1.

Introduction

Sulfur dioxide is produced industrially in greater quantities than any other

single sulfur compound. It is generated as the first stage in the
manufacture of virtually all the sulfuric acid used by industry, irrespective
of the basic raw material. It may arguably be considered the linchpin of
the sulfur sector of the heavy inorganic chemicals industry.

Sulfur dioxide is generated intentionally by combustion of elemental

sulfur or by oxidation of sulfides. Since most carbonaceous or
hydrocarbon fuels and the ores of many nonferrous metals contain
sulfidic sulfur, vast volumes are also produced involuntarily in large fuel-
consuming installations, such as power stations, and in nonferrous metal
smelters. On account of its chemical and physiological properties,
coupled with the scale on which it is produced, waste sulfur dioxide is
now considered as one of the gravest chemical threats to the global
environment. Progress has been made in controlling atmospheric
discharges of sulfur dioxide from the most amenable sources — an aim
towards which much of the technological development relating to sulfur
dioxide has been directed in the past decade. However, sulfur dioxide
pollution, particularly where it crosses national boundaries, continues to
be a matter for argument and debate within and between governments
and industry.

Sulfur dioxide has been known since ancient times for its bleaching and
disinfecting effects [1]. These days it finds widespread use in a rather
diverse range of chemical and other industries, including wood pulping
and paper production, textile manufacture and food processing, where it
serves as reagent, bleaching agent, preservative or solvent. However,
the outstanding use of sulfur dioxide — about 98 % of the total amount
used by industry — is in the manufacture of sulfuric acid.
2. Properties

Physical Properties. Sulfur dioxide, SO2, is a colorless, nonflammable,

toxic gas with a characteristic pungent smell and acid taste. Its most
important physical properties are summarized below [2-8]:

Molecular mass 64.06

Melting point (101.3 kPa) –75.5 °C
Latent heat of fusion (at mp) 115.6 J/g
Dynamic viscosity at 0 °C 368 Pas
Density at – 10 °C 1.46 g/cm3
Critical density 0.525 g/cm3
Critical pressure 78.8 bar
1
 Critical temperature 157.5 °C
Boiling point (101.3 kPa) –10.0 °C
Latent heat of vaporization (at bp) 402 J/g
Standard density at 0 °C (101.3 kPa) 2.93 kg/m3
Density relative to air (0 °C, 101.3 kPa) 2.263
Molar volume (0 °C, 101.3 kPa) 21.9 L/mol
Standard enthalpy of formation –4636 J/g
Specific heat capacity cp (101.3 kPa):
0 °C 586 J kg–1 K–1
100 °C 662 J kg–1K–1
300 °C 754 J kg–1K–1
500 °C 816 J kg–1K–1
cp/cv (15 °C, 1013 mb) 1.29

Graphical representations of the temperature functions of the various

physical properties can be found in [2], [6], [9]. Interpolation formulas for
thermodynamic characteristics such as specific heat are given in [2], [5],
[7], [10], [11]. The vapor pressure curve of liquid sulfur dioxide as a
function of temperature is shown in Figure 1. Figure 1 also shows the
dewpoints of gases of varying sulfur dioxide contents at different
pressures. These are important in the production of pure sulfur dioxide
by condensation (see Chemical Production from Sulfur Trioxide).

Figure 1. Vaporization curves for

sulfur dioxide

Curve for 100 % SO2 shows

relationship of boiling point to
equilibrium vapor pressure. Curves for
impure SO2 mixtures show
dependence of dewpoint on SO2
partial pressure

2
The solubility of sulfur dioxide in water rises in proportion to the SO2
partial pressure in good conformity with Henry's law and is increased by
lowering the temperature, as represented graphically in Figure 2 [12]. In
solution, sulfur dioxide is present mainly as SO2 molecules, but Raman
spectroscopy confirms the presence of minor amounts of the species
, and H2SO3 [5]. The latter, sulfurous acid (the anhydride of
which is sulfur dioxide), exists only in aqueous solution. Aqueous
solutions of alkaline compounds will absorb much more sulfur dioxide
than pure water because of the formation of hydrogensulfite and sulfite
ions [12].

Figure 2. Solubility of sulfur dioxide in

water [12]

3
The solubility of sulfur dioxide in sulfuric acid is of industrial importance.
It first decreases with rising sulfuric acid concentration, reaching a
minimum at an H2SO4 concentration of ca. 85 %; beyond this
concentration it increases again (Fig. 3) [13].

Figure 3. Solubility of sulfur dioxide

in sulfuric acid at [13]

Liquid sulfur dioxide is miscible in all proportions with liquid sulfur

trioxide. Sulfur dioxide also dissolves readily in most organic liquids (e.g.,
methanol, ethanol, benzene, acetone, or carbon tetrachloride). With
ether, carbon disulfide, chloroform, and glycol it is completely miscible,
even at low temperatures [2], [5].

4
Chemical Properties. Sulfur dioxide is very stable; thermal dissociation
becomes significant only above 2000 °C. It can be decomposed by
shock waves, irradiation with ultraviolet or X-rays, or by electric
discharge [2], [5].

The reaction of sulfur dioxide with oxygen to form sulfur trioxide is

industrially the most significant of all its reactions because of its
importance in sulfuric acid production. In the gas phase, it only takes
place at elevated temperatures and, for a satisfactory yield of sulfur
trioxide, it requires the presence of a catalyst (see  Sulfuric Acid and
Sulfur Trioxide). In aqueous solution, sulfur dioxide is oxidized to sulfuric
acid at low temperature by air in the presence of activated coke or
nitrous gases, or by oxidizing agents such as hydrogen peroxide.

The reduction of sulfur dioxide with hydrogen, carbon or carbon

compounds such as methane or carbon monoxide is also of industrial
interest [2], [5]. These reactions require high temperatures, or catalysts,
or both. They result in mixtures of elemental sulfur and hydrogen sulfide.
If carbon or a carbon compound is used as reducing agent,
carbon- containing species such as carbon dioxide, carbonyl sulfide, and
carbon disulfide are formed as well.

Sulfur dioxide reacts with hydrogen sulfide in accordance with the Claus
reaction (Eq. 1). This is the basis of the main method by which elemental
sulfur is recovered from hydrogen sulfide in industry (see also
 Sulfur – The Claus Process)

(1)

Sulfur dioxide generally oxidizes metals at elevated temperatures,

simultaneously forming metal sulfides and oxides [2], [5].

Liquid sulfur dioxide is a relatively efficient solvent with some waterlike

properties [2], [3]. Polar inorganic compounds are usually insoluble or
only sparingly soluble in liquid sulfur dioxide [2], whereas covalent
inorganic and organic compounds are often dissolved, mostly forming
stable solvates. The fact that aromatic hydrocarbons dissolve more
readily than aliphatics in sulfur dioxide is exploited on an industrial scale
for the extraction of aromatics from crude oil (Edeleanu process) [5].
3. Raw Materials for Sulfur Dioxide Production

5
Sulfur dioxide is manufactured industrially from the following raw
materials:
1. Elemental sulfur
2. Pyrite
3. Sulfide ores of nonferrous metals
4. Waste sulfuric acid and sulfates
5. Gypsum and anhydrite
6. Hydrogen sulfide-containing waste gases
7. Flue gases from the combustion of sulfureous fossil fuels

Elemental Sulfur. Elemental sulfur is the most important raw material for
sulfur dioxide production worldwide. For information on sources of
elemental sulfur and production methods, see  Sulfur.

Today, elemental sulfur used for sulfur dioxide production is considered

not only as a raw material but also, to an increasing extent, as an energy
source. Modern sulfur-burning plants are normally connected to a
turbo – alternator combination for electrical power production; these are
known as cogeneration units. Elemental sulfur is the raw material of
choice for production of 100 % sulfur dioxide and of industrially
significant sulfites, such as calcium hydrogensulfite solution, which is
used in cellulose production. Of the 33.5×106 t of elemental sulfur
produced worldwide in 1981, almost 28×106 t were burnt to sulfur dioxide
for conversion to sulfuric acid [14]. The proportion of industrial sulfur
dioxide production that is based on elemental sulfur varies in different
countries. Of the total amount of sulfur-in-all-forms used for producing
sulfur dioxide-containing gases for sulfuric acid production, the share of
elemental sulfur was ca. 80 % in the United States, about 50 % in
Germany, and probably less than 25 % in the former Soviet Union.

Pyrites. Pyrites and other iron sulfide ores still constitute today, in some
countries, an important raw material basis for sulfur dioxide production,
especially as the primary stage for sulfuric acid manufacture [15]. The
most important iron sulfide minerals are pyrite and marcasite, both FeS2,
as well as pyrrhotite or magnetopyrite, the compositions of which
correspond approximately to the formula Fe7S8. They occur in varying
purity as sedimentary rocks in all formations, and as minor constituents
in coal deposits. Pyrites are usually slate gray or sometimes greenish in
color, depending on the deposit. Other constituents, apart from gangue
materials that are present in varying proportions in pyritic ores include
sulfides of nonferrous metals, especially of copper, zinc, and lead as well
as, to a lesser extent, cobalt, nickel, manganese, bismuth, silver and

6
gold. In some cases, pyrites may also contain arsenic, antimony,
tellurium, and thallium as well as fluorides and chlorides.

The production of pyrites, either as crude pyrites or as beneficiated

pyrites, has decreased drastically within the last years. This was a result
of a depressed fertilizer market, lower sulfur prices, and increasing
amounts of involuntarily produced sulfuric acid. Flotation pyrites are,
however, obtained in markable quantities as a byproduct of nonferrous
metal sulfide beneficiation plants. The sulfur content of pyrites
concentrates may be as high as 50 %. Average analyses of pyrite ores of
different origins are presented in Table 1.

Table 1. Typical analyses of various types of pyrites

Country Location Grade* Typical content, wt %

S Fe Cu Zn Pb As

Germany Meggen flot. 43 – 37 – 0.03 – 0.4 – 0.4 – 0.06

48 40 0.1 1.2 0.6
Cyprus Kalavassos flot. 45 – 40 – 0.2 – 0.1 0.1 0.01
48 44 0.3
Limni flot. 44 – 37 – 0.2 – 0.1 0.1
46 40 0.3
Finland Outokumpu flot. 43 – 40 – 0.2 – 0.1 – 0.3 0.03
48 44 05 1.1 0.06
Greece Kassandra flot. 47 – 41 – 0.1 – 0.6 – 0.5 – 0.5
49 44 0.3 1.5 0.7 1.8
Italy Maremma flot. 48 – 44 – 0.02 0.09 0.02 0.045
50 45
former Bor flot. 44 – 41 – 0.25 – 0.04 – 0.04 – 0.03
Yugoslavia 49 43 0.6 0.1 0.1 0.07
Norway Orkla fine 41 – 37 – 2.2 – 1.6 – 0.05 – 0.04
46 40 2.7 2.0 0.06 0.08
Folldal flot. 48 – 44 – 0.06 – 0.07 – 0.04 – 0.03
50 45 0.07 0.08 0.05 0.05
Portugal Aljustrel fine 45 – 39 – 0.65 – 2.0 – 1.2 – 0.4
47 40 1.33 3.0 1.4 0.5
Romania Baia Sprie fine 44 – 37 – 0.6 – 4 – 5 1.0 – 0.03
45 40 0.7 1.1 0.05
Sweden Boliden flot. 49 – 45 – 0.04 – 0.2 – 0.04 – 0.05
7
 50 46 0.15 0.4 0.05 0.2
Spain Rio Tinto flot. 48 – 43 – 0.15 – 0.08 – 0.08 – 0.26
49 44 0.4 1.0 1.0 0.4
Tharsis flot. 46 – 41 – 0.7 – 1.3 – 0.5 – 0.4
48 43 1.1 1.6 0.6 0.5
San Telmo fine 45 – 40 – 1.9 – 1.0 – 0.05 – 0.1
47 41 2.5 2.6 0.15 0.2
former flot. 46 – 39 – 0.9 – 0.1 – 0.1 – 0.05
Soviet 48 41 1.0 0.2 0.2 0.1
Union

*
Flot. = flotation pyrite; fine = fine pyrite.

Depending on the grain size, a distinction is made between flotation

pyrites (grain size < 0.1 mm), fine pyrites (0 – 6 mm), coarse pyrites
(1 – 12 mm and more), and lump pyrites (10 – 100 mm and above).

Whereas in Japan, for instance, all the pyrite ores produced are
processed exclusively in the country, other countries still export pyrites
as raw materials for sulfuric acid production, especially the former Soviet
Union, Spain, and Norway. In 1982, the Soviet Union exported about
178 000 tonnes S, Spain 145 000 tonnes and Norway 82 000 tonnes S in
the form of pyrites [16]. From 1965 until about 1975, world pyrites
extraction remained almost constant at an annual rate of about
10 – 12×106 t sulfur content. But, as industrial sulfur demand increased
substantially during that period, pyrite was in fact declining consistently in
its relative importance among the sulfur raw materials. Processing of iron
sulfide containing gold and silver ores has increased in importance,
resulting in additional involuntarily produced SO2.

Nonferrous Metal Ores. Nonferrous metal sulfide ores are another

important raw material source for sulfur dioxide production. In the
pyrometallurgical processing of sulfide ores for extraction of copper,
nickel, zinc, and lead waste gases containing sulfur dioxide are inevitably
produced. In most countries it is no longer permissible, for environmental
reasons, to discharge these gases directly into the atmosphere because
of the large plant capacities used nowadays. The development of the
various smelting processes is, therefore, strongly influenced by the dual
needs to capture the released sulfur dioxide as completely as possible
and to generate waste gases with high sulfur dioxide concentrations so
that they can be processed as economically as possible [17].

8
The rate at which sulfur dioxide is produced in the processing of
nonferrous metal sulfide ores can be estimated approximately on the
basis of metal smelter production statistics, assuming characteristic
mean sulfur/metal weight ratios in the ore concentrates of each metal.
For 1992 a figure of ca. 20×106 t of SO2 can be given. A comparison of
smelter productions for the most important nonferrous metals in 1979
[18], which can be assumed to be based mainly on sulfide ores or
concentrates enriched by flotation, is given in [18].

Other Materials. Sulfur dioxide is normally only recovered for

environmental reasons from waste sulfuric acid and sulfates, from
hydrogen sulfide- containing industrial waste gases and from power
station flue gases. Calcium sulfate, in the form of natural gypsum or
anhydrite, was formerly used in a few small plants as raw material for
sulfur dioxide and sulfuric acid production, but on account of their high
energy consumption such plants have either been shut down or modified
to process waste gypsum (e.g., from phosphoric acid manufacture).

All sulfur raw materials, except for elemental sulfur and pyrites, are
summarized in the statistics under the general item of “sulfur-in-other-
forms”. The worldwide production of “sulfur-in-other-forms” (with
nonferrous metal ores playing the dominant role) amounted in 1982 to
about 8.2 million tonnes, of which approximately 7.5 million tonnes were
consumed in sulfuric acid production, corresponding to 18 % of the total
amount of sulfur used in the production of sulfur dioxide in sulfuric acid
manufacture. The most important sources of sulfur-in-other-forms
accounted for the following quantities in 1982 (in 103 t of sulfur) [16]:

former Soviet Union 1700

United States 1378
Japan 1371
Canada 430
Federal Republic of Germany 408
Finland 221
Australia 160
Italy 152

4. Production from Elemental Sulfur

To burn sulfur in accordance with Equation (2), equimolar quantities of

sulfur and oxygen are required

9
 (2)

With air, which contains 20.95 vol % O2, sulfur dioxide concentrations of
ca. 20.5 vol % can in theory be obtained. But in practice, to ensure
complete combustion of the sulfur, air is generally supplied in excess, as
the problems which might result from unburnt sulfur condensing and
solidifying in colder parts of the plant are not worth risking.

As a prerequisite for complete sulfur combustion and for maintaining

constant sulfur dioxide concentrations under industrial conditions, it is
essential for the sulfur to be finely and evenly dispersed and intimately
mixed with the combustion air during burning. The general practice to
use liquid sulfur at 140 – 150 °C, at which temperature its viscosity is low
enough to enable it to be sprayed through nozzles. The high surface
area of the resulting fine droplets aids rapid evaporation (the normal
boiling point of sulfur is ca. 440 °C).

During the combustion process, it is S2 molecules in the vapor phase that

are actually oxidized. Evaporation of liquid sulfur initially produces mainly
S8 molecules, which are only decomposed to S2 to an appreciable extent
above 600 °C [19]. More than 60 % of the heat of reaction (about
9280 kJ/kg S) liberated in the combustion of sulfur to sulfur dioxide is
theoretically required for preheating the air and sulfur and for evaporation
and decomposition of the sulfur at 600 °C. Therefore, favorable
combustion is dependent not only on maximum mass transfer velocity
but also, very markedly, on intensive heat transfer in the initial phase.

When solid sulfur is used, the lowest ignition temperature in air under
atmospheric pressure is ca. 246 °C. However, sulfur dust clouds may
ignite spontaneously at temperatures as low as 190 °C, a phenomenon
which has to be borne in mind when handling and storing solid sulfur
[20].

Sulfur combustion in industrial oxygen or in oxygen-enriched air is

generally uneconomic on account of the high cost of oxygen, and it is not
currently used on an industrial scale.

The practicable upper concentration limit for the industrial production of

sulfur dioxide gases by burning liquid sulfur with air is at about 18 vol %
SO2. Normally, however, the gases from sulfur combustion furnaces
associated with sulfuric acid plants contain 10 – 12 vol % SO2,
corresponding to an O2/SO2 ratio of ca. 1 : 1, as required in double
10
absorption plants (see  Sulfuric Acid and Sulfur Trioxide).

Normally the combustion air used in the industrial production of sulfuric

acid plant feed gases from elemental sulfur is dried with concentrated
sulfuric acid. It is then compressed by a blower to1.3 – 1.5 bar to provide
the driving force needed to overcome pressure drop within the rest of the
plant. Drying and compression raise the temperature of the air to about
60 –100 °C.

Theoretically, the combustion temperature is ca. 1000 °C when liquid

sulfur is burnt at an initial temperature of 140 °C in dried air originally at
60 – 80 °C to produce a gas containing 10 vol % SO2. As can be seen
from Figure 4, the combustion temperature rises in direct linear
proportion to the sulfur dioxide concentration. It reaches about 1600 °C
when the combustion gas sulfur dioxide concentration is 18 vol %.

