The growth of industrialization makes it imperative to reduce the amounts of sulfur dioxide emitted into the atmosphere. This article describes various processes for cleaning flue gases, and gives details of new methods being investigated.
Wet scrubbing with water, though widely practised, has many disadvantages. Scrubbing with zinc oxide, feasible in zinc works, is more satisfactory.
Dry methods use a solid absorbent; they have the advantage of a high emission temperature.
Other methods are based on the addition to the fuel or the flue gases of substances such as activated metal oxides, which react with the sulfur to form compounds less harmful than sulfur dioxide. Also being investigated are a two-stage combustion system, in which the sulfur dioxide is removed in the first stage, and the injection of activated powdered dolomite into burning fuel; the resulting sulfates being removed by electrostatic precipitation.
A wet catalysis process has recently been developed.
Most of the cleaning processes are not yet technically mature, but first results show good efficiency and relatively low cost.
For many years air hygienists have been in general agreement that flue gases containing sulfur should not be emitted into the atmosphere without previous cleaning, at least in the larger installations. The cleaning of effluents is not widely practised and, in addition to the economic difficulties, it presents technical problems until recently insoluble. Analyses of air in various countries show that the concentration of SO₂ reaches and sometimes exceeds the maximum permissible concentrations, especially in large cities and highly industrialized areas. In Germany, the basic tolerance level is established at 0.4 mg/m3, but this figure may be greatly exceeded in periods of poor atmospheric exchange. Thus, in the Ruhr area in December 1962, the SO₂ concentration reached 4 mg/mi, and a level of 2-3 mg/m3 was measured over a period of several days; similarly high values have been recorded in London and elsewhere. Increasing urbanization and industrialization, with the consequent increase in the SO₂ concentrations at ground level, mean that this problem will be a serious menace in the near future, unless methods are found for tackling it. Fortunately, some processes now being developed appear to meet the requirements for efficient sulfur removal.
1 Scientific Counsellor, Bundesgesundheitsamt, Institut fur Wasser-, Boden- und Lufthygiene, Dusseldorf, Germany.
Reduction of air pollution can be accomplished only by reducing the emissions themselves, as there is no way of preventing their dangerous accumulation once they have been discharged into the atmosphere. The ground level concentrations can be reduced by measures aimed at improving emission conditions: increasing the height of chimney stacks, raising the temperature or speed of the flue gas, or diluting the flue gas with air. These measures, however, are effective only in favourable weather conditions; with atmospheric stagnation causing the flue gas to sink to ground level or to accumulate in the atmosphere, even the highest chimneys and the best emission methods may be ineffective.
Sulfur dioxide is produced by some chemical processes but mainly by combustion reactions. The only method of reducing pollution from domestic heating, with its innumerable sources, seems to be the provision of fuels with a restricted sulfur con-tent, which should be as low as possible. Industrial sources of SO₂, though fewer numerically than the domestic sources, often produce far greater amounts of pollution. But the use of sulfur-free fuel in industry is unlikely to be accepted in practice, as it is far more costly to remove sulfur from the fuel than to cleanse the burnt gases. Removal of sulfur from solid fuel appears to be impracticable, so that reasonable methods for the removal of sulfur oxides from flue gases will continue to be of importance. Although devices for cleaning gases are very expensive, the benefit to the public of reducing emissions of injurious gases may possibly justify the expenditure involved.
The removal of gaseous SO₂ from flue gases may be achieved by chemical or physical means, using gaseous, liquid or solid substances. The most successful combinations of these processes have proved to be wet scrubbing with liquids and dry methods with solid substances. It is possible that a cleaning method using gaseous substances may be found effective.
