CONTINUOUS FUSION PROCESS Donald
F. Othmer, Joseph J. Jacobs, Jr., and Arthur
C. Pabst
Polytechnic Institute, Brooklyn,
N.Y.
METHODS OF ANALYSIS The more convenient method of analysis was devised t o facilitate control during continuous operation, and to make possible accurate material balances and determination of yields. The solution formed by taking up the fusion mass in water contains sodium carbonate, sodium hydroxide, sodium oxalate, sodium acetate, and sodium formate, as well as suspended organic matter. To determine these materials the following steps are carried out:
ased upon the determination of optimum conditions, a fusion trough has been developed which allows the continuous fusion of caustic soda and sawdust. Direct fusion w i t h calcium hydroxide or w i t h mixtures of sodium hydroxide and calcium hydroxide i s not feasible. Continuous Fusion gives higher yields and shorter fusion times than batch operation, due t o the better control which i s possible and the relatively small amount OF material handled at any one time. These yields, o n pilotplant equipment, are as high as 79 per cent oxalic, 18.9 per cent acetic, and 3.86 per cent formic acid, all based on the dry weight of sawdust used. A method of precise control for plant operations has been developed based on the color changes of the mass during the reaction. A particular shade indicates optimum conditions; and as the color variation on either side i s wide, exact control is possible. A new system has been developed, tested, and standardized for complete analysis of the complicated systems of materials encountered. A method for the recovery OF the salts OF oxalic, acetic, and formic acids i s given, and a flow sheet of projected operation i s presented.
1. The humus is largely removed by filtering the solution through activated carbon, which is then washed well with hot water. The solution and washings are made up to a standard volume. 2. An aliquot portion is diluted with considerable water to reduce the subsequent precipitation of calcium hydroxide. Calcium chloride or calcium acetate solution is added until there is no more precipitation. If the precipitate is voluminous, it should be allowed to settle in the beaker, the liquid decanted, and the residue shaken with 200 cc. of warm distilled water. This precipitate is filtered on a weighed Gooch crucible and washed thoroughly with warm distilled water, until the filtrate is no longer alkaline to phenolphthalein. It is then dried, and contains only calcium carbonate and calcium oxalate if all the calcium hydroxide has been washed out. It is brought to dull red heat with a Meeker burner; the following reactions take place:
+CaO + COZ --+CaO + GO2 + CO
CaC08 CaCaO,
The blasted Gooch crucible may then be weighed to determine the amount of lime; and the total calcium ion is calculated as an equivalent of the sum of carbonate and oxalate. 3. Another aliquot portion is taken; and the same procedure is followed, except that it is not necessary t o be so careful about washing out the calcium hydroxide precipitate. The Gooch crucible is placed in a lar e excess of 2:l sulfuric acid (300 to 400 cc.) and heated to 90" This sample is titrated with potassium permanganate to determine oxalic acid. When this is calculated to calcium oxide and subtracted from total calcium oxide, the total carbonate equivalent may be obtained. An alternate method, which avoids contamination of the precipitate with the large amounts of calcium carbonate in this step, consists in acidifying the aliquot portion with hydrochloric acid and boiling to eliminate carbonates before the precipitation of the oxalate. Any excess hydrochloric acid can be neutralized with ammonium hydroxide. 4. An aliquot sample of the fusion mass solution is taken again and titrated t o the phenolphthalein end point with standard hydrochloric acid. This is equivalent to all of the sodium hydroxide plus half of the sodium carbonate. Half of the carbonate equivalent, as obtained in the last step, is subtracted, and the value for sodium hydroxide is obtained. 5. The method for determining acetic and formic acids depends upon their volatility. An aliquot sample is taken in a distilling flask, 2:l sulfuric acid is added through a separator funnel, and the acetic and formic acids ?re distilled over w i d some water. Additional water is added in small amounts until the distillate shows no acid. The distillate is made up to 500 pc. in a volumetric flask. An aliquot is taken and titrated with standard sodium hydroxide to determine the total acidity. Another aliquot is taken and treated with mercuric chloride, and the precipitate of mercurous chloride is weighed by the method mentioned in the first paper, page 262. This depends on the reducing property of formic acid and is shown by the following equation: 2HgClp f HCOOH +Hg2Cla 4-COa f 2HC1
8.
