Note: Descriptions are shown in the official language in which they were submitted.
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
CATALYSTS, SYSTEMS AND METHODS OF STEAM REFORMING, AND METHODS
OF MAKING STEAM REFORMING CATALYSTS
FIELD OF THE INVENTION
The invention relates to catalysts, systems and methods of steam reforming;
and methods
of making steam reforming catalysts.
BACKGROUND OF THE INVENTION
Steam reforming is a chemical process in which a hydrocarbon is reacted with
steam
(Hz0) to form hydrogen (HZ), CO and CO2. For decades, steam reforming has been
the principal
industrial process for making hydrogen. More recently, steam reforming has
attracted great interest
as a possible means to supply hydrogen for fuel cells. There have been
intensive research efforts
over many years to improve the steam reforming process. Despite these efforts,
problems continue
to exist with catalyst performance and/or cost, and the need for catalyst
replacement or
regeneration due to the rather harsh conditions (high temperature and steam)
in which steam
reforming is typically conducted.
Known steam reforming catalyst support materials include gamma alumina doped
with a
stabilizing element such as magnesium, lanthanum, and barium. Incorporation of
a stabilizer could
delay the irreversible transformation of gamma alumina into alpha alumina.
However, under
severe steam reforming conditions the supports continually sinter, resulting
in a permanent loss of
active sites.
The effect of rare earth elements and alkaline earth elements as dopants for
alumina has
been investigated. See Church et al., Appl., Catal., A, 101 (1993) 105.
Wachowski, et al., Mater.
Cheyo. and Phys., 37 (1994) 29. Some studies have indicated that there is a
correlation between
final surface area of alumina treated at high temperatures (1200°C) and
ionic radius of the dopant.
Mizukami et al., in A. Cruca (Ed.), Studies ira sufface sciefTCe arad
catalysis, Vol 71, Elsevier
Amsterdam; 1991, 557, reported that La3+ and Ba'+ are the most effective
dopants among rare
earths and alkaline earths, which are large ions with high ionic charges.
Yeung et al., J. Membrane
Sci., 131 (1997) 9, described procedures for making mesoporous alumina
membranes by slip
casting a 1 M alumina sol (6 wt%) containing 3 atom% of a nitrate of Y, Ba,
La, and Ce. Das et al.
in Appl. Catal. A: 207, 95-102 (2001), reported that a more stable La-doped
support could be
obtained by processing in ethanol rather than water. Martin et al. in Appl.
Catal. A: 131, 297-307
(1995) described experiments in which Rh w%as sintered on alumina under H~ at
temperatures of
700 to 900 °C; under these conditions, the mean Rh particle size varied
from 1.1 to 2.1 nm.
Temperature Programmed Reduction (TPR) data was also reported. Schaper et al.,
in SirrterifZg-
-1-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
Theory and Pi°actice, Proceedings of the St'' International Round Table
Conference on Sintering,
Material Science Monographs, vol. 14, pp173-176, reported on the influence of
Mg, Cr, La and Zr
on the stability of gamma alumina catalyst supports and concluded that only
lanthanum resulted in
a significant improvement in stability against sintering.
Treatment in 10% flowing H20 has been used to pre-age a support for a
catalytic
convertor. McCabe et al. in J. Catal. 151, 385-393 (1995), reported that
presteaming alumina at
1223 K with 10% HBO in flowing air for 24 hours prior to Rh impregnation
eliminated irreversible
occlusion of Rh during subsequent hydrothermal aging.
Wang et al. in US published patent application 2003/00317105 described methods
and
catalysts for steam reforming. In one preferred embodiment, the catalyst
includes Rh on a Mg-Al
spinel. This reference does not describe pretreatment of the support at
elevated steam pressure.
Despite extensive efforts over many years, there remains a need for catalytic
systems and
steam reforming methods that have hlgh performance and stability under the
hydrothermal
conditions that are typical of steam reforming. Also, despite extensive
efforts, there remains a need
for better methods of making steam reforming catalysts.
SUMMARY OF THE INVENTION
In a first aspect, the invention provides a method of making a steam reforming
catalyst
composition, comprising: providing a coating composition comprising an alumina
precursor;
applying the composition onto a substrate to form a coated substrate;
providing an alumina stabilizer; drying the coated substrate to form a dried
support; and
hydrothermally pre-aging the dried support in a gas atmosphere comprising at
least 1 atm H20
(partial pressure) and a temperature of at least 850°C to form a
hydrothermally stabilized support.
In another aspect, the invention provides a steam reforming catalyst,
comprising:
a Mg and A1 containing spinel support; and Rh on the surface of the support;
and
being characterizable by a stability and reactivity such that, when tested in
a flowing stream of
methane and water in a water:methane ratio of 3 at 15 atm, a test reactor wall
temperature of 880
°C (this temperature is measured by a thermocouple attached to a 0.020
inch (0.051 cm) thick
metal wall (InconelOO 617 if available) opposite a point where the catalyst
(or finned catalyst
support) is in direct contact with the wall; this temperature is believed to
be within 10 °C of the
actual catalyst skin temperature), and a contact time of 5 ms or less that is
adjusted to obtain a
methane conversion (after 100 hours time-on-stream (TOS)) of 70% at a pressure
of 15 atm (total
combined pressure of methane and steam), and then maintaining the same
conditions for 1000
hours, there is a continuous period of at least 400 hours (during the 1000
hours) in which the
methane conversion changes by 3% or less. Methane conversion may be
conveniently measured by
-2-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
gas chromatography. To measure this property, it is important to provide
sufficient heat to
maintain temperature. This should be done by placing the catalyst inside a
half flattened metal rod
with a rectangular channel that is 5 cm long, 0.3 inch (7.6 mm) wide and 0.02
inch (0.5 mm) high,
and loading with a length of 0.7 inch (1.8 cm) of catalyst. One side of the
metal rod has been
removed to place a thermocouple adjacent to the catalyst chamber (thus, the
"flattened" form). If
the form of the catalyst makes this placement impossible, catalyst can be
tested in a packed bed in
a more conventional tube, if necessary with dilution of the catalyst with
alumina in order to
prevent cold spots. The invention further includes a catalyst system that
includes this catalyst,
preferably in the presence of steam and a hydrocarbon.
For purposes of this invention, the protocol for analyzing catalyst properties
uses the
technique for temperature measurement that is described in the aspect above.
The test reactor for
measuring catalyst properties should be the half flattened metal rod that is
described in the
Examples section. This test procedure is chosen because it best conforms with
the data; in a less
preferred embodiment, the catalyst properties can be alternatively defined as
being measured at a
catalyst skin temperature of 870 °C.
Unless stated otherwise, "conversion percent" refers to absolute conversion
percent
throughout these descriptions. "Contact time" is defined as the total catalyst
chamber volume
(including the catalyst substrate volume) divided by the total volumetric
inlet flowrate of reactants
at standard temperature and pressure (STP: 273K and 1.013 bar absolute).
Catalyst chamber
20, volume includes any volume between a catalyst coating (or other flow-by
catalyst arrangement)
and the opposite wall of a reaction channel.
In a further aspect, the invention provides a composition comprising a Mg and
AI
containing spinel material. This composition has a hydrothermal stability such
that, when tested by
treating the composition at 925 °C and 15 atm H20, for I 00 hours, the
support has a surface area
of at least 10 m'/g. This composition is intended for use as a catalyst
support or a precursor for a
catalyst support. In a preferred embodiment, the composition is a catalyst
composition comprising
Rh on the surface.
In another aspect, the invention provides a method of methane steam reforming
that
comprises conducting a reaction in which methane and steam contact a catalyst
at a temperature of
at least 800 °C for at least 1000 hours of continuous operation (i.e.,
operation without
regeneration), wherein the catalyst comprises a stabilized alumina support and
a catalytically
active material; and fin-ther characterized by one or more of the following
sets of conditions and
process characteristics: (a) a contact time of 15 ms or less while achieving
an average approach to
equilibrium conversion of greater than 80%, preferably greater than 90%, still
more preferably
greater than 95%, and most preferably greater than 98%; (b) a contact time of
10 ms or less while
-3-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
achieving an average approach to equilibrium conversion of greater than 70%,
preferably greater
than 80%, still more preferably greater than 90%, and most preferably greater
than 95%; or (c) a
contact time of 5 ms or less while achieving an average approach to
equilibrium conversion of
greater than 65%, preferably greater than 75%, still more preferably greater
than 85%, and most
preferably greater than 90%.
For all of the above conditions, the approach to equilibrium conversion is the
ratio of
measured hydrocarbon conversion to equilibrium hydrocarbon conversion, as
shown below:
(znoles hydracaz°bon in-moles hydrocarbon oat
nzoles hydrocarbon. in measured
(moles hydrocarbon itz)yzzeasuf°ed-moles hydrocarbon out)e~uilibriunz
moles hydrocarbon in measured
The equilibrium composition (or moles hydrocarbon out at equilibrium) is based
upon the
measured average pressure of the inlet and outlet of the reactor zone and the
inlet molar
composition. The equilibrium distribution (or composition) for a given
temperature, pressure, and
inlet mole fraction distribution can be calculated using Gibbs free energies
with programs such as
IS theNASALEWIS equilibrium code or FACTSAGE.