Figure 4. Theoretical sulfur Figure 4. Theoretical sulfur

combustion temperature as a function combustion temperature as a
of combustion gas SO2 concentration function of combustion gas SO2
concentration
(air temperature = 80 °C, sulfur
temperature = 140 °C) (air temperature = 80 °C, sulfur
temperature = 140 °C)
[Full View]

4.1. Sulfur Burning Processes

In industry nowadays sulfur combustion furnaces with atomizing burners
for liquid sulfur are used almost exclusively. Figure 5 shows the layout of
a Lurgi sulfur combustion furnace. The furnace comprises a cylindrical
steel shell lined with several layers of insulating and refractory bricks. It
is normally arranged horizontally, but when space is limited a vertical
version is available. The sulfur burner is mounted at one end of the

11
combustion chamber and a waste heat boiler is connected directly to the
gas outlet at the opposite end. The specific furnace size is very much
dependent on the type and efficiency of the burner used.

Figure 5. Lurgi sulfur furnace Figure 5. Lurgi sulfur furnace

a) Rotary atomizer; b) Sulfur furnace a) Rotary atomizer; b) Sulfur

furnace
[Full View]

Burner Types. There are various ways in which liquid sulfur can be
atomized to produce the fine spray needed for the most efficient possible
combustion. The main commercial burner types are the pressure-nozzle
burner, in which sulfur is injected under pressure through a spray nozzle
into the furnace chamber; the two- component burner, in which the sulfur
is released through a nozzle into a turbulent jet of air; and the rotary
burner, in which the sulfur is distributed centrifugally from the edge of a
spinning cup into a stream of air [21], [22].

With the simple pressure-nozzle burner, the liquid sulfur is atomized with
a single spray head. The relatively high pressure of 8 – 15 bar needed to
drive this type of burner is produced by a sulfur pump of special design
[23]. A steam-heated double shell around the nozzle pipe serves the
purpose of keeping the sulfur molten when the burner is not operating
and preventing overheating when it is. The main disadvantage of this
type of atomizer is its relatively low flexibility: major changes in the
throughput can only be accommodated by changing the nozzle. The
smaller-diameter nozzles are also prone to clogging.

The maximum sulfur throughput of a pressure-nozzle burner is ca. 72 t/d.

In the relatively large plants commonly used nowadays, it is, therefore,
necessary to install several nozzle systems with separate measuring and
control instruments. This also increases the flexibility of the system.

Two-component burners, in which the sulfur is chiefly atomized by the

12
combustion air, can be used for extremely low sulfur throughputs of
< 7 t/d, for which pressure-nozzle burners are inherently unsuitable. The
combustion air flows at high velocity from an annular space surrounding
the sulfur nozzle. The sulfur, supplied to the nozzle through steam-
jacketed pipes, is entrained in the air stream and finely distributed by
turbulence in the combustion air induced by deflector plates. The
necessary sulfur pressure is only ca. 2 bar. Modern two- component
burners are designed for sulfur throughputs of up to ca. 170 t/d. The turn-
down limit is usually around 15 % of capacity, although a compact design
intended for small-scale applications such as pulp mills is reported to be
able to run at as little as 5 % of its nominal capacity of 5 t/d. This is
because a variable-speed sulfur pump is used instead of the more
normal check valve arrangement to control the flow of sulfur [24].

The rotary burner, represented by the Lurgi-developed Luro Burner, is

nowadays the preferred type for large plants. The design of such a
burner is shown schematically in Figure 6. The sulfur passes at virtually
atmospheric pressure into an open cup rotating at 4000 – 6000 rpm. The
sulfur film on the inner wall is dispersed from the edge by centrifugal
force and is further atomized by entrainment in the so- called primary air,
which emerges at high velocity through the narrow annular space
between the rotating cup and the atomizer hood. “Secondary air,”
constituting the majority of the combustion air, enters the furnace through
an annular space between the atomizer hood and the burner casing. A
system of adjustable air guide vanes in this space imparts a tangential
component to the flow of secondary air as it enters the conical
combustion space between the burner and the main furnace chamber.
The resulting spiral path of the gases through the furnace not only
effectively keeps the flame central but at the same time prevents
unvaporized sulfur droplets from impinging on the furnace wall.

Figure 6. LURO rotary sulfur burner

a) Sulfur distributor; b) Atomizing cup; c) Atomizer

hood; d) Burner housing; e) Seal air; f ) Burner shaft; g)
Electric motor; h) Belt transmission; i) Sulfur inlet

13
A highly advantageous characteristic of rotary burners is that the
throughput can be varied continuously from the rated capacity down to
20 % capacity or less without appreciable change in atomizing intensity.
Rotary burners can be designed for sulfur throughputs ranging from 5 to
400 t/d. The highly turbulent atomizing effect ensures swift and safe
combustion of the sulfur, which makes it possible to design for extremely
high specific furnace loads (ca. 4 – 8×106 kJ m–3 h–1) and small specific
furnace volumes.

4.2. Heat Recovery

The heat released in the combustion furnace is utilized in a waste heat
boiler immediately downstream, generally producing high-pressure
steam. In the case of a sulfuric acid plant, gases of 10 – 12 vol % SO2
leave the furnace at ca. 1100 °C and are cooled to the converter inlet
temperature of ca. 420 – 440 °C. The waste heat boiler is coupled with
the waste heat system of the contact plant and, in conventional double-
absorption contact plants, it accounts for ca. 58 –60 % of the total energy
recovered as high-pressure steam. On the whole, ca. 3.8 – 4.0 t of steam
at 40 bar and 400 °C can be produced per tonne of liquid sulfur. If the
sulfur is supplied in solid form, about 0.5 t of saturated steam at 5 bar is
required per tonne of sulfur for melting, heating the sulfur filters, and
tracing the sulfur piping.

4.3. Impurities and their Removal

The 10 – 12 vol % SO2 combustion gases produced in a sulfuric acid
plant contain minor amounts of sulfur trioxide (3 – 6 g/m3) [25] and up to
20 ppm nitrogen oxides. Other impurities mainly depend on the quality of
the sulfur used.

Much of the Frasch and recovered sulfur commercially available

14
nowadays, which is supplied in Western Europe and in the United States
mainly in liquid form, is adequate for direct burning without requiring
additional cleaning, but various impurities do occur in sulfur, particularly
where it is handled and transported in solid form.

Bitumen and hydrocarbons, which occur especially in the so- called dark
sulfur grades, burn to carbon dioxide and water. The correspondingly
higher moisture content of the sulfur dioxide-containing contact gases
can lead to sulfuric acid mists, which are difficult to absorb. In sulfuric
acid plants running on dark sulfur, additional equipment, such as candle
filters downstream of the sulfur trioxide absorbers, may be necessary to
avoid corrosion and emission problems.

Any ash in the sulfur may lead to deposits in the combustion furnace and
waste heat boiler. However, the major part is entrained as fine dust in the
sulfur dioxide-containing gases and passes into subsequent parts of the
plant. In a sulfuric acid plant, it is mostly caught in the first catalyst bed.
This leads to a progressive increase in the gas pressure drop and may
also cause a loss in the sulfur dioxide conversion efficiency. Regular
screening of the catalyst may be necessary. Hot gas filters were formerly
placed in the gas line upstream of the converter to separate the dust. But
they are hardly used any more because they add to the pressure drop
and are only partially effective. Sedimentation tanks or settling pits
cannot be relied on alone to remove all the solid components from the
molten sulfur.

Filtration of the liquid sulfur [26] is the most effective means of separating
higher ash contents, which may occur especially in solid sulfur. The
sulfur is melted and heated by using low-pressure steam of about 5 bar
to a temperature of ca. 140 – 150 °C, at which its viscosity is at a
minimum. It is then pumped through vertical or horizontal filters of up to
ca. 150 m2 filtering area. Normally, these are of the pressure leaf type,
comprising a battery of stainless steel wire gauze elements suspended in
a containing vessel. Each element consists of two fine-mesh filter gauzes
suspended in a ring frame and supported and held apart by a third,
coarse mesh layer. The intermediate space between the two fine gauzes
of each element is connected to the outlet of the filter. In operation, the
outside of the filter elements is first coated with a filter aid such as
kieselguhr by passing it as a suspension into the filter vessel and
recirculating the effluent from the outlet back into the filter feed tank or
settling pit. Once prepared in this way, the filter is ready for service.
Periodically, as it becomes clogged with ash-forming material from the
sulfur, the filter is cleaned and recoated.

15
Apart from the additional operations and equipment required, sulfur
filtration poses a problem to the sulfur user of what to do with the filter
cake, which contains ca. 70 % sulfur. As a result, the responsibility for
sulfur filtration is increasingly being shifted to the sulfur producer.

4.4. Two-Stage Sulfur Combustion

For producing gases of high sulfur dioxide concentration and extremely
low nitrogen oxides (NOx) content, Lurgi has developed a double-stage
process with substoichiometric sulfur combustion [27]. As explained
earlier, the combustion temperature increases in proportion to the sulfur
dioxide concentration of the combustion gas (see Fig. 4). Above the
usual combustion temperature of 1000 °C, formation of nitrogen oxides
increases in proportion to the temperature, even though the free oxygen
content remaining in the gas is lower when the sulfur dioxide content is
higher if the combustion air is not enriched. Only above a concentration
of ca. 18 vol % SO2 is the free oxygen content low enough for thermal
NOx formation to decline drastically. Nitrogen oxides formation is also the
factor limiting the extent to which it is possible to preheat the combustion
air in ordinary sulfur burning systems, as this raises the combustion
temperature.

Nevertheless, formation of nitrogen oxides can be avoided, even at a

combustion temperature of 1750 °C, if the combustion of sulfur is divided
into two stages. In the first, the sulfur is burnt with a stoichiometric
deficiency of oxygen. It is particularly important in that stage to ensure
that the gases are extremely well mixed so that all the oxygen is
consumed, and nitrogen oxides cannot form. The combustion gas,
containing ca. 20 vol % SO2 and some unconverted sulfur vapor, is
cooled in a waste heat boiler to ca. 620 – 650 °C and then passed into
the second stage, an afterburning chamber (see Figure 7). There,
enough dried air is added to convert the remaining sulfur vapor
completely to sulfur dioxide, leading to a final gas concentration of ca.
18 vol % SO2. On account of the low concentration of sulfur vapor in the
second stage, the combustion temperature is too low for nitrogen oxides
to form. The gas mixture leaves the afterburning chamber at a
temperature not exceeding 700 °C and is cooled in another waste heat
boiler (usually designed as a steam superheater) before being passed on
to, for example, a sulfur dioxide liquefying or a bisulfite plant. This type
of system has been proved in years of satisfactory service in large-scale
plants.

Figure 7. Lurgi substoichiometric Figure 7. Lurgi substoichiometric

sulfur combustion process sulfur combustion process

16
a) Sulfur furnace; b) Waste heat a) Sulfur furnace; b) Waste heat
boiler; c) Afterburner boiler; c) Afterburner
[Full View]

The concentration of nitrogen oxides observed in the gases containing

18 vol % SO2 produced in this double-stage sulfur burning system are
actually lower than those of the 10 % SO2 gases originating from single-
stage sulfur burning plants
5. Production by Pyrite Roasting

5.1. Importance of Impurities

In the production of sulfur dioxide-containing gases by roasting of pyrites,
the constituents occurring besides sulfur and iron in the flotation
concentrates or fines exert a substantial influence on the selection of the
roasting process to be applied and its economics. Pyrite contains ca.
30 – 50 % sulfur, 26 – 46 % iron and up to 2.7 % copper, 3 % zinc, 1.4 %
lead, up to 10 % arsenic, and a number of other metals in small
quantities (see Table 1). Some of these impurities are entrained with the
sulfur dioxide in the roaster gases, and to remove them efficient gas
cleaning is necessary.

Impurities are also of prime concern in disposal or utilization of the iron

oxide residue of pyrite roasting, the so-called cinder. Each tonne of pyrite
concentrate yields about 0.7 t of cinder. High purity is a prerequisite
where the cinder is to be used for steelmaking. Cinders within the
following specification are generally considered acceptable [28]: Fe
65 %, Cu 0.1 %, Zn 0.1 %, Pb 0.02 %, As 0.06 %, S 0.5 %.

If a pyrite concentrate contains arsenic in larger quantities, special

roasting methods have to be used so as to obtain sufficiently low residual
arsenic values in the cinder (see Section Roasting of Arsenical Pyrite).

Nonferrous metals such as Cu and Zn, which are still present in flotation
pyrites, are again found in the cinder. At high nonferrous metal contents,
the cinder can be treated by chlorination and leaching.

17
5.2. Thermodynamics and Kinetics
Four main reaction steps (Eqs. 3 – 6) make up the overall pyrite roasting
reaction (represented by Eq. 7). The specified enthalpy values [7] refer to
298 K:

The endothermic decomposition of pyrites gives FeS or magnetopyrite,

liberating sulfur vapor (Eq. 3). If oxygen is available in adequate
quantities, the sulfur is immediately oxidized to sulfur dioxide (Eq. 4) and
simultaneously the sulfur content of the FeS is converted to sulfur
dioxide as well, in accordance with the equally exothermic reaction
(Eq. 5). At the temperatures prevailing in the roasting process, the iron(II)
oxide (FeO) formed as an intermediate phase is oxidized to magnetite
(Fe3O4) or hematite (Fe2O3), depending on the oxygen ratio (Eq. 6). The
overall pyrite roasting reaction (Eq. 7) is strongly exothermic, yielding ca.
13 000 kJ/kg pyrite sulfur. (The corresponding energy yield of burning
elemental sulfur is only ca. 9300 kJ/kg of sulfur). The roasting rate
depends on the temperature, grain size, and mineralogical composition
of the pyrites as well as the oxygen partial pressure.

Iron(II) oxide and sulfide form an eutectic mixture containing 56.2 % FeS
and 43.8 % FeO, which melts at 940 °C. It is not possible to increase the
reaction velocity by raising the temperature above this value because
any molten material deposits on the surface of the pyrite particles. This
renders unreacted material inaccessible to oxygen and thus inhibits
further reaction, particularly with larger particles.

The ignition temperature of pyrite lies between 350 °C and 550 °C. The
finer the grain size and the higher the nonferrous metal content are, the
nearer the ignition temperature is to the lower limit. A temperature of at
least 800 °C is required for complete reaction; the optimum range is

18
850 °C to 940 °C, depending on the grain size. A few pyrites decrepitate
during roasting. The fine particles thus formed influence the roasting
velocity and increase dust formation.

The residual sulfur content of the cinder depends not only on the amount
of sulfide remaining unconverted but also on sulfate sulfur. The latter
forms chiefly when the pyrite contains nonferrous metal impurities such
as copper, zinc, lead, cadmium, magnesium, and calcium, which remain
in the cinder as sulfates.

The maximum attainable sulfur dioxide content of the roaster gas

depends on the pyrite composition as well as the atmosphere in which it
is burnt. In air it is ca. 16 vol % in the most favorable case, although with
the slight excess of air usually used values of 13 – 14 vol % are
common. In oxygen or oxygen-enriched air, higher sulfur dioxide
concentrations can be attained. In spite of the thermal and process
advantages that this would have with respect to sulfur dioxide and
sulfuric acid production, oxygen is rarely used for pyrite roasting because
of the high costs.

5.3. Roasting Processes

Various types of roasting equipment have been used in the past for
pyrite roasting, including shaft furnaces, multiple hearth roasters, rotary
kilns, and dust roasters [29], [30]. Nowadays fluid-bed roasters are
preferred, as they are much superior to the other types in terms of
process technology, throughput rates, and economy. As a further step,
circulating fluid bed technology was introduced for roasting special ores
that containing economically viable quantities of gold.

Basically a fluid-bed roaster is in the form of two upright cylinders, the

upper being of a larger diameter than the lower, to which it is joined by a
frustoconical section. The steel shell is lined with insulating and
refractory bricks. At the lower end is a horizontal bar grate or a nozzle
grate, below which is a conical wind box, at the apex of which the
roasting air is introduced. In this way the roasting air is uniformly
distributed over the entire cross section of the lower portion of the
roaster. Above the grate, a bed of cinder is maintained in a fluidized state
by the roasting air. The pyrite is fed into the roaster directly above or in
the uppermost zone of the fluid bed. The depth of the fluid bed is kept
constant by an overflow. In the wider upper section of the roaster, the
velocity of the gas leaving the fluid bed is reduced. This serves to
stabilize the top zone of the fluid bed and reduces entrainment of solids
in the gas flow. This section also acts as an afterburning chamber. The
roaster gases pass from the top of the roaster to a waste heat boiler

19
directly downstream.

Fluid-bed technology for pyrite roasting was first introduced at the

beginning of the 1950s almost simultaneously by Dorr Oliver in the
United States and BASF in the Federal Republic of Germany.

Dorr Oliver FluoSolids Reactor. In its original concept, the FluoSolids

reactor was designed on the wet feeding principle [31], [32]. Finely
ground ore or flotation pyrites are mixed with water to form a viscous
pyrite sludge with a 70 – 80 % solids content. This is forced continuously
into the roaster by means of a pump. The height of the cinder bed is up
to 1 m in the unfluidized state and 1.6 m when fluidized at a gas velocity
in the bed of 0.2 – 0.6 m/s. The specific daily throughput is around
4 – 5 t/m2 of roaster area. The bed temperature is adjusted to
600 – 900 °C by adiabatic evaporation of the water introduced with the
pyrite and (where appropriate) of additional water injected separately. A
later version of the FluoSolids roaster was designed for dry feeding,
using indirect cooling and heat recovery [33].

The BASF – Lurgi fluid bed furnace roaster was developed [34-36] for
dry pyrite feeding and designed to recover the maximum amount of heat
for steam production. In this roaster the temperature of the fluid bed is
kept constant by indirect cooling; the surplus heat of reaction is removed
in the fluid bed by immersed cooling elements which form part of a waste
heat system for the production of high-pressure steam. Figure 8 shows a
typical design of a modern fluid bed furnace with waste heat system
according to the Lurgi principle.

Figure 8. Fluid bed roaster with waste Figure 8. Fluid bed roaster with
heat system waste heat system

a) Fluid bed furnace; b) Waste heat a) Fluid bed furnace; b) Waste

boiler; c) Pre-evaporator; d) heat boiler; c) Pre-evaporator; d)
Superheater; e) Evaporator; f ) Steam Superheater; e) Evaporator; f )
drum Steam drum
[Full View]

20
The grain size of the pyrites to be processed may be up to 10 mm,
depending on the stability of the particle structure during the roasting
process. Usually the particles have diameters of up to 6 mm. The
coarser the grain is, the higher the gas velocity must be to maintain the
bed in a fluidized state. Even flotation pyrites with grain sizes < 0.2 mm
can be processed in the fluid bed roaster, but in that case the moisture
content is important. Although the pyrites must be dry enough to be fed
in the solid state, enough moisture must be present to ensure that the
particles are sufficiently agglomerated to maintain a stable fluid bed.
Otherwise the particles will be too light and will be largely blown out of
the reaction in the gas flow.

