The Sulfite Process
( Originally Published 1920 )
THE second chemical method for pulp making was the sulfite process, invented by Benjamin C. Tilghman, who experimented with the effect of sulfurous acid on wood shortly after the Civil War. His digester was a rotating, horizontal cylinder lined with lead and about 50 feet long by 3 feet in diameter. A worm made of lead plate was supposed to carry the pulp through in one direction, while the cooking liquor was forced through in the other. The product he obtained was satisfactory, but leakages and other difficulties led to his abandoning the development.
In 1870 Fry and Ekman in Sweden carried the work further and their improved process came into use, at first secretly, but finally openly in England about 1880. This process was the one used in the first American mill, which was that of the Richmond Paper Co., built in 1882 at East Providence, R. I. This mill had a capacity of about 15 tons of book and newsprint per day; it was operated only about 5 years because of financial difficulties.
Other variations of the sulfite process were quickly developed. In 1876 there was the Mitscherlich process, using lower pressures, longer times, and indirect heating by means of steam in copper coils within the digester. Between 1878 and 1882 another modified process was developed by Ritter and Kellner in Austria. The time of cooking was considerably reduced by admitting steam directly to the digester contents, hence it came to be known as the "quick cook" process.
About 1885 the patent rights for the Mitscherlich process were obtained for use by a group of mills in Wisconsin, and in 1887 the first commercially successful sulfite mill in America was built by G. N. Fletcher in Alpena, Michigan. This mill continued in active production until 1940.
The patent rights for the Ritter-Kellner process were acquired about 1886 by the American Sulfite Pulp Co. These patents covered the digester, the method of making the acid cooking liquor, and all features of the system.
In a trade journal of 1888 is an interesting note that the Bremaker-Moore Paper Co. at Louisville, Ky., had four of the Bremaker patent sulfite boilers in successful operation, and that they employed a female chemist who was an expert in her profession and gave entire satisfaction in her work.
Even with all these small beginnings the pulp making industry had not yet gathered much momentum, and in 1891 the Forestry Division of the U. S. Department of Agriculture estimated that the total wood pulp production of all kinds in America amounted to about 2000 tons per day. Doubtless much the greater part of this was groundwood, and the remainder chiefly soda fiber. Since that time sulfite production has grown in importance until it far exceeds that of the soda process, but recently it has had to yield first place to the sulfate process.
The sulfite pulping process depends on two principal reactions, which probably take place simultaneously to a considerable ex-tent. These are the splitting of the cellulose-lignin complex of the wood by hydrolysis, and the combination of the lignin with the calcium bisulfite to form calcium lignosulfonate, which is soluble in the hot cooking liquor, and can be easily washed out at the end of the cook. The hydrolytic reaction also causes the easily decomposable hemi-celluloses to dissolve with the formation of sugars. Fortunately the chief portion of the cellulose is relatively stable under conditfons of cooking, though it is not entirely unattacked.
These reactions are less drastic than those of the alkaline pulping processes, so the wood must be selected more carefully. There is no saponification reaction in the sulfite cooking process, so resins in the wood remain largely unattacked and insoluble. This limits the process to the less resinous woods and prevents its use for the highly resinous southern pines, unless they are cut before the formation of heartwood commences. The woods most commonly used for sulfite pulp are spruce, hemlock, balsam and the true firs; the woods of many deciduous trees can be cooked by this process, but are not largely used because their short fibers impart low strength to papers in which they are employed.
Because of the lower solvent power of the sulfite liquor the presence of bark and knots is likely to cause more dirt in the pulp than in the alkaline processes. Decayed wood is also bad for the same reason, though it seems to have less influence on the strength and yield of pulp than it does in alkaline cooks. Except for these notes the preparation of the wood is sufficiently covered in Chapter III of this book.
The cooking liquor for the sulfite process is made by burning sulfur to form sulfur dioxide, and absorbing the latter in water in the presence of calcium hydroxide or carbonate. This causes the formation of calcium sulfite, which is relatively insoluble, but which dissolves as calcium bisulfite when more sulfur dioxide is passed in. The solution containing calcium bisulfite and an excess of sulfur dioxide forms the usual cooking liquor, or "acid," for the sulfite process. In some plants a high magnesia lime, containing about 40 per cent magnesia, is used; in others a high calcium lime which contains about 92—95 per cent calcium oxide. An acid pre-pared with a sodium base instead of calcium can be used equally well, and even has an advantage in cooking those woods which do not permit ordinary acid to penetrate easily. An all-magnesia base acid is being used in a few mills, as is also an ammonia base acid. These last three—soda, magnesia and ammonia—enable the waste liquors to be treated for recovery of the base constituent, and so aid in avoiding stream pollution, which is becoming a more and more serious problem.
