Zevenhoven & Kilpinen SULPHUR 13.6.

2001 3-1

Chapter 3 Sulphur
3.1 Introduction

Sulphur is probably the “easiest” pollutant from a gas clean-up point of view. A first
reason for that is certainly that the chemistry involved is relatively simple and (hence)
well understood. In combustion plants sulphur in fuel will eventually be oxidised to
sulphur dioxide SO2 and some sulphur trioxide SO3, be it that small amounts may be
bound by ash-forming minerals as sulphates. At the same time, a long history of SO2
capture with limestone and other calcareous rocks exists, which is accomplished at
relatively low costs due to the abundance of these materials in nature. Hence, it is no
surprise that in response to stricter legislations introduced during the 1970-1980s,
significant sulphur oxides (“SOx”) emission reductions of were rapidly implemented
(Ponting, 1991). In Europe, sulphur emissions dropped by 53% since 1980 (Acid
News, 2000a). This in contrast with the ongoing struggle to reduce nitrogen oxides
(NOx) emissions from the same facilities.

In gasification or pyrolysis plants hydrogen sulphide H2S will be the most important
sulphur compound. Methods for its control have been developed first by the petro-
chemical industry, while later methods based on (regenerable) sorbents were specially
developed for coal gasification units. Upon release to the environment, H2S will be
oxidised to SO2.

One fact remaining, though, is that sulphur must eventually be disposed of in some
chemical form. Unlike fuel-related nitrogen, that without any environmental concern
may be released partly as molecular nitrogen N2, there is no such “escape route” for
sulphur. Fuel-related sulphur that is captured may end up as molecular sulphur,
sulphuric acid, or bound in flue gas desulphurisation (FGD) residues like gypsum, and
partly in bottom ashes and fly ashes. In other cases, such as cement kilns, the cement
product may contain a significant amount of the sulphur from the fuels used (coals,
petcoke, alternative fuels) without causing problems.

For utilities and power plants firing sulphur-containing solid fuels SO2 capture can
be implemented with guaranteed success, but a different situation exists for the fuels
derived from oil. As a result of tight limits on sulphur contents of transport fuels like
gasoline and light fuel oils, it is difficult for the petro-chemical industry to find an
outlet for the sulphur that is obtained from oil refining. Currently much of that
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-2

sulphur is found in heavy fuel oils used on ships without emissions control, that is,
at those seawaters where high-sulphur fuels for shipping are accepted ("Cleaner
Shipping", 1997) (L section 3.15).

Sulphur emissions to the environment may be divided in being the result of human
activities, basically the combustion of fuels for heat and power or transport, plus a
contribution from natural sources. Volcanoes emit significant amounts of sulphur as
well. Emissions of sulphur dioxide are correlated with deposition of sulphur
compounds, often in the form of dilute sulphuric acid, part of which is known as
“acid rain”. Figures 3.1 and 3.2 give an impression of the situation in Europe during
the last decade. It is seen that the sulphur emissions result in deposition mainly in the
densely populated and industrialised “band” from England via Germany to Central
and South-East Europe. The countries in Scandinavia can be considered as importers
of sulphur-related pollution, whilst countries like Spain, Portugal, Ireland and Russia
appear to export most of their sulphurous pollutants.

Figure 3.1 Annual emissions of SO2 in Europe in 1994, in Figure 3.2 Annual depositons of
tonnes (picture from Ågren and Elvingson, 1997) sulphur in Europe in 1994, in mg S/m2
(picture from Ågren and Elvingson,
1997)

Reasons for controlling SO2 emissions are related to its contribution to acid rain, and
more recently it was recognised that sulphate-containing aerosols form a health
hazard as well, causing problems to the respiratory and cardio-vascular systems. With
oil and gas processing, and gasification processes H2S is the most important sulphur
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-3

species to be controlled, being not only highly poisonous and bad smelling, but also
resulting in SO2 after oxidation in the environment.

In this chapter, the routes from sulphur present in fuels to pollutants and methods
to avoid these from being emitted to the atmosphere are discussed. After a short
discussion on in what forms sulphur occurs in fuels, the essentials of the chemistry
of sulphur during combustion and gasification are given. This is followed by some
emission standards for SO2, for Finland and abroad. A general summary of
approaches to sulphur control is given, which is followed by a rather extensive
treatment of the most important technique: flue gas desulphurisation (FGD) for SO2.
Wet, semi-dry and dry FGD with non-regenerable or regenerable sorbents are
discussed. High temperature SO2 capture is discussed addressing furnace sorbent
injection and in situ SO2 control during fluidised bed combustion (FBC). After this the
point of focus switches to H2S. First, low temperature processes for H2S removal
from gases are considered, followed by the famous Claus process. Methods based on
H2S capture with calcium-based and regenerable sorbents as used in relation to (coal)
gasification, sometimes refered to as high temperature gas clean-up for sulphur, are
discussed next. After that the control of sulphur compounds emissions from
transport vehicles is addressed.