WET SCRUBBING
Wet scrubbing processes need not necessarily use water as a solvent: other liquids such as alcohols, amines, etc., could also be used. Ideally, the solvent used should be cheap, have a high capacity to dissolve SO₂, and form a substance that could be sold economically. Its vapour pressure should be very low, to avoid losses due to evaporation. Until now, it has been impossible to find such a substance, and the choice has always fallen on the cheapest and most readily available substance, water. However, it is now recognized almost everywhere in the world that wet scrubbing processes requiring huge quantities of water, such as those employed in the United Kingdom and the USSR, are unsatisfactory. The low washing capacity has proved to be a great disadvantage, as this means the use of very large quantities of water and a deterioration in emission conditions. For example, 75 tons of water are required per ton of oil, and even then 5% of the SO₂ remains in the flue gas emitted into the atmosphere; moreover, it is now in a much worse condition, since the gas has lost its buoyancy through the cooling action of the washing and under adverse weather conditions it may descend to ground level and cause very high ground concentrations. Thus scrubbed gases, if incompletely cleaned, may give rise to more nuisance than unwashed gases. The removal of SO₂ by scrubbing with water is limited by the decreasing solubility of SO₂ with diminishing concentration: as the washing process proceeds, larger quantities of water are required to remove the same amount of SO₂. The volume of water needed to remove the last low percentages is particularly great, and this is the reason why, in many water scrubbing processes, some SO₂ remains in the flue gas. This could be avoided by, for example, reheating the flue gas, but this would increase the cost. However, there is undoubtedly a compromise possible between the scrubbing process and the reheating of the gases which would be acceptable. Such a compromise, in fact, has been made at a large power station: the flue gases are washed free of 70 % of SO₂ and then reheated to 105°C before being emitted. If the SO₂ is completely removed, no reheating is necessary.
Water washing processes such as the Battersea process cause water pollution as well as air pollution. In London the problem is solved by returning the water to the Thames, which is tidal at that point and is not used downstream of the power station. The recovery of sulfur from water is usually un-profitable, so that this valuable material is wasted.
Zinc works generally have an accompanying sulfuric acid plant. Here the most rational method for removing the sulfur oxides from the flue gases would be the use of ZnO, if wet scrubbing is the cleaning process. The sulfur oxides are absorbed and the resulting compound, containing insoluble zinc sulfite, insoluble excess zinc oxide and soluble zinc sulfate, is used as a wetting agent for the molten pyrites and blendes, which have to be wetted before roasting in the sintering furnace. If the amount of water used is more than that needed as a wetting agent, part of the clear ZnSO4 solution remaining after separation of the insoluble matter may have fresh ZnO added and be re-used for washing. The efficiency of the ZnO method is almost 100%, and only small traces of SO₂ are left in the flue gases. The primary cost of erecting this type of scrubber is less than that for a pure water scrubber, as the dimensions of the ZnO device are considerably smaller for an equal scrubbing capacity.
DRY PROCESSES
In view of the disadvantages common to all wet scrubbing processes, scientists, especially in the Federal Republic of Germany, the United Kingdom, and the United States, are attempting to find dry materials capable of removing SO₂ from flue gases. A dry process would have the great advantage of maintaining a high emission temperature, whereas wet scrubbing necessarily lowers the temperature of the gases to that of the washing water. The aim of the dry processes, like that of the wet ones, is to prevent the emission of the noxious and offensive SO₂ gas into the atmosphere. The principle of dry processes is similar to that of wet methods: the adsorbent material is circulated in counter current to the flow of gases to be cleaned. The adsorbent substance must have a high adsorbing efficiency if the plant is not to be excessively large. High efficiency means short contact time, large throughput and small reactor volume. In addition, the solid matter must be relatively hard with minimum rub off. If the reaction is catalytic, the catalyst in the reactor need not be removed.
OTHER METHODS
Other methods suitable for sulfur removal are based on different principles. It is known that sulfur dioxide and sulfur trioxide reduce the dew point of the flue gases and, particularly in the presence of corrosive catalysts such as vanadium pentoxide or sodium sulfate, cause heavy corrosion of boiler materials and a consequent lowering of fuel efficiency. For high-pressure boilers in particular, the fuel should have a low sulfur content. Sulfur removal, which is necessary for thermotechnical reasons, is effected by adding, either to the fuel or to the flue gas, substances which will react with the sulfur to form compounds less harmful than SO₂.
Such substances are, for example, ammonia, soda, dolomite, or metallic oxides such as Fe₂O₃, ZnO and others. Although until now this method has not proved technically successful, two new processes seem to offer improved possibilities.
Metal oxides
Basic studies on the sorption efficiency of various substances (metals and metal oxides) have been made in recent years in the United States by Bien-stock et al. (1958). In addition, in 1955 Johswich (1962) found that activated iron oxide is a suitable adsorbent for SO₂ and SO at temperatures between 100°C and 300°C. The substance he used consisted of 75 % iron ore, Giulini mass or Lux mass, and 25 % soda. The price of this mixture was about $50 per ton. This substance is able to adsorb sulfur gases to almost 100% until it is exhausted, usually at about 21 %; after this, the SO₂ passes the adsorbent. Assuming that the soda is completely spent, this means that only 10% of the iron oxide reacted. At a sulfur content of 3 %, the cost for cleaning one ton of oil would be $15, which is relatively high. Regeneration of the waste product, which could reduce the cost, has not been tried.
Two other methods that make use of alkalized metal oxides are based on the application of activated alumina and manganese oxides.
Activated alumina gel has been prepared by precipitating aluminium sulfate with a soda solution: the sulfate-free gel, after washing and drying, consists of 60% Al203 and 30% Na2O, and is capable of adsorbing SO2 gas at temperatures of 300°C to 330°C. It takes up 90 % ofthe SO₂ until it is exhausted, which happens when it has adsorbed 19 % to 24% of its own weight. The sulfur oxides are catalytically oxidized to sulfates, so that the product is mainly sodium sulfate. Regeneration of the adsorbent material is carried out at a temperature of about 600°C by a reducing gas such as H2, CO, water-gas or producer-gas, in combination with CO.
This process causes the sodium sulfate to form soda, the sulfate being reduced to a mixture of H2S, COS and elementary sulfur which is fed into a Claus reactor. Fig. 1 shows the process schematically.
FIG. 1 REMOVAL OF SO2 BY ALKALIZED ALUMINA GEL
For one ton of oil with 3% sulfur content, the cost of this process (in addition to the basic construction costs) would be about $6, namely, $2.25 for the adsorbent, S3 for the reducing gas and $0.75 for operation. If the primary investment costs-fairly high, as part of the equipment must be made of stainless steel-are taken into account, the over-all costs would come to about $6.50 per ton of oil.
The manganese oxide process begins with manganese sulfate (MnSO4), which is treated with sodium hydroxide (NaOH). Further treatment of the sediment is the same as that used for the aluminium oxide gel. The adsorption capacity for SO₂ is slightly better, being 25 % to 37 % by weight. In the presence of oxygen, of which there is almost always a certain excess in combustion processes, the reaction with Mn₂O₈ produces MnSO4. Recovery of the manganese oxide must be carried out separately, so that a two-stage system results. The schematic representation of this process is shown in Fig. 2. The sodium sulfate obtained could be sold as such or electrolysed to form sulfuric acid, hydrogen and utilizable sodium hydroxide. The main disadvantage of this process is that the flue gas must be completely free of dust, otherwise the activity of the adsorbent (manganese oxide) is gradually blocked. Owing to the high efficiency of the adsorbent, this method promises to be technically significant for oil furnaces, sulfuric acid plants and other emission sources without dust.
FIG. 2 REMOVAL OF SO2 BY PRECIPITATED MANGANESE OXIDE
A relatively simple method for removing SO₂ has been developed by Heitmann & Sieth (1963) at the Siemens-Schuckert plant in Erlangen. In the first stage the method is similar to the alumina gel and manganese oxide processes: the SO₂ is adsorbed at temperatures from 400°C (start) to 200°C on to iron oxides or on to the contact mass of a carrier with precipitated iron oxide. However, regeneration is carried out by raising the temperature of the desorber to 600°C; at this temperature the iron sulfate formed separates from SO₂, which can be used in the manufacture of sulfuric acid, while the re-formed iron oxide mass is returned to the absorber. This process has given satisfactory results on a laboratory scale, and is to be tried in a pilot plant. The contact mass of the carrier has proved to adsorb more SO₂ than Fe₂0₃ can take up stoichiometrically: this shows that the carrier also adsorbs SO₂.
Reinluft process
The Reinluft process, based on physical and chemical principles, began by using activated charcoal, long recognized as a good SO2 adsorbent. If oxygen is present, charcoal catalytically oxidizes SO₂ to SO₃, which combines with the water vapour from the combustion process to form sulfuric acid.