To ensure uniform, controlled feeding of the pyrites, table feeders or

weight- controlled vari-able-speed conveyors are used. The gas velocity
in the fluid bed is ca. 0.8 – 2.5 m/s at 1 bar and 850 °C. In the top part of
the roaster the velocity is ca. 0.5 – 1.8 m/s. In the unfluidized state, the
bed has a rest depth of 700 – 800 mm.

In contrast with all the earlier roasting systems, the material distribution
in a fluid bed system is very rapid, and both the composition of the bed
and its temperature are very uniform. Therefore, it is possible to select
and maintain the roasting process conditions within a narrow optimum
range. On account of the good heat transfer characteristics (800 –
1200 kJ m–2 h–1 K–1) the temperature can be controlled satisfactorily by
indirect cooling using immersed, water- cooled elements with a relatively
small heat exchange surface area.

The bed temperature is normally maintained in the range 650 – 900 °C.
In the gas space above the fluid bed, the temperature may be somewhat
higher (up to 950 °C) on account of the afterburning effect when flotation
pyrites are being processed. Above the fluid bed the pressure is kept
21
slightly below atmospheric pressure to avoid leakage of sulfur dioxide to
the outside. The roaster gases are drawn off from the furnace by the
main gas fan of the subsequent contact plant or by an intermediate gas
fan in the gas cleaning system.

Abrasion under the turbulent bed conditions and decrepitation of the

pyrite particles create a substantial amount of fines, so that a sizeable
amount of the solids leave the bed in the roaster gas in the form of dust.
This dust is precipitated in the waste heat boiler and also in subsequent
cyclones and hot gas electrostatic precipitators. When fine pyrites are
roasted, the portion of cinder leaving the roaster as entrained dust may
be around 50 % and in the case of flotation pyrite roasting it may even be
as high as 90 % or more.

Being predominantly iron oxide, the dust has catalytic activity for the
oxidation of sulfur dioxide to sulfur trioxide. This reaction is highly
undesirable at this stage because any sulfur trioxide formed reacts with
moisture in the gas to form sulfuric acid, which condenses in cool parts
of the plant and causes corrosion. In the roaster itself the oxidation of
sulfur dioxide is inhibited by the high prevailing temperature and by
controlling the air supply to minimize the residual oxygen content in the
roaster gas. Having a low oxygen content is of even greater importance
in the waste heat boiler, where the gas is cooled to temperatures at
which sulfur trioxide formation is thermodynamically more favored. It is
also of importance to cool the gases here as quickly as possible to
temperatures at which the reaction is kinetically inhibited. Nevertheless,
some sulfur trioxide is inevitably formed, and since neither the pyrites nor
the roasting air are dried, the roaster gas temperature at the outlet of the
hot gas electrostatic precipitator must be maintained above the dewpoint
of sulfuric acid. This limits the amount of heat that can be recovered from
the gases in the waste heat boiler. Generally, the roaster gases are
cooled to about 350 °C before they enter the dust separation cyclone.

Figure 9 shows a typical energy flow diagram for the roasting of flotation
pyrites containing 47 – 48 % sulfur in a Lurgi fluid bed roaster system.
Approximately 70 % of the total energy is recovered from the fluid bed
and roaster gas in the form of high-pressure steam. This corresponds to
the production of about 1.5 t of steam of 40 bar and 400 °C per tonne of
48 % S pyrite from feed water at a temperature of about 104 °C. The
remaining 30 % of the total energy is represented by the heat content of
the roaster gases at an exit temperature of 350 °C, the heat content of
the cinder, and radiation losses. The power consumption amounts to a
25 – 30 kWh/t pyrites. Most of it is used in the roasting air fan. The
sulfide sulfur content of the cinder from pyrites originally containing
22
47 – 48 % sulfur is about 0.5 %. The sulfur dioxide content of the roaster
gases at the furnace outlet is ca. 14 – 14.5 vol %. The use of an
automatic control system to maintain the oxygen content of the roaster
gas at the furnace outlet at a minimum substantially reduces sulfur
trioxide formation in the waste heat boiler and with it the rate of scrubber
acid accumulation in the wet gas cleaning system.

Figure 9. Energy flow diagram for fluid

bed pyrite roasting

The specific capacity of Lurgi fluid bed roasters corresponds to a daily

throughput of about 20 t/m2 of roaster area. Today, fluid bed roasters are
designed for pyrites throughputs of up to about 1000 t/d. Their turndown
limit is around 70 – 80 %. For large plant capacities, in most cases two
furnaces are connected in parallel to provide better flexibility.

During start-up, the fluid bed made up of cinder is preheated with oil
burners. As soon as the bed temperature reaches ca. 600 °C, pyrites can
be fed in and the oil burners are turned off.

On account of the high thermal capacity of the cinder bed, the roaster
can be restarted without requiring additional heating after standstills of
up to about 24 h.

Circulating Fluid Bed Roasting [37]. A further step towards more efficient
roasting — especially of gold-containing ores — was the introduction of
the circulating fluid bed process.

The circulating fluid bed (CFB) technology is known from the calcination
23
of hydrated alumina or clay and combustion of coal, and was adapted for
roasting of gold-bearing minerals. After extensive pilot plant tests the first
industrial scale plant was built by Lurgi and started up in 1989 in
Australia.

The aim of roasting gold-containing minerals is to obtain a maximum

gold recovery from the calcine in subsequentleaching steps. Such
conditions can only be achieved within a narrow range of roasting
temperatures and oxygen partial pressures in the roaster gas
atmosphere. The CFB process is capable of providing such conditions.

Contrary to a “normal” stationary fluid bed, circulating fluid bed roasters

are operated at higher gas velocities in the range of 3 – 6 m/s. Because
of the velocity the fine solids (calcine) are entrained with the gas and in
part recycled to the roaster via a cyclone.

The main features of the CFB process are:

1. Treatment of a wide range of minerals
2. High heat and mass transfer rates
3. Uniform temperature profile over the whole roaster/cyclone system
4. Accurate temperature control by cooling surfaces, direct water
injection, or direct fuel injection
5. Stable oxygen partial pressure all over the system
6. Turndown ratio of 50 %

Ores with sufficient sulfur and iron content can be roasted autothermally;
ores with low sulfur and iron contents can be treated with additional
heating; and even the most sensitive refractory gold ores can be roasted
successfully in the CFB. Air or oxygen-enriched air can be employed.
The high heat and mass transfer rates guarantee almost complete
combustion of sulfur and carbon. The uniformity of the temperature
profile avoids the partial overheating of particles known from stationary
fluid beds. Overheating of particles may cause substantial drop in gold
recovery. The temperature is uniform in the whole system due to the high
amount of circulated material and the effect of cooling (or heating) is
transferred to the total inventory of the system. Because of the high
velocities and mass transfer rates an exact and stable control of the
oxygen partial pressure is possible to obtain optimum roasting results. A
turndown ratio of up to 50 % offers a high operational flexibility compared
to normal fluid bed roasting.

Key components of CFB plants are the roaster and the recycling cyclone.

24
The roaster consists of a cylindrical shell with insulating and refractory
bricks. At the lower end a nozzle grate is installed through which the
roasting air is introduced. At the upper end the recycling cyclone is
connected to the roaster. The ore is fed to the CFB roaster and roasted
with air or oxygen-enriched air. A portion of the hot calcine which is
separated in the cyclone is recycled to the roaster, a second portion is
transferred to a fluid bed cooler where part of the heat contained in the
calcine is removed. The hot gas leaving the cyclone is cooled and
treated in downstream plant elements.

The associated equipment must be selected according to the specific

requirements of each plant; some of such elements are listed below:
1. Dry feed or slurry feed
2. Preheating and drying of feed with recovered heat from the calcine
3. Preheating of roasting air with recovered heat from the calcine
4. Direct injection of additional fuel (if required)
5. Cooling of the reaction material by integrated cooling elements, direct
water injection or recirculation of cooled calcine
6. Cooling of the calcine in a fluid bed cooler, or direct quenching
7. Heat recovery from roaster gas in a waste heat boiler or gas – gas
heat exchangers
8. Further treatment of the off-gas (bag house, electrostatic precipitator,
acid plant, . . .)

A flowsheet showing the key equipment and some alternatives for

associated equipment is given in Figure 10.

Figure 10. CFB-roasting of gold-bearing ores

EP = electrostatic precipitator

25
Roasting in a Cyclone Furnace. Pilot-plant trials have been carried out in
Poland on a method of roasting pyrites in a cyclone furnace, which is
claimed to facilitate the subsequent recovery of nonferrous metals [38].
Flotation pyrites, preheated air, and (as necessary) fuel are introduced
into the upper section of a cyclone. The operating temperature is such
that the burnt particles of pyrite cinder melt, coalesce, and run down the
walls of the cyclone. The melt is drained into a quench bath, where it
forms granules. The gas leaving the cyclone contains entrained iron
oxide droplets and part of the original nonferrous metals content of the
pyrites, especially zinc and lead, both as particles and vapors of oxides
and metals. It is cooled in a fluid bed cooler with immersed cooling coils,
which recover the heat partly as high-pressure steam and partly by
preheating the furnace air. The entrained material and vapors of
nonferrous metals are largely arrested by agglomeration in the fluid bed.
The material overflowing from this bed is stated to contain a substantially
higher concentration of nonferrous metal values than the cinder from
other types of roasting furnace.

5.4. Roasting of Arsenical Pyrite

Arsenic occurs in pyrites in the form of arsenical pyrite (FeAsS;
FeAs2 · FeS2). Under the usual roasting conditions, with a slight excess
of oxygen in relation to the stoichiometric requirement for converting the
26
iron content of the pyrite to hematite, the majority of the arsenic is
oxidized to As(V) and will react with the hematite to form stable iron
arsenate (FeAsO4). In this way about 70 – 80 % of the arsenic content in
the pyrite feed becomes bound to the cinder. The remaining arsenic,
which is oxidized to As(III) as shown in Equation (8), is volatilized and
entrained by the roaster gas and must be separated in the wet gas
cleaning system

(8)

If the cinder is to be used in steelmaking, it must not contain too much

arsenic. In such cases high-arsenic pyrites can be processed in a two-
stage fluid-bed roasting system. In the first stage, operating under
oxygen-deficient conditions, neither hematite nor iron arsenate is formed.
The arsenic is thus volatilized almost completely and discharged with the
roaster gases. In the second fluid bed stage, the separated cinder is
oxidized to hematite. This principle is applied on an industrial scale in the
Boliden roasting process and the BASF two-stage process.

Both the Boliden and BASF processes are capable of diminishing the
residual arsenic content of the cinder to less than 0.05 %.

Boliden Process. The Boliden process [28] was specifically developed for
processing arsenical flotation pyrites. In the fluid bed roaster, the
roasting air/pyrites ratio is adjusted so that magnetite forms at ca.
850 °C. Most of this is blown out of the furnace in the roaster gas. It is
separated at an efficiency of ca. 90 % in a hot-gas cyclone immediately
following the furnace. The gas leaving the system contains, besides
residual dust, the arsenic and ca. 10 g/m3 of sulfur vapor. The sulfur
vapor is burnt in a special combustion chamber at the inlet of the waste
heat boiler with a controlled amount of additional air. More air is added at
the waste heat boiler outlet to complete the combustion of residual traces
of sulfur before the gases pass on to the hot gas electrostatic
precipitator. The cinder separated from the gas in the hot gas cyclone is
treated in a second fluid bed roaster with a controlled volume of air. In
the process, the residual sulfur level is diminished substantially, and the
magnetite is partly converted to hematite (Eq. 9)

(9)

27
As in the first roaster, the surplus heat of reaction is recovered by
generating steam in cooling coils immersed in the fluid bed. The
discharged cinder is cooled to ca. 200 °C in a separate fluid bed cooler,
which is connected into the steam system as a preheater. Up to ca. 1.6 t
of high-pressure steam at approximately 425 °C are produced per tonne
of pyrites.

The arsenic is removed from the dedusted roaster gases in the wet gas
cleaning system (see Section Cleaning Systems). Plants built under
license by Lurgi processing up to 500 t/d pyrites have been operated
successfully in Sweden, Spain, and Japan.

BASF Process. The BASF two-stage process [39] is especially suited for
processing relatively coarse pyrites. Partial roasting in a first fluid bed
unit at 900 °C under oxygen-deficient conditions removes only the
loosely bound polysulfide sulfur of the pyrite, giving FeS.

(10)

Arsenic, antimony, and lead are carried out of the system with the
roaster gases, mainly as volatile sulfides or oxides. Coarse ferrous
sulfide cinder withdrawn directly from the bed overflow and hot dust
separated in the high-efficiency cyclone are passed on to a second fluid
bed unit. The iron(II) sulfide, which is virtually free from arsenic, is
roasted at about 800 °C with excess oxygen to obtain hematite.

The roaster gas from the first stage leaves the hot gas cyclone with
about 18 vol % SO2. It is first passed with additional air through an
afterburning chamber to oxidize any remaining sulfur vapors and residual
sulfides. The gases are then cooled in the waste heat boiler. The dust
separated there and in the hot gas electrostatic precipitator immediately
downstream is rich in arsenic, antimony, and lead and can be treated
further to recover them. But the main part of the arsenic remains in the
gases as volatile arsenic trioxide and is separated in the wet gas
cleaning system.

The gases originating from the second roaster stage, with about 10 vol %
SO2, are cooled in another waste heat boiler and dedusted in a cyclone
followed by a hot gas electrostatic precipitator. The gas streams from the
28
two stages are subsequently united (upstream or downstream of the wet
gas cleaning system).

In the low-arsenic cinder and dust coming from the second stage, the
iron is present as hematite (Fe2O3).

5.5. Production in Nonferrous Metal Sulfide Ore Processing

5.5.1. Copper Smelting ( Copper – Pyrometallurgical Principles,
 Copper – Traditional Bath Smelting,  Copper – Autogenous Smelting,
 Copper – Discontinuous Matte Conversion,  Copper – Continuous
Matte Conversion)
Copper ore concentrates consist mainly of copper pyrites or chalcopyrite
(CuFeS2), the most frequent copper mineral, but may also contain pyrite,
bornite, cuprite, chalcocite and others. A typical concentrate composition
is 27 % Cu, 28 % Fe, 32 % S. They are processed almost exclusively by
pyrometallurgical methods. Pyrometallurgical copper extraction is based
on the decomposition of complex iron – copper sulfide minerals into iron
and copper sulfides, followed by selective oxidation and separation of the
iron portion, and finally oxidation of the remaining copper sulfide.
Virtually all the sulfur in the original concentrate ends up as sulfur
dioxide. The overall process is conducted in the three discrete stages of
roasting, smelting and conversion.

In the roasting stage, the higher sulfide components of the so- called
green concentrate are initially decomposed thermally to iron(II) and
copper(I) sulfides and elemental sulfur (Eq. 11) [40], [41]. In the
prevailing oxidizing atmosphere, this sulfur is instantly converted to sulfur
dioxide

(11)

The iron(III) sulfide also undergoes exothermic oxidation. To attain a

good separation of iron and copper in the subsequent smelting stage one
must ensure that the copper does not become converted to oxide.
Therefore, the amount of oxygen is limited so that some of the iron(II)
sulfide remains unchanged. Under such conditions magnetite is the
predominant oxide (Eq. 12) rather than hematite (cf. Eq. 7)
29
 (12)

The roasted ore, alone or with green concentrate (depending on the

concentrate composition and the type of process used), then passes to
the smelting stage. Here, in the presence of a siliceous flux, it is heated
by the combustion of a fuel, by electricity, or autogenously by oxidation
with oxygen to a temperature at which the unconverted
sulfides — mainly copper(II) sulfide — are molten. The iron oxide
component reacts with the siliceous flux to form a molten slag. Being
appreciably less dense, this floats on top of the heavier sulfide phase,
known as copper matte. The two can thus be tapped separately.

In the conversion or blowing stage, air or oxygen is blown through the

matte in two steps. In the first step (slag blowing), the residual iron
sulfide portion is converted to iron(II) oxide, which is removed as a slag
from the concentrated matte of copper(I) sulfide. In the second step
(copper blowing), the copper(I) sulfide is converted to crude copper
(blister copper):

(13)

The development of the copper recovery processes during the last few
decades has been dominated by two principal objectives. One is the
need to economize on energy by making maximum use of reaction heat
liberated in the individual process steps. The other is the need to lower
the volume and increase the sulfur dioxide concentration of the waste
gases as much as possible so as to facilitate environmental pollution
control. The trend is increasingly in favor of oxygen technology.

Off-Gas Composition from Specific Process Types. The reverberatory

furnace process is still used to a considerable extent for smelting copper
concentrates, although it suffers from the economic disadvantage of high
energy consumption. The design and mode of operation of reverberatory
furnaces is described in [42]. Partially preroasted concentrates and
green concentrates of low sulfur content are the most frequent feed
materials. Copper matte is produced with a copper content about 10 %
higher than that of the feed material. The most serious technical
disadvantage of this process is the large volume of combustion gas that
30
results from the high consumption of fossil fuels. This dilutes the sulfur
dioxide concentration of the total smelter off-gas to ca. 1 – 2 vol %. It is
economically scarcely feasible to process this into sulfuric acid.

A comparison of the various pyrometallurgical copper extraction

processes from the point of view of the waste gases produced, based on
calculations of SCHWARTZ and EISERT [43], emphasizes the problems
inherent in the reverberatory furnace process.

By using oxygen-enriched air in oxyfuel burners, the sulfur dioxide

content of reverberatory furnace waste gases can be raised to about
6 – 8 vol % and the specific waste gas rate is considerably diminished.

The electric furnace process can be used to produce copper matte when
electric power is available at low cost. The connected load of such
furnaces used in copper metallurgy is up to 50 MW. The process is in
many respects similar to that of the reverberatory furnace except that the
specific waste gas rate can be kept relatively low if the furnace is
adequately sealed. The waste gas composition depends on the feed
material characteristics. The sulfur dioxide content may reach
3 – 10 vol %. Carbon dioxide and carbon monoxide also occur in small
concentrations in the waste gas owing to burning of the carbon
electrodes.

The Outokumpu flash smelting process [44] is one of the standard

pyrometallurgical processes used most frequently today. The flash
smelting furnace (Fig. 11) consists of an upright cylindrical reaction shaft
designed for simultaneous roasting and smelting atop one end of a
hearth, over the other end of which is a vertical waste gas shaft. Fine-
grained dry concentrate is injected into the reaction shaft with air
preheated to 450 –750 °C or with oxygen-enriched air.

Figure 11. Outokumpu flash smelting Figure 11. Outokumpu flash

furnace smelting furnace

a) Concentrate burner; b) Reaction a) Concentrate burner; b) Reaction

shaft; c) Settling trough; d) Waste gas shaft; c) Settling trough; d) Waste
shaft gas shaft
[Full View]

31
Since the process is not usually autogenous, supplementary heating is
necessary, not only in the hearth but also in the reaction shaft, except
when oxygen-enriched air is used. In that case, the addition of fossil
fuels to the reaction shaft can be dispensed with. The droplets forming in
the reaction shaft, consisting of copper matte and slag, accumulate in the
hearth, while the gases leave the flash smelting furnace through the
waste gas shaft at 1200 – 1300 °C.

Depending on the concentrate composition and the copper concentration

of the matte obtained in the flash smelting furnace, about 50 – 70 % of
the sulfur content of the concentrate is converted to sulfur dioxide. When
operating the furnace without oxygen enrichment, the waste gases
contain 10 – 13 vol % of sulfur dioxide, 1 vol % of oxygen, and ca.
0.05 vol % of sulfur trioxide at specific off-gas rates of about
6000 – 6500 m3/t copper.

Most flash smelting furnaces are operated with reaction air enriched to
45 – 60 % O2, which substantially reduces the off-gas rate and increases
the SO2 concentration. The relationships shown below are obtained
when feeding a concentrate containing 21.9 % Cu, 30.3 % Fe and
32.0 % S.

O2 content of
feed air, vol % 21.0 28.0 31.9 50.0
Inlet temperature of
feed air, °C 450 200 25 25
Specific off-gas volume,
m3/t Cu 6016 4593 4087 2363
Off-gas SO2 content,
32
 vol % 13.4 17.8 19.8 28.5

Oxygen has already been used continuously on an industrial scale in the

Inco flash smelting process. Inco of Canada developed a flash smelting
furnace with a daily throughput of up to 1500 t of concentrate. After
pretreatment in a fluid bed dryer, the concentrates are injected into the
reaction space with 95 % oxygen. The process operates without
supplementary heating. The specific off-gas rate is as low as 1090 m3/t
Cu at about 80 vol % SO2. After adequate gas cleaning, the sulfur
dioxide- containing gases are chiefly converted to liquid sulfur dioxide.

The flame cyclone processes for intensive smelting with oxygen

represent another practical option. On an industrial scale, a combination
of a flame cyclone reactor with an electric resistance furnace has been
used in the Kivcet process developed at the Irtysh metal combine at
Glubokoye in the former Soviet Union [45]. The feed material is copper
concentrate or complex zinc – lead – copper concentrate, which must be
of fine grain size and pourable. Tonnage oxygen is used for the smelting
process. The off-gases leave the furnace at about 1250 – 1350 °C. They
contain 80 – 85 vol % SO2. The specific off-gas rate ranges from 800 to
2000 m3/t Cu, depending on the concentrate composition.

Composition of Converter Off-Gases. Converter processes are generally

used for blowing air or oxygen-enriched air through copper matte to
convert it to crude copper. The blowing procedure is not continuous, nor
are, therefore, the composition and volume of the off-gas. During slag
blowing, the sulfur dioxide content of the waste gases at the converter
mouth is up to ca. 10 vol %, but during copper blowing it may rise to as
much as 20 vol %.

If oxygen-enriched air is used, correspondingly higher sulfur dioxide

concentrations are achieved, but precautions must be taken due to the
increased generation of heat. However, it is not only the continually
fluctuating sulfur dioxide concentrations which are typical of converter
processes but also the heavy fluctuations in the gas flow rates.

Because air leaks into the exhaust hoods of the converters, the converter
off-gases are diluted considerably. The dilution factors are between ca.
1.8 and ca. 5 [46]. In the Hoboken syphon converter (Fig. 12), operated
at Metallurgie Hoboken Overpelt in Belgium, air inleakage is effectively
avoided [47].

33
Figure 12. Hoboken syphon Figure 12. Hoboken syphon
converter converter

[Full View]

Developments to Overcome Off-Gas Processing Problems. All sulfur

dioxide-containing gas streams from copper smelting operations,
including converter off-gases, must be processed to prevent more than
traces of sulfur dioxide entering the atmosphere. The key to successful
purification of sulfur dioxide-containing gases to the required standards is
continuous stable operation of the purification process, be it a sulfuric
acid plant (as is most common at smelters) or some other kind of
abatement process. Gas streams which are liable to wide fluctuations in
both volume and composition, such as converter off-gases, are quite
incompatible with this criterion. These fluctuations can be mitigrated to
some extent by conducting the conversion procedure in several
converters in a phased sequential operation and combining their off-
gases [48]. It is also common practice to mix them with the steadier,
more concentrated gas streams from the roasting and smelting stages to
produce a combined gas stream within the concentration range needed
to maintain autothermal operation of a sulfuric acid plant [49]. As an
example, the routing of off-gases from a modern copper smelter
incorporating an Outokumpu flash smelting furnace and Pierce Smith
[50] converter operation is schematically represented in Figure 13.

Figure 13. Sulfur dioxide production Figure 13. Sulfur dioxide

and processing in a modern copper production and processing in a
smelter modern copper smelter

a) Smelting furnace; b) Waste heat a) Smelting furnace; b) Waste

boiler; c) Hot electrostatic precipitator; heat boiler; c) Hot electrostatic
d) Pierce Smith converter; e) Wet gas precipitator; d) Pierce Smith
converter; e) Wet gas cleaning
34
cleaning system; f ) Sulfuric acid plant system; f ) Sulfuric acid plant
[Full View]

The chief aim of current development work is to unite the individual

stages of crude copper extraction into one continuous process, avoiding
any discontinuous converter stages. This has been achieved on an
industrial scale by Mitsubishi. In the Mitsubishi process, the individual
process steps are carried out in separate furnaces which are connected
with each other via launders to form a continuously operating system
[51].

5.5.2. Nickel Smelting ( Nickel – Pyrometallurgy of Nickel

Concentrates )
The mineral used most commonly for nickel production is pentlandite
(Ni, Fe)9S8, which often occurs in conjunction with pyrrhotite (Fe7S8) and
chalcopyrite (CuFeS2). Nickel flotation concentrates contain 6 – 12 % Ni,
2 – 10 % Cu, 25 –35 % Fe and 25 – 32 % S.

As in copper metallurgy, the basic operations of nickel extraction are

roasting, smelting, and blowing [52].

During concentrate roasting, initially pyrrhotite is oxidized (Eq. 14):

(14)

Roasting can be performed in sinter machines, multiple-hearth furnaces

or, in modern plants, in fluid bed furnaces.

At Inco, at Copper Cliff, Ontario, Canada, ca. 1350 t/d of nickel

concentrate are roasted in a fluid bed furnace (diameter 5.5 m, height
7.9 m). About 40 000 m3/h of roasting air is injected through 566 nozzles;
35
the fluid bed has a height of 1.2 m. Roasting proceeds at 650 °C. The
roaster gases contain ca. 12 vol % SO2.

The Outokumpu flash smelting furnace yields an off-gas containing

10 – 12 vol % SO2 when the roasting air is preheated to 500 °C. Oxygen
enrichment of the reaction air to 27 – 35 vol % O2 allows throughput to be
increased from 400 to 700 t/d of nickel concentrate and simultaneously
raises the sulfur dioxide content in the furnace off-gas to 17 – 18 vol %
[44].

A smelting furnace operated by Inco with tonnage oxygen achieves

throughputs of 1500 t/d of nickel concentrate. From the furnace off-
gases, containing about 80 vol % SO2, liquid sulfur dioxide is recovered.
The annual liquid sulfur dioxide production amounted to ca. 110 000 t in
1970 [53].

The smelting and blowing operations of most process steps of nickel

smelting are technically comparable to those of copper metallurgy.

5.5.3. Zinc Blende Roasting ( Zinc)

Zinc blende (ZnS) is the most important primary material for zinc
production. Zinc blende is enriched by flotation to obtain a concentrate
with a typical composition of 55 % Zn, 8 % Fe, 32 % S.

The first step in all zinc recovery processes is the roasting of zinc blende
to convert zinc sulfide to zinc oxide, simultaneously removing the bound
sulfur [54], [55]. In the exothermic roasting reaction, the sulfur is liberated
as sulfur dioxide.

(15)

To produce a calcine which can easily be leached for electrolytic zinc

recovery, which accounts for about 85 % of world production, especially
severe demands are made on the roasting process. The calcine should
contain not more than 0.3 % sulfide sulfur and 1.8 % sulfate sulfur. The
formation of zinc ferrite (ZnO · Fe2O3), which would disturb the leaching
process, must be avoided as far as possible by maintaining constant,
optimum roasting temperatures. At the same time, the heat released in
the roasting process should be utilized as efficiently as possible in steam
generation. Also, to permit economic processing of the sulfur dioxide
roaster gases, the sulfur dioxide concentration should be kept as high

36
and steady as possible. These requirements can be fulfilled with the aid
of fluid bed technology originally developed for the roasting of pyrite.

Vieille Montagne – Lurgi Process. Figure 14 shows the layout of a fluid

bed roasting plant for zinc concentrate according to the process of Vieille
Montagne and Lurgi, which is nowadays used worldwide [56]. The
concentrate, containing ca. 8 % water, is continuously fed into the fluid
bed furnace by slinger conveyors. Roasting air is injected through a
nozzle grate. The gas velocity in the fluid bed is ca. 0.5 m/s. Cooling
elements incorporated in the fluid bed maintain the roasting temperature
between 900 °C and 1000 °C. The dissipated heat is recovered in steam
generation.

Figure 14. Vieille-Montagne – Lurgi zinc concentrate fluid Figure 14.

bed roasting plant Vieille-
Montagne – L
a) Slinger conveyor; b) Fluid bed roaster; c) Blower; d) urgi zinc
Waste heat boiler; e) Cyclone; f ) Fluid bed cooler; g) Ball concentrate
mill; h) Hot-gas electrostatic precipitator fluid bed
roasting plant

a) Slinger
conveyor; b)
Fluid bed
roaster; c)
Blower; d)
Waste heat
boiler; e)
Cyclone; f )
Fluid bed
cooler; g) Ball
mill; h) Hot-
gas
electrostatic
precipitator
[Full View]

Depending on the composition of the zinc concentrate, the roaster gas at

the furnace outlet contains 9 – 11 vol % SO2 and 5 – 6 vol % O2. A
characteristic feature of fluid bed furnaces is that gas flow rates and gas
compositions are relatively constant, which is advantageous from the
point of view of subsequent processing. In a downstream waste heat
boiler, the roaster gases are cooled to ca. 350 °C and then passed on to
37
the hot gas electrostatic precipitators. About 1 t of high-pressure steam
at 400 °C is produced per tonne of concentrate processed.

All other cooling and cleaning operations are as practised in conventional

sulfur dioxide gas cleaning systems (see Chap. Cleaning of Crude Sulfur
Dioxide Gas Streams).

The zinc oxide calcine is recovered at various points of the roasting plant
in the following approximate proportions:

Fluid bed discharge 20 – 30 %

Waste heat boiler discharge 30 – 40 %
Cyclone discharge 22 – 48 %
Electrostatic precipitator discharge 2 – 5 %

The precise distribution is influenced by the grain size range of the

concentrate and by controlled wetting and agglomeration of the
concentrate before it is fed onto the slinger belts.

Some characteristic data for the largest zinc blende fluid bed furnace
based on the Vieille Montagne – Lurgi process, which has been in
operation since 1975 at Electrolytic Zinc in Risdon, Australia [57] are
presented below:

Grate area 123 m2

Furnace volume 2680 m3
Concentrate throughput (dry) 800 t/d
Roasting air rate 70 000 m3/h
Roasting temperature 920 ± 20 °C
Roaster off-gas (dry) 67 000 m3/h
Roaster gas SO2 content 10 – 11 vol %
Steam generation 1 t/t of concentrate
Steam grade 40 bar/400 °C
Sulfide sulfur in calcine 0.15 – 0.20 %
Sulfate sulfur in calcine 1.4 – 1.6 %

Sanki – Dorr Process. In the Sanki – Dorr fluid bed process, the
concentrate is fed to the roaster as an aqueous suspension containing
ca. 78 % solids. The fluid bed temperatures are controlled by injecting
additional water. This wet feeding method yields substantially less steam
than dry feed processes (ca. 0.68 t per tonne of concentrate). It also
38
produces about ten times as much dilute scrubber acid, which must be
disposed of [58].

Characteristic data for a Sanki – Dorr plant operating at Mitsui Mining

and Smelting, Hiroshima, Japan, are listed below:

Grate area 66 m2
Furnace height 10.4 m
Concentrate throughput (dry) 250 t/d
Roasting air 18 960 m3/h
Roasting temperature 930 °C
Concentrate contents
Zn 50.6 %
Fe 9.6 %
Pb 1.04 %
Cu 0.42 %
Cd 0.28 %
S 30.7 %
Sulfide sulfur in calcine 0.22 %
Sulfate sulfur in calcine 1.63 %

The roasting air is injected into the furnace through 1025 nozzles and the
fluid bed is about 1.2 – 1.5 m high.

Oxygen Enrichment. The effect on the throughput of adding oxygen in

zinc blende fluid bed roasters has been tested, and several plants are
now being operated successfully with oxygen-enriched air. The
conditions in the phase diagram of the system Zn – Fe – S – O [59]
suggest that there is a limit on the extent to which it is practical to enrich
the roasting air with oxygen on account of undesirable sulfate and ferrite
formation. According to information from Electrolytic Zinc Co., which has
operated a 360 t/d nominal capacity fluid bed furnace with oxygen-
enriched air, the optimum was at ca. 24 vol % O2. By this means it was
possible to raise the throughput by ca. 25 % and the sulfur dioxide
content of the roaster gases to 12 – 12.5 vol %.

5.5.4. Lead Smelting ( Lead – Smelting Lead – Roast Reaction

Processes Lead – Direct Smelting Reduction Processes)
Lead sulfide ore flotation concentrates, containing ca. 60 % Pb and 17 %
S, are roasted almost exclusively in travelling-grate sinter plants.

The oxidation of the lead sulfide is initiated by chemisorption of oxygen

39
on the particle surface; the lead sulfate formed then reacts with the
remaining lead sulfide and oxygen to form basic sulfates and, finally,
litharge (PbO), releasing sulfur dioxide [60], [61]

(16)

(17)

(18)

(19)

Metallic lead can be formed in side reactions

(20)

(21)

so the sinter product frequently contains entrained lead droplets as well

as litharge.

The sulfur dioxide concentrations of lead concentrate roaster gases are

in the range 4 –6 vol % under optimum operating conditions when
sintering machines with gas recycling are used.

Sintering roasting of mixed zinc – lead concentrates, as used in the

Imperial Smelting shaft furnace for simultaneous recovery of zinc and
40
lead, gives only slightly higher sulfur dioxide concentrations of
5 – 7 vol %.

The sulfur dioxide concentration of sintering roaster gases is subject to

large and irregular fluctuations and may remain at extremely low levels
over extended periods. This makes autothermal operation of a double-
absorption sulfuric acid plant difficult. The oxygen content of the roaster
gases discharged from the sinter plant corresponds already to more than
the required minimum O2/SO2 ratio for a contact sulfuric acid plant.

The sulfur trioxide content is influenced to a certain extent by the

sintering temperature and by the presence of other materials such as
iron oxide. It may amount to as much as 5 – 10 % of the sulfur dioxide
content.

The dust content of lead concentrate roaster gases is ca. 5 – 10 g/m3; in

the case of zinc – lead sinter roasting it may even be as high as
20 – 30 g/m3. Therefore, an efficient gas cleaning system is required.
Sinter roasting gases from the processing of flotation concentrates also
usually contain varying amounts of hydrocarbons, which originate from
additives used in the flotation process. These are only partially burnt in
the sintering process, and they pass almost completely through the
normal gas cleaning systems, only being oxidized when they reach the
dryer of the sulfuric acid plant. The colloidal carbon formed by oxidation
colors the acid black. Lighter hydrocarbons even reach the converter,
where they burn.

In sintering technology attempts have been made to raise the roasting

capacity and increase the sulfur dioxide content of the roaster gas by
oxygen enrichment [62]. Although the throughput and sulfur dioxide
concentration could be raised by a small extent, there are decisive
factors in the sintering process, such as the porosity of the mix and the
optimum specific air load for heat transport in the hot sintering zone
which limit the increase in throughput [63]. The cost of using technical
oxygen is out of all proportion to the potential benefits [39], so this
procedure has not been adopted in industrial practice.

Apart from these problems in processing roasting gases with low and
fluctuating sulfur dioxide concentrations, dust emissions and dedusting
within the sinter plant itself present particular difficulties. Development
work on pyrometallurgical lead extraction has, therefore, been directed
towards avoiding the complicated sintering process and carrying out the
roasting and reduction procedures in a single reaction unit so as to
produce waste gases with a higher concentration of sulfur dioxide.
41
Outokumpu Oy, for instance, has processed lead concentrates on a pilot
scale in a flash smelting furnace, applying principles analogous to those
of copper technology. Off-gases with sulfur dioxide concentrations of
11 vol % have thus been obtained [64].

In Italy, a plant for the production of 80 000 t/a lead is in operation based
on the Kivcet process, which is used in the former former Soviet Union
for copper extraction. By using technical oxygen, gases containing
20 vol % SO2 are obtained from the smelting of lead concentrates.

In the QSL process (the Queneau – Schuhmann process as developed

further by Lurgi) [65], (see Fig. 15), the lead concentrate is continuously
fed into a horizontal cylindrical reactor with flue dust and flux additives in
the form of green pellets. In the oxidation zone the lead sulfide is partially
oxidized by injection of technical oxygen (Eq. 22)

Figure 15. QSL lead smelting process

a) Mixer; b) Pelletizing disk; c) Reactor; d) Waste heat

boiler; e) Hot gas electrostatic precipitator

(22)

The lead oxide is taken up by the slag and is reduced by carbon injected
into the subsequent reduction zone. The gases leaving the reactor at ca.
1200 °C reach concentrations of 15 – 20 vol % SO2. QSL plants are in

42
operation in Germany and Korea, with capacities of 75 000 t/a and
60 000 t/a lead, respectively.