For purposes of this invention, the temperature used to describe a method,
such as the
above-described aspect, is defined as the peak temperature (that is, the
highest temperature) in the
catalyst (sometimes also referred to by the equivalent teen "peak catalyst bed
temperature").
"Average approach to equilibrium conversion" is based on this peak
temperature. Also, other
~ variables such as contact time and productivity, when used to describe a
method in conjunction
with a temperature, are based Oll 0171y those parts of the catalyst that are
within 20 °C of the peak
temperature. When used to describe the inventive methods, temperatures refer
to catalyst skin
temperatures that can be measured directly (this is preferred) or calculated
by techniques known in
the art. In the tests described in the Examples section, the temperature at
any time in the catalyst
chamber was essentially isothermal (that is, the temperature at any instant is
believed to have
varied by less than 20 °C).
In some preferred embodiments, the maximum catalyst temperature is 1050
°C, more
preferably 1000 °C or less, more preferably 950 °C or less, and
in some embodiments, 900 °C or
less. Additional, or narrower parameters, that may be used in preferred
embodiments are described
in the Detailed Description section.
In some embodiments, the temperature along the length of a catalytic reforming
channel
(preferably a microchannel) may be nearly isothermal, as defined within 20
°C from inlet to
outlet. In some alternate embodiments, the temperature may vary along the
length of the
-4-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
microchannel. For example, the temperature at the inlet of the catalyst zone
of a microchannel
may be considerably cooler than the outlet of the microchannel. In one example
for methane
steam reforming, the inlet temperature of the catalyst zone may be around 650
°C and the end of
the catalyst section may be around 850 °G. The temperature may rise
monotonically in a linear
fashion or may rise more quickly near either the front or end of the catalyst
bed. Thus, in some
examples, the section of the catalyst containing microchannel that exceeds 800
°C may only
include the final 75%, or 50%, or 25%, or 10% of the catalyst bed, or any
value therewithin. For
the example of a temperature proftle ranging from 650 °C to 850
°C, the reaction may equilibrate
near 840 °C and demonstrate an approach to equilibrium greater than 80%
as defined by the peak
temperature. The equivalent contact time spent in the reaction zone that
exceeds 800 °C may be
considerably less than the overall reaction contact time as defined by the
entire reaction channel
volume (i.e., the volume of the channel containing catalyst). As an example,
the contact time
within the entire reaction channel volume may be 5 ms, but only 1 ms in the
reactor section that
exceeds 800 °C. It is envisioned that the contact time spent in the
reaction zone exceeding 800 °C
required to achieve at least 80% approach to equilibrium defined at peak
temperature will exceed
0.1 ms. In some embodiments, the temperature of the catalyst-containing
microchannel may be
highest near the end of the reaction zone, or, alternatively, may be higher at
the front or middle of
the reactor rather than near the end of the reaction zone.
In another aspect, the invention provides a method of methane steam reforming,
comprising: passing a steam comprising methane and water in a steam to carbon
ratio of 3 or less
through a catalyst at a contact time of 10 ms or less and maintaining the same
conditions for at
least 1000 hours, wherein there is a continuous period of at least 500 hours
in which the methane
conversion changes by 3% or less; and wherein at least 2 mol of methane are
converted per gram
catalytically active material each minute. '
In a further aspect, the invention provides a steam reforming catalyst
comprising: a
stabilized alumina support; and rhodium. This catalyst is characterizable by a
stability such that,
when tested according to Test Procedure 1 (see Detailed Description), after
595 hours of
continuous operation, the conversion of methane is at least 80% of equilibrium
and the conversion
is diminished by less than 5%.
A "stabilized alumina suppou" comprises a stabilizing element that enhances
stability in
steam reforming conditions. As is conventional patent terminology,
"comprising" means including
and when this term is used the invention can, in some narrower preferred
embodiments, be
described as "consisting essentially off' or in the narrowest embodiments as
"consisting of."
Aspects of the invention described as "comprising a" are not intended to be
limited to a single
component, but may contain additional components.
-5-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
In a further aspect, the invention provides a method of making a steam
reforming catalyst,
comprising: providing a support material comprising alumina and at least one
stabilizing agent
selected from the group consisting of magnesium, lanthanum, barium, and
strontium; and heat
treating the support material by heating in the range of 850 to 1100
°C, in the presence of oxygen;
and depositing a steam reforming catalyst metal. The step of heat treating is
conducted in an
atmosphere containing less than 5 % H20.
In another aspect, the invention provides a steam reforming catalyst
comprising: a
stabilized alumina suppol-t; and catalytically active particles on the surface
ofthe stabilized
alumina support; wherein the catalytically active (preferably rhodium)
particles have a mean
particle size in the range of 2.0 to 4.0 nm, more preferably 2.5 to 3.5 nm, as
measured by electron
microscopy, where the mean is based on the number (not mass) of particles. In
some embodiments,
the catalyst comprises a stabilized alumina support with Rh particles on the
surface of the
stabilized alumina support; wherein the rhodium particles have an average size
of 2.0 to 3.0 nm.
In a further aspect, the invention provides a method of making a steam
reforming catalyst
comprising: providing a stabilized alumina support; depositing Rh on the
stabilized alumina
support; and reducing the Rh in a HZ-containing stream at a temperature of 200
to 300 °C.
In some aspects, the present invention is defined as catalysts or catalyst
systems containing
stabilized alumina and Rh that are characterized by surprisingly good
stability and conversion or
selectivity. It may be subsequently discovered that other supports or catalyst
metals may perform
equivalently if substituted for the Mg stabilized alumina and/or Rh 111 these
catalysts or catalyst
systems; however, the present inventors are not presently aware of any such
equivalent materials.
In some other aspects, the invention is described more broadly; for example,
the invention
discloses superior methods (both synthetic methods as well as steam reforming
methods) that
encompass a broader range of materials.
The invention includes methods of steam reforming, reactors, and fuel
processing systems
that use the catalysts described herein. Furthermore, methods of making
catalyst according to the
present invention include: 1 ) methods of making highly stable suspended
catalyst particles in
liquid medium, 2) methods of modifying the liquid suspension properties to
achieve an adherent,
uniform coating on a reactor wall or on catalyst substrates such as felts,
foams, ceramics and metal
flat surfaces, 3) coating methods to achieve uniform catalyst coatings with
excellent adhesion to a
substrate in short catalyst preparation time (e.g., in dip coating, reduced
dips).
Various embodiments of the invention can provide numerous advantages such as
one or
more of the following: improved catalyst performance, high stability under
steam reforming
conditions (steam, high temperature and high pressure), high conversions at
relatively short contact
times, selectivity control, lower cost, ease of manufacturing, and low
temperature operation.
-6-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
BRIEF DESCRIPTION OF THE FIGURES
Figure 1 is a representational, schematic, cross-sectional view of reactor
apparatus that can
be utilized in some embodiments of the present invention.
Figure 2 shows of X-ray diffraction patterns of 6%Mg/A1~03 treated
under (a) thermal and (b) hydrothermal conditions.
Figure 3 is a plot of BET surface area vs. thermal aging time for sol-based 3%
La/A1203,
10% Mg/AIz03; and 7% Ba/Ah03, 5% La1A1~03 using : ;-Ah03 as starting material.
The sol-
based material was thermally treated at 1050°C while others were
treated at 1075°C in air at
atmospheric pressure.
Figure 4a shows steam reforming activity of a finned, coated catalyst system
whose
support has been hydrothermally treated at near ambient pressure (about 0.7
atm HBO). Methane
conversion (0) is indicated in the range of 80 to 90% equilibrium selectivity,
CO selectivity (~) in
the range of 60 to 70% and carbon balance (o) near zero.
Figure 4b shows steam reforming activity including methane conversion (0), CO
selectivity (~), and carbon balance (o) over 600 hours time on stream of a
Rh/LalAlZ03 catalyst
whose support was thermally treated at 1050 °C for SO 110urS. The
vertical lines indicate system
upsets during testing.
Figure 5 shows the results including methane conversion (0), CO selectivity
(0), and
carbon balance (o) of testing a catalyst of Rh/MgA120~ that was prepared by
sol coating a finned
substrate followed by steam pre-aging at 25atm, 925°C for 100 hours
under Hz0/He=2/1. Methane
steam reforming was conducted for 2800 hours of operation at l5atm, 3 to l
steam-to-carbon ratio,
4.5-3.0 ms contact time, and reactor wall temperature of 869 °C. A step
change in contact time
(vertical line) was introduced at 2150 hours to move further away from the 840
°C equilibrium
conversion.
Figure 6 shows the results including methane conversion (0), CO selectivity
(0), and
carbon balance (o) of testing a catalyst of Rh/MgAl20d that was prepared by
sol coating a finned
substrate followed by steam pre-aging at 25atm, 925°C for 100 hours
under H~O/He = 211. The
metal loading was 20 mg/in' (25wt% Rh oxide). Methane steam reforming was
conducted at
27atm, 3 to 1 steam-to-carbon ratio, 3.0-4.2 ms contact time, and reactor wall
temperature of 870-
873°C.