3.2 Sulphur in fuel

In solid fossil fuels, such as coals ranking from brown coals and lignites up to
anthracite, peat, oil shales etc., sulphur is present in two inorganic forms, being pyritic
sulphur (FeS2), and sulphates (Na2SO4, CaSO4, FeSO4) and as organic sulphur.
Organic sulphur is bound in these fuels as sulphides, mercaptanes, bisulphides,
thiophenes, thiopyrones etc., see Figure [Link] organic compounds are also found
in unrefined oil and heavy fuel oils. In natural gases as well as in raw fuel gases from
gasification processes sulphur is generally present as hydrogen sulphide (H2S) and
carbonyl sulphide (COS).

Biomass fuels such as waste wood, straw, bagasse (residual from sugar cane
processing) and energy crops (Salix, Miscanthus, etc.) contain very small amounts of
sulphur. The relatively large amount of sulphur in bark is apparently correlated with
the exposure of this part of the tree to the surrounding atmosphere and occasional
acid rain and other downfall. In waste-derived fuels, sulphur can be found in car-tyre
scrap and electronic scrap as a vulcaniser, in food-derived wastes (onions and garlic)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-4

Figure 3.3 Typical sulphur compounds found in fuels

etc. Specific for pulp and paper processing is black liquor, for which the large sulphur
content, mainly as Na2SO4, is recovered in reduced form as Na2S in so-called recovery
boilers (Hupa, 1993). Typical values for the sulphur content of various fuels are given
in Table 3.1.

Table 3.1 Typical values for the sulphur content of fuels (%-wt, dry)

Fossil fuels Biomasses & waste derived fuels

Coal 0.2 - 5 Wood < 0.1
Bark <2
Oil 1-4 Straw ~ 0.2
Natural gas 0 - 10
Light fuel oil < 0.5 Sewage sludge ~ 0.2
Heavy fuel oil <5 Car tyre scrap ~2
Municipal solid waste (MSW) 0.1 - 1
Refuse derived fuel (RDF) 0.1 - 1
Peat <1 Packaging derived fuel (PDF) ~ 0.2
Auto shredder residue (ASR) ~ 0.3
Petroleum coke, “petcoke” ~5 Leather waste 1-2
Estonian oil shale ~2
Orimulsion™ ~4 Black liquor solids ~5
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-5

3.3 Formation of sulphur compounds during combustion and gasification

In short, the formation of sulphuric pollutants during combustion and gasification

can be described by the following chemistry :

Fuel devolatilisation / “pyrolysis”

Fuel-S (s) + heat 6 H2S + COS + ... + Char-S (s) (R3-1)

Char oxidation
Char-S (s) + O2 6 SO2 (R3-2)
Char-S (s)+ CO2 6 COS (R3-3)
Char-S (s) + H2O 6 H2S (R3-4)
Gas phase reactions
H2S + 1½O2 6 SO2 + H2O (R3-5)
CO2 + H2S W COS + H2O (R3-6)
CO + H2S W H2 + COS (R3-7)
COS + H2S W CS2 + H2O (R3-8)
CS2 W C(s) + 2/x Sx (s) (R3-9)
Sulphuric oxide formation
SO2 + ½ O2 W SO3 (R3-10)
H2O + SO3 W H2SO4 (R3-11)

The reactions involving O2 lead to the major sulphuric product during combustion,
SO2 via (R3-2) and (R3-5). Although the thermodynamics of (R3-10) are in favour
of SO3, the oxidation of SO2 to SO3 is a slow chemical reaction. It requires
temperatures above 1100EC if it is to proceed as a gas phase reaction, or a catalyst
such as oxides of vanadium, nickel and iron for the reaction to become important
already at temperatures as low as 500EC.

This catalytic effect explains why during the early 20th century when oil-derived
transport fuels were introduced corrosion problems arose that were not seen that
strongly with coal. The sulphur in the transport fuels was oxidised to SO2 and further
to SO3, catalysed by traces of metallic elements in the fuel, which did not occur that
strongly with coal. Upon cooling, this SO3 which had bound with water to form
H2SO4, formed highly corrosive liquid sulphuric acid. The “dew point” of H2SO4, i.e.
the temperature at which condensate is formed is determined by the concentrations
of water and SO3 in the gas. Figure 3.4 (Backman et al., 1983) gives the dewpoint of
sulphuric acid.
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-6

Figure 3.4 The temperature at which sulphuric acid

condensation occurs in flue gas as function of SO3
and water concentration (after Backman et al., 1983)

For gasification processes the situation is more complicated. The most important
sulphur species will be H2S, which is in equilibrium with certain amounts of COS

Figure 3.5 Carbon, hydrogen and sulphur species from combustion (air factor
>1) and gasification (air factor < 1) according to thermodynamic
equilibrium calculations (after Iisa, 1992, 1995)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-7

(carbonyl sulphide), CS2 and elemental sulphur Sx (with S8 as a common form) via
reactions (R3-6,7,8,9). H2S control is one option here, besides removing the sulphur
as SO2 after an oxidation if that occurs downstream in the process, via reaction (R3-
5). Figure 3.5 gives some results of thermodynamic equilibrium calculations for
sulphur compounds in combustion (air factor$1) and gasification (0<air factor<1)
reactors (Iisa, 1992, 1995). As mentioned above, the SO3 levels shown generally do
not occur due to slow SO2 oxidation kinetics, leading to slightly higher SO2 levels than
what thermodynamics indicate.