5.6. Production by Decomposition of Sulfuric Acid and Sulfates

5.6.1. Waste Sulfuric Acid Processing [66]
Sulfur dioxide- containing gases can be produced by decomposing
sulfuric acid. This procedure has been used industrially since 1932 as a
means of disposing of heavily contaminated spent sulfuric acid while
recovering its sulfur values. The various processes can be classified into
two groups [67], [68]. Sulfur dioxide is produced either by thermal
decomposition at 800 – 1300 °C or by reducing decomposition at
200 – 600 °C. The input to the process is usually spent sulfuric acid so
heavily contaminated with organic substances that it cannot be restored
to a useful level of purity by other processes (see  Sulfuric Acid and
Sulfur Trioxide – Regeneration of Spent Sulfuric Acid). In some cases
the organic contaminants may be of such a nature and be present in
such concentration that the acid is more of a sludge than a liquid. In that
case it will have a relatively high heating value on account of the
impurities, which is advantageous in supplying a portion of the heat
required for decomposition. Decomposition is the most reliable way to
recover high-grade sulfuric acid or oleum from such wastes, as the
impurities are removed completely [69-71].

The sulfur dioxide- containing gases obtained by decomposition of spent

sulfuric acid are generally converted back into sulfuric acid in a contact
plant after purification (see Chap. Cleaning of Crude Sulfur Dioxide Gas
Streams). The most important prerequisite for a decomposition process,
therefore, is that it should generate a gas with a high enough sulfur
dioxide concentration for use in a contact plant. This means that it should
contain at least 4 vol % of sulfur dioxide after its O2/SO2 ratio has been
adjusted to the requisite value by addition of air [72]. The sulfur dioxide
concentration depends chiefly on the H2SO4 content of the waste acid
and the heating value of the oxidizable impurities in the spent acid being
decomposed.

Thermal Decomposition. Gas-phase thermolysis of sulfuric acid is

represented by the overall equation

(23)

To ensure complete decomposition of the acid and combustion of the

43
organic impurities, decomposition is carried out under oxidizing
conditions at ca. 1000 °C [69]. This is a strongly endothermic process,
and the more dilute the waste acid the more additional energy is needed
to evaporate water before decomposition of sulfuric acid begins. To
decompose acid of initial concentration 70 % at 1000 °C, the theoretical
energy requirements are ca. 6.4×106 kJ/t H2SO4 [69]. About 46 % of this
energy is required for dehydration of the 70 % sulfuric acid, evaporation
of the water, heating to the maximum boiling temperature of about
340 °C, and evaporation of H2SO4. Another 31 % is required for
decomposing H2SO4 into SO2, H2O and O2; the remaining 23 % is
required for heating the decomposition products to the reaction
temperature of 1000 °C.

However, this theoretical energy demand of ca. 6.4×106 kJ/t H2SO4

applies only in the case of indirect heat transfer [69]. Furthermore, there
are no construction materials for the heat exchangers that are resistant
to corrosion under these reaction conditions. Therefore, the heat
required for acid decomposition is in practice supplied directly by means
of hot flue gases from the combustion of fossil fuels. Inevitably the heat
requirements are far higher than in the case of indirect heating, as the
flue gases must be heated to the decomposition temperature (1000 °C).
But a large part of the additional heat can be recovered when the gases
leaving the decomposition reactor are cooled from 1000 °C to about
350 °C.

The energy requirement for acid decomposition depends mainly on the

composition of the spent acid. It rises steeply with decreasing H2SO4
concentration in proportion to the amount of water to be evaporated and
it is diminished by the heating value of the organic impurities. In the case
of dilute spent acids, it is always energetically more favorable to
preconcentrate to ca. 60 – 75 % H2SO4 so as to separate the bulk of the
water before decomposition rather than to decompose the dilute acid
directly.

The sulfur dioxide-containing gases produced in the decomposition of

sulfuric acid are heavily diluted by the combustion gases introduced for
direct heating. However, the amount of flue gas can be substantially
reduced by preheating the combustion air. To obtain gases from the
decomposition with a high enough sulfur dioxide content for processing
in a contact sulfuric acid plant when using fuel oil, the spent acid must
contain at least 60 % H2SO4 even when the combustion air is preheated
to 450 °C.

Another method of reducing external energy input and increasing SO2

44
concentrations is the use of oxygen-enriched air.

Figure 16 shows the operation window for the spent acid decomposition
process.

Figure 16. Operation window for Figure 16. Operation window for the
the spent acid decomposition spent acid decomposition process
process

[Full View]

The decomposition of spent sulfuric acid at high temperatures was

developed in the 1930s by Lurgi [65]. Figure 17 shows the flow diagram
of a modern spent acid regeneration plant. The decomposition unit itself
consists of the empty decomposition furnace and the waste heat boiler
system. Downstream of this is a gas cleaning system for further cooling
and purifying the decomposition products (see Chap. Cleaning of Crude
Sulfur Dioxide Gas Streams) and a double-absorption plant producing
concentrated sulfuric acid and oleum. The application of such an empty
decomposition furnace is, however, limited to acids with low residue
levels and low organics concentrations that do not result in excess heat
beyond the thermal energy requirements of the process.

Figure 17. Sulfuric acid regeneration plant

a) Decomposition furnace; b) Air preheater; c) Waste heat boiler; d)

Venturi scrubber; e) Gas cooler; f ) Electrostatic precipitator; g) Gas
heat exchanger; h) Converter; i) Intermediate heat exchanger; j) Fan; k)
Drying tower; l) Oleum tower; m) Intermediate absorber; n) Final
absorber

45
Spent acid is fed to the top of the vertical, brick-lined furnace and is finely
distributed by means of a rotary atomizer [69] specifically developed by
Lurgi for atomizing viscous spent sulfuric acid. The atomizing medium is
cold air, injected at ca. 1.35 bar. Fuel oil or gas is burnt in several
combustion chambers attached laterally to the top of the furnace. The
separation of the combustion chamber from the furnace permits
complete mixing and combustion of the fuel with the combustion air
outside the acid decomposition zone proper. A 10 – 30 % excess of air
over the amount actually required for combustion of the fuel is
maintained. This is because the amount of oxygen liberated in the
decomposition of the acid is not normally enough to burn the oxidizable
inorganic and organic components completely. The flue gases enter the
furnace tangentially on the same level as the combustion chambers at
ca. 1800 °C. A part of their heat content is transferred to the furnace top
by radiation, but the majority is transferred to the acid because of the
intensive mixing which takes place in the vortex formed near the outlet
openings of the combustion chambers. The furnace temperature of ca.
1000 °C required for acid decomposition is adjusted by controlling the
fuel and combustion air rates.

The sulfur dioxide-containing gases leaving the furnace are cooled to ca.
350 °C in an energy recovery system. Depending on the intended use of
the recovered energy, different waste heat systems may be used for this
purpose. Figure 18 shows three typical decomposition plant concepts
with different energy recovery systems [69].

46
Figure 18. Typical sulfuric acid regeneration plant concepts with
different energy recovery systems [19]

The system in Figure 18 A (analogous to Fig. 17) is designed to produce

high-pressure steam (60 – 65 bar). The heat content of the gases
emerging from the furnace is used to generate saturated steam. The
steam is then heated to ca. 450 °C in a superheater fired by fuel oil. The
residual heat of the flue gas is recovered in a gas – gas heat exchanger
which preheats combustion air for the decomposition furnace to about
350 °C. Enough high-pressure steam is produced by this system to
power a back-pressure turbine drive on the main blower of the
downstream sulfuric acid plant, which — as is usual for a “cold gas”
plant — does not generate steam of its own. The low-pressure steam
exhausting from the back-pressure turbine may be used to provide heat
for preconcentrating the spent sulfuric acid.

This process concept is specifically suited for larger plants with

capacities > 100 t/d H2SO4. It has, for example, been used in a Lurgi-built
47
plant regenerating 360 t/d of 65 % waste sulfuric acid.

In the system of Figure 18 B, the decomposition furnace is followed by a

gas – gas heat exchanger, in which the decomposition products are
cooled indirectly from 1000 °C to ca. 350 °C. Simultaneously, the furnace
combustion air is heated to ca. 450 °C. In general, only about one third of
the hot air generated in this heat exchanger is required in the fuel oil
burners of the decomposition furnace. The remaining two thirds can be
utilized, for instance, for generating steam at 10 bar in a waste heat
boiler. A waste heat system of this type has been installed in a twin-
stream decomposition plant built by Lurgi for processing ca. 900 t/d of
ammonium sulfate- containing spent sulfuric acid resulting from acrylic
acid production.

The system of Figure 18 C has the same gas – gas heat exchanger
arrangement for cooling the decomposition gas and preheating air as
system B. But in this case the surplus hot air at about 450 °C is used,
along with the dry tail gases of the sulfuric acid plant, for
preconcentrating the spent sulfuric acid. In this way, 43 % sulfuric acid
can be concentrated to 70 %. The preconcentration unit and the
downstream sulfuric acid plant can be integrated so as to allow waste
heat from the acid plant absorption system to be utilized for preheating
the contaminated acid. In this case spent acid even as dilute as 33 %
H2SO4 can be concentrated to 70 % using only waste heat [69].

In another decomposition process, developed by Chemical Construction

Corporation (Chemico) in co-operation with Consolidated Chemical Co.
(now Stauffer Chemical) [73], the acid is decomposed at 1000 – 1200 °C
in a horizontal furnace in which the fuel is burnt directly. For this process,
exact control of the oxygen content of the furnace gases is necessary. If
the gases contain less than about 2 % free oxygen, there is a risk that
sulfur dioxide may be reduced to elemental sulfur, which is precipitated
in the subsequent colder parts of the plant. The spent acid is injected
through nozzles into the furnace. Acid sludges with a high content of
organic impurities are mixed with spent acids from other sources to
reduce their viscosity so that they can be atomized more easily.

Spent acids with widely varying physical and chemical properties can be
decomposed in the Grillo rotary kiln process. The Grillo process (Fig. 19)
is a multistage decomposition process which is particularly suitable for
the decomposition of spent sulfuric acid with widely varying physical and
chemical properties.

Figure 19. Lurgi Grillo decomposition

48
process

a) Rotary kiln decomposition furnace; b)

Air preheater; c) Intermediate combustion
chamber; d) Post combustion chamber;
e) Waste heat boiler

The Grillo acid decomposition plant consists of a rotary kiln

decomposition reactor, a downstream intermediate combustion chamber,
and two post-combustion chambers.

The waste acid is injected virtually unpressurized onto a red hot coke
bed in the rotary kiln with relatively coarse atomization. The coke bed
serves the following three main functions:
1. Uniform distribution of the liquid feed, which offers advantages in
terms of reaction rate and uniform reaction conditions
2. Heat storage and regulation
3. Most of the solids content of the waste acid is retained in the coke
bed.

In the rotary kiln, the spent acid is burnt under substoichiometric

conditions at 900 – 1000 °C, so that the decomposition gases contain
H2S and COS in addition to CO. By adding combustion air to the
intermediate combustion chamber downstream of the rotary kiln, the
organic decomposition gas components are completely oxidized at
1250 °C.

Advantages of this process, which is relatively simple to control, are the

drastic reduction of the NOx levels in the decomposition gas, the
complete destruction of PCB- containing oils due to the high reaction
temperatures, and the retention of the acid-borne solids in the form of
coking duff.
49
Even when using residue oils with high PCB levels as an energy source,
this process meets all environmental standards.

For treating spent acids with high residue concentrations dry metal
sulfates, and all mixtures of sulfur- containing solids with spent acid, the
fluid bed decomposition (Fig. 20) process is particularly suitable.

Figure 20. Decomposition in fluid bed

reactor

a) Waste heat boiler; b) Hot

electrostatic precipitator

The stationary fluid bed process for sulfate and spent acid decomposition
has been successfully employed on a commercial scale for more than 10
years.

A further development of this process is circulating fluid bed technology,

which achieves improved decomposition efficiencies due to staged
reaction control in the decomposition furnace.

In thermal sulfuric acid and sulfate decomposition, staged process

control will gain in significance, as this is the simplest method to
suppress NOx formation. The decomposition temperatures are in the
range of 900 – 1000 °C.

Where feasible, the decomposition of spent acids/sulfates can be

combined with the roasting of pyrites. The rate at which acid can be
processed is limited by the available pyrite roasting reaction heat. When
the spent acid is of about 70 % concentration, the ratio of the sulfur
dioxide produced from pyrite to that produced from the acid is about
3 : 1. This means that the sulfuric acid output of the downstream plant is

50
four times the H2SO4 content of the spent acid feed to the fluid bed
reactor. To increase the decomposition capacity of the pyrite roaster, fuel
oil, sulfur and oxygen-enriched air may be introduced.

Reductive Decomposition. In a process developed by Hechenbleickner,

spent sulfuric acid is decomposed at 200 – 600 °C after mixing with a
reducing material such as tar sludge or coal. The reduction product is
sulfur dioxide, and coke is generated as a byproduct [68]. This process is
chiefly used for processing acid sludge from lubricating oil and kerosene
production plants associated with oil refineries. In the original process
concept, the sludge, which contains up to 50 % organic matter and has a
comparatively high calorific value, is fed through a rotary kiln in
countercurrent with hot flue gas. The most important variants of this
process are known under the names of Chemico [73], [74],
Ross – Wilde, and Miley [68]. A further development of the process has
been designed to process a wide range of spent acids containing from
0.5 to 60 wt % C and 40 – 90 % H2SO4 [75].

In all process variants of the reductive acid decomposition, the sulfur

dioxide-containing raw gas obtained contains, besides carbon dioxide
and moisture, varying proportions of carbon monoxide and hydrocarbons
[68]. Since the hydrocarbons, especially, are detrimental in a
downstream contact plant, generally post- combustion at about 800 °C is
required before the gas cleaning stage.

5.6.2. Sulfate Decomposition

Calcium Sulfate. All industrial processes for producing sulfur
dioxide- containing gases by decomposition of calcium sulfate, the so-
called gypsum – sulfuric acid processes, are based on the
Müller – Kühne process [76], [77]. In this process the calcium sulfate is
decomposed in a rotary kiln at 900 – 1400 °C to give sulfur
dioxide- containing gases and cement clinker. The sulfur dioxide is
processed to sulfuric acid in an associated contact plant. The clinker is
used for making Portland cement, whose sale recovers part of the
extremely high costs.

Though originally the raw materials for this process were natural
anhydrite or natural gypsum, the process has been employed for
decomposing waste gypsum from chemical processes, especially the
phosphogypsum, produced as a waste product in the manufacture of
phosphoric acid by the wet process (sulfuric acidulation of phosphate
rock). In addition, coke is required as a reducing agent and additives
such as clay, sand, and iron oxide are needed to convert the lime into a
clinker that can be utilized in cement manufacture.

51
The Müller – Kühne process is based on the reductive decomposition of
calcium sulfate [78] with carbon above 700 °C. Part of the calcium sulfate
is reduced to calcium sulfide, which reacts above 900 °C with the
remaining calcium sulfate to form calcium oxide and sulfur dioxide
(Eq. 24 – 26)

(24)

(25)

(26)

It is assumed that the reduction to calcium sulfide takes place not as a

solid-state reaction but by way of an intermediate stage involving carbon
monoxide [79]. The silica-, alumina- and iron- containing additives
accelerate the release of the sulfur dioxide by reducing the
decomposition temperature and by undergoing side reactions with the
calcium sulfate [77]. They also react at 1200 – 1400 °C with the resulting
calcium oxide to form cement clinker. The clinker quality depends
essentially on exact proportioning of the carbon in relation to the calcium
sulfate [78], [80]. Although the stoichiometric molar C/ CaSO4 ratio for
decomposing calcium sulfate is 1 : 2, a carbon excess of about 20 % is
applied in practice because part of it will burn directly and will therefore
not be available for reduction [79]. If not enough carbon is present, the
clinker will contain unconverted sulfate and will melt; if there is too much,
the clinker will contain sulfide and will disintegrate. Under strongly
reducing conditions in the bed, elemental sulfur may form and vaporizes
under the prevailing conditions. It can be burnt in the gas phase merely
by injecting additional air, but this is undesirable because the nitrogen
from the extra air dilutes the sulfur dioxide in the product gases.

The primary materials have to be prepared specifically for the process.

Gypsum must first be calcined to remove its water of crystallization and
convert it to anhydrite, a process which requires additional energy. The
anhydrite, coke, sand, and alumina are finely ground, dried, and fed into
52
a mixer by an automatic metering device. In some cases, the raw meal is
pelletized [81]. The average composition is 78 – 80 % anhydrite,
9 – 10 % clay, 4 – 5 % sand, and 6 – 7 % coke. The material is fed in at
the upper end of the slightly inclined rotary kiln and passes in
countercurrent with the hot gases to the lower, heated end, from which
the clinker product is discharged.

The rotary kiln (normally about 70 m long and about 3.2 m inside
diameter) is refractory lined with magnesite bricks at its lower end, in the
clinkering zone, where a temperature of about 1400 °C prevails.
Formerly, pulverized coal was used for heating but those plants which
are still in operation preferentially use fuel oil.

The raw gas leaving the kiln at about 350 – 420 °C contains ca.
7 – 10 vol % SO2, 0.7 –1.6 vol % O2, 15 – 20 vol % CO2, and 23 vol %
H2O [78]. It is efficiently cleaned in a gas cleaning system (similar to
those used at roasting plants; see Chap. Cleaning of Crude Sulfur
Dioxide Gas Streams) and is then processed in a contact sulfuric acid
plant after addition of enough air to bring the O2/SO2 ratio to the required
value. The total sulfur yield is about 85 %.

The hot cement clinker leaving the kiln is cooled with air in the clinker
cooler and is then ground with about 2 % of added gypsum to obtain the
cement product [82]. The attainable quality compares with that of high-
grade Portland cement. With this process, cement and sulfuric acid are
produced in equal quantities. Per tonne of cement or sulfuric acid, the
following raw material quantities are required [77]: 1.6 t anhydrite or 2.2 t
natural gypsum, 0.3 t clay, 0.1 t sand, and 0.1 t coke. Furthermore, a
total of 9.6 – 11.7×106 kJ of energy in the form of fuel oil are consumed.
This includes that required in the raw material preparation system.

