Lloyd L. Handbook of Industrial Catalysts

Подождите немного. Документ загружается.

378 Chapter 9

and about twice the normal volume of catalyst had to be used to compensate for

the lower activity of iron sulfide.

The catalysts were originally produced from readily available materials

such as copperas, the ferrous sulfate waste product from steel works, and

chrome tan, an industrial form of chromic acid. The process involved precipita-

tion from ferrous sulfate solution with sodium carbonate. The precipitate was

carefully washed, to remove soluble impurities but early catalysts still contained

up to 1-2% insoluble ferric and chromic hydroxyl sulfates, which formed hydro-

gen sulfide during catalyst reduction. It was possible for the concentration of

hydrogen sulfide in the process gas to reach more than 250 ppm, gradually fall-

ing over a period of several days, before the concentration fell to a an equilibri-

um level. Of course, on most occasions hydrogen sulfide levels were lower than

this. When catalysts were used in naphtha based town gas plants during the

1960s and, later, in single stream ammonia plants with copper based LTS cata-

lyst, the HTS catalyst production method had to be modified. With low sulfur

catalysts it was possible to produce sulfur free gas following reduction in less

than twelve hours.

A further undesirable feature of catalysts containing chromium is that dur-

ing the calcination stage of preparation, a proportion of the chromium oxide is

oxidized to hexavalent chromium at temperatures in the range 250°–350°C. The

concentration of hexavalent chromium falls to an acceptable level of less than

1% when the calcination temperature is increased to about 430°C. High levels of

hexavalent chromium in the catalyst lead to a significant exotherm during reduc-

tion, which can lower the activity of the catalyst. Furthermore, handling of the

catalyst during manufacture can be a very dusty procedure, and contact with the

dust containing hexavalent chromium is hazardous to the workforce, since it

believed to have carcinogenic properties.

9.5.2.1 Operating Conditions

In modern, single stream ammonia plants there is little scope in the design to

make significant changes to the operating conditions in any of the individual

catalyst reactors. Operating conditions for the carbon monoxide conversion reac-

tion are shown in Table 9.14. The only practical variable is operating tempera-

ture which can be slowly increased as catalyst loses activity.

The composition of the catalyst can affect its performance in many ways.

For example, changes in the formulation to lower the amount of hydrogen sul-

fide evolved during reduction led to a significant physical weakening of the

catalyst, which became much more prone to damage caused by the condensation

of water or contamination by potash. Any water that condensed during plant

start-up could also wash out any soluble chromium and lead to loss of stability.

Most problems led to an increase in pressure drop or maldistribution of gas. Cat

Synthesis Gas

379

TABLE 9.14. Carbon Monoxide Conversion Catalysts Operation.

Inlet HTS Outlet HTS Inlet LTS Outlet LTS

Gas rate (m

/h) 133,000 146,000

Steam ratio 0.6 0.46

Pressure (atm) 29 28

Catalyst volume (m

) 60 64

Temperature (°C) 360

20 430

20 210

10 230±10

Gas composition (%):

CO 13 3.5 <0.5

8 16 18

56 60 61

22 20 20

/A ~0.3 0.5 0.5

alysts were usually, however, able to operate for a period ranging from two to

four years.

With the drive to more energy-efficient ammonia plants in the 1980s, the

steam ratio fed to the primary reformer was significantly decreased. Apart from

the effects in the reformer which have already been described, a number of

changes were detected in the operation of the HTS catalyst. The lower steam

ratio was still greater than the level calculated from the thermodynamics at

which magnetite could be reduced to the metallic state (Figure 9.5). However,

the lower concentration of steam resulted in a higher level of carbon monoxide

in the process gas, as described earlier in this chapter, and this increase in the

carbon monoxide/carbon dioxide ratio gave rise to a more reducing atmosphere.

This, in turn, led to the formation of more carbon by the disproportionation of

carbon monoxide, also known as the Boudouard reaction. The end result was an

increase in pressure drop, and further disintegration of the catalyst.

Carbon monoxide also reacted with magnetite forming carbides which cata-

lyzed the production of hydrocarbons by Fischer-Tropsch reactions. It has since

been shown that both reactions can be suppressed by the addition of copper to

the existing iron-chromium catalyst. This has allowed operation of HTS cata-

lysts in existing plants down to a steam ratio as low as 0.4 compared with about

0.6 in the early single stream plants.

In more recent ammonia plant designs, in which the large steam reforming

furnace has been replaced with gas-heated reformers or combined autothermal

reformers, the HTS catalyst can be replaced by a temperature resistant copper

catalyst that can operate at temperature as low as 260°–270°C.