3.4 Emission standards for SO2

The maximum allowable SO2 emissions are dependent on the location, the type of
fuel and the size and type of the facility from which they arise. Tabled data is
presented and frequently updated by various authorities, following changes which, in
general, show lower allowable emissions with each update. Typically, new and existing
facilities are distinguished in order to reduce economic burden on older facilities. For
coal (and peat), SO2 emission standards for Finland (1995) and the European
Community (1988) are given in Tables 3.2 and 3.3. The World Bank suggests a
worldwide emission limit for all new coal-fired units of 2000 mg/m³STP (dry) @ 6 %
O2, or 500 tonnes SO2 /day (Soud, 1991, McConville, 1997).

For waste firing, the SO2 emission standard for Finland (as of 1.8.1994) is 50
mg/m³STP (dry) @ 10 % O2 (Finland, 1994). The same value recently came into force
for the EU15 countries for waste firing as well as cement plants.

Table 3.2 SO2 emission standards for Finland (1995)

Type of plant New / Plant size Emission standard Comments

Existing (MWth) (mg/m3STP dry 6% O2)
Combustor, peat New > 300 378
Utility, hard coal New < 50 2150
Utility, hard coal New 50 -150 620
Utility, hard coal New > 150 380
Combustor, New and all sulphur content # 1 % plant without de-
hard coal Existing sulphurisation equipm.
Utility , hard coal Existing < 200 2105 lifetime > 15000 hours
after 1.1.1994
Utility, hard coal Existing > 200 620 ibid.
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-8

Table 3.3 SO2 emission standards for the European Community (1988)

Type of plant New / Plant size Emission standard Comments

Existing (MWth) (mg/m3STP dry 6% O2)
Combustion, coal New * 50-100 2000
Combustion, coal New * 100-500 2400 - 4×P P = plant size in MWth
Combustion, coal New * > 500 400
Combustion, coal New * $ 400 800 operation < 2200 h/yr
Combustion, coal New * 100 - 166 40 % removal domestic high or
variable sulphur coal
Combustion, coal New * 167 - 500 (15+0.15×P) % removal P = plant size in MWth
domestic high or
variable sulphur coal
Combustion, coal New * > 500 90 % removal domestic high or
variable sulphur coal
* construction licence after July 1 1988

3.5 Options for sulphur emissions control

Several routes are open for reducing the sulphur emissions from thermal processes
that fire or treat sulphur-containing fuels or materials.

1. Substitute the fuel.

It is obvious that replacing a high-sulphur fuel with a low-sulphur or sulphur-free fuel
is an option that should never be forgotten. This depends on the fuel-flexibility of the
unit, the economics of operation with another fuel and the adjustments that may be
needed, and the contracts with the fuel (or whatever is thermally treated) supplier. It
must be noted that a fuel switch may have more benefits, for instance less ash
formation or lower taxation when e.g. a renewable fuel is chosen. At the same time,
it should not be forgotten that sulphur may have positive effects on ash behaviour in
boilers and furnaces, and on the performance of an electrostatic precipitator (ESP)
for dust emissions control (L Chapter 5). In some cases, fuel substitutions may also
be enforced by local authorities.

2. Fuel cleaning.
A second method that prevents the sulphur from entering the facility is to apply fuel
pre-treatment. Inorganic fuel sulphur, especially pyritic sulphur, FeS2, can be relatively
easily removed to up to 90% by physical methods based on gravity, flotation,
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-9

agglomeration or flocculation. Chemical methods may be able to remove organic fuel

sulphur as well, involving treatment with alkaline or caustic solutions, oxidative
leaching or chlorinolysis with chlorine-based chemicals. The price for these chemical
methods is that part of the material is oxidised, which reduces its potential for use as
a fuel. A third method is biological treatment, capable of removing inorganic a well
as organic sulphur. It was found that Thiobacillus ferrooxidans converts FeS2 into water-
soluble FeSO4,whilst a mutant of Pseudomonas, called CB1 (coal bug 1) consumes
organic sulphur in thiophene groups. These techniques may require very small
particles, are time-consuming, and may need close process control. All methods
mentioned are quite expensive and still may have a limited effect. On the other hand
many low-grade coals are “washed” in order to reduce their mineral content which
results in some level of sulphur removal as well. A discussion on coal desulphurisation
was given by Celik and Somasundran (1994). In a recent paper by Rubiera et al. (1999)
it was argued that biological treatment does not have a substantial effect on the
combustion behaviour of coal.

