Fluorochemical
Decomposition
Processes
­
April
4,2001
David
A.
Dixon
Theory,
Modeling,
and
Simulation,
William
R.
Wiley
Environmental
Molecular
Sciences
Laboratory,
Pacific
Northwest
National
Laboratory,
Richland
WA
Overview
This
document
describes
results
on
potential
degradation
pathways
for
fluorochemicals
in
incineration
processes
based
on
modeling
of
the
thermodynamics
of
the
processes.
The
primary
goal
of
this
work
is
to
evaluate
if
perfluoroalkyl
sulfonates,
or
potential
precursors
of
these
compounds,
are
likely
emission
products
of
the
incineration
of
perfluoroalkyl
sulfonates
or
of
monomers
or
polymers
containing
perfluoroalkyl
sulfonamides.
Ab
initio
electronic
structure
theory
was
used
to
predict
the
thermodynamics
of
various
degradation
pathways
for
model
sulfonamides.
Key
conclusions
from
this
work
include:

1.
If
the
temperature
in
the
incineration
process
(e.
g.
incinerator,
cement
kiln,
etc.)
is
high
enough,
the
fluorocarbon
will
completely
decompose.
Thus,
the
engineering
aspects
of
the
incineration
system
in
terms
of
mixing
and
residence
time
will
be
the
dominant
issue
in
minimizing
by­
product
formation
as
long
as
the
incineration
temperature
is
high
enough.
One
must
also
have
the
proper
fluorocarbodfuel
ratio.

2.
If
CF4,
the
most
stable
perfluoroalkyl
compound,
were
added
as
a
tracer
to
the
incineration
system,
its
complete
destruction
would
show
that
the
incineration
conditions
should
fully
destroy
all
other
perfluoroalkyl
compounds.

Thermal
Decomposition
Processes
In
an
incineration
system
as
well
as
in
a
cement
kiln,
the
waste
material
is
heated
to
a
high
temperature
(temperatures
up
to
­3000"
F
­
1650'
C
­
1925
K
based
on
a
burner
flame
temperature
of
3500
to
4000
'
F)
in
an
air
stream
that
provides
0
2
as
the
oxidizing
agent
together
with
a
hydrocarbon
fuel.
Typical
residence
times
are
­3
sec
at
temperatures
between
2000'
and
3000'
F
and
­8
sec
at
temperatures
above
1500
'
F
and
below
2000'.
Combustion
can
then
occur
leading
to
the
destruction
of
the
starting
material
and
under
the
proper
conditions,
the
formation
of
products
with
low
or
zero
environmental
impact.

The
material
can
either
burn
as
a
solid
which
tends
to
be
a
slower
process
(and
even
this
is
often
dominated
by
vaporization)
or
it
can
be
vaporized
leading
to
a
faster
process.
We
have
focused
on
the
vapor
phase
reactions
of
the
materials
of
interest
as
these
are
likely
to
be
the
dominant
processes
at
high
temperature.
Once
vaporized,
the
compound
can
start
to
decompose
and
react
with
oxygen
(or
with
other
radicals
generated
by
the
fuel
that
is
present).
The
decomposition
reactions
usually
occur
by
bond
scission
(2­
center)
or
by
a
3­
center
or
4­
center
elimination
process
and
continue
via
a
radical
chain
mechanism.
If
bond
breaking
is
the
initial
reaction,
one
must
determine
the
bond
that
breaks
first.
Different
bonds
with
different
strengths
will
have
different
dependencies
on
the
temperature.
The
weakest
bond
will
break
at
the
lowest
temperature
1
BACK
TO
MAIN
and
at
very
high
temperatures
there
will
be
less
dependence
on
the
bond
strengths
and
a
broader
distribution
of
radicals
species
can
be
formed.

The
compounds
of
interest
to
the
present
study
are
predominantly
sulfonamides
of
the
form
RfS02NRR"
where
Rf=
n­
CsF17
or
other
normal
or
branched
perfluoro
alkyl
chains,
R'
=
CH3
or
C2H5,
and
R"
is
CH2CH20H
or
it
is
often
a
very
complex
organic
linked
to
the
N
by
a
(N)­
CH2CH20­
C(
0)
1
linker
group.
We
show,
based
on
high
level
ab
initio
electronic
structure
calculations
at
the
density
hnctional
theory
and
molecular
orbital
theory
levels,
that
the
most
likely
bond
to
break
in
compounds
of
the
form
RfS02R'
is
the
C­
S
bond.
We
focus
on
the
differences
in
the
bond
energies
of
the
C­
S
and
the
C­
C
bonds
in
the
Rf
fragment.
We
used
Rf
=
n­
C3F7
and
CF3
in
our
model
studies
and
examined
R'
=
OH,
F,
NH2,
and
0.
In
all
cases
where
a
neutral,
nonradical
compound
was
studied,
we
find
that
the
average
C­
S
bond
energy
was
­
64
kcal/
mol
and
the
average
C­
C
bond
energy
was
­85
kcal/
mol.
This
is
the
energy
required
for
breaking
the
CrC(
S02)
R'
bond.
(C­
F
bond
energies
are
much
higher,
near
120
kcal/
mol.)
The
average
bond
energies
of
C­
C
bonds
in
the
perfluoroalkyl
fragment
further
from
the
SO2
group
are
likely
to
be
on
the
order
of
95
kcal/
mol
based
on
other
studies
of
perfluorocarbon
bond
energies.
Bond
energies
of
simple
unsubstituted
hydrocarbon
alkane
substituents
are
expected
to
be
on
the
order
of
85
to
90
kcal/
mol.
Substituents
that
can
stabilize
radical
carbon
centers
will
clearly
lead
to
lower
C­
C
bond
energies.

If
a
C­
C
or
a
C­
S
bond
breaks,
the
decomposition
chemistry
of
the
perfluoroalkane
radical
formed
in
the
process
will
follow
the
mechanism
described
below.
If
0
2
intercepts
the
radical
(a
likely
process),
then
a
peroxy
radical
ROO.
is
formed
which
quickly
decomposes
to
an
alkoxy
radical
ROO.
The
R@*
radical
will
undergo
a
p­
scission
process
as
shown
in
Reaction
(1)
to
form
carbonyl
fluoride
COF2
and
a
one
carbon
shorter
perfluoroalkyl
radical.
The
chain
decomposition
will
then
continue,
as
above,
until
CF3
is
formed.

Rf`
CF203Rf`
+
F2C(
O)
(1)
The
=CF3
can
react
with
0
2
to
eventually
give
the
alkoxide
CF3O.
which
usually
abstracts
a
hydrogen
atom
from
the
hydrocarbon
fuel
that
is
present
to
form
the
alcohol
CF30H.
The
alcohol
can
undergo
4­
center
HF
elimination
to
form
carbonyl
fluoride
COF2
again.
Thus
the
main
products
from
the
decomposition
of
the
fluorocarbon
chain
should
be
COF2
and
HF.
The
COF2
will
react
with
any
water
present
and
decompose
to
C02
and
HF
and,
of
course,
as
in
all
combustion
processes
with
hydrocarbons
present,
some
H20
will
be
formed.
As
in
all
incineration
processes,
the
completeness
of
the
reaction
will
depend
on
the
temperature
and
residence
time.
CF4
is
probably
the
most
difficult
fluorinated
compound
to
decompose
in
a
normal
incineration
system.
'
In
order
to
detect
whether
incomplete
incineration
is
occurring
for
other
fluorocarbon­
derived
species,
one
can
add
a
small
amount
of
CF4
to
the
system.
If
no
CF4
is
present
in
the
exit
stream
from
the
incineration
furnace,
it
is
likely
that
all
of
fluorinated
material
has
been
combusted.

We
have
used
density
functional
theory
(DFT)
at
the
local
and
nonlocal
levels
as
well
as
molecular
orbital
theory
at
the
G3
and
G3­
MP2
levels2
to
calculate
the
various
bond
energies.
We
used
correction
factors
from
the
G3
and
G3­
MP2
levels
to
correct
the
non­
local
DFT
results.
We
obtain
the
following
reaction
energies.

2
BACK
TO
MAIN
Bond
Cleavage
Products
C3F7*
+
*S02NH2
C2F5*
+
*CF2S02NH2
C3F7*
+
SO2
C2F5*
+
CF2S02
(singlet)

C3F7*
+
SO3
(singlet)
C2F5*
+
*CF2S03*
(triplet)

C3F7*
+
*S02OH
C2F5*
+
*CF2S020H
C3F7*
+
*S02F
C2F5.
+
aCF2S02F
B.
E.(
kcal/
mol)

63
85
7
68
20
84
64
85
64
84
78
89
The
conditions
in
most
cement
kiln
incineration
systems
are
very
high
temperatures
(up
to
­1900
K).
As
the
C­
S
bonds
are
of
the
lowest
energy,
they
will
break
first
under
these
high
temperatures
and
will
be
the
first
to
break
at
the
lower
temperatures
expected
in
the
temperature
ramp
region
in
the
incineration
system.
This
will
lead
to
formation
of
fluorocarbon
radicals
that
will
then
decompose
as
discussed
above.
If
the
S02R'
group
is
SO3,
then
the
C­
S
bond
is
extremely
weak,
­20
kcal/
mole,
and
the
compound
will
readily
decompose.
The
C­
C
bond
energy
is
still
on
the
order
of
85
kcal/
mole.
Also
note
that
this
is
dissociation
to
the
singlet
ground
state
of
SO3.
The
triplet
is
calculated
to
be
44
kcal/
mole
higher
in
energy
than
the
singlet
which
places
the
S­
C
bond
energy
in
the
same
range
as
those
of
other
compounds.
(Our
calculated
singlet­
triplet
energy
difference
is
likely
be
too
low.)
The
C­
C
bond
for
CF$
F*­
CF$
03
to
form
triplet
CF2S03
is
of
a
comparable
energy
to
other
C­
C
bonds
in
this
series.
If
the
S­
R'
bond
breaks
first,
for
example
in
RfS02N(
CH3)
C2H40R
(however,
see
below
for
a
discussion
of
this
point),
then
the
radical
RrSO2
will
be
formed.
If
an
RS02*
fragment
is
formed,
the
C­
S
bond
is
very
weak,
only
21
kcal/
mole
in
CH3S02
and
an
even
lower
7
kcal/
mol
for
CF3S02.
Cleavage
of
the
C­
S
bond
in
RfS02
will
quickly
lead
to
the
fluorocarbon
radical
chain
decomposition
described
above
and
SO2
formation.
The
C­
C
bond
energy
in
RfSO2
is
again
much
higher
than
the
C­
S
bond
energy
although
lower
than
other
C­
C
bond
energies.
If
the
S02R'
group
is
SO3­,
then
the
C­
S
bond
is
about
10
to
15
kcal/
mole
stronger
than
in
the
other
compounds
noted
above
but
still
­10
kcal/
mole
weaker
than
the
C­
C
bond.

We
also
have
calculated
­S02­
N
bond
anti
N­
C
bond
energies
in
related
model
systems.
These
models
were
chosen
to
get
higher
accuracy
without
having
to
use
additivity
corrections.
The
N­
C
bond
strength
in
CH3NH2
is
85
kcal/
mol
at
298
K
based
on
the
latest
results
from
experiment
and
theory.
For
the
model
system
CF3S02NH2,
we
obtain
the
following
bond
energies
at
the
G3­
MP2
Level:

3
BACK
TO
MAIN
CF3S02NH2
+
CF3.
+
*S02NH2
68.3
kcal/
mol
3
NH2.
+
.SO:?
CF3
79.5
kcal/
mol
The
C­
S
bond
energy
is
consistent
with
the
values
shown
above
(slightly
stronger
as
the
CF3
radical
is
not
as
stabilizing
as
the
C3F7
radical)
and
the
S­
N
bond
energy
is
clearly
stronger
than
the
C­
S
bond
as
expected.
For
the
model
system
CF3S02NHCH3,
we
find
the
following
at
the
G3­
MP2
level:

CF3S02NHCH3
+
CF3S02NH.
+
.CH3
94.3
kcal/
mol
3
CF3S02.
+
aNHCH3
78.4
kcal/
mol
These
results
are
consistent
with
those
discussed
above.
The
model
systems
might
lead
to
changes
of
a
few
kcal/
mol
due
to
remote
substituent
effects
but
no
larger
changes
are
expected.

The
hydrocarbon
amine
decomposition
chemistry
and
the
alcohol
amine
decomposition
chemistry
can
be
worked
out
by
similar
methods
as
can
the
acrylate
chemistry
(e.
g.,
polymers
made
by
polymerizing
RfS02N(
CH3)
C2H20C(
O)
CH=
CH2
with
methacrylates
or
other
acrylates).
As
long
as
the
fluorocarbon
group
degrades
completely,
it
is
likely
that
any
of
the
other
substituents
bonded
to
the
SO2
will
also
fully
degrade,
especially
if
they
have
a
significant
hydrocarbon
component.

One
must
also
consider
the
possibility
that
reactions
with
H
or
0
atoms
or
the
OH
radical
are
occurring
in
the
incineration
regime.
'The
likely
radical
chain
carriers
are
H
atoms
and
OH
radicals.
The
H
atom
will
usually
abstract
an
F
or
an
H
depending
on
the
position
and
the
OH
radical
can
abstract
any
hydrogens
added
to
the
fluorocarbon
chains
during
the
combustion
process.
One
can
also
abstract
other
atoms
depending
on
bond
strengths.
The
other
possibility
that
also
must
be
considered
is
3­
center
elimination
of
HF
if
H
is
added
to
a
perfluorocarbon
chain.
This
can
be
a
lower
energy
process
than
bond
breaking.
However,
one
does
not
have
to
wisrry
about
0­
F
bonds
being
formed
as
these
are
very
weak.
Based
on
the
bond
energies,
the
temperatures
in
incinerators
and
in
cement
kilns
are
high
enough
(­
peak
1900
K
for
cement
kilns)
that
most
decomposition
processes
should
readily
take
place.
The
concern
about
byproduct
formation
except
for
CF4
is
not
so
much
the
temperature
but
the
mixing
conditions
in
the
incinerator.
Flow
patterns
that
allow
the
fluorocarbons
to
bypass
the
flame
zone
and
to
have
a
short
(4
s)
contact
time
even
at
high
temperatures
can
lead
to
the
formation
of
unwanted
byproducts
if
only
partial
combustion
occurs.
This
is
essentially
an
engineering
problem,
not
a
chemistry
problem.
In
addition,
we
note
that
Tsang
et.
al.
'
provide
estimated
temperatures
to
achieve
destruction
of
99.99%
of
a
fluorocarbon
compound
in
1
sec
based
on
unimolecular
decomposition
reactions.
The
highest
temperature
needed
for
a
fluorocarbon
system
was
­1715
K
for
CF4.
For
ii
fluorocarbon
radical
such
as
C2F5,
a
temperature
of
­1200
K
was
needed
(based
on
C­
C
bond
scission)
and
for
C2F6,
a
temperature
of
­1200
K
was
needed.
These
results
are
consistent
with
our
conclusions
that
the
temperature
is
not
the
likely
culprit
in
the
formation
of
any
byproducts
but
rather
the
amount
of
mixing
and
the
residence
time
in
the
high
temperature
region.
The
4
BACK
TO
MAIN
temperature
does
however
need
to
be
well
above
­lOOO°
C.
Lower
temperatures
can
and
will
lead
to
the
formation
of
undesired
by­
products.
Thus,
cement
kilns
will
need
to
be
operated
at
as
high
a
temperature
as
possible
in
order
to
guarantee
complete
decomposition
of
the
fluorocarbons
except
for
CF4
and
at
the
highest
operating
temperatures
even
CF4
should
decompose.
An
additional
important
concern
is
the
fluorocarbodfiel
ratio.
If
the
amount
of
fluorinated
material
gets
above
a
few
percent,
it
is
likely
to
also
lead
to
more
by­
product
formation
as
there
will
not
be
as
good
mixing
and
less
of
the
fluorinated
material
may
pass
through
the
high
temperature
combustion
zone
where
efficient
decomposition
can
take
place.

Photolytic
Processes
For
the
model
compound
CF3S02NH2,
we
also
calculated
the
ultraviolet­
visible
(UV­
Vis)
absorption
spectra
by
using
time­
dependent
density
functional
theory
(TD­
DFT).
TD­
DFT
is
an
approach
with
reasonable
computational
cost
and
reasonable
accuracy
for
the
calculation
of
the
energies
of
excited
states
of
molecules
that
we
have
been
testing
for
use
in
the
development
of
fluorinated
resist
materials
for
157
nm
photo
resist^.^
We
have
previously
shown
that
this
a
good
level
at
which
to
calculate
such
values
for
a
broad
range
of
inorganic
and
organic
compounds.
The
predicted
values
show
that
these
types
of
compounds
will
not
absorb
visible
light
and
decompose.
If
the
primary
perfluoroalkylsulfonamides
get
into
the
atmosphere,
they
would
only
decompose
photolytically
in
the
upper
stratosphere.
In
other
words,
they
are
extremely
stable
in
terms
of
photolysis.

TD­
DFT
Calculated
Excitation
Energies
of
CFsS02NH2
Excitation
Excitation
Oscillator
Energy
(eV)
Strength
f
ho
177
7.00
0.0008
161
7.69
0.0228
157
7.90
0.0066
156
7.93
0.005
1
Conclusions
following
conclusions
may
be
drawn:
Based
on
the
above
assessment
of
the
thermal
destruction
of
fluorochemicals,
the
1.
The
C­
S
bond
should
break
first
under
high
temperature
conditions
leaving
the
substituted
S02R'
or
the
Rf
SO2
radicals
and
the
Rfand
R'
radicals
respectively.

2.
The
Rf
radical
will
then
follow
normal
fluorocarbon
combustion
pathways
via
C(
O)
F2
3.
The
S02R'
radical
will
likely
decompose
to
SO2
and
R'
with
the
R'
radical
following
its
normal
decomposition
path.
to
form
C02
and
HF.

5
BACK
TO
MAIN
4.
If
RfSO3
is
formed,
it
will
quickly
decompose
to
Rf
and
SO3
at
very
low
temperatures.

5.
At
the
upper
operating
temperatures
of
cement
kilns
(­
1900
K),
CF4
should
be
efficiently
decomposed
so
that
it
can
serve
as
a
marker
for
complete
decomposition.
If
CF4
was
added
a
tracer
gas
and
no
CF4
is
present
at
the
outlet,
it
is
unlikely
that
any
significant
fluorinated
by­
products
are
generated.

6.
The
engineering
aspects
of
the
incineration
system
in
terms
of
mixing
and
residence
time
will
be
the
dominant
issue
in
minimizing
by­
product
formation
as
long
as
the
incineration
temperature
is
high
enough.
One
wants
to
have
the
longest
possible
contact
time
to
ensure
complete
degradation.
One
must
also
have
the
proper
fluorocarbodfuel
ratio.
In
addition,
it
will
be
best
if
the
cement
kiln
is
run
near
the
upper
end
of
its
operating
temperature
range.

References
1.
Tsang,
W.
S.;
Burgess,
Jr.,
D.
R.;
Babushok,
17.
Combust
Sci
and
Tech.
1998,
139,
385.

2.
Curtiss,
L.
A.;
Raghavachari,
K.;
Redfern,
P.
C.;
Pople,
J.
A.
J.
Chem.
Phys.
1997,
103,
1063;
Curtiss,
L.
A.;
Raghavachari,
K.;
Redfern,
P.
C.;
Pople,
J.
A.
J.
Chem.
Phys.
1998,
lU9,7764.

3.
Matsuzawa,
N.
N.;
Mori,
S.;
Yano,
E.;
Okazaki,
S.;
Ishitani,
A.;
Dixon,
D.
A.
Proc.
SPIE
2000,
3999,
375;
Matsuzawa,
N.
N.;
Ishitani,
A.;
Dixon,
D.
A.;
Uda,
T.
J.
Phys.
Chem.
A,
in
press
(2001).

6
BACK
TO
MAIN