If phosphogypsum is used, it is much more difficult to produce high-

grade cement, and the cleaning of the sulfur dioxide- containing gases is
much more complicated than when natural gypsum is used [74]. On
account of its high moisture content (up to 40 %), phosphogypsum
requires a large amount of calcining energy. The precise contents of
residual phosphate and fluoride depend on the phosphoric acid
production process and the rock phosphates used, but they are
practically always considerably above the very low levels allowable either
in cement clinker or in sulfuric acid plant feed gas. Therefore, they have
to be removed by a specific, elaborate pretreatment of the
phosphogypsum. Material analyses and consumption figures for the
OSW-Krupp process, developed by Oesterreichische Stickstoffwerke

53
(now Chemie Linz) are shown in Table 2 [77]. Material requirements per
tonne of clinker or sulfuric acid are as follows:

Gypsum (as calcined anhydrite) 1.6 – 1.7 t

Clay 0.2 – 0.7 t
Sand 0.1 t
Fuel oil 11.7 – 13.8×106 kJ

Table 2. Material analyses and consumption figures for the

OSW – Krupp calcium sulfate decomposition process [77]

wt % in raw wt % in
Constituent meal clinker

SiO2 9.33 19.35

Al2O3 2.8 5.81
Fe2O3 0.93 1.94
CaO 30.45 63.19
SO3 42.86 2.00
F 0.15 0.30
P2O5 0.50 1.00
C 4.50 0
H2O 5.26 0

The gypsum – sulfuric acid process was developed by MÜLLER and

KÜHNE in Leverkusen, Germany, during World War I, from 1915 to 1917
[76] as a means of making sulfuric acid from indigenous raw materials
independent of imports. It has remained of interest in countries with
natural gypsum resources where, for strategic or foreign exchange
reasons, it is desirable to avoid dependence on imported sulfur.
Furthermore, a Müller – Kühne plant operating on recycled
phosphogypsum is, in theory, an ideal means of avoiding waste disposal
problems in phosphoric acid production. Nowadays, though, its very high
capital costs, as well as the cost of its very substantial energy
requirement, have made the economics of the Müller – Kühne process
so unfavorable as to preclude its use in any but a handful of special
cases. The capital costs for a complete gypsum – sulfuric acid plant are
54
about five times as high as those of a sulfuric acid plant based on sulfur
combustion. And while the gypsum-based plant is a substantial net
energy consumer, a sulfur-based plant produces a substantial energy
credit.

The first gypsum – sulfuric acid plant (40 t/d H2SO4) operated in
Leverkusen from 1916 to 1931. Other plants running on natural gypsum
or anhydrite were built in the United Kingdom (Billingham, Widnes,
Whitehaven), France (Miramas), Germany (Wolfen, Coswig), Poland,
and Austria (Linz) [77]. The French plant shut down as early as 1952 and
the plants in Great Britain all closed down in 1972 – 1975. The plant in
Linz (230 t/d H2SO4) was modified in 1969 to run on phosphogypsum. In
1972, a plant with a capacity of 350 t/d H2SO4, processing
phosphogypsum according to the OSW – Krupp process, started up at
Phalaborwa in South Africa.

Iron(II) Sulfate. Iron(II) sulfate can be decomposed into sulfur dioxide and
iron oxide in fluid bed furnaces, multiple hearth furnaces, or rotary kilns
[72]. Iron(II) sulfate is obtained in large quantities as its heptahydrate
(green salt) during the regeneration of pickling liquors from the metals
industry and in the production of titanium dioxide pigment by the sulfate
process, [68], [73] which produces 3 – 4 t of green salt for every tonne of
titanium dioxide. In 1978, titanium dioxide production capacity in the
Western Hemisphere based on the sulfate process amounted to ca.
1.6×106 t/a [83]. Disposing of the enormous amount of co-produced
green salt is a major problem. Decomposition is one of the few remaining
options for keeping this largely unwanted material from entering the
environment.

Iron(II) sulfate heptahydrate is normally separated from sulfuric acid

solutions by crystallization. In most cases, it is subjected to preliminary
heat treatment to drive off the majority of the water of crystallization
before it is decomposed [74], since the direct decomposition of the
heptahydrate is energetically unfavorable. In this first treatment step, the
heptahydrate is dehydrated at 130 – 200 °C by flue gases in spray dryers
or fluid bed dryers to a monohydrate or mixed hydrate of the empirical
formula FeSO4 · 1.5 H2O [84]:

(27)

In the second step, the sesquihydrate is decomposed at ca. 900 °C

55
 (28)

Formerly, decomposition was conducted in sintering machines or rotary

kilns with coke as reductant [73], and the gases thus obtained contained
ca. 7 vol % of sulfur dioxide. Today it is common practice for iron(II)
sulfate to be decomposed in a fluid bed pyrite roasting furnace at 850 °C
or more. Elemental sulfur, coal, or fuel oil may be used as supplementary
fuels [84]. The sulfate is fed in above the fluid bed or iron(III) oxide. The
sulfur dioxide- containing gas leaving the furnace is cooled in a waste
heat boiler to 350 – 400 °C and is subsequently passed on to the gas
cleaning system. The cleaned gases are fed to the sulfuric acid plant.
The mixture of metal sulfates and sulfuric acid resulting from the
concentration of acid wastes of titanium dioxide production can similarly
be processed in a fluid bed reactor [85]. In individual cases, iron(II)
sulfate is also decomposed in multiple-hearth furnaces with flue gases
from fuel oil or natural gas combustion [72], [85], [86].

5.7. Production from Waste Gases

5.7.1. Combustion of Hydrogen Sulfide
The main source of hydrogen sulfide-containing gases is fossil fuel
upgrading. Large-volume gas streams with high hydrogen sulfide
contents such as those produced in plants for natural gas purification, in
crude oil refineries and in plants for the liquefaction or gasification of
solid fuels are in most cases treated by the Claus process to convert the
hydrogen sulfide to elemental sulfur (see also  Sulfur – The Claus
Process). With small-volume gas streams such as coke plant off-gases
the hydrogen sulfide is burnt to sulfur dioxide, which is converted to
sulfuric acid in a wet contact plant. Gas streams which cannot be
processed by the Claus process because they contain too little hydrogen
sulfide or too high a concentration of gaseous constituents that interface
with the Claus process are usually incinerated to oxidize the hydrogen
sulfide to sulfur dioxide. Even extremely dilute gas streams, if they
contain any hydrogen sulfide or other sulfur-containing compounds such
as carbonyl sulfide, carbon disulfide, and thiols, cannot be discharged to
the atmosphere because of their high toxicity. The maximum allowable
concentration of hydrogen sulfide that can be emitted in Germany, for
example, is 10 mg/m3. The sulfur dioxide-containing gases formed during
incineration are either processed to sulfuric acid by special wet catalysis
methods or, at very low sulfur dioxide concentrations, are emitted
directly.

56
The hydrogen sulfide-containing gases are burnt with an excess of air:

(29)

Depending on the calorific value of the gases and their content of air, the
combustion temperature may range from 500 to 1000 °C. The calorific
value is a function of the concentrations and the heats of combustion of
the hydrogen sulfide and of the oxidizable accompanying compounds
such as carbonyl sulfide, carbon disulfide, carbon monoxide, ammonia,
hydrogen cyanide, hydrogen, and organic sulfur compounds. In cases
where the combustion temperature required for complete oxidation of the
noxious substances cannot be reached, a pilot flame or supplementary
natural gas or fuel oil firing must be provided. Gases with a hydrogen
sulfide content of 10 vol % or more or with a corresponding calorific
value can normally sustain combustion without supplementary firing.

Gases with fluctuating and, especially, extremely low hydrogen sulfide

contents cannot be oxidized completely without supplementary heating.
For them it is preferable to use catalytic incineration. Special catalysts
are available which completely convert all hydrogen sulfide at initiation
temperatures of ca. 200 °C. The essential advantage of catalytic
incineration over thermal incineration of low-grade hydrogen sulfide-
containing gases is that it consumes no additional fossil fuel. This in itself
is a saving, and another advantage is that the sulfur dioxide-containing
gas formed has a lower moisture content and is not diluted by
combustion gases. In addition, the catalyst presents a large specific
surface area for the reaction so that the hydrogen sulfide can be
completely oxidized with a lower oxygen excess. Therefore a smaller
volume of oxidized gas with a correspondingly higher sulfur dioxide
concentration is formed.

The sulfur dioxide concentration is determined by the concentration of

hydrogen sulfide and other sulfur compounds and the relative volume of
air. Where the oxidized gases are destined for subsequent processing to
sulfuric acid, the air necessary to provide the required O2/SO2 ratio of
1.1 – 1.3 in the contact gases is introduced directly into the hydrogen
sulfide combustion zone. Starting from a gas with 20 vol % hydrogen
sulfide, for instance, a contact gas with 6 vol % sulfur dioxide is obtained.
The water content of the sulfur dioxide-containing gas is inherently high
because the oxidation of hydrogen sulfide produces an equimolar
57
quantity of water, to which is added water from the combustion air and
water generated by combustion of any other hydrogen-containing
substances present in the gas or added as auxiliary fuel.

The hydrogen sulfide-containing gases are normally burnt in a horizontal

cylindrical furnace, lined with several layers of refractory bricks and
equipped with special safety devices such as explosion flaps, as
mixtures of air with hydrogen sulfide in the concentration range
4.3 – 45.5 vol % H2S are explosive [87]. The hot gases leaving the
furnace are preferably cooled in a fire-tube waste heat boiler to generate
steam. It must be ensured that the wall temperatures do not fall below
the relatively high dewpoint of the moist gases, which is partly
determined by the sulfur trioxide content.

5.7.2. Recovery from Flue Gases

The waste gases produced by combustion of sulfur-containing fossil
fuels such as coal or crude oil in power stations (“flue gases”) have quite
low sulfur dioxide contents. The sulfur dioxide content of flue gas from a
plant fired by coal containing 1 % S, for instance, is about 700 mL/m3.
Consequently these gases are inherently unsuitable for processing by
methods used for treating metallurgical waste gases, for example.
However, on account of the large installed power of such stations the
volume of flue gas is such that the emission rate of sulfur dioxide is
extremely high in absolute terms and is nowadays environmentally
intolerable. A 600 MW power station fired with coal containing 1 % S, for
instance, emits about 2×106 m3 flue gas containing ca. 4 t/h of sulfur
dioxide.

Most industrial countries have introduced exacting regulations relating to

environmental protection in recent years, and a large number of
dedicated processes for flue gas desulfurization have been developed,
some of which are already in use in the industrial countries on a large
scale. In most processes, the sulfur dioxide is ultimately converted to
calcium sulfate, which is partly disposed of as a waste material and
partly utilized in the construction industry. Apart from these, there are the
so-called regenerative processes in which the sulfur dioxide is recovered
in a concentrated form from the flue gases (“rich gas”) or in the form of
sulfuric acid (Desonox-process). Typical regenerative processes are the
BF process of Bergbau-Forschung and the Wellman – Lord process of
Davy McKee (see Absorption Processes.).

The BF process is a dry flue gas cleaning process based on sulfur

dioxide sorption on activated coke with thermal regeneration [88]. A
simplified flow diagram of this process is shown in Figure 21. The

58
dedusted flue gases from the electrostatic precipitator of the power
station are conveyed at about 130 °C to the sulfur dioxide adsorber,
which operates as a moving-bed reactor. The bed is composed of
activated coke and moves downwards in crosscurrent with the flue gas.
About 80 – 95 % of the sulfur dioxide is adsorbed on the activated coke;
simultaneously, residual flue dust is arrested. The cleaned flue gas is
drawn through by means of a fan, which discharges into the stack. In
contrast to wet flue gas desulfurization processes, in this case the
purified gas does not need reheating.

Figure 21. Bergbau-Forschung Figure 21. Bergbau-Forschung

regenerative flue gas desulfurization regenerative flue gas
process desulfurization process

a) Absorber; b) Coke cooler; c) a) Absorber; b) Coke cooler; c)

Regenerator; d) Cyclone; e) Screen; f ) Regenerator; d) Cyclone; e)
Burner Screen; f ) Burner
[Full View]

The sulfur dioxide adsorbed on the activated coke is oxidized in the

presence of adsorbed moisture by the residual oxygen available in the
flue gas. The resulting sulfuric acid is stored in the coke pores. The laden
coke is discharged continuously from the adsorber into the regenerator,
where it is heated to 600 – 650 °C by the addition of hot sand at ca.
700 °C. The sulfuric acid is decomposed to water vapor and sulfur
trioxide. The sulfur trioxide is immediately reduced to sulfur dioxide by
the coke, which is itself partly converted to carbon dioxide. Accordingly,
the regenerated sulfur dioxide contains substantial amounts of carbon
dioxide and moisture.

The regenerator is designed as a moving bed reactor. The mixture of

sand and regenerated activated coke discharged at the bottom is
59
separated by means of a vibratory screen. The coke is directly cooled to
about 120 – 140 °C by injected water and is recycled to the top of the
adsorber. Any carbon losses due to the chemical reaction in the
regenerator are compensated by the continuous addition of fresh
activated coke. The sand is heated in a fluid bed by combustion of fuel
gas and is conveyed pneumatically to the top of the regenerator in the
hot flue gases.

Typical analysis values for the rich sulfur dioxide- containing gas
discharged from the regenerator are: 25 vol % SO2, 15 vol % CO2,
60 vol % H2O, and traces of N2, SO3, HCl, and HF.

The chlorine and fluorine concentrations in the rich gas are about 300
times as high as the concentrations in the original flue gas [88]. Before
subsequent processing, the impurities must be removed in a gas
cleaning system similar in principle to the type used for cleaning roaster
gases (see Section Cleaning Systems). The purified and cooled rich gas
contains 50 – 60 vol % SO2 and 30 – 35 vol % CO2, the remainder
consisting of water vapor and nitrogen.

The BF process is suitable for the desulfurization of flue gases with sulfur
dioxide concentrations of up to 0.4 vol % [88]. The sulfur dioxide-rich gas
produced can be utilized for the production of liquid sulfur dioxide, sulfur
or sulfuric acid. In 1974 a large-scale pilot plant using the process was
started up in Lünen. It was designed to treat a flue gas volume of about
150 000 m3/h, the rich gas being processed to elemental sulfur [88].
6. Cleaning of Crude Sulfur Dioxide Gas Streams

6.1. Constituents of Roaster and Metallurgical Gases

The crude sulfur dioxide-containing gases obtained by pyrites roasting,
pyrometallurgical nonferrous metal extraction from sulfides, and thermal
decomposition of waste sulfuric acid and sulfate must be cooled and
cleaned efficiently before being processed to obtain sulfuric acid or liquid
sulfur dioxide.

The impurities which may be present can be classified into four

categories according to their state of aggregation and the process step of
the gas cleaning system in which they are removed [43].

Primary dusts are particles discharged in solid form from the roasting or
metallurgical extraction process at the so- called flue dust, purely
mechanically entrained by the waste gases. The proportion and grain
size of such impurities depend on the furnace feed and the flow
conditions.

60
Fumes are sublimed substances which are volatilized under the furnace
conditions but have resolidified in the form of an aerosol by the time they
enter the dedusting system inlet at 300 –400 °C. Their grain size is
0.01 – 2 µm. Frequently they consist of zinc and lead compounds such
as zinc oxide and lead sulfate or basic lead sulfate.

Volatiles are substances which are still gaseous when they enter the dry
dedusting system but may precipitate in the cold areas of the gas
cleaning system. These include arsenic and selenium compounds,
mercury, sulfuric acid mists, and hydrocarbons.

Gaseous substances include, for instance, hydrogen chloride, hydrogen

fluoride, and lower hydrocarbons. Normally, carbon dioxide and small
quantities of carbon monoxide originating from the combustion gases of
supplementary heating systems are not harmful.

Whereas the primary dusts and fumes can in most cases be separated
mechanically at elevated temperatures in a dry gas cleaning step, the
gas must be cooled and passed through a wet cleaning stage to ensure
complete separation of the volatile and gaseous impurities.

6.2. Cleaning Systems

The simplified flow diagram of a Lurgi gas cleaning system for sulfur
dioxide- containing gases from a fluid bed roasting plant is shown in
Figure 22. The same combination of individual process steps is also
used for cleaning sulfur dioxide-containing gases produced in other
metallurgical processes.

Figure 22. Gas cleaning plant (Lurgi system)

a) Cyclone; b) Hot-gas electrostatic precipitator; c) Venturi scrubber;

d) Settling tank; e) Pump tank; f ) SO2 stripper; g) “Star” gas cooler; h)
Wet electrostatic precipitator

61
Dry Dedusting. After cooling in a waste heat boiler to 320 – 400 °C, the
furnace waste gases, laden with 50 – 150 g/m3 of dust, pass through a
cyclone for preliminary separation of the coarse particles. Generally, the
collection efficiency is 85 – 90 % with a pressure drop of about
5 – 10 mbar.

The majority of the dust remaining in the hot gas is then removed in a
hot-gas electrostatic precipitator. The gas passes between two
electrodes, one of which (the collector(điện cực lắng) electrode) is
grounded and the other (the discharge electrode) is maintained at a
sufficiently high negative potential to ionize the surrounding gas
molecules, resulting in a so-called corona discharge. Dust particles
passing through the corona acquire an electrostatic charge of the same
polarity as the discharge electrode and are, therefore, repelled towards
the collector electrode, where they are discharged.

Because corona discharges form much more readily around charged

objects which have a very small surface area, and particularly from
pointed objects, the discharge electrode is usually in the form of a wire or
needle. The collector electrode, however must have as large a surface
area as possible to minimize the chances of any charged particles
escaping. Hot-gas precipitators are usually of the plate type, in which the
gas passes through the channels between a battery of vertically mounted
flat collector electrodes. Rows of discharge electrodes are suspended in
the gas spaces between the plates. Dust accumulating on the surface of
the collector electrodes falls by gravity into a hopper below the
precipitator. The applied d.c. voltage between the discharge and
62
collector electrodes is 10 – 80 kV. The specific power requirements are
1.5 – 2.5 kWh/1000 m3.

Electrostatic precipitators have the considerable advantage of being able

to separate more than 99 % of the dust at a pressure drop of  1 mbar.
The residual dust content is typically 20 – 200 mg/m3.

The size and mode of operation of the hot gas electrostatic precipitator
depends not only on the volume of gas but also on its dust loading, the
size distribution of the entrained particles and, above all, its resistivity.
The gas velocity through the unit must be kept below the value at which
significant re-entrainment of the discharged particles occurs. The
optimum temperature range for operation is determined by the dust
characteristics. By means of adequate insulation and temperature
control, especially in the “dead” corners, it can be ensured in practical
operation that the temperature does not drop below the dewpoint, which
would give rise to corrosion by acid condensation [89].

The separated dust is normally recycled to the roasted material handling

system.

Wet Separation. The remaining impurities are removed by scrubbing the

gases. Adiabatic water evaporation simultaneously cools the gas to
60 – 80 °C so that the major part of impurities such as selenium, arsenic
trioxide, and sulfuric acid condense. Some of the impurities are not
washed out in the scrubber but remain in the gas in the form of minute
solid or liquid particles. These are subsequently removed in wet
electrostatic precipitators (see below). Gaseous impurities such as
hydrogen chloride or fluoride are also separated in the scrubber, at least
in part.

The scrubbing liquid is normally recirculated, so the separated impurities

accumulate in it. Solid impurities can be removed as a slurry by
sedimentation in a settling tank. Liquid impurities with a relatively low
vapor pressure, such as sulfuric acid, increase the acidity of the
scrubbing liquid, hence the term wash acid. Part of the wash acid is
continuously bled off to hold the concentration in the circuit at or below a
maximum of 30 – 50 % H2SO4. It is replaced by fresh water and
recirculated condensate from the sub-sequent wet gas cleaning systems.
In some cases it may be necessary to bleed the wash acid at a lower
sulfuric acid concentration so as to hold the concentrations of other,
more volatile impurities such as hydrogen chloride and hydrogen fluoride
down to levels at which they are still efficiently absorbed. Arsenic (as
As2O3) first condenses in the scrubber and then dissolves in the wash
63
acid. The limit of solubility may be exceeded in the circulating wash acid,
and crystalline arsenic trioxide can be separated from the supersaturated
solution in settling tanks. The discharged wash acid is normally
neutralized after air stripping to remove sulfur dioxide.

The scrubbing and adiabatic cooling procedure is usually performed in

an unpacked scrubber tower, a Venturi scrubber, a radial flow scrubber,
or a combined scrubbing and cooling tower of the Peabody type.

The moisture content of the gases at the temperature at which they leave
the scrubber (60 – 80 °C) is too high for production of concentrated
sulfuric acid in a contact plant. The gas therefore must be cooled to
reduce its moisture content.

Various systems can be used for gas cooling. One method is to pass it
through an additional, packed scrubbing tower in which it is cooled by
circulating scrubbing liquid. The warm liquid leaving the tower is cooled
in a scrubbing acid cooler and is then pumped back to the top of the
tower. In another method, the gas is cooled to the requisite gas
temperature (30 – 40 °C) in indirect gas coolers (star coolers). The
resulting condensate is passed on to the scrubber where it is used for
balancing the evaporation losses that occur there.

When the crude gas comes from the roasting or pyrometallurgical

treatment of an ore concentrate with a high fluorine content, it may also
be necessary to install a so- called fluorine tower to remove any
hydrofluoric acid still present after scrubbing and gas cooling. This tower
is packed with a material with a high content of silica, which reacts with
hydrogen fluoride to form fluosilicic acid.

The solid and liquid impurities still present in the gas after scrubbing are
removed almost completely in wet electrostatic precipitators.

To ensure that no unpurified gas reaches the contact plant, even in case
of a wet precipitator failure, two precipitator steps are normally installed
in series (primary and secondary cleaning). The two stages together
have a collection efficiency of > 99 %. At the outlet the gas is optically
clear when viewed against a light source.

Two patterns of wet precipitator are in use. The older plate type is made
of lead. The gas passes along the space between a battery of extremely
smooth vertical plates which are the collecting electrodes. The discharge
electrodes are suspended between the collector electrodes, from which
they are isolated.
64
Nowadays, though, the plate type of precipitator has practically been
replaced by the tube type, in which the collector electrodes are tubular
and the discharge electrodes are suspended concentrially inside them. In
these precipitators the tubes are made of plastics: the requisite
conductivity is attained by preparing the inside of the tube so that the
separated acid mist will form a uniform liquid film flowing downwards. In
another version, the collection tubes are replaced by a plastic honey
comb element.

For the best possible separation effect, the voltage must be

maintained — as in the dry precipitators — just below the spark-over
voltage of the gas, which may fluctuate, depending on the process. An
automatic voltage control system is used for this purpose.

A narrow negative electrode is required only for establishing the corona

discharge. The subsequent repulsion of charged dust particles away
from the negative electrode is, in fact, more efficient if the electrode
diameter is larger and there is no discharge. Some tubular precipitators
are therefore designed on a two-stage principle, in which a relatively
short corona discharge needle is mounted on the end of a wider,
nondischarging electrode rod. There are thus discrete ionization and
collection zones [90].

The wet precipitator casing may also be made from plastic, but steel
casings with internal rubber-linings or homogeneous interior lead
coatings are used as well.

The gas-side pressure drop of such electrostatic precipitators is ca.

2 mbar. The high-tension set requires about 0.8 kWh per 1000 m3 gas.

The gases at the gas cleaning system outlet must be optically clear
(checked by means of a light source) to guard against problems in
subsequent processing. Any residual moist mists are difficult to separate,
even in the dryer of a sulfuric acid plant, so they may cause substantial
operational and corrosion problems.

Mercury Removal. Any mercury contained in the concentrates is

volatilized almost completely, passing on to the scrubbing plant as
mercury vapor. Depending on the partial pressure, metallic mercury may
occur in the scrubbing system, but part of it always remains as vapor in
the gas.

As the volatilized mercury still contained in the SO2 gas will end up in the
65
sulfuric acid production, special gas treatment has to be provided for the
removal of the mercury in order to obtain acceptable mercury contents in
the acid. Alternatively, if the mercury vapor remains in the gas, the
sulfuric acid produced has to be treated to eliminate the mercury.

Both wet and dry methods have been used to remove mercury from
roaster gas [91]. Dry methods include filtration through a porousmaterial
impregnated with selenium, with which the mercury reacts to form
mercury(II) selenide, and activated carbon adsorption. Wet methods
include introducing elemental selenium into the wet scrubber of the gas
cleaning plant and washing with mercury(II) chloride solution to convert
the mercury to insoluble mercury(I) chloride (removed from the scrubbing
liquor by filtration). Scrubbing with concentrated sulfuric acid removes
mercury in the form of mercuric sulfate.

If mercury is not removed from sulfuric acid plant feed gas, it is absorbed
in the drying tower and therefore appears in the product acid. Though a
certain amount may be precipitated in the form of mercury(II) sulfate,
which can be filtered out, for some purposes it is necessary to purify the
acid to a higher standard. One method [91] is to treat the acid in the
drying circuit with thiosulfate, which decomposes to give colloidal sulfur.
This in turn reacts with mercury to give insoluble mercury(II) sulfide.
Because this technique can only be used when the acid concentration is
below 85 %, a preliminary drying tower is installed operating on acid of
70 – 85 % concentration. The main drying tower operates on full-strength
product acid.
7. Production of Pure Sulfur Dioxide

One way of producing pure sulfur dioxide is to chemically reduce sulfur

trioxide with sulfur. However, since sulfur trioxide is itself usually
produced in a sulfuric acid plant processing moderately concentrated
sulfur dioxide-containing gases, physical or chemical separation may
offer a more direct and attractive route to the pure product. From an
economic point of view, the higher the sulfur dioxide concentration is in
the raw gas from which it is to be separated the better. Therefore, gases
produced by combustion of sulfur or in pyrometallurgical processing of
sulfides ores are preferred. But it is possible, by means of regenerative
flue gas scrubbing methods, to produce pure sulfur dioxide from low-
concentration sources such as power station combustion gases. In view
of the very large volumes of these gases that are produced and the
current need in many locations to remove their sulfur dioxide content for
environmental reasons, this method cannot always be dismissed on
simple economic grounds alone.

66
7.1. Chemical Production from Sulfur Trioxide
Pure sulfur dioxide is formed when sulfur and sulfur trioxide undergo the
following weakly exothermic reaction [92]:

(30)

Practical processes using this reaction are distinguished by the form in

which the reactants are converted.

Despite its age, the process developed in Germany by IG

Farbenindustrie [93] is still in use today. Solid sulfur is fed into a reactor
containing oleum at ca. 110 °C.

Simultaneously, preheated oleum with 26 –30 % free SO3 is fed in at the

bottom of the reactor. The oleum rises in countercurrent to the slowly
descending sulfur, leaving the reactor at the top with a residual free SO3
content of 1 – 5 wt %. The sulfur dioxide gas formed in the reactor still
contains 5 – 10 vol % SO3, which is converted to further sulfur dioxide in
a column filled with lump sulfur. The last traces of sulfur trioxide are
finally removed from the sulfur dioxide by scrubbing with 98 % sulfuric
acid. The dilute oleum leaving the reactor flows to the oleum tower of the
sulfuric acid plant for reconcentration. As it contains impurities introduced
in the sulfur, it must be discharged partially and replaced by fresh oleum.

In modern processes, the sulfur trioxide is not fed as concentrated oleum

but in dilute liquid form, and the sulfur is preferably molten. The reaction
medium is either oleum of extremely low SO3 content [94], a liquid
mixture of sulfur dioxide and sulfur trioxide under pressure [95], liquid
sulfur [96] or oleum of 25 % strength [95].

Figure 23 shows the flow diagram of the Schweizerhall process [97].

Molten sulfur and liquid sulfur trioxide in stoichiometric proportions are
agitated in a reactor with oleum of ca. 25 % strength. The reactor bottom
is equipped with a double shell, through which cooling water is pumped
to provide temperature control. In normal operation, only slight cooling is
required. For start-up, steam heating is necessary. The reaction
temperature is adjusted to maintain a sufficiently high pressure of the
sulfur dioxide formed that it liquefies in a water- cooled condenser
without further refrigeration. The sulfur dioxide leaving the reactor is
passed through a secondary reactor filled with solid sulfur, in which the
residual sulfur trioxide is converted completely to sulfur dioxide, before
67
entering the condenser. The sulfur dioxide produced by this process is of
high purity.

Figure 23. Schweizerhall liquid sulfur dioxid process Figure 23.

Schweizerhall
a) Reactor; b) Thermostat; c) Post-reactor; d) Lump liquid sulfur dioxid
sulfur; e) SO2 condenser; f ) Tank for liquefied SO2 process

a) Reactor; b)
Thermostat; c)
Post-reactor; d)
Lump sulfur; e)
SO2 condenser;
f ) Tank for
liquefied SO2
[Full View]

The principal advantages of this process are its relatively low capital cost
and the simplicity of operation. The sulfur dioxide production costs are
essentially determined by the sulfur price and (especially) the cost of
liquid sulfur trioxide, which are normally very high. For 1t of liquid sulfur
dioxide, ca. 20 kWh of electric power and around 30 m3 of cooling water
at 20 °C are required.

7.2. Separation from Sulfur Dioxide-Containing Gases

Condensation processes for the production of pure sulfur dioxide are
based on the separation of sulfur dioxide from gases by liquefaction. In
accordance with the vapor pressure equilibrium between gaseous and
liquid sulfur dioxide (see Properties), low temperatures, or high
pressures, or both are required for this purpose.

In a plant operated by Bayer, roaster gases with 6 – 7 vol % of sulfur

dioxide are compressed to ca. 5 bar and then passed through a number
of countercurrent coolers. These are cooled by liquid ammonia and by
the waste gas of the process, the temperature of which has been
reduced to – 70 °C by stepwise flashing. The sulfur dioxide content of the
waste gas is ca. 0.7 vol % [98]. The energy consumption per tonne of
liquid sulfur dioxide of a plant designed for a production of 25 – 30 t/d of
liquid sulfur dioxide is ca. 680 kWh.
68
The so- called partial condensation process is quite commonly used.
Normally, 30 – 60 % of the fed sulfur dioxide is condensed. Partial
condensation plants are frequently operated in connection with sulfuric
acid plants, as the residual sulfur dioxide content of the depleted gas is
in most cases still relatively high, and so it requires subsequent
treatment.

A simplified diagram of this process is shown in Figure 24. The sulfur

dioxide is partially condensed under normal pressure by cooling in
indirect exchange with an evaporating coolant. The ratio of the volume of
liquid sulfur dioxide produced to the residual sulfur dioxide content in the
gases is determined by the sulfur dioxide vapor pressure at the
condensation temperature, which is controlled in the coolant circuit.

Figure 24. Liquid sulfur dioxide production by partial Figure 24.

condensation Liquid sulfur
dioxide
a) Gas heat exchanger; b) SO2 condenser; c) Expansion production by
valve; d) Coolant compressor; e) Coolant condenser; f ) partial
Pump; g) Expansion valve; h) Coolant expansion vessel; i) condensation
Liquid SO2 tank
a) Gas heat
exchanger; b)
SO2
condenser; c)
Expansion
valve; d)
Coolant
compressor;
e) Coolant
condenser; f )
Pump; g)
Expansion
valve; h)
Coolant
expansion
vessel; i)
Liquid SO2
tank
[Full View]

Dry, sulfur dioxide-rich gases are pre-cooled by the cold residual gases
69
in a heat exchanger and passed on to the condenser. The temperature in
the sulfur dioxide condenser is held between – 70 and – 40 °C by
evaporation of the coolant in the shell. The depleted gases leaving the
condenser are passed on to a separator, which removes entrained
droplets of liquid sulfur dioxide, and are then conveyed to a sulfuric acid
plant or other plant for processing sulfur dioxide after reheating in the
gas – gas heat exchanger.

Fluorinated hydrocarbons are mostly used as coolants. The coolant

vapor leaving the sulfur dioxide condenser is compressed to the
pressure required to restore it to the liquid state after cooling; the exact
pressure therefore depends on the temperature of the available cooling
water. The liquefied coolant is expanded in several stages to the
pressure corresponding to the desired evaporation temperature before
being recycled to the sulfur dioxide condenser.

The economics of the partial condensation process are essentially

determined by the sulfur dioxide content of the gases; 16 – 18 vol % is
the ideal concentration. In a partial condensation plant with 12 vol % SO2
in the inlet gas and 6 vol % SO2 in the residual gas, ca. 260 kWh of
power and 90 m3 of cooling water at 20 °C are consumed per tonne of
sulfur dioxide liquefied [97].

On account of its economy and operational reliability, the partial

condensation process is one of the most frequently used methods of
recovering pure sulfur dioxide. It is especially suited for large production
capacities. The largest plant built by Lurgi so far, for example, is
designed for the production of 160 t/d of liquid sulfur dioxide.

Absorption Processes. In absorption processes, sulfur dioxide is

separated from the SO2-containing gases by reversible absorption into a
liquid. This may either be a solvent, in which it dissolves physically, or a
scrubbing solution containing a carefully selected active component, to
which it becomes loosely and reversibly chemically bound. It is then
regenerated in concentrated form by subjecting the rich solution to
conditions of temperature and pressure under which the sulfur dioxide is
desorbed.

Absorption in Water. One of the oldest absorption processes is

atmospheric water washing according to the Hänisch – Schröder process
of 1884. It is still used today in a modified form, but usually only to
recover a part of the sulfur diox-ide content of the gas being processed;
therefore, it must be associated with some ancillary process to treat the
residual gases. For satisfactory operation the crude gas should contain
70
at least 10 –11 vol % SO2.

The sulfur dioxide-containing gas is fed into the base of a packed tower
and rises in countercurrent with water introduced at the top. The water
leaving the bottom of the tower may contain 1 – 2 wt % of sulfur dioxide,
depending on the temperature. It is passed to a desorption tower, in
which the sulfur dioxide is driven off by low-pressure steam. The water at
the exit of the desorber still contains ca. 0.01 wt % of sulfur dioxide. The
recovered mixture of sulfur dioxide and water vapor is passed through a
water-cooled condenser to remove as much water as possible, after
which the sulfur dioxide is dried with sulfuric acid.

The process has relatively high utility requirements: for partial

absorption, it uses about 120 m3 of process and cooling water, 3 t of
steam and 90 kWh of power per tonne of liquid sulfur dioxide [97].

A pressurized sulfur dioxide water wash (the Grillo system) is

schematically represented in Figure 25. Purified gas with at least 9 vol %
SO2 is compressed to ca. 5 bar in a compressor. To dissipate the heat of
compression, the gas is cooled to about 50 °C in heat exchangers. It is
then scrubbed in a water-irrigated packed tower, in which its sulfur
dioxide content is reduced to < 0.1 vol %. The residual gas is heated
against the compressed fresh gas and subsequently expanded to
recover up to 35 % of the compressor energy.

Figure 25. Production of pure sulfur dioxide by Figure 25.

pressurized water absorption (Grillo system) Production of
pure sulfur
a) Compressor; b) Expansion turbine; c) Gas heat dioxide by
exchanger; d) Absorber; e) Liquid heat exchanger; f ) SO2 pressurized
stripper water
absorption
(Grillo system)

a)
Compressor;
b) Expansion
turbine; c) Gas
heat
exchanger; d)
Absorber; e)
Liquid heat
exchanger; f )
SO2 stripper
71
 [Full View]

The sulfur dioxide-laden wash water extracted from the absorption tower
is first preheated in a liquid heat exchanger and then passed to a
desorption column, in which the sulfur dioxide is stripped by injection of
low-pressure steam. The column is connected at the top with a
dephlegmator in which the water vapor is disengaged from the sulfur
dioxide by irrigation with a partial stream of cold, sulfur dioxide-laden
water from the absorption tower. In some cases, instead of the
dephlegmator, a separate condensation tower irrigated with water from
an indirectly cooled circuit is used. The cooled sulfur dioxide is dried in
the same way as in the atmospheric water wash. The wastewater leaving
the desorption column sump is used to indirectly preheat the incoming
rich scrubbing solution before being discharged.

Even though part of the compression energy is recovered in the

expansion unit, the consumption of electric power is extremely high in
comparison with the atmospheric water wash. Production of 1 t of liquid
sulfur dioxide from a raw gas containing 10 vol % of sulfur dioxide
consumes ca. 300 kWh [97]. Furthermore, ca. 1 t of low-pressure steam
and ca. 50 m3 of processed cooling water are required as well.

Both the pressurized and atmospheric water wash processes have the
advantage of being able to run on almost any quality of water (except
water containing volatile materials).

Absorption in Alkaline Solutions. Alkaline solutions absorb much more

sulfur dioxide than water. In some instances strong bases such as
caustic soda and ammonia are used, but to an increasing extent these
are being replaced by alkaline salts of weak inorganic and organic acids
(e.g., sodium carbonate or citrates), or (occasionally) aliphatic or
aromatic amines. This is because regenerating the sulfur dioxide is
easier and less energy-intensive with these weaker bases. The principle
of the process corresponds essentially to that of the atmospheric water
wash, except that the absorbing solution is always recirculated from the
desorber to the absorber.

In the Wellman – Lord process [99], [100], the sulfur dioxide is extracted
by means of an aqueous solution of sodium sulfite and sodium hydrogen
sulfite. The absorbent is regenerated, and concentrated sulfur dioxide is
released by indirect heating with steam in a separate step, which
thermally decomposes the sodium bisulfide. The process is chiefly used
for desulfurizing flue gases from power stations and tail gases from
72
Claus sulfur plants (see  Sulfur – The Claus Process), but in a few
cases it has also been used for purifying the waste gas from contact (i.e.,
single-absorption) sulfuric acid plants.

Figure 26 shows a flow diagram of a typical Wellman – Lord plant for flue
gas desulfurization. Flue gas which has already been passed through dry
electrostatic precipitators or other dust control apparatus is sprayed with
water in a saturator section at the bottom of the absorber. This not only
cools the gas adiabatically by evaporation of some of the water but also
removes any remaining particulates. A bleed stream of the scrubbing
water is extracted from the circuit so as to keep the solids content in the
circulating liquor down to ca. 5 % and is fed to the wastewater-treatment
system. The gas next enters the main absorption zone and is washed in
countercurrent with a saturated aqueous solution of sodium sulfite and
sodium bisulfite, which is recirculated in several steps. At the top of the
absorber, a droplet separator arrests entrained droplets. After reheating
to about 140 °C, the moisture-laden flue gas is discharged to the stacks.

Figure 26. Wellman – Lord process for sulfur Figure 26.

dioxide production Wellman – Lord
process for sulfur
a) Absorber; b) Crystallizer; c) Centrifuge; d) Dryer; dioxide production
e) Forced- circulation evaporator/crystallizer; f )
Condenser; g) Thickener; h) Dissolution tank a) Absorber; b)
Crystallizer; c)
Centrifuge; d) Dryer;
e)
Forced- circulation
evaporator/crystalliz
er; f ) Condenser; g)
Thickener; h)
Dissolution tank
[Full View]

73
The principal mechanism by which sulfur dioxide is absorbed is its
interaction with sodium sulfite to form sodium hydrogensulfite

(31)

In a side reaction, part of the sulfite is oxidized to sulfate by residual

oxygen present in the flue gas. Sulfate is also formed to a lesser extent
by reaction of the scrubbing solution with traces of sulfur trioxide in the
flue gas.

The hydrogensulfite-rich scrubbing liquid is regenerated in a forced-

circulation evaporator, indirectly heated by steam. There, the
hydrogensulfite is thermally decomposed by the reverse of the
absorption reaction (Eq. 31). Some of the moisture content of the
saturated sulfur dioxide liberated in the evaporator is removed by indirect
cooling with water in the subsequent condenser.

Because the solution becomes oversatured in the evaporator, sodium

sulfite is precipitated in crystalline form. The suspension of crystals
passes to a thickener, the overflow from which is withdrawn as waste
liquor. The slurry of crystals from the underflow is diluted in a tank with
condensate and fresh water to extent necessary to turn it back into a
saturated sulfite – bisulfite solution. To compensate for bleed losses,
sodium hydroxide or carbonate is added.

The sulfates formed in the side reactions mentioned above are not
decomposed in the evaporator. To prevent their concentration rising to
values at which the absorptive capacity of the solution is impaired and
there is a risk of clogging by accumulations of precipitated solids, part of
the scrubbing solution is continuously withdrawn from the circuit [99].
The water solubility of the sodium salts in the scrubbing system
decreases in the order bisulfite, sulfite, sulfate. The bleed stream is
therefore taken off directly after the sulfur dioxide absorber, where the
bisulfite concentration is the highest and the sulfite content is the lowest,
and is passed through a water-cooled crystallizer, in which crystals
consisting mainly of sodium sulfate are precipitated. The sulfate-rich
solid is separated in a centrifuge from the residual solution, which is
returned to the absorption circuit. The separated solid is dried by indirect
steam heating and discharged from the system for further processing
[99].

Other Absorption Processes. Sulfur dioxide absorption processes using

74
aqueous amine solutions have the advantage over processes using
sodium sulfite – hydrogensulfite solutions that the amine solutions can
absorb more sulfur dioxide and the rich solutions can be regenerated
more easily. That is why, in former times, aliphatic and also aromatic
amines were used preferably. However, amines are disadvantageous in
respect of their vapor pressure, which is high enough for appreciable
amounts to be discharged from the absorption system in the cleaned
waste gas. Such emissions are today considered detrimental to health,
quite apart from having an objectionable odor.

The Sulfidin process, commonly used in the past, employed an

absorbent consisting of an aqueous solution of xylidine and toluidine. It is
no longer used for environmental reasons.

In the United States, a process developed by American Smelting and

Refining Co., which uses N,N-dimethylaniline, is still used [101].

The UCAP process of Union Carbide Corporation was specifically

developed for cleaning the tail gases from Claus plants, to which the
pure sulfur dioxide is recycled. The absorbent used is an aqueous
solution of triethanolamine [102].

During the development of flue gas and exhaust air cleaning processes
physical absorption processes have been introduced. An example of
such a process for the removal of SO2 from flue gases or other SO2-
containing gas steams is the Solinox process introduced by Linde, in
which poly(ethylene glycol) dimethyl ethers are used as absorbent. The
SO2 is recovered in form of a 80 – 90 % SO2 gas [103].
8. Construction Materials

Dry gases containing sulfur dioxide — and even those containing water
vapor as well, if at a temperature safely above the dewpoint — barely
attack the steel grades used normally.

Standard steel St 37.2 (material no. 1.0114) and boiler plate HII (material
no. 1.0425) are generally preferred. For extreme stress conditions, St
35.8 (material no. 1.1035) or low-alloy steel grades such as St 45.8
(material no. 1.0405) and 15 Mo 3 (material no. 1.5413) are used. Above
550 °C, heat-resistant ferritic (e.g., material no. 1.4713) or austenitic
steels (e.g., material no. 1.7841 and 1.7828) are used. On account of
their sensitivity to possible traces of sulfur at the furnace outlet, nickel-
rich steel grades are not appropriate, especially at elevated
temperatures.

75
For moist sulfur dioxide-bearing gases, especially in gas cleaning plants,
lead was the classical material. It can be used at temperatures up to
110 °C and is also resistant to sulfuric acid at concentrations up to ca.
80 %. Lead (DIN 1719) is mainly used in the form of refined copper lead
(Pb 99.9 %, Cu 0.05 – 0.08 %; material no. 2.3040) as sheet, castings,
or homogeneous lead coatings on steel.

Where there is likely to be thermal cycling stress, antimony-alloyed lead

(DIN 28 058) is preferable. The limited creep resistance of lead and lead
alloys must be compensated by adequate support. In homogeneous lead
coatings, the adhesion of the lead layer is frequently enhanced by initially
tinning the steel surface before the lead coating is applied.

Nowadays plastics are normally applied for piping and ducting and for
certain other equipment for wet SO2 gas at moderate temperatures.
Plastics such as polyvinyl chloride (PVC), rigid polyethylene (rigid PE)
and polypropylene (PP) are chemically resistant to sulfur dioxide, dilute
sulfuric acid, and other inorganic compounds which may occur as
impurities in the sulfur dioxide gas cleaning system. But with plastics,
factors like ageing by sunlight, cold embrittlement, and temperature
cycling have to be taken into account. The upper temperature limit for
practical application is ca. 60 °C for PVC, ca. 80 °C for rigid PE, and ca.
90 °C for PP, but for composites of these materials with fiberglass-
reinforced plastic (FRP), the upper temperature limits are somewhat
higher: for FRP/rigid PE, 100 °C, and for FRP/PP, ca. 120 °C.
Fluorinated plastics (with or without reinforcement) can be used up to ca.
120 °C.

For wash acids containing sulfur dioxide, lead, rigid polyethylene, and
polypropylene are used for piping, pumps, pump tanks, valves, and
fittings. With lead, pitting corrosion is likely to occur at higher chloride or
mercury contents.

For sulfur dioxide-saturated, dilute wash acids, plastics and silicon-

containing cast iron are most often used for valves, fittings, and pumps.

Commercial acid-resistant bricks are used for lining Venturi scrubbers,

scrubbing towers, and vessels. For highly fluoride-laden systems, carbon
bricks are used.

Impregnated graphite, especially graphite impregnated with phenolic

resin, is normally limited to special applications in heat exchangers,
valves, fittings, and pumps.

76
Special stainless steels are suitable for use in scrubbers, heat
exchangers, etc. provided well-defined and stable conditions are
maintained. This refers specifically to impurities such as HCl, HF and
reducing components, which may disturb the passivating layer on the
steel.
9. Industrial Output and Uses

Production for Sulfuric Acid Manufacture. The majority of the sulfur

dioxide produced industrially in the world is directly used for sulfuric acid
production. Sulfur dioxide requirements for other purposes are relatively
small, and they are usually satisfied either by combustion of sulfur at the
consumer's works or by pure sulfur dioxide supplied in liquid form. In
1979, for instance, in the Federal Republic of Germany, the consumption
for sulfuric acid manufacture was about 3.3×106 t SO2, whereas other
uses accounted for only about 40 000 t [104]. In the same year the
United States used ca. 23.6×106 t of sulfur dioxide for sulfuric acid
production and only about 150 000 t for other purposes [105].

Miscellaneous Uses. Apart from sulfuric acid manufacture, sulfur dioxide

has uses in the following industrial areas [106], [107].
Paper and pulp industry ( Paper and Pulp – Sulfite Process)
Production of cooking solutions for the sulfite pulping process
Reduction of sodium chlorate in the production of chlorine dioxide for
pulp bleaching
Pulp treatment after the first (peroxide) stage of a two-stage bleaching
process to eliminate hydrogen peroxide before hydrogensulfite
treatment
Chlorine scavenger, especially in sulfate pulp factories after the
chlorine bleaching step

Chemical and pharmaceutical industries

Production of sulfites, hydrogensulfites, disulfites, thiosulfates,
dithionites, and hydroxyalkanesulfinic acid salts
Production of alkane sulfonates by sulfoxidation of sulfochlorination
Production of gelatins and glues
Reductant or solvent
Chemical processing of manganese ores and recovery of, for instance,
Co, Se, and Te from waste

Petrochemical and mineral oil industries

Solvent for separation of aromatics from hydrocarbon mixtures
(Edeleanu process)
Removal of sulfur-containing impurities from mineral oils
77
 Deodorization of lubricating greases

Textile and leather industries

Bleaching agent

Food and beverage industries

Preserving agent and disinfectant (e.g. sterilization in breweries,
fermentation control, and preservation in viniculture)

Sulfur dioxide is also required for water and wastewater treatment. Liquid
sulfur dioxide was formerly utilized for refrigerating machines and plants.
However, in this area of application it has been replaced almost
completely by fluorinated hydrocarbon refrigerants.

10. Analysis

Analytical techniques for sulfur dioxide are covered in literature reviews

[108-110].

Qualitative Detection. Sulfur dioxide can be recognized even in extreme

dilution by its pungent smell. Solutions of sulfur dioxide, sulfites, or
hydrogensulfites decolorize iodine and permanganate on account of their
reducing effect. Sulfur dioxide and sulfites are reduced by zinc in a
hydrochloric acid solution to hydrogen sulfide, which is readily detected
by its smell.

Quantitative Analysis. Sulfur dioxide is determined chemically by

iodometry, titrimetry, gravimetry, or colorimetry. On account of the large
number of quantitative legal restrictions relating to sulfur dioxide that are
now in force, accuracy and reliability are a prerequisite of any analytical
method, and definitive procedures and equipment are often laid down by
official bodies. In Germany, for example, they are governed by VDI
regulations [111-115].

Continuous Monitoring. Routine in-plant performance monitoring usually

entails continuous measurement of the sulfur dioxide content of, for
instance, roaster gases, contact gases, and tail gases in a sulfuric acid
plant by automatic recording analysers, based on a physical property
such as spectroscopic absorption in the infrared [114], [116], or
ultraviolet [117], or electrical conductivity [115]. Strict adherence to set
sampling procedures is essential for accuracy and reliability of in-plant
monitoring.

78
Iodometric Method. REICH's classical, manual iodometric determination
is still widely used in industrial practice. A partial stream of the sulfur
dioxide- containing gas is drawn through an iodine solution with starch as
an indicator until it is discolored. The sulfur dioxide concentration is
calculated from the gas volume and the amount of iodine originally
present [108]. In another version of the method [111], an excess of
iodine is used, and a predetermined volume of the gas is drawn through
the solution. The unused iodine is then determined by titration with
sodium thiosulfate after addition of starch. Nitrogen oxides interfere with
readings from the Reich method; this disturbing influence can, however,
be avoided by addition of acetate [108] or aminosulfonic acid [118].

Titrimetric and Gravimetric Methods. For titrimetric or gravimetric

determination of sulfur dioxide, the gases to be analysed are passed
through a hydrogen peroxide solution, which oxidizes sulfur dioxide to
sulfuric acid. This sulfuric acid is either titrated with caustic soda solution
against bromophenol blue [112] or is precipitated as barium sulfate and
weighed [113]. The analyses are disturbed by other acid or
sulfur- containing components. If sulfur trioxide or sulfuric acid are
present in the gas, they are removed first by absorption in concentrated
sulfuric acid in a wash bottle. On account of the solubility of sulfur dioxide
in sulfuric acid, this wash acid must first be saturated with sulfur dioxide.

Colorimetric Method. In the colorimetric determination of sulfur dioxide

according to WEST and GAEKE the gas is passed through a solution of
sodium tetrachloromercurate [119]. After addition of formaldehyde and p-
rosaniline, the intensely coloured p-rosaniline methylsulfonic acid is
formed in a Schiff reaction. The sulfur dioxide concentration is
determined by a comparison of the extinction at 548 nm with a calibration
curve based on standard solutions.
11. Environmental Aspects and Toxicology

Environmental Aspects. A substantially larger amount of sulfur dioxide

than is utilized industrially is produced by the combustion of sulfureous
fossil fuels and is discharged into the atmosphere with the flue gases.
The calculated total emission of sulfur dioxide from power stations,
traffic, households, industry, and trade in the Federal Republic of
Germany in 1975 amounted to ca. 3.6×106 t [120]. In contrast, only about
2.7×106 t were used for that year's sulfuric acid production of ca.
4.2×106 t H2SO4 [121]. In the last years the SO2 emissions from power
plants in Germany have been reduced continuously with the installation
of flue gas desulfurization plant and with the utilization of low-sulfur fuels,
but emissions from traffic, households, industry, etc. in 1990 still account
for 620×103 t/a. Emissions from power stations amounted to 320×103 t/a.
79
For worldwide atmospheric emissions, only rough estimates are
available. For 1970, while global emissions were estimated at 157×106 t
SO2 [122], only about 61×106 t were consumed in the production of ca.
94×106 t of sulfuric acid [121].

Large-scale emission of sulfur dioxide close to ground level has

indisputably been the cause of some spectacular environmental
problems in the past. A large area around Sudbury, Ontario, was
completely sterilized by sulfur dioxide from primitive ore roasting
operations around the turn of the century, and much of it is still barren as
a result of the ensuing soil erosion. Calamitous sulfuric acid fogs
occurred in Donora, Pennsylvania, and London, England, in 1948 and
1952, when adverse climatic conditions prevented sulfur dioxide from
industrial sources and domestic coal fires from dispersing. The causes of
these episodes were correctly identified, and the action taken to control
ground level smoke and sulfur dioxide concentrations has effectively
prevented them from recurring. In the case of industrial sources,
however, this has often been achieved by merely building taller chimneys
to disperse the sulfur dioxide over a wider area, and as the consumption
of fuels has increased over the years it has been argued that the
problem has merely been shifted from the location where the sulfur
dioxide is produced to other locations downwind.

Considerable publicity has been given to the apparently increasing

problem of acid rain, and there does, indeed, seem to be a correlation
between the incidence of acid rainfall and the atmospheric transport of
pollutants from major industrial locations [123-127]. Establishing in any
degree of certainty to what extent sulfur emissions are responsible for
acid rain is, however, complicated by incomplete knowledge of the
magnitude of sulfur emissions from natural sources, the atmospheric
chemistry of sulfur, and the importance of other potential acidulants such
as nitrogen oxides. It is also, unfortunately, hampered by politics,
particularly in those cases where there is a possibility that pollution from
one country could be contributing to acidic rainfall in another [128], [129].
Much of the research is being conducted by or for national agencies
which have the power to publish or play down the results as may be
expedient. No government purporting to be dedicated to the topical
causes of controlling inflation and combatting unemployment at home will
lightly order a major programme of sulfur dioxide emission control, with
its almost devastating economic implications, just for the purpose of
ameliorating the environment in some other country if there is yet room
for argument. Nevertheless, the evidence linking sulfur dioxide pollution
to acid precipitation has been mounting gradually [130], [131]. A report
by the U.S. National Academy of Sciences, for example, has stated that
80
there appears to be a direct proportionate relationship between sulfur
dioxide pollution and the amount of acidic sulfates in precipitation [132].

Toxicology. Ill-effects of sulfur dioxide on humans and animals are

mainly related to irritation and damage of the moist mucous membranes
by the formation of sulfurous acid. The odor threshold of sulfur dioxide in
air is between 0.3 and 2.5 ppm. In most human beings, concentrations of
5 – 10 ppm will lead to irritation of the respiratory tract; in sensitive
people they may produce spasms of the bronchi. Higher concentrations
will cause heavy irritative coughing, while breathing sulfur dioxide at
concentrations above 400 – 500 ppm, even for only a few minutes, is
dangerous to life. The maximum workplace concentration is defined as
5 ppm (14 mg/m3).

Long-term exposure may possibly lead to bronchopneumonia and, in

extreme cases, to toxic pulmonary edema with dyspnoea, cyanosis and
cardiac-circulatory failure. Short-term exposure to extremely high
concentrations may cause glottal spasm, gasping or inability to breathe.
Symptoms experienced after long-term exposure are loss of appetite,
constipation, loss of taste, red tongue, convulsions, eye inflammation,
emphysema and nasal and bronchial mucus secretion, which may be
blood-stained. Contact with liquid sulfur dioxide causes extremely painful
conjunctivitis, conjunctival edema and opacity of the cornea in the eye as
well as frostbite on skin and mucous membranes.

Aqueous sulfur dioxide solutions (sulfurous acid) may burn eyes and
skin. Swallowing, which is rather improbable on account of the intense
warning smell, may cause irritation or burns in the esophagus and
stomach, depending on the quantity and concentration swallowed,
followed by nausea, gastric pains, vomiting, and diarrhoea. Sulfurous
acid absorbed by the body is converted to sulfate and discharged in the
urine, as evidenced by an increase in urine acidity.

Effects on Vegetation. Plants are impaired by sulfur dioxide at even

lower concentrations than human beings. Sulfur dioxide levels of
1 – 2 ppm are enough to cause acute damage to the leaves of plants in
only a few hours, followed by necrosis, probably resulting from impaired
photosynthesis. For protection of the environment, the BimSchG law in
Germany lays down the following emission values: 80 µg/m3 related to a
dust value of more than 150 µg/m3, and 120 µg related to a dust value of
less than 150 µg/m3.

Further information on the toxic effects of sulfur dioxide is available in

publications [107], [133-136].
81
12. Acknowledgement

Based on the corresponding article in Ullmann, 4th ed., written by

ULRICH SANDER and ULRICH ROTHE (English Edition: British Sulphur
Corporation, 1984)

82