Desulphurisation of petroleum
fractions has been successfully
applied for a long time in the
field of petro-chemicals. Using
H2 and a Co/Mo catalyst,
sulphur is removed from
petroleum fractions by
conversion to H2S. Examples of
“hydrocracking” reactions are
Figure 3.6 Cracking reactions during hydro-
desulphurisation (picture from van den Berg and de Jong, given in Figure 3.6 (van den
1980) Berg and de Jong, 1980).

3. Abatement techniques.
Since eventually the sulphur that enters the facility will come out in some form, these
methods are based on lowering the sulphur input, with the fuel, to the process. This
implies increasing the thermal efficiency of the fuel conversion process by increasing
the steam parameters (supercritical or ultra-supercritical steam). Alternatively, the
facility can be upgraded by “repowering” - examples are: transforming a pulverised
coal fired unit to a (pressurised) fluidised bed unit, implementing combined-cycle
technology or even the use of fuel cells (7 Chapter 2). In practically all cases this
implies rather expensive changes with many risks.
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-10

4. End-of-pipe control technologies.

The last but most widespread option is to remove the sulphur as SO2 or H2S after it
has been formed in the process. In fluidised bed reactors, sulphur can be captured in
situ in the bed by a calcium-based sorbent (L section 3.11) - for older plants existing
flue gas desulphurisation equipment may become obsolete after repowering to a
fluidised bed furnace. The rest of this chapter will concentrate on techniques that
accomplish that. In all cases this brings with it the production of significant amounts
of by-products, such as spent lime and dolomite, gypsum, elemental sulphur or
sulphuric acid.

3.6 Flue Gas Desulphurisation (FGD) 1: Non-regenerable processes

Worldwide, at those power utilities that have sulphur emissions control equipment
installed, in more than 80% of the cases the sulphur is removed from the flue gas
using FGD with a non-regenerable (“throwaway”) sorbent based on calcium. In
general lime (CaO), limestone (CaCO3) and ashes with a high calcium+magnesium
content are used, in an aqueous solution. The product is then an FGD gypsum
(CaSO4@2H2O) that finds large-scale application as a building or filling material.
Typically FGD can be applied to processes involving sulphur contents up to ~ 10 %-
wt in the fuel or input material. The most important FGD processes as listed by IEA
Coal Research (Soud, 1995) are given in Table 3.4.

Table 3.4 The most important FGD systems with non-regenerable sorbents

Process Sorbent End/by - product

Wet scrubbers Lime / Limestone Gypsum, Calcium sulphate/sulphite
Lime / Fly ash Calcium sulphate/sulphite/fly ash
Spray-dry scrubbers Lime Calcium sulphate/sulphite
Dual - alkali Primary: sodium hydroxide Calcium sulphate/sulphite
Secondary: lime
Seawater Primary: seawater Waste seawater
Secondary: lime
Walther Ammonia Ammonia sulphate

Lime, CaO, is produced by limestone burning, i.e. “calcining” the CaCO3 at

temperatures above 800EC under release of CO2.
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-11

Limestone/lime scrubbing of SO2 is based on the following chemistry, at

temperatures of the order of 70-90EC:

limestone scrubbing
CaCO3 (s) + H2O + 2SO2 6 Ca2+ + 2HSO3- + CO2 (R3-12)
CaCO3(s) + 2HSO3- + Ca2+ + H2O 6 CaSO3 @ 2H2O(s) + CO2 (R3-13)
CaSO3 @ 2H2O(s) + ½O2 6 CaSO4 @ 2H2O(s) (R3-14)
lime scrubbing
CaO (s)+ H2O 6 Ca(OH)2 (s) (R3-15)
SO2 + Ca(OH)2 (s) + H2O 6 CaSO3@ 2H2O(s) (R3-16)
CaSO3 @ 2H2O (s) + ½O2 6 CaSO4 @ 2H2O(s) (R3-14)

Figure 3.7 Process scheme for a typical wet flue gas desulphurisation process (picture from
Stultz and Kito, 1992)

A typical process scheme of FGD for a combustion flue gas is shown in Figure 3.7.
The flue gas is confronted with the aqueous lime (or limestone) solution in a scrubber;
the liquid collected by the scrubber is partly recycled and partly fed to the thickener
where calcium sulphite is further oxidised to calcium sulphate (reaction R3-14). This
can be applied with or without air added to the thickener, which are referred to as
natural or forced oxidation, respectively. The latter gives higher gypsum quality that
finds wider application, whilst the larger crystals produced make de-watering easier.

Figure 3.8 shows a flue gas desulphurisation absorber for a 600 MW power plant.
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-12

Figure 3.8 Flue gas desulphurisation absorber (picture Figure 3.9 A venturi scrubber (picture
from OECD/IEA&ETSU, 1993) from Cooper and Alley, 1994)

As an illustration a venturi scrubber is given in Figure 3.9. In principle it is very well

suited for the removal of other gases and dust particles at the same time.

Recently, the integration of the FGD scrubber process into the off-gas stack was
presented by Klitzke and Bengtsson (1999).

An FGD process that can be

applied in coastal regions is Seawater To stack
scrubbing with seawater as
Seawater
illustrated by Figure 3.10. scrubber
Condenser
Particle
Seawater, that is claimed to collector
Power FLUE FLUE
possess sufficient natural plant GAS GAS
alkalinity for SO2 removal is Lime
Water
preheated in the condenser treatment Air
of the steam cycle of the plant

power plant shown in the

Effluent seawater
Figure, and is used to scrub
fly ash particles as well as Figure 3.10 Principle of an FGD system based on seawater
scrubbing, using the natual alkalinity of seawater (after Soud,
SO2. 1995)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-13

In the water treatment plant the pH of the scrubbing liquid is adjusted before disposal
back into the sea (Soud, 1995). A seawater scrubbing plant with in situ oxidation (in
the bottom part of the absorber) is operating in Indonesia (Scheuch, 2000)

An important issue in limestone / lime scrubbing is the pH ( = - 10log [H3O+]) of the

solution. For lime scrubbing the optimum value is pH~8, for limestone scrubbing
pH~6. This difference is illustrated by the following gas/solid/water equilibrium
chemistry (Flagan and Seinfeld, 1988):

SO2 is absorbed in the water: SO2 + 2H2O W H3O+ + HSO3- (R3-17)

for limestone:
CaCO3 goes into solution: CaCO3(s) + H3O+ W Ca2+ + HCO3- (R3-18)
CaSO3 is formed: Ca2+ + HSO3- + 3H2O
W CaSO3@2H2O(s) + H3O+ (R3-19)
CO2 is released: H3O+ + HCO3- W CO2 + 2H2O (R3-20)
for lime:
CaO goes into solution: CaO (s) + H2O W Ca2+ + 2OH - (R3-21)
CaSO3 is formed: Ca2+ + HSO3- + 3H2O
W CaSO3@2H2O(s) + H3O+ (R3-19)
water is formed: H3O+ + OH - W 2H2O (R3-22)

The excess H3O+ concentration with limestone scrubbing explains the slightly acidic
solution, the excess OH -concentration with lime scrubbing explains the slightly
alkaline solution. In both cases, a sulphite product is formed that has to be further
oxidised, for which the oxygen is available in the flue gas that is cleaned (natural
oxidation). Alternatively the calcium sulphite can be further oxidised in a separate
reactor to which air is fed (forced oxidation). From a product quality point of view
a calcium sulphite concentration of 0.25 %-wt besides 90-95 %-wt calcium sulphate
is considered acceptable: gypsum quality specifications for its use in wallboard are
given in Figure 3.11 for several countries (Carpenter, 1998).

As argued by Sandelin et al. (1997), HF can interact with Al that may be present in an
FGD unit (as impurity of the lime(stone) or from fly ash) to form aluminium fluoride
complexes that hinder the dissolution of the lime(stone) (R3-18 or R3-21).

It is clear that limestone/lime FGD is very sensitive to other strongly acidic or

alkaline other components in the flue gas. For example HCl present in the flue gas
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-14

Figure 3.11 Several national quality specifications for wallboard gypsum (taken from
Carpenter, 1998)

reduces the pH, changes the equilibrium chemistry given above and consumes part
of the calcium by forming CaCl2. A pre-scrubber that operates at low pH (~1) can be
applied for the removal of HCl and HF (and fly ash and mercury, Hg), improving
gypsum quality. Alternatively, organic buffers can be applied that maintain a nearly
constant pH of the solution. This results also in smaller absorber size and less
pumping power needed.
The development of
Clean flue gas to stack the so-called dual-alkali
process was initiated by
the low solubility of
lime and limestone in
water and gypsum
Thickener
scaling problems in wet
FGD. Several
commercial-scale units
have been applied in
the USA. The solubility
Figure 3.12 Principle of the dual-alkali process (after Cooper and
of sodium salts is much
Alley, 1994) better, which led to this
concept based on a
scrubbing liquor with NaHCO3 and Na2SO3 as the SO2-binding compounds. The
principle of the process is given in Figure 3.12.
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-15

The chemistry of the process, again with FGD gypsum as the by-product, can be
summarised as follows:

in the scrubber:
Na2SO3 + SO2 + H2O 6 2NaHSO3 (R3-23)
2NaOH + SO2 6 Na2SO3 + H2O (R3-24)
in the precipitation tank:
2NaHSO3 + CaO(s) + H2O 6 Na2SO3 + CaSO3 @ 2H2O(s) (R3-25)
NaHSO3 + CaO(s) 6 CaSO3 @ 2H2O(s) + NaOH (R3-26)
CaSO3 @ 2H2O(s) + ½O2 6 CaSO4 @ 2H2O(s) (R3-27)
in the regeneration tank:
NaHSO3 + Na2CO3 6 Na2SO3 + NaHCO3 (R3-28)
NaHSO3 + NaHCO3 6 Na2SO3 + H2O + CO2 (R3-29)

Since some sodium is lost from the system with the CaSO3/CaSO4 product, some
make-up Na2CO3 (soda) or Na2CO3 " NaHCO3 (trona) is continuously added to the
regeneration tank. The actual SO2 capture is accomplished with a more concentrated
sodium-based washing liquid with limestone/lime scrubbers, combined with
precipitation and regeneration in separate devices.

