Memorandum
Date:
February
22,
2006
Subject:
Effect
of
Benzene
Control
on
Gasoline
Quality
From:
Lester
Wyborny,
Engineering
Specialist
Assessment
and
Standards
Division
Office
of
Transportation
and
Air
Quality
U.
S.
Environmental
Protection
Agency
To:
Docket
EPA­
HQ­
OAR­
2005­
0036
The
refinery­
by­
refinery
modeling
tool
used
to
estimate
the
cost
of
benzene
control
for
the
proposal
estimates
the
change
in
benzene
levels,
but
does
not
estimate
the
changes
in
other
fuel
qualities.
1
While
their
effects
are
not
expected
to
be
large,
changes
in
other
fuel
qualities
need
to
be
understood
to
account
for
their
impact
on
emissions.
The
purpose
for
this
note
is
to
summarize
an
analysis
conducted
for
to
assess
the
changes
in
other
gasoline
qualities
associated
with
a
benzene
control
standard.
This
information,
along
with
the
change
in
benzene
levels
in
each
PADD,
was
used
to
estimate
changes
in
total
air
toxics
emissions
in
the
air
quality
analysis.

The
analysis
to
estimate
the
effect
of
benzene
control
on
other
fuel
qualities
relies
on
the
LP
refinery
modeling
work
conducted
by
Abt
Associates.
Based
on
the
capacities
of
gasoline
producing
units
within
refineries,
the
estimated
volumes
of
fuel
produced
and
the
constraints
placed
on
fuel
quality,
the
LP
refinery
model
estimates
the
volumes
of
gasoline
blendstocks
used
to
manufacture
gasoline.
Within
the
LP
refinery
model,
each
gasoline
blendstock
is
assigned
the
Complex
Model
qualities
including
RVP,
percent
benzene,
total
aromatics,
olefins,
and
oxygen,
and
E200
and
E300
(
percent
of
gasoline
which
boils
off
at
200
and
300
degrees
Fahrenheit).
When
blending
up
several
different
gasoline
types,
the
LP
refinery
model
volume­
weights
the
blendstocks
together
in
a
way
that
ensures
that
each
gasoline
type
complies
with
all
specifications
and
standards.
When
the
refinery
model
is
establishing
the
blendstock
mix
for
each
gasoline
type,
it
simultaneously
is
estimating
the
Complex
Model
gasoline
qualities
for
each
gasoline
type.
The
gasoline
types
modeled
by
the
LP
refinery
model
include:

­
Conventional
gasoline,
­
Ethanol
blended
conventional
gasoline,
­
7.8
psi
Low
RVP
gasoline,
­
7.0
psi
Low
RVP
gasoline,
­
Reformulated
gasoline
with
MTBE,
and
­
Reformulated
gasoline
with
ethanol.

1
Consult
Chapter
9
of
the
RIA
to
further
understand
the
refinery
modeling
techniques
used
to
model
benzene
control
conducted
for
this
proposed
rulemaking..
The
LP
refinery
modeling
study
modeled
benzene
reductions
in
summertime
gasoline
in
regions
called
Petroleum
Administrative
Districts
for
Defense
(
PADD).
The
PADDs
modeled
include
PADD
1,
PADD
2,
PADD
3,
and
PADD
4
and
5
together.
For
each
PADD,
the
refinery
model
first
estimates
gasoline
production
in
a
future
year
without
the
benzene
control
 
this
case
is
called
a
reference
case.
A
benzene
control
case
was
modeled
incremental
to
the
reference
case
which
in
this
case
is
an
average
0.50
volume
percent
benzene
standard.
The
change
in
gasoline
qualities
between
the
reference
case
and
benzene
control
case
can
be
used
to
estimate
the
change
in
gasoline
quality
associated
with
benzene
control.
The
benzene
control
standard
did
not
match
the
proposed
benzene
control
standard
so
the
change
in
gasoline
quality
in
the
modeling
runs
could
not
be
used
to
directly
project
changes
in
gasoline
quality
associated
with
the
proposed
benzene
control
standard.
Instead
correlations
were
made
between
benzene
reductions
and
changes
in
other
gasoline
qualities
that
would
be
used
with
the
modeled
benzene
reductions
associated
with
the
proposed
rule.
The
other
regulations
under
which
this
benzene
control
case
was
run
are
consistent
with
the
regulatory
situation
in
2004.
This
scenario
includes
state
MTBE
bans
in
California,
New
York
and
Connecticut,
but
MTBE
use
occurs
in
the
rest
of
RFG
and
on
the
East
and
Gulf
Coast.
Also
in
this
scenario,
there
was
no
renewable
fuels
standard.
This
scenario
is
consistent
with
the
regulatory
assumptions
which
formed
the
basis
for
the
fuel
quality
information
used
in
the
air
quality
analysis.

Tables
1
­
4
provide
the
gasoline
qualities
for
the
various
gasoline
types
making
up
the
gasoline
pool
for
the
reference
case
and
0.50
volume
percent
benzene
control
case
in
each
PADD.
Table
5
contains
the
weighted­
average
gasoline
qualities
for
PADDs
1
 
4.
To
understand
the
correlations
between
changes
in
benzene
levels
and
changes
in
the
other
fuel
qualities,
we
provide
a
ratio
for
the
change
in
each
of
the
other
fuel
qualities
divided
by
the
change
in
benzene.
We
provide
the
correlations
for
each
PADD
for
RFG,
CG
and
for
a
volume­
weighted
average
of
RFG
and
CG.
We
also
provide
the
correlations
for
a
volumeweighted
average
of
all
the
PADDs
for
RFG,
CG
and
the
entire
pool.
We
did
not
provide
these
correlations
for
the
low
RVP
gasoline
blends
because
their
very
small
volumes
usually
result
in
very
poor
correlations
due
to
large
changes
in
gasoline
blending
in
the
control
case
relative
to
the
reference
case.
Instead,
we
volume­
weighted
the
gasoline
qualities
for
the
low
RVP
cases
with
the
rest
of
conventional
gasoline.
Table
1
­
Summary
of
Reference
Case
and
Benzene
Control
Case
Gasoline
Qualities
and
Correlations
for
Benzene
Control
for
PADD
1
2010
Reference
Case
(
no
Additional
Benzene
Control)
2010
Control
Case
with
0.50
vol%
Benzene
Control
Standard
Change
in
Qualities
per
Change
in
Bz
Conventional
Gasoline
Reformulated
Gasoline
CG
&
RFG
Conventional
Gasoline
Reformulated
Gasoline
CG
&
RFG
9.0
RVP
7.8
RVP
CG
Avg.
MTBEBlended
Etoh­
Blended
RFG
Avg.
Avg.
9.0
RVP
7.8
RVP
CG
Avg.
MTBEBlended
Etoh­
Blended
RFG
Avg.
Avg.
CG
RFG
Pool
RVP
8.7
7.6
8.41
6.8
7.0
6.9
7.5
8.7
7.6
8.4
6.8
7.0
6.9
7.5
0
0
0
Oxygen
0
0
0
2.1
3.5
2.6
1.6
0
0
0
2.1
3.5
2.6
1.6
0
0
0
Aromatics
34.5
36.0
34.9
21.5
20.2
21.1
26.5
34.5
36.0
34.9
21.5
19.4
20.8
26.4
0
2.39
0.73
Benzene
0.88
0.87
0.88
0.59
0.59
0.59
0.70
0.50
0.50
0.50
0.48
0.48
0.48
0.49
­
­
­

Olefins
15.1
14.8
15.0
13.0
14.0
13.3
14.0
14.7
16.0
15.0
13.0
14.0
13.3
14.0
­
0.04
0
­
0.03
Sulfur
27
12
23
27
29
28
26
24
22
24
27
29
27.7
26.0
­
0.99
0
­
0.69
E200
43.5
39.7
42.5
48.7
46.7
48.0
45.9
42.8
42.8
42.6
48.2
46.2
47.5
45.6
­
0.37
4.55
1.15
E300
81.9
77.4
80.7
82.3
83.6
82.7
81.9
83.4
75.0
81.2
81.8
83.3
82.3
81.9
­
1.29
3.95
0.33
Volume
251
88
339
349
171
520
859
251
88
339
349
171
520
859
­
­
­

Table
2
­
Summary
of
Reference
Case
and
Benzene
Control
Case
Gasoline
Qualities
and
Correlations
for
Benzene
Control
for
PADD
2
2010
Reference
Case
(
no
Additional
Benzene
Control)
2010
Control
Case
with
0.50
vol%
Benzene
Control
Standard
Change
in
Qualities
per
Change
in
Bz
Conventional
Gasoline
RFG
Pool
Conventional
Gasoline
RFG
Pool
CG
RFG
Pool
9.0
RVP
CG
w
Etoh
7.8
RVP
7.0
RVP
CG
Avg.
Avg.
Avg.
9.0
RVP
CG
w
Etoh
7.8
RVP
7.0
RVP
CG
Avg.
Avg.
Avg.

RVP
8.7
8.7
7.6
6.8
8.4
6.8
8.2
8.7
8.7
6.8
7.6
8.3
6.8
8.2
0.10
0
0.10
Oxygen
0
3.5
0
0
0.8
3.5
1.1
0
3.5
0
0
0.8
3.5
1.1
0
0
0
Aromatics
34.0
19.0
34.0
35.0
30.7
22.7
29.8
29.1
33.2
25.0
34.0
29.6
21.5
28.6
1.31
3.64
1.42
Benzene
1.37
1.37
1.37
1.37
1.37
0.83
1.31
0.50
0.50
0.50
0.50
0.50
0.50
0.5
­
­
­

Olefins
8.9
12.0
15.2
12.2
10.8
4.1
10.0
11.5
6.2
11.8
14.9
10.5
4.1
9.8
0.24
0
0.23
Sulfur
18
25
25
25
21
13
20
23
13
25
25
21
13
20.2
­
0.23
0
­
0.22
E200
43.8
54.7
46.6
42.4
46.6
51.9
47.2
44.4
54.5
40.0
40.0
45.7
49.2
46.1
1.00
8.18
1.35
E300
78.7
88.4
86.1
87.3
82.4
81.7
82.4
83.5
79.5
85.3
79.3
82.7
82.3
82.7
­
0.29
­
1.82
­
0.37
Volume
1006
391
282
86
1768
237
2005
1009
391
282
86
1768
237
2005
­
­
­

Table
3a
­
Summary
of
Reference
Case
Gasoline
Qualities
for
PADD
3
2010
Reference
Case
(
no
Additional
Benzene
Control)

Conventional
Gasoline
RFG
Pool
CG
w
MTBE
CG
7.8
RVP
7.0
RVP
CG
Avg.
RFG
w
MTBE
RFG
no
Oxy
RFG
w
Etoh
RFG
Avg.
Avg.

RVP
8.7
8.7
7.6
6.8
8.32
6.8
6.8
6.8
6.8
8.0
Oxygen
0.3
0
0.5
0.1
0.28
2.1
0
3.5
2.3
0.7
Aromatics
34.0
34.0
34.0
34.0
34.0
22.0
22.0
22.0
22.0
31.6
Benzene
0.94
0.83
0.94
0.73
0.91
0.58
0.57
0.58
0.58
0.84
Olefins
8.2
13.0
12.2
0.2
9.2
12.5
11.8
8.2
11.6
9.7
Sulfur
18
25
25
4
20
25
25
19
24
20.4
E200
42.0
42.0
43.9
46.1
42.7
52.2
44.9
51.2
51.7
44.5
E300
79.0
81.1
87.9
77.4
81.1
83.2
80.8
82.1
82.9
81.4
Volume
1928
558
732
266
3484
656
34
173
863
4347
Table
3b
­
Summary
of
Benzene
Control
Case
Gasoline
Qualities
and
Correlations
for
Benzene
Control
for
PADD
3
2010
Control
Case
with
0.50
vol%
Benzene
Control
Standard
Change
in
Qualities
Per
Change
in
Bz
Conventional
Gasoline
RFG
Pool
CG
RFG
Pool
CG
w
MTBE
CG
7.8
RVP
7.0
RVP
CG
Avg.
RFG
w
MTBE
RFG
no
Oxy
RFG
w
Etoh
RFG
Avg.
Avg.

RVP
8.7
8.7
7.6
6.8
8.3
6.8
6.8
6.8
6.8
8.0
0
0
0
Oxygen
0.3
0
0.4
0.6
0.3
2.1
0
3.5
2.3
0.7
­
0.04
0
­
0.04
Aromatics
34.0
34.0
34.0
34.0
34.0
22.0
22.0
22.0
22.0
31.6
0
0
0
Benzene
0.49
0.49
0.49
0.49
0.49
0.46
0.46
0.46
0.46
0.48
­
­
­

Olefins
9.2
12.9
13.3
0.2
10.0
9.7
0.2
7.5
8.9
9.8
­
1.85
22.9
­
0.21
Sulfur
20
25
25
5
21
22
8
18
20.6
21
­
2.84
27.18
­
0.86
E200
42.0
42.0
45.4
45.0
42.9
47.8
42.2
50.6
48.1
44.0
­
0.56
30.06
1.47
E300
79.0
80.4
80.9
98.5
81.1
82.1
97.6
81.4
82.6
81.4
­
0.07
2.65
0.11
Volume
1928
558
732
266
3484
656
34
173
863
4347
­
­
­
Table
4
­
Summary
of
Reference
Case
and
Benzene
Control
Case
Gasoline
Qualities
and
Correlations
for
Benzene
Control
for
PADDs
4
and
5
2010
Reference
Case
(
no
Additional
Benzene
Control)
2010
Control
Case
with
0.50
vol%

Benzene
Control
Standard
Change
in
Qualities
per
Change
in
Bz
Conventional
Gasoline
Conventional
Gasoline
CG
CG
w
Etoh
7.8
RVP
Avg
CG
CG
w
Etoh
7.8
RVP
Avg
Pool
RVP
8.7
8.7
7.6
8.5
8.7
8.7
7.6
8.5
0
Oxygen
0
3.5
0
0.5
0
3.5
0
0.5
0
Aromatics
29.2
26.9
33.4
29.7
29.8
26.8
25.0
28.4
0.90
Benzene
2.00
1.09
2.00
1.86
0.50
0.50
0.50
0.50
­

Olefins
8.9
15.5
13.3
10.7
10.9
15.8
10.0
11.5
­
0.54
Sulfur
19
25
25
21.1
22.0
25.0
21
22.3
­
0.89
E200
50.0
50.0
49.0
50.0
48.4
50.0
44.7
47.9
1.37
E300
87.4
80.2
87.1
86.2
87.0
79.7
85.9
85.7
0.42
Volume
476
110
138
724
476
110
138
724
­

Table
5
 
Summary
of
Reference
Case
and
Benzene
Control
Case
Gasoline
Qualities
and
Correlations
for
Benzene
Control
for
the
U.
S.

2010
Reference
Case
(
no
Additional
Benzene
Control)
2010
Control
Case
with
0.50
vol%
Benzene
Control
Standard
Change
in
Qualities
Per
Change
in
Bz
2010
Reference
Case
(
no
Add.

Benzene
Control)
2010
Control
Case
with
0.50
vol%
Bz
Control
Std
Change
in
Qualities
per
Change
in
Bz
CG
RFG
CG
RFG
CG
RFG
CG
&
RFG
CG
&
RFG
CG
&
RFG
RVP
8.4
6.8
8.4
6.8
0.04
0
8.06
8.04
0.04
Oxygen
0.2
1.3
0.2
1.3
0.08
0
0.87
0.87
­
0.01
Aromatics
32.6
21.8
32.2
21.5
0.71
1.77
30.42
30.00
0.77
Benzene
1.14
0.62
0.49
0.47
­
­
1.04
0.49
­

Olefins
10.1
11.1
10.6
9.6
­
0.69
9.89
10.31
10.38
­
0.11
Sulfur
20
23
21
21.7
­
1.34
11.71
20.95
21.29
­
0.62
E200
44.6
50.6
44.3
48.1
0.50
16.73
45.82
45.06
1.39
E300
82.0
82.7
82.1
82.4
­
0.07
1.49
82.17
82.16
0.01
Volume
6315
1620
6315
1620
­
­
7935
7935
­
The
results
from
the
LP
modeling
summarized
in
Tables
1
 
5
were
then
used
to
develop
the
correlations
for
how
fuel
qualities
vary
with
changing
benzene
levels
(
i.
e.,
by
RFG
and
CG
separately
or
as
a
single
pool,
PADD
by
PADD,
or
nationally).
We
attempted
to
develop
correlations
between
benzene
and
total
aromatics,
RVP,
sulfur,
olefins,
E200
and
E300.
Unfortunately,
as
we
closely
evaluated
the
correlations,
many
were
inconsistent
with
refinery
practices,
fuel
standard
constraints,
or
how
benzene
reduction
is
expected
to
affect
gasoline
quality,
with
the
exception
of
aromatics.
As
a
result,
while
we
discuss
the
other
parameter
results
as
well,
a
realistic
correlation
was
only
possible
to
be
developed
between
benzene
reduction
and
the
change
in
total
aromatics
(
of
which
benzene
is
a
small
component).

Total
Aromatics
In
PADD
1,
total
aromatics
decrease
only
in
RFG,
despite
the
larger
decrease
of
benzene
in
CG.
In
PADD
2,
total
aromatics
decrease
in
both
pools,
while
for
PADD
3,
aromatics
don't
decrease
at
all.
Aromatics
decrease
in
the
CG
in
PADD
4
and
5.
The
inconsistent
decrease
in
total
aromatics
relative
to
the
decrease
in
benzene
in
RFG
compared
to
CG
is
likely
to
be
caused
by
refinery
modeling
overoptimization.
Because
the
LP
model
acts
as
one
big
refinery
in
each
PADD,
it
is
common
for
the
refinery
model
to
switch
gasoline
blendstocks
freely
between
the
two
pools.
However,
while
some
refineries
produce
both
RFG
and
CG,
many
produce
only
one
of
the
two
gasoline
types.
For
example,
in
PADD
1,
despite
the
large
fraction
that
RFG
comprises
of
the
gasoline
pool
there,
three
refineries
only
produce
CG
and
another
three
predominantly
produce
CG.
It
is
likely
that
these
CG
refineries
would
experience
some
impact
on
their
gasoline
aromatics
levels
concurrent
with
benzene
reductions
in
their
gasoline.
Thus,
it
would
be
very
likely
for
there
to
be
a
change
in
CG
aromatic
levels
in
PADD
1.
It
is
also
likely
that
PADD
3
would
have
changes
to
the
total
aromatic
levels
in
the
RFG
and
CG
produced
within
that
PADD.
Because
of
these
inconsistencies,
the
correlations
in
total
aromatics
reductions
with
benzene
reductions
were
established
based
on
pool
averages
of
RFG
and
CG
and
across
PADDs
­
from
the
last
column
in
Table
5.
Thus,
for
each
1
percent
reduction
in
benzene
in
a
refinery's
gasoline
pool,
there
is
a
0.77
percent
reduction
in
total
aromatics
(
i.
e.,
nonbenzene
aromatics
increase
by
0.23
per
each
1.0
change
in
benzene).
Thus,
the
only
secondary
impact
on
gasoline
quality
that
we
are
considering
in
our
emissions
modeling
work
is
aromatics
­
for
every
1
percent
decrease
of
benzene
in
gasoline,
aromatics
concentration
decreases
by
0.77
percent.

RVP
and
Sulfur
Both
of
these
gasoline
qualities
will
be
established
by
refiners
using
technologies
separate
from
that
used
to
reduce
benzene.
We
fully
expect
that
refiners
will
use
these
technologies
to
just
meet
the
applicable
standard
to
minimize
cost
and
maximize
gasoline
production.
RVP
of
the
gasoline
pool
is
established
by
blending
back
butane
into
the
gasoline
pool
up
to
the
RVP
standard.
Gasoline
sulfur
levels
will
be
established
by
how
hard
refiners
operate
their
naphtha
hydrotreaters.
For
this
reason,
no
changes
in
RVP
or
sulfur
are
anticipated
in
practice
to
result
from
benzene
changes,
though
the
modeling
allowed
for
such
changes.

Olefins
Reducing
benzene
levels
would
not
be
expected
to
affect
olefin
levels
because
the
refining
unit
principally
involved
in
producing
olefins
(
the
fluidized
catalytic
cracker
unit)
is
not
the
unit
that
refiners
would
adjust
or
treat
to
reduce
benzene.
The
inconsistent
results
across
the
PADDs
may
simply
be
evidence
of
the
limitations
of
the
refiner
model
for
use
in
this
situation.
Changes
in
olefins
are
associated
with
gasoline
sulfur
changes,
however,
so
it
is
possible
that
the
olefins
changes
resulted
from
the
projected
changes
in
sulfur
content.

Oxygen
Oxygen
levels
increased
only
in
PADD
3
through
the
addition
of
a
small
amount
of
MTBE.
However,
this
result
is
inconsistent
with
recent
expectations
of
the
industry.

E200
&
E300
E200
levels
decreased
somewhat
in
each
PADD.
Most
of
the
benzene
reduction
is
expected
to
occur
through
benzene
saturation
or
benzene
precursor
rerouting.
When
benzene
is
saturated
or
if
the
benzene
precursors
are
routed
around
the
reformer,
cyclohexane
would
replace
benzene
in
the
gasoline
pool.
Cyclohexane
and
benzene
both
boil
below
200
degrees
Fahrenheit
(
and
at
almost
an
identical
temperature),
thus
this
switch
in
gasoline
components
is
not
expected
to
impact
either
E200
or
E300.
Benzene
extraction
could
impact
E200
and
E300,
but
not
significantly
beyond
the
base
case
and
not
nationwide.
Thus,
we
did
not
consider
the
E200
and
E300
effects
estimated
by
the
LP
refinery
model
to
be
a
valid
model
outcome.