9.5.3. Low Temperature Carbon Monoxide Conversion

As the industry moved towards cleaner feedstocks, such as naphtha and naturals

gas, and the use of increasingly efficient gas purification systems, the synthesis

380

gas fr

made

conver

er ox

200°–

to 0.2

–

catalys

ation

cedure

chlori

import

lyst w

large a

obvio

Chapter 9

Figure 9.5.

conventional

at H

~400

C. Rep

sion of M. T

m the reform

t possible to

sion and to si

de-zinc oxid

50°C, lowere

–

0.3%. Conditi

t from the con

f the new cop

was improve

e, was under

nt part of the

s important t

ich can oper

monia plant

s penalty is t

s cost is the o

inimum ratio

HT shift cataly

> 2x10

inted from Cat

igg.

rs contained

onsider the

lify plant d

catalyst in 1

the concentr

ns were only

ensation of

per catalyst w

however an

tood the low

ammonia synt

understand t

te at maximu

. In a typical

e daily cost

erall effect o

of steam to hy

ts. Fe

eco

t ~550

C and

lyst Handboo

nly trace am

se of copper

sign even furt

963, operatin

tion of carbo

limited

y th

ater at low te

as not reliabl

the effect o

temperature s

esis process.

e commercial

conversion

1000 tonnes p

f an unexpe

an increase i

drogen for red

es stable with r

bout half this

ed., by kin

unts of catal

catalysts for

her. The intro

at temperat

monoxide in

need to prev

perature. Un

. Once the m

catalyst poi

ift (LTS) ca

significance

or long, predi

r day ammo

ted plant clo

the volume

ction of

espect to

value at

permis-

st poisons. T

arbon monox

uction of a c

res in the ra

the synthesis

nt damage to

ortunately, op

nufacturing p

ons,

articula

alyst became

f an active c

ctable

eriods

ia plant the m

ure. Anther l

f carbon mon

er-

ide leaving the low temperature reactor. One volume of hydrogen is lost for every

Synthesis Gas

381

again an extra three volumes of hydrogen are lost as the carbon monoxide is

removed in the methanator.

CO + H

O Æ H

+ CO

1 vol. H

not made (9.13)

CO + 3 H

Æ CH

+ H

O 3 vol. H

consumed (9.14)

The additional methane formed must then be purged from the synthesis loop

which leads to even more hydrogen loss as significant hydrogen is also removed

during the purge. Overall, the equivalent of failing to convert just 0.1% of car-

bon monoxide is more than 4000 tons of ammonia every year, or about 1% of

the design production.

Every catalyst in an ammonia plant is important but, because low-

temperature carbon monoxide conversion catalyst is likely to be the most sensi-

tive to poisons, good operation of this catalyst is probably the most significant.

Apart from using the best available catalyst, it is usual to make an allowance for

catalyst poisoning by increasing the catalyst volume in the reactor.

9.5.3.1 Operation

An LTS catalyst should be sufficiently active to give a high conversion for a

given volume of catalyst at the minimum practicable temperature. It should also

be thermally stable and operate for the design period with maximum carbon

monoxide conversion. With proper design and good upstream poisons removal,

a typical catalyst lifetime is about three years.

The catalyst must be activated by careful reduction to convert the copper

oxide component to metallic copper before use (Figure 9.6). The reduction reac-

tion is exothermic and the reduction process must be carried out using an inert

carrier gas to which a low concentration of hydrogen has been added, so that the

temperature of the bed does not exceed 250ºC. After reduction and when a new

charge of catalyst is commissioned, the bed should be heated to a temperature

greater than the dew point of the feed gas so that liquid water does not condense

onto the catalyst when the feed is first admitted. The temperature is then in-

creased gradually until the reaction begins. At this stage, the temperature can be

adjusted until the required degree of conversion is achieved and a satisfactory

temperature profile is seen in the bed.

As the catalyst ages, the peak temperature moves slowly down the bed

(Figure 9.7). The rate of movement depends both on the thermal stability of the

catalyst and on the total amounts of different poisons that accumulate on the

bed. The temperature at the inlet to the bed can be increased gradually to com-

pensate for deactivation of the catalyst. Eventually the catalyst must be changed

because the concentration of carbon monoxide leaving the bed at equilibrium

volume of carbon monoxide which is not converted by the LTS catalyst and

Figure 9.6. Sch

Catalyst Handb

matic arrangement

, 2

ed., by kind

for the

eduction of

ermission of M. Tw

LT shift catalyst us

gg.

ng a typical once-t

rough system. Repr

inted from

382 Chapter 9

increa

the de

eratu

life ca

any co

into th

no wa

guard

oison

oxide

seriou

The u

carbon

Figure 9.7. T

53-1 LT shift

kind permissi

es as the temp

ree of slippag

e profile in t

be estimate

densation of

active layer.

s possible to

of removin

ed can be ins

can be remo

as used as a

poison for L

e of catalyst

monoxide co

ypical temperat

catalyst. Repri

on of M. Twigg

erature is incr

e of carbon m

e bed should

or any malo

ater in the c

inimize the e

them and a

alled in a sep

ed. This bed

guard to rem

S catalyst is

s guard has

version.

re profile thro

ted from Catal

ased, and a p

noxide is not

e measured

eration of th

talyst bed mi

ffects of pois

short catalyst

rate reactor

eed to be cha

ve sulfur co

hlorine, ordi

he advantage

Synt

gh a bed of ICI

st Handboo

, 2

int is eventua

economically

egularly so t

plant detect

ht wash solu

ns in the

roc

life is unacc

efore the mai

ged

egularly

pounds but,

ary LTS catal

of always en

hesis Gas

Catalyst

ed., by

ly reached w

viable. The te

at the remain

d. For exam

le poisons do

ss gas if ther

ptable. A s

bed so that

. Originally, z

ecause the m

yst is prefera

uring maxim

le,

all

le.

384 Chapter 9

9.5.3.2 Catalyst

The first LTS catalysts were based on copper and zinc oxides, (CuO)/2ZnO),

and were prepared by calcination of the precipitated hydroxycarbonates of zinc

and copper. By 1965, it was found that incorporation of some alumina into the

formulation led to an increase in thermal stability and improved resistance to

poisons. Depending on the plant conditions, the operating life of the catalyst was

increased from about six months to more than two years. The most successful

catalysts were produced by simultaneous precipitation of the copper, zinc and

aluminum hydroxycarbonates, at about pH = 7, to optimize the particle size and

distribution of the oxides. They were similar in composition to the successful

low-pressure methanol catalysts produced in the same way although the copper

content was significantly lower. Other catalysts were introduced which had high

copper content in an effort to achieve a high activity. High copper catalysts were

more susceptible to poisoning in some plants. Another catalyst, based on copper

and zinc oxides but with high chromium oxide content, was expected to be re-

generable, to be able to compensate for the effects of poisons. However, it was

very susceptible to poisoning by sulfur, needed a long period off line for regen-

eration, and its use was not successful. This catalyst has now proved to be an

acceptable chlorine guard for a much more stable copper-zinc-aluminum cata-

lyst. Details of approximate catalyst compositions are given in Table 9.15.

Most spent catalysts contain varying amounts of sulfur, chlorine and silica

poisons depending on the age of the catalyst and the purity of synthesis gas. The

effects of each poison can therefore be considered separately in relation to the

differing operating conditions. In some cases, up to 3–4% sulfur has been meas-

TABLE 9.15. Carbon Monoxide Conversion Catalyst Compositions.

Catalyst Composition

Low Temperature Shift:

Composition (wt%) loss free Low Copper High Copper Chromium

Copper oxide 32 42 20

Zinc oxide 55 47 35

Alumina 13 10

Chromia – – 45

Ignition loss 900°C

<20 <10 <10

Bulk density (kg liter

-1

) 1 1.4 1.2

High Temperature Shift:

Composition (wt%) loss free High Steam Ratio Low Steam Ratio

Ferric oxide 89–91 85–89

Chromia 9–10 7–11

Water soluble CrO

Copper oxide – 2

Ignition loss 900°C

<10 <10

Bulk density (kg liter

-1

) 1.3 1.3

Synthesis Gas

385

ured in samples from the top of catalyst beds when the life has been as long as

seven years. The level of chloride in the same samples has been less than 0.02-

0.04%, and it can be concluded that sulfur is held firmly by the catalyst and may

not be a serious problem. A possible explanation for this observation is that alt-

hough sulfur compounds are initially adsorbed by the copper crystallites, they

are rapidly transferred to the small zinc oxide crystals. Zinc sulfide, which is

thermodynamically more stable, is easily formed. This mechanism is dependent

upon the manufacturing process to provide small crystallites in the catalyst

structure. Similarly, silica deposits of up to 1.5wt% at the top of the LTS cata-

lyst bed have been detected after satisfactory lives of two to four years.

On the other hand, while chlorides accumulate near the top of the catalyst,

they are more mobile and can be detected in significant concentrations, up to

0.05%, at all levels in a deactivated bed. Although reasonable lives of at least

two years can often be achieved in the presence of chloride there is more rapid

movement of the peak in temperature profile, and the concentration of carbon

monoxide in the outlet gas increases more rapidly. Surface chlorides, which are

formed by reaction with zinc oxide, are mobile and sinter the catalyst surface.

Chlorides are also soluble in condensed steam and can be washed down onto

lower, more active catalyst layers.

Sulfur and chlorine are both present in the hydrocarbon feed, the process

steam and the lubricating oils used while sulfur may also come from the high

temperature shift catalyst. A major source of chlorine is from the air used in the

secondary reformer. Silica is present in process steam but also comes from the

refractory linings or support materials used in the reforming section.

9.6 METHANATION

Carbon monoxide and carbon dioxide cause the temporary deactivation of am-

monia catalysts. Carbon dioxide can also lead to further problems because it

forms ammonium carbonate in the make-up gas compressor and the synthesis

loop. The removal of these impurities is, therefore, a vital step in the purification

of synthesis gas. Removal of carbon dioxide has generally been via absorption

in some suitable solvent, whereas at the present time, the concentration of car-

bon monoxide is reduced to a low level by reaction with steam in the water gas

shift reaction, prior to almost complete removal by an additional procedure.

Methanation, the conversion of carbon monoxide and dioxide to methane,

was described by Sabatier in 1905, during his work on hydrogenation catalysts,

and since then it has been used in several industrial processes. Since about 1955,

as the use of large, single-stream\m ammonia plants became more widespread,

the methanation reaction has become the preferred way to remove the final trac-

es of carbon oxides from the process gas.

386

Fig

Thi

9.6.1

When

0.7%

A sim

250°–

ides o

for ev

Chapter 9

re 9.8. Photo

vessel contain

d., by kind per

Operation

the plant is o

xides of carb

le, single-

20°C dependi

carbon (Figur

ry 0.1% carb

raph of a met

d about 25 ton

ission of M. T

erating norm

n, which need

reactor can

g on the cat

9.8). There i

n monoxide

anator in a typ

e catalyst. Repr

igg.

lly, the synt

to be remove

e used at an

lyst activity a

an adiabatic

nd 6°C for e

ical ammonia

nted from Cata

esis gas cont

during the

inlet tempera

d the concen

emperature ri

ery 0.1% car

ynthesis plant.

lyst Handboo

ins around 0

ethanation sta

ure in the ra

ration of the

e of about 7.4

on dioxide c

3–

°C

Synthesis Gas

387

verted. Although the catalyst should not normally be operated at temperatures

exceeding 450°C it is frequently allowed to overheat to temperatures as high as

600ºC during periods of maloperation, but seems to incur little deactivation..

There is probably more chance of damage to the reactor in these situations. At a

space velocity in the range 5000-7000 hours

-1

methanation is virtually complete

at a space velocity in the range 5000-7000h-1 and the outlet carbon oxides level

is usually less than 5ppm. Good methanation catalysts have a normal life of

about ten years.

The most common catalyst poisons are derived from traces of solvent car-

ried over from carbon dioxide removal process. These simply block the catalyst

pores and, providing no sulfur or arsenic compounds are used, can easily be re-

moved by washing the catalyst with water to restore activity. If sulfur enters the

catalyst bed from any other source it will rapidly poison the methanation cata-

lyst. The low temperature shift catalyst, up-stream of the methanator, usually

acts as a very efficient sulfur guard!

9.6.2. Catalyst

Different suppliers have produced methanation catalyst in a wide range of

shapes and sizes, ranging from spheres and pellets to extrusions and granules,

containing between 15–30% nickel oxide. Some typical support materials react

with nickel oxide to form relatively stable spinels, which cannot readily be re-

duced to give an active catalyst, and these materials are not used. It is better if

the nickel oxide can be precipitated in solid solution with a thermally stable ma-

trix similar to those supports used in reforming catalysts. Typical supports are

alumina, silica and magnesia, sometimes bound by alumina cement. The most

important catalyst property is that it can be reduced easily at a temperature at or

below 300°C to give an active and stable catalyst. The nickel can be combined

with the support by either impregnation or precipitation, provided that the final

catalyst is thermally stable.

The nickel oxide component of the catalyst is reduced to the metal with

process gas, by gradually increasing the temperature of the bed to about 300ºC.

If necessary, the reduction procedure can be assisted by allowing a small propor-

tion of the process gas to bypass the low temperature shift converter. This results

in a temporary increase in the concentration of the oxides of carbon in the reduc-

ing gas to around 1–2%, which are then hydrogenated into methane over the

active part of the catalyst. The resulting exotherm from the reaction increases

the temperature of the bed to about 400

C, so that any remaining nickel oxide

can be reduced fully. During the reduction period the carbon monoxide and car-

bon dioxide content of outlet gas is analyzed at regular intervals. Full operation

can start when the level falls to less than 20 ppm. This normally takes 10–15

hours. If the LTS reactor cannot be by-passed the catalyst reduction should be