HE first paper indicated that the production of oxalic acid and other products from sawdust by fusion with caustic soda might be feasible. In determining the optimum conditions, it was found that satisfactory yields of
T
oxalic, acetic, and formic acids and methanol could be obtained. However, efficient recovery and recycle of the caustic soda is necessary. It was felt that the development of a continuous fusion process would result in a more efficient use of the caustic soda as well as better control, a decrease in labor and heat costs, and similar advantages usually resulting from continuous processing. An important additional need is an analytical method which will allow the determination of the several constituents individually and more readily and accurately than by conducting the sequence of operations as before. The analytical methods following gave more accurate results than the method of the fist paper; the earlier analysis was made by securing final products, whereas in this case it was made on aliquot portions. Furthermore, the higher yields reported here over those shown in the previous paper may be due to the losses in processing when the preceding method of evaluation was used and to the return of salts of acids in the mother liquor. Repeated cycling of these liquors would probably have shown yields comparable to those reported here. The previous results were consistent among themselves, and while low in absolute values, may be regarded as adequately supporting the general conclusions.
The value for formic acid,is thus obtained and, when subtracted from the titration value, gives the amount of acetic acid. 268
269
INDUSTRIAL AND ENGINEERING CHEMISTRY
March, 1942
A solution containing known amounts of the constituents was made up, and the method outlined was used to determine them; the results follow: Sodium oxalate Sodium hydroxide Sodium oarbonate Sodium formate Sodium aoetate
Amount Added
Found by Analysis
0.500 0.400 0,1044 0.100 0.632
0.504 0.393 0.100
O.OQ8 0.526
The method appeared satisfactory and was adopted in all subsequent determinations. PRELIMINARY EXPERIMENTS In the previous paper a method for carrying out the fusion operation and for recovering the pure salts was proposed. This involved taking up the fusion mass in water, filtering off the humus, evaporating the solution to 38' BB.,and cooling. The sodium salts of the organic acids were precipitated and the mother liquor could then be recycled. The salts were taken up in water, treated with milk of l i e , and evaporated until the calcium salts of the acids were precipitated and recovered. The regenerated sodium hydroxide could ,then be re-used. The calcium salts were treated with sulfuric acid, and the volatile acids were distilled. The residue liquid contained the free oxalic acid which could then be crystallized out. Several alternatives to this procedure suggested themselves as being worthy of study. The substitution of lime as the fusion medium and conse quent elimination of caustic would be desirable. Several batch runs were attempted using lime alone or mixtures of lime and caustic soda as a fusion medium. No fusion was obtained with calcium hydroxide. As the controlled fusion progresses readily, it was thought that the addition of slaked lime in a slurry form at the end of the fusion would precipitate calcium salts of acetic and oxalic acids in the strong caustic solution. This would simplify the process since it would eliminate the necessity for taking up the batch in water, evaporating, recovering the crude sodium salts, dissolving them, and then treating them with lime. The run was made, the mass cooled, and a measured volume of water added, together with a slight excess of slaked lime, over the theoretical amount needed. This slurry
stance (d. w. s.); oxalic acid in cake (as sodium oxalate) 8.2 per cent of d. w. s.; total, 61 per cent of d. w. s. The results showed that only about one seventh of the total oxalate formed has been precipitated as the calcium salt. This conversion is too small for practical operation. The reaction is as follows: Na2C20d Ca(OH)2 e 2!Na(IH
+
+ Cacaon
From a consideration of the equilibrium conditions, it is evident that the high concentration of caustic due t o the large excess present prevents the reaction from going very far in the righbto-left direction. Since the analytical methods used in reporting yields in the previous paper gave low absolute yields, it was decided to check the yield of oxalic acid under controlled fusion conditions to give a basis for comparison of the larger scale work to follow. Accordingly, 1.666 grams of sawdust (1.500 grams d. w. 8.) were weighed out into each of four small baking tins. Exactly 5.00 grams of caustic soda and 5.00 grams of water were weighed in, and the total weight was checked. These pans were then placed in a small gas-heated oven where the temperature was accurately maintained at 204' C. A t the end of an hour one pan was removed, the loss in weight recorded, and the whole mass dissolved in water. The oxalic acid present was determined, and this value was taken as the criterion of the progress of the reaction. A pan was removed each half hour, and the same procedure was carried out. The following data were recorded and plotted in Figure 1: srp .of Original Total
wt.
