Recovery of Sulfur Dioxide from Waste Gases Regeneration of the Absorbent by Treatment with Zinc Oxide H. F. JOHNSTONE AND A. D. SINGH University of Illinois, Urbana, 111.
I
' N PREVIOUS papers of this
Chemical regeneration can be A new process for the regeneration of accomplished by the use of any series' equilibrium partial sulfite-bisulfite solutions used for absorbing material that will react to repressure data have been sulfur dioxide from waste gases is demove the sulfur dioxide from the reported for a number of sulfitescribed. . I t is based on the precipitation solution. For example, the adbisulfite systems. Their appliof zinc sulfite, followed by decomposition dition of lime to a sodium sulfitecation to the cyclic process of bisulfite solution would precipiabsorbing sulfur dioxide from of the dried solid in a flash calciner to give tate calcium sulfite and remove dilute waste gases, followed by pure sulfur dioxide and zinc oxide, the latter the sulfur dioxide. I n this case, heating t o regenerate the solbeing recycled. A thorough study in the however, the net result would . vent, has been described. I n laboratory and pilot plant has given imbe the same as if the gases were such a process a limitation is portant information on several of the unit washed directly with lime slurry, set by the extent to which the and an equivalent amount of sulfur dioxide can be removed operations. The paper includes a discuslime mould be consumed. The from the solution without exsion of the design of flash calciners based two-stage process (8) has the cessively increasing the heat r e on the transfer of heat to particles susdistinct advantage of circulating quired for the regeneration. I n pended in gas streams. Data are presented a clear solution through the t h e c a s e of t h e a m m o n i u m on the rate of oxidation of sulfite-bisulfite scrubber without any possibility sulfite-bisulfite so 1u t i o n s t h e of scaling the surfaces, which is stripping cannot be carried besolutions in the recovery process from an inherent difficulty in the lime yond a certain ratio of sulfite to actual plant operation, and a new flow sheet process (18). A similar reaction bisulfite even by using large for desulfation is described. quantities of heat or steam, owtakes place when other metallic oxides which form insoluble ing to the volatility of ammonia over solutions of ammonium sulfite a t high temperatures. sulfites are added to the solution. Of these, zinc oxide Since regeneration by heating must always depend on vaporis most suitable, not only because it is readily available but especially because zinc sulfite is easily decomposed by heating ization of the sulfur dioxide, some small equilibrium vapor pressure must remain over the regenerated solution. When this to about 260' C. (500' F.). The decomposition yields zinc - solution is cooled before being returned to the scrubber, the oxide, sulfur dioxide, and water. By combining the two cycles of absorption, regeneration, and calcination, sulfur vapor pressure of sulfur dioxide is decreased but remains finite dioxide can be removed from the gases and recovered in so t h a t the gases leaving the scrubber cannot be freed completely of the sulfur dioxide. Regeneration by heating, therepractically pure form. The only other metallic sulfite t h a t fore, cannot be used for very dilute gases. Solvents which give compares with zinc is that of magnesium. A study of i t s decomposition, however, has shown that it must be heated to a high-temperature coefficient of the vapor pressure of sulfur dioxide would be most desirable. Unfortunately, with the ex650" C. (1200' F.) and a t this temperature about 20 per cent ception of certain two-phase systems which present difficulties is converted into sulfate. Calcium sulfite decomposes a t in handling, those that have been suggested as having this about 1050' C. (1920' F.) to give the sulfide and sulfate, a s property have too low capacity for use with dilute gases. well as the oxide and sulfur dioxide (4). Zinc oxide, thereThe substitution of a strictly chemical method of regenerafore, is the most desirable regenerating agent of this type. tion for the heating step would remove the disadvantages of the latter and would, in turn, have the following advantages: The Flow Sheet (16) faster absorption of the sulfur dioxide due to more complete The flow sheet of the process is shown in Figure 1. The regeneration, removal of the necessity for nearly saturating waste gases are scrubbed in a suitable scrubber with an aquethe solution in order to regenerate economically, more comous solJtion of sodium sulfite and bisulfite. The sulfur diplete removal of the sulfur dioxide from the gases, the possioxide absorbed exists in solution as an increased ratio of bibility of using the process on more dilute gases, and the possisulfite to sulfite by virtue of the reaction: bility of using sodium sulfite-bisulfite solutions rather than ammonium solutions and thus reducing the cost of wastage SOL SOs-H20 2HS03losses. After passing through the clarifier, which separates the sus1 Previous papers in this series of articles appeared in IND. ENQ.CEEM.,37 pended solids removed from the gases along with the sulfur 587,659 (1935); 41, 286, 1396 (1937); SO, 101 (1938). '
I
+
1037
+
INDUSTRIAL AND ENGINEERING CHEMISTRY
1038
VOL. 32, NO. 8
5090 Lhl W G A S 450
H,O
LbsOIL
im,wo CU.FT Q 300 D e G F
5129 LhDRY GAS
9.6 '
CoSO.ZH.0
1.6 Lbr ASH
FIGURE 1. FLOW SHEETFOR SULFUR DIOXIDERECOVERY, USIXGZINC OXIDEFOR REGENERATIOX Figures given represent quantities flowing in one minute; quantities in parentheses are make-up for losses. dioxide, the solution is sent to a mixer where it is treated with zinc oxide. Here the following reactions take place: ZnO
+ NaHSOa + 21/,H20 +ZnSOs.21/.H20 + NaOH NaOH + NaHS03 NazSOs + HzO
and the original ratio of bisulfite to sulfite is restored. The mixture is agitated to promote the growth of the crystals and is then filtered in a continuous filter. Make-up water is added here as wash water to compensate for that evaporated in the scrubber and lost a t other points. Incidental losses of solution are made up by addition of soda ash, which is immediately converted to sodium sulfite. The zinc sulfite cake is sent through a dryer and then calcined in a flash calciner heated by indirect heat. The hot combustion gases from the calciner serve as drying gases in the dryer so that efficient use is made of the fuel. The gases evolved from the calciner consist of approximately 40 per cent water and 60 per cent sulfur dioxide by weight. After passing through suitable coolers, dryers, and compressors, the product may be obtained as liquid sulfur dioxide in nearly pure state. X o carbon dioxide can pass over into the product since this gas is not soluble in the absorbing solution under the conditions used. The zinc oxide obtained from the flash calciner is in the form of a light, fluffy, and extremely active material. It is sent directly back to the cycle. I n the operation of the process some of the dissolved sulfur dioxide is inevitably oxidized to the sulfate. Sulfate also accumulates in the solution from the small percentage of sulfuric acid vapor in the gases and by oxidation of the zinc sulfite. Since zinc sulfate is soluble, the net effect of this oxidation is a decrease in the sodium concentration available for forming the sulfite and bisulfite. This can be overcome by discarding the solution in the underflow from the first clarifier and making up with soda ash, or some process of removing the sulfate may be used. A cost estimate on these two alternatives has shown that the latter is the less expensive.
The following process for desulfating has been found satisfactory: The correct portion of the solution is withdrawn as a side stream with the underflow from the clarifier. The underflow contains, in addition to the ash particles, precipitated calcium sulfite, the source of which is described below. This thin slurry is treated with sufficient sulfur dioxide to convert all of the sulfite to bisulfite. Under these conditions the sulfate is precipitated as calcium sulfate. The presence of the ash facilitates the crystallization, and the two solids are removed together in a small centrifugal or in a filter. The desulfated solution, containing calcium ions, is now treated with lime in sufficient quantity to convert part of the bisulfite to calcium sulfite. This restores the side stream to approximately the same pH as the main stream. Since calcium sulfite is insoluble, the treatment ensures against calcium being precipitated along with the zinc sulfite in the regenerator. The calcium sulfite slurry, with the desulfated solution, is then returned to the clarifier along with the scrubber efffuent. The calcium sulfite crystals settle rapidly and aid in removing the fine ash particles from the suspension. The quantity and concentration of the underflow removed is fixed by the amount of sulfate formed per cycle of the absorbing solution. Each operation in the process has been thoroughly investigated, both in the laboratory and in a small pilot plant. The work has resulted in new information on the principles underlying clarification, filtering, drying, and calcination. The reader interested in these subjects should refer to the details published elsewhere (15). I n this paper only the results on the flash calcination and on the oxidation and desulfation will be given. The former represents the development of a new unit operation on which important information has been obtained. The work on oxidation and desulfation has an important bearing on the economic possibilities of the process and is inevitably connected with any process of sulfur dioxide recovery.
AUGUST, 1940
1059
INDIISTRLAI. A 4 D K Y ( i l Y W K I N G CHEMISTHY
iilent.ii:al x-ray pLttel-trs ( Y i R t i n 2c) in a r m t p i ~ sulfitef siiriirirn ratio l)et,wem 0 . i and 1.0. (4) At 35' C. a double salt, Z i ' a ~ ~ I ) ~ . : ~ Z n ~ O ~is . 4 Hforiried ~ O , as large granular I titls wit11 many cleavage plane?, at coiiccntratioiis of iirii a h w e 1.7.5 nrdea per 100 moles of water and over a wick range of thc rnbio (Figure 2c). This salt was riot f o u ~ ~ d
Precipitation of Zinc sulfite The reaction hetween zinc oxide and the sidfite-biaultitr: solution inay be accomplished without occlusion of thc oxide aiiil with rirarly theoret,ieal effieieiicybctween certain limits OS operation. A careful s t u d y of tlie phase equilibria ha.s in& cateii that iLt least, four types of soliil compounds can he forrrirxi as follorr-s: ( I ) At ratios of sulfiw dioxiiic to sodiarn betwvccn 0.5 arid 0.05, a hasic sulfite having the composition ZnO: 2ZnSOa.:jIlr0 is formed with fairly defiiiite mil-shaped crysralr, as slrowii in Figiirc 2a. (2) A normal sulfite, apparentlv iiaviirr t,he composition ZoSO3.2€&O, exists its ncixlldiki; crystals at slightly higher ratios of siilfitn t,o sodiiirn at :%3' (,'. (%OF.) hit. was mt foiiiiii n t 50" (!. (122'F.). (3) Thr i i i o c o mmmon Zr~'il)~.%' .K3O eaiPts i o several erystsl hehits with
at 50"C. bclow 3.0 moles (if sodium per 100 inolei of water but is iitidoubtcdly formed at higher concent.rations. Since, on tlie one hand, the Eormatioii of t,Iw basic salt is unilesirable I m m i s c of t h e iinfamrahlc h i p c of tin! crystals for filtering mil t,hc lower sulfur dioxide content, and, on the otlier hand; the iloiib!e salt is umilesirable becansc of excessive oxidation of the combined soiiiurn sulfite during calcinatioii, the operating h i t s for tho solution are effectively set as being below 3 irioles of sodilini per 100 mrdw of ivvxtiir arid diove w iulfitc-
c
d
b
d
c
e
INDUSTRIAL AND ENGINEERING CHEMISTRY
1040
TABLE I. RESULTS OF
C.4LCINATION O F ZINC SULFlTE I N
Feed Direction of Gqs Flow in Calciner
Rate of Feed Lb./(MiA.) (Sq. Ft.) A. 6.8 6.6 6.9 6.8 7.4
~-II\JCH FLASH CALCINER Product
Lb.
Lb. free oxidized Lb. free Wall Screen SO? per SO2 per SO? per Temp., Size, % 100 I b . 103 1 b. 103 lb. Run h-0. c. Mesh ZnO ZnO ZnO ZnO Wall Temp., 450' C . ; Av. Feed Rate, 6.9 Lb./(Min.)(Sq. Ft.) 40-60 35.7 Down 29 450 78.7 0.0 58.0 30 Down 425 60-80 37 0 78.7 0.0 61.8 80-100 37.2 Down 450 78.7 0.0 56.3 31 32 Down 450 100-140 37.4 78.7 0 0 61.4 450 >140 36.2 Down 33 78.7 0 0 59.2 B. Wall Temp., 600' C.; Av. Feed Rate, 6.9 Lb./(Min.)(Sq. Ft.) Down 34 so0 6.8 40-60 35.7 78.7 0.0 33.6 090 7.0 60-80 37.0 35 Down 78.7 0.0 31.9 600 6.7 80-100 Down 37.2 78.7 0.0 25.8 36 590 6.6 100-140 37.4 78.7 0.0 37.0 37 Down >140 600 7.6 36.2 78.7 Down 0.0 20 1 38 C. Wall TemD.. 600' C.; AT. Feed Rate. 11.1 Lb./(hfin.)(Sa. . . . . _ Ft.) . 610 11.4 60-80 37.0 78.7 Down 0.0 42.1 44 590 10.1 80-100 37.2 78.7 0.0 42.6 Down 45 600 11.1 100-140 37.4 78.7 0.0 42.3 Down 46 600 12.5 >140 36.2 78.7 0.0 35.6 Down 47 D. Wall Temp., 740' C.: AI-. Feed Rate, 3.6 Lb./(hfin.) (Sq. Ft.) Rlixedo 750 2.3 45.7 73.7 Down 51 30-40 740 4.2 34.8 49.2 9 40-60 740 4.2 35.1 51.0 8 55.5 3.6 37.2 60-80 740 7 UP 100-140 60.8 740 3.9 39.9 4 UP > 140 740 4.2 40.0 65.0 3 UP > 140 65.0 740 3.6 40.0 Down 2 > 140 65.0 740 3.6 40.0 1 UP 2.6 36.2 > 140 78.7 750 48 UP E . Wall Temp., 740' C.: Av. Feed Rate, 6.6 Lb./(Min.)(Sq. Ft.) 7.7 740 6.0 40-60 35.7 78.7 Down 0.0 24 60-80 37.0 78.7 10.6 0 0 750 6.0 Down 25 2 6.5 80-100 37.2 0 . 0 730 6 . 2 78.7 Down 26 100-140 37.4 6.6 78.7 25.0 0.0 730 Down 27 4 .8 100-140 40.4 1 5 . 9 740 6 . 3 62.8 5 UP >140 36.2 13.8 7.6 78.7 Down 0.0 730 28 3 . 7 >140 36.2 0 . 0 780 7 . 0 78.7 49 UP F. Wall Temp., 780' C.; Feed Rate, 12.3 Lb./(Min.)(Sq. Ft.) >140 36.2 78.7 0.0 18.1 50 UP 780 12.3 a Pure zinc sulfite (partially dehydrated), feed-milled but not screened.
3
sodium ratio of 0.65 during regeneration. Previous work ( I S ) showed that the vapor pressure of sulfur dioxide over these solutions is practically zero a t sulfite-sodium ratios of 0.85. Thus, the solution may have a capacity of 0.60 mole of sulfur dioxide per 100 moles of water, or roughly 2 per cent by weight, without exerting any vapor pressure of sulfur dioxide at any part of the process. I n the pilot plant the crystals obtained settled readily and were easily dewatered on a continuous filter or centrifuge. The range of operation in which the normal pentahemihydrate was formed was somewhat larger than that indicated above, and the crystal habit was evidently influenced by the impurities present in the solution. A typical photomicrograph is shown in Figure 2d, along with the x-ray pattern indicating similarity to 2c.
Calcination of Zinc Sulfite Vapor pressure measurements made on pure dry ZnSO8.-
21/2H20 indicate that the hydrate has an appreciable vapor pressure of water a t 70" C. (158" F.). This increases rapidly with temperature until the salt becomes deliquescent a t about 95" C. (203" F.). A t this temperature, also, decomposition of the sulfite can be detected. The vapor pressure of sulfur dioxide reaches atmospheric pressure at approximately 260" C. (500" F.). The results of batch decompositions of small samples in a muffle furnace are shown in Figure 3 . The first attempt a t continuous calcination of the sulfite was made using a small rotary kiln heated externally. Difficulties from clogging could not be overcome even when the feed was thoroughly dried. Apparently, a plastic stage was reached either by the formation of liquid water during dehydration or by fusion of the anhydrous material before decomposition was complete. These difficulties suggested that
VOL. 32, NO. 8
LO.
oxidized SO? per 100 Ib. ZnO
Release of Free
0.0 0.0 0.0 0.0 0.0
26.3 21.5 28.4 22.0 24.7
0.0 0.0
0.0 0.0 1.2
57.4 59.6 67.2 52.8 73.0
0.0 0.0 0.0 0.0 4-1.5
0.0 0.0 0.0 0.0
46.7 45.8 46.3 55.6
0.0 0.0 0.0 0.0
5.9 26.8 25.0 22.5 20.0 8.4 12.8 11.2 5.0
90.2 67.1 90.8 95.1 90.6 95.4 95.4 97.7 82.2
4-2.6 -5.7 -5.3
7.0 6.5 0.0 2.3 16.0 5.2 4 5
81.5 78.3 66.4 65.5 92.2 76.1 89.8
+S.8 +8.2
2.8
73.6
+3.6
%
so2
%.
Oxidation of Free
so1
0.0 0.0 0.0 0.0 0.0
-1.2
+3.4 -S.2 -1.4 -3.8 4-6.3
o.a
+2.8 +0.2 +6.6 +5.7
the calcination could be accomplished best in a space reaction in which the particles are heated while falling through a vertical tube (25). This method has recently been developed in Europe for heating blast furnace dust for briquetting, and its application to smelting copper and zinc ores and for the production of soluble phosphates has been suggested (23). The method, as finally developed in the pilot plant, gave such excellent results and a product of such unique character that a thorough investigation of the rates of heat transfer to falling particles has been made in order to have a basis for the design of large flash calciners for this and other purposes. At the pilot plant a gas-heated flash calciner was constructed consisting of a 10-foot section of 4-inch 0. d., ll-gage stainless steel tubing. This tubing was surrounded by cylindrical firebrick tile, of 6-inch 0. d. and 1-inch thickness, for protection against the direct action of the flames. The outer wall of the furnace was constructed of 4.5-inch firebrick with nine burner ports arranged for tangential flames. The temperature of the wall of the calciner was recorded by a sixpoint recording pyrometer operating on thermocouples peened into the metal. The variations of temperature from one point to another along the tube were not greatex than 100" C. (180" F.). I n the normal operation of the pilot plant the ~ n sulfite c from the dryer was passed through a small hammer mill before being sent to the calciner. The average rate of feed was 0.4 to 0.7 pound per minute, and the wall temperature was maintained a t 675" to 740" C. (1245" to 1365" F.). The effect of the particle size, wall temperature, and rate of feed on the extent of the calcination are shown by a portion of the results as reported in Table I. Column 2 shows the direction of flow of the sulfur dioxide and water vapor in the calciner. It is evident that downward flow gives a shorter time of exposure than upward flow, since the particles are swept
AUGUST, 1940
IPU'DUSTRIAL AND ENGINEERIXG CHEMISTRY
through the tube by the flow of gas. Upward flow, however, results in considerable material being lost from the top. I n the tests this was not collected and therefore is not included in the analyzed product.
same degree of completion. Upward flow of the gas provides greater decomposition of the particles with fall through the calciner. I n these tests the oxidation of the sulfite was erratic, which indicated that it was probably caused by leakage of air into the bottom of the calciner. When the original material v a s already partially oxidized, decomposition of the sulfate actually occurred a t high temperatures. The product of the flash calciner is a fluffy material varying in color from dark gray to n-hite, depending on the purity of the zinc sulfite. The bulk density of the dry product is as follon-s : Compression Lb./sq. m. 0 1.4 4.9
I
I
lb 15 20 25 TIME OF HEATING, HOURS
FIGURE 3. BATCHDECOJfPOSITION
30
15
O F Z I S C SULFITE
1041
Bulk Density Lb./cu. ft. 1.8 5.4 8.1
Compression Lb./sq. in. 11.8 23.4 46.6
Bulk Density Lb./cu. ft. 11.2 12.5 13.7
Under the microscope no crystalline structure of the calcined oxide can be detected, but examination by x-rays shows definite crystalline arrangement. The fineness of the powder accounts for its great reactivity. Its heat conductivity has been measured and found to be lower than that of cork and other insulating materials.
Design of Large Flash Calciners
The basis of the flash calciner is the rapid transmission of heat to the falling particles. Since no information on this Column 6 gives the actual percentage of zinc oxide in the subject was available and the results of Table I were inadefeed and indicates the purity, since the calculated percentage quate for design of a large-scale unit, measurements were in ZnS03.21/2H20is 42.7. Column 7 gives the weight ratio of made on the rate of heat transfer to several sizes of quartz, sulfcr dioxide present as zinc sulfite to zinc oxide, and column Carborundum, and Aloxite particles falling through air or 8 gives the ratio of sulfur dioxide originally present as zinc carbon dioxide (1). These substances were chosen because sulfite but oxidized before calcination. Columns 9 and 10 they are stable a t high temperatures and inert to the hot gases give the corresponding ratios after calcination. The last two used as well as to the mater in the calorimeter. The heat columns, respectively, give the percentage of the available transferred was measured by collecting the hot particles in a sulfur dioxide released and oxidized during calcination. A calorimeter and observing the temperature rise. negative value in the last column indicates that some of the I n order to calculate the over-all coefficient from the quanoxidized sulfur dioxide was removed. This could take place tity of heat transferred or, conversely, to calculate the amount either by decomposition of the sulfate or by reaction of the of heat transferred to a particle a t any point in a hot tube, it sulfate with carbon originating in the flue gases. is necessary to know the velocity of fall and the area of the Interpretation of the data in Table I must be made iyith particles. Rates of fall were determined by elutriation care. The degree of calcination depends upon the competing measurements in which the terminal velocity was observed rates of heat input and of fall through the heated zone. I n for several closely sized fractions of each material studied and the upper part of the tube the heat input is rapid, and there is for zinc sulfite and zinc oxide. The method of Martin (20) rapid acceleration of the particle. I n the lower part the was used in these measurements. 17hen a particle reaches particle is losing weight rapidly, but the gases in downward flow tend to smeer, i t onward. Two zones of reaction a t constant temperature exist-dehydration and decomposition. The latter requires the larger amount of heat. The degree of completion of the decomposition will depend upon the location of this zone in the tube. A few inches increase in the length of the tube may be sufficient to double the decomposition. The decomposition of large particles is small because of their high velocity. The rate of heat input depends principally upon the temperature of the wall, the ratio of wall area to particle area, and the screening effect of the cloud of particles on radiation, as determined by the diameter of the tube and the number of particles per unit volume. The results of the tests on the 4-inch calciner indicate that the percentage decomposition in the 10-foot length increases with wall temperature and decreases with increase in particle size and rate of feed. With the walls a t 600" C. ( 1 l l O O F.) the decomposition is 90 per cent complete for particles smaller than 80 mesh and for feed rates of 2.5 pounds per minute per square foot (not shown in Table I). At 740" C. the feed rate of this size particles may be increased to 6.5 pounds n i t h the THROUGH AIR FIGURE 4. DRAGCOEFFICIENTS FOR PARTICLES FALLING
1042
INDUSTRIAL AND ENGINEERING CHEMISTRY
its terminal velocity, the resistance due to the passage through the gas is equal to the force of gravity. In terms of the nomenclature adopted,
or
With the exception of the calcined zinc oxide, the density of the particles was determined by displacement of water or kerosene. I n the case of the porous zinc oxide it was calculated from the observed weights and diameters, assuming spherical shapes. The relation between the drag coefficient and Reynolds number for a number of materials, including the zinc sulfite from the dryer, is shown in Figure 4. It is apparent that only for the smallest size particles does Stokes' law hold. The mean diameter of the particles was determined by measurements of a large number on a petrographic microscope, using a slide specially prepared to give random orientation ( 2 7 ) . The area of the particles available for heat transfer was measured by projecting these slides and tracing the projected area of the images with a planimeter. The surface area was taken as four times the projected area. From the data on drag coefficients and diameters, the velocity of a particle falling from rest was calculated by graphical integration of the equation:
The resulting velocity-time curves were likewise integrated t o give the corresponding distance-time curves. Correlation of the heat transfer data indicates that transmission to falling particles takes place both by radiation from the hot walls and conduction-convection through the gases. I n small tubes with wall temperatures up to 650" C. radiation plays a relatively small part, accounting for, a t most, not more than 20 per cent of the total heat transferred; but with larger cross sections, particularly those of rectangular shapes, i t becomes increasingly important. The following equations represent the experimental data on heat transfer fairly accurately and are recommended for the design of large flash calciners: 1. The over-all heat transfer coefficient from the furnace wall to the particle is given by
2 . The heat transfer coefficient from the wall to the gas is given approximately by Rice's equation for natural convection (19):
The constant is dimensionless. The values of the physical properties of the gas are determined at the arithmetic mean temperature of the film. A t , must be determined by trial and error from a rate balance. 3. The heat transfer coefficient from the hot gas to the particle is high. It may be approximated by the theoretical equation: (4)
The proportionality constant is approximately unity as indicated. 4. The radiation heat transfer coefficient is given by the formula :
VOL. 32, NO. 8
The factor F A E includes the emissivity, size, and concentration of the particles, the length of the radiant beam, and the shape of the hot walls. It was estimated on the basis of the following considerations: a. The emissivity of finely ground zinc sulfite was assumed to be 0.5. Hild ( 7 ) reports values for several oxide powders, including zinc oxide. A t temperatures below 870" C. (1600" F.) the values vary between 0.25 and 0.40 for particle sizes between l and 120 microns. The values increase slightly with increasing grain size and temperature. b. The fraction of the radiation emitted from the wall which strikes a particle depends on the concentration and projected area of the particles, and on the shape of the enclosure. Based on considerations similar to those used by Nusselt ($1) and Jakob (9) for the amount of radiation reaching any point in a gas-filled space, curves were derived for a cloud of swpended particles in cylinders of varying diameter and between infinite plates of varying pitch. It was assumed that the values for a square-shaped cross section would be slightly greater than those for cylinders. Those for rectangular cross sections were interpolated between the two curves according to the ratio of the sides of the rectangle. The final curves used in the calculations on the calciner are shown in Figure 5.
e IO 12 14 DIAMETER (OR PITCH) IN INCHES
6 X,
16
I8
0
FIGURE5. RECOMMENDED ABSORPTIVITYFACTORS FOR FURNACES O F VARIOUS SHAPES AND SIZES Abscissas should be multiplied by the emissivity of the particles, by the number of particles per cubic inch, and b y the projected area of the particles in square inches.
For the design of the flash calciner for zinc sulfite the application of the above equations represents a t best an approximation. In order to compare the estimated length of a cylindrical calciner similar t o that used a t the pilot plant, calculations were made for 200-micron particles fed a t the rate of 6.6 pounds per minute per square foot and a wall temperature of 725" C. (1335" F.). The total heat required for dehydration and decomposition and for sensible heat is 6.3 x B. t. u. per particle. The estimated performance of the calciner is shown in Figure 6, The particle is heated to the dehydration stage within a few inches. The dehydration requires approximately 2 feet. Another half foot is required to raise the particle to the decomposition temperature. The removal of the sulfur dioxide requires 5 feet. The total estimated length of the calciner is 8 feet. These conditions correspond closely to those in runs 26 and 27 in Table I, in which the decomposition of the sulfite was 66 per cent. According to the curves this percentage should have been reached in 6.25 feet. In the operation of the calciner the lowest thermocouple, located on the tube wall 6 inches above the end and a t a point in the refractory bottom of the furnace, showed several hundred degrees below the temperature of the other five. I n view of this uncertainty of temperature and considering the complexity of the design equations, the agreement is considered surprisingly close. It is interesting to note that the estimated total time of exposure of the particles in the cal-
AUGUST, 1940
INDUSTRIAL AND ENGINEERING CHEMISTRY
1043
in the table enables comparisons to be made. The results are fairly consistent and may be summarized as follows : 1. Oxidation of the solution in the tanks, clarifiers, and thickeners is negligible compared to the oxidation in the scrubber. Not only is the interfacial area available for oxygen absorption much less than \\-hen the scrubber is included, but also the rate of oxygen absorption per unit area by quiescent liquids is very slow. 2. The highest rate of oxidation was obtained when air was passed through the scrubber in contact with the solution. This was to be expected from the higher oxygen concentration, although the difference between the rate observed under these conditions and that when flue gases were being scrubbed was not great. In these tests the carbon dioxide content of the gases was approximately 7 per cent, corresponding to 14 per cent oxygen. The sulfur dioxide content of the stack gases was 0.08 t,o 0.15 per cent. 3. The oxidation apparently increases as the acidity increases. This may he ascribed to the catalytic action of the iron salts. The ferric ion is precipitated as hydroxide a t the pH corresponding to the normal sodium sulfite. The work in connection xith the Battersea project ( 6 ) has shown that suspended iron salts are catalysts for the oxidation, but iron in solution is 2.8 times as effective as iron in suspension. There is some indication that the oxidation is decreased when the solution is being regenerated by the addition of zinc oxide. This also would he in agreement with the theory of catalytic action by iron salts, since ferric ions Lyould tend to he removed from solution in the reaction tank due to the decrease in acidity.
- DISTANCE FROM
I
TOP,
FEET
FIGURE6. CALCGLATED PERFORMANCE OF CYLINDRICAL CALCINER AT PILOT PLANT Wall temperature 728' C. (1335' F.): rate of feed 3 . 2 grams/iminute) (square centimeter) or 6.6 pounds/ (minute) (square foot) : particle size, 200 microns; internal diameter of cylinder, 9.54 centimeters or 3.76 inches: actual heated length, 274 centimeters or 9 feet
ciner is only 2.1 seconds and the actual decomposition occurq within 1 second. Radiation accounts for approximately 70 per cent of the heat transfer. The estimated heights of calciners with different cross sections and wall temperatures are shown in Figure 7. These curves are based on a uniform feed rate of 11.7 pounds per minute per square foot, with particles just passing 100 mesh.
* W
F __ ---- WALL _TEMPERATURE- 12000 lOOOD F , I1
I/
I
1
I
Oxidation of Absorbing Solution Laboratory tests on the rate of oxidation of sulfite solutions are of little value in predicting the rate of formation of the sulfate. The rate is known t o be influenced by both catalysts and inhibitors that are present in flue gases (IO). The extent and nature of the absorbing surfaces must also play a n important part. Table I1 summarizes a series of tests on the rate of oxidation made at the pilot plant. These show the effect of agitation of the solution, the presence of a n inhibitor, circulation through the grid scrubber, oxygen content of the gases, concentration of the solution, and the acidity of the solution The flow of solution in series 1 was from a wooden sump tank, 33 inches in diameter, through a high-speed centrifugal pump to the scrubber packed with mood grids. The effluent solution flowed through a steel ash clarification cone, 38.5 inches in diameter, to the reaction tank similar t o the sump tank, and thence through a similar settling cone back to the sump tank. I n series 2 the ash clarification cone was omitted. I n series 3 , 4 , and 5 , only the tower and sump tank were included in the cycle. These modifications gave the effect of changing the surface area of the quiescent, or slowly moving solution. In the table the letters following the series number indicate the sequence of the runs. Thus the previous history of any solution is identified. The classification of conditions shown
I
~
8 12 I6 20 24 X , DISTANCE BETWEEN WALLS - I N C H E S
FIGURE 7. HEZTTRASSFER COEFFICIENTS A N D CALCTL d T E D HEIGHTS O F R E C T A N G U L A R FURNACES FOR c.4LCISATIOS OF ZIXC SULFITE
Rate of feed, 5.73 gramsi(minute) (square centimeter) or 11.7 pounds/(minute) (square foot) ; particle size. 200 microns
1. Oxidation is definitely less in the more concentrated solutions. This is consistent with the decreasing solubility of oxygen as the concentration increases. 5 . The results show that t h e presence of hydroquinone has no effect on the rate of oxidation. Explanation for this is not apparent unless the inhibitor is eit,her lost from the solution during continued circulation or its action is masked by the effect of the many impurities taken up from the stack gases. Test 3i was made immediately after the concentration of hydroquinone was doubled. Although the oxidation was less than that in test 3h. it is not significantly less than that when no hydroquinone was present .
INDUSTRIAL AND ENGINEERING CHEMISTRY
1044
TABLE 11.
SUMMARY O F
RESULTS OK
VOL. 32, NO. 8
O X I D A T I O X O F SCRUBBING SOLUTION
S o h . Concn., Moles per 100 Moles Water Run NO.
Duration of Test, Hr.
Vol. of Soln., Gal.
lb Id
12.0 41.8 17.5 15.5 44.2 15.8 94.0 16.0 38.2 16.0
430 430 430 430 430 125 125 125 125 125
:fh
le 3a 3c 3e 30 4c
Total Na
SOr SO&-2SOn-A . Solution Quiescenta
so2
+ HSOa-
PH
...
.. . ... ... ...
...
6.0 4.8 5.4
Hydroquinone. P. P. .M. 540 540 540 540 540 600 600 600 600 None
Solution Circulated through Towerb, No Air Flow 1.20 0 54 0.04 0 4s ... 540 0.28 0.50 ... 3.71 1.58 600 C. Solution Circulated through Towerb with Air Floningc 2.36 1.12 0 67 le 1.0 430 0.33 ... 540 4.16 0 63 6.5 0.51 2.12 3d 5.0 125 600 0 94 3.50 5.5 125 4.4 1.03 600 1.31 1.80 0 68 6.2 5.35 5.0 125 1.35 600 1.87 0 72 6.15 3i 5.0 125 1200 6.0 1.73 0 61 5.5 4a None 2.24 0.52 4.71 5.0 125 0 70 0.6 None 5.70 4b 2.0 125 3.13 0.90 0.87 6.50 4d 5.2 3.0 125 None 3.53 1.21 D. Solution Circulated throuah Towerb with Flue Gases Flowinnc 0.51 1.14 0.53 0.05 540 IC 2.48 0 78 .. . 540 1.21 0.49 10 0.57 2.41 1.21 0.16 600 2a 2.37 0.83 ... 1.52 0.28 2b 600 0.94 2.40 1.49 0.41 ... 600 2c 5,25 0.72 5,s h-one 2.68 0.64 4e 0 63 4.70 2.54 0.34 6.5 None 5a 4.2 125 0.82 5.4 None 3.0 126 7.05 4.60 0.74 5b E . Solution Circulated Flue Gases FlowingC and Regeneration with Pure ZnO 0.92 ... 540 430 0.69 li 7.2 0.92 ... 540 430 0.70 1.0 Ij 0.92 ... 540 430 lk 2.3 0.82 0.85 ... 540 430 Im 5.S 0.79 0.88 ... 5.0 600 0.80 350 2d 0.90 600 350 6.8 2e 1.12 0.85 600 1.35 350 2.9 21 a Surface area of solution in tanks in runs l b to l e , 36 s q . ft.; in all others, 14 s q . f t . b Average solution rate, 3.5 gal. per min.; contact area in tower, 375 s q . f t . 8 Average gas rate, 900 cu. f t . per min.
Oxidation Rate, Lb. Moles S O & - - Formed per Hr. 0.000 0.003
0.003 0.004 0.000 0.000 0.000
0.000 0.000 0.000
B.
la 3b
1.0 2.5
430 125
:i
... ...
... ...
6. As a general statement of the over-all oxidation under normal operating conditions in the grid-packed tower, the rate of formation of the sulfate may be taken as 0.05 mole per hour when the circulation rate is 240 gallons per hour and the size of the scrubber is sufficient t o absorb 99 per cent of the sulfur dioxide. Since the recommended ratio of gas to liquid is approximately 400 cubic feet of hot gas per gallon (14), this rate corresponds to the oxidation of 320 cubic feet of sulfur dioxide per million cubic feet of gas.
It is of interest to compare these results with those obtained in other washers. I n the development of the lime process for sulfur dioxide removal, Pearson, Nonhebel, and Ulander (22) found that the extent of oxidation of the sulfite depends on the catalysts in the ash and lime. With a grid-packed washer, similar t o that used a t the pilot plant, they estimate that 50 per cent oxidation is a fair assumption when dealing with gases containing 0.11 per cent sulfur dioxide by volume. This corresponds to the oxidation of 550 cubic feet of sulfur dioxide per million cubic feet of gas. However, the lime process requires a circulation of 1 gallon of slurry for every 10 to 12 cubic feet of gas. The rate of oxygen absorption is known to increase with liquor rate (24). I n the preliminary work in connection with the Battersea Station (6),in which absorption was accomplished by passing the gases through sprays and over wetted steel surfaces, the oxidation, when the effluent solution was not further aerated, was approximately 33 per cent of the total sulfur absorbed. I n this case the gases contained approximately 0.03 per cent sulfur dioxide so that about 100 cubic feet of sulfur dioxide was oxidized per million cubic feet of gas. This low rate of oxidation in spray washers agrees with the observation previously reported for a wet cyclone scrubber ( l a ) . All of these results indicate that the extent of the oxidation depends primarily on
0.004 0.004 0,090 0.100
0.072 0.086 0.012 0.016 0 044 0.001
0.076 0.042 0.01s 0.025
0.037 0.031 0 . on2
0.017 0.021 0.000 0.026 0.033
0.024 0 051 0.079
the type of scrubber used and is independent of the concentration of sulfur dioxide in the gas. Oxidation of the zinc sulfite likewise leads to accumulation of sulfate in the solution. I n the pilot plant, because of the discontinuous nature of the tests and the necessity of handling the solids manually, a larger percentage of oxidation was encountered than would be found in large-scale operation. In small batches as much as 10 per cent of the wet sulfite would be oxidized after standing for 90 hours exposed to air. Oxidation in the dryer itself was hardly detectable. I n the dry state the solid was stable for weeks. The results of the calciner tests shown in Table I indicated that the oxidation there would not be serious.
Desulfating the Solution The use of lime for desulfating the absorbing solution in sulfur dioxide recovery processes has been suggested several times ( 2 , 11). The particular flow sheet outlined above has several advantages which were apparent after a careful laboratory study of the equilibria involved. I n order to determine the efficiency of lime when added to a solution of sodium bisulfite and sodium sulfate containing various amounts of excess free sulfur dioxide, mixtures of these mere rotated in a thermostat for periods varying from 20 minutes to 68 hours. The solutions >\-ereanalyzed for sulfur dioxide and sulfate and calcium ions. The precipitates were examined under a microscope for evidence of the purity of the precipitated sulfate and particularly for the presence of calcium sulfite, the crystals of which could be distinguished easily. The observed efficiency of the lime was considerably lower than corresponds to the reported solubility of the sulfate in water. It is well known that calcium sulfate crystallizes slowly from water. At room
AUGUST, 1940
INDUSTRIAL AND ENGIKEERING CHEMISTRY
temperature the slow rate of crystallization was a serious problem, but above 40' C. (104' F.) the crystallization was rapid, After 20 minutes the efficiency of the lime reached a maximum and the calcium content of the solution could not be decreased by seeding with crystals of the sulfate. Examination of the crystals showed that in every case, when the ratio of sulfur dioxide to sodium exceeded 1.0 (i. e., the bisulfite point) in the final solution, no calcium sulfite was formed. The low efficiency of the lime may be explained by the increased solubility of calcium sulfate due to the presence of t h e sodium bisulfite in solution. A comparison of this effect with that of other salts without a common ion, such as sodium and ammonium chloride, and potassium, ammonium, and magnesium nitrates, is shown in Figure 8, in which the reciprocal of the geometric mean of the calcium- and sulfate-ion concentrations, in moles per liter, are plotted against the square root of the ionic strength. I n this figure low experimental values correspond to high concentrations of calcium sulfate-i. e., supersaturation due to incomplete precipitation. High experimental values correspond to low concentrations, or unsaturation, due to incomplete dissolution of crystals. The laboratory data on the crystallization from sodium bisulfitesulfurous acid solutions fall, for the most part, between the curves for ammonium nitrate and magnesium nitrate solutions, and indicate t h a t the solutions were approximately at equilibrium with the crystals. The data from the pilot plant, however, showed some supersaturation. The solubility increases from 0.0151 mole per liter in pure water t o nearly 0.10 mole in a 4 N solution of sodium bisulfite, for which the ionic strength, when saturated with calcium sulfate, is approximately 4.4. The effect of temperature on the solubility in this range is negligible.
1045
the theoretical quantity is used, irrespective of the extent of oxidation of the solution, but it is greatest when the solution contains the larger amount of sulfate. Removal of over 90 per cent of the sulfate can be accomplished only in concentrated sulfate solutions and when a large excess of lime is used.
Advantages of Proposed Flow Sheet for Desulfating As shown by the above discussion, the treatment of a partly oxidized sodium bisulfite solution with lime inevitably leaves a small concentration of calcium ions in solution. If the acidity of the solution is then reduced, such as by mixing with the main body of the absorbent, calcium sulfite, being quite insoluble, is precipitated. One of the principal advantages of the flow sheet as shown in Figure 1,in which the desulfated solution is returned to the liming tank rather than immediately mixed with the solution being treated with zinc oxide, is t h a t this afterprecipitation is prevented. Thus no calcium can be carried down along with the zinc sulfite. Further-
I
I
I
I
1
I
3
FIGURE 9. EFFECT OF RATIOOF LIJfE TO SULFATE I N DESULF.4TION PROCESS
FIQURE
8.
SOLVBILITY OF CALCIUM SULFATE I N AQUEOUS SALT SOLUTIONS
With the data of Figure 8 as a basis, calculations were made
to determine the effect of the ratio of lime added to sulfate present on the efficiency of the lime and on the percentage removal of the sulfate. Two curves for two solutions having the same active sodium concentration, 2.15 moles per 100 moles of water, but with different sulfate concentrations are shown in Figure 9. The efficiency of the lime apparently passes through a maximum when only 60 to 65 per cent of
more, no precipitation of calcium salts can take place in the scrubber. A second advantage in the flow sheet is the decrease in the amount of sulfur dioxide required for the desulfation as compared with that required in any processes previously described. If lime is added directly to a solution to be desulfated, followed by the addition of sulfur dioxide to bring the mixture to the bisulfite point, 2 moles of sulfur dioxide must be added for each mole of lime, in addition to 1 mole for each mole of sodium sulfite present. By adding precipitated calcium sulfite formed from the highly acid desulfated solution, only 1 mole is required for each mole of lime. Thus, the quantity of sulfur dioxide returned t o the solution before regeneration is accomplished is reduced almost 50 per cent. In Figure 9 curves are also plotted showing the moles of sulfur
AUGUST, 1940
INDUSTRIAL AND ENGINEERIKG CHEMISTRY TABLE111. REICLTEO F
PILOT PLAXT
TESTSON
SULFTJR
1047
DIOXIDE REMOVAL FROM
STACK
GASES
(Grid-packed t o w e r : height of parking, 10 feet, including grill and slat distributor: cross section, 1.56 square feet; interfacial area, 375 aquare feet) Date Time
11:4A~.M.
Entering gases Dry-bulb temp., ' F. Humidity, lb./lb. ,302, 7" by vol.. a-et basis Rate of flow, cu. f t . / m h a
165 0.0290 0.0644
Leaving gases Dry-bulb temp., F Humidity, lb./lb. SO2. soby vol.. wet basis
103 0.0460
684
0.0011
Entering solution Temp., O F. .ictive Xa, molesilO0 moles H20 SOz/active Na Liquor rate, gal./min.
1 174 0.80
Leaving solution Temp., F. Active Ka, moles/100 moles H i 0 Sodactive S a
1 116
Results
70SO2 removalh Total lb. moles absorbed Total lb. moles ZnO added Ratio SOz/ZnO Lb. Hz0 evapd. Lb. Hz0 lost by spray Total vol. gas scrubbed, 1000 c u i t . Total vol. liquor circulated, 1000 gal. Q
b
96
3.4
101
0.67
9s
2 0 0
0
8/9/37 1:46P.M. 3:46P.M. 162
162 0,0278 0,0770 924
0.0290 0,0744 910
102 0.0446 0.001s
102 0,0444 0.0072
9s 1,130 0.8s 3.4 100 1.299 0.94 97.6
0 172 0 230
0.76 134
0
...
0 0
99 0.38
6:46P.>f. 165 0.0270 0 0709
8/10/37 9 : 3 0 ~ . ~ .11:30A.M. 1:3OP.>f!.
2:30P.~.
860
165 0.0250 0 0599 920
170 0.0260 0.0740 920
169 0.0260 0.0795 920
169 0.0200 0.0795 920
102 0 0445 0,0048
99 0.0416 0.0081
102 0.0445 0.0101
102 0.0445 0.0089
102 0.0445 0.0069
1.363 0.69 3.6
1 173
0.70
0.98s 0.72
96 0.876
4.0
4.0
101 0.802 0.82 3.6
101 0.802 0.82 3.0
100 1.365 0 80
100 1.117 0 62
96 0.916 0.86
100 0.905 0.97
99 0.827 0.95
90.2
93 I 0 52i 0 664 0 77 410
86.3 0 615 0 816 0 75 463 64 372 1 52
81.8 0.803 1.060 0.76 619
86.5 0.960 1.392 0.70 781
482 1.96
596 2.44
SS
0.347 0.466 0.74 269 39 210 0 81
57 316 1 29
0.76 3.6
...
99 0.827 0.95 88.5
1.081
1.515
0.71 860
656 2.64
At 60° F 30 inches Hg. Caloulateh from percentages on the dry basis in inlet and outlet gases.
with the sulfur dioxide, although no measurements were made on dust loadings of the gases. Pearson, Sonhebel, and L-lander (82) report better than 90 per cent removal of the dust in a similar grid-packed scrubber operating on gases from a powdered fuel boiler. The gases from the scrubber were approximately saturated at the wet-bulb temperature, about 102" F. I n the summer months the discharge was practically invisible, but in cold weather there was a white plume from the short stack. S o precipitation was ever observed near the stack. Since a direct measurement of the spray in the discharge \vas difficult, an indirect method was used based on t h e amount of make-up water required to keep the level in the sump tank constant. The amount of evaporation in the scrubber was calculated from the gas flow and from wet- and dry-bulb temperature readings on the inlet and outlet gases taken at frequent intervals. The former was measured by a calibrated integrating flowmeter. Corrections were made for the moisture content of solid discharges and for pump drips. This method of estimating the spray loss can give only an approximation, but the accumulated values over a long run should be fairly reliable. The complete data on one of the tests made on the absoiptioii and regeneration part of the cycle are shown in Table 111. They are representative of the results obtained in several other tests except that in this case the sulfur dioxide content of the gases was slightly lower than usual. The removal efficiency varied from 81.8 to 98.2 per cent. The predicted efficiency (14) of the tower was 98.8. The lower percentage removal was obtained a t the time the ratio of sulfite to active sodium in the effluent solution was 0 . 9 i . At this point the solution had a slight equilibrium pressure of sulfur dioxide, and the scrubber could not operate efficiently. Over the 12hour test period, during which 650,000 cubic feet of gas were scrubbed, the absorption averaged 90.5 per cent and the residual sulfur dioxide was 64 p. p. m. The results of the test further show that the losses by spray were approximately 0.5 per cent of the total solution circulated, in spite of the fact that the gas velocity was over 12 feet per second. The ratio of sulfur dioxide removed to zinc oxide added n a s approximately 0 . i . Although this would indicate a low regenerating efficiency of the oxide, analysis of the sulfite crystals showed a 1 : l ratio. This indicates that the low ratio was due to making a material balance
over too short a time, considering the volume of solution in circulation. I n the operation of the pilot plant over a period of 2 years, there has been no indication of scale formation in the grid scrubber. Some zinc sulfite has been carried over as a result of poor separation of the solids in the settling cone. This eventually accumulated in the sump tank, the bottom of the scrubber, and the ash clarification cone. Although the system operates in the slightly acid range between p H 5.5 and 6.5, no precautions have been taken against corrosion. All of the liquor lines were ordinary black iron pipe, and the settling cones were fabricated of steel. Brass valves were used. The grid sections in the tower were nailed together. The absence of corrosion undoubtedly can be explained by the absence of dissolved oxygen in the solution, since this is a coiitrolling factor in this pH range. In apite of this experience, it is recommended that in actual operation all steel surfaces be protected from contact with the solution since the conditions are potentially corrosive; furthermore, steel surfaces are known to promote oxidation of the sulfite. On the other hand, the use of acidproof masonry does not seem to be necessary.
Cost Estimates Thorough consideration has been given t o the investment and operating costs of the process. They will be published elsewhere (16). On the basis of the experimental results an estimate was made of the type and size of the equipment required for a plant recovering 21.7 tons of sulfur dioxide per day from waste gases containing 0.3 per cent sulfur dioxide. The cost of the process compares favorably with those used in the English installations for gases from low-sulfur coals and is much less than the cost of the English processes (6, 2.2) when applied to high-sulfur coals. It is estimated that the installed cost of the equipment on a plant of this size, treating 100,000 cubic feet of gas per minute a t 300" F., would be approximately $140,000. This would include the machinery for liquefaction of the sulfur dioxide. The fixed charges and operating costs amount to approximately $15.40 per ton of
1048
INDUSTRIAL AND ENGINEERING CHEMISTRY
sulfur dioxide, or $1.08 per ton of coal burned. Uncertainty in these cost figures lies in the undetermined loss of chemicals which would exist in a large plant. The above estimates are based on an assumed loss of 1 per cent per cycle. This accounts for nearly 30 per cent of the total cost.
Conclusions The advantages inherent in the process have already been enumerated. Criticism will emphasize the large amount of equipment required. On this point, however, realization should be given not only t o the quantity of gas being treated but also to the amount of sulfur dioxide recovered. This is a sizable quantity not exceeded in many heavy chemical industries even nhen the processing is considered simple. Furthermore, the actual cost of the equipment exceeds by only 30 per cent that required by other processes now in use on dilute waste gases and producing only large quantities of waste material. From a technical standpoint the most serious difficulty lies in the oxidation of the circulated materials; this produces sulfates which must be removed from the solution. This disadvantage is not peculiar to this process alone but applies to every method of sulfur dioxide recovery. Comparison of the results obtained on different types of scrubbers indicates that the wet cyclone produces less oxidation than other scrubbers and is therefore more desirable from this standpoint. The improved method of desulfation described above depends on returning a part of the product to a portion of the circulated solution and thereafter removing the sulfate with lime. When dealing with 0.3 per cent gases, the total oxidation is estimated to be approximately 10 per cent of the sulfur dioxide absorbed. The removal of this amount requires the return of 20 per cent of the product; i. e., the regeneration part of the process must operate on 120 per cent of the sulfur dioxide removed. Although this is not serious when dealing with gases containing 0.3 per cent or more, it becomes increasingly serious for more dilute gases and thereby increases the unit cost of producing the sulfur dioxide. For 0.06 per cent gases, the oxidation would amount to 50 per cent of the sulfur dioxide. Probably somewhere between this concentration and 0.1 per cent the recovery process would no longer be feasible, and straight neutralization should be substituted. If neutralization is to be accomplished by treatment with lime, the use of the indirect method would have several advantages over direct absorption of the sulfur dioxide in lime slurry. I n this case only a clear sodium sulfite solution would be circulated through the scrubber. Regeneration and removal of sulfate would be accomplished outside by means of lime, and sufficient time for crystallization would be allowed so that scaling of the scrubber surfaces would be impossible. The possibility of commercial development of the recovery process depends on several economic factors. For the production of relatively small quantities of liquid sulfur dioxide, such that there would be no disturbance of market conditions, the process should show a profit. Furthermore, according to the best available estimates of the production costs, byproduct sulfur dioxide could actually undersell sulfur dioxide produced from crude sulfur a t the present price of brimstone. On the other hand, production of large tonnages of sulfur dioxide mill inevitably reduce the selling price to near the value of the sulfur equivalent-i. e., $8 t o $10 per ton. At this point, transportation and storage costs become serious, and reduction to elemental sulfur would have to be considered for a t least a part of the product. There is no indication that even pure dry gaseous sulfur dioxide can be produced from dilute waste gases a t these prices if the above cost estimate is accepted. Any reduction in the price of pure sulfur dioxide which will even approach these figures will obviously
VOL. 32, NO. 8
affect the economics of a large part of the chemical and related industries. The total annual production of liquid sulfur dioxide in the United States a t present is about 12,000 tons. Most of it is used for refrigeration, as a selective solvent, as an antichlor, as a bleaching agent, and, to a limited extent, as a chemical reagent. The present uses themselves probably could not absorb a much greater quantity even a t lower prices. The possibility of developing new uses for cheap sulfur dioxide exists, however, and this must be considered. Edwardes (3) and Hasche ( 5 ) showed that cheap sulfur dioxide can displace sulfur in the paper industry. Of special interest are the new sulfur dioxide-olefin resins (26) containing nearly 50 per cent sulfur dioxide produced by reaction of liquid sulfur dioxide with certain waste refinery gases. These resins are thermoplastic, moldable, acidproof, and stable up to 300" C., and have high dielectric strength. Because of the nature of their ingredients they are among the cheapest of the purely synthetic fabricating materials. Other possibilities include new uses of sulfur dioxide in certain metallurgical processes and as a raw material for heavy chemicals. The general conclusion to be reached, therefore, regarding the development of the zinc oxide process and other processes for recovering sulfur dioxide from waste gases, is that largescale development will come in this country only as new uses are found for the product so that the disposal problem is automatically solved. The broad program of the investigation of the sulfur problem obviously includes research along these lines.
Acknowledgment This paper contains part of the results obtained in the cooperative research project, Case 34, with the Utilities Research Commission of Chicago. It is published by permission of the Director of the University of Illinois Engineering Experiment Station. Parts of the investigation were used as thesis subjects for advanced degrees in the Graduate School of the University. These contributions and the assistance of several former students in chemical engineering in the laboratory and plant work are gratefully acknowledged.
Nomenclature Consistent units are used except where designated in the text. A = surface area; subscript p refers to the average particle area, w t o the area of the heating wall P - coefficient of thermal expansion c = thermal capacity of gas C n = drag coefficient d = arithmetic mean diameter of particles D = equivalent diameter of channel * F A E = radiation factor as defined of gravity acceleration '9 h = heat transfer coefficient; subscript c p refers t o that from the gas t o the particle, cw t o that from the wall to the gas, m to the over-all coefficient based on the area of the particle, and r t o the radiation coefficient based on the area of the particle k = thermal conductivity of gas K = volume factor, equal to volume/(diameter)3 n = total number of particles in suspension at any time P = viscosity of the gas n - = ratio of circumference to diameter P = density; subscript p refers to particle and Q t o gas At = mean temperature difference; subscript p refers t o that between the wall and the particle, w t o that between the wall and gas T = absolute temperature; subscript w refers t o the all, p t o the particle e = time velocity of particles; subscript t refers to terminal velocity.
v=
AUGUST, 1940
INDUSTRIAL AND ENGINEERING CHEMISTRY
Literature Cited
1049
H. F., a n d Singh, A. D., Univ. Illinois, Eng. Ezpt. Stu. Bull. 324 (1940). Johnstone, H . F., a n d Singh, A. D. (to Commonwealth Edison Co.), U.S. P a t e n t 2,161,056 (June 6, 1939). Johnstone,
C h a p i n , J. H., Univ. Illinois, P h . D . thesis, 1939. Clark, A. M. ( t o Imperial C h e m . I n d . ) , U. S. P a t e n t s 2,086,379 (July 6, 1937), 2,128,027 (Oct. 23, 1938). Edwardes, V. P . , P u l p Paper Mag. Can., Aug. 5, 1920. Foerste?, F., a n d Kubel, K . , Z . anorg. allgem. C h . ,139, 261 (1924). Hasche, R. L., Pacific P u l p &- Paper I n d . , 1930,33. Hewson, G . W., Pearce, S. L., Pollitt, A., a n d Rees, R. L., SOC. C h e m . I n d . , Proc. Chem. Eng. Group, 15, 67 (1933). Kild, K., Mitt.Kaiser-Wilhelm I n s t . Eisenforsch., 16, 59 (1932). Howard, H . , a n d S t a n t i a l , F. G . , U. S. P a t e n t 1,271,899 (July 9, 1918). .Jakob, AI., “ D e r Chemie Ingenieur”, VoI. 11, p. 301. Leipzig, Akademische Verlagsgesellschaft, 1933. ,Johnstone, H . F., Combustion, 5 , 19 (1933). .Johnstone, H . F. (to Commonwealth Edison Co.), U. S. P a t e n t 2,082,006 ( J u n e 1, 1937). Johnstone, H . F., a n d Kleinschrnidt, R. V., Tram. Am. Inst. Chem. Engrs., 34, 181 (1938). Johnstone, H . F., R e a d , H . J., a n d Blankmeyer, H. W., IND. ESG. CHEM.,30, 101 (1938). .Johnstone, H . F., and Singh, A. D., Ibid., 29, 286 (1937).
ENQ.CHEM., 31, 993 Johnstone, H . F., a n d Williams, G. C., IND. (1939). Lessing, R., J. SOC.Chem. I n d . , 57, 374 (193s). McAdams, W. H., “ H e a t Transmission”, p. 254, New York, McGraw-Hill Book Co., 1933. M a r t i n , Geoffrey, T r a n s . Inst. Chem. Engrs. (London), 4, 164 (1926). Nusselt, W., 2. V e r . deut. I n g . , 70, 273 (1926). Pearson, J. L., Nonhebel, C . , a n d Ulander, P. H. N., J. Inst. Fuel, 8, 119 (1935). S t . Jacques, E. C . , IND. ENG.C H E W ,News E d . , 15, 29 (1937). Sherviood, T. K . , “Absorption a n d Extraction”, p. 182, New k‘ork, McGraw-Hill Book Co., 1937. Singh, A. D. (to Commonwealth Edison Co.) U. 5. P a t e n t 2,141,228 (Dec. 27, 1938). ENG.C H E M . ,30, 176 (1938). Snow, R. D., and Frey, F. E., IND. Tooley, F. V., Univ. Illinois, P h . D . thesis, 1939. PRESENTED before the Division of Industrial and Engineering Chemistry a t the 99th Meeting of the American Chemical Society, Cincinnati, Ohio.
Utilization of Waste Lignin Current Chemical Research ELWIN E. HARRIS Forest Products Laboratory, Madison, Wis.
T
HE utilization of lignin in Chemical research on the fundamental A REVIEW of some of the aspects of lignin from waste products is characteristics of lignin gives us the wastes incurred in the a picture, somewhat incomplete logging, milling, and conaiding in the solution of this problem. A in places, but nevertheless helpversion of the forest crop has been a problem for many years. knowledge of the various external groupsful in the solution of the problem hydroxyl, methoxyl, and unsaturatedof what to do with lignin. Sulfite liquors have not only aids in foretelling the reactions of lignin. For isolating lignin, wood or been a waste but a nuisance in Hydrogenation of lignin suggests a way of other plant material is subthe national waterways. Pulping liquors alone are capable jected to a hydrolytic reaction converting lignin wastes into valuable prodaided by the use of any one of of supplying annually 1,500,000 the following agents: sulfuric tons of lienin. Forest and sawucts* acid (13); hydrochloric acid mill was& constitute an addi(25); other mineral acids ( 2 2 ) ; organic acids, either those tional lignin source of several million tons annually. present in the wood (15) or added (6) pulping chemicals; Such conditions have stimulated extensive research directed bases ( 1 8 ) ; fungi ( 1 ) or hydrolytic enzymes present in wood toward finding methods of utilizing lignin. A few sulfite pulp mills have in recent years worked out means for the disposal ( 2 ) . There is no example known where lignin has been isolated without the application of hydrolytic reactions. or recovery of this waste; the sulfate and soda pulp mills dis“Native lignin” (12), the alcohol-soluble lignin obtained pose of lignin waste by evaporating and burning. Current from wood without the use of added acid, has been thought by recovery or disposal methods a t pulp mills entail the installasome investigators to exist as such in wood. Table I contains tion of equipment a t considerable expense to precipitate the the results of experimental work to determine the amount of lignin or to evaporate the pulping liquor preparatory for connative lignin that can be obtained from wood treated to preversion into other products or for burning. vent the action of hydrolytic enzymes, as compared with the More than 100 years have passed since lignin was first recamount obtained from air-dried wood or wood attacked by ognized as a constituent of plant material, but its structure is blue stain, both of which promote hydrolysis. One sample still unknown. This is because lignin does not readily split was placed in 95 per cent alcohol and extracted immediately up into identifiable building units, as is the case n.ith its assoafter cutting, the other was allowed to air-dry and then was ciate cellulose. Because of the lack of adequate means of extracted. Blue stain developed in one sample that was being identifying lignin, a great deal of time was consumed in seekstored under normal conditions before extraction. Fats and ing a satisfactory method for the isolation of unchanged ligoils were removed from the dried extract by ether and carbon nin. Within the last decade lignin investigators have suctetrachloride. The residue was triturated with 70 per cent ceeded in finding the number and types of reactive groups and sulfuric acid a t 15” C. for 4 hours and otherwise treated as in certain cleavage products. This knowledge has made it possible to correlate many of the cqnflicting ideas about lignin the lignin determination. The dried residue was considered and to carry out a more definite program of research for its the lignin yield. It is recognized that compounds other than utilization. lignin may be present which would make these values higher
