Date:
4/
28/
03
Subject:
Butane
Blending
Technical
Analysis
From:
Stuart
Romanow
To:
Docket
EPA
is
considering
a
modification
to
its
regulations
which
would
allow
blending
of
butane
into
reformulated
gasoline
(
RFG)
outside
of
the
high
ozone
season,
with
relaxed
sampling
and
testing
requirements.
This
modification
would
make
butane
blending
into
finished
RFG
at
terminals
a
more
feasible
option.
This
memorandum
examines
the
potential
effects
of
blending
butane
into
RFG.
The
analysis
contained
in
this
memorandum
supports
the
following
conclusions:

When
butane
is
blended
into
RFG,
the
resultant
blend
is
likely
to
conform
to
the
toxics
and
NOx
performance
standards
applicable
to
non­
VOC­
controlled
RFG.
Performance,
as
measured
by
the
applicable
compliance
model,
the
winter
complex
model,
may
actually
improve
after
butane
blending.

Blending
butane
into
RFG
is
likely
to
substantially
increase
the
Reid
Vapor
Pressure
(
RVP)
of
the
gasoline,
particularly
when
the
"
before
blending"
gasoline
has
an
RVP
characteristic
of
VOC­
controlled
gasoline.

The
winter
complex
model
does
not
consider
the
effects
of
RVP
on
emissions
performance.
Increases
in
RVP,
in
addition
to
causing
increases
in
non­
exhaust
VOC
and
non­
exhaust
toxics
emissions,
may
also
cause
increases
in
exhaust
emissions,
including
NOx.
Therefore,
the
winter
model
may
not
fully
characterize
the
adverse
emissions
impacts
of
butane
blending.

Although
the
potential
for
adverse
environmental
impacts
resulting
from
such
emissions
increases
is
small,
the
potential
can
be
further
reduced
if
butane
is
not
blended
into
RFG
at
terminals
during
the
shoulder
seasons.

Analysis
RFG
performance
is
evaluated
by
computing
emission
reductions
relative
to
a
baseline
gasoline,
using
a
specified
compliance
model.
The
Phase
II
complex
model,
the
applicable
compliance
model
for
Phase
II
RFG
is
described
in
40
CFR
80.45.
There
are
two
versions
of
this
model;
a
summer
model
used
to
determine
the
compliance
of
VOC­
controlled
gasoline
required
during
the
summer
high
ozone
season,
and
a
winter
model
used
for
non­
VOC­
controlled
gasoline.
Certain
fuel
properties
are
input
parameters
for
the
complex
model.
The
parameter
values
for
the
summer
and
winter
baseline
gasolines
are
also
specified
in
the
regulation.
Non­
VOC­
controlled
Phase
II
RFG
per­
gallon
performance
standards
for
toxics
and
NOx
respectively,
are
20%
and
0.0%
reductions.
The
averaged
standards
for
toxics
and
NOx
are
21.5%
and
1.5%.
In
order
to
demonstrate
that
RFG
will
continue
to
comply
with
these
standards
after
butane
blending,
sets
of
properties
representative
of
RFG
before
blending
were
selected,
property
changes
from
butane
blending
were
estimated,
and
the
winter
complex
model
was
used
to
confirm
that
the
resultant
blends
still
comply
with
these
standards.
Both
"
winter
gasoline"
and
"
summer
gasoline"
before­
blending
properties
were
selected.
Although
the
relaxed
sampling
and
testing
will
apply
to
blending
of
butane
into
RFG
only
outside
of
the
VOC
season,
these
summer
properties
may
be
representative
of
gasoline
at
terminals
at
the
close
of
the
VOC
season.

Calendar
year
2000
RFG
Survey
data
were
used
to
determine
representative
sets
of
RFG
properties.
Property
averages
for
Sussex
County
Delaware
were
selected
as
parameters
for
this
analysis.
Although
Sussex
County's
average
RFG
performance
was
substantially
better
than
performance
standards,
the
area
ranked
at
or
near
the
bottom
of
areas
surveyed
in
year
2000
for
both
winter
and
summer
toxics
and
NOx
performance.
Parameter
values
are
shown
in
table
1,
below:

Table
1­
Sussex
County,
DE
Average
Property
Values­
Year
2000
RFG
Surveys
Summer
2000
Winter
2000
Total
oxygen
(
wt%)
2.01
2.00
MTBE
(
wt%
oxygen)
1.65
1.47
ETBE
(
wt%
oxygen)
Ethanol
(
wt%
oxygen)
TAME
(
wt%
oxygen)
0.33
0.50
SULFUR
(
ppm)
134
225
RVP
(
psi)
6.43
E200
(%)
45.5
53.5
E300
(%)
81.6
82.8
AROMATICS
(
vol%)
24.87
20.33
OLEFINS
(
vol%)
11.04
12.03
BENZENE
(
vol%)
0.730
0.652
Although
TAME
contributed
a
substantial
portion
of
total
oxygen
and
small
quantities
of
oxygenates
other
than
MTBE
and
TAME
were
present,
it
was
assumed
for
this
analysis
that
all
oxygen
was
from
MTBE
(
i.
e.
total
oxygen
weight
was
assigned
to
MTBE
for
complex
model
calculatios.)
This
has
some
effect
on
complex
model
toxics,
but
none
on
NOx.
The
winter
RFG
surveys
do
not
measure
RVP,
since,
for
compliance
determinations,
RVP
is
set
to
a
value
of
8.7
psi
in
the
winter
complex
model.
Data
from
the
Alliance
of
Automobile
Manufacturers
(
AAM
)
winter
surveys
for
year
2000
were
used
to
select
an
RVP
value
for
the
winter
gasoline.
Since
Philadelphia,
PA
is
an
RFG
area
geographically
close
to
Sussex
County,
DE,
13.3
psi,
the
average
RVP
value
for
regular
grade
gasoline
from
the
AAM
Philadelphia
survey,
was
selected.
1
Sussex
County
had
the
lowest
average
RVP
of
areas
surveyed,
so
8
v%
butane
blending
is
about
the
maximum
amount
of
butane
that
could
be
blended
to
meet
the
11.5
psi
class
C
limit.
See
http://
www.
epa.
gov/
otaq/
regs/
fuels/
rfg/
properf/
rfgperf.
htm
for
a
survey
data
summary.

2
The
potential
effect
of
butane
blending
on
the
emissions
performance
of
gasoline
is
largely
due
to
the
impurities
that
may
be
present
in
the
butane.
For
example,
if
butane
with
high
sulfur
and
olefin
content
were
blended
with
gasoline
with
lower
sulfur
and
olefin
content,
the
resultant
blend
could
have
poorer
NOx
emission
performance
than
the
before­
blending
gasoline.
Consequently,
the
non­
commercial
olefin
maximum
was
used
in
this
analysis,
to
represent
a
"
worst
case"
condition.
The
olefin
maximum
does
not
change
after
January
1,
2004,
however
the
sulfur
maximum
decreases
from
140
ppm
to
30
ppm.
Since
this
rule
is
unlikely
to
become
final
in
time
to
have
any
substantial
impact
prior
to
2004,
the
30
ppm
value
was
used.
It
was
assumed
that
RVP
would
blend
linearly
on
mole­
fraction
basis
when
butane
was
blended
with
gasoline.
A
relative
density
of
0.73
and
a
molecular
weight
of
105
were
assumed
for
the
gasoline
before
blending.
Based
on
a
table
of
physical
constants
of
paraffins
51.7
psi
was
used
as
the
butane
RVP,
since
this
was
the
listed
vapor
pressure
at
100
F
for
n­
butane.
With
these
assumptions,
if
butane
is
blended
at
3
volume
percent
with
13.3
psi
gasoline,
the
resultant
blend
RVP
is
around
14.9
psi.
According
to
ASTM
Standard
D­
4814,
the
highest
winter
gasoline
vapor
pressure/
distillation
class
for
Delaware
is
E,
with
a
maximum
RVP
of
15.0
psi.
Therefore,
blending
butane
at
3
volume
percent
into
this
winter
gasoline
should
be
a
representative
blending
case.
The
highest
ASTM
volatility
class
standard
for
Delaware
during
the
latter
part
of
September
is
C,
with
a
maximum
RVP
of
11.5
psi.
With
the
above
blending
assumptions,
if
butane
is
blended
at
8
volume
percent
with
6.43
psi
RVP
gasoline,
the
resultant
blend
RVP
is
around
11.5
psi,
consequently
blending
butane
at
8
v%
into
this
summer
gasoline
should
also
be
a
representative
blending
case.
1
Other
properties
of
the
blend
were
calculated
assuming
linear
volumetric
blending
for
aromatics,
olefins,
benzene,
E200
and
E300,
and
gravimetric
blending
for
sulfur
and
oxygen.
It
was
also
assumed
that
the
butane
used
for
blending
had
some
olefin,
aromatic
and
sulfur
content.
The
values
for
these
parameters
are
the
maximums
for
noncommercial
grade
butane,
applicable
after
January
1,
2004.2
Tables
2
and
3
show
the
before
and
after
blending
property
values
for
these
two
cases.
Table
2­
Properties
Before
and
After
Butane
Blending
at
3
v%

Winter
gas
Gasoline
Butane
After
blending
3v%

MTBE
(
wt%
oxygen)
2.00
0
1.95
ETBE
(
wt%
oxygen)
0
Ethanol
(
wt%
oxygen)
0
TAME
(
wt%
oxygen)
0
SULFUR
(
ppm)
225
30
220
RVP
(
psi)
13.3
51.7
14.9
E200
(%)
53.5
100
54.9
E300
(%)
82.8
100
83.3
AROMATICS
(
vol%)
20.33
2
19.78
OLEFINS
(
vol%)
12.03
10
11.97
BENZENE
(
vol%)
0.652
0.03
0.633
relative
density
0.73
0.584
molecular
weight
105
58
Table
3­
Properties
Before
and
After
Butane
Blending
at
8
v%

Summer
gas
Gasoline
Butane
after
blending
8
v%

MTBE
(
wt%
oxygen)
2.01
0
1.88
ETBE
(
wt%
oxygen)
0
Ethanol
(
wt%
oxygen)
0
TAME
(
wt%
oxygen)
0
SULFUR
(
ppm)
134
30
127
RVP
(
psi)
6.43
51.7
11.5
E200
(%)
45.5
100
49.9
E300
(%)
81.6
100
83.1
AROMATICS
(
vol%)
24.87
2
23.04
OLEFINS
(
vol%)
11.04
10
10.96
BENZENE
(
vol%)
0.730
0.03
0.674
relative
density
0.73
0.584
molecular
weight
105
58
The
winter
complex
model,
used
for
compliance
calculations
outside
of
the
high
ozone­
3
The
RVP
levels
of
the
fuels
in
the
database
used
to
develop
the
complex
model
ranged
from
6.5
to
10
psi,
lower
than
the
typical
range
for
winter
fuels.
Since
data
on
the
exhaust
emission
effects
of
fuels
with
winter
RVP
levels
under
winter
conditions
were
very
limited,
EPA
could
not
model
the
effects
of
winter
RVP
levels
on
exhaust
emissions.
Consequently,
although
RVP
is
likely
to
have
some
effect
on
wintertime
exhaust
emissions,
EPA
opted
to
assume
no
effect
in
the
winter
complex
model.

4
The
winter
model
also
assumes
that
non­
exhaust
VOC
emissions
differences,
which
would
be
affected
by
RVP
differences
as
well
as
ambient
temperatures,
are
zero.

5
As
noted,
Sussex
County
Delaware
RFG
ranked
near
the
bottom
in
average
NOx
and
toxics
emissions
performance
in
year
2000.
Both
the
sulfur
and
olefin
content
of
the
RFG
in
this
analysis
exceeded
the
levels
in
the
butane,
so
blending
reduced
these
parameters,
with
a
beneficial
effect
on
NOx.
If
this
impure
butane
were
blended
into
RFG
with
lower
sulfur
and/
or
olefin
content
than
the
butane,
these
parameter
values
in
the
blend
could
be
higher
than
the
beforeblending
gasoline
values.
This
may
result
in
an
increase,
rather
than
a
decrease
in
NOx
after
blending.
However,
gasoline
with
low
sulfur
and
olefin
content
would
be
expected
to
have
superior
NOx
performance.
Thus,
blending­
related
increases
in
NOx
emissions,
as
measured
by
the
winter
complex
model,
would
be
expected
only
where
there
is
extreme
overcompliance
in
the
before­
blending
RFG.

6
Although
this
analysis
addressed
butane
blending
into
RFG,
the
same
general
conclusion
holds
for
blending
into
conventional
gasoline;
i.
e.
butane
blending
is
likely
to
reduce
winter
season,
is
used
with
a
default
RVP
value
of
8.7
psi
for
both
the
winter
baseline
fuel
and
the
fuel
being
compared
to
the
baseline.
Use
of
the
same
RVP
for
both
fuels,
rather
than
the
actual
RVP,
zeros
out
the
effects
of
RVP
change
on
exhaust
emissions.
3
The
winter
model
ignores
nonexhaust
benzene
emissions,
which
are
a
component
of
total
toxics
in
the
summer
model.
4
Using
the
winter
complex
model
to
evaluate
the
performance
of
the
before
and
after­
blending
formulations
shows
that
these
after­
blending
formulations
not
only
still
comply
with
the
toxics
and
NOx
performance
standards,
but
that
the
performance
for
both
toxics
and
NOx
improves.
5
This
is
shown
in
table
4,
below.
(
Reductions
from
baseline
fuel
emissions
are
shown
as
negative
numbers,
so
larger
negative
numbers
denote
better
performance.)

Table
4­
Winter
Complex
Model
Performance
(
Both
cases)

Winter
Complex
Model
Performance
(%
change
from
baseline)
Using
8.7
psi
RVP
Before
blending
Blending
@
3v%
Blending
@
8
v%
winter
gas
Toxics
(
exhaust)
­
24.81
­
25.58
not
calculated
NOx
­
4.86
­
5.15
not
calculated
summer
gas
Toxics
(
exhaust)
­
22.15
not
calculated
­
24.60
NOx
­
8.13
not
calculated
­
8.60
This
shows
that
butane
blending
at
the
terminal
is
unlikely
to
result
in
non­
VOC
controlled
RFG
which
fails
to
comply
with
toxics
or
NOx
performance
standards.
6
However,
the
above
complex
model
exhaust
toxics
and
NOx
emissions.
analysis
ignores
any
effect
that
the
increase
in
RVP
may
have
on
emissions
Therefore,
one
cannot
unequivocally
conclude
that
toxics
and
NOx
emissions
will
not
increase
as
a
result
of
butane
blending
into
RFG.
To
further
investigate
if
any
adverse
emission
effect
(
even
one
that
would
not
affect
compliance)
is
likely
as
a
result
of
butane
blending,
the
summer
complex
model,
which
considers
RVP
effects,
was
used
to
evaluate
the
"
summer
gas"
case.
Results
are
shown
in
table
5,
below:

Table
5­
Summer
Complex
Model
Performance
(
8v%
case)

Summer
Complex
Model
Performance
(%
change
from
Baseline)
Before
blending
Blending
@
8v%
Summer
gas
exhaust
VOC
­
16.37
­
2.17
non­
exhaust
VOC
­
49.86
139.46
total
VOC
­
28.15
47.65
exhaust
toxics
­
26.95
­
28.03
total
toxics
­
29.87
­
28.53
NOx
­
7.62
­
6.66
The
summer
complex
model
predicts
that
exhaust
toxics
performance
will
improve
after
blending,
but
that
NOx
performance
will
get
worse.
The
model
predicts
that
non­
exhaust
VOCs,
which
are
a
function
of
RVP,
will
increase
substantially,
and
that
total
toxics
will
increase
slightly
(
Non­
exhaust
benzene
emissions
increase
because
of
the
RVP
increase.)
Exhaust
VOC
emissions
increase
as
well.

Alternative
NOx
models,
which
EPA
developed
to
evaluate
California's
RFG
oxygenate
waiver
request,
confirm
that
NOx
emissions
could
increase
after
butane
blending.
These
results
are
shown
in
table
6,
below:

Table
6­
Alternative
NOx
Model
Emission
Estimates
Before
and
after
Blending
(
8v%
case)

Alternative
NOx
Model
Results­
Before
and
After
Blending
Model
2
Model
3
EPA­
3
Model
5
Model
6
Model
7
average
Summer
gas
before
blending
(
gm/
mile)
0.75892
0.75201
0.75364
0.75828
0.75973
0.75674
After
blending
8v%
butane
(
gm/.
mile)
0.79797
0.80409
0.79496
0.79049
0.79468
0.78184
%
change
before
to
after
5.15%
6.93%
5.48%
4.25%
4.60%
3.32%
4.95%

However,
in
general,
these
models
also
indicate
that
the
NOx
emissions
of
the
after­
blending
formulations
are
lower
than
the
emissions
of
the
1990
winter
baseline
fuel
as
measured
by
the
same
models.
Consequently,
these
models
indicate
that
the
NOx
performance
after
blending,
even
considering
the
effect
of
RVP
increase,
is
likely
to
be
as
good
as
or
better
than
1990
winter
7
The
actual
RVP
for
winter
baseline
gasoline
(
11.5
psi)
was
used
in
these
calculations.
These
alternative
models
use
T50
and
T90
distillation
parameters.
T50
and
T90
values
were
200F
and
333F
for
the
winter
baseline,
189.8F
and
328.0F
for
the
3%
blend,
and
200.1F
and
329.1F
for
the
8%
blend.
baseline
gasoline,
consistent
with
the
intent
of
the
RFG
regulations
7:

Table
7­
Alternative
Model
Comparisons
to
Winter
Baseline
Emissions
(
both
cases)

Alternative
NOx
Model­
Comparison
of
After­
Blending
and
90
Baseline
Model
2
Model
3
EPA­
3
Model
5
Model
6
Model
7
average
90
winter
bl
gasoline
0.85733
0.85246
0.85396
0.84416
0.84878
0.85196
Blending
at
3%
into
winter
gas
0.84920
0.87024
0.85078
0.83688
0.84370
0.82471
%
change
from
winter
baseline
­
0.95%
2.09%
­
0.37%
­
0.86%
­
0.60%
­
3.20%
­
0.65%

Blending
at
8%
into
summer
gas
0.79797
0.80409
0.79496
0.79049
0.79468
0.78184
%
change
from
winter
baseline
­
6.92%
­
5.67%
­
6.91%
­
6.36%
­
6.37%
­
8.23%
­
6.74%

These
results
should
be
viewed
with
some
caution.
It
is
uncertain
that
either
the
complex
model
or
any
of
these
alternative
NOx
models
accurately
predicts
the
effect
of
RVP's
in
this
range
on
NOx
emissions.
The
upper
limit
of
the
acceptable
RVP
range
for
the
complex
model
is
10.0
psi
for
RFG
and
11.0
for
conventional
gasoline.
Observations
with
RVP>
10
were
excluded
from
the
data
used
to
develop
these
alternative
NOx
models.

Butane
blending
downstream
of
oxygenate
blending
will
dilute
the
oxygen
content.
Thus,
there
could
be
an
increased
probability
of
oxygen
survey
series
failures.
There
is
also
the
possibility
that
some
RFG
would
fall
below
the
per
gallon
minimum
for
oxygen.
Neither
situation
is
likely
to
have
any
significant
adverse
environmental
effect.

In
summary,
butane
blending
will
have
no
compliance
consequences
for
non­
VOC
­
controlled
RFG,
other
than
its
effect
on
oxygen
content.
However,
blending
butane
into
gasoline,
could
degrade
its
emissions
performance
in
ways
not
considered
in
the
winter
complex
model.
Butane
blending
could
significantly
increase
the
RVP
of
gasoline
at
terminals
at
the
close
of
the
ozone
season.
Consequently,
the
exhaust
and
non­
exhaust
VOC
emissions
performance
of
the
after­
blending
gasoline
would
be
substantially
worse
than
before
blending.
NOx
emissions
performance
is
likely
to
be
adversely
affected,
as
well.
Toxics
emissions
may
also
be
adversely
affected,
because
of
non­
exhaust
benzene
increases
not
included
in
the
winter
model,
but
butane
blending
also
benefits
exhaust
toxics
emission
performance
(
e.
g.
through
dilution
of
benzene
and
aromatics
content)
offsetting
any
"
uncounted"
non­
exhaust
increase.
8
The
Mobile
Source
Air
Toxics
rule
will
help
maintain
RFG's
toxics
overcompliance
and
the
Tier
2
sulfur
requirements
should
maintain
or
increase
NOx
overcompliance.
Since
butane
blending
will
not
contribute
to
noncompliance
with
performance
standards
outside
of
the
VOC
control
season,
and
since
there
is
already
substantial
overcompliance
with
both
toxics
and
NOx
performance
standards,
it
is
unlikely
that
any
uncounted
NOx
or
toxics
increases
would
have
a
substantial
adverse
environmental
effect
during
this
time
period.
8
VOC
performance
is
controlled
and
a
stricter
standard
applies
to
NOx
during
the
VOC
control
season
because
these
pollutants,
in
summertime
atmospheric
conditions,
are
involved
in
the
photochemical
reactions
that
form
ozone.
Additionally,
for
a
given
fuel
composition,
summertime
ambient
temperature
profiles
are
likely
to
result
in
greater
non­
exhaust
benzene
emissions
than
wintertime
profiles.

However,
summertime
conditions
can
exist
outside
of
the
VOC
control
season.
Consequently,
the
presence
of
VOC­
controlled
RFG,
which
may
be
in
the
distribution
system,
could
potentially
provide
additional
benefit
relative
to
non­
VOC­
controlled
RFG.
This
benefit
may
be
reduced
if
butane
blending
into
this
RFG
occurs
at
terminals.
Both
the
likelihood
of
VOC­
controlled
RFG
in
the
distribution
system
and
the
likelihood
of
any
additional
benefit
from
this
RFG
would
be
greatest
in
the
shoulder
seasons.