Oxalio Aoid Formed, % of D. W. S.
37.0 46.0 46.0 46.0
18.3 69.3 81.3 80.3
Loss,
Sample Pan 1
2
a
4
Time Hourb
1.0 1.5 2.0 2.5
Figure 1. Loss i n Weight and Oxalic A c i d Yields in ShallowPan Fusions
Here again the high yields from runs having a thin fusion mass are apparent; and because of the improved method of analysis which shows all of the oxalic acid produced, the yields are much higher than those previously reported. The time required for maximum yield in these experiments is believed, however, to be greater than is necessary in other equipment, owing to the slower heat transfer to, and consequent evaporation of water from, the mass. The drying curve indicates that under these conditions the loss of water in evaporation and of gases, methanol, and other volatile materials comes to about 46 per cent of the original weight of charge materials relatively early in the process, and then does not increase. In other words, any subsequent losses must be balanced fairly closely by increases due to absorption of oxygen from the air. Also of interest is the fact that this total loss in weight happened to be almost identical with the weight of the water used. Thus any water and oxygen from the air entering into the final product, either chemically or mechanically, must have been about equal to the weight of gases, methanol, and other volatile materials discharged.
was agitated for 30 minutes. A sample was taken, filtered, and analyzed for oxalic acid, since the presence of oxalic acid in the a t r a t e would indicate incomplete precipitation. The latter was shown by the following analyses: Oxalic acid in the filtrate (as sodium oxalate) 52.8 per cent of dry wood sub-
FUSION TROUGH The following proportions of raw materials had been found to give optimum yields: 3 parts caustic soda, 1 part sawdust, 3 parts water. The fusion period was 3 hours, although this depended upon the time necessary to bring the reactants up to the temperature at which the exothermic fusion took
90
80 70 60
6 50
-
3:
d 40 30
20
IO 0 0
I
2
Time (hours)
210
INDUSTRIAL A N D ENGINEERING CHEMISTRY
Vol. 34, No. 3
The operation of the fusion trough can best be illustrated by typical runs. CONTINUOUS RUNS
place. Carrying out the fusion in thin layers or by blowing air over the mass also resulted in high yields. All of these factors were kept in mind when the continuous fusion trough was designed. The fusion apparatus as finally developed consisted of a trough 20 feet long, 7 inches wide, and 8 inches deep, constructed of sheet steel. It was modified from a standard Link-Belt screw conveyor. Forward motion of the fusion mass was obtained by forty vanes of 1/4-inch steel bolted t o a 2-inch drive shaft on the axis of the trough. These vanes or pushers were set in the shaft at an angle which caused all together to form a continuous helix. The shaft was supported by two end-plate bearings and a center hanger bearing. It was driven by a 2-horsepower motor through a worm-gear speed reducer and chain sprocket drive, set to turn the shaft a t approximately 7 r. p. m. The fusion trough was heated by gas burners made by drilling many l/ra-inch holes in lengths of 1-inch pipe. Gas was supplied from a 2-inch manifold by Venturi air mixers. A line of these burners extended the length of the trough, about 0 inches below its bottom. Valves were placed a t appropriate intervals, two in the first 10 feet from the feed end and four in the last 10 feet. Since fusion was expected to occur somewhere in the last 10 feet of the trough, more accurate control was desired in this range and the larger number of separately controllable sections of burner were supplied. An apron of light sheet steel hung on either side of the entire length of the trough, with an opening between it and the outer edge of the trough’s flange for combustion gases to escape. It served to enclose and direct the heating from the burner. This apparatus is shown in Figure 2, and the discharge end is reproduced in Figure 3. The unit was installed so as to be several inches lower a t the discharge end (the drive end) than a t the feed end. Discharge of material was through a 3-inch hole in the bottom of the trough, several inches from the drive end. This hole could be partly or completely closed by an external gate which was adjustable through a door in the apron, as shown in the figures. No provision was made in this apparatus to collect volatile materials given off during fusion, since the combustible gases were insignificant from the standpoint of heating value; and while the methanol would certainly be worth recovering in full-scale operations, the additional complications in this experimental unit would not be warranted.
Many preliminary runs showed the operation of the fusion trough t o be satisfactory when using less water than the amount required to make a 50 per cent solution of the caustic, which had been found necessary in previous batch runs. This was doubtless due to better agitation and more uniform heating than were previously obtained. Subsequent operation also showed that the use of less than a 3 to 1 weight ratio of caustic and sawdust was feasible. Quantities of sawdust and flake caustic were weighed out and added a t the feed end of the trough. The required amount of water was then poured over this mass. Thus, enough of each of the three ingredients for a half-hour operation would be weighed into containers near the feed end. The operator had scoops of such size that one scoopful of sawdust would be required for each measure of caustic soda; and the required amount of water was added by its appropriate measure. The three materials were added continuously in this way, in almost the correct ratios and in exactly the weighed ratios over the timed interval. The caustic and sawdust absorbed the water rapidly, and the mixture was quite uniform. At approximately 5 feet from the feed end of the trough the mass was a smooth, pasty conglomerate. Fusion usually took place approximately 15 feet from the feed end, and the fused mass was discharged in a semiplastic state. All of the mass produced from the known weight of charged materials was collected and weighed; and these required the clearing of the trough of all material after each run. A weighed sample of the mass was analyzed for oxalic, acetic, and formic acids. From these data and the material balance the yields were calculated. For this reason, however, the yields reported are averages for a run and include the poorly fused material left on the trough a t the end of the run; they are not representative of maximum yields which could be attained and doubtless were attained, at various stages of the fusion. I n a continuous run of any length, the trough could no doubt be controlled to give the much higher yields, approaching the values indicated by spot samples. After trial runs on the methods of operation, many runs were made to produce salts continuously; a few examples
TABLE I. RESULTSOB CONTINUOUS RUN A Raw materials, pounds Sawdust Caustic soda Water Yields Total fused mixture ib. Oxalic acid, % d. w.’ s. Volatile acids, % d. w. 8 .
95 (85.5 d. w. 285 65
8.)
350
41.6
22.2
TABLE11. RESULTS OB CONTINUOUS RUN B Raw materials, pounds Sawdust Caustic soda Water Yields Total fused mixture lb Oxalic acid 7 of d ’w ‘s Acetic acid’ $ of d: w: a: Formic acid, % of d. w. s.
40 (36 d. 91 45 125 65.7 18.9 3.86
W.
s.)
March, 1942
INDUSTRIAL AND ENGINEERING CHEMISTRY
are cited. In. one of these rum the quantities of raw materials shown in Table I were used and were fed directly, individually, simultaneously, and as uniformly as possible over a 1.5-hour period; the time required for the completion of the run was 2.5 hours. Another run, using a lower rate of feed, gave the results shown in Table 11. COLOR OF REACTION Some definite color changes took place as the wet mixture changed to the final plastic mass discharged from the fusion trough. Since these color changes are indicative of the progress of the reaction, and provide a convenient and simple control method for the operator, samples were taken at various places in the trough during a run and analyzed for oxalic acid. Samples were also taken of a run which was known to be badly overheated, to show the full range of the reaction and the accompanying colors (Table 111). TABLE 111. COLORCHANGES DURINQ REACTION Feet Oxalic from Acid Feed T:m$., Color of Reaction Mass (% of D. W. 9.) End 145 Very dark red, particles of sawdust still visible 13.4 3 160 particles swollen and yellow brown 20 0 5 180 Light ellow brown, no individual particles of 8 saw&st visible 27.2 185 Stiff mass, light yellow, almost dry in appearance 37.2 10 200 Pale yellow tinqed with green, exothermio reac13 tion mass fluid 48.0 210 Pale d i v e qreen, optimum color for discharge, 15 exothermic reaotion complete, quiet fluid mass 79.2 225 Dark walnut brown. wet mass, gone too far 29.2 18 240 Almost black; dry mass, worthless residue 9.7 20
I n this particular run too much heat was supplied; and charring took place a t the discharge end. Data are given to indicate color of mass and percentage of oxalic acid in the overheated material. When operated with slightly lower flames
Figure
3. Discharge End of
Fusion Trough
271
of the u n 8 e r E i g gas heaters at the same feed rate, the darker color disappeared from the discharge end, the optimum color moved down the trough away from the feed end, and the discharged material contained more oxalate. This indicated that 75 per cent yields of oxalic acid are possible with accurate control of operating conditions (as indicated by the definite color changes) so that the product is always discharged a t the correct stage in the fusion. It was not possible in the experimental operations to operate long enough to permit optimum conditions to be maintained continuously for test runs; and since all material fed into the trough was taken as a basis for calculations, it follows t h a t the first and last could not have been under optimum conditions. Therefore the over-all average yields were reduced. RESULTS
OF CONTINUOUS FUSION
Table I1 shows that satisfactory yields may be obtained using approximately 2.5 parts of caustic soda per part of sawdust. This is less than the 3 parts regarded as optimum heretofore. It was also found that 65 per cent solution of caustic soda was preferable for efficient operation of the trough, which means that a lower amount of heat would be required for the evaporation of the excess water. These improvements over the optimum conditions originally determined were due to use of the continuous fusion trough, which permitted closer control and more intimate contacting of the materials, The increase in yields above indicated for the continuous operation as compared to the batch operation were more apparent, possibly, then real; for the better analytic methods showed all of the yield as compared to the poorer methods previously used. Nevertheless, the better yields were probably due in some degree to the excellent aeration of the mass, caused by the lifting and mixing motion of the vanes on the screw conveyor in the fusion trough. The gas consumption for heating was measured and found to be from 10 to 20 cubic feet of gas (550 B. t. u. per cubic foot) for every pound of sawdust charged. This would obviously be much less on a large-scale unit, particularly if designed to minimize heat losses. Since the continuous fusion process is in a n open trough with immediate access of air, all volatile products (of which the most important is methanol) are lost. It would be possible to have a closed trough with air circulated through and a condenser and water scrubber for the gases discharged, to remove the methanol. This would complicate the feeding, discharge, and control of the process and might present some hazard due to the possibility of the presence of explosive mixtures of the air with carbon monoxide, methanol, and other idammables. This loss of production of methanol and other volatile materials would be balanced somewhat by the high yields of acids, excellent control, and other advantages of continuous operation. The capacity of the apparatus is probably determined largely by the rate at which heat may be supplied and mechanical motion secured under conditions of proper agitation, aeration, and control. It is believed that the upper capacity limit was not even approached in the experiments on the unit, and would be comparatively greater on a larger plant-operated unit. These experiments, however, showed a feed capacity of from 20 to 30 pounds of sawdust per hour on this standard trough, or 0.5 to 1 pound per square inch effective cross section of trough.
212
INDUSTRIAL AND ENGINEERING CHEMISTRY Sodium hydroxide 285.0 74.5 Water Sawdust (d. w. s.) 85.5 95 lb. water and volatile material
FUSION
Sodium oxalate Sodium acetate and formate Sodium carbonate Sodium hydroxide Humus (difference)
Analysis of Product 37.8 26.1
NaOH Equivalent 22.5 12.7
76.2 193.0
57.5 193.0
1B.g 350.0
285.7
Figures are in pounds
Figure
+
In projecting to larger or industrial-scale operation, the questcon o f eqGipmentcosts and labor charge; would be of utmost importance. It is apparent that the continuous system would be much more economical from both standpoints. MATERIAL BALANCE OVER FUSION TROUGH
T o determine the disposition of materials in this process, a complete material balance was made on one run. The run of Table I was used, because large amounts of material were available which allowed the use of fairly large-size laboratory equipment, and because the low yields probably represent the worst conditions which would be encountered in plant operation. The complete data are given in Table IV and in Figure 4. TABLE IV.
MATERIAL BALANCE DATA 95 (85.5 d. w. 285 65 370.5 210 2.5 350
8.)
ANALYSISOR FUBION MIXTURB Lb. of Acids Sodium Salts Based on D. W. S. 37.8 41.6 26.1 22.2 76.2 193.0 16.9 14: s
v0
..
360.0
RECOVERY
oxalate acetate and formate hydroxide carbonate Total
4. Material Balance of Fusion Operation
Sawdust, Ib. Caustio lb. Water, ib. Total input of dry solid3 lb. Temperature of fusion C. Time of complete run 'hr. Total material reoove;ed, Ib.
and was then centrifuged. It was found that a centrifuge is not the proper type of equipment for this filtration, as 2.5 hours were required to centrifuge this small batch. At the end of this time the cake still contained considerable water, and thus a higher percentage of caustic soda remained in the cake than would ordinarily be the case if a more suitable separating machine were used. The cake recovered weighed 21 pounds and the filtrate 59.5 pounds. Samples of each were analyzed. Since the starting material had been analyzed, a material balance of all the products involved could be made around the centrifuge operation :
Water Sodium Sodium Sodium Sodium Humus
-
OF SODIUM SALTS
I n the method originally projected for recovering the pure salts, the leached solution was evaporated to 38" BB. and cooled, and the crop of crystals was recovered by centrifuging. Since this served as an analytical method as well as a proposed plant operation, it was decided to carry out a material balance on this operation with the larger quantities of material available and with the improved analytical methods. Twenty-five pounds of the fused material from the above run were treated as outlined previously to obtain a slurry having a gravity of 38" BB. The slurry was allowed to cool
Vol. 34, No. 3
Lb. Charged 35.0 2.7 1.86 13.8 5.4 1.2 __ 60
Lb. in Filtrate 24.20 0.22 0.72 13.50 0.75
Lb. in Cake 11.04 2.60 1.16
39.39
21.00
... -
0.40
4.60 1.20
PRECIPITATION OF OXALATE ACETATE, AND FORMATES A L ~ S
Since the evaporation of the slurry t o 38" BB. and the recovery and reslurrying of the salts before liming represented an extra step in the process, it was decided to attempt the precipitation of a t least the calcium oxalate by direct addition of lime to the original solution of the fusion mass. Fifty pounds of fusion mixture were dissolved in 85 pounds of water, and 3.5 pounds of slaked lime were added. The slurry was cooked for 30 minutes a t 110" C. and filtered hot. The specific gravity of the filtrate was 58" BB. The filter cake was placed in the kettle, cooked with 16 pounds of water, and filtered again. The gravity of the second filtrate was 26" BB. After two washes with hot water, the gravity of the filtrate was down to 5" BB. The weight of the precipitate was 13 pounds. According to analysis, the 50 pounds of starting material should have contained 7.5 pounds of oxalic acid. This would have formed 8.8 pounds of calcium oxalate. The precipitate contained 40.0 per cent calcium oxalate or 5.5 pounds. The filtrate and washings contained 1.49 per cent oxalic acid calculated as calcium oxalate, or 2.5 pounds. This oxalic in the filtrate is doubtless sodium oxalate which, as mentioned previously, will not react with the lime in the presence of the large excess of caustic soda. Of the 37.5 pounds of caustic soda in the starting material, 34 pounds were found in the filtrate plus the washings, and the remaining 3.5 pounds must have remained in the cake. Analysis of the cake revealed that it contained 22.2 per cent sodium hydroxide or 2.78 pounds. This represents a loss of * 2.78 pounds of caustic for every 5.5 pounds of oxalate formed. A more efficient filter and better washing would cut this loss considerably. Analysis of the cake showed that it also contained calcium carbonate due to a reaction between the sodium carbonate formed during fusion and the added lime. This calcium carbonate would remain with the calcium oxalate and require considerable sulfuric acid for its decomposition. A5 much sodium carbonate as possible should be eliminated before precipitation of calcium oxalate. From the above results the following method appeared to be a logical and simple method of separating the ingredients without the necessity of a distillation step. This is shown on the flow sheet of Figure 5. The fusion mixture is dissolved, heated and, if necessary, evaporated so that the solution boils a t 126' C. (38" Be). It is then cooled and filtered. The filtrate contains the reclaimed caustic soda and some sodium acetate and sodium formate. The filter cake contains sodium oxalate, sodium