The Walther process is based on scrubbing with ammonia water: sulphur is obtained
mainly as (NH4)2SO4. Due to the risk of formation of ammonium sulphate aerosols
a fuel sulphur content below 2%-wt is required. One installation in Germany operates
at efficiencies ~90%. The technology is nowadays owned by the German LLB (Lurgi
Lentjes Bischoff) who changed the concept to a two-stage scrubber and an electrostic
precipitator for aerosol particle removal (Scheuch, 2000).

3.7 Flue Gas Desulphurisation (FGD) 2: Regenerable processes

One important disadvantage of FGD with throwaway sorbents is that large amounts
of (by-)products and spent sorbents are being generated. Although this is driven by
economics (abundance of limestone and other calcareous rock), wet FGD methods
have been developed using a regenerable sorbent as well. Production of a valuable
sulphur-based product, reducing gypsum scaling problems or the possibility to
combine SO2 control with the control of other pollutants provide the basis for these
processes. The most important FGD processes using regenerable sorbents, as listed
by IEA Coal Research (Soud, 1995) are given in Table 3.5.
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-16

Table 3.5 The most important FGD systems with regenerable sorbents

Process Sorbent / principle End/by - product

Wellman - Lord Sodium sulphite (Na2SO3) Concentrated SO2
Bergbau Forschung / Uhde Activated carbon Concentrated SO2
Linde SOLINOX Physical absorption (amine) Concentrated SO2
Spray-dry scrubbing Sodium carbonate (Na2CO3) Elemental sulphur
MgO process Magnesium oxide (MgO) Concentrated SO2

The Wellman-Lord process (the most widely used wet, renewable sorbent FGD
process) has similarities to the dual alkali process mentioned above. A simplified
process scheme is given in Figure 3.13. Also here, sodium sulphite, Na2SO3 is the
actual sorbent, binding sulphur as NaHSO3. However, the sulphur is not passed on
to calcium, but it is released again as SO2, which is extracted as a concentrated
mixture with water (~85%
SO2, 15% water) for
further processing to
sulphuric acid. Some
sodium sulphite/sulphate
is formed as a by-product:
make-up soda or trona is
needed to balance this.
Pre-scrubbing for HCl etc.
prevents sorbent
Figure 3.13 Principle of the Welmann-Lord process deactivation. Process
(after Soud, 1995) chemistry can be
summarised as follows:

in the scrubber:
Na2SO3+ SO2 + H2O 6 2NaHSO3 (R3-23)
Na2SO3 + ½O2 6 Na2SO4 (R3-30)
in the regenerator:
NaHSO3 6 Na2SO3+ SO2 + H2O (R3-31)
make-up soda or lime:
Na2CO3 + SO2 6 Na2SO3 + CO2 (R3-32)
2NaOH + SO2 6 Na2SO3 + H2O (R3-33)
in the gaseous product:
SO2 + ½O2 + H2O 6 sulphuric acid (R3-34)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-17

A process that allows for the

control of SOx and NOx
(nitrogen oxides NO and NO2,
L Chapter 4) at the same time
is the fluidised bed copper
oxide process (Frey, 1993). This
process needs methane, CH4 as
well as ammonia, NH3. In a
fluidised bed absorber, SO2 is
captured by copper oxide CuO
at ~400EC. At the same time,
NO is reduced by ammonia to
nitrogen, for which copper
oxide is a very good catalyst.
The copper sulphate CuSO4
product is transported to the
Figure 3.14 The fluidised bed copper oxide proces (picture two-stage heater/regenerator
from Frey, 1993) section where it is reduced to
Cu at the expense of methane at
~500EC. A simplified process scheme is shown in Figure 3.14. In short, the chemistry
involved here is as follows:

in the absorber:
CuO (s) + SO2 +½ O2 6 CuSO4 (s) (R3-35)
CuO (s) + SO2 6 CuSO3 (s) (R3-36)
CuO (s) + SO3 6 CuSO4 (s) (R3-37)
4 NO + 4 NH3 + O2 6 4 N2 + 6 H2O (R3-38)
2 NO2 + 4 NH3 + O2 6 3 N2 + 6 H2O (R3-39)
in the transport section:
Cu (s) + ½ O2 6 CuO (s) (R3-40)
CuSO3 (s) + ½ O2 6 CuSO4 (s) (R3-41)
in the heater / regenerator:
CuSO4 (s) + ½ CH4 6 Cu (s) + SO2 + ½ CO2 + H2O (R3-42)
CuSO3 (s) + ¼ CH4 6 Cu (s) + SO2 + ¼ CO2 + ½ H2O (R3-43)
CuO (s)+ ¼ CH4 6 Cu (s) + ¼ CO2 + ½ H2O (R3-44)
in the off-gas treatment:
CH4 + 2 SO2 6 2/x Sx (s) + 2 H2O + CO2 (R3-45)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-18

The sulphur is released as concentrated SO2 that can be used to produce sulphuric
acid, similar to the Wellman-Lord process. The Cu sorbent is transported back to the
fluidised bed using air, during which it is oxidised to CuO. The SO2 product gas may
contain some methane that is converted via reaction (R3-45). Currently the FB copper
oxide is receiving more attention from an NOx control point of view (e.g. Jeong and
Kim, 2000) whilst earlier SO2 was the main target pollutant.

3.8 In-duct sorbent injection at low temperatures

As alternatives for the wet FGD methods discussed above, dry or semi-dry in-duct
sorbent injection methods can be used more upstream in the flue gas channel, at
temperatures of the order of 150-400EC. For a combustion plant this means between
the air pre-heater and the cold-side (electro-) filter system for dust control (L
Chapter 5). In dry systems a dry sorbent powder, usually lime, is dispersed in the flue
gas channel, where it reacts with the SO2 via the overall reaction

CaO (s) + ½O2 + SO2 6 CaSO4 (s) (R3-46)

The CaSO4 particles are then removed from the flue gas by the (electro-) filter for fly
ash emissions control further downstream. This approach is relatively cheap and can
easily be applied to older facilities with dust control but without proper sulphur
control, such as can be found in Eastern Europe and in large parts of Asia.

In general, the sorbent used is lime which needs additional humidification, which is
not needed when a (more reactive) sodium-based sorbent is used, such as sodium
bicarbonate NaHCO3. With semi-dry methods, a calcium-based sorbent is sprayed
into the flue gas duct as a water slurry, which, after the water has been vaporised
renders a more reactive sorbent when compared to dry methods. Similar activation
of the sorbent can be obtained by feeding water to the dust a little downstream of the
point where the actual sorbent is fed. The unreacted lime will be converted to calcium
hydroxide :
CaO (s) + H2O 6 Ca(OH)2 (s) (R3-47)

which is more reactive to SO2, giving calcium sulphite

Ca(OH)2 (s) + SO2 6 CaSO3 (s) + H2O (R3-48)

Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-19

that will be (partly) oxidised to calcium sulphate. A drawback of the semi-dry

methods is the corrosion of the humidification duct.
Typically these
processes are operated
at Ca/S ratios near 2,
with typical SO2
concentrations of a few
1000 ppmv from coal
combustion. Another
very important process
parameter is the so-
called approach
temperature, which is
how far the gas
Figure 3.15 Principle of in-duct sorbent injection (picture from
temperature is above the
Ciemat, 1998) saturation temperature
for the water in the gas.
With a lower approach temperature more time is needed for water evaporation,
resulting in a higher sulphur uptake by the sorbent (Ciemat, 1998).

A simplified process scheme for this is shown in Figure 3.15. Often the performance
when it comes to conversion of lime is rather poor (50 ~ 70%) even after spent
sorbent recycle, and better results are obtained with high-temperature sorbent
injection methods (L section 3.9).

Figure 3.16a Main features of the NID process Figure 3.16b NID process at Laziska power
(picture from MPS, 1997) plant in Poland (picture from MPS, 1997)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-20

A recent low temperature process is the Alstom Power (earlier ABB Fläkt) NID (New
Integrated Desulphurisation) process. The principle is illustrated in Figure 3.16. After
hydration to Ca(OH)2, lime is mixed with some more water and part of the dust cake
from a fabric filter. This wet solid is fed to the flue gas, by which it is rapidly
transported to the fabric filter. In the filter the solids form a cake that is reactive to
SO2, which is periodically removed by pressure pulses (L chapter 5 for more detail
on fabric filters). One NID unit is operated at a 120 MW pulverised coal-fired boiler
in Poland (Figure 3.16b), removing 90-95% of the SO2 at 600-1200 ppm (MPS, 1997).
Other facilities are operated in Germany and the UK (Åhman and Wiktorsson, 2000).

CFB-FGD technology (Baege and Sauer, 2000) is FGD technology based on

absorption on hydrated lime at ~70EC in a circulating fluidised bed. Also HF and HCl
are removed from the flue gas. The process operates at Ca/S (molar) ratios in the
range 1.1~1.5, depending on the SO2 concentration etc. In order to cut the costs of
the lime an alternative is to inject limestone into the upper zone of the boiler at 850-
1000EC for calcination. Upon entering the CFB reactor the lime is reactivated by
water injection.

3.9 High temperature SO2 capture by furnace sorbent injection

A second method that is attractive for older boilers with a relatively short remaining
lifetime is to inject a sorbent for SO2 into the furnace or the hot part of the flue gas

Figure 3.17 A typical furnace sorbent injection process (picture from Ciemat, 1998)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-21

channel. The principle of furnace sorbent injection is shown in Figure 3.17. As with
in-duct sorbent injection at lower temperatures this approach involves relatively low
capital costs, achieving sulphur removal efficiencies of the order of 60-75 % with
Ca/S ratios 2-4. Higher efficiencies can be reached when water is sprayed into the
flue gas duct before the dust filter (which reactivates the unreacted sorbent for SO2
that is still in the flue gas), by spent sorbent recycling and by careful optimisation of
the position of sorbent injection with respect to the temperature.

The relation between

temperature and the SO2
removal that can be
obtained is shown in Figure
3.18 for a sorbent such as
limestone or hydrated lime.
The temperature maximum
at around 1200EC is related
to the kinetics of CaSO4
formation on one hand and
its thermodynamic stability
Figure 3.18 SO2 removal efficiency versus temperature for lime
sorbent injection (picture from Ciemat, 1998) on the other: CaSO4 is not
stable above around 1250EC
in typical flue gases from
coal-firing, containing up to
5000 ppm SO2.

The difference between the

use of CaCO3 and Ca(OH)2
is illustrated by Figure 3.19,
corresponding to Zone 1 in
Figure 3.18.

Figure 3.19 Sulphur uptake efficiency for limestone and

hydated lime versus temperature (picture from Iisa, 1995)
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-22
A simplified view on the
chemistry of SO2 capture by
hydrated lime or limestone is
shown in Figure 3.20. As
shown, particle size - related
effects such as fragmentation
play a major role.

An advantage of furnace
(and also in-duct) sorbent
injection is that it more
efficiently removes SO3 than
a wet FGD scrubber does,
Figure 3.20 Chemical processes during in-duct sorbent and that HCl and HF are
injection for SO2 removal (picture from Ciemat, 1998) captured as well:

formation of CaO:
CaCO3 (s) W CaO (s) + CO2 (R3-49)
Ca(OH)2 (s) W CaO (s) + H2O (R3-50)
sorption reactions:
CaO (s) + ½O2 + SO2 W CaSO4 (s) (R3-46)
CaO (s) + SO3 W CaSO4 (s) (R3-51)
CaO (s) + 2HCl W CaCl2 (s) + H2O (R3-52)

Furnace sorbent injection may require some additional soot-blowing to remove solids
accumulated on boiler tubes, which is not problematic, though (Ciemat, 1998).

One typical example is the LIFAC (Limestone Injection into the Furnace and
Activation of unreacted Calcium) process, developed in Finland during the 1970s and
1980s by Tampella Power Oy (the current Kvaerner Pulping Power Division). The
technology is nowadays owned by the Finnish Fortum Power and Heat (the Former
Imatran Voima Oy). One unit in is operated Finland (at the Inkoo power plant),
about 10 other units exist in North-America, Russia and China (LIFAC, 2000,
Ryyppö and Ekman, 2000) - see Figure 3.22.

In the LIFAC process (Figure 3.21), limestone sorbent is injected into the upper
furnace near the location of the superheater. It calcines (reaction R3-49) and partly
reacts with SO2 (reaction R3-46). Further downstream, in the activation reactor, the
Zevenhoven & Kilpinen SULPHUR 13.6.2001 3-23

Figure 3.21 Schematic of the LIFAC process (from LIFAC, 2000)

unreacted calcium oxide reacts with water at ~400EC to

f o r m c a l c i u m h y d r o x i d e
(reaction R3-47) which is more reactive to SO2 (reaction
R3-48). The solid reaction products are collected
together with fly ash in the electrostatic precipitator
filter (L chapter 5) and are partly re-injected. Sulphur
removal efficiencies of 65-85% were reported,
depending on Ca/S ratio (typically 2-2.5), approach
temperature, and sorbent fineness, for a boiler fired
with coal with 1.4 - 2.8 %-wt sulphur (LIFAC, 2000).

The LIFAC system at Richmond Power & Light at

Whitewater Valley Station, Unit No. 2, shown in Figure
Figure 3.22 LIFAC unit in 3.22 (on the right side of the stack), is the first to be
Richmond (VI) USA (from applied to a power plant using high-sulfur (2.0-2.9%)
LIFAC, 2000)
coal.

A system referred to as LIMB (Limestone Injection Multi-stage Burner) by Babcock

& Wilcox Co is based on injection of limestone, lime or dolomitic limestone in the
upper furnace, together with fuel in a Low-NOx burner (L chapter 4). This method
is reported to give 60~70 % sulphur removal plus ~50 % NOx reduction (US DOE,
2000).