Figure 7 shows the results including methane conversion (0), CO selectivity
(0), and
carbon balance (o) of testing a catalyst of Rh/MgAh04 that was prepared by sol
coating a finned
substrate followed by thermal pre-aging at 1050 °C for 50 hours.
Methane steam reforming was
conducted at 15 atm, 3 to I steam-to-carbon ratio, 3.0-4.5 ms contact time,
and reactor wall
temperature of about 840 to 950 °C.
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
Figure 8 illustrates testing results including methane conversion (0), CO
selectivity (~),
and carbon balance (o) of a Rh/MgA1,04 catalyst that was prepared by sol
coating a finned
substrate, and without thermal or hydrothermal pretreatment of the support.
Figure 9 shows a plot representing Hz consumption as a function of temperature
on fresh
Rh/MgA1,04 catalyst that was oxidized 500°C.
Figure 10 illustrates the SMR activity of certain powder catalysts measured as
a function
of reduction temperature.
Figure 11 is a plot of feed rate per gram Rh (N) versus equilibrium methane
conversion
approach temperature difference (~TapProach)~
DESCRIPTION OF PREFERRED EMBODIMENTS
A steam reforming process requires a hydrocarbon (or hydrocarbons) and steam
(HZO). A
reactant mixture can include other components such as CO or nonreactive
diluents such as nitrogen
or other inert gases. In some preferred processes, the reaction stream
consists essentially of
hydrocarbon and steam. In some preferred embodiments, the steam to carbon
ratio in a reactant
stream is 3 to 1 to 1 to 1, and in some embodiments 1.5 to 1 or less.
Hydrocarbons according to the present invention include: alkanes, alkenes,
alkynes,
alcohols, aromatics, and combinations thereof including fuels such as
gasoline, kerosene, diesel,
JP-8. Preferably, the hydrocarbon is an allcane or a fuel. Preferred alkanes
are C~ - Coo alkanes,
such as methane, ethane, propane, butane, and isooctane.
The catalyst requires catalytically active surface sites that reduce the
kinetic barrier to the
steam reforming reaction. The catalyst comprises one or more of the following
catalytically active
materials: ruthenium, rhodium, iridium, nickel, palladium, platinum, and
carbide of group VIb.
Rhodium is particularly preferred. The catalytically active materials are
typically quite expensive,
therefore it is desirable to minimize the amount used to accomplish the
desired performance. In
some preferred embodiments, the catalyst (including all support materials)
contains 0.5 to 10
weight percent, more preferably 1 to 3 wt% of the above-mentioned
catalytically active materials.
In some preferred embodiments, the sum of stabilized alumina support plus
catalytically active
materials contains 0.5 to 10 weight percent, more preferably 1 to 7 wt%, and
still more preferably
2 to 5 wt% catalytically active materials.
The catalyst also contains an alumina support for the catalytically active
materials. An
''alumina support" contains aluminum atoms bonded to oxygen atoms, and
additional elements can
be present. Preferably, the alumina support comprises stabilizing element or
elements that improve
the stability of the catalyst in hydrothermal conditions. Stabilizing elements
typically are large,
highly charged cations. Examples of stabilizing elements are Mg, Ba, La, and
Y, and combinations
_g_
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
of these. In this invention, a "stabilized alumina support" is an alumina
support containing at least
one stabilizing element. Preferably, the stabilized alumina support contains 1
to 10, more
preferably 3 to 7 weight percent of stabilizing elements. Several workers have
reported that La is a
better stabilizer than Mg; with Schaper et al., reporting that while La
stabilized the alumina, Mg
had no effect. We have surprisingly found that Mg works better than La in
stabilizing a support in
hydrothermal conditions. Therefore Mg-AI spinel is a particularly preferred
support for steam
reforming.
Preferably, the catalytically active materials (such as Rh) are present in the
form of small
particles on the surface of the stabilized alumina support. More preferably,
the particles of
catalytically active material have a mean particle size in the range of 2.0 to
4.0 nm, more
preferably 2.5 to 3.5 nm, as measured by electron microscopy, where the mean
is based on the
number (not mass) of particles.
The catalyst can be in the form of particles, preferably having diameters less
than 4 mm,
more preferably less than 1 mm. More preferably, the stabilized alumina forms
a layer (of
agglomerated particles or a continuous i~ihn) having a thickness less than 4
mm, more preferably
less than I mm, and still more preferably a thickness of less than 40 p.m.
This layer is preferably
disposed on a porous substrate. Preferably the catalyst contains an alumina
layer disposed on a
thermally conductive surface. The surface could be, for example, a porous
substrate or reaction
chamber wall(s).
In some embodiments, a stabilized alumina layer is coated over, and preferably
in direct
contact with, a high surface area material such as alumina, preferably (gamma)-
alumina. This
coal-iguration provides high surface area for good metal dispersion and/or
high metal loadings and
also provides a stabilized alumina layer for excellent stability. The high
surface area material is
porous; the meaning of a stabilized alumina "disposed over" or "coated over" a
high surface area
material means that the stabilized alumina may also coat crevices and cavities
within a high
surface area material or within a large pore substrate.
In some preferred configurations, the catalyst includes an underlying large
pore substrate.
Examples of preferred large pore substrates include commercially available
metal foams and, more
preferably, metal felts. Prior to depositing any coatings, a large pore
substrate has a porosity of at
least 5%, more preferably 30 to 99%, and still more preferably 70 to 98%. In
some preferred
embodiments, a large pore substrate has a volumetric average pore size, as
measured by BET, of
0.1 ~.m or greater, more preferably between 1 and 500 pm. Preferred forms of
porous substrates
are foams and felts and these are preferably made of a thermally stable and
conductive material,
preferably a metal such as stainless steel or FeCrAIY alloy. These porous
substrates can be thin,
such as between 0.1 and 1 mm. Foams are continuous structures with continuous
walls defining
-9-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
pores throughout the structure. Felts are nonwoven fibers with interstitial
spaces between fibers
and includes tangled strands like steel wool. Felts are conventionally defined
as being made of
nonwoven fibers. In one embodiment, the large-pore substrate has a corrugated
shape that could be
placed in a reaction chamber (preferably a small channel) of a steam reformer.
Various substrates
and substrate configurations are described in U.S. Patent Applications Ser.
No. 09/640,903 (filed
Aug. I 6, 2000), U.S. Patent No. 6,680,044, which is incorporated by
reference. Another preferred
substrate is a finned substrate that is characterized by the presence of fins
(such as square-wave
type fins) on the substrate's surface. Alternatively, the catalyst may take
any conventional form
such as a powder or pellet.
A catalyst with a large pores (and including the alumina-supported
catalytically active
sites) preferably has a pore volume of 5 to 98%, more preferably 30 to 95% of
the total porous
material's volume. Preferably, at least 20% (more preferably at least 50%) of
the material's pore
volume is composed of pores in the size (diameter) range of 0.1 to 300
microns, more preferably
0.3 to 200 microns, and still more preferably 1 to 100 microns. Pore volume
and pore size
distribution are measured by mercury porisimetly (assuming cylindrical
geometry of the pores)
and nitrogen adsorption. As is kIlOWll, mercury porisimetly and nitrogen
adsorption are
complementary techniques with mercury porisimetry being more accurate for
measuring large pore
sizes (larger than 30 nm) and nitrogen adsorption more accurate for small
pores (less than 50 nm).
Pore sizes in the range of about 0.1 to 300 microns enable molecules to
diffuse molecularly
through the materials under most gas phase catalysis conditions.
In some preferred embodiments, the catalyst comprises a metal, ceramic or
composite
substrate having a layer or layers of a catalyst material or materials
deposited thereon. The porosity
can be geometrically regular as in a honeycomb or parallel pore structure, or
porosity may be
geometrically tortuous or random. Preferred porous support materials include
felts (nonwoven
fibers or strands), foams (including a foam metal or foam ceramic), ins and
honeycombs. In
embodiments employing a porous substrate, the average pore size (volume
average) of the catalyst
layers) is preferably smaller than the average pore size of the substrate.
In a particularly preferred embodiment, alumina is deposited on a finned metal
substrate
and a catalytic metal is deposited on the alumina. Preferably, this support is
a thermally conductive
metal that is sized to fit within a microchannel. Alternatively, the finned
support could be
fabricated directly within the microchannel and be integral to the
microchannel. One method of
fabrication within a microchannel comprises the use of a slitting saw, partial
etching using a
photochemical process, or a laser EDM. This type of support provides numerous
advantages
including: high heat flLlX Wlth Short heat transfer distances, high surface
area, and low pressure
drop. Preferably, the support has a height (including ins) of less than 5 mm
and preferably less
-10-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
than 2 mm and a fin-to-fin separation of 1000 p.m or less, and in some
embodiments, a fin-to-fin
separation of 150 to 500 pm. The fin structure can be integral with a reaction
chamber (and thus
coated in situ), or as a separate insert that can be coated prior to being
inserted into a reaction
chamber.
In some embodiments, the catalyst, including the presence of catalytically
active surface
sites, as measured by BET, has a volumetric average pore size of less than 0.1
micrometer (p.m).
The catalyst, including the presence of catalytically active surface sites, as
measured by BET,
nitrogen physisorption, preferably has a surface area of more than 10 m'/g,
and in some
embodiments a surface area of 20 to 500 mz/g.