This adheres tightly to the active surface of the charcoal, especially at low temperatures, and desorption requires temperatures of 350°C to 450°C. At these temperatures, relatively low for desorption, all the sulfur is recovered as SO₂; other undesirable sulfur compounds, such as COS or CS2, do not occur; SO, is reduced to SO₂. Some CO2 is produced from the charcoal, as a result of the reaction C + 2SO = CO2 + 2SO2. Although the chemical loss (CO2) is not very high, the cost of activated charcoal, which is very expensive, would alone reach about 55 per ton of oil with 3 % sulfur content. However, certain half-cokes possess almost the same properties for adsorbing SO₂. Half-cokes are those products obtained at temperatures of about 600°C from bituminous coal or other carbonaceous mate-rials. Though at first half-coke has only a low SO₂ adsorbency, in time it becomes activated through repeated adsorption and desorption, until it achieves an efficiency almost as high as that of the best activated charcoal. Successful application has been made of the fact that coke immersed in sulfuric acid and then dried has very good SO₂ adsorb-ency. The use of half-coke is economically very much cheaper than the use of prepared activated charcoal, so that this modified method becomes competitive with others. The raw material for making half-coke should have good briquetting qualities, and the half-coke itself should be capable of being oxidized. Not all bituminous coals are suitable for the production of half-coke, as some of them contain admixtures, the amount of vanadium present appearing to have considerable effect on the properties of the coal.
FIG. 3 REMOVAL OF SO BY THE REINLUFT PROCESS WITH ACTIVATED HALF-COKE
The Reinluft process is shown schematically in Fig. 3. The flue gas to be cleaned (containing SO₂ and SO₃) first passes through E2, where it is cooled to 200°C and led through the adsorber A for the first time: here SO₂ is removed (some SO₂ is oxidized to SO, and adsorbed), and the flue gas is further cooled to 100°C by being passed through El before being led through the adsorber for the second time for the removal of SO₂. The cleaned gas, now at a temperature of 100°C to 120°C because of the heat released by the adsorption reaction, is then emitted into the atmosphere. Regeneration of the half-coke is carried out in the desorber B at temperatures of about 400'C, using reducing gases, the most convenient being that from the adsorber. At 400°C, H₂SO₄ splits into H2O and SO, the latter being reduced to SO₂. The result is that all the sulfur is recovered in the form of SO₂ as a 40% to 50 % gas mixture, which can be sold economically in that state or used as raw material for the production of sulfuric acid. The sulfur-removal efficiency of the process can be designed as high as required, but for economic reasons it is preferable not to adsorb more than about 90 %. Small amounts of dust do not impair the adsorbency of the coke: the dust is first retained on the surface of the coke, then separated by a sieve and removed from the process with the rub-off. Large quantities of dust, however, destroy the activated surface, and should be removed by an electric precipitator installed between the gas outlet from the boiler and E2. The extent to which the coke surface would be affected would depend mainly on the nature of the dust.
The process has been tested in a pilot plant at the Volkswagen plant in Wolfsburg and has given satisfactory results. The tests were made with an oil-burning furnace from which 2000 m3 NTP of flue gas with a sulfur content of 3 to 6.5 g/m0 NTP were drawn off and examined for (1) waste of contact mass, (2) regeneration, and (3) operating conditions.
The table below shows that any required efficiency of cleaning may be obtained by appropriate adjustment of the gas flow rate, the adsorbent flow rate and the adsorbent temperature.
Although adsorption is most efficient at the lowest temperatures, the process cannot in practice be operated below 160°C because of the heavy corrosion that results at temperatures below the dew point of SO₂. The efficiency of the adsorbent mass decreases with increasing temperature: at 160°C the amount of SO bound is only 40 % of that bound at 70°C. On the other hand, 53 is completely removed even at temperatures as high as 250°C. The pilot plant, made of normal open-hearth steel, has been in continuous operation, which means that it has been exposed to heavy operating conditions.
Corrosion appeared first in those parts of the equipment charged with flue gas at temperatures below the dew point of the gas, as might be expected in view of the repeated action of dilute sulfuric acid on the material. After three weeks the pipe of the desorber was so damaged as to be unserviceable, and some other components had to be reconstructed and improved. The part of the plant operated at temperatures of about 100°C was in perfect order and showed no corrosion.
The tests on the pilot plant could not provide any data on costs, but they showed that the Reinluft process is probably near to technical maturity. In view of the results achieved, the Volkswagen company is planning a full-scale plant with a capacity of 60000 m3 NTP/h for a sulfur content of 5 g/m3 NTP and with an adsorbent efficiency of 80%. The maximum load capacity of the adsorber will be 80 000 ms NTP/h, but at this figure the efficiency is expected to drop to 70%.
Positive results are also expected from the Rein-luft process installed in an iron sintering plant in Duisburg. An interesting fact is that here the dust in the flue gas has not reduced the adsorbency of the half-coke. The same process has been tested at the Warren Spring Laboratory in Stevenage, England. It should be pointed out that half-coke adsorbs not only SO₂ but also other gaseous compounds and aerosols such as H2S, COS, CS2 and nitrogen oxides. In addition, organic matter such as tar, resinous products and soot are removed from the flue gases, as they are destroyed by the high temperatures and the sulfuric acid.
Two other full-scale plants are being built in Germany, one at a chemical works in Cologne (Glanzstoffwerke-mainly H2S and CS2-60 000 ml NTP/h) and the other at a coal dust-fired power station (33 000 ml NTP/h) in Lunen, Westphalia. The plants are expected to come into operation
in 1965.
Apart from the construction costs, which are about 10% of the total costs of a new power station, the operating cost per ton of oil with a 2% sulfur content, for a purification factor of 95 %, would come to $1.50, which is relatively low.
Combustion methods
An additive process recently developed by Wickert (1963) burns oil in two stages; in the first, oil is burned in a pre-burner with only one-third of the air needed for complete combustion. Under these conditions the liquid partially separates into a gaseous form and also forms soot. Because of the very strong reducing atmosphere, the sulfur, mainly organic, is converted into hydrogen sulfide, which is easier to remove from flue gas than sulfur dioxide. The H2S is drawn off before the fuel enters the main combustion chamber, so that no SO₂ occurs. The diagram of the process is given in Fig. 4.
After leaving the pre-burner, the gaseous fuel containing H2S is passed into the reactor B, which is fed by CaCO3. This substance is separated, at temperatures of about 1200°C, into CO2 and CaO, the latter reacting with the H2S to form CaS. The fuel is then led via a dedusting device into the main combustion chamber. In principle, the reactor could be fed in countercurrent with CaCO2 granules or concurrently with CaCO3 dust. In practice, the dust is preferable, since its adsorbency is better than that of granules, which also have the disadvantage of having to be melted for regeneration. The reactant (CaS) is stirred with water to form a sludge (F) and treated with CO2 to re-form CaCO3. The dried and melted CaCO3 (H) is circulated again in the reactor, while the H2S is fed into the Claus reactor. The final sulfur product of the oil is elementary sulfur.
In order to reduce the amount of soot produced in the pre-burner, a finely dispersed desooting catalyst should be added to the oil. The substances used for desorption of the H2S-namely, CaO and MgO-are suitable as desooting catalysts. In the presence of these catalysts, soot reacts at high temperatures with water vapour to give CO and H2. A special after-burner for the soot is unnecessary, as the H2S-adsorbent mass itself is a soot combustion catalyst and any soot which reaches B is burned there. The marked adsorption of H2S starts in B at about 600°C and increases with increasing tem-perature, up to a maximum of about 900°C. The reaction
CaO + H2S ↔ CaS + H2O
is reversible, and in the absence of water vapour it lies almost entirely on the right-hand side of the equation; if, however, water vapour is present, the reaction may be in reverse. The amount of water formed in the pre-burner is usually too small to impair the adsorbency of the CaO, and the balance is predominantly on the right-hand side of the equa-tion. The separate stages of this process have been worked out thoroughly in a pilot plant, and the results obtained are now being used in the design of a full-scale plant for the Pintsch-Bamag company.
FIG. 4 REMOVAL OF SULFUR FROM OIL BEFORE COMBUSTION
FIG. 5 REMOVAL OF SULFUR FROM OIL AFTER COMBUSTION
Wickert (1962) also investigated the reaction of powdered dolomite in a flame at high temperature. The experiment is not a new one, but departed from A previous practice in activating the dolomite with alkali. The activated and dried powder is injected through six to eight jets about 4 m above the oil/air input, as shown in Fig. 5. The high temperature, Jets and the catalytic property of the iron oxide always 0 0 present in natural dolomite in the ratio of at least 1 %, cause the sulfur to oxidize to SO₃, which then binds to MgO and CaO to form MgSO4 and CaSO4. These sulfates have to be removed from the flue gas by an effective electrostatic precipitator. The reaction of SO₂ with CaO occurs without any catalyst, but for the reaction with MgO the presence of Fe2O3 is necessary. If the flue gas contains oxygen, SO₂ is bound as a sulfate; if not, CaO reacts with the SO₂, forming CaSO3, which is stable up to about 600°C; above this temperature, CaSO3 separates into CaS and CaSO4. At 600°C, CaSO3, CaSO4 and CaS have all been found.
In practice, the amount of sulfur removed depends greatly on the temperature. Experiments have shown that the best temperature for using CaO and CaO.MgO is about 900°C, though the range is slightly different for the two compounds: for CaO it is 800°C-1 100°C, for CaO.MgO, 750°C-1000°C. Below and above these temperatures there is a sharp drop in catalytic efficiency; very little SO₂ is bound at 1200°C, and none at all at 1300°C. The best method proved to be that of blowing the dust into the 1400°C zone, in order to ensure its complete distribution at 1200°C. Attention must be given to the sintering temperature of the dust, which usually lies below the injection zone temperature (about 1250°C, depending on the mineralogical deposit). In the experiments described, dolomite dust with a sintering temperature of 1230°C did not sinter when blown into a temperature zone of 1400°C, since the time during which the dust was subjected to this high temperature was not long enough. The contact time for reaction must be at least one second, but this presents no problem, as the particles usually persist much longer. Symmetrical spreading, which is very important, is best achieved by blowing the dust in through several nozzles at right angles to the flame direction.
Full-scale studies have been made at the Mobil Oil works in Bremen and at the Volkswagen plant in
Wolfsburg, the second of these being particularly valuable. Removal of SO₂ has been carried out using 20.4 kg of dolomite hydrate per ton of oil per per-centage of sulfur; the catalyst was blown in through four nozzles. Since the oil had a sulfur content of 1.9%, the amount of dolomite hydrate used was 38.8 kg per ton of oil. The maximum sulfur removal was 91 %, attained with a dust particle diameter between 10 u and 60 ,u and a contact time in the flame of about 5 seconds. The process can be applied to coal-fired furnaces as well as to oil burners. The total cost of sulfur removal would chiefly depend on the price of the dolomite: it might be $0.25 per ton of oil per percentage of sulfur.
Wet catalysis
It has already been said that a wet scrubbing process using water would be satisfactory only if almost complete removal of the SO₂ were obtained, leaving only small traces of this noxious material in the flue gas. Such a process has been recently developed by Pauling at the Esso company in Hamburg and adopted by the Lurgi company of Frankfurt under the name of the Sulfacid Process. It is based on the wet catalysis of SO₂ to H₂SO₄. The flue gas containing SO₂ is passed through a fixed bed of a granular carbon-containing catalyst, where the SO₂ is oxidized to SO₃, which then con-denses with water to form H₂SO₄. If the flue gas, the catalyst, the SO₂ content, the oxygen, the water and the other components of the process are in the right ratio, the dilute H₂SO₄ drops from the catalyst. Additional water and oxygen must be provided to the reactor. The dilute acid is passed to a concentrator fed in countercurrent by the flue gas: the acid heats and leaves behind some of its water content. Fig. 6 is a diagram of the process. In the concentrator the flue gas temperature drops to about 60°C and H₂SO₄ is produced as a 75% solution. Of this, 70% is removed from the concentrator and the remaining 30% forms an aerosol which is captured by a special demisting device. Dust up to 100 mg/r8 present in the flue gas would also be removed in the concentrator. The operating cost is estimated to be $2-S2.50 per ton of oil with a sulfur content of 2.5%.
FIG. 6 REMOVAL OF SO2 BY THE SULFACID PROCESS
This process and also those for dry removal of SO₂ are in various stages of development, some of them, as shown, having gone beyond the bench-test and the pilot study. Definite statements as to the applicability of these processes, however, can be made only when the experience gained from full-scale plants becomes available.
Source: Helmut Kettner
The 10 largest coal producers and exporters in Indonesia:
