Steam Reforiming Process

10.
Synthesis gas and hydrogen
This chapter covers the technologies for the

production of hydrogen. In the first part, the
individual process steps used for the conversion
of hydrocarbon feedstock to hydrogen and
carbon oxides are discussed. This part is relevant
for the production of both hydrogen and
synthesis gas. Subsequently, industrial process
concepts specifically for the production of
hydrogen are described, with emphasis on
concepts based on the steam reforming of light
hydrocarbons.
ammonia, methanol and dimethylether (DME; see

Chapter 10.4).
The technologies to be discussed are:
a) feedstock purification; b) adiabatic
prereforming; c) fired tubular and heat exchange
steam reforming; d ) adiabatic, oxidative
reforming; e) autothermal reforming; f ) other
technologies such as Catalytic Partial Oxidation
(CPO), Ceramic Membrane Reforming (CMR) and
Hydrogen Permeable Membranes (HPMs).
Feedstock purification
10.3.1 Gas conversion

technologies
Feed characteristics and purification

requirements
The two main categories of technology which can

be used to prepare synthesis gas and hydrogen are
gasification and reforming. This terminology could
be ambiguous. Gasification is the term normally
used to describe processes for the conversion of
solid or heavy liquid feedstock, while reforming is
used for the conversion of gaseous or light liquid
feedstock. Some technologies, particularly
high-temperature partial oxidation, can be used
for a whole range of feeds and the literature may
then refer to gasification to include methane
reforming. In a recent text book on the topic
of gasification (Higman and Burgt van der, 2003),
the term gasification includes what is classified as
adiabatic oxidative reforming.
In the following, the most important
technologies used for the conversion of natural gas
or other light hydrocarbon feedstock to gas rich in
hydrogen and carbon monoxide are described. The
description is general and relevant not only for
hydrogen production, but also for the production of
Fischer-Tropsch (FT) products and, for example, of
The most important impurities to be considered

in the feedstock purification unit are H2S and other
S compounds since these compounds are poisons
for downstream catalysts. In many cases, details
about the type and concentration of individual S
compounds are not available or, if they are given,
they will often be unreliable and prone to
variations over time. However, it must be expected
that organic S compounds such as COS,
mercaptans and also heavier compounds such as
sulphides, di-sulphides and thiophenes may be
present in low concentrations perhaps too low to
allow detection, but still high enough to be harmful
to catalysts.
There is no absolute value for an acceptable
concentration of S compounds in the purified gas.
However, for certain types of downstream
catalysts, very low concentrations, preferably
single-digit ppb, are necessary to ensure an
acceptable lifetime. It may be noted that it is not a
trivial task to determine S concentrations at these
levels. Sampling is difficult because most surfaces
may adsorb and later desorb S compounds
VOLUME II / REFINING AND PETROCHEMICALS
469
BULK PRODUCTS AND PRODUCTION LINES IN THE PETROCHEMICAL INDUSTRY
depending on the gas phase concentrations. Special

materials and procedures are required to obtain a
representative sample of the gas.
Principles of gas desulphurization
The typical process concept for the

desulphurization of natural gas and similar
feedstock is a two-step process based on the
hydrogenation of organic S compounds (and Cl
compounds if they are present) and the subsequent
adsorption of H2S (and HCl). A typical layout for
the desulphurization of gas with a relatively high S
concentration is shown in Fig. 1.
The feedstock is mixed with a small amount of
hydrogen or hydrogen-rich gas, preheated to
350-400C, and passed to a first reactor containing
a hydrogenation catalyst, typically based on cobalt
and molybdenum (CoMo) or nickel and
molybdenum (NiMo). After the hydrogenation
reactor, the gas passes to two S adsorbers in series,
both typically containing zinc oxide (ZnO), which
adsorbs the H2S formed in the hydrogenation
reactor. In special cases, (for instance, when the
feed contains high concentrations of CO2 and very
low concentrations of S compounds) a special
adsorption material may be required downstream
the ZnO. If organic Cl compounds are present in
the feed, HCl will be formed in the hydrogenator.
The HCl can react with ZnO and form ZnCl2.
Under normal operating conditions in the S
adsorber, ZnCl2 will sublimate and deposit on
downstream catalysts or heat transfer surfaces.
Therefore, HCl should be removed using a special
adsorbent, normally an activated alumina, placed
in a vessel between the hydrogenator and the first S
adsorber.
Variations of the process concept shown in
Fig. 1 may be used. In the simplest case, only one
vessel is used with a top layer of hydrogenation
catalyst and a bottom layer of ZnO. In other
feed
adsorber
(ZnO)
adsorber
(ZnO)
hydrogenation
reactor
H2
purified gas
Fig. 1. A typical layout for desulphurization of gas with a
relatively high S concentration.
470
cases, after the hydrogenation reactor, the gas

passes through only one S adsorber. In special
cases, a final layer of an adsorption material with
high chemisorption capacity may be installed in
the bottom of the last S adsorber or in a separate
vessel after the last S adsorber. In the layout
shown in Fig. 1, valves and piping are installed
so that the first S absorber can be isolated in
order to change the ZnO as needed, while the
plant is running. In an alternative layout with two
S adsorbers, known as swing operation, valves
and piping are installed so that the vessel with
fresh ZnO will always be the last of the two
adsorbers. This arrangement is not well-suited for
cases with stringent requirements for
purification, since some sulphur will always be
adsorbed on piping etc. upstream the adsorber,
and part of this sulphur will, after a swing, be
passed to downstream catalysts.
Steam reforming
Steam reforming of hydrocarbons is the
principal process for the manufacture of hydrogen,
especially in refineries. The typical feedstock
ranges from natural gas and Liquified Petroleum
Gas (LPG) to liquid fuels including naphtha and in
some cases, kerosene. Steam reforming is also
used, often in combinations with various oxygen or
air-blown partial oxidation processes, in the
production of synthesis gas for ammonia, methanol
and various petrochemical products. In recent
years, steam reforming has further been seen as an
option for converting the primary feed into a gas
suitable for a fuel cell. In this case, the choice of
steam reforming technology depends on the type of
fuel cell.
Steam reforming alone is not the preferred
technology for the production of synthesis gas in
large-scale applications, e.g. in the production of
synthesis gas for ammonia, methanol, or FT
synthesis. Large-scale steam reformers have poor
economy of scale compared to processes based on
partial oxidation and air separation.
Steam reforming is carried out in several
different types of reactors. Each of these is
optimized for a specific application. The main
types of reactors are: adiabatic prereformers,
tubular or primary steam reformers and various
types of heat exchange reformers.
Steam reforming reactions
Steam reforming converts hydrocarbons into

hydrogen and carbon oxides. The key reactions
involved are given in Table 1 along with the
ENCYCLOPAEDIA OF HYDROCARBONS
SYNTHESIS GAS AND HYDROGEN
enthalpy of reaction and an expression for the

equilibrium constants.
Reaction [1] in Table 1 is the steam reforming
of methane. The reaction results in gas expansion,
and high temperature and low pressure are needed
to give high equilibrium conversions. High values
of the steam to methane ratio in the feed are
required to give high conversions especially at
elevated pressure. Steam can be partially
substituted by carbon dioxide to perform CO2
reforming (reaction [2] in Table 1). This reduces
the H2/CO ratio in the product gas, which in some
cases may be more economical, especially if a
source of low-cost carbon dioxide is available.
Steam (and CO2) reforming is always accompanied
by the further conversion of CO to CO2, also called
water gas shift reaction, or more simply shift
reaction (reaction [3] in Table 1), which is
generally fast and may be considered in
equilibrium under most conditions.
The methane steam reforming reaction is
strongly endothermic as illustrated by the high
negative value of the standard enthalpy of reaction.
The heat required to convert a 1:2 mixture of
methane and steam from 600C to equilibrium at
900C is 214 kJ/mol CH4 at 30 bar. Reaction [4] in
Table 1 is the steam reforming of higher
hydrocarbons (hydrocarbons with two or more
carbon atoms). This is also a highly endothermic
reaction, which may be considered irreversible
under all conditions of practical interest.
Adiabatic prereformers
Adiabatic prereforming has, during the last

twenty years, become an integrated step in modern
reforming technology. Adiabatic gasification of
naphtha was already used during the first part of
the 1960s for the production of town gas and
substitute natural gas. However, these applications
are mainly of historical interest. Today adiabatic
prereforming is used widely in the chemical
process industry; it is well-established in the

ammonia and methanol industries, for hydrogen
production in refineries, and in the production of
synthesis gas for a variety of other chemicals
(Christensen, 1996).
The adiabatic prereformer converts higher
hydrocarbons in the feedstock into a mixture of
methane, steam, carbon oxides and hydrogen,
according to the reactions in Table 1. All higher
hydrocarbons are quantitatively converted by
reaction [4] assuming sufficient catalyst activity
(Christensen, 1996). This is accompanied by the
equilibration of the exothermic shift (reaction [3])
and methanation, the reverse of methane steam
reforming (reaction [1]).
The operating conditions in an adiabatic
prereformer depend on the type of feedstock and
the application. The typical inlet temperature is
between 350C and 550C. The selection of the
operating conditions is in many cases dictated by
the limits of carbon formation on the catalyst. For a
given feedstock and pressure, the adiabatic
prereformer must be operated within a certain
temperature window. The formation of a whisker
type of carbon will occur above the upper
temperature limit. Operation below the lower
temperature limit may result either in a polymeric
type of carbon formation (gum) or insufficient
catalyst activity. The formation of carbon on the
catalyst is further discussed below.
The overall prereforming process is often
exothermic for heavy feedstock such as naphtha.
For lighter feedstock such as LPG and natural gas,
the overall reaction may be endothermic,
thermoneutral or exothermic.
The reactor is an adiabatic vessel with
specially-designed catalysts typically based on
nickel. The low operating temperature requires a
catalyst with a high surface area in order to obtain
sufficient activity and resistance to poisoning,
especially by sulphur. The optimal shape of the
Table 1. Key reactions in steam reforming
Reaction
[1] CH4 H2O
CO 3H2
[2] CH4 CO2

2CO 2H2

[3] CO H2O CO2 H2
m
) H2
[4] CnHm n H2O
n CO (n 1
2
Standard enthalpy of reaction

, kJ/mol
DH298
Equilibrium constant
lnKpAB/T*
A
206
30.420
27,106
247
34.218
31,266
41
3.798
4,160
1,109**
* Standard state: 298 K and 1 bar; ** For n-C7H16.
471
catalyst particle depends on the specific

application and on the plant capacity. In many
cases, catalyst particles of a cylindrical shape, 3-5
mm, are used (Christensen, 1996). This particle
provides a large surface area for the access of gas
into the pore system. The pressure drop over the
prereformer is often low (0.4 bar) for small or
medium-scale plants, even with such particles
giving low void. For large-scale plants, a shapeoptimized catalyst will be an advantage and
particles in the form of rings or large cylinders
with axial holes are usually the preferred choice
for minimum pressure drop and high activity
(Christensen, 1996).
Deactivation of the prereformer catalyst may
take place during operation. The cause is typically
sulphur poisoning but sintering and other poisons
may also play a role (see below). The deactivation
of the catalyst can be observed as a progressive
movement of the temperature profile. Resistance to
deactivation is an important aspect in the design of
adiabatic prereformers.
commissioned by ICI. Less than five years later, a

Topse reformer was operating at 40 bar. At this
time, the basic types of reformers (see below) had
already been developed and industrialized
(Slack and James, 1973) and, in spite of major
improvements in performance and cost, all later
developments have not basically changed the
design.
The tubular reformer is an energy converter
(Rostrup-Nielsen, 1984a), since most of the energy
input for many processes is added via the reformer
with the hydrocarbon feed and the fuel. The energy
is transferred into hot, steam-containing synthesis
gas, and hot flue gas. The synthesis gas may
subsequently be processed further and converted
into products mainly by exothermic reactions, thus
releasing more heat. The latent heat of the process
streams and of the flue gas must be recovered to
achieve high energy efficiency and this requires
tight integration of the reformer with other parts of
the process.
Mechanical design
Fired tubular reformers
Steam reforming is, in industrial practice,

mainly carried out in reactors referred to as steam
reformers, which are essentially fired heaters with
catalyst-filled tubes placed in the radiant part of
the heater. The process may also be carried out in
reactors referred to as heat exchange reformers.
These are essentially heat exchangers with
catalyst-filled tubes. Heat exchange reformer
design is further discussed below. Abundant
literature is available on steam reforming and the
design of steam reformers (Slack and James, 1973;
Kawai et al., 1984; Rostrup-Nielsen, 1984a;
Rostrup-Nielsen et al., 1993; Dybkjr, 1995a, b;
Refining [], 1995; Rostrup-Nielsen et al., 1997,
2002; Rostrup-Nielsen and Rostrup-Nielsen, 2002).
Steam reforming is a mature process
(Rostrup-Nielsen, 1984a). The first industrial
steam reformer was installed at Baton Rouge in
Louisiana (USA) by Standard Oil of New Jersey
and commissioned in 1930. Six years later a steam
reformer was commissioned at ICI (Imperial
Chemical Industries) in Billingham (UK). During
the 1950s, metallurgical developments made it
possible to design reformers for operation at
elevated pressures. This improved the energy
efficiency of the overall process, because higher
pressure facilitates heat recovery and leads to
savings in compression energy in ammonia and
methanol plants. In 1962, two tubular reformers
operating at around 15 bar and using high
molecular weight hydrocarbon as feed were
472
The main elements in the design of a steam

reformer furnace are: a) the tube and burner
arrangement; b) inlet and outlet systems; c) tube
design; d) burner characteristics.
The tube and burner arrangement
Tubes and burners ideally should be arranged

in the furnace box in such a way that adjustment
and control of the temperature and heat flux along
the tube length can be obtained. Reformers are
designed with a variety of tube and burner
arrangements. Basically there are four types of
reformers as illustrated in Fig. 2 (Kawai et al.,
1984; Rostrup-Nielsen, 1984a; Rostrup-Nielsen et
al., 1993; Dybkjr, 1995b).
The radiant wall or side-fired reformer
contains tubes mounted in a single row along the
centreline of the furnace. In larger installations,
two such furnaces are erected side by side with
common inlet and outlet systems and with a
common fuel supply, flue gas duct and waste heat
section. Burners are mounted at several levels in
the furnace walls and the flames are directed
backwards towards the walls. The tubes are heated
by radiation from the furnace walls and the flue
gas and to a minor extent by convection. The flue
gas leaves the furnace at the top so that the flow of
process gas and flue gas is counter-current.
The top-fired reformer features a furnace box
with several rows of tubes. Burners are mounted in
the furnace ceiling between the tube rows and
between the tubes and the furnace wall. From the
Fig. 2. Tube and burner

arrangements.
radiant wall
burners, long flames are directed downwards, and

the tubes are heated by radiation from the flames
and the hot flue gas and by convection. The flue gas
leaves the furnace box at the bottom, so that the
flow of process gas and flue gas is co-current.
The bottom-fired reformer has easy access to
the burners and gives an almost constant heat flux
profile along the length of the tube. Since the tubes
are hot at the bottom, a substantial margin is
required in the tube design temperature, limiting
the outlet temperature. Today, the bottom-fired
design is considered outdated.
The terrace wall-fired reformer is a
modification of the bottom-fired type, having
slightly lower tube wall temperatures. Problems
can arise at the pinch point in the middle of the
furnace where the tubes are subject to both
radiation from the burners and to enhanced
convection from the flue gas.
Inlet and outlet systems
Inlet and outlet systems should be designed to

allow thermal expansion of the hot parts without
any risk of excessive stress.
As far as the gas inlet is concerned, the feed
gas to the reformer is normally distributed to the
individual reformer tubes from an inlet header
through so-called pigtails. The design will
depend on the type of tube support and on the
temperature of the feed gas. For reformers
designed for traditional inlet temperatures, an
inlet hairpin is connected to the top of the
reformer tube. The hairpin is designed to absorb
the thermal expansion of the catalyst tube and the
thermal expansion of the inlet header. Due to the
flexibility of the hairpin it can easily be swung
aside to give free access to the catalyst tubes for
loading or unloading the catalyst. In cases where
there is high preheating of the gas inlet
temperature, the inlet hairpin may be connected
top-fired
bottom-fired
terrace wall
to the side of the reformer tube. This design

allows a lowering of the temperature of the
flange assembly to a level where tightness can be
ensured by means of insulation inside the top of
the reformer tube. This type of reformer will
normally produce conditions for the reformer
tubes where spring support arrangements are
needed in order to avoid the tubes bending.
Where the gas outlet system is concerned, this
should in all cases be flexible enough to withstand
thermal expansion and contraction during start-up
and shutdown as well as during normal operation.
The outlet system connecting the catalyst tubes to
the manifolds is critical in most reformer designs.
A considerable number of failures in
manifolds/pigtails have been reported, and
therefore a thorough analysis of stress in these
areas is important.
In the cold collector system the reformed gas
is introduced directly from the individual tubes to
the transfer line which is located outside the
furnace. In this system, special devices are
required to connect the hot parts to the cold
collector.
In the subcollector system, groups of catalyst
tubes are connected through pigtails or hairpins to
hot subcollectors, and the subcollectors are joined
to the main collector. In relatively small reformers,
the pigtails may be connected directly to the
transfer line. The pigtails are designed to relax
thermal stress. This concept is predominant in
reformers designed for low outlet temperatures
(below about 900C).
Tube design
A steam reformer may contain up to about

1,000 high alloy steel tubes filled with catalyst.
The outer diameter is typically 100-180 mm, the
tube wall thickness is 8-20 mm, and the heated
length may be 10-14 m. Reformer tubes
473
traditionally used to be made of HK 40 (25 Cr-20

Ni) or IN-519 (24 Cr-24 Ni-1Nb; Kawai et al.,
1984; Mohri et al., 1993; Hosoya et al., 2001).
However, the demand for better creep rupture
properties led to the development of new alloys
with greater strength. Today, the preferred tube
materials are the so-called microalloys, typically
25 Cr-35 Ni-Nb-Ti (Mohri et al., 1993). The
other types of materials are only used in cases
where the strength of the microalloys cannot be
fully exploited, as there is a lower limit for the
wall thickness of the tubes.
Reformer tubes are normally designed
according to the remaining life assessment
technique, API-530 (API, 1996), for an average
lifetime, before creep rupture, of 100,000 h. The
main parameters in the design are the design
pressure, the design temperature and the creep
rupture strength of the material used. However, the
determination of these parameters can be
ambiguous, and each reforming technology
licensor applies in-house procedures to determine
the parameters and to introduce necessary design
margins.
The calculation of the design temperature is
demanding. Firstly, it requires detailed
understanding of the heat transfer in the furnace.
This includes heat transfer to the individual tube by
radiation from the furnace internals including
furnace walls and neighbouring tubes, as well as
heat transfer by convection from gas to tube wall,
by conduction through the tube wall and by
convection from the inner tube wall to the catalyst
and the reacting gas. Secondly, an understanding of
reaction kinetics, catalyst ageing, and heat and
mass transfer (radial and axial) in the catalyst bed,
etc. is required as well as knowledge about the
interplay between catalyst, reacting gas and
reformer tube; essential for the prediction of the
limits for carbon formation.
Burner characteristics
Reformer burners are a speciality product

supplied by few companies. Originally, burner
developments were mainly directed towards control
of radiation heat transfer by means of controlling
the flame shape and combustion intensity. Lately,
environmental issues (NOx control) have become a
major concern, and this has, to some extent,
changed the combustion characteristics.
Computational Fluid Dynamics (CFD) modelling
of the furnace side including correct simulation of
combustion is important in the analysis of the
effect of this development on reformer design and
performance.
474
Heat exchange reformers
Basically, a heat exchange reformer is

a steam reformer where the heat required for
the reaction is supplied predominantly by convective
heat exchange. The heat can be supplied by a flue
gas or a process gas or in principle by any other
available hot gas. The various types of heat
exchange reformers are discussed below, according
to their heat source and mechanical concept.
When the heat and mass balance on the process
(catalyst) side only is considered, there is no
difference between heat exchange reforming and
fired tubular reforming, where the heat transfer
occurs predominantly by radiation. This means that
all process schemes using heat exchange reforming
will have alternatives where the function of the
heat exchange reformer is performed in a fired
reformer. The process schemes differ only in the
amount of latent heat in the flue gas and/or process
gas and how this heat is utilized.
Models for the design and simulation of heat
exchange reformers are combinations of models
for steam reformer catalyst tubes and models for
convective heat transfer (as used in the design and
simulation of normal gas/gas heat exchangers).
Types of heat exchange reformers
Three different concepts for heat exchange

reformer design have been commercialized by
various companies. The three concepts are
illustrated in Fig. 3. Types A and B in Fig. 3 can be
used with all types of heating gas, whereas type C
can only be used when the desired product gas is a
mixture of the heating gas and the product gas
from the catalyst in the heat exchange reformer. A
comprehensive review of available types of heat
exchange reformers is given in Aasberg-Petersen et
al. (2004).
Process concepts
Heat exchange reformers heated by process gas

are, of course, always installed in combination with
another reformer, which may be a fired tubular
reformer or an air or O2-blown secondary or
autothermal reformer. Evidently, a significant
number of possible combinations exist. If there is
more than one feedstock like, for example, in Gas
To Liquid (GTL) plants, where recycled tail gas
from the synthesis may be used as additional feed
to adjust the gas composition, or in synthesis gas
or carbon monoxide plants with the importation or
recycling of carbon dioxide, the number of
possible process concepts increases further. The
use of a prereformer may also be considered, which
consequently increases the number of possible
feed
feed
product gas
heating
gas outlet
feed
mixture of
product gas
and heating
gas outlet
heating
gas outlet
product gas
heating
gas inlet
heating
gas inlet
A
heating
gas inlet
B
Fig. 3. Main types of heat exchange reformers: concept with straight-through tubes (A); concept with bayonet tubes (B);
concept with mixing of heating gas and product gas before heat exchange (C).
determined by the exit conditions from secondary

reforming.
Parallel arrangements. For obvious reasons, the
two-in, one-out concept can only be used in
parallel arrangements, i.e. process arrangements,
where the feed gas is split into two streams. One
goes directly to a conventional reformer, while the
other goes to a gas heated reformer heated by the
outlet gas from the conventional reformer or by the
mixed outlet gases from the two reformers. In
parallel arrangements, either the GHR or two-in,
two-out design or the Gas Heated Heat
Exchange Reformer (GHHER) or two-in,
one-out design can be used. In principle, with a
GHR, it is possible to produce two different
product gases, whereas the GHHER, for obvious
reasons, allows only the production of one product
gas the mixture of product gases from the two
reformers. The heat exchange reformer may be
steam
GHR
feed
SMR
process concepts. In the following, an attempt is

made to systematically describe the possible
combinations of heat exchange reformers with
either a fired tubular reformer or a secondary or
autothermal reformer. Only cases with one feed,
which most frequently is natural gas, are
considered. The cases may be divided into two
main types: series and parallel arrangements.
Series arrangements. In series arrangements,
all the process feed gas first passes through a
heat exchange reformer and then through a
second reformer, and the product gas from the
second reformer supplies heat to the heat
exchange reformer. The second reformer in the
series arrangement may be a fired tubular
reformer. This process concept is referred to as
gas heated prereforming (Fig. 4). Alternatively,
the second reformer may be an air or oxygenblown secondary reformer. This concept, which
is often referred to as Gas Heated Reforming
(GHR), is equivalent to two-step reforming and
could be called two-step reforming with GHR
(Fig. 5); in two-step reforming, a fired tubular
reformer is operated in a similar way, in series
with an air-fired secondary reformer for the
production of ammonia synthesis gas or with an
oxygen-blown secondary reformer for production
of synthesis gas for methanol or FT synthesis.
There is an alternative where only part of the
feed passes through the GHR, while the rest is
bypassed direct to the secondary reformer. This
could be called combined reforming with GHR
(Fig. 6). The operating conditions, for instance
the steam to carbon ratio (H2O/C), may in these
concepts be limited by steam reforming in the
GHR, whereas the final gas composition will be
Fig. 4. Gas heated prereforming: fired
tubular reformer (SMR, Steam Methane Reformer)

and Gas Heated Reform (GHR)
in series arrangement.
475
steam
natural gas
product gas
ATR
GHR
oxidant
Fig. 5. Two-step reforming with GHR: AutoThermal
Reformer (ATR) and Gas Heated Reformer (GHR) in series

arrangement.
coupled with a tubular reformer or with an air or

O2-blown autothermal reformer. Operating
conditions (steam to carbon ratio) in the two
reformers may be different; the final gas
composition is determined by the exit conditions
from the two reforming catalyst beds. The four
possible schemes are shown in Figs. 7-10.
Metal dusting
In all process concepts using heat exchangers

heated by process gas, the problem of avoiding
metal dusting corrosion on the heat transfer
surfaces is a challenge. Metal dusting is a
carburization phenomenon, which may occur on
metal surfaces under specific conditions, especially
at temperatures between 400C and 800C and
when in contact with a gas of high carbon activity.
natural gas
GHR
ATR
oxidant
product gas
Fig. 6. Combined reforming with GHR:

AutoThermal Reformer (ATR) and Gas Heated Reformer
(GHR) in series arrangement with partial by-pass of feed
over the GHR.
476
Metal dusting is well known in the process

industry, but unfortunately it has not yet been
completely understood. It results in loss of
material, in some cases as metal dust, a mixture of
metal, carbides and/or carbon. In severe cases the
material wastage can be very fast, leading to
catastrophic equipment failure. In recent years,
especially, significant efforts have been undertaken
to clarify the mechanism and causes of metal
dusting and to identify means of coping with the
challenge. Reviews of the status of knowledge with
comprehensive literature surveys are given in
Grabke (2000) and Jones and Baumert (2001).
On alloys protected by oxide layers, e.g.
stainless steels and Ni-based alloys, metal dusting
involves the breakdown of the protective oxide
layer, transfer of carbon into the base alloy,
formation of internal carbides, and disintegration
of the matrix. The metal particles generated by the
disintegration of the Fe or Fe-Ni-matrix act as
catalysts for carbon-formation, resulting in coke
growing from the corrosion pit (Chun et al., 1999).
The mechanism on Fe and Fe-Ni-based alloys
seems to be different, causing a different visual
appearance of the corroded metal and different
types of corrosion products. Full understanding
and agreement concerning the mechanism and the
kinetics of the attack have not been achieved (Chun
et al., 1999; Grabke, 2000). The role of gas
composition, temperature, and pressure is
understood in some detail, but knowledge which
allows a reliable prediction of the risk of metal
dusting on different alloys under different
conditions and of the progress of a possible attack
is not sufficiently detailed. Studies of metal
dusting are therefore empirical. Experimental work
is difficult because of the demanding conditions
required, and because significant time lag is
sometimes seen before an attack takes place. The
surface condition of the metal and its history may
also have an influence on its susceptibility to
attack.
Metal dusting can be prevented by adding
sulphur to the process gas (Grabke, 2000), by
choosing correct construction materials (Baker et
al., 2002), or by applying protective coatings
(Rosado and Shtze, 2003), especially Al2O3,
so-called alonizing (Bayer, 2001) or aluminizing
(Smith et al., 2001). Equipment can be designed to
avoid dangerous temperatures. Special surface
treatments, including pre-treatment with air or
steam at high temperature may be advantageous,
while other treatments such as traditional pickling
should, if possible, be avoided. But in spite of these
possibilities, metal dusting will remain a challenge
product
gas 1
process
steam
prereformer
natural gas
product
gas 1
oxidant
GHR
ATR
product
gas 2
Fig. 8. AutoThermal Reformer (ATR) and
Gas Heated Reformer (two-in, two-out, GHR)

in parallel arrangement.
in the design of processes and equipment until the

ultimate materials or protection methods are
developed and a complete understanding of this
phenomenon is acquired.
GHR
natural
gas
steam
product
gas 2
SMR
Fig. 7. Fired tubular reformer (SMR)

and Gas Heated Reformer
(two-in, two-out, GHR) in parallel
arrangement.
by sections on sulphur poisoning and carbon

formation.
Steam reforming catalysts
Most steam reforming catalysts are based on

nickel as the active material. Cobalt and noble
metals also catalyse the steam reforming reaction
but are generally too expensive for widespread use,
although both ruthenium and rhodium have higher
activity per metal area than nickel (RostrupNielsen, 1984a; Rostrup-Nielsen and Hansen,
1993). A number of different carriers including
alumina, magnesium-aluminium spinel, and
zirconia are employed.
The intrinsic activity of the catalyst is
proportional to its nickel surface area. The catalyst
consists of a huge number of small nickel particles
on the ceramic carrier. The active surface area
ANi(m2g1) may be calculated from the equation
below, when the average nickel particle diameter,
dNi() and the Ni-loading XNi (g/g) are known:
6,800 XNi (g/g)
ANi (m2g1) 1111121
dNi ()
This equation applies to spherical nickel
particles.
As mentioned above, the catalyst should be
optimized for the given application. A key criterion
for designing a catalyst for a heated reformer is to
maximize the heat transfer at low pressure drop. A
high heat transfer coefficient may reduce the tube
wall temperature and thereby minimize the
required wall thickness. High heat transfer is not
needed for a prereformer catalyst due to the
adiabatic nature of this reactor. In all cases the
catalyst should be optimized for low pressure drop
and high strength.
[1]
Steam reforming catalysis
The optimal steam reforming catalyst depends

to some extent on the application. In an adiabatic
prereformer, high activity and a high surface area
are desirable to increase resistance to deactivation
and poisoning. This activity is not as important in
primary reformers and heat exchange reformers
since the reactor design and volume is typically
determined by heat transfer and mechanical
criteria. In these cases catalyst activity is normally
much higher than needed. Instead, the catalysts are
designed for maximum heat transfer and minimum
pressure drop. However, in all cases the catalyst
should ensure equilibrium conversion throughout
its lifetime and sufficient resistance to carbon
formation and poisoning.
The typical properties and types of steam
reforming catalysts are described below, followed
Deactivation by sintering and poisoning
Catalysts in a reforming reactor may lose

activity for a number of reasons. In all cases the
477
natural
gas
steam
product
gas
natural gas
ATR
GHHER
oxidant
corresponding to a sulphur uptake of 440 mg S per

m2 of Ni surface. The maximum sulphur uptake on
a reformer catalyst is proportional to the nickel
surface area and inversely proportional to the
extent of sintering. Formation of nickel sulphide
only takes place at much higher sulphur levels than
those normally found in a reformer.
The fraction of the nickel surface area covered
with sulphur at equilibrium (the sulphur coverage, s)
depends on temperature and the PH2 S /PH2 ratio
(Rostrup-Nielsen, 1975; Rostrup-Nielsen, 1984a;
Rostrup-Nielsen et al., 1993; Christensen, 1996):
[3]
Fig. 10. AutoThermal Reformer (ATR) and Gas Heated Heat
Exchange Reformer (two-in, one-out, GHHER) in parallel

arrangement.
design of the reactor must take into account the

effect of progressive deactivation during the
catalysts lifetime.
Sintering causes the loss of surface area of the
active species of catalyst. The mechanism for
sintering is the migration and coalescence of nickel
particles on the carrier surface (Sehested, 2003).
Sintering is a complex process influenced by
several parameters including chemical
environment, catalyst structure and composition,
and support morphology. Factors that enhance
sintering include high temperature and high steam
partial pressure (Sehested, 2003).
Steam reforming catalysts are susceptible to
sulphur poisoning. Under steam reforming
conditions, all sulphur compounds are converted
into H2S, which is chemisorbed on the nickel
surface:
[2]
H2S Nisurface
S NisurfaceH2
Sulphur forms a well-defined structure with a

S:Ni stoichiometry of approximately 0.5,
478
GHHER
product
gas
process
steam
SMR
and Gas Heated Heat Exchange

Reformer (two-in, one-out, GHHER)
in parallel arrangement.
prereformer
Fig. 9. Fired tubular reformer (SMR)
s 1.459.53105 T4.17105T lnPH2 S /PH2
This expression is not valid for s close

to 0 and close to 1. In Fig. 11, the equilibrium PH2 S /PH2
ratio is plotted as a function of the sulphur coverage
for three different temperatures. The equilibrium
partial pressure of H2S is extremely low at the lower
temperature. This means that any sulphur in the feed
will be quantitatively withheld in an adiabatic
prereformer, thus protecting downstream catalysts
against sulphur poisoning.
Poisoning by sulphur takes place as shell
poisoning due to pore diffusion restrictions,
indicated in the insert in Fig. 11. The average
coverage of the particle will be much lower than in
the shell and it may take years before the
chemisorption front has moved to the centre of the
particle (Christensen, 1996).
Sulphur is a strong poison for any reforming
catalyst (Rostrup-Nielsen, 1984b). The intrinsic
activity of a catalyst decreases rapidly with sulphur
coverage as expressed below:
[4]
(1 )3
Rsp(s) Rsp
s
are the specific reaction

where Rsp (s) and Rsp
rates for the sulphur-poisoned and the sulphur-free
catalysts, respectively.
Other poisons include silica and alkali metals.
Silica may substantially reduce the activity of the
catalyst by acting as a pore mouth poison

(Christensen and Rostrup-Nielsen, 1996). The
alkali metals reduce the reaction rates in some
cases by orders of magnitude.
carbon, formed by the cracking of higher

hydrocarbons. This often occurs if the catalyst in
the top part is deactivated due to sulphur
poisoning. The temperature of the tubes at the
point of carbon formation increases because
pyrolytic carbon thermally isolates the tubes and
encapsulates the catalyst pellets resulting in no
activity and no consumption of the supplied heat.
Encapsulating carbon (gum) may be formed
when reforming heavy feeds with a high content of
aromatic compounds. The rate of gum formation is
enhanced by low temperatures and a high final
boiling point of the hydrocarbon mixture.
Encapsulating carbon is a thin CHx film, which
covers the nickel particles and leads to the
deactivation of the catalyst bed.
Carbon formation
Steam reforming involves the risk of carbon

formation. In general, three types of carbon are
concerned in steam reforming reactors like those
shown in Table 2.
Whisker carbon is the most destructive form of
carbon formed in steam reforming over nickel
catalysts. It may be formed from higher
hydrocarbons or from methane in equilibrated gas
if the overall H2O/C ratio is too low. Carbon
whiskers grow as a result of the reaction of
hydrocarbons on one side of the nickel particle and
the nucleation of carbon as a whisker on the other
side of the nickel particle as shown schematically
in Fig. 12.
The carbon whisker has higher energy than
graphite (Alstrup et al., 1981). This means that it
may be feasible, when operating in conditions at
which thermodynamics predict formation of
graphite, that there will be no carbon formation on
the catalyst. The carbon limit also depends on the
crystal size of the nickel particle. Smaller nickel
crystals are more resistant towards carbon
formation. The temperature at the onset of whisker
carbon formation is approximately 100C higher
for a catalyst with small nickel crystals
(around 7 nm) than for one with large crystals
(around 100 nm).
Pyrolytic carbon results from the exposure of
higher hydrocarbons to high temperatures. Reddish
zones known as hot bands on the walls of tubular
reformers are, in many cases, the result of pyrolytic
Fig. 11. Equilibrium
In adiabatic oxidative reforming the heat for the

reforming reactions is supplied by internal
combustion. This is in contrast to fired tubular
reforming and heat exchange reforming, where the
heat is supplied by heat exchange from an external
source. Initially it may be worth considering what
the benefits are of using adiabatic oxidative
reforming rather than steam reforming processes
such as fired tubular reforming.
In steam reforming, where the hydrocarbon
feed is made to react with steam alone, the
composition of the synthesis gas is governed by the
steam reforming reaction and the shift reaction
only. When an oxidant is used to supply heat
for the reforming reactions by internal combustion,
additional reactions are introduced. The overall
reaction is adiabatic, meaning that there is no
exchange of heat with the surroundings except a
1.0
700C
relative sulphur
content
1.0 . 101
1.0 . 102
1.0 . 103
1.0 . 104
1.0 . 105
1.0 . 106
1.0 . 107
1.0 . 108
1.0 . 109
1.0 . 1010
1.0 . 1011
1.0 . 1012
equilibrium PH2S/PH2 ratio (ppm)
PH2S /PH2 ratio as a

function of the sulphur
coverage. The inserted
figure gives the sulphur
profile in a strongly
sulphur-poisoned
catalyst.
Adiabatic oxidative reforming
600C
1.0
0.5
500C
0
0.1
0.2
0.3
0.4
0.5
0.6
2,000
distance (m)
0.7
0.8
0.9
4,000
1.0
sulphur coverage
479
Table 2. Overview of the routes to carbon formation in a reformer

Critical parameters
Characteristics
Whisker carbon
Low H2O/C
Low H2O/CnHm
High T
Dissociation of hydrocarbons at
the nickel surface and formation
of a carbon whisker at the back
of the nickel crystal
No immediate deactivation but

mechanical disintegration of the
catalyst and increased DP
Pyrolytic carbon
High T
Low activity
(sulphur
poisoning)
High partial pressure
of CnHm
Non-catalytic cracking of higher

hydrocarbons to form carbon on
catalyst and tube
Carbon formation in tube and

on catalyst.
Formation of hot bands and
increased DP
Encapsulating carbon
Low T
Heavy feed
Low H2O/C
Low H2O/CnHm
Encapsulating of the nickel

particles by a CHx film
Deactivation of the catalyst
CH4
Ni
1)
(11
ite
ph
gra
Fig. 12. Schematic illustration of the process by which
carbon whiskers are formed on the nickel particle during

steam reforming.
very limited heat loss, and the composition of the

synthesis gas can be predicted by a heat and mass
balance over the reactor. It should be noted that the
combustion reactions are all irreversible. Under the
conditions applied (high temperature and the
addition of an under-stoichiometric amount of
oxidant), all oxygen will be consumed, because this
is the limiting reactant.
Steam reforming alone will produce a
hydrogen-rich gas. When CH4 is the feed, and full
recycle of CO2 from the synthesis gas back to the
steam reformer is applied, a ratio between
hydrogen and carbon monoxide of 3.0 will be
obtained. This concept is often used in HYCO
(HYdrogen-CO) plants, where both H2 and CO are
480
Effects
products (Vannby et al., 1993). Alternatively, CO

may be converted to CO2 utilizing the water gas
shift reaction, producing a 4 to 1 mixture of H2 and
CO2. By removing CO2 and other undesired
components in, for example, a Pressure Swing
Adsorption unit (PSA unit), pure H2 may then be
produced. This is the dominating process
in hydrogen production (Dybkjr and Madsen,
1997-1998).
The synthesis gas produced by adiabatic
oxidative reforming is often characterized by the
ratio between H2 and carbon oxides or by the
stoichiometric number (SN) or module
M(H2CO2)/(COCO2). The values of module
M and the H2/CO ratios which can be obtained by
autothermal reforming of CH4, using O2 as the
oxidant, are illustrated in Fig. 13 (Christensen et al.,
1998a).
There are, of course, also other advantages (and
disadvantages) related to the use of an oxidant as
an additional reactant in reforming reactions. The
main advantages are related to economies of scale
and to size of equipment; much larger single
stream units are possible with adiabatic oxidative
reforming than with steam reforming and adiabatic
oxidative reformers are very compact units
compared to fired reformers. Furthermore,
reformer tube materials limit the outlet temperature
from fired reformers to a maximum of about
950C, while the adiabatic oxidative reforming
processes easily exceed 1,000C. This makes
higher conversion of the feed possible, even at a
low steam to carbon ratio.
The main disadvantage of adiabatic oxidative
reforming, especially with O2 as an oxidant, is that
it requires an oxygen source. Oxygen plants are

expensive, and the associated investment
constitutes a major part of the total investments in
a synthesis gas generation unit. However, the
advantages usually outweigh this disadvantage,
especially for the very large capacities required
for FT synthesis units and for modern
mega-methanol plants.
The process concepts for adiabatic oxidative
reforming may be split into 3 categories
considering the type of chemical reactions taking
place in the reactor. These can be homogeneous
reactions, heterogeneous reactions or a
combination of homogeneous and heterogeneous
reactions.
Furthermore, oxidative adiabatic reforming
processes may be characterized by the type of
feed. If the feed comes directly from a
desulphurization unit or from a prereformer, and
the reactions are carried out homogeneously
without the aid of a reforming catalyst, then the
oxidative adiabatic reformer is referred to as a
gasification or partial oxidation unit. If the
reactions are carried out heterogeneously on one
or several catalysts, they are referred to as
Catalytic Partial Oxidation (CPO). If they are
initiated by homogeneous reactions, for
example, in a burner, and completed by
heterogeneous catalysis, then the reactor is
called an AutoThermal Reformer (ATR). If the
feed has been partly reformed in a fired tubular
reformer, the ATR reactor is most often called a

secondary reformer, which is almost always
based on a combination of homogeneous
reactions (combustion) followed by catalytic
conversion.
Finally, the processes based on adiabatic
oxidative reforming may be characterized
according to the type of oxidant. Here, the
distinction is between processes based on the use
of air or enriched air in one group and the use of
oxygen in another. It is also possible to integrate
the process layouts with a heat exchange
reformer, where the heat of the gas leaving the
adiabatic oxidative reformer is used to either
reform the feed for the oxidative reformer or to
reform a separate feed stream. Such process
schemes are described in Figs. 5-10. The various
process concepts and their principal
characteristics and applications are described
below. A survey is shown in Table 3.
Processes based on homogeneous reactions
H2/CO, 950C
H2/CO, 1,050C
2.0
M
H2CO2
, 950C
COCO2
M
H2CO2
, 1,050C
COCO2
1.0
steam to carbon ratio

Fig. 13. M and H2/CO versus steam to carbon in raw gas
from ATR.
2.0
Production of synthesis gas based on

homogeneous reactions alone is referred to as
gasification or Partial OXidation (POX). The
oxidant and the hydrocarbon feedstock are
mixed in a reactor, where they are allowed to
react homogeneously at very high temperatures,
typically 1,300-1,400C. The resulting gas is
quenched or cooled by steam production, and
carbonaceous by-products like soot are removed
by washing. The formation of these by-products
has an influence on carbon efficiency.
Gasification is a versatile process which can
convert a wide range of feedstocks to synthesis
gas. The entrained flow gasification technologies,
especially, have been adapted to operate on liquid
or gaseous feed. Both Shell and Texaco have
supplied technology for natural gas conversion by
gasification for decades, and lately Lurgi has also
been promoting a Multi-Purpose Gasification
(MPG) process which is available in a version
adapted for operation on natural gas feed. The
information available in the literature about these
technologies is scarce (Philcox and Fenner, 1997;
Liebner and Ulber, 2000; Menon et al., 2001;
Beaudette et al., 2002; Hoek, 2003; Schichtling,
2003).
When operating on gas, the gasifiers will most
often be in boiler mode, where the product gas
from the gasifier is cooled by steam production in
a boiler which is specially designed to tolerate the
aggressive process conditions and the presence of
soot. The exit temperature from the gasifier is
high, typically 1,300-1,400C, to minimize soot
481
formation and to ensure close to complete

conversion of the feedstock.
Typical operating conditions for a gasification
unit are shown in Table 4. The feedstock is
assumed to be light natural gas; the product gas
is in equilibrium with respect to the steam
reforming reaction but equilibrated with respect
to the shift reaction at a lower temperature, about
1,125C; the reaction runs homogeneously
backwards at the highest temperatures in the
cooling unit. The H2/CO ratio in the product gas
is slightly below 2.0; lower values can be
achieved by recycling CO2, while steam additions
(beyond the small amounts added for burner
cooling) can not be tolerated due to increased
soot formation. In GTL applications tail gas
(CO2-rich) may be recycled to the gasifier to
improve carbon yield. If this is carried out, the
gas production in the POX unit must be
supplemented by gas production in a steam
reformer as in the Shell Middle Distillate
Synthesis process (De Graaf and Schrauwen,
2002). In the Shell Bintulu GTL plant, each

gasifier has a capacity corresponding to about
3,000 bbl/d FT products. However, Shell claims
that the gasification units may be scaled up to
about 8,000 bbl/d for new projects (Hoek, 2003).
Similar capacities are quoted for the Lurgi Multi
Purpose Gasification (MPG) process.
Processes based on heterogeneous reactions
The production of synthesis gas based only on

heterogeneous catalytic reactions is normally
referred to as CPO. The oxidant and the
hydrocarbon feedstock are premixed in a mixer
before the feed enters the catalytic bed. The
catalyst bed consisted, in older designs, of an
ignition catalyst followed by a reforming catalyst.
In some modern versions, only one catalyst is
employed and ultra short residence time is used.
The risk of auto-ignition limits the preheat
temperature, which results in a higher
consumption of oxidant. The process is reported
to be able to handle a wide range of feedstocks
Table 3. Survey of process concepts and characteristics

Secondary Secondary Autothermal Autothermal
reformer
reformer
reformer
reformer
(air)
(O2)
(O2)
(air)
Burner/mixer type
Burner
Hydrocarbon feed
Process
CH4 content (dry mole%)
Burner
gas1
10-15
PDU
(O2)
PDU
(air)
Burner
Burner
Mixer
Mixer
NG2
NG
NG/diesel
NG/diesel
10-40
85-1003
85-1003
85-1004
85-1004
1-103
1-104
1-104
Process
gas1
CnHm content (dry mole%)
None
None
1-103
Oxidant feed
Air
Oxygen/steam
Oxygen/steam
Air
Oxygen/
enriched air
Air
Process gas temperature C
770-830
750-810
350-650
350-650
200
AIT5
H2O/C ratio, mole/mole
2.5-3.6
1.5-2.5
0.4-3.5
2-4
0-2
2.06
Feed ratio gas/oxidant, mole/mole
3-4
7-12
1-3
0.5-1.0
1-4
1-5
Flame temperature (C)

Peak
Adiabatic7
2,000
1,200
2,500
1,200-1,500
2,500-3,500
1,300-2,100
2,200-2,500
1,200-1,500
Catalyst temperature (C)

Exit
970-1,020
950-1,020
850-1,100
1,000-1,100
750-1,300
750-1,200
Syngas for
FT synthesis
Fuel cells
Syngas for
FT synthesis
Typical products
Ammonia
Methanol
synthesis gas synthesis gas
Hydrogen
Hydrogen/
carbon monoxide
Syngas for
FT synthesis
Partly converted process gas from primary reformer.

Naphtha and LPG feed can also be converted directly in an autothermal reformer.
3 For natural gas. In case of naphtha and LPG feed, approx 100% C H and no methane content.
n m
4 For natural gas. In case of diesel, 100% and no methane content.
5 Preheat depends on the auto-ignition temperature of the fuel.
6 High H O/C ratios quench the reactions.
2
7 For the reaction: CH 3/2O
4
2 CO2H2OQ. Higher hydrocarbons react by means of a similar reaction scheme.
2
482
including heavy aromatic-containing

hydrocarbons like diesel. However, for GTL
applications the feed will, of course, be natural
gas. For CPO, see below.
below 1.0, become the preferred route to

synthesis gas for these syntheses, especially in
very large capacities (Olsvik and Hansen, 1998;
Haugaard and Holm-Larsen, 1999; AasbergPetersen et al., 2003; Dybkjr, 2003; Srensen
and Holm-Larsen, 2003).
Processes based on combined homogeneous

and heterogeneous reactions
Secondary reforming with air or oxygen

as the oxidant
Autothermal reforming and secondary

reforming
Processes in which the reactions are initiated

homogeneously, for example, in a burner, and
completed heterogeneously in a catalyst bed are
called AutoThermal Reforming (ATR) or secondary
reforming.
In ATR, the feed goes directly from
desulphurization or from a prereformer to the
adiabatic oxidative reformer. ATR is used mainly
with O2 as an oxidant, although it has been used
with air or enriched air for production of NH3
synthesis gas in the past. In order to make a
stoichiometric NH3 synthesis gas (H2/N23.0) by
ATR with enriched air, 30-40% vol O2 is required
in the oxidant (Dybkjr, 1995a).
As shown in Fig. 13, it is not possible to
produce synthesis gas for the production of
methanol, DME or FT synthesis directly
by autothermal reforming pure feed. It is
necessary to adjust the composition of the gas
either by removing carbon oxides or by adding
H2 (for methanol synthesis gas and similar) or
carbon-rich gas (for low temperature FT
synthesis). The carbon-rich gas may be recovered
or imported CO2, or carbon-rich tail gas from the
synthesis. However, the advantages of ATR are
such that it has, especially after the development
of processes operating at steam to carbon ratios
Table 4. Typical operating conditions for a POX unit
operating on natural gas feedstock

Oxygen/NG ratio (mol/mol)
0.55-0.65
H2O/C ratio (mol/mol)
0-0.15
Exit pressure (kg/cm2g)

Exit temperature (C)
25-40*
1,300-1,400
Exit gas characteristics

H2/CO ratio
CO2/CO ratio
CH4 content (%mol)
*Higher pressure is feasible, but probably not relevant
for GTL.
1.6-1.9
0.05-0.1
0.1
Air-blown secondary reforming is the principal

process for the manufacture of synthesis gas for
NH3 production from natural gas and naphtha
(Dybkjr, 1995b). The process concept is used by
all important suppliers of technology for NH3
production. Natural gas is desulphurized, mixed
with process steam, and passed to a fired tubular
reformer, the primary reformer. The product gas
from the primary reformer is made to react with air
in the secondary reformer to produce synthesis gas,
which is processed further by shift conversion,
removal of CO2 and methanation to give the final
synthesis gas, a 3 to 1 mixture of H2 and N2 with
small amounts of inerts, mainly CH4 and Ar. The
amount of air added to the secondary reformer is
adjusted to give the correct ratio of H2 and N2 in
the synthesis gas. Variations of the process scheme
exist (Dybkjr, 1995a).
The secondary reformer is a refractory lined
vessel with a mixer/burner (normally a
multi-nozzle design), a combustion chamber,
where homogeneous reactions take place, and a
bed of Ni-based reforming catalyst, where the
shift reaction and the reforming reaction are
equilibrated by a heterogeneous reaction on the
catalyst. When N2 is an undesired constituent in
the synthesis gas, O2 may be used as the oxidant
in the secondary reformer. This is the case in the
production of synthesis gas for methanol, DME
and high temperature FT synthesis. A process
scheme employed by Haldor Topse for the
production of methanol synthesis gas by so-called
two-step reforming uses prereforming, fired
tubular reforming, and O2-blown secondary
reforming. The design of the secondary reformer
is very similar to the design used in air-blown
processes. However, the operating conditions in
the secondary reformer are more severe than in
the air-fired concept, and a multi-nozzle burner
design can not be used. Instead, a design similar
to the design of burners for autothermal reformers
is used.
The advantage of using O2 instead of air as an
oxidant is obvious. The presence of N2 as an inert in
the final synthesis gas is avoided. The use of air as an
oxidant has been suggested in the production of
483
synthesis gas for methanol (Reducing [], 1999) and

FT synthesis (Christensen et al., 1998b; Jess et al.,
1999). However, this is not economically feasible
(Dybkjr and Christensen, 2001), since the presence
of large amounts of N2 about 50% vol in the dry
synthesis gas makes recycling concepts in the
synthesis gas impossible as they lead to low overall
efficiency. Furthermore, compression of the large
amounts of air required consumes more power than is
necessary for the production and compression of O2 in
O2-blown concepts.
Autothermal reforming (ATR)
The ATR process has been used to produce
hydrogen-rich and carbon monoxide-rich synthesis
gas for decades. In the 1950s and 1960s, autothermal
reformers were used to produce synthesis gas for
ammonia production and methanol (Hydrogen [],
1962; Topse-SBA [], 1962). In ammonia plants,
hydrogen production was maximized by operating at
high steam to carbon ratios ranging from 2.5 to 3.5
on molar basis, while in the methanol units carbon
dioxide removal adjusted the synthesis gas
composition. In the early 1990s, the technology was
improved and operation at much lower steam to
carbon ratios was achieved (Christensen and
Primdahl, 1994).
In the production of CO-rich synthesis gas for
the production of pure CO and H2, and CO mixtures
as feed for methanol or Fischer-Tropsch synthesis,
operation at low steam to carbon ratio feed ratios are
beneficial. Operation at H2O/C ratio of 0.6 has been
demonstrated in pilot scale (Christensen and
Primdahl, 1994; Christensen et al., 1998b;
Christensen et al., 2001) and industrial scale (Ernst
et al., 2000).
Chemical reactions
ATR is a combined combustion and catalytic

process carried out in an adiabatic reactor.
A mixture of natural gas and steam is partially
converted by pressurized combustion under
fuel-rich conditions and the final conversion of
hydrocarbons into an equilibrated synthesis gas is
carried out over a fixed catalyst bed. The overall
chemical reactions taking place in the ATR reactor
are described in equations [5-7], below:
484
[5]
Combustion
CH4 3/2O2 CO 2H2O

519 kJ/mol
DH298
[6]
Reforming
CH4H2O CO 3H2
206 kJ/mol
DH298
[7]
Shift
COH2O CO2H2
41 kJ/mol
DH298
The ATR reactor can be divided into three

zones: the combustion zone, the thermal zone and
the catalytic zone.
The combustion zone is a turbulent diffusion
flame, where hydrocarbon molecules and oxygen
are gradually mixed and combusted. The
combustion reactions are exothermic and very
fast and, from a global point of view, it can be
assumed that they progress as mixed is burnt.
Combustion in an ATR is sub-stoichiometric
with an overall oxygen to hydrocarbon ratio of
0.55 to 0.6, but when simplified as a one-step
model the flame zone can be described as a
single reaction [5] of CH4 to CO and H2O with
an O2/CH4 ratio of 1.5. Although simplified in
the single reaction [5] above, the combustion of
hydrocarbons consists of a large number of
homogenous radical reactions. The local
stoichiometry in the flame zone will vary from
very fuel-lean to very fuel-rich. The core of the
flame will have parts that are close to
stoichiometric, and therefore the centre of the
flame will be very hot as the adiabatic flame
temperature for the stoichiometric mixture
exceeds 3,000C.
In the thermal zone above the catalyst bed,
further conversion occurs by homogeneous
gas-phase reactions. These reactions are slower
reactions, like CO oxidation and pyrolysis
reactions involving higher hydrocarbons. The main
overall reactions in the thermal zone are the
homogeneous gas-phase steam methane reforming
reaction [6] and the shift reaction [7]. Reactions
between N2 and hydrocarbon radicals, which may
form by-products like HCN and NH3 also proceed
in the thermal zone.
In the catalytic zone the final conversion of
hydrocarbons takes place through heterogeneous
catalytic reactions including steam methane
reforming reaction [6] and shift reaction [7].
Fuel-rich combustion in partial oxidation
processes involves the risk of incomplete
combustion. Methane combustion under fuel-rich
conditions mainly proceeds through reaction steps
with C2 radicals as intermediates, which may react
to soot precursors, like Poly-Aromatic
Hydrocarbons (PAH) and soot particles (Warnatz
et al., 1996). ATR operation is soot-free in normal
circumstances, and this is achieved by means of a
carefully designed burner, by the addition of small
amounts of steam and finally by the catalytic
conversion of soot precursors; therefore the exit
gas contains no other hydrocarbon than methane.
Soot formation is unwanted; it reduces the carbon
efficiency of the process; moreover the soot
particles also need to be removed from the

synthesis gas.
ATR process and reactor design
The ATR reactor design consists of a burner, a

combustion chamber, a fixed catalyst bed section,
refractory lining, and a reactor pressure shell as
illustrated in Fig. 14. The key elements in ATR
technology are the burner and the catalyst:
The burner provides proper mixing of the feed
streams, and the fuel-rich combustion takes
place as a turbulent diffusion flame. Intensive
mixing is essential in order to avoid soot
formation.
The catalyst equilibrates the synthesis gas and
destroys soot precursors. The catalyst particle
size and shape is optimized to achieve high
activity and low pressure drop, in order to
obtain a compact reactor design.
A carefully designed process burner and
combustion chamber is required in order to avoid
excessive temperatures and the formation of soot
particles. Furthermore, the detailed design and
construction of the whole ATR reactor, including
refractory and catalyst bed, is of utmost importance
for ensuring the safe design and operation of the
synthesis gas unit. Predictions and design are
facilitated by reactor models based on the use of
Computational Fluid Dynamics (CFD) and
chemical kinetics.
combustion chamber and protect the refractory

from the hot flame core (Christensen et al., 1994).
The temperature of the gas circulating along the
walls and into the catalyst bed is reduced to the
range between 1,100-1,400C by the endothermic
reactions proceeding in the thermal zone.
Optimized recirculation will also provide a
homogeneous distribution of gas and temperature
at the entrance to the catalyst bed. Inhomogeneity
will result in larger distance to equilibrium, and the
methane concentration in the exit gas being
increased. The more even the distribution of gas to
the catalyst bed, the better is the utilization of the
catalyst activity. An even flow and temperature
distribution is obtained by careful design of the
geometry of the combustion chamber.
The flow velocities in the burner nozzles can be
selected within large ranges. Highly turbulent
mixing intensity of the diffusion flame is obtained
with high velocities in the nozzle gaps. For
applications with oxygen or enriched air as an
oxidant, the flame speed will be faster than for
similar air flames (Baukal, 1998). The position of
an oxygen flame will be very close to the burner
oxygen
natural gas
steam
ATR burner and combustion chamber
The ATR process burner is a key element of the

technology. The burner mixes the hydrocarbon feed
and the oxidant in a turbulent diffusion flame. The
flame core may exceed 3,000C. It is essential to
minimize heat transfer by thermal radiation or hot
gas recirculation to the burner parts.
In the design of the burner and the combustion
chamber the following reaction-engineering
aspects must be considered in order to ensure
optimal reactor performance, safe operation and
the satisfactory working life of the equipment:
a) effective mixing at the burner nozzles;
b) low-metal temperature of the burner;
c) soot-free combustion; d ) homogeneous gas and
temperature distribution at the catalyst bed
entrance; e) protection of the refractory from the
hot flame core.
Recirculation of the reacted gas from the
thermal zone to the burner can protect the
refractory and the burner from the hot flame core
and gases from the combustion zone. Efficient
external recirculation will enhance the position of
the flame core along the centreline of the
burner
combustion
chamber
tiles
refractory
lining
catalyst
pressure
shell
syngas
Fig. 14. Illustration of an ATR reactor.
485
486
nozzles especially at highly turbulent mixing

intensities. A turbulent diffusion flame is
in steady-state seen over a certain time period, but
an inherent feature of the turbulent flame is that the
flame is dynamic and changes position within
short time frames.
A process burner concept, the CTS
(Controlled Thermal Severity) burner
(Christensen et al., 1994), was developed in 1991
by Haldor Topse. CTS burners have been in
successful operation since 1992 in both ATR
reactors and oxygen-blown secondary reformers
for production of a variety of products ranging
from methanol, hydrogen, and ammonia to pure
CO and FT products. A new and optimized
version of the CTS burner has been developed
with a special focus on operation at low steam to
carbon ratios with high flame temperatures. The
flow patterns and mechanical design of the
burner nozzles have been optimized by the use of
tools like CFD, stress-analysis with Finite
Element Analysis (FEA) and real-environment
testing in pilot scale and in full-size
demonstrations.
Burners performing in secondary reformers
and autothermal reformers are exposed to high
operating temperatures. Industrial operation has
from time to time faced problems ranging from
burner wear without serious process implications
to severe failure. However, safe operation and
long burner lifetime have been obtained as a
result of an accurately designed burner whereby
severe consequences to equipment and production
loss have been avoided (Christensen et al., 1994;
Ernst and Berthelsen, 2004). An example of a
severe failure of an oxygen-blown secondary
reformer refractory and reactor shell following a
burner-related incident is described by Mossgas
(Shaw et al., 1994). Burners for high temperature
reformers can be designed with a focus on
mechanical and thermal integrity in combination
with the combustion chamber design (Christensen
et al., 1994). CFD can be used to predict the flow
pattern and avoid unwanted behaviour.
reactor wall is reduced to 100-200C at normal

operation.
In air-blown secondary reformers it is common
practice today to use a design with two refractory
layers. In older designs only one layer was applied,
but such a design was sensitive to cracks in the
refractory layers, which resulted in gas flow and
transfer of heat to the shell and thereby in so-called
hot spots on the pressure shell (Sterling and
Moon, 1974). In oxygen-blown secondary
reformers and ATR reactors the operating
conditions are more severe, including a higher
operating temperature in the combustion zone. In
ATR reactors, a refractory design with three layers
of refractory is commonly used today. The inner
layer has high thermal resistance and stability and
is typically a high-density alumina brick layer. The
installation of the refractory lining is important and
involves skilled craftsmen.
The transport of hot gases from the high
temperature combustion chamber and the catalyst
bed through the refractory layers to the reactor wall
does not happen with good refractory design and
installation. However, it is a potential risk and may
lead to increased temperatures at the reactor walls
which could develop into hot spots, where the
design temperature of the vessel is approached or
exceeded. The risk of gas bypass through the
refractory is especially high in the combustion
chamber where the temperature is highest, and at
the catalyst bed where a pressure drop is created,
which may force the gas through weak zones of the
refractory.
The outside of the reactor vessel is either noninsulated and exposed to the atmosphere or
covered in a water jacket. In non-insulated designs
a thermally-sensitive paint is applied to the reactor
surface. Such paint allows visual monitoring and
inspections for colour changes indicating increased
temperature in case of hot spot development. With
a water jacket it is more difficult to detect hot
spots, but this can be done to a certain degree by
monitoring the increased evaporation rate from the
water jacket.
Vessel and refractory
Catalyst bed
The reactor vessel is lined on the inside with

refractory. The refractory insulates the steel wall of
the pressure vessel from the high temperature
reaction environment and it is usually made up of
several layers of various materials with different
insulation properties. Efficient refractory design
ensures that reasonably low mechanical
temperature design temperatures can be applied for
the pressure shell. Typically, the temperature of the
The hydrocarbons are only partly converted in

the combustion chamber, and the gas leaving the
combustion chamber contains a significant amount
of methane and traces of other hydrocarbons. In
the catalytic zone the final conversion of methane
and other hydrocarbons takes place. The reactions
are heterogeneous steam reforming reactions over
the fixed catalyst bed. The methane steam
reforming reaction is endothermic, and the
temperature will decrease typically from 1,2001,300C at the inlet to the catalyst bed, to approx
1,000C at the exit of the catalyst bed. The catalyst
bed operates adiabatically with only a small heat
loss to the surroundings. The gas composition is
ideally in total equilibrium when the catalyst has
sufficient activity. However, the gas distribution to
the catalyst bed and especially the temperature and
homogeneity of the inlet gas composition will
affect the average product gas composition.
A layer of protective tiles is often placed on top
of the catalyst bed to protect it from the very
intense turbulent flow in the combustion chamber.
The radiation from the flame and the circulation
velocities in the combustion chamber require that
the tiles have high thermal stability and can resist
the thermal shocks during start-up and trips.
The requirements for the catalyst include: high
thermal stability, sufficient activity to reach
equilibrium, and low pressure drop to avoid the
bypass of gas through the refractory.
The preferred catalyst for autothermal
reforming is a nickel-based reforming catalyst. The
catalyst is exposed to high operating temperatures
and the nickel metal is subject to a high degree of
sintering, resulting in low intrinsic catalyst activity.
The support of the nickel catalysts should also
have high thermal stability to achieve sufficient
strength at high operating temperatures. In ATR
and secondary reformers carriers of both alumina
(a-Al2O3) and magnesium alumina spinel
(MgAl2O4) are used. Spinel has a higher melting
point and, in general, higher thermal strength and
stability than the alumina-based catalysts.
The shape of the catalyst pellet is an important
design parameter for the catalyst bed. The pressure
drop should be kept low in order to reduce the risk
of bypass around the catalyst bed through the
refractory. Bypass would have the consequence
that flow through the refractory layers transports
heat into these layers, which could increase
temperature on the pressure shell and result in an
increased risk of local hot spots. A shapeoptimized catalyst should be selected. For example,
seven-axial holes with a low pressure drop
(Christensen et al., 1994). In shape optimization, a
pellet shape with high catalyst pellet activity can
be selected. The optimal loading of the catalytic
fixed bed may consist of several layers of different
types of catalysts.
Industrial operation at low steam
to carbon ratio
ATR operation at a low steam to carbon ratio of

around 0.6 has become the state-of-the-art
synthesis gas technology for FT applications.

An industrial unit based on ATR technology for the
production of CO-rich synthesis gas has been in
operation in Europe since 2002. Two units with a
combined production of synthesis gas of 430,000
Nm3/h (H2+CO) were inaugurated at Sasolburg,
South Africa, in 2004. The synthesis gas is being
used for the production of chemicals and waxes by
FT synthesis. One large purpose-built GTL plant
started up in 2005 in Qatar and another will start
up in Nigeria in 2007 (Jay, 2003). Both feature two
synthesis gas units based on ATR at
steam/carbon0.6 with a combined production of
about 1.5 million Nm3/h synthesis gas.
The first operation in an industrial scale was a
demonstration test in an ATR unit operating with
approximately 20,000 Nm3/h hydrocarbon feed gas
(Ernst et al., 2000). This unit is one of 16
autothermal reformers owned and operated by
Sasol at their site in Secunda, South Africa. These
reforming units were originally revamped in 1997
to use burner technology from Haldor Topse
(Ernst and Berthelsen, 2004). One of the units was
modified specifically for the demonstration run at
a steam to carbon ratio of 0.6 instead of the normal
high ratio of 1.5-2.0. During the demonstration
run, which lasted 5 weeks, satisfactory and stable
operation at the low steam to carbon ratio was
demonstrated without soot formation. The
demonstration test pushed forward the limits for
the application of ATR technology.
Other technologies
Most, if not all of the technologies for the
production of synthesis gas described above are in
commercial use today. However, substantial efforts
to develop new technologies are being undertaken
due to the fact that the most capital-intensive part
of a GTL plant is the Syngas Generation Unit
(SGU). The focus of many of these technological
developments is to reduce or eliminate the use of
oxygen and/or reduce the size of the primary
reactor in the SGU. In the following text the
production of synthesis gas by CPO and Oxygen
Membrane Reforming (OMR) will be briefly
described.
Catalytic partial oxidation
The principle of CPO is illustrated in Fig. 15.

The hydrocarbon feed and the oxidant are mixed in
an inlet zone upstream the catalyst bed. In the
catalytic section, heterogeneous reactions in the
mixture take place (i.e. partial and total
combustion along with the methane steam
487
reforming and shift reactions). The catalyst is

normally based on noble metals and the space
velocity is in many cases very high. Catalysts in
the form of pellets, monoliths, and foams have
been used to perform the reaction.
At the reactor exit, the methane steam
reforming and shift reactions are typically at, or
close to, equilibrium. It has been claimed
(Choudhary et al., 1992) that methane reacts by
partial oxidation according to reaction [8] below
and that methane conversions above those,
corresponding to the thermodynamic equilibrium
of the methane steam reforming reaction, can be
obtained.
[8]
CH4 1/2O2 CO 2H2

However, in practice the products of reaction

[8] will further oxidize and fundamental studies
have shown that partial oxidation is only
kinetically favoured at temperatures above
900C (Schmidt, 2001). Gas compositions
indicating higher conversions than
thermodynamic equilibrium most likely reflect
the temperature of the catalyst (Rostrup-Nielsen,
1993).
CPO differs from ATR especially due to the
fact that no burner is used. Instead, all the chemical
reactions take place in the catalytic zone. Total
combustion takes place to some extent in the first
part of the catalyst layers making the catalyst very
hot in this region. Laboratory measurements have
indicated that this temperature may be higher than
1,100C (Basini et al., 2001). In order to avoid
overheating the gas upstream of the catalyst,
a thermal shield is often employed, as indicated
in Fig. 15. It should be noted that the gas
temperature remains relatively low compared to the
catalyst surface temperature in the inlet zone
(Basini et al., 2001).
CPO has been investigated extensively for
many years. Before 1992 most studies were
carried out at moderate or low space velocities at
a residence time of 1 s or above (Basini et al.,
2001). However, in later years CPO was carried
out, at least in the laboratory, at very short
contact times between 0.1 and 10 ms, in some
cases, without preheating the feedstock and with
no steam addition. Additional information
regarding research, mainly of a fundamental
nature, can be found, for example, in a series of
papers by L.D. Schmidt and co-workers
(Hickman and Schmidt, 1992; Hickman and
Schmidt, 1993; Hickman et al., 1993; Torniainen
et al., 1994; Dietz and Schmidt, 1995; Schmidt,
2001).
488
Both air and oxygen may, in principle, be used

as oxidants in a CPO reactor. Experiments with
CPO and air as an oxidant have been conducted at
the Topse pilot plant in Houston, Texas. In all
cases, methane conversion corresponds closely to
the equilibrium of the methane steam reforming
reaction.
The presence of a flammable mixture in the
inlet zone upstream the catalyst may, in some cases,
make use of CPO at high inlet temperatures
difficult, especially at elevated pressures. In Table 5
the auto-ignition temperature is given for natural
gas in air as a function of pressure.
The auto-ignition temperatures are lower with
oxygen as an oxidant. For safety reasons, the inlet
feed temperatures of the hydrocarbon feedstock
and oxidant must be kept low. This increases the
oxygen consumption as shown in Table 6. Higher
oxygen consumption increases the Air Separation
Unit (ASU) investment and the level of inerts in
the synthesis gas.
Oxygen membrane reforming
Significant efforts are being undertaken to

develop OMR, the principle of which is shown in
Fig. 16; see references (Shen et al., 2003), for
example, for a recent overview. Air is introduced at
one side of a ceramic membrane through which
oxygen in the form of ions is transported
selectively to the other side where the oxygen
reacts with the hydrocarbon feedstock to produce
synthesis gas. The concept simultaneously avoids
the capital cost of the ASU and a high content of
inert nitrogen in the synthesis gas. The catalyst on
hydrocarbon
feed
oxidant
mixing
layer
radiation
shield
noble
metal
catalyst
syngas
CO, H2, CO2, CH4, H2O
Fig. 15. Catalytic partial oxidation.
the synthesis side of the membrane may be in the

form of pellets or directly attached to the
membrane itself.
The membrane is made out of a ceramic
material often in the form of a perovskite or a
brownmillerite. The driving force across the
membrane is proportional to the logarithm of the
ratio of the partial pressures of oxygen on the two
sides. Hence, in principle, air may be introduced at
ambient pressure to supply oxygen to the other side
at elevated pressure, because the oxygen partial
pressure on the process side is extremely low. The
temperature should probably be above 750-800C
for sufficient oxygen flux.
The membrane material must enable high flux,
probably in the range of more than 10 Nm3
O2/m2h. The membrane should also withstand
reducing gas on one side and air on the other.
Various types of composite membranes have been
proposed and research is also being undertaken to
place a thin membrane on a stronger porous
support.
A key challenge in the development of OMR is
the absolute pressure difference across the
membrane. It may render the process
uneconomical if air must be compressed to ensure
similar pressures on the two sides of the
membrane. It sets great demands on the
mechanical integrity of the membrane if ambient
pressure air is used. In any case, the process may
be best suited for small or medium-scale
applications as the scaling factor of the membrane
unit will be close to unity.
Hydrogen permeable membranes
The steam reforming process, as it is carried

out today, faces a number of constraints
(Rostrup-Nielsen and Christiansen, 1997). First
of all, thermodynamics require high exit
temperatures to achieve high conversion of
methane, which is in contrast to the potential of
the catalyst, showing activity even below 400C
(Rostrup-Nielsen and Christiansen, 1997). This
Table 5. Auto-ignition temperatures for natural gas
has led to efforts to circumvent constraints by the

use of a selective hydrogen membrane installed
in the catalyst bed (Lgsgaard-Joergensen et al.,
1995; Aasberg-Petersen et al., 1998). Hydrogen
is continuously extracted from the reaction,
thereby pushing the equilibrium to higher
conversion at a lower temperature. Reactor
simulations and experiments (Aasberg-Petersen
et al., 1998) have shown that the reformer exit
temperature can indeed be reduced to below
700C while maintaining the same conversion.
The economy of this scheme depends on
membrane cost versus savings by elimination of
the CO shift reaction and the PSA unit
(Lgsgaard-Joergensen et al., 1995). However,
the produced hydrogen with this concept is at low
pressure and must be compressed to the usual
delivery pressure of 20 bar. This renders the
process uneconomical except, possibly, in
specific scenarios with very low electricity prices
(Aasberg-Petersen et al., 1998), or in cases where
hydrogen is used as a feedstock for a fuel cell or
as low pressure fuel.
This scheme for hydrogen fuel production with
CO2 sequestration is much simpler than a scheme
based on conventional reforming followed by shift,
CO2 separation and CO2 compression. If CO2
sequestration becomes accepted the membrane
reforming scheme shown in Fig. 17 may become
the preferred process for the production of
hydrogen fuel.
State-of-the-art commercial membranes
(Pd-type) are still much too expensive to make the
membrane reformer scheme attractive, although
progress is being made to prepare membranes with
palladium films of a thickness approaching one
micron. Furthermore, the current achievable flux
with commercially available membranes is much
too low. Moreover, if significant amounts of
hydrogen were to be produced by membrane
reforming, the world supply of palladium would
soon be exhausted. Hence, there is a need for new
membrane types.
10.3.2 Hydrogen technology
in air (Geerssen, 1988)

Pressure (bar)
Auto-ignition
temperature (C)
465
313
20
267
40
259
There is a growing need for hydrogen and a future

hydrogen economy is high on the political agenda.
But where should the hydrogen come from?
The sustainable routes are still too expensive.
Steam reforming of hydrocarbons is the most
feasible route today. If CO2 sequestration is
accepted, fossil fuels may play an important role at
the start of a hydrogen economy in the future. This
489
Table 6. Relative oxygen and natural gas consumption for CPO or ATR-based GTL front ends for the production
of hydrogen and carbon monoxide (Aasberg-Petersen et al., 2003)

Steam/Carbon
ratio
Hydrocarbon feed
temperature,
reactor inlet (C)
Oxygen consumption
(relative)
Natural gas
consumption
(relative)
ATR
0.6
650
100
100
CPO
0.6
200
121
109
ATR
0.3
650
97
102
CPO
0.3
200
114
109
Reactor
CH4, H2O
membrane
air
2O2
2e
Po2
CO, H2, CO2

CH4 1/2O2
CH4 H2O
CO H2O
N2 (O2)
CO 2H2
3H2 CO
CO2 H2
Fig. 16. Principle of oxygen membrane reforming.
will happen through the use of the reforming

technologies (Rostrup-Nielsen et al., 2002).
Hydrogen manufacturing
Hydrogen is an important raw material for the
chemical and the refinery industry, and it may play
a future role in the energy sector. In 1998, the total
hydrogen market was 390109 Nm3/yr with the
addition of 110109 Nm3/yr as coproduction. The
present use of manufactured hydrogen is primarily
for the production of ammonia and methanol
(about 51% in mixtures with nitrogen or carbon
oxides) followed by hydrotreating in refineries
(44% including coproduction, pure hydrogen).
Pure hydrogen is also used for a number of
hydrogenation reactions (4% of total consumption)
such as hydrogenation of unsaturated hydrocarbons
(including hardening of edible oil) and of
490
aromatics, hydrogenation of aldehydes and ketones

(e.g. oxo-products), hydrogenation of nitrogen
compounds (e.g. for the manufacture of aniline).
Other present uses (1%) of hydrogen are related to
the food industry, the semiconductor industry and
the metallurgical industry (e.g. direct reduction of
iron ore).
Mixtures of hydrogen and nitrogen are used in
ammonia synthesis, and mixtures of hydrogen and
carbon oxides (synthesis gas) for the synthesis of
methanol, liquid hydrocarbons (for instance by FT
synthesis), and the synthesis of higher alcohols
(hydroformylation), etc. This section deals only
with the technology for the manufacture of pure
hydrogen.
The refinery hydrogen balance problem
The environmental objectives for providing

better transportation fuels may lead to significant
changes in the refinery industry (Rostrup-Nielsen,
1995). Specifications for reformulated gasoline
have meant fewer aromatics and olefins
and constraints on light hydrocarbons and sulphur.
New legislation for diesel requires deep
desulphurization to 10-50 ppm S. This is done by
reacting the sulphur compounds with hydrogen
into hydrogen sulphide, which is removed from the
hydrocarbon stream. The requirement for removing
sulphur may be accompanied by a wish to remove
aromatics.
In general, these trends result in an increasing
atomic ratio H/C of the fuels approaching 2 (De
Jong, 1996), while available oil resources become
heavier with higher contents of sulphur and metals.
This has created a great necessity for more
hydrotreating hydrodesulphurization (HDS),
hydrodenitrogenation (HDN),
hydrodemetallization (HDM) and hydrocracking.
Traditionally, most of the hydrogen
consumption in refineries was covered by
hydrogen produced as a by-product from other
CO2 free gas

steam
air
combustion of H2
feed
HDS
heat transport from flue gas

to steam reforming
steam reforming
H2
H2
hydrogen permeable
membrane
H2 fuel
CO2 for
sequestration
compression
hydrogen
product
Fig. 17. Membrane reforming. Hydrogen plant with CO2 sequestration.
refinery processes (110109 Nm3/y), mainly

catalytic reforming (platforming). A main reaction
in catalytic reforming (not to be confused with
catalytic steam reforming) is the conversion of
paraffins into aromatics and hydrogen. As
aromatics are not wanted in reformulated fuels,
this means that less hydrogen will become
available from catalytic reforming. Similarly, the
gasoline and diesel fractions from catalytic
crackers are highly unsaturated.
In conclusion, there is a fast-growing need for
increased hydrogen production capacity in
refineries. This need is being met mainly by the
installation of steam reforming-based hydrogen
plants.
Hydrogen as an energy vector
Due to its environmentally benign properties,

hydrogen has been discussed as a future energy
vector. Key applications for hydrogen are as a
carbon-free fuel and as a fuel for hydrogen-driven
fuel cells for automotive applications.
Many technologies for the production of
hydrogen that do not coproduce CO2 are being
considered. Hydrogen production using non-fossil
energy for the electrolysis of water is one example.
These schemes have not been introduced, primarily
due to reluctance concerning nuclear power and the
low efficiency of the electrolysis process.
Hydrogen from bio-fuels, wind energy or solar
energy, is still expensive leaving fossil fuels as the
most feasible feedstock for hydrogen generation in
the near term.
Significant efforts are being made to develop
technologies for hydrogen production based on
fossil fuels that allow storing the CO2 (CO2
sequestration) while the hydrogen can be used as a
carbon-free fuel, for instance, in gas turbines for
power production. The envisioned schemes for
CO2 sequestration include precombustion
decarbonization and post combustion CO2 capture.

A group of oil companies has joined forces in a
CO2 Capture Project (CCP) (Middleton et al.,
2002). The aim of the CCP is to develop effective
methods to capture significant amounts of CO2
emitted from power generation and industrial
sources and store the CO2 in geological formations
below the Earths surface.
Hydrogen is being used in fuel cells typically in
units of capacity 50 kW to 1 MW. The application
of fuel cells has not grown as fast as predicted
because of the high investment costs and
competition from advanced gas turbines. Lately,
hydrogen-driven fuel cells (30-50 kW) have
attracted great interest for mobile applications. The
issue is then where to produce the hydrogen: in
large centralized plants, at the gas stations or in the
car. Difficulties in storing sufficient hydrogen
mean that commercial vehicles will probably
evolve hydrogen generation on board the vehicle
from hydrocarbons or methanol
(Rostrup-Nielsen, 2001).
Hydrogen production technologies
There are several production routes for
hydrogen (Table 7). The choice depends on the
size of production and the cost of available
feedstocks. The most important method is
catalytic conversion (steam reforming) of
hydrocarbons followed by gasification of coal,
tar sands etc. For small-scale production,
investment is predominant and simple equipment
may be preferred over high-energy efficiency.
Electrolysis of water accounts for less than 5%.
For large-scale production, steam reforming of
natural gas (or refinery off-gases) has become
the preferred solution. Gasification of heavy oil
fractions may play an increasing role, as these
fractions are becoming more available because of
491
falling demands. Some refineries have installed

gasification units for power production and
cogeneration of hydrogen (Pitt, 2001).
In areas with a high cost of hydrocarbon
feedstocks, methanol may be considered as an
alternative. One possible scheme involves the
production of methanol in an area with very
inexpensive natural gas, with subsequent
transportation of the methanol to the hydrogen
plant location. A methanol-based hydrogen plant is
a simple unit (Rostrup-Nielsen, 2001) and less
costly than a natural gas and naphtha-based plant
with a steam reformer.
Hydrogen by steam reforming
The principal process for converting

hydrocarbons into hydrogen is steam reforming
(Rostrup-Nielsen, 1984a; Aasberg-Petersen et al.,
2001), which involves the following reactions:
[9]
[10]
[11]
CH4 H2O
CO 3H2
206 kJ/mol
DH298
CO H2O
CO2 H2
41 kJ/mol
DH298
m
) H2
CnHm n H2O
n CO (n 1
2
1,109
DH298
kJ/mol for n-C7H16
Reaction [9] is the steam reforming of

methane. It is reversible and strongly
endothermic, and according to the principle of Le
Chtelier it must be carried out at a high
temperature, a high steam to methane ratio and
low pressure to achieve maximum conversion.
The design of the steam reforming process is, in
part, dictated by these constraints. The
equilibrium composition of gas out of the steam
reformer is shown in Fig. 18 as a function of
steam reformer outlet temperature under typical
industrial conditions (26 bar with a feed steam to
methane ratio of 2.5).
Process layouts
Modern hydrogen plants will almost invariably

be designed using a low steam to carbon ratio. A
high steam to carbon ratio (4-5 mol H2O/C-atom)
results in higher conversion of the hydrocarbons,
but a low steam to carbon ratio (typically 2.5 or
lower) reduces the mass flow through the plant and
thus the size of the equipment. The lowest
investment is therefore generally obtained for
plants designed for a low steam to carbon ratio,
which also results in a more energy-efficient plant
and thus in lower operating costs. In principle, a
low steam to carbon ratio increases the methane
leakage from the reformer, but this can be
492
Table 7. Hydrogen production routes

Feedstock
Process
Natural gas
Refinery off-gases
LPG
Naphtha
Kerosene, gas oil
Steam reforming
Methanol, DME, NH3
Cracking
Coal
Biomass
Gasification
Water
Electrolysis
compensated for by increasing the reformer outlet

temperature to typically 920C.
Operation at a low steam to carbon ratio
requires the use of a non-iron containing carbon
monoxide conversion catalyst, i.e. a copper-based
Medium Temperature Shift (MTS) catalyst. This
catalyst is very similar to the conventional low
temperature shift catalyst, which is optionally used
in combination with an iron-based High
Temperature Shift (HTS) catalyst. The iron-based
catalyst can only be used above a certain steam to
dry gas ratio since it will, under more severe
conditions, develop activity for FT synthesis
because of its potential to form iron carbide
(Hjlund-Nielsen and Hansen, 1989). For further
description of shift catalysts see Chapter 10.4.
Today, PSA is normally used for final hydrogen
purification. This layout gives a high purity
hydrogen product (99.9% or higher) and more
efficient operation than traditional layouts with
CO2-adsorption (Dybkjr and Madsen,
1997-1998).
A typical process layout of a feedstock-flexible
hydrogen plant operating at 25 bar on refinery gas,
natural gas and naphtha is given in Fig. 19.
Refinery gas, containing large amounts of
hydrogen, is sent to a PSA unit where pure
hydrogen is extracted. The off-gas from the PSA,
containing non-converted methane is compressed
and used as feed in the hydrogen plant. In this way,
low-grade refinery gas is used as feed to a
hydrogen plant, thereby substituting more
expensive natural gas or naphtha. PSA off-gas is
mixed with natural gas or vapourized naphtha, and
the gas mixture is preheated, desulphurized (over
CoMo-catalyst and ZnO), mixed with process
steam and further heated before entering the
adiabatic prereformer. Typical inlet temperatures
are in the range 450-550C, depending on the

feedstock and the steam to carbon ratio. The
prereformed gas is then heated to 650C before
entering the tubular reformer where final
conversion to equilibrium of methane into
hydrogen, carbon monoxide and carbon dioxide
takes place at 850-950C depending on the layout.
The reformed gas is cooled by producing steam
before entering the shift converter, typically
containing an MTS (210-330C). Over the
copper-based shift catalyst, more hydrogen is
produced by converting carbon monoxide and
steam to carbon dioxide and hydrogen [10].
The shifted gas is cooled to ambient temperature
before entering the second PSA unit, the off-gas
from which is used as fuel in the tubular reformer
supplemented with fuel gas.
Energy efficiency of the process
As energy prices increase, the operating cost

becomes an increasing part of the total hydrogen
production cost. In todays hydrogen plants the
operating cost (hydrocarbon cost steam credit)
amounts to about 2/3 of the hydrogen production
cost and it is therefore important to have a very
energy-efficient process. In recent years, the
hydrogen process has been improved significantly.
It is important to understand the thermodynamic
limit for energy efficiency that one is up against
(Rostrup-Nielsen, 2005).
Consider an ideal process in which a
stoichiometric mixture of CH4 and H2O is
converted completely to H2 and CO2 by [9] and
[10] and the heat of reaction is supplied by the
combustion of CH4. For this ideal process it is
possible to determine the theoretical CH4 (Lower
Heating Value, LHV) energy required to produce
hydrogen. The theoretical efficiency of hydrogen
dry gas composition (%)
100
CO
80
CH4
CO2
60
40
H2
20
0
400
500
600
700
800
900
reformer outlet temperature (C)

Fig. 18. Equilibrium composition of gas out
of a steam reformer at 26 bar with a feed steam

to methane ratio of 2.5.
1,000
production from methane by steam reforming is

2.59 Gcal/1,000 Nm3 H2 when starting from water
vapour, and 2.81 Gcal/1,000 Nm3 H2 when starting
from liquid water as in the real process.
Hydrogen plants designed with old technology
utilize reforming temperatures well below 900C
and high steam to carbon ratios (above 2.5). These
plants are characterized by poor energy efficiency,
as significant amounts of process steam have to be
condensed by large air coolers and water coolers.
Also investment costs are high as large volumetric
process flows have to be handled.
Modern hydrogen plants utilize the new
developments in steam reforming and shift
technology, thus allowing them to be designed with
reforming temperatures above 900C and steam to
carbon ratios below 2.5 or even below 2.0. These
conditions are easily handled by the side-fired
reformer and using an MTS catalyst. For many
years CO and oxo-gas plants have operated using
side-fired steam reformers with outlet
temperatures of about 950C, and many hydrogen
plants have operated with reformer outlet
temperatures of about 920C. The reasons for
using these advanced steam reforming conditions
are improved energy efficiency and reduced
hydrogen production cost (Dybkjr and Madsen,
1997-1998).
Advanced steam reforming conditions (high
reforming temperature combined with low steam to
carbon ratio) have made it possible to operate
hydrogen plants with an energy efficiency of less
than 3 Gcal/1,000 Nm3 H2, allowing for export
steam credit (feedfuel amounts to 3.5
Gcal/1,000 Nm3 H2), which is only 6% more than
the theoretically possible 2.81 Gcal/1,000 Nm3 H2.
The only losses are heat losses from the equipment
and heat losses with the flue gas. This is illustrated
in Fig. 20.
Radiant steam reformers utilize about 50-60%
of the fired heat for the steam reforming reaction,
the rest of the energy is used for process heating
and steam production, making these plants steam
exporting.
Since steam export is not always desirable, it
has become a challenge to design a reformer and
an energy-efficient process which does not
export steam. The Haldor Topse Convective
Reformer (HTCR) developed by Topse utilizes
about 80% of the fired heat for the steam
reforming process and the product gas is at a
reduced temperature due to a bayonet design of
the reformer tube. With the convective steam
reformer it is possible to reach energy
efficiencies of about 3.15 to 3.25 Gcal/1,000
493
steam export
tubular
reformer
MT shift
prereformer
H2
process
steam
refinery gas
S-removal
PSA
PSA 2
H2
NG feed
naphtha
flue gas
combustion air
DMW
fuel gas
Fig. 19. Process layout of a typical multi-feedstock H2-plant (Haldor Topse).
DMW: DeMineralized Water.
Nm3 H2 without steam export (Dybkjr and

Madsen, 2004). This is significantly less than the
3.5 Gcal/1,000 Nm3 H2 obtained with a radiant
steam reformer when steam credit is not taken
into account. The lower overall natural gas
consumption obtained with the convective
reformer results in hydrogen production with
significantly reduced CO2 emissions.
The industry has been active in the
development of steam reforming technology,
processes and catalysts, and significant effort is
continually being spent on these developments to
ensure that a range of options are available to the
industry so that the optimum process concept can
be identified for each project and customized for
each location (Goland and Dybkjr, 1995). One of
Topses latest developments in tubular steam
reforming process technology to meet this goal is
the Topse Advanced Steam Reforming process
(Dybkjr and Madsen, 1997-1998).
(optional); f ) high combustion air preheat

(optional; up to 550C has been industrially
proven).
These process parameters are possible with the
Topse side-fired reformer design and with the
catalysts developed by Topse.
The low steam to carbon ratio reduces the mass
flow through the plant, and thus the size of the
equipment. A major source of increased operating
costs and energy inefficiency is the condensation
of unconverted process steam after the shift
converter. By operating at low steam to carbon
ratio this loss is minimized. The increase in
methane leakage from the reformer, typically
associated with operation at low steam to carbon
ratio, is compensated for by operating at a high
reformer outlet temperature. Furthermore, the low
steam/carbon ratio will result in excessive
by-product formation if an iron-based
shift catalyst is used, such as the typical HTS.
This problem is avoided by switching to a special
copper-based shift catalyst, such as the MTS.
Advanced Steam Reforming process
Final plant layout optimization
This concept offers a low investment cost plant

with low operating cost. However, final layout
selection takes place after further analysis of
specific unit costs. The Topse Advanced Steam
Reforming process for hydrogen production
is characterized by: a) low steam to carbon ratio
(typically below 2.5); b) high reformer outlet
temperature (typically 920-930C); c) high average
heat flux reformer (typically above 80,000
kcal/m2h); d) non-iron containing shift (e.g.
copper-based MTS); e) adiabatic prereforming
There is great flexibility in designing a modern

hydrogen plant to obtain the overall best operating
economics. Depending on the customers
requirements and on the cost of feed and fuel and
the credit for steam export, the final plant design
can be selected. The key design parameters are:
a) steam to carbon ratio; b) use of a prereformer;
c) reformer inlet temperature; d) reformer outlet
temperature; e) layout of the shift section;
f ) combustion air preheat; g) pressure around the
reformer; h) efficiency of the PSA unit.
Process options
494
temperature
net energy consumption 2.96 Gcal/1,000 Nm3 H2
cooling curve
heating curve
enthalpy
Fig. 20. Pinch diagram for a hydrogen
plant using advanced steam reforming

conditions. The heating and cooling curves
are ideally matched, meaning that all heat
contained in cooled process streams is used
to heat other process streams and no
air or water cooling is necessary.
The following illustrates how these parameters can

be used to custom-design a hydrogen plant. For more
details, see references (Goland and Dybkjr, 1995).
Example where feed is more expensive than
fuel. For a plant that has to operate with expensive
feed and cheap fuel, it is necessary to minimize the
feed consumption at the expense of fuel
consumption. This is achieved by selecting options
that utilize the feed in the best possible way, i.e.
with a high steam/carbon ratio, a high reformer
outlet temperature, a two-stage shift system, and
high-efficiency PSA.
Example with minimum and maximum steam
export. Energy-efficient hydrogen plant designs
can be optimized for a minimum or maximum
steam export, by a scrupulous selection of the
design parameters. Of course, it is possible to
achieve any steam export quantity in between these
extremes. Table 8 highlights the key differences
between plants designed for minimum and
maximum steam export respectively.
Table 8. Key design differences for minimum
and maximum steam export

Steam export
Plant design
Minimum
Maximum
Zero steam export. It is possible to reduce the

steam production from a hydrogen plant based on
tubular steam reforming (Goland and Dybkjr,
1995). Introduction of a prereformer with reheat
(Fig. 21) increases the thermal efficiency for
reforming from 50% to about 60%. Another part of
the flue gas heat content can be used for preheating
combustion air. However, it is not possible to
completely eliminate the steam export. This can be
done by using a convective heat exchange reformer
in which the flue gas as well as the hot product gas
are heat exchanged with the process gas, so that
they leave the reformer at about 600C (Dybkjr
and Madsen, 1997-1998). The amount of waste
heat is reduced from 50% in the conventional
design to about 20% of the fired duty in the heat
exchange reformer. This means that the steam
generated from the remaining waste heat just
matches the steam needed for the process in such a
way that export of steam can be eliminated.
Convective reformers are industrially proven and
are preferred for smaller units due to their
compactness.
Feedstock flexibility
Traditionally, feedstocks for hydrogen plants

have been naphtha, natural gas and LPG. However,
today hydrogen plants are often designed to be able
to utilize a number of different refinery off-gas
streams as feedstock.
This includes hydrogen-rich streams as well as
streams with high concentrations of olefins.
This increased demand on plant flexibility
requires the development of process designs and
catalysts that will ensure the reliable operation of
the hydrogen plant.
Multiple feedstocks
Designing a hydrogen plant so that it is able to

operate on multiple feedstocks creates flexibility in
using the most economical feedstock at any given
time.
A key in the design of this type of plant is the
prereformer. Topse has designed several plants
operating on a number of feedstocks as indicated
in Table 9.
The Indian ammonia plant of the TaTa group
operates with an automatic Mixed Feed Control
Strategy (MFCS) system, which allows automatic
switch from one feed to the other.
Combustion air preheat
Yes
No
Prereformer
Yes
Optional
Olefin-containing feedstock
Low
When designing hydrogen plants to operate on

olefin-containing feedstock, it is necessary to pay
special attention to hydrogenation. Olefins are
Reformer
inlet temperature
High
495
hydrogenated by hydrogen over the hydrogenation

catalyst used in the desulphurization section.
The hydrogenation of olefins is an exothermic
reaction and the resulting temperature increase
over the hydrogenation reactor is dependent on
the concentration of olefins. When the associated
temperature increase exceeds about 100C it is
necessary to design the hydrogenation section to
limit the increase. One effective way of limiting
the temperature increase is to introduce a recycle
loop with cooling around the hydrogenation
reactor.
Topses hydrogenation catalyst, for example, is
successfully used in three Italian hydrogen plants
operating on refinery off-gas containing from 0 to
25% olefins. The hydrogenation section can handle
rapid changes in olefin concentration by
controlling the recirculation flow by frequent
analysis of the feed gas composition.
the purpose being to bridge the gap between

existing cars and fuel cell cars with on-board
hydrogen storage. The DOE stopped the
programme in the summer of 2004, the reasons it
announced being that technical targets had not
been met, competing hybrid technology
had bridged the gap, and interest from
automakers in the technology had diminished
(Ho and Lightner, 2004).
Centralized production (large capacity)
Hydrogen can be produced in large centralized

plants, using todays technology as described
above at a reasonable cost. However, there is a
high cost associated with the distribution of
hydrogen, for instance, to fuelling stations,
increasing the cost of hydrogen by about 3 times,
whereas gasoline cost has only increased to about
1.4 times production cost (European Commission,
2003). This is especially true during the early
stages of the introduction of a hydrogen economy,
where hydrogen consumption is low and
dispersed. There are also many issues concerning
the safety of a hydrogen distribution system.
Therefore, it is reasonable to assume that in the
early stages of a transition to a hydrogen economy,
hydrogen would be produced locally at the fuelling
station using the natural gas distribution system
that exists in many places.
Challenges for the future

The emerging hydrogen economy wrestles
with a number of issues (National Research
Council, 2004). For example, whether hydrogen
production should be centralized or localized, or
whether it will be necessary to consider CO2
sequestration, what the best energy carriers are
from which to produce hydrogen, and other
questions. However, it is evident that hydrogen
production by steam reforming of hydrocarbons
will play an important role for many years to
come in a transition to renewable energy sources,
although the steam reforming process will be
faced with many new challenges depending on
the distribution scheme chosen, which today is
unknown. For example, in the US, the effort
supported by the Department
Of Energy (DOE) to develop an on-board
fuel-processing base on steam reforming for
automotive applications can be acknowledged;
tubular reformer
CO2 (optional)
feed from HDS

process steam
496
Decentralized hydrogen production for

localized power production and hydrogen fuelling
stations depends upon small scale plants with
many new requirements regarding size,
by-products, start-up time, production cost, etc.
In areas where natural gas is not available,
hydrogen could be produced from hydrogen
carriers such as methanol, via electricity (e.g. from
renewable energy such as wind or solar) or from
biofuels, etc.
prereformer
Fig. 21. A, installation of a

prereformer with reheat in a
hydrogen plant. B,
prereformer shown in front
of a tubular reformer in a
70,000 Nm3/h hydrogen plant
at SK Corporation, Korea.
Localized production (small capacity)
waste
heat
channel
Table 9. Examples of multiple feedstock plants designed by Topse

Plant location
Korea
USA
India
Plant type
Hydrogen
Hydrogen
Ammonia
Feedstocks
Light and naphtha
Naphtha
Naphtha
Off-gases
Off-gases
Off-gases as fuel
Off-gas from PSA
Natural gas
Natural gas
LPG
LPG
These requirements present a challenge and

necessitate the development of new types of
steam reformers. There are many options,
ranging from designs inspired by conventional
reformers to designs based on plate heat
exchanges utilizing catalyzed surfaces, etc. Many
groups are working on the development of
different designs of small-scale steam reformers
(Ogden, 1999).
CO2 sequestration
The global environmental debate may lead to

the necessity for hydrogen produced from fossil
fuels, which will have to include CO2
sequestration.
As already mentioned, CCP has the aim of
developing technology to produce hydrogen fuel
and capture the CO2 for storage in geological
formations. A study was carried out by Topse for
CCP (Middleton et al., 2002), analysing the
application of a natural gas-fed membrane steam
reformer and coal gasifier synthesis gas-fed
membrane shift reactor as a cost-effective means of
producing hydrogen with CO2 sequestration.
Although these membrane reactors have a potential
for meeting the requirements, further development
of membrane technology would be required,
thereby lowering the cost. With this type of
technology, even coal could be used to generate
hydrogen with CO2 sequestration.
Topse conducted a study with Norsk Hydro
ASA, determining the state-of-the-art
technology for the production of 1,200 MW
power from natural gas with CO2 sequestration
(Solgaard-Andersen et al., 2002). The study
focussed on three options: precombustion
decarbonization, combustion with pure oxygen,
and post-combustion decarbonization. The study
found that the most realistic scheme was
precombustion decarbonization. Here, a
hydrogen/nitrogen fuel mix is produced by
autothermal reforming followed by
CO2-removal, which is inspired by Topses
ammonia plant front end. Norsk Hydro ASA

concluded that the technology was feasible
without extensive further development.
Finding a solution for delivering hydrogen
economically in a hydrogen economy is
challenging, especially if one has to consider CO2
sequestration. It is a dilemma between centralized
and decentralized hydrogen production.
Centralized large-scale hydrogen production fits
into a hydrogen economy in combination with a
hydrogen distribution system.
The centralized production offers reasonable
options for CO2 sequestration, for example, by
membrane reforming. However, it suffers
from high distribution costs. The alternative is
localized hydrogen production, which fits better
with fuel distribution systems, but CO2
sequestration is not easy.
References
Aasberg-Petersen K. et al. (1998) Membrane reforming for
hydrogen, CatalysisToday, 46, 193-201.
Aasberg-Petersen K. et al. (2001) Technologies for largescale gas conversion, Applied Catalysis A. General, 221,
379-387.
Aasberg-Petersen K. et al. (2003) Recent developments in
autothermal reforming and pre-reforming for synthesis gas
production in GTL applications, Fuel Processing
Technology, 83, 253-261.
Aasberg-Petersen K. et al. (2004) Synthesis gas production
for FT synthesis, in: Studies in surface science and catalysis,
152, Amsterdam, Elsevier, 258-405.
Alstrup I. et al. (1981) High temperature hydrogen sulphide
chemisorption on nickel catalyst, Applied Catalysis, 1,
303-314.
API (American Petroleum Institute) (1996) Recommended
practice 530, Washington (D.C.), API.
Baker B.A. et al. (2002) Selection of nickel-base alloys for
metal dusting resistance, in: Ammonia plant safety & related
facilities, 42, New York, American Institute of Chemical
Engineers, 257.
Basini G. et al. (2001) Catalytic partial oxidation of natural
gas at elevated pressure and low residence time, Catalysis
Today, 64, 9-20.
497
Baukal C.E. Jr. (edited by) (1998) Oxygen-enhanced

combustion, Boca Raton (FL), CRC.
Bayer G.T. (2001) Surface engineered coatings for metal
dusting, in: Corrosion 2001. Proceedings of the National
Association of Corrosion Engineers international annual
conference, Paper 01387, NACE.
Beaudette T.M. et al. (2002) Project success, Hydrocarbon
Engineering, 7, 41.
Choudhary V.R. et al. (1992) Low temperature oxidative
conversion of methane to syngas over NiO-CaO catalyst,
Catalysis Letters, 15, 363-370.
Christensen T.S. (1996) Adiabatic prereforming of
hydrocarbons. An important step in syngas production,
Applied Catalysis A. General, 138, 285-309.
Christensen T.S., Primdahl I.I. (1994) Improving syngas
production using autothermal reforming, Hydrocarbon
Processing, 73, 39.
Christensen T.S., Rostrup-Nielsen J.R. (1996) Catalyst
deactivation in adiabatic prereforming. Experimental
methods and models for prediction of performance, in:
Deactivation and testing of hydrocarbon processing
catalysts. Proceedings of the American Chemical Society
210th national meeting, Chicago (IL), August 1995.
Christensen T.S. et al. (1994) Burners for secondary and
autothermal reforming design and industrial performance,
in: Ammonia plant safety & related facilities, 35, New York,
American Institute of Chemical Engineers, 205.
Christensen T.S. et al. (1998a) Development in autothermal
reforming, in: Studies in surface science and catalysis,
119, Amsterdam, Elsevier, 883.
Christensen T.S. et al. (1998b) Synthesis gas preparation by
autothermal reforming for conversion of natural Gas to
Liquid Products (GTL), in: Proceedings of the Monetizing
stranded gas reserves annual conference, San Francisco
(CA), 14-16 December.
Christensen T.S. et al. (2001) Process demonstration of
autothermal reforming at low steam-to-carbon ratios for
production of synthesis gas, in: Proceedings of the American
Institute of Chemical Engineers annual meeting, Reno
(NE), 4-9 November.
Chun C.H. et al. (1999) Relationship between coking and metal
dusting, Materials and Corrosion-Werkstoffe und
Korrosion, 50, 634-639.
De Graaf W., Schrauwen F. (2002) World scale GTL,
Hydrocarbon Engineering, 7, 55-58.
De Jong K.P. (1996) Efficient catalytic processes for the
manufacturing of high-quality transportation fuels,
CatalysisToday, 29, 171-178.
Dietz A.G., Schmidt L.D. (1995) Effect of pressure on three
catalytic partial oxidation reactions at millisecond contact
times, Catalysis Letters, 33, 15-29.
Dybkjr I. (1995a) Ammonia production processes, in: Nielsen
A., Aika K. (editors) Ammonia. Catalysis and manufacture,
Berlin, Springer, 199.
Dybkjr I. (1995b) Tubular reforming and autothermal
reforming of natural gas. An overview of available processes,
Fuel Processing Technology, 42, 85-107.
Dybkjr I. (2003) Synthesis gas technology, in:
Fundamentals of gas to liquids, London, Petroleum
Economist, 16.
Dybkjr I., Christensen T.S. (2001) Syngas for large scale
conversion of natural gas to liquid fuels, in: Studies
498
in surface science and catalysis, 136, Amsterdam, Elsevier,

435.
Dybkjr I., Madsen S.W. (1997-1998) Advanced reforming
technologies for hydrogen production, International Journal
of Hydrocarbon Engineering, 3, 56-65.
Dybkjr I., Madsen S.W. (2004) Compact hydrogen plants,
International Journal of Hydrocarbon Engineering,
November.
Ernst W.S., Berthelsen A.C. (2004) Operational reformer
improvement. Experience at Sasol Synfuels 1997-2003, in:
Ammonia plant safety & related facilities, 43, New York,
American Institute of Chemical Engineers, 172.
Ernst W.S. et al. (2000) Push syngas production limits,
Hydrocarbon Processing, 79, 100/C-100/J.
European Commission (2003) Market development of
alternative fuels. Report of the alternative fuels contact
group, December.
Geerssen T.M. (1988) Physical properties of natural gas,
Groningen, N.V. Nederlandse Gasunie.
Goland J.N., Dybkjr I. (1995) Options for hydrogen
production, Hydrogen Technology International Quarterly,
Summer, 27.
Grabke H.J. (2000) Corrosion by carbonaceous gases,
carburization and metal dusting, and methods of prevention,
Materials at High Temperatures, 17, 483-487.
Haugaard J., Holm-Larsen H. (1999) Recent advances in
autothermal technology. Reducing production cost to
prosper in a depressed market, in: Proceedings of the World
Methanol Conference, San Diego (CA), 29 November-1
December.
Hickman D.A., Schmidt L.D. (1992) Synthesis gas formation
by direct oxidation of methane over Pt monoliths, Journal
of Catalysis, 138, 267-282.
Hickman D.A., Schmidt L.D. (1993) Syngas formation
by direct catalytic oxidation of methane, Science, 259,
343-349.
Hickman D.A. et al. (1993) Synthesis gas formation by direct
oxidation of methane over Rh monoliths, Catalysis Letters,
17, 223-237.
Higman C., Burgt M. van der (2003) Gasification,
Amsterdam, Elsevier; Burlington (MA), Gulf.
Ho D., Lightner V. (2004) Future directions for DOE fuel
processing R&D. Results of the on-board fuel processing
go/no-go decision for DOEs office on hydrogen, fuel
cells & infrastructure technologies, in: Fuel cell:
technology, markets and commercialization. Proceedings
of the fuel cells seminar, San Antonio (TX), 1-5
November, 37.
Hoek A. (2003) The Shell middle distillate synthesis process,
in: Gas to liquids. When the rubber meets the road.
Proceedings of the CatCon, Houston (TX), 5-6 May.
Hjlund-Nielsen P.E., Hansen B.J. (1989) Conversion
limitations in hydrocarbon synthesis, Journal of Molecular
Catalysis, 17, 183-193.
Hosoya K. et al. (2001) Better furnace performance with new
alloy radiant coil, Petroleum Technology Quarterly,
Autumn, 115.
Hydrogen process broadens feedstock range (1962), Chemical
Engineering, 69, 88.
Jay M. (2003) Progress at Escravos GTL, in: Fundamentals
of gas to liquids, London, Petroleum Economist, 45.
Jess A. et al. (1999) Fischer-Tropsch. Synthesis with nitrogenrich syngas. Fundamentals and reactor design aspects,
Applied Catalysis A. General, 186, 321-342.
Jones R.T., Baumert K.L. (2001) Metal dusting. An overview
of current literature, in: Corrosion 2001. Proceedings
of the National Association of Corrosion Engineers
international annual conference, Paper 01372, NACE.
Kawai T. et al. (1984) Design features, material selection and
performance of steam reforming furnace, in: Proceedings
of the International Plant Engineering Conference, Bombay,
14-16 November.
Lgsgaard-Joergensen S. et al. (1995) Steam reforming of
methane in a hydrogen permeable membrane reactor,
Catalysis Today, 25, 303-307.
Liebner W., Ulber D. (2000) MPG-Lurgi multi purpose
gasification. Application in gas-gasification, in:
Proceedings of the Gasification Technologies Conference,
San Francisco (CA), 8-11 October.
Menon R. et al. (2001) Industrial gas partnership enables
upgrade at BP refinery, Oil & Gas Journal, 99, 70.
Middleton P. et al. (2002) Hydrogen production with CO2
capture using membrane reactors, in: The hydrogen planet.
Proceedings of the 14th World hydrogen energy conference,
Montreal (Canada), 9-13 June, 45.
Mohri T. et al. (1993) Application of advanced materials for
catalyst tubes for steam reformers, in: Ammonia plant safety
& related facilities, 33, New York, American Institute of
Chemical Engineers, 86.
National Research Council (2004) The hydrogen economy.
Opportunities, costs, barriers, and R&D needs, Washington
(D.C.), The National Academies Press.
Ogden J.M. (1999) Hydrogen energy systems studies, in:
Proceedings of the 1999 US DOE Hydrogen Program
Review, Lakewood (CO), 4-6 May.
Olsvik O., Hansen R. (1998) High pressure autothermal reforming
(HP ATR), in: Proceedings of the 5th Natural gas conversion
symposium, Taormina (Italy), 20-25 September, 875.
Philcox J.E., Fenner G.W. (1997) A Texas project illustrates
the benefits of integrated gasification, Oil & Gas Journal,
95, 41-46.
Pitt R. (2001) Gasification provides power and hydrogen,
World Refining, January, 6.
Reducing methanol production costs. The Starchem process
(1999), Hydrocarbon Asia, 9, 56.
Refining reforming technology (1995), Nitrogen, 214, 38.
Rosado C., Shtze M. (2003) Protective behaviour of newly
developed coatings against metal dusting, Materials and
Corrosion, 54, 831-853.
Rostrup-Nielsen J.R. (1975) Steam reforming catalysts. An
investigation of catalysts for tubular steam reforming of
hydrocarbons. A contribution from the Research Laboratory
of Haldor Topse A/S, Copenhagen, Teknisk Forlag.
Rostrup-Nielsen J.R. (1984a) Catalytic steam reforming, in:
Anderson J.R., Boudart M. (edited by), Catalysis, science
and technology, Berlin, Springer, 1981-1996, 11v.; v. VI.
Rostrup-Nielsen J.R. (1984b) Sulfur-passivated nickel
catalysts for carbon-free steam reforming of methane,
Journal of Catalysis, 85, 31-43.
Rostrup-Nielsen J.R. (1993) Production of synthesis gas,

Rostrup-Nielsen J.R. (1995) Innovation and the catalytic
process industry. The science and the challenge, Chemical
Engineering Science, 50, 4061-4071.
Rostrup-Nielsen J.R. (2001) Conversion of hydrocarbons
and alcohols for fuel cells, Physical Chemistry Chemical
Physics, 3, 283-288.
Rostrup-Nielsen J.R., Christiansen L.J. (1997) Tubular
steam reforming, in: Tominaga H., Tamaki M. (edited by),
Chemical reaction and reactor design, Chichester, John
Wiley, Chapter 5.2.
Rostrup-Nielsen J.R., Hansen J.H.B. (1993) CO2-Reforming
of methane over transition metals, Journal of Catalysis,
144, 38-49.
Rostrup-Nielsen J.R., Rostrup-Nielsen T. (2002) Large
scale hydrogen production, Cattech, 6, 150-159.
Rostrup-Nielsen J.R. et al. (1993) Steam reforming
opportunities and limits of the technology, in: de Lasa
H.J. et al. (editors), Chemical reactor technology for
environmentally safe reactors and products, Dordrecht,
Kluwer, 249.
Rostrup-Nielsen J.R. et al. (1997) Reforming of hydrocarbons
into synthesis gas on supported metal catalysts, Journal
of the Japan Petroleum Institute, 40, 366.
Rostrup-Nielsen J.R. et al. (2002) Hydrogen and synthesis
gas by steam reforming, Advances in Catalysis, 47, 65139.
Rostrup-Nielsen T. (2005) Manufacture of hydrogen,
Schichtling H. (2003) Update on Lurgi syngas technologies,
in: Proceedings of the Gasification technologies conference,
San Francisco (CA), 12-15 October.
Schmidt L.D. (2001) Modeling millisecond reactors, in: Natural
Gas Conversion VI. Proceedings of the 6th Natural Gas
Conversion symposium, Girwood (AK), 17-22 June.
Sehested J. (2003) Sintering of nickel steam-reforming
catalysts, Journal of Catalysis, 217, 417-426.
Shaw G. et al. (1994) Commissioning of the worlds largest
oxygen blown reformers, in: Ammonia plant safety & related
facilities, New York, American Institute of Chemical
Engineers, 35, 315.
Shen J. et al. (2003) Applications of ceramic-membrane
technology, in: Fundamentals of gas to liquids, London,
Petroleum Economist, 24.
Slack A.V., James G.R. (edited by) (1973) Ammonia, New
York, Marcel Decker, 1973-1979, 4v.; v.I, 145.
Smith A.B. et al. (2001) The use of aluminide diffusion coatings
to improve carburization resistance, in: Corrosion 2001.
National Association of Corrosion Engineers international
annual conference, Paper 01391, NACE.
Solgaard-Andersen H. et al. (2002) O2-Production and
abatement in power generation: Integrated reforming
combined cycle IRCC, in: Future energy systems and
technology for CO2 abatement. Proceedings of the
international symposium, Antwerpen (Belgium), 17-20
November.
Srensen E.L., Holm-Larsen H. (2003) Flexibility in design
of large-scale methanol plants, in: Proceedings of the annual
World Methanol Conference, Phoenix (AR), 10-12
November.
499
Sterling M.B., Moon A.J. (1974) Failure in secondary

reformer vessel, in: Ammonia plant safety & related
facilities, 17, New York, American Institute of Chemical
Engineers, 135.
Topse-SBA autothermal process (1962), The Journal of World
Nitrogen, May.
Torniainen P.M. et al. (1994) Comparison of monolithsupported metals for the direct oxidation of methane to
syngas, Journal of Catalysis, 146, 1-10.
Vannby R. et al. (1993) Steam reforming for carbon monoxide
500
production, Hydrocarbon Technology International.

Warnatz J. et al. (1996) Combustion. Physical and chemical
fundamentals, modelling and simulation experiments,
pollutant formation, Berlin, Springer.
Ib Dybkjr
Thomas Rostrup-Nielsen
Kim Aasberg-Petersen
Haldor Topse
Lyngby, Denmark

Steam Reforiming Process

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Steam Reforiming Process

Diunggah oleh

Hak Cipta:

Format Tersedia

10.

Synthesis gas and hydrogen

This chapter covers the technologies for the

ammonia, methanol and dimethylether (DME; see

10.3.1 Gas conversion

Feed characteristics and purification

The two main categories of technology which can

The most important impurities to be considered

VOLUME II / REFINING AND PETROCHEMICALS

BULK PRODUCTS AND PRODUCTION LINES IN THE PETROCHEMICAL INDUSTRY

depending on the gas phase concentrations. Special

The typical process concept for the

cases, after the hydrogenation reactor, the gas

Steam reforming converts hydrocarbons into

SYNTHESIS GAS AND HYDROGEN

enthalpy of reaction and an expression for the

Adiabatic prereforming has, during the last

process industry; it is well-established in the

Table 1. Key reactions in steam reforming

[2] CH4 CO2

Standard enthalpy of reaction

* Standard state: 298 K and 1 bar; ** For n-C7H16.

VOLUME II / REFINING AND PETROCHEMICALS

BULK PRODUCTS AND PRODUCTION LINES IN THE PETROCHEMICAL INDUSTRY

catalyst particle depends on the specific

commissioned by ICI. Less than five years later, a

Fired tubular reformers

Steam reforming is, in industrial practice,

The main elements in the design of a steam

Tubes and burners ideally should be arranged

SYNTHESIS GAS AND HYDROGEN

Fig. 2. Tube and burner

burners, long flames are directed downwards, and

Inlet and outlet systems should be designed to

VOLUME II / REFINING AND PETROCHEMICALS

to the side of the reformer tube. This design

A steam reformer may contain up to about

BULK PRODUCTS AND PRODUCTION LINES IN THE PETROCHEMICAL INDUSTRY

traditionally used to be made of HK 40 (25 Cr-20

Reformer burners are a speciality product

Heat exchange reformers

Basically, a heat exchange reformer is

Three different concepts for heat exchange

Heat exchange reformers heated by process gas

SYNTHESIS GAS AND HYDROGEN

VOLUME II / REFINING AND PETROCHEMICALS

determined by the exit conditions from secondary

process concepts. In the following, an attempt is

Fig. 4. Gas heated prereforming: fired

tubular reformer (SMR, Steam Methane Reformer)

BULK PRODUCTS AND PRODUCTION LINES IN THE PETROCHEMICAL INDUSTRY

Fig. 5. Two-step reforming with GHR: AutoThermal

Reformer (ATR) and Gas Heated Reformer (GHR) in series

coupled with a tubular reformer or with an air or

In all process concepts using heat exchangers

Fig. 6. Combined reforming with GHR:

Metal dusting is well known in the process

SYNTHESIS GAS AND HYDROGEN

Fig. 8. AutoThermal Reformer (ATR) and

Gas Heated Reformer (two-in, two-out, GHR)

in the design of processes and equipment until the

Fig. 7. Fired tubular reformer (SMR)

by sections on sulphur poisoning and carbon

Most steam reforming catalysts are based on

s 1.459.53105 T4.17105T lnPH2 S /PH2