Sulfur dioxide was formerly prepared by burning iron pyrites, but in America this has been almost entirely replaced by sulfur. Like all other equipment used in papermaking, sulfur burners have undergone tremendous changes with the advances in chemical technology. The old type of flat retort burners, into which the sulfur was fed by hand through a door at one end, gave gas of irregular composition because of the sudden inrush of air when the door was opened. Modern sulfur burners are of several types, of which the rotary and spray burners are examples. The rotary burner receives its supply of sulfur from a hopper, or from a melting tank, from which the sulfur flows to the burner through a steam jacketed pipe. The rotation of the burner carries the sulfur up its sides and thus presents a continually renewed surface for combustion. At the end of the burner is a large combustion chamber in which the burning of any sublimed sulfur is completed.
In the spray burner, molten sulfur is pumped through a spray nozzle into a horizontal combustion chamber lined with brick. A waste-heat boiler is usually installed in connection with this type of burner to cool the gases somewhat, as well as to utilize some of the heat for generating steam. Both the spray and rotary burners are far superior to the old flat burners with regard to speed of starting and stopping, as well as in giving gas with a higher percentage of sulfur dioxide and less sulfur trioxide.
In operating any type of sulfur burner the most important factor is the regulation of the air supply. For complete combustion a pound of sulfur requires a pound of oxygen, which is the amount present in about 54 cubic feet of air. If much more is used, especially if it contains much moisture, there is greater formation of sulfur trioxide, which causes loss of both lime and sulfur. If too little air is admitted after the burner has become thoroughly heated, sublimation of sulfur may take place, and unless this is properly taken care of in the combustion chamber it may contaminate the acid with free sulfur, or lead to the formation of thiosulfuric and polythionic acids in the cooking liquor. All of these conditions may lead to the formation of calcium sulfate in the digester, with possible circulation troubles and inferior pulp. The actual operation of the burner is judged largely by the appearance of the flame; if it is operating satisfactorily it is blue, sometimes tipped with white; if brown fumes show, it is an indication of sublimed sulfur caused by too hot a furnace.
The gases leave the combustion chamber at the end of the burner at a temperature ranging from about 1300° to 1800° F., and immediately pass through iron pipes to a cooler. Up to this point the gas is hot and dry enough to have little effect on iron, but for the cooler, and all piping beyond, lead must be used. The cooler usually consists of a series of pipes through which the gas passes hack and forth. The pipes are placed in a trough of running water, or are so arranged that a thin film of water flows over them. A more modern method of cooling is to introduce the hot gases from the sulfur burner into the top of a tower lined with acid-proof tiles and at once spray them with cold water. This cools the gas almost instantaneously to about 155 degrees F., and the cooling water is used over again after passing it through heat exchangers. Uniform and regular cooling of the gas is important because both the rate of absorption and the amount dissolved in the next step in the process depend largely on the temperatures—the lower the temperature the better the results obtained. It would, of course, be desirable to use liquid sulfur dioxide if it could be obtained at a low enough price. This would insure a cool, clean, dry gas free from sulfur trioxide and the absorption would be excellent. Unfortunately cost considerations rule it out in nearly every case.
The next step in the preparation of the cooking acid is the absorption of the gas in water and its conversion into bisulfite. There are two types of apparatus for doing this, of which the oldest and most generally used consists of towers packed with large lumps of limestone. Early towers were of wood, but modern ones are generally of reinforced concrete with acid-resisting tile linings. Mitscherlich type towers are 6 to 10 feet in diameter and from 100 to 150 feet in height, and the stone rests on heavy oak beams placed about 6 to 10 feet from the bottom. Such towers are frequently built in groups of four with the whole surrounded by a structure with stairs, platforms, and stone hoist. Often each tower is divided into sections by timber gratings to aid in filling with stone and regulating the flow of gas. Some towers act as chimneys and require no artificial draft, but with others it may be desirable to force the draft by placing a fan between the sulfur burners and the base of the tower.
Ritter-Kellner towers are smaller than the Mitscherlich and are usually built in pairs. The acid from the base of one tower is pumped to the top of the second, while the gas from the top of the second is returned to the base of the first, thus avoiding any appreciable loss of gas. This is essentially the principle of the Jenssen system, which is widely used in America. In this system the towers are so cross-connected that either can be used for making strong acid, while the other recovers the gas not absorbed in the first.
In both these tower systems an ordinary dense limestone is used, preferably one as low in magnesia and as free from dirt, iron and silica as possible. Water is discharged over the stone by spray pipes, or similar device at the top of the tower, and forms a thin film on the lumps as it passes downward. This film absorbs the ascending gas very rapidly, forming sulfurous acid which reacts with the limestone to form calcium sulfite. This is then dissolved by more sulfurous acid. Both these reactions take place at the same time and if the balance between gas, water and temperature is maintained at the proper point little difficulty is encountered. Sometimes an insufficient flow of water, or a gas low in sulfur dioxide content, may cause the formation of crusts of calcium sulfate, or monosulfite, which may seriously interfere with the satisfactory operation of the tower.
The other type of apparatus for making acid is the milk-of-lime absorption system. In the Barker apparatus, which may be considered typical, this consists of a high tank or tower made of steel, lined with acid-proof brick, and divided into three or more sections by horizontal, perforated partitions. A continuous flow of milk of lime is supplied to the upper compartment where it meets gas bubbling up through the perforations of the false bottom. The weak liquor thus formed passes down to the next compartment through an overflow pipe, where it absorbs more gas. After passing in this way through the upper sections of the tower it flows onto a distributing plate and thence into the lower part of the tower. This part is filled with some type of brick or checker-work filling to expose a large surface of the liquor to the strong gas, which enters the system below a gas distributing plate under the checker-work. The finished acid leaves the system through a pipe at the bottom. To maintain proper conditions in the tower the top section is connected with an exhaust pump which removes unabsorbed gases and keeps up a sufficient vacuum to draw the gas in at the bottom of the tower.
The milk of lime for such a system as this is prepared by slaking well-burned lime, diluting and screening through a 60-mesh brass sieve. The lime must be cold when it goes to the absorption towers. It is customary to use dolomitic lime containing about 40 per cent of magnesia, but ordinary high-calcium limes can be used without difficulty.
In whichever way the acid liquor is prepared it is necessary to store it until it is desired for use, and this is usually done in wooden tanks made of southern pine or Douglas fir heartwood without lining. As it enters these tanks from the absorption towers it contains too little sulfur dioxide for cooking. This is corrected by passing the gas relieved from the digesters into the storage tanks until the proper strength is reached. This also serves to keep the liquor warm, which is desirable because of the subsequent saving of steam in the cooking process.
The digesters used for cooking by the sulfite process have passed through many stages of development before reaching their present condition. The acid acts so destructively on iron that some form of protective lining is necessary. Early attempts to avoid linings by using digesters of cast bronze led to several disastrous explosions before their use was abandoned. Lead resists the action of the acid satisfactorily, but when heated it expands nearly twice as much as iron, and on cooling it does not return to its original size. This leads to "crawling" and "buckling," and cracks are likely to appear after a short time. In vertical digesters there is also a general creeping downward of the lead lining, causing the upper part to stretch, become thin and finally give way.
The Mitscherlich lining included a coating of tar or pitch applied directly to the digester shell, then a thin lead lining with the edges burned together, and finally two layers of dense vitrified bricks, with tongues and grooves, which were sometimes laid in Portland cement. This is the first lining in which bricks were used.
Modern digester linings are usually of acid-proof bricks, 2 to 3 inches thick, backed by about an inch of cement next the shell. The cement for backing and pointing the first layer of bricks varies more or less in composition, but includes Portland cement, crushed and sifted acid-proof brick or quartz sand, and silicate of soda, mixed to the right consistency. The inner layer of bricks is pointed with a mixture of sand, Portland cement, litharge and glycerine. In some installations this has proved more durable than the bricks. Lead sheathing back of the bricks and cement is no longer used because the steel shell is pierced with numerous small telltale holes which enable a leak in the lining to be located approximately.
The form and size of the digesters have also undergone considerable change since the development of the process. As mentioned, Tilghman's first digester was a horizontal, rotating vessel 50 feet long by 3 feet in diameter. The digester in the Fletcher mill is said to have had a capacity of 25 cords of wood. Rotary, globe digesters were also used in some mills, but the trend has been to stationary, vertical digesters of larger and larger size, and generally with top and bottom cones, though the top is sometimes dished, as in soda digesters. The sizes vary greatly in different mills, and with the age of the installation; they may range from 10 to 18 feet in diameter inside the shell and from 25 to 65 feet in height. Because the lining occupies considerable space, the capacity in cubic feet is not that which would be calculated from the dimensions given. Actual capacities may vary from about 1,000 to 14,000 cubic feet, and assuming that a space of 450 cubic feet is needed to produce a ton of dry pulp, the productive capacity of such sizes would be 2.2 and 31.1 tons respectively.
The method of carrying out a sulfite cook depends on whether the Mitscherlich (slow cook) or Ritter-Kellner (quick cook) process is used. In the Mitscherlich process the digester is filled with chips which are then steamed for several hours with direct steam, but no pressure. The condensed water runs to waste as a brownish liquid. After steaming, all valves except that leading to the liquor storage are closed and the partial vacuum caused by condensation of steam draws the cooking liquor in rapidly. Steam is next admitted to the coils of copper, hard lead, or stainless steel which are placed in the bottom of the digester, and which supply the only source of heat for the cook. Because of the slow transfer of heat and the large size of the digester it may require as much as 12 hours to raise the temperature of the charge to 230°F. When pressure is reached a valve is opened two or three times in the next hour and air and gas allowed to escape. Since no liquor escapes, the gas can go to the storage tank, or recovery system, to reclaim the sulfur dioxide. The temperature of the charge is gradually raised to 240°F. for the rest of the cooking period, during which time the pressure should not exceed 80 pounds. Shortly before the end of the cook the steam is shut off and the pressure reduced by relieving gas. The liquor is then blown out by the remaining pressure and the pulp washed with water once or twice in the digester before removing it through the bottom openings.
The original very long cooks of this process have been gradually shortened to 20 to 30 hours depending on the pressure and temperature. In modern European practice the actual time for cooking varies from 12 to 22 hours, with an additional 6—7 hours for filling with chips, steaming, adding acid, blowing off gas to reduce pressure and washing. The advantages of the process are a high yield and extra strong fiber.
The quick-cook process is generally used in America. This differs from the Ritter-Kellner process, from which it was derived, in blowing out the entire charge under pressure, rather than washing in the digester. The chips are run into the digester from the top, as in the soda process, and the cooking acid is pumped into the bottom, usually starting its admission before all the chips are in. The volume of acid varies somewhat in different mills, but a fair average would be about 1,200 gallons per cord of wood. When the digester is filled the head is closed and steam introduced into the bottom of the lower cone; in many mills it is also added through nozzles near the top of the cone, directed upward to give better circulation. The steam required is much greater than for a soda cook of the same number of cords; it varies greatly in different mills, and in cold climates with the time of the year, the minimum being around August and the maximum about February. American mills use 4 to 4 1/2 pounds of steam per pound of fiber produced while European mills require about 2 to 3 pounds; the amount also depends to some extent on the method of acid recovery used, being less with the modern hot-acid systems.
In steaming a cook it is very important to raise the pressure slowly, so that the liquor may penetrate the chips completely before the temperature reaches a point which will cause their centers to become hard and dark brown in color. The success of sulfite cooking depends very greatly on the control of the steam supply throughout the entire cook, and there are two methods of operation. In one the steam is admitted at a rate to give the temperature shown on a predetermined schedule, and the digester pressure is maintained by hand operation of the relief valve; in the other the digester pressure is controlled automatically—as the relief valve is opened the steam valve also opens to maintain the desired cooking pressure.
No hard and fast rule for cooking can be given; each mill has a method which experience has proved satisfactory for its particular requirements. Merely as an example, and not in any way as a general procedure, the following schedule may be given as that producing an easy-bleaching pulp: Steam at such a rate as to reach 75 pounds pressure in 2 to 3 hours, then open the relief valve and by further steaming bring the temperature to 240°F. in about an hour. Close both steam and relief valves for about 1% hours, then turn on the steam and open the relief valve slightly so that the maximum temperature of 300°F. will be reached in 10 hours while the maximum pressure is still kept at 75 pounds. Shut off steam at this point but continue the relief until the pressure drops to 50 pounds in about 1 to 2 hours; then discharge the digester through the bottom blow valve into the blow pit.
Relieving, which has been mentioned several times, is necessary for two reasons. The large amount of steam which is condensed during the cooking would fill the digester and prevent further heating unless some of the liquor were removed. Also heating the acid, and the chemical reactions which take place, cause gas to be evolved and this builds up pressure in the digester beyond the safe working point. In practice relieving is done from the top or the side of the digester, or sometimes from both points. The schedule for relieving varies in different mills and with different types of cook and no general rule can be given, except that it must be sufficient to keep the digester pressure at the point considered desirable for the type of fiber being made.
The material relieved from the digester is partly fluid and partly gaseous and in the interests of economy recovery of the useful portions is necessary. Again there are various ways of accomplishing this but essentially they consist in sending the fluid portion which is richest in sulfur dioxide to the tanks containing the acid from the absorption towers, and in using the gaseous portion to build up the acid to the strength desired for cooking. In another method, known as the "hot acid" system, both liquid and gas are collected in a pressure accumulator, which is a vessel capable of withstanding heat, pressure, and the corrosive action of the acid. This system also involves certain changes in the filling of the digester with acid, which are said to be beneficial by aiding penetration of the liquor into the chips and by giving cooks of greater uniformity.
The acid used in cooking contains free sulfur dioxide (SO2) as well as that combined with the base present, and it is not possible to say that there is any one "best acid." The acid made in the towers before being enriched by recovered gas, and the finished acid ready to use, may fall within the following limits expressed in grams per 100 cubic centimeters.
Tower Acid Finished Acid
A high amount of free sulfur dioxide aids the penetration of the liquor into the chips, but if the combined SO2 drops below 0.9 per cent the cooks are likely to be raw and incomplete. Between 0.9 and 1.1 per cent combined SO2 is a favorable condition for easy-bleaching pulp, while with a higher percentage of combined SO2 the pulp becomes harder to bleach unless the cooking time is considerably increased.
It is obvious that making fiber by the sulfite process is a complicated procedure and that if uniformity of product is desired very careful control tests must be made all along the line. Such tests start with the analysis of the sulfur burner gases to see if conditions are right for the highest strength of gas. Tests after the cooler follow, to indicate possible leaks of air into the gas; then the "raw" acid from the towers must be analyzed, and also the enriched cooking acid, to see that each is uniform and up to strength. Finally the end of the cook is determined by chemical analysis of a sample of the liquor drawn from the digester, as well as by comparing the color of this liquor with a color standard, which is sometimes made from an infusion of coffee, but may be of oil, or iodine, or any solution of the right color, which is only slightly influenced by long standing and use. Some mills use for comparison a sample of the liquor from a preceding cook, which proved to be of satisfactory quality.
Prior to the completion of the cook the admission of steam is stopped and the relief is continued at such a rate that the pressure in the digester is reduced to the desired point at the same time the tests indicate that the cook is done. A pressure of 50 pounds or less has been found satisfactory, but for some grades of pulp, "blowing," or discharging the digester, is started at the full pressure used in cooking. If blown at too high a pressure, undercooked chips are likely to be blown to pieces and to form undesirable "shives" in the pulp.
The blowing of the digester is done by opening a valve in a large pipe connected with the bottom of the digester and with the blow pit, into which the digester contents are forced by the residual pressure. The blow pit is constructed with a false bottom, sometimes of perforated tiles, sometimes of wood covered with coco matting, and sometimes of perforated stainless steel plates. All of these serve to retain the fiber and allow the liquor to drain away, either to waste or to a recovery system if waste products, or the base used, are to be reclaimed. Before blowing, the blow pit drain valve is closed and the pit flooded with water to a depth of about a foot. The blow valve is then opened and the digester con-tents discharged against a "target" in the end of the pit opposite the blow pipe. This target is a heavy bronze or hard cast iron plate and it serves to break up the fibers from their chip form, and to prevent wear on the blow pit wall. The top of the pit is connected to a flue, or stack, large enough to permit all steam and gas to escape quickly without causing too much pressure on the pit.
After the blowing operation, the fiber in the pit is washed by flooding it with water and draining it several times until it is sufficiently free from the dissolved impurities to be screened for removal of dirt, knots, slivers and partly cooked chips. After this it is ready to be used in the unbleached condition for certain grades of paper, or to go to the bleaching system to be made into pulp for use in white papers.
The waste liquor from the sulfite process is a cause of serious stream pollution. For every ton of fiber produced, approximately a ton of material is contained in solution in the waste liquor, and as this is greatly diluted during blowing and washing, any process of disposal which involves evaporation is necessarily an expensive one. Many efforts have been made to prepare useful by-products from the waste liquor, but none has developed to a point where it takes more than a small portion of the total which has to be disposed of in some way. Among the uses for waste liquor, or the products made from it, may be mentioned its employment as a road binder, as an adhesive in preparing coal briquettes, and as a constituent of adhesives for laying linoleum and wall coverings. One mill is producing vanillin from it, but it supplies enough of this flavoring agent to satisfy practically all the needs of the country. The fermentable sugars in the liquor can be converted to alcohol, or used for growing yeast; tanning materials can be made from the waste; and numerous other products of limited use can be prepared, but no reasonably profitable process of wide application has yet been developed.