
7.00
0.70
0.35
Phenol
0.034
11.24%
5.62
0.56
0.28
Xylenes
0.030
9.95%
4.97
0.50
0.25
Formaldehyde
0.028
9.18%
4.59
4.59
4.59
POMs
(
naphthalene+)
0.027
9.02%
4.51
0.45
0.23
Cresol
0.015
4.96%
2.48
0.25
0.12
Total
50.00
18.31
16.55
From
the
results
presented
in
Table
1,
it
appears
that
it
is
feasible
for
a
carbon
adsorption
system
to
meet
a
20
ppmv
VOC
(
as
hexane)
emission
limit.
This
evaluation
used
emissions
data
showing
the
highest
acetaldehyde
emissions
of
all
of
the
data
available
for
these
sources.
As
acetaldehdye
is
not
well
controlled
by
carbon
adsorption,
we
believe
that
this
represents
a
reasonable
"
worst­
case"
circumstance
that
could
be
expected
at
a
green
sand
iron
foundry.

Note,
however,
that
foundries
that
use
chemically­
bonded
mold
sand
may
have
significantly
higher
exhaust
stream
concentrations
than
those
expected
from
green
sand
foundries.
We
do
not
have
enough
data
to
indicate
whether
or
not
the
mixture
of
organics
released
from
chemically­
bonded
sand
mold
systems
would
be
significantly
different
from
the
green
sand
mold
system.
We
do
know,
however,
that
at
least
one
iron
foundry
that
uses
chemically­
bonded
mold
sand
employs
a
regenerative
thermal
oxidizer
(
RTO)
that
achieves
the
20
ppmv
exhaust
stream
outlet
concentration.
Information
submitted
by
several
commenters
suggests
that
other
alternatives,
such
as
advanced
oxidation
systems
and
binder
system
reformulations,
can
also
effectively
reduce
cooling
and
shakeout
emissions
from
iron
foundries
to
below
20
ppmv
VOC
as
hexane.

In
summary,
the
20
ppmv
VOC
emission
limit
appears
to
be
a
reasonable
expectation
of
performance
for
cooling
and
shakeout
operations
at
new
sources.
This
level
should
be
achievable
by
a
carbon
adsorption
system
at
most
foundries.
Site­
specific
conditions,
however,
may
require
incineration
or
a
carbon
adsorber
in
combination
with
other
alternative
emission
reduction
techniques
in
order
to
achieve
the
20
ppmv
VOC
emission
limit.
3
References
CERP,
1998.
Foundry
Process
Emission
Factors:
Baseline
Emissions
from
Automotive
Foundries
in
Mexico,
Casting
Emission
Reduction
Program,
McClellan
Air
Force
Base,
California,
November
24,
1998.

U.
S.
Environmental
Protection
Agency.
1993.
Internal
Instruction
Manual
for
ESD
Regulation
Development
­
Carbon
Adsorption
Control
for
Organic
Emissions,
Office
of
Air
Quality
Planning
and
Standards,
Research
Triangle
Park,
NC.
Emissions
Standard
Division.
March
1993.