Conventional catalysts typically undergo large decreases in surface area when
exposed to
hydrothermal conditions. Preferred catalysts of this invention have a surface
area, as measured by
N~ adsorption BET, of at least 5 mz/g, more preferably at least 10 m'/g, and
in some embodiments
5 to about 20 m'/g; and preferably maintain these surface areas after exposure
to 25 atm steam for
I 00 hours.
Certain aspects of the invention can best be described in terms of properties
such as
stability, conversion or selectivity. Both the catalysts and methods can be
characterized in terms of
hydrocarbon conversions and selectivities in steam reforming processes.
Hydrocarbon conversion
is preferably at least 50%, more preferably at least 80% and still more
preferably at least 90%.
The foregoing conversion values can be either absolute or equilibrium
conversions. If not
specified, it is conventional to consider conversion values to be absolute
conversions. Under
conditions where conversion approaches 100%, absolute and equilibrium
conversion is the same.
"Equilibrium conversion" is defined in the classical manner, where the maximum
attainable
conversion is a function of the reactor temperature, pressure, and feed
composition. For the case of
hydrocarbon steam reforming reactions, the equilibrium conversion increases
with increasing
temperature and decreases with increasing pressure. In some embodiments,
hydrocarbon
equilibrium conversion is in the range of 70 to 100%.
Hydrogen selectivity, defined as moles H atoms in H~ in the product gas
divided by moles
H in all product gases, is preferably at least 50%, more preferably at least
60%, still more
preferably at least 85%, and yet still more preferably at least 95%. For some
embodiments, carbon
monoxide selectivity is preferably less than 65%, more preferably less than
40%.
In some preferred embodiments, the catalyst is characterizable by a stability
and reactivity
such that, when tested in a flowing stream of methane and water in a
water:methane ratio of 3 at 15
atm, a catalyst temperature (actually a reactor wall temperature as explained
above) of 880 °C, a
reactor pressure of 15 atm (note that in the reactor design described in the
Examples section,
pressure drop through the reactor is negligible) and a contact time of 5 ms or
less that is adjusted to
-11-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
obtain a methane conversion (after 100 hours TOS) of 70%, and maintaining the
same conditions
for 1000 hours, there is a continuous period of at least 400 hours (more
preferably at least 600
hours, and in some embodiments from 400 to about 2800 hours TOS) in which the
methane
conversion changes by 3% or less. To measure this property, it is important to
provide sufficient
heat to maintain temperature. This should be done by placing the catalyst
inside a metal tube with
a rectangular channel cut through the long axis of the metal tube, the channel
being 5 cm long, 0.3
inch (7.6 mm) wide and 0.02 ltlch (0.5 mm) high, and loading into tile
c11a1111e1 a length of 0.7 inch
( 1.8 cm) of catalyst. If the form of the catalyst makes this impossible,
catalyst can be tested in a
more conventional tube, if necessary with dilution of the catalyst with
alumina in order to prevent
cold spots. The invention further includes a catalyst system that includes
this catalyst, preferably in
the presence of steam and a hydrocarbon.
" In some preferred embodiments, the catalysts, systems and methods can be
characterized
by a stability such that after 595 hours of continuous operation in
hydrocarbon steam reforming
conditions , the conversion of hydrocarbon is at least 70%, more preferably at
least 80%, still more
preferably at least 85%, and in some embodiments 80 to about 90% of the
equilibrium conversion,
and where the conversion has diminished by less than 10%, more preferably less
than 5%, and still
less preferably less than 3% over the 595 hours of continuous operation.
Selectively to CO is
preferably essentially unchanged over the course of the 595 hours of
continuous operation.
"Systems" are formed by a catalyst in conjunction with a reactor, preferably a
microreactor.
Some preferred embodiments of the inventive catalysts and methods may also be
described in terms of their exceptionally high activity. Preferably, the
catalyst possesses a
catalytic activity such that, when exposed to a stream of CH4 and HBO at a
steam-to-carbon ratio of
3, at a catalyst temperature of 840 °C and a pressure of 27 atm at a
flow rate of 6000 SLPM
(standard liters per minute measured at 0 °C and 1 atm) total flow per
gram of catalytically active
material, the catalyst convents methane to products at a rate of at least 2
mol CHd/g catalytically
active material, more preferably at least 10, still more preferably at least
25, and in some
embodiments 3 to about 50 mol CH4/g catalytically active material. This
activity is preferably
maintained for at least 400 hours, more preferably at least 600 hours.
In another aspect, the invention provides a method of methane steam reforming,
comprising: passing a steam comprising methane and water in a steam to carbon
ratio of 3 or less
through a catalyst at a contact time of l 0 ms or less and maintaining the
same conditions for at
least 1000 hours, wherein there is a continuous period of at least 500 hours
in which the methane
conversion changes by 3% or less; and wherein at least 2 mol of methane are
converted per gram
of catalytically active material each minute.
-12-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
In some embodiments, a catalyst's properties (such as stability, conversion
and selectivity)
are defined by the fol lowing test procedure (referred to as "Test Procedure 1
"). In this test
procedure, about 7 mg of catalyst is loaded into a reaction chamber that is a
0. I 62 inch x 0.021
inch x 2.0 117C11 (4.1 x 0.53 x 51 mm) slot in a 0.5 inch (1.3 cm) diameter
solid Inconel 617 rod. If
necessary, the catalyst can be crushed, cut or scraped from a larger piece of
catalyst to obtain a
catalyst that will fit in the slot. The catalyst can be on a finned support
(as described in the
examples) or in other forms such as pellets or powders. The catalyst is
reduced to an optimal
extent, as can be determined through routine experimentation. The reduction is
under flowing
hydrogen, typically in the range of 300 to 500 °C. Then the reactor is
pressurized to 15 atm under
an inert atmosphere, and then heated in nitrogen to the reaction temperature.
In this test procedure,
the steam reforming reaction is carried out at 850 °C, a steam:methane
ratio of 3:1, at 15 atm and a
contact time of 4.5 milliseconds. Catalyst packing can be varied to obtain
this pressure and contact
time; if the exact conditions cannot practically be obtained, results can be
extrapolated to estimate
results at the specified conditions. In this procedure, pressure drop through
the reaction chamber is
I S preferably 10 psig or less. In preferred embodiments of a catalyst tested
with this procedure there
is at least 85% methane conversion, and, in some embodiments 80 to about 90%
methane
conversion. CO selectivity is preferably at least 50% and remains constant.
The present invention also provides methods of steam reforming in which a
hydrocarbon
is reacted with water vapor at short residence times (or alternatively,
described in contact times)
over the catalysts described herein. The residence time is preferably less
than 0.1 s. Short contact
times are preferably 5-100 milliseconds (cosec), in some embodiments, 10-25
cosec.
The steam reforming reaction is preferably carried out at more than 400
°C, more
preferably 500-1000°C, and still more preferably 650-900°C. The
reaction can be run over a broad
pressure range from sub-ambient to very high, in some embodiments the process
is conducted at a
pressure of from 10 atm to 30 atm, more preferably 12 atm to 25 atm. The HBO
partial pressure is
preferably at least 0.2 atm, in some embodiments at least 2 atm, and in some
embodiments in the
range of 5 atm to 20 atm.
The catalyst support is made from a composition that includes an alumina
precursor. An
"aluminum precursor" is any form of aluminum (such as an alumina slurry) that
can be used to
form solid alumina. The catalyst may be made starting from an alumina sol or
solid alumina.
Suitable, commercially available materials include colloidal alumina suspended
in aqueous
medium from Sasol, or Engelhard alumina ground to a particle size of 70-I 00
mesh.
Starting from alumina particles, a slurry can be prepared by milling in water
or organic
phase. Stabilizing elements can be added as solids or in solution. It has been
discovered that
adding acid to adjust the pH to 3 to 6, preferably 5 to 6, results in a stable
slurry even for solid
-13-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
contents greater than 20 wt%, and in some cases solid content in the range of
I 5 to 25 wt%.
Adding a binding agent to the milled alumina can improve slurry
characteristics.
Once alumina has been stabilized, it can be dried before a subsequent pre-
aging treatment.
Drying can be accomplished by removing or evaporating a liquid phase ofthe
slurry; alternatively
the slurry can be coated on a porous substrate and then dried prior to a
subsequent pre-aging
treatment.
A stabilized alumina is preferably subjected to a pre-aging treatment
comprising a
hydrothernal or thermal treatment, prior to applying a catalytically active
material. In a
hydrothernal treatment, the support material is exposed to steam at high
temperature. Based on the
results for which we have comparative data, we have discovered that thermal
treatment of La-
stabilized alumina at ambient conditions produces superior results as compared
to hydrothernal
treatments of La-stabilized alumina conducted at atmospheric pressure. In
thermal pre-aging, the
support materials are thermally treated at high temperatures in the absence of
added water.
Preferred conditions for the thernal treatment include heating to a
temperature of from I 000 to
I 150 °C. A particularly preferred set of conditions for a
nonhydrothernal treatment is heating at
I 050-1100°C for 24-50 hours in air at atmospheric pressure.
On the other hand, hydrothermal treatment of Mg-stabilized alumina at elevated
pressure
resulted in supports that are superior to both ambient pressure hydrothertnal
treatments and
thermal treatments. Preferred hydrothernal pre-aging conditions include at
least 5 atm partial
pressure of steam, and at least 900 °C, for a duration of at least 80
hours. More preferred
conditions include at least 10 atm steam partial pressure, and in some
embodiments 5 to 25 atm
steam; temperatures of at least 900 °C, more preferably at least 950
°C, and in some embodiments
900 to 1100 °G; and times of at least 100 hours, and in some
embodiments 50 to 300 hours.
When an underlying substrate is used, an alumina slurry or sol can be coated
over the
substrate at any stage in the preparative process. For example, particles of a
stabilized and heat
treated alumina can be slurry coated onto the substrate followed by
depositing, drying and
activating a metal via the impregnation method. Alternatively, a vapor coat or
soluble form of
alumina (or other high surface area material) could be applied onto a
substrate prior to addition of
a catalytic metal. In another embodiment, the substrate may be coated with a
buffer layer formed
111 SltL1 Llslllg chemical vapor deposition. The buffer layer may not have a
high surface area, but
may be used to create a layer with a CTE (coefficient of thermal expansion)
between the base
metal substrate and the higher surface area catalyst support to promote good
adhesion of the layers.
The buffer layer may also be used to inhibit corrosion of the base metal
substrate by creating a
near dense and almost pin-hole free coating. Although solution (such as spray
coating) or slurry
coating is typically less expensive, vapor coating of the various materials
could also be employed.
-14-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
The hydrothermal or thermal pre-aging treatment can be conducted either before
or, preferably,
after applying the alumina to a substrate.
A catalytically active material can be deposited onto alumina using known
techniques such
as the incipient wetness technique. Preferably, a precursor of the
catalytically active material is
added after the pre-aging treatment.
As described in the Examples section, the reduction temperature range for an
alumina/Rh
catalyst was studied by a temperature-programmed reduction technique. Although
sintering might
be expected at higher temperatures, it was surprisingly discovered that the
activity of a catalyst
reduced at 300°C was better as compared to an identical catalyst
reduced at 125 °C. In some
preferred embodiments, catalysts are reduced in a hydrogen atmosphere at
temperatures in the
range of 200-400 °C, more preferably 250-350 °C; reduction at
these temperatures results in
substantially improved activity. It was found that the metal dispersion of a
catalyst, as measured
using a hydrogen chemisorption technique, showed higher dispersion on the
catalyst reduced at
300°C as compared to an identical catalyst reduced at 125 °C.
The present invention includes methods and systems in which a steam reforming
catalyst
is disposed within a microchannel reaction channel. The height and/or width of
a reaction
microchannel (wall-to-wall, not counting catalyst) is preferably 5 mm or less,
and more preferably
2 mm or less, and in some embodiments 50 to 1000 pin. Both height and width
are perpendicular
to the direction of flow. The length of a reaction channel is parallel to flow
through the channel
and is typically longer than height and width. Preferably, the length of a
reaction chamber is
greater than 1 cm, more preferably in the range of 1 to 100 cm. Typically, the
sides of the reaction
channel are defined by reaction channel walls. These walls are preferably made
of a hard material
such as a ceramic, an iron based alloy such as steel, or a nickel-based alloy.
In some preferred
embodiments, the reaction chamber walls are comprised of stainless steel or
inconel which is
durable and has good thermal conductivity.
In addition to the reaction channel(s), additional features such as
microchannel or non-
microchannel heat exchangers may be present. Microchannel heat exchangers are
preferred.
Adjacent heat transfer microchannels enable temperature in the reaction
channel to be controlled
precisely to promote steam reforming and minimize unselective reactions in the
gas phase. The
thickness of a wall between adjacent process channels and heat exchange
channels is preferably 5
mm or less. Each of the process or heat exchange channels may be further
subdivided with parallel
subchannels. The heat exchange fluids can be gases or liquids and may include
steam, liquid
metals, or any other known heat exchange fluids. Especially preferred heat
exchangers include
combustors in which a fuel is oxidized to produce heat for the steam reforming
reaction. The
incorporation of a simultaneous exothermic reaction to provide an improved
heat source can
-15-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
provide a typical heat flux of roughly an order of magnitude above the
convective cooling heat
flux. The flow of hot fluid through a heat exchanger may be cross flow,
counter-flow or co-flow.
The reactors preferably include a plurality of microchannel reaction channels
andlor a
plurality of adjacent heat exchange microchannels. A plurality of microchannel
reaction channels
may contain, for example, 2, I 0, 100, 1000 or more channels. In some
preferred embodiments,
multiple heat exchange layers are interleaved with multiple reaction
microchannels (for example,
at least 10 heat exchanger layers interleaved with at least 10 layers of
reaction microchannels.
Typically, flow into and/or out of some or all of a plurality of reaction
channels passes through a
manifold or manifolds that combines the fluid flow. In preferred embodiments,
microchannels are
arranged in parallel arrays of planar microchannels.
Preferred reactors usable in the present invention include those of the
microcomponent
sheet architecture variety (for example, a laminate with microchannels).
Examples of integrated
combustion reactors that could be used in the present invention are described
in U.S. patent
application serial no. 10/222,196, filed Aug. 15, 2002, which is incorporated
herein by reference.
Some other suitable reactor designs and methods of making reactors are
disclosed in U.S. patent
application serial no. I 0/306,722, filed November 27, 2002, and I 0/408,744,
filed April 7, 2003,
which are also incorporated herein, in full, by reference.
The catalyst can fill up a cross-section of the reaction channel (a flow-
through catalyst) or
only occupy a portion of the cross-section of a reaction channel (flow-by).
The use of a flow-by
catalyst configuration can create an advantageous capacity/pressure drop
relationship. In a flow-by
catalyst configuration, gas preferably flows in a 0.1-2.0 mm gap adjacent to a
porous insert or a
t11111 layer of catalyst that contacts the microchannel wall (preferably the
microchannel wall that
contacts the catalyst is in direct thermal contact with a heat exchanger,
preferably a heated fluid or
exothermic reaction process stream contacts the opposite side of the wall that
contacts the
catalyst).
The reaction channel contains a steam reforming catalyst. Suitable catalyst
structures
include (but are not limited to) porous catalyst materials, fins, washcoats,
pellets, and powders. In
one preferred embodiment, a reaction channel contains a catalyst material that
defines at least a
portion of at least one wall of a bulk flow path. In this preferred
embodiment, the surface of the
catalyst defines at least one wall of a bulk flow path through which the
mixture passes. During
operation, the mixture flows through a microchannel, past and in contact with
the catalyst in the
microchannel. The term "bulk flow path" refers to an open path (contiguous
bulk flow region)
within the reaction chamber. A contiguous bulk flow region allows rapid gas
flow through the
reaction chamber without large pressure drops. In preferred embodiments there
is laminar flow in
the bulk flow region. Bulk flow regions within each reaction channel
preferably have a cross-
-16-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
sectional area of 5 x 10-8 to 1 x 10-2 m', more preferably 5 x I 0-~ to 1 x 10-
ø m'. The bulk flow
regions preferably comprise at least 5%, more preferably 30-80% of either 1 )
the internal volume
of the reaction chamber, or 2) the cross-section of the reaction channel. One
example of a bulk
flow path is the space between fins in a finned catalyst. When a combustion
reaction is used to
heat the steam reforming reaction chamber in an integrated combustion reactor,
the combustion
reaction preferably contains a bulk flow path having the properties discussed
above.
In some embodiments, the catalyst is provided as a porous insert (such as a
foam
monolith) that can be inserted into (or removed from) each channel in a single
piece; preferably the
porous insert is sized to fit within a microchannel with a width of less than
2 tnm. In some
embodiments, the porous catalyst occupies at least 60%, in some embodiments at
least 90%, of a
cross-sectional area of a microchannel. In an alternative preferred
embodiment, the catalyst is a
coating (such as a washcoat) of material within a microchannel reaction
channel or channels.
One embodiment of a reactor 2 is shown in cross-section in Fig. 1. The
reaction chamber
4 contains catalyst 6 and has an inlet 8 and outlet 10. In Fig. I, the
catalyst is shown on the top
and bottom of the reaction chamber with an open channel from the reactor inlet
to the outlet - this
configuration is called "flow-by." Other configurations, such as "flow-
through" where flow is
directed through a porous catalyst, are, of course, possible. To improve heat
transfer, a
microchannel heat exchanger 12 can be placed in contact with the reaction
chamber. The
microchannel heat exchanger 12 has channels 14 for passage of a heat exchange
fluid. These
channels 14 have at least one dimension that is less than 2 mm. The distance
from the channels 14
to catalyst 6 is preferably minimized in order to reduce the heat transport
distance. In preferred
embodiments, a reaction chamber 4 is connected to fuel tank 16 such that fuel
from the tank can
flow into the reaction chamber. Although a fuel tank is shown in Figure l, it
should be recognized
that any fuel source, such as a pipeline could be used. A liquid fuel stream
may flow through a
separate vaporizer or be vaporized within a section of the steam-reforming
reactor. In some
preferred embodiments a fuel is vaporized in a microchannel vaporizer and/or
preheated in a
microchannel preheater. The product gases (including H~) then may either flow
into hydrogen
input device 22, or the product of the reforming reactor may flow into a water
gas shift reactor to
convert some of the carbon monoxide into carbon dioxide and additional
hydrogen. Additionally,
the product stream may flow into a secondary clean up process to further
purify hydrogen or
reduce carbon monoxide. The secondary clean-up process may include a
preferential oxidation
reactor, membrane separation of either hydrogen or carbon monoxide, a sorption
based separation
system for either hydrogen or carbon monoxide, and the like. These elements
form a highly
simplified fuel processing system 30. In practice, fuel processing systems
will be significantly
more complex. Typically, heat from a combustor will be used to generate heat
for other processes
-17-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
such as generating steam (not shown) that can be utilized for steam reformer
and/or water gas shift
reactor. The hydrogen input device 22, can be, for example: a fuel cell, a
storage tank, a refueling
station, a hydrocracker, hydrotreater, or an additional hydrogen purifier. In
a fuel cell, the HZ is
combined with OZ to generate electricity. Various fuel cells are well-known
and commercially
available and need not be described here.
EXAMPLES
To test stability and surface area, supports were prepared from a slurry in a
two-step
process and from a sol in a one-step process. The resulting supports were
analyzed after various
pretreatments.
SlurrX Coating
Support powder was wet-milled in water or organic phase to a uniform particle
size of 0.5
um. The pH was adjusted before the wet milling to 3-6, more specifically, 5-6.
At this pH, the
viscosity of slurry was lowered, allowing a higher solid concentration in the
slurry while
IS maintaining an effective milling. At these conditions, the catalyst
particles are slightly charged and
repelled from each other. The resulting catalyst slurry is highly stable with
minimum particle
agglomeration. Therefore, sedimentation rate was retarded, and uniform
suspension in the liquid
was maintained for days without the need of constantly agitating the slurry.
After the wet milling, binding agents such as alumina sol or surfactant were
added to the
slurry to improve interparticle adhesion. Addition of binders was done after
the wet milling
process, since binder could increase the slurry viscosity, reducing the
overall effectiveness of wet
milling. The viscosity of the slurry is adjusted once more just prior to
coating the substrates by
adding dilute nitric acid after the milling. This can improve flow during the
coating process to
achieve uniform coverage of catalyst slurry on the substrate with a high solid
loading slurry. With
high solid loading slurry, the number of dip coating cycles can be minimized
to achieve the target
catalyst loading on the substrates.
Dip coating or spin coating was used for coating substrates. Dip coating
applies to foam or
felt substrates with inner interstitial spaces, while spin coating was used
for flat surfaces such as a
ceramic plate or metal foil. In the dip coating process, uniformity of coating
was achieved by
withdrawing the substrate that was fully immersed in the slurry vertically at
a fixed rate, between 5
cm/min and 20 cm/min. The coated substrate was then dried under controlled
drying to achieve a
crack-free coated catalyst layer. The dip coating process was repeated until
the target weight gain
was achieved. After drying the last coat, the substrate was annealed at high
temperature. This step
was carried out to improve the adhesion and to decompose the surfactant, if
surfactant was added
as binding agent.
-18-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
Example I : A 17wt% -22wt% solid loading catalyst slurry was prepared by
mixing powder
with deionized water. Two different powders were used for the ball mill,
standard Mg0-A1,03
spinet and hydrothermally treated Mg0-A1,O3 spinet. The hydrothermally treated
spinet has a
higher density than the fresh spinet due to the formation of alpha alumina
during the hydrothermal
treatment. Solid/liquid mixture was milled in a 2" ball mill bottle along with
3 mm alumina
grinding balls for 24 or 62 h, depending on the type of the starting material.
Standard MgO-A1z03
required 24h of milling while the hydrothermally treated Mg0-A1~03 required 62
h to achieve an
average particle size that is less than 1 p.m. After the milling, particle
size analysis showed average
particle sizes of 0.5 p.m and 0.8 pm, for standard and hydrothermally treated
catalysts,
respectively. Nitric acid was added to adjust the pH to 4, and 1 wt% of PQ
A1203 is added as
binding agent. Changing the pH to 4.5-5 significantly reduced the slurry
viscosity. Therefore, by
modifying the pH, the slurry can be prepared with higher solid loadings, which
reduces the number
of dipping or spin coating in the coating steps, without compromising coating
quality. In addition,
it was found that the sedimentation rate of the higher solid loading slurries
( 17 to 22 solids%) was
comparable to a pH-unmodified slurry with a much lower solid loading of 1 I %.
A catalyst slurry with 13.8 wt% solid loading was separately prepared by
mixing slurry
from Example I (above) with PQ A1ZO3, surfactant polyoxyethylene 10-lauryl
ether (C~ZEO~o) and
2 wt% HNO~. The final weight ratio of slurry was: support powder :PQ Ah03:
C,ZEO~o= 100:5:8,
pH = 3.4.
Metal foils and metal felts were coated with the slurries. FeCrAIY
intermetallic alloy in
the form of felt (2"x0.35"x0.01") structure was fired at 900°C for 2h
in a box furnace. At this
temperature, A1 migrates to the surface and oxidizes to form a native alpha
A1~03 scale. The
formation of a uniform layer of metal oxide scale on the intermetallic alloy
substrates significantly
improves the adhesion between the oxide coating and the alloy substrate. It is
believed that the
scale reduces the thermal expansion mismatch between the ceramic coating and
the metal
substrate, which could lead to spalling of catalyst layer. Upon cooling to
room temperature, the felt
was coated with the slurry from Example 1. The felt was first immersed
vertically in the slung,
then withdrawn at a constant rate of 5-10 cm/min, dried at room temperature
for 10 min, then dried
at I 10°C in vacuum oven. The dip coating step was repeated 3 times to
achieve the target weight
loading density of 0.09 g cat/in''. After the final drying, the substrate was
calcined at 5 °C /min to
500 °C, and held at that temperature for 2h. A control piece was made
using the slurry without any
modifications, dip coating without controlling the rate of substrate
withdrawal. By modification of
slurry pH, the slurry became less viscous, allowing a higher solid
concentration without affecting
the flow characteristic during dip coating. As a result, a uniform catalyst
layer can be obtained in
minimum number of dips. The dip cycle was reduced from 5 to 3 with an increase
in solid loading
-19-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
of from I 1 wt% to 22.20 wt%. Examining the coated substrate under a
microscope showed a
uniform coating layer with minimal cracks. In addition, adhesion was tested by
sonicating the
samples in petroleum ether for 30 min showing significant enhancement in
adhesion for the slurry
modified with 1 wt% of PQ Ah03. The weight loss at the end of the sonication
test was
negligible.
FeCrAIY foil (0.02" thick) was cleaned in acetone and isopropanol prior to
firing to 900°C
for 2h. A layer of native alpha A1z03 formed due to diffusion of A1 to the
surface. The substrate
was then spin coated at 600 rpm for 30 seconds with slurry, dried at 200-
300°C on a hot plate for
0.5-1 min, the spin coating was repeated for 3-6 times to achieve the target
weight loading, then
calcined at 500 °C for 2 hrs. A hydrothermal test was performed on this
foil at 900 °C with 60
HBO for 100h at 12 atm in a sealed Inconel tube. After the hydrothermal test,
the coated foil then
underwent thermal cycling from 100 °C to 900 °C at a heating and
cooling rate of 5 °C/min for 10
times. No spalling of the catalyst powder was observed at the end of both
tests.
Catalysts for comparing pre-aging treatments were prepared as described below.
The
catalyst was reduced under a 10% H~ in N~ stream at 300 °C for 1 hour.
A finned substrate was coated with both slurry (see above) and sol (see below)
coating
techniques. The finned substrate, which is more stable than felts in prolonged
steam reforming
conditions, were a thin FeCrAIY plate with surface ftns. Width at the fin base
is 0.16 inch (4.1
nnn), length is 0.7 111CI7 (1.8 mm), channel width is 0.15 inch (0.38 mm) and
depth is 0.14111ch (3.6
mm); the fins are separated by 0.10 inch (2.5 mm) wide and 0.14 111ch (3.6 mm)
tall walls; forming
a total of 6 channels through the finned support. These supports were cleaned
and heat-treated to
grow thin aluminum oxide layer. Each substrate was cleaned ultrasonically in 2-
propanol and then
20wt% nitric acid for 20 minutes. After rinsing with deionized water, the fins
were dried at 100°C
for 30 minutes. The fins were then placed in a furnace in air and heated to
1000°C at a rate of
3.5°C/min and held at 1000°C for 8 hours, then allowed to cool
to room temperature slowly
(~3.5°C/min).
Solution (Sol) Coating Method:
A plate with surface fins was solution coated with alumina sol (AIOOH) and a
solution containing
stabilizing ion as follows:
I .Alumilia sol (SASOL 14N4-25) was coated onto the fin surface by submerging
the
flll 111t0 the sol (ex-situ or if2-situ). Remove the excess sol.
2.Dry the coated fin at 100°C at a rate of 3.5 °C/min for 1
hour.
3.Calcine the fin by heating to 450°C at a rate of 3.5°C/min and
hold at 450°C for 4
hours, then allowed to cool to room temperature (3.5°C/min).
4.Stabilizing ion such as lanthanum ion, barium ion or magnesium ion was
impregnated onto the alumina coated fin using 1 Owt% aqueous solution of
lanthanum nitrate, barium acetate, or magnesium nitrate by submerging the fin
into tl7e SOhlt1011. Excess solution was removed.
-20-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
S.Dry the coated fins at 100°G at a rate of 3.5 °C/min for
1 hour.
6.Calcine the fin by heating to 1000°C at a rate of 3.5°C/min
and hold at 1000°C for 4
hours, then cooled to room temperature (3.5°C/min).
7.Pre-age (if utilized) the fins using the procedures in the following
section.
8.hnpregnate Rh metal onto the pre-aged fin by dipcoating into a l Owt%
solution of
rhodium nitrate.
9.Dry the coated fins at 100°C at a rate of 3.5 °C/min for 1
hour.
1 O.Calcine the fin by heating to 950°C at a rate of 3.5°C/min
and hold at 950°C for 1
hour, then allow them to cool to room temperature slowly (3.5°C/min).
An X-ray diffraction pattern of the thermally pre-aged material was compared
to that of
hydrothermally treated (at 1 atm) material in Figure 2. Figure 2(a) shows X-
ray diffraction pattern
of 6%Mg/A1,0~ thermally treated at 1100°C for 24 hours whereas Figure
2(b) shows the pattern
for the same composition material hydrothermally treated at 900°C and
l2atm for I 00 hours in a
flow of steam and nitrogen mixture at a ratio of 2 to 1. It was found that the
major phases consist
of alpha alumina and spinel phases. The relative intensity of peaks indicating
the spinel phase was
higher on the thermally treated sample (a) than the hydrothermally treated
sample (b).
Crystallinity of both samples was similar to each other, which is in agreement
with the BET
surface area results.
Surface Area
In thermally pre-aged samples, La was found to be more effective than Mg in
stabilizing
the surface area. See Fig. 3. After aging at 1100 °C for 24h, XRD
analysis of a La stabilized A1203
sample showed a combination of delta, theta, and gamma alumina phases. Gamma
alumina has
the highest surface area among different phases of alumina.
Thermal treatment of sol-based lanthanum-, barium-, and magnesium-doped
alumina was
carried out in a powder form and BET surface area was measured. The thermally
stabilized La-
doped alumina was found to be more stable than the sample that was
hydrothermally treated at
ambient pressure. This trend was not observed for the material that was
hydrothermally treated at
25 atm - in the samples hydrothermally pre-aged under increased pressure, the
Mg-doped support
was found to be the most stable. The Table below shows surface area stability
of variously
prepared alumina powders. The sol-derived powder exhibited the greatest
surface area after
hydrothermal treatment. Among three formulations (one-step MgAh04, two-step
Mg/A1z03, and
two-step LalAh03), the Mg-containing supports had similar surface area around
44 m2/g but the
La-containing support had 59mz/g. Steam aging at 25 atm decreased the surface
area by 86% and
94% on two-step 6%MglAh03 and two-step 3%La/Ah03, respectively. However, the
degree of
surface area reduction was much smaller (54%) on one-step 6%MgAhOd. This could
be possibly
due to the formation of more spinel (less isolated y-alumina which transforms
to a-alumina with
increasing temperature) in the one-step method because of the intimate contact
of Mg nitrate with
-21 -
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
AIOOH that can be achieved through this method. See the following table in
which the atmosphere
comprises 60% water, so H20 partial pressure at 25 atm is 15 atm.
S.A.
Treatment One-step 6%MgAlz04 reduction
m2/g (%)
1000C, 4hrs, air, 92.01
I atm
925C, 20hrs, steam,65.06 29.3
I atm
925C, 1 OOhrs, steam,44.17 52.0
I atm
925C, 100hrs, steam,
25atm 20.15 78.1
1050C, 50hrs 39.84 56.7
1050C, 100hrs
S.A.
Two-step 6%Mg/A1z03reduction
mz/g (%)
1000C, 4hrs, air, 81.97
I atm
925C, 1 OOhrs, steam,46.24 43.6
1 atm
975C, 200hrs, steam,46.29 43.5
I atm
925C, 100hrs, steam,
25 atm 6.37 92.2
1050C, 50hrs 27.50 66.5
1050C, IOOhrs 12.20 85.1
Two-step ' S.A.
0.4wt%Mg/A1~03,surfreduction
m'/g (%)
1050G, 100hrs 48.84
S.A.
Two-step 3%La1A1203reduction
m'/g (%)
1000C, 4hrs, air, 105.30
1 atm
925C, I OOhrs, steam,59.30 43.7
1 atm
925C, IOOhrs, steam,
25atm 3.28 96.9
1050C, 50hrs 68.09 35.3
1050C, 100hrs 79.50 24.5
S.A.
Two-step 3%La/A1203/surfreduction
m'/g (%)
1000C, 4hrs, air, 65.33
1 atm
1050C, 100hrs 54.60 16.4
Performance And Stability In Methane Steam Reformine Conditions
The FeCrAIY fins were used as support substrate for washcoat materials. After
heat-
treatment of the fins to form a surface oxide, they were coated with a layer
of Dispal 14N4-25
alumina sol obtained from SASOL and a layer of doping solution. The synthesis
method has been
-22-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
described earlier. Steam methane reforming (SMR) reaction tests were carried
out on fresh,
hydrothermally-treated and thermally-treated catalysts at l5atm, 850°C,
S/C=3/I, CT=4.5ms. A
fresh (not preaged) RhlLa/A1~03 catalyst (25mg/in'') tested over 750 hours of
life testing, showed
steady methane conversion and carbon monoxide selectivity for approximately
250 hours;
however, activity of the catalyst showed a continual decline over the next 500
hours. Figure 4a
shows activity of catalyst whose support has been hydrothermally treated at 1
atm. The support
(La/AhO;) was pre-treated under steam and nitrogen mixture at 2 to 1 ratio at
near ambient
pressure (about 0.7 atm Hz0) and heated (at a ramp up rate of 3.5
°C/min) to 950°C, l5atm, and
held at these conditions for I 00 hours. Then, the support was cooled at about
3.5 °C/min. The total
loading of the catalyst was 30.3 mg/in2. Over 500 hours on stream, methane
conversion fell by
6%. Figure 4b shows the SMR activity over Rh/La/AIz03 catalyst whose support
was thermally
treated (at a ramp up rate of 3.5 °C/min) to 1050 °C and held at
1050 °C for 50 hours. Then the
support was cooled at about 3.5 °C/min. The total catalyst loading was
31.4 mg/in 2. Over 600
hours on stream at the same conditions, the activity of the catalyst decreased
by less than 3% and
did not experience the deactivation that was observed in fresh or
hydrothermally treated (at
ambient pressure) catalysts. This catalyst also survived system upsets
(inadvertent shut downs and
start ups) without any loss of catalyst performance.
A slurry of an La-doped alumina was coated onto a metal foil and tested for
850 hours at
12 atm, 825 °C, S/C=2/1, CT = 9 ms, showed a drop in methane conversion
from 60% to 48% over
the first 600 hours on stream. Thus, the catalysts prepared from a sol possess
superior stability as
compared with the slurry-derived catalyst.
Fig. 5 shows the results of testing a catalyst of Rh/MgAhO4 whose support was
prepared
by sol coating a finned substrate followed by steam pre-aging at 25 atm,
925°C for 100 hours
under H~O/He=2/1. The fin was then coated with Rh nitrate solution (I% Rh) and
calcined at
950°C for 1 hour. The metal loading was 13.6 mg/in' (9.4 wt% Rh oxide
measured by weight gain
after calcinations assuming all Rh in form of R1y03). Methane steam reforming
was conducted for
2800 hours of operation at 15 atm, 3 to I steam-to-carbon ratio, 4.5-3.0 ms
contact time, and
reactor wall temperature of 869 °C. A step change in contact time was
introduced at 2150 hours to
move further away from the 840 °C equilibrium conversion. There were no
system upsets which
contributed to a smooth operation for 2800 hours TOS. As can be seen, there
was no change in
activity fox over 2000 hours of continuous operation. Notably, methane
conversion remained at
70% ~3% for at least 600 hours (between 2200 and 2800 hours) at a catalyst
operating temperature
of less than 900 °C and a contact time of 5 ms or less.
Fig. 6 shows the results of testing a catalyst of RhlMgA1,04 whose support was
prepared
by sol coating a finned substrate followed by steam pre-aging at 25atm,
925°C for 100 hours under
- 23 -
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
H~O/He = 2/I . The metal loading was 20 mg/in2 (25wt% Rh oxide). Methane steam
reforming was
conducted at 27 atm, 3 to I steam-to-carbon ratio, 3.0-4.2 ms contact time,
and reactor wall
temperature of 870-873 °C. At 650 hours time-on-stream (TOS), the steam
to carbon ratio dropped
due to a water pump malfunction and resulted in a loss of activity; however,
after conditions were
restored, the catalyst reached another steady-state level at 64%. Thus, two
periods of at least 400
hours of steady state operation were observed; one at 70% conversion and one
at 64% conversion.
In Fig. 6, the horizontal lines at 76 and 78% are equilibrium conversions at
840 and 850 °C,
respectively. The vertical line at 90 hours indicates the change to 3 ms
contact time.
Performance of Thermal Pre-aced Catalyst
Fig. 7 shows the results of testing a catalyst of Rh/MgA1~04 whose support was
prepared
by sol coating a finned substrate followed by thermal pre-aging at I 050
°C for 50 hours in air. The
metal loading was 3 I mg/in' (20 wt% Rh oxide) (a thicker coating than the
last example). Methane
steam reforming was conducted at I 5 atm, 3 to 1 steam-to-carbon ratio, 3.0-
4.5 ms contact time,
and reactor wall temperature of about 840 °C. In Fig. 7, the two
vertical lines at about 100 hours
and the vertical line at about 350 hours indicate system upsets. After
demonstrating a steady
methane conversion at 88% for 600 hours, the conditions were changed to
reflect a kinetically
controlled regime by lowering the contact time to 3.0 ms at 625 hours TOS
without compensating
for the drop in temperature due to the higher flow rate used. The activity was
steady for 200
hours. At 800 hrs on stream, furnace temperature was increased by 20°C
(while maintaining the 3
ms contact time) to match the temperature of the reactor at 4.5 ms contact
time, and over the next
500 hours, the catalyst steadily deactivated, dropping from 84% conversion to
78% conversion.
After 1300 hours, the conditions were changed back to the initial conditions
(875 °C, 4 ms) and it
was found that the methane conversion level had dropped from 88% to 80%. In
order to recover
the initial conversion level of methane (~88%), it was necessary to raise the
furnace temperature
by 55°C.
Performance of Non-aced Catalyst
Fig. 8 shows the results of testing a catalyst of Rh/LaA1204 whose support was
prepared
by sol coating a finned substrate followed by drying. The metal loading was 25
mg/in'' Methane
steam reforming was conducted for 750 hours. Up to 250 hours, conditions were
changed several
times (vertical lines) at 850 or 875 °C and contact times between 2.5
and 6 ms with one system
shut down at 25 hours. At 250 hours, conditions were set to 15 arm, 3 to I
steam-to-carbon ratio,
4.5 ms contact time, and catalyst reactor wall temperature of about 875
°C. By 750 hours TOS,
methane conversion dropped 8%, probably due to a phase change of y-A1,0~ to a-
A1z03 under the
reaction conditions, which in turn reduced the surface area significantly,
thus, encapsulating some
of Rh sites.
-24-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
Effect of Reduction Temperature
The temperature range over which Rh/Mg0-AI203 catalysts (fresh and thermally
stabilized) was reduced was studied using temperature-programmed reduction
(TPR) technique
with 5% Hz in Ar. Figure 9 presents a plot of Hz consumption as a function of
temperature on the
fresh catalyst that was oxidized 500°C. There is a broad peak starting
at 70°C and tailing off to a
baseline at about 280°C with multiple shoulders, indicating hydrogen
adsorption/absorption on the
surface and the bulk of the metal. The plot reveals that most of catalyst
reduction is completed at
200°C. Similar results were obtained on the pre-stabilized catalyst.
Metal dispersion of the fresh and pre-stabilized (PS) SMR catalysts reduced at
125°C and
300°C were measured using hydrogen chemisorption technique and
presented in Table 4.
Table 4 Variation on dispersion % on catalysts supported on
fresh and re-stabWzedmel
s mth
function
of
reduction
tem
erature
Reduction T 10% Rh on fresh 5% Rh on PS spinel
spinel
(C)
125 17.0 15.8
3 00 24.0 15.9
On 10%Rh on fresh spinet, the dispersion increased by 7% while on the pre-
stabilized support,
there was only a small improvement observed. This could be explained by the
significant
difference in specific surface areas of the two supports (158 mz/g (fresh) vs
l Omzlg ( pre-
stabilized)).
The activity of the powder catalysts, which were diluted with alpha alumina by
20wt%,
were measured as a function of reduction temperature as shown in Figure 10.
Both catalysts
(10%Rh/fresh spinet and 5%Rh/pre-stabilized spinet) showed an improvement in
catalyst activity
when they were reduced at 300°C rather than 125°C. The relative
increase in methane conversion
rate for fresh and pre-stabilized powder catalysts were 22% and l3%,
respectively.
Particle Size
On supports made of a hydrothennally-treated magnesia-stabilized alumina
support having
a surface area about 10 m'/g, Rh nitrate solution was added using the
incipient wetness technique.
The Rh was reduced at 300 °C. Hydrogen chemisorption experiments showed
the metal dispersion
(i.e., the percent of the total metal that is exposed on the surface) of 1 %
Rh, 5% Rh, and 10% Rh
on the pre-aged support to be 23%, 15%, and 10%, respectively. Transmission
Electron
Microscopic analysis of the catalysts showed that the Rh particles on the 10%
Rh/spinel catalyst
were about 1.5 nm in size. The Rh particle size on the 5% Rhlpreaged spinet
catalyst ranged from
1.4 to 4.0 nm with an average particle size of about 2.7 nm.
-25-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
CatalXst Activity and Productivity
Catalyst activity and productivity were compared against the best prior art
literature
(Wang et al., US published patent application 2003100317105). The approach
temperature
difference for the catalyst and the gram moles of methane fed per gram
catalytically active material
(in preferred embodiments, per gram Rh) per minute can be compared to the
upper and lower
limits of the operating space described by the function below:
1°CxN<~T°p~,~wp~n, <50°CxN
where N is the gram moles of methane fed per minute per gram catalytically
active material (in
preferred embodiments, per gram Rh). For example, the values ofN and
~Taupr°a~,, were calculated
for one condition of this study as follows: 0.122 SLPM (or 0.00545 mol/min) of
methane was fed
with steam at a molar ratio of 3:1 steam to carbon to an engineered catalyst
loaded with 0.00024
grams Rh catalyst. Thus, for this example, N = (0.00545 mol GH4lmin) l
(0.00024 g Rh) = 23 mol
CH4lmin/g Rh. The average of the measured inlet and outlet pressure across the
engineered
catalyst at the time the data was taken was about 198 psig (14.7 bar). The
calculated peak catalyst
bed temperature at this same condition was 856 °C. The measured methane
conversion at this
condition (feeding 3:1 steam to carbon at 14.7 bar and 891 °C at 23 mol
CH4/min/g Rh) was
85.3%. Using the NASA Lewis equilibrium code to calculate gas phase
equilibrium for a 3:1
steam to carbon mixture at 14.7 bar, it is found that the equilibrium methane
conversion to CO and
CO~ is 85.3% at 825 °C. In other words, chemical equilibrium of a 3:1
steam:C mixture at 14.7
bar and 825 °C represents a conversion of 85.3% of the methane in the
original mixture to carbon
monoxide and carbon dioxide. Therefore the ~TaPProach for this example is 856
°C - 825 °C = 31
°C.
The "approach temperature difference" (~Tapproach) is the difference between
the "peak
catalyst bed temperature" (as defined in the Summary section for methods of
steam reforming) and
the "apparent equilibrium conversion temperature". The difference is a measure
of catalyst
performance, with the difference increasing as the chemical reaction kinetics
of a given catalyst
become slow enough that the equilibrium composition is not reached at the
average measured
reaction temperature.
The "apparent equilibrium conversion temperature" is the apparent temperature
based
upon the conversion of the reactant in question. It is the temperature
required to produce an
equilibrium composition equivalent in reactant conversion to the measured
experimental
conversion. The equilibrium composition is based upon the measured average
pressure of the inlet
and outlet of the reactor zone and the inlet molar composition. It is helpful
to obtain a curve of
equilibrium conversion versus temperature to find the apparent equilibrium
conversion
-26-
CA 02536462 2006-02-21
WO 2005/003025 PCT/US2004/020221
temperature. The equilibrium distribution (or composition) for a given
temperature, pressure, and
inlet mole fraction distribution can be calculated using Gibbs free energies
with programs such as
the NASALEWIS equilibrium code or FACTSAGE. The fractional conversion of a
reactant is the
difference between the molar flow rate into and the molar flow rate out of the
reactor zone divided
by the flow rate into the reactor zone as shown below.
Fractional conversion = (moles methane in - moles methane out) / (moles
methane in)
Moles methane in = inlet flowrate of methane at STP / (22.4 L/mol)
Moles methane out = [outlet flowrate of total product dry gas / (22.4 L/mol)]
x % methane in dry
gas GC analysis
The data plotted in Fig. I 1 shows lines corresponding to the bounds set forth
in equation
1.1 as well as data points from this work (0) and the literature (~). Some
preferred embodiments of
the invention can be defined by the OTaPproach in the expression:
1°~X~~C~. ~;~ra~u ~1.~ ~C Y c. C~°~.''~:
more preferably '~'~rt'~e
and in some embodiments by the expression:
~ . ~5° ~.'' ~l' ~ ~ ~,,~;.~ar~ '~ 1 ~ ° ~'' ~ ~~T
_ ?~ _
