Speight J.G. Natural Gas: A Basic Handbook
Подождите немного. Документ загружается.
7.2
Physical Solvent Processes
169
Physical solvent processes can be easily configured to take advantage
of
their high hydrogen sulfide/carbon dioxide selectivity together
with high levels
of
carbon dioxide recovery.
The Selexol process has been used since the late
1960s.
The process
solvent is a mixture
of
dimethyl ethers
of
polyethylene glycol
[CH3(CH,CH,0),CH3], where
n
is between
3
and
9.
The solvent is
chemically and thermally stable, and has a low vapor pressure that
limits its losses
to
the treated gas. The solvent has a high solubility for
carbon dioxide, hydrogen sulfide, and carbonyl sulfide. It also has
appreciable selectivity for hydrogen sulfide over carbon dioxide.
The Selexol process can be configured in various ways, depending on
the requirements for the level
of
hydrogen sulfide/carbon dioxide
selectivity, the depth
of
sulfur removal, the need for bulk carbon
dioxide removal, and whether the gas needs
to
be dehydrated
(Figure
7-2).
The gas stream from the low-pressure flash is combined
with the acid gas from the regenerator. This combined gas stream is
then sent to a sulfur recovery unit. However, the hydrogen sulfide
content could be too low for use in a conventional Claus plant.
Figure
7-2
The
Selexol
process.
170
Chapter
7
Processes
The Selexol process can, however, be configured to give both a rich
acid-gas feed to the Claus unit as well as to provide for bulk carbon-
dioxide removal.
7.2.1
Rectisol Process
This process is the most widely used physical solvent gas-treating pro-
cess in the world. It uses chilled methanol (methyl alcohol,
CH,OH)
at a temperature
of
about
-40°F
to
-80°F.
The selectivity (by meth-
anol) for hydrogen sulfide over carbon dioxide at these temperatures
is
about
6/1,
a little lower than that
of
the Selexol process at its usual
operating temperature. However, the solubility
of
hydrogen sulfide
and carbonyl sulfide
in
methanol, at typical process operating tem-
peratures, are higher than in Selexol and allow for very deep sulfur
removal. The high selectivity for hydrogen sulfide over carbon
dioxide, combined with the ability to remove carbonyl sulfide, is the
primary advantage of the process. The need to refrigerate the solvent
is the main disadvantage of the process, resulting
in
high capital and
operating costs.
There are many possible process configurations for the Rectisol pro-
cess, depending on process requirements. Different process layouts
are used for selective hydrogen sulfide removal, and carbon dioxide,
and for non-selective carbon dioxide and hydrogen sulfide.
The process is designed for the bulk removal of carbon dioxide and
nearly all of the removal of hydrogen sulfide and carbonyl sulfide
occur in the bottom section of the absorber. The methanol solvent
contacting the feed gas in the first stage
of
the absorber is stripped in
two
stages of flashing via pressure reduction. The regenerated solvent
is virtually free of sulfur compounds but contains some carbon
dioxide. The acid gas leaving the first stage solvent regenerator is suit-
able for a Claus plant. The second stage of absorption is designed for
the removal of the remaining sulfur compounds and carbon dioxide.
The solvent from the bottom of the second stage of the absorber is
stripped deeply in a steam-heated regenerator and is returned to the
top of the absorption column after cooling and refrigeration.
7.2.2
Sulfinol
Process
This process, developed in the early
1960s,
is a combination process
that
uses
a
mixture
of
amines
and
a
physical
solvent.
The
solvent
con-
sists
of
an aqueous amine and sulfolane. In operation, this process is in
7.2
Physical Solvent Processes
171
many respects identical to the familiar amine method, and its equip-
ment components are similar to those found in amine units. The main
difference is that while the conventional amine process employs a
fairly diluted concentration of amine in water to remove acid gases by
chemical reaction, the Sulfinol system uses a mixture of highly con-
centrated amine and a physical solvent to remove the acid gases by
physical and chemical reactions. The concentrations
of
the amine and
the physical solvent vary with the type of feed gas
in
each application.
Common Sulfinol mixtures are
in
the range of
40%
amine (also called
DIPA),
40%
sulfolane (an organic solvent) and
20%
water.
Sulfinol has a good affinity for most
of
the acid gases and has the
ability to release these gases in the regenerator upon pressure reduction
and heat application. When operating under suitable conditions, it is
capable of removing twice as much acid gas as a
20%
MEA
solution.
The Sulfinol-D process uses diisopropanolamine (DIPA), while
Sulfinol-M uses methyl diethanolamine (MDEA). The mixed solvents
allow
for
better solvent loadings at high acid-gas partial pressures and
higher solubility
of
carbonyl sulfide and organic sulfur compounds
than straight aqueous amines.
The Sulfinol-D process is primarily used
in
cases where selective removal
of
hydrogen sulfide
is
not
of
primary concern, but where partial removal
of organic sulfur compounds (mercaptans, RSH, and carbon disulfide,
carbon disulfide) is desired, typically in natural gas and refinery applica-
tions.
The Sulfinol-D configuration is also able to remove some carbonyl
sulfide via physical solubility in sulfolane and partial hydrolysis to
hydrogen sulfide induced by the diisopropanolamine (DIPA). However,
complete removal of carbonyl sulfide by Sulfinol-D cannot be guaran-
teed. Unlike solvents that use other primary and secondary amines
(monoethanolamine,
MEA,
diethanolamine, DEA), Sulfinol-D is claimed
not to be degraded by these sulfur compounds.
Sulfinol-M is used when a higher degree of hydrogen sulfide selec-
tivity is needed. Hydrogen sulfide selectivity in Sulfinol-M is con-
trolled by the relative reaction rate of the reaction of hydrogen sulfide
with methyl diethanolamine and by the physical solubility of
hydrogen sulfide and carbon dioxide
in
the solvent.
172
Chapter
7
Processes
7.3
Metal
Oxide
Processes
7.3.1
Iron
Sponge
Process
Hydrogen sulfide scavengers have been used for many years. An
example is the iron sponge process or dry box process, which is the
oldest and still the most widely used batch process for sweetening
of
natural gas and natural gas liquids (Duckworth and Geddes, 1965;
Anerousis and Whitman,
1984;
and Zapffe,
1963).
The process was
implemented during the
19th
century and has been
in
use in Europe
and the United States for more than
100
years. Hydrogen sulfide scav-
engers are appropriate for use at the low concentrations
of
hydrogen
sulfide, where conventional chemical absorption and physical sol-
vents are not economical. During recent years, hydrogen sulfide scav-
enger technology has been expanded with many new materials
coming on the market and others being discontinued. Overall, the
simplicity
of
the process, low capital costs, and relatively low chem-
ical (iron oxide) cost continue
to
make the process an ideal solution
for hydrogen sulfide removal.
The
sponge
consists
of
wood shavings impregnated with a hydrated
form
of
iron oxide. The wood shavings serve as a carrier for the active
iron-oxide powder. Hydrogen sulfide is removed by reacting with iron
oxide
to
form ferric sulfide. The process is usually best applied to
gases containing low to medium concentrations
(300
ppm)
of
hydrogen sulfide or mercaptans. This process tends to be highly selec-
tive and does not normally remove significant quantities of carbon
dioxide. As a result, the hydrogen sulfide stream from the process is
usually high purity. The use
of
iron sponge process for sweetening
sour gas is based
on
adsorption
of
the acid gases
on
the surface
of
the
solid sweetening agent followed by chemical reaction
of
ferric oxide
(Fe203)
with
hydrogen sulfide:
2Fe20,
+
6H2S +ZFe2S3
+
6H20
The reaction requires the presence
of
slightly alkaline water and a
temperature below
43°C
(110"
F);
bed alkalinity should be checked
regularly, usually daily.
A
pH level
on
the order of
8-10
should be
maintained through the injection
of
caustic soda
with
the water.
If
the gas does not contain sufficient water
vapor,
water
may
need
to
be
injected into the inlet gas stream.
7.3
Metal
Oxide
Processes
173
The ferric sulfide produced by the reaction
of
hydrogen sulfide with
ferric oxide can be oxidized with air
to
produce sulfur and regenerate
the ferric oxide:
2Fe,S3
+
30,
+2Fe,03
+
6s
s,
+
20,
+2SO,
The regeneration step, i.e., the reaction with oxygen is exothermic,
and air must be introduced slowly
so
the heat
of
reaction can be dis-
sipated.
If
air is introduced quickly, the heat
of
reaction may ignite
the bed. Some
of
the elemental sulfur produced
in
the regeneration
step remains in the bed. After several cycles this sulfur will cake over
the ferric oxide, decreasing the reactivity
of
the bed. Typically, after
10
cycles the bed must be removed and a new bed introduced into
the vessel.
In
some designs the iron sponge may be operated with continuous
regeneration by injecting a small amount
of
air into the sour gas feed.
The air regenerates ferric sulfide while hydrogen sulfide is removed by
ferric oxide. This process is not as effective at regenerating the bed as
the batch process and requires a higher-pressure air stream (Arnold
and Stewart,
1999).
In
the process (Figure
7-3),
the sour gas should pass down through
the bed.
In
the case where continuous regeneration is to be used, a
small concentration
of
air is added to the sour gas before it is pro-
cessed. This air serves
to
continuously regenerate the iron oxide,
which has reacted with hydrogen sulfide; although it serves to extend
the on-stream life
of
a given tower, it probably serves to decrease the
total amount
of
sulfur that a given weight
of
bed will remove.
The number
of
vessels containing iron oxide can vary from one to
four.
In
a two-vessel process, one
of
the vessels would be
on
stream
removing hydrogen sulfide from the
sour
gas, while the second vessel
would either be
in
the regeneration cycle or having the
iron
sponge
bed replaced.
When periodic regeneration is used,
a
tower
is
operated until the
bed
is saturated with sulfur and hydrogen sulfide begins to appear
in
the
174
ChaPter
7
Processes
Figure
7-3
Typical iron-oxide process
flow
sheet (Maddox,
1974).
sweetened gas stream. At this point the vessel is removed from ser-
vice, and air is circulated through the bed to regenerate the iron
oxide. Regardless of the type
of
regeneration process used, a given
iron-oxide bed will lose activity gradually and eventually will be
replaced. For this reason the reactor vessels should be designed to
minimize difficulties
in
replacing the iron sponge in the beds. The
change out
of
the beds is hazardous. Exposure to air when dumping a
bed can cause a sharp rise in temperature, which can result in sponta-
neous combustion
of
the bed. Care must be exercised in opening the
tower to the air. The entire bed should be wetted before beginning the
change-out operation.
Iron-oxide reactors have taken on several configurations from conven-
tional box vessels to static tower purifiers. The selection depends upon
the process application. Static tower purifiers are used in high-pressure
applications, because a longer bed depth gives a greater efficiency, and
the total pressure drop is a smaller fraction of the available pressure.
Static tower purifiers are also fitted with trays to eliminate compaction
for
lower pressure applications (Gollmar,
1945).
7.3
Metal
Oxide
Processes
175
Conventional boxes consist
of
large rectangular vessels, either built
into the ground, or supported
on
legs
to
save footprint. They are com-
posed
of
several layers, with a typical bed depth
of
at least
two
feet.
Iron-oxide suspensions, like iron-oxide slurries, rely upon hydrated
ferric oxide as the active regenerable agent. However, iron-oxide sus-
pensions react
in
a basic environment with an alkaline compound,
followed by the reaction
of
the hydrosulfide with iron oxide to form
iron sulfide (Kohl and Nielsen,
1997).
The iron is then regenerated
by
aeration.
Generally, the iron-oxide process is suitable only for small
to
mod-
erate quantities of hydrogen sulfide. About
90%
of the hydrogen sul-
fide can be removed per bed, but bed clogging by elemental sulfur
occurs, the bed must be discarded, and the use
of
several beds
in
series
is not usually economical. Removal
of
larger amounts of hydrogen
sulfide from gas streams requires a continuous process, such as the
Ferrox process or the Stretford process. The Ferrox process is based
on
the same chemistry as the iron-oxide process, except that
it
is fluid
and continuous. The Stretford process employs a solution containing
vanadium salts and anthraquinone disulfonic acid (Maddox,
1974).
The natural gas should be wet when passing through an iron sponge
bed because drying of the bed will cause the iron sponge to lose its
capacity for reactivity.
If
the gas is not already water-saturated or
if
the
influent stream has a temperature greater than 50°C (about
120"F),
water with soda ash is sprayed into the top
of
the contactor
to
main-
tain the desired moisture and alkaline conditions during operation.
7.3.2
Other Processes
SZuny
processes
were developed as alternatives to iron sponge. Slurries
of
iron oxide have been used
to
selectively absorb hydrogen sulfide
(Fox,
1981;
and Kattner et al.,
1986;
Samuels,
1988).
The chemical
cost for these processes is higher than that for iron sponge process,
but this is partially offset by the ease and lower cost with which the
contact tower can be cleaned out and recharged. Obtaining approval
to dispose
of
the spent chemicals, even if they are non-hazardous, is
time consuming.
An
example
of
such
a
process
is
the
Sulfi-check process,
which selec-
tively removes hydrogen sulfide and mercaptans from natural gas
in
176
Chapter7
Processes
the presence of carbon dioxide (Bhatia and Allford,
1986;
Dobbs,
1986).
The process uses sodium nitrite (NaNO,).
Gas streams with elevated oxygen levels with the Sulfa-Check process
will produce some nitrogen oxides in the gas stream (Schaack and
Chan,
1989).
Removal of hydrogen sulfide is not affected under short
contact times because the reaction is almost instantaneously
(Hohlfeld,
1979).
Sodium hydroxide and sodium nitrite are consum-
ables in the processes and cannot be regenerated. This process is
accomplished in a one-step single vessel design using an aqueous
solution of sodium nitrite buffered to stabilize the pH above
8.
Also,
there is enough strong base to raise the pH of the fresh material to
12.5.
The reaction with hydrogen sulfide forms elemental sulfur,
ammonia, and caustic soda:
NaNO,
+
3H,S
t)
NaOH
+
NH,
+
3s
+
H,O
Other reactions forming the oxides of nitrogen do occur (Burnes and
Bhatia,
1985),
and carbon dioxide in the gas reacts with the sodium
hydroxide to form sodium carbonate and sodium bicarbonate. The
spent solution is slurry of fine sulfur particles in
a
solution of sodium
and ammonium salts (Manning and Thompson,
1991).
The
SuZfuTreutprocess
is a batch-type process for the selective removal
of
hydrogen sulfide and mercaptans from natural gas. The process is
dry, using no free liquids, and can be used for all natural gas applica-
tions where
a
batch process is suitable. The SulfaTreat system
is
a
more recent development using iron oxide on
a
porous solid mate-
rial. Unlike the iron sponge process, the SulfaTreat material is non-
pyrophoric and has
a
higher capacity than iron sponge on
a
volu-
metric or mass basis.
In the process, the inlet gas stream enters the unit through a small
inlet separator to separate out any entrained liquid, salt water, and
any liquid hydrocarbon in the gas stream. The sour feed gas then
passes to the top of the contactor where the SulfaTreat product
(a
dry,
granular, free-flowing material of uniform porosity and permeability)
only reacts with sulfur-containing compounds. This eliminates any
side reactions with carbon dioxide that which could reduce its effi-
ciency,
As
the SulfaTreat bed dries out, the rate
of
reaction decreases.
Optimum efficiency may require measuring water content
of
the gas
7.4
Methanol-Based Processes
177
and injecting a sufficient amount
of
water into the influent gas to
maintain a water-saturated gas and evidence of free water in the
outlet gas.
7.4
Methanol-Based
Processes
Methanol is probably one
of
the most versatile solvents
in
the natural
gas processing industry. Historically, methanol was the first commer-
cial organic physical solvent and
has
been used
for
hydrate inhibition,
dehydration, gas sweetening, and liquids recovery (Kohl and Nielsen,
1997).
Most
of
these applications involve low temperatures, where
methanol’s physical properties are advantageous compared with other
solvents that exhibit high viscosity problems or even solids formation.
Operation at low temperatures tends to suppress methanol’s most sig-
nificant disadvantage, high solvent loss. Furthermore, methanol is rel-
atively inexpensive and easy to produce, making the solvent a very
attractive alternate for gas-processing applications.
Methanol has favorable physical properties relative to other solvents
except for vapor pressure. The benefits of the low viscosity
of
meth-
anol at low temperature are manifested
in
the pressure drop improve-
ment in the cold box
of
injection facilities and improved heat
transfer. Methanol has a much lower surface tension relative to the
other solvents. High surface tension tends to promote foaming prob-
lems in contactors. Methanol processes are probably not susceptible
to foaming. However, the primary drawback
of
methanol is the high
vapor pressure that is several times greater than that of the glycols or
amines.
To
minimize methanol losses and enhance water and acid-gas
absorption, the absorber or separator temperatures are usually less
than
-20°F.
The high vapor pressure
of
methanol may initially appear to be a sig-
nificant drawback because
of
high solvent losses. However, the high
vapor pressure also has significant advantages. Although often not
considered, lack
of
thorough mixing
of
the gas and solvent can pose
significant problems. Because
of
the high vapor pressure, methanol is
completely mixed
in
the gas stream before the cold box. Glycols,
because they do not completely vaporize, may require special nozzles
and nozzle placement
in
the cold box
to
prevent freeze-up. Solvent
carry-over to other downstream processes may also represent a signifi-
cant problem. Because methanol is more volatile than glycols, amines,
and other physical solvents including lean oil, methanol
is
usually
rejected in the regeneration step
of
these downstream processes. The
178
ChaDter
7
Processes
stripper concentrates the methanol
in
the overhead condenser, where
it can be removed and further purified. Unfortunately,
if
glycols are
carried over to amine units, the glycol becomes concentrated in the
solution and potentially starts to degrade and possibly dilute the
amine solution.
The use of methanol has been further exploited
in
the development of
the Rectisol process either alone or as toluene-methanol mixtures are
used to more selectively remove hydrogen sulfide and slip carbon
dioxide to the overhead product (Ranke and Mohr, 1985). Toluene has
an additional advantage insofar as carbonyl sulfide is more soluble in
toluene than
in
methanol. The Rectisol process was primarily devel-
oped to remove carbon dioxide and hydrogen sulfide (along with
other sulfur-containing species) from gas streams resulting from the
partial oxidation of coal, oil, and petroleum residua. The ability of
methanol to absorb these unwanted components made it the natural
solvent
of
choice. Unfortunately, at cold temperatures, methanol
also has a high affinity for hydrocarbon constituents of the gas
streams. For example, propane is more soluble
in
methanol than is
carbon dioxide. There are two versions of the Rectisol process
(Hoochgesand, 1970)-the two-stage and the once-through. The first
step of the two-stage process
is
desulfurization before shift conver-
sion; the concentrations
of
hydrogen sulfide and carbon dioxide are
about
1
and 5% by volume, respectively. Regeneration of the meth-
anol following the desulfurization
of
the feed gas produces high
sulfur feed for sulfur recovery. The once-through process is only
applicable for high-pressure partial oxidation products. The once-
through process is also applicable when the hydrogen sulfide to
carbon dioxide content is unfavorable, in the neighborhood of
150
(Esteban et al.,
2000).
Recently, a process using methanol has been developed that has the
simultaneous capability to dehydrate, to remove acid gas, and to con-
trol hydrocarbon dew point (Rojey. and Larue, 1988; Rojey et al.,
1990). The
IFPEXOL-I
is used for water removal and hydrocarbon
dew point control; the
IFPEXOL-2
process is used for acid gas
removal. The novel concept behind the IFPEXOL-1 process is to use a
portion
of
the water-saturated inlet feed to recover the methanol
from the aqueous portion of the low temperature separator. That
approach has solved a major problem with methanol injection in
large facilities, the methanol recovery via distillation. Beyond that
very simple discovery, the cold section
of
the process is remarkably
similar to a basic methanol injection process. Modifications to the