March, 1942
INDUSTRIAL AND ENGINEERING CHEMISTRY
carbonate, sodium acetate, and sodium formate, and whatever sodium hydroxide is in the retained mother liquor. The c:hke is dissolved in sufficient boiling water to dissolve all the ingredients. Slaked lime is added slowly until all of the oxalate has been precipitated as the calcium salt. (Small samples are withdrawn, filtered, and tested with calcium chloride in a faintly ammoniacal solution until free of turbidity.) The product is kept at boiling temperature for 30 minutes, filtered hot, and washed well with hot water until free of caustic. The filtrate is evaporated to throw down sodium acetate and formate as crystals, which are separated. The liquid portion contains the balance of the caustic which may be added to the primary filtrate. The calcium oxalate is treated with sulfuric acid and recovered in the regular manner. The sodium acetate and formate may be purified and marketed or may be made into the mixed acids and purified. WOOD FATTY ACIDS FORMED DURING FUSION Certain volatile wood oils are formed during the fusion, but owing to the construction of the trough, no volatile materials were recovered during the continuous runs. During the fusion, however, sodium soap was formed as a result of the action of caustic on the natural fatty acids existing in the wood. This soap was found in both the filtrate and the cake during the first filtration. The major part was in the cake and continued through the process until it was introduced into the lead tank containing the sulfuric acid. Here it was liberated as a resin-type fatty acid. It floated to the top of the liquor where it was skimmed off and could be recovered as a by-product. The soaps formed approximated 2 to 3 per cent of the weight of the dry sawdust. A pure sample of this acid was obtained in the following manner: 1000 grams of the fusion product were dissolved in water, and hydrochloric acid was added until it was acid to methyl orange. (This destroyed the soap and liberated the free fatty acids which floated to the top of the solution.) The liberated fatty acid was extracted with ether, and the
w
Projected Flow Sheet for Recovery of Salts from Continuous Fusion
Figure 5.
273
ether extraction washed with distilled water until the washings were neutral to methyl orange. (This washed out any inorganic acids present bu2, left insoluble organic material in the ether layer.) The ether layer was evaporated, and the fat dried overnight on a steam bath. This experiment gave a yield of 12 grams of a dark, sticky, resinous acid with the following constants: Acid value Ester value Saponification No.
194.1 44.5 238.6
Because of its high saponification value this acid might form soaps that would be of value in the manufacture of grease, textile oils, and specialty products. PLANT DESIGN From the data collected in the laboratory on the continuous fusion apparatus, it can be assumed that operation on a large scale would result in yields of 50 to 75 per cent oxalic acid, based on the dry weight of the sawdust used. The lowest figure may be taken as conservative; and i t is thought that as experience is acquired in plant operation, the process can readily be run to yield 65 per cent. A flow sheet of operation projected upon the above method of operation is shown in Figure 5. SUMMARY The gas-heated, continuous fusion apparatus was built around standard screw conveyor parts, although the screw was not used. The trough was 20 feet long, 7 inches wide, and 8 inches deep. Forward motion of the fusion mass was obtained by the use of forty 0.25-inch vanes bolted a t an angle to form a continuous helix, to a 2-inch drive shaft turning a t approximately 7 r. p. m. The fusion mass was a mixture of sodium hydroxide, sodium carbonate, sodium oxalate, sodium formate, sodium acetate, and humus. I n order to evaluate products, control production, and make material balances, a method of analysis was developed for this complicated mixture. Under continuous operation, the amount of caustic used could be decreased from that found necessary in the previous investigation, and much higher yields of oