Lloyd L. Handbook of Industrial Catalysts

Подождите немного. Документ загружается.

368 Chapter 9

The presence of excess steam in the process gas to the reformer results in

the formation of carbon dioxide by the water gas shift reaction. Thus the gas

leaving the steam reformer also contains between 7 and 15% carbon dioxide:

CO + H

O ↔ CO

+ H

(9.9)

Typical operating conditions for a modern steam reformer are given in Ta-

ble 9.9 although these can be slightly different depending on the overall plant

design. The table also includes typical conditions for the steam reformers in

methanol and hydrogen plants which use the same catalysts. These examples

illustrate the wide variations in gas composition which can be achieved by

changing the operating steam ratio, temperature and pressure, to increase me-

thane conversion.

The long catalyst tubes are suspended vertically in rows within the furnace

(Figure

9.3). The hydrocarbon/steam mixture, preheated to about 500°C, passes

into the reformer through a header pipe connected to the tops of each tube by

pigtails or flexible joints. Following reaction, the products leave the reformer

via a manifold pipe to the secondary reformer. Heat can be supplied to the fur-

nace from burners at the top, sides or bottom of the insulated box. The intensity

of the heat, or heat flux, in the furnace can be varied depending on the demands

of the reaction but is generally highest at about one third from the top of the

tube, where most of the endothermic reaction takes place.

Furnace design is complex and is fully described in specialized publica-

tions.

Catalyst operation at extremely high temperatures is critical in the pro-

duction of ammonia and full instructions on handling and operation are issued

by all catalyst suppliers.

TABLE 9.9. Steam Reforming Operating Conditions with Natural Gas.

Process Ammonia

1000 t/day

Methanol

1000 t/day

Hydrogen

30 mm cfd

Catalyst volume (m

) 17–20 20–25 20

Steam/carbon 3–5 5 6

Inlet pressure (atm) 35 20 20

Temperature inlet (°C)

525 515 520

Temperature outlet (°C)

790 850 780

Outlet gas composition (vol%):

Nitrogen <2 Trace Trace

Hydrogen 70 73 75

Carbon monoxide 8 14

Carbon dioxide 12 8 12

Methane 10 4–5 2–3

Addition of CO

required before reforming or synthesis.

CO removed by catalysts up to methanation or pressure swing absorption after HTS converter.

9.4.2

Early

magne

ound

Cimen

strengt

mixtur

The co

I. G.

Oil in

reform

ressu

the ref

zation

Figure

top-fir

gas pr

ed., by

eformin

eforming cata

ia and kaoli

and olefins i

fondu was t

and giving

of precipita

mposition of

arben catalyst

1932 was pr

ng catalysts

e drop. Durin

rming proces

nd zinc oxid

9.3. Arrangeme

d furnace for t

duction. Repri

kind permissio

tal

sts

ysts were ma

. These were

the feed gas

en found to b

etter resistanc

ed nickel oxi

number of ea

to be used i

bably produc

ere found b

g the 1930s,

. This led to

absorption pr

t of burner an

e primary ref

ted from Catal

of M. Twigg.

e by mixing

active but e

obtained fro

a useful add

to sulfur poi

e with magn

ly catalysts is

the large-sca

d as three-qu

ICI to give

CI also bega

he modern co

cedures.

Synt

tubes in a typi

rmer in synthe

st Handboo

recipitated n

sily

oisone

coal hydroge

itive by incre

oning. The b

sia, kaolin a

shown in Ta

e plants oper

arter inch cu

better perfor

to desulfuriz

alt molybdat

hesis Gas

sis

ckel oxides

by sulfur co

ation process

sing the catal

st catalyst wa

d ciment fon

le 9.10. The f

ted by Stand

es. Raschig r

ance and lo

the feed gas

hydrodesulf

ith

es.

s a

rst

370 Chapter 9

TABLE 9.10. Examples of Early Reforming Catalyst Composition.

First Catalysts Raschig Ring

Catalyst

Wt% A B C D

Nickel 20 10–16 20 15

Magnesia 20 0–5 15 12

Kaolin 60 – 35 40

Alumina – 10–16 – –

Ciment Fondu – 68–75 30 35

Early catalysts were slowly developed to improve resistance to carbon formation.

A Operated well with pure methane but when poisoned by sulfur and unsaturated hydrocarbons led

to carbon deposition.

B Ciment fondu acted as a filler and partly inhibited carbon formation.

C Gave activity as high as A and less carbon than B but still sensitive to unsaturated hydrocarbons.

D Could be used with most natural gases and off gas, providing that poisons were removed. Some

early charges operated for 10–20 years with occasional steaming to clean catalyst.

The early catalyst formulations, with only minor changes, were still being

used until about 1965 when methane was first reformed at pressures above about

15atm. Silica is volatile in steam at the reforming temperature in high-pressure

plants and it is then deposited at a lower temperature in the waste heat boilers, or

on to the high-temperature shift catalyst. This led to the development of silica-

free reforming catalysts made from alumina, rather than kaolin or magnesia, and

alumina cement, rather than ciment fondu. Although alumina based catalysts

were active, it was found that they lost strength at high temperature. This led to

some breakage and caused hot or patchy tubes. Attempts were made to improve

strength by incorporating titanium dioxide into the catalyst recipe rather than

alumina although these were not really successful. The best solution was found

to be the use of silica free preformed rings, impregnated with nickel nitrate,

which could be decomposed to give nickel oxide. Problems in the early days in

the selection of a support that was stable in the severe operating conditions led

to the introduction of several different catalysts. The supports ranged from pure

alpha alumina to calcium aluminate and magnesium aluminate. Operational

problems were experienced in the early days, arising from poor thermal stability

and loss of activity when using α–alumina supports. Carbon was deposited about

one third of the way down from the top of the tube and hot bands formed in

high-heat-flux converters.

A partial solution to the low activity problem was to use shorter rings to

improve heat transfer although this did, of course, this did increase the pressure

drop through the tubes. Hot bands could also be avoided by using alkalized cata-

lysts in the top part of tubes to control carbon deposition. It was clear that im-

proved reforming catalysts were required, not only with a higher, stable activity

but also, as it turned out, with a different shape.

By the time most ammonia plants were being operated beyond the limits of

their design capacity, new catalyst shapes were eventually developed. The origi-

Synthesis Gas

371

nal intention was to increase the geometric surface area of the ring and thus to

increase activity. A further benefit was, however, found to be a reduction of the

pressure drop through the tubes and improved heat transfer from the tube wall to

the reacting gases. Once again a variety of different shapes were introduced

ranging from rings, spoked wagon wheels and cylinders with either four or sev-

en holes. Part of the range of catalyst compositions and shapes available is

shown in Table 9.11.

9.4.3. Reformer Operation

The catalyst should be loaded carefully into the reformer tubes to avoid

breaking the rings. The normal procedure is to fill socks or soft tubes with a

bottom flap which can be lowered into each of the tubes before use. The sock is

then withdrawn from the tube and the flap opens to leave the catalyst in place.

To ensure even packing, the same weight of catalyst is placed in each tube and

the pressure drop measured. Gentle tapping, at the bottom of the tubes, may be

needed to pack the catalyst properly. In common with all other tubular reactors

charging catalyst into the tubes of a steam reformer is a laborious process.

The catalyst must be activated by reduction with a hydrogen-rich gas added

to the steam passing through the tubes. A steam to hydrogen ratio of about six is

used during reduction at operating temperature. The procedure can be completed

in about six to eight hours. Feed may then be gradually introduced and operation

begins. Catalyst breakage can result from shortage of steam or thermal shock

during operation. This leads to an uneven gas flow with hot tubes and possible

TABLE 9.11. Reforming Catalysts: Composition and Shapes.

Calcium

aluminate

-aluminate

Magnesium

aluminate

Composition (wt%):

NiO 15–20 15–20 15–20

65–70 80–85 60–65

CaO 10–15 – –

MgO – – 15–20

Bulk density (kg liter

-1

) ~1.0 ~0.8–0.9 ~1.0

Surface area (m

-1

) 10–20 2–5 10–20

Shapes Size: diameter x length x wall

Raschig Rings mm 16x16x6

16x16x6

Rings 16mm 7 or 9 ribbed rings

Rings 16mm 4 to 9 holes

372 Chapter 9

carbon deposition. In badly affected tubes the catalyst must be replaced. The

leakage of sulfur or chlorine from the purification stage also results in poisoning

and deactivation of the catalyst. Any decrease in activity leads to a high methane

content in the reformed gas from the reformer tubes which results in hot spots in

the tube and the deposition of more carbon. A mild occurrence of sulfur or chlo-

rine poisoning can generally be reversed by increasing the steam ratio for a short

period. Arsenic is a permanent poison to the catalyst. If the arsenic concentration

in the catalyst at the top of the tubes exceeds about 150 ppm, then the catalyst

must be changed, and the tubes cleaned very carefully. The arsenic usually

found its way into the reformers in the early days either as a component of the

feedstock, or through some maloperation of the Vetrocoke system, used for the

absorption of carbon dioxide. A typical early warning of catalyst poisoning is a

gradual increase in the methane content of gas leaving the reformer and, in seri-

ous cases, an increase in pressure drop.

The introduction of the first 1000 tons per day high-pressure steam reform-

ers in single stream plants led to some operating difficulties. For example, the

new high heat flux reformers developed overheated bands, about one third of the

way down the reformer tubes. Hot bands, as they were called, resulted from the

deposition of carbon from the thermal cracking reaction as the catalyst lost ac-

tivity. Carbon can most easily form from methane cracking:

↔ C + 2 H

(9.10)

This normally takes place when flow of steam stops for any reason but is also

thermodynamically possible during full-scale operation of a reformer before the

gas temperature reaches about 650°C.

Fortunately, two other reactions lead to the removal of carbon under normal

reforming conditions:

C + CO

↔ 2 CO Boudouard reaction (9.11)

C + H

O ↔ CO + H

Water-gas reaction (9.12)

Thus, any carbon formation arising from the cracking reaction at the tube inlet is

controlled to some extent by the two carbon removal reactions, which tend to be

rather faster than the cracking reaction. Unfortunately, the cracking reaction

becomes the faster reaction at temperatures above about 650ºC and this tempera-

ture corresponds to the position in the reactor tube where the hot bands tend to

form. The rate of carbon deposition is greater than its rate of removal, unless the

activity of the catalyst is sufficient to provide sufficient hydrogen by the reform-

ing reaction to prevent the deposition of carbon in the first place.

It was clear that a more active and thermally stable catalyst was needed. For

some time the situation was controlled by the use of shorter rings at the top of

Synthesis Gas

373

the tube. This provided more geometric surface area and increased reforming

activity. Even more successful operation was possible if the shorter rings were

alkalized. The addition of potash to catalysts under development for the reform-

ing of naphtha was introduced by ICI during the 1960s and was a major factor in

the control of the formation of hot bands. The potash had the effect of accelerat-

ing the steam carbon reaction and inhibiting cracking and polymerization on the

catalyst surface.

During the 1970s and 1980s the cost of oil had increased substantially and

production of ammonia in the older plants was becoming uneconomical. This

led to the revamping of many units and the use of more energy efficient process-

es. A steam reformer uses about one third of the total energy required by a large

ammonia plant. It was therefore suggested that the steam ratio could be de-

creased, from 3.5–4.0 to less than 3.0, even though this could mean that more

carbon would deposit on the catalyst and that higher tube wall temperature

would lead to shorter tube life. A penalty of operating a reformer with hot tubes,

for any reason, is that a temperature only 10°C above the design level can re-

duce the tube life by up to 50%. The cost of tube failures resulting from the use

of low activity catalysts or decreasing the steam ratio was very high.

The new catalyst shapes that were under development gave some compen-

sation for these disadvantages. Higher activity, particularly at lower temperature,

could help to avoid carbon formation. Carefully optimized shapes also gave a

lower pressure drop and provided better heat transfer within the tubes to de-

crease the temperature of the tube wall. The use of manaurite tubes also helped

to improve operation.

Some operators decided to use a small pre-reformer able to ease the load on

the primary reformer and to give further energy savings. The small adiabatic

reactor was loaded with a stable high activity catalyst operating at 530°C. The

alkalized nickel alumina catalyst, introduced by British Gas in their CRG Pro-

cess during the 1960s, can operate continuously for about two years. All traces

of high hydrocarbons were removed in the prereformer and methane was partial-

ly reformed to give about 20% hydrogen in the feed gas entering the primary

reformer. This allowed the minimum practicable steam ratio to be used.

An overall fuel saving of 7–10% was possible, and the tube wall tempera-

ture of the reformer tubes can be reduced by about 20°C by decreasing the firing

rate and heat flux in the reformer. This leads to a longer tube life. Alternatively,

up to 5–10% extra ammonia could be produced by increasing the feed rate.

When mixed feeds were used, as in some hydrogen plants, steadier operation

was possible. Brief details of the CRG pre-reforming process operation are giv-

en in Table 9.12.

374 Chapter 9

TABLE 9.12. Prereforming of Natural Gas or Refinery Feeds

Process Composition (dry outlet gas)

Feed Natural Gas Butane Naphtha

Steam ratio 0.3 1 1.5

22–24 30–34 20–22

66–68 50–55 60

+ <10 ppm <10 ppm –

Balance CO/CO

CO/CO

Temperature inlet (°C)

530 530 450–530

ΔT (°C)

–64 +11 +40

Pressure Reformer design

9.4.4. Secondary Reforming

When coke was the basic raw material for the production of ammonia synthesis

gas, the hydrogen and nitrogen required was supplied by mixing water gas and

producer gas. Several small ammonia plants were, in complete contrast, built in

the United States during the 1920s to use electrolytic hydrogen.

The operators

found it relatively easy to introduce the necessary nitrogen by burning some of

the hydrogen in air. When the war time methane steam reformers were built by

the US Government, the same procedure was modified by burning primary re-

former outlet gas in a high temperature adiabatic reactor containing more prima-

ry reforming catalyst. The additional reactor became known as the secondary

reformer. Eventually a catalyst with lower nickel content, and an even higher

thermal stability, was developed to withstand the very high temperature reached

at the top of the bed.

In modern ammonia plants, air containing the design volume of nitrogen

needed for ammonia synthesis is added to the secondary reformer. Two im-

portant reactions can take place over the heat-resistant nickel catalyst. Two im-

portant reactions take place over the heat resistant nickel catalyst in the second-

ary reformer. Firstly, the oxygen component of the air is consumed by combus-

tion with hydrogen and possibly carbon monoxide. Secondly, some of the resid-

ual methane from the primary reformer undergoes further reforming. Both reac-

tions reach equilibrium at a bed outlet temperature of about 1000°C. The me-

thane content of gas leaving the secondary reformer is usually in the range 0.2–

0.5% while oxygen is completely removed.

The main problems in the operation of a secondary reformer are associated

with the high temperatures generated by the combustion reactions. The gas dis-

tributor must be well designed so that the process gas and air are mixed as rapid-

ly and thoroughly as possible. The catalyst bed itself is protected from the very

high temperatures generated by the homogeneous combustion of air, by a layer

of refractory material that is placed on top of the large, temperature-resistant

catalyst particles. The rest of the bed is filled with secondary reforming catalyst.

Synthesis Gas

375

TABLE 9.13. Secondary Reforming Catalyst Operation.

Gas Composition Secondary Reformer

Inlet Outlet

(dry vol%) 9.4 0.2

(vol%) 11.6 8.8

CO (vol%) 8.3 11.5

(vol%) 0.5 22.1

(vol%) 70.2 57.1

Air (vol% of dry reformed gas) ~38

Pressure (atm) 30 29

Temperature (°C)

790 971

Catalyst volume (m

)(1000 tonnes/day plant) 30

The catalyst can operate for several years provided that the reacting gases

are well mixed and that no hot spots (which in extreme cases could reach

1500

C) are allowed to develop in the catalyst bed. The gas distributor plays an

important part in protecting the catalyst. The temperature of the catalyst bed in

the secondary reformer falls as gas passes through the bed. At the inlet to the

bed, the temperature is extremely high due to the highly exothermic combustion

reaction in which all of the oxygen is consumed. The temperature of the bed

towards the outlet then falls as the endothermic reforming reaction takes place.

It is interesting to note that when the temperature of the primary reformer outlet

falls and the slippage of methane increases, the temperature of the outlet of the

secondary reformer also falls. Typical operating conditions are shown in Table

9.13.

Secondary reforming catalyst contains about 7% nickel oxide, supported on

temperature-resistant α–alumina or calcium aluminate, in the form of raschig

rings. Additional catalyst used at the top of the bed as a heat guard is in the form

of large solid cylinders of α-alumina containing 5% nickel oxide.

9.5. CARBON MONOXIDE REMOVAL

Carbon monoxide leaving the secondary reformer is converted to useful hydro-

gen by the water gas shift reaction in a two-stage recovery section. The original

high temperature shift reactor is now combined with a low temperature reactor

filled with a copper catalyst. The removal of carbon monoxide is shown in

Fig-

ure 9.4. Since copper catalysts are extremely prone to poisoning by sulfur and

chlorin

e compounds, it is therefore essential that the concentration of these con-

taminants is reduced to an absolute minimum.

376

Figu

lytic

Twi

9.5.1.

In 191

y rea

the fir

ased

work

has be

stabili

operat

togeth

use, th

and at

state.

mody

the me

Chapter 9

e 9.4. Typical

beds. Reprinte

h Temper

he water-g

tion with ste

t commercial

n catalysts d

ating from the

ce then the

n iron oxide,

ing the cataly

ng conditions

e catalyst as

r with a vari

catalyst mus

the same tim

nder the typi

amically-stabl

allic state. So

ariations of ca

from Catalyst

ature Carbo

s shift reactio

m and to incr

ammonia pla

scovered by

late 1800s.

asic formulat

stabilized wi

ts and produc

ave changed

upplied consi

ble amount

be activated

e, any chrom

al conditions

state for the

e HTS catal

bon monoxide

Handboo

, 2

Monoxide C

has been us

ase the poten

t began oper

ild of BAS

on of high te

h chromium

ing it in large

ts mainly of

f the hexaval

by reduction

um trioxide i

for the shift

xides of, and

sts may cont

levels in HT a

ed., by kind p

nversion

d to remove c

tial hydrogen

tion in 1913.

in 1912

perature shi

xide, althoug

quantities hav

ematite and

nt chromium

o magnetite u

also reduce

eaction, mag

there is no fu

in traces of s

d LT shift cata

rmission of M

arbon monoxi

production si

The process

ile following

t catalyst (H

the methods

e been refined

hromium oxi

trioxide. Bef

ing process g

to the trival

etite is the th

ther reductio

lfu

, particula

as,

er-

Synthesis Gas

377

if the catalyst has been manufactured from ferrous sulfate. When commissioning

such a catalyst, most of the sulfur is reduced to hydrogen sulfide by the process

gas and this amount would be sufficient to destroy the LTS catalyst if it were

allowed to pass into the catalyst bed. It is normal practise, therefore, to flare the

gas product until it is free from sulfur, before allowing the low temperature shift

to come on line.

The forward shift reaction that is the conversion of carbon monoxide into

hydrogen and carbon dioxide is quite strongly exothermic, and in common with

all exothermic reactions, the level of conversion of carbon monoxide to products

at equilibrium is greater at lower temperatures. The equilibrium constant is in-

dependent of pressure, but higher conversions are also obtained at higher steam

ratios. Unfortunately, the preferred iron catalyst is not active at low temperature

and plants operate in the temperature range 350° - 500°C. Furthermore, the more

active catalysts based on copper that are active at low temperatures are not suffi-

ciently stable to be able to withstand the exotherm associated with the shift reac-

tion when the feed gas contains high levels of carbon monoxide. The exact con-

ditions for operation of the high temperature shift converter are therefore deter-

mined by the carbon monoxide content of the gas entering the reactor, and the

steam ratio used in the primary reformer.

At the time when natural gas was introduced as feed to a steam reformer

from about 1940 onwards, most ammonia plants were designed to use single

beds of HTS catalyst and the concentration of carbon monoxide in the outlet gas

was about 2%. The synthesis gas from the older plants, which used coal as feed-

stock, contained higher concentrations of carbon monoxide, and it was often

necessary to control the temperature by splitting the reactor into two or more

separate beds, with inter-bed cooling or by the incorporation of a quench system.

More recently, when the first ammonia plants based on natural gas were first

operated at the higher pressures, a two bed/inter-cooled reactor design was also

used and the concentration of carbon monoxide in the exit gas was lowered to

about 1% It was then practicable to use a methanator to hydrogenate the residual

carbon monoxide to methane, rather than operating with the copper liquor

scrubbing stage which had previously been used for the final removal of carbon

monoxide.

9.5.2 High Temperature Conversion Catalysts

The first HTS catalysts were reported to operate for about two years before re-

placement was required. As production techniques were developed, however,

catalyst lives improved so that by 1940, lives of more than 14 years were regu-

larly achieved. There were few poisons which affected the catalyst performance

although sulfur, which was the most common impurity in early plants, did sul-

fide the magnetite. This reaction was, nevertheless, reversible. If hydrogen sul-

fide levels exceeded about 300 ppm, sulfided catalysts could not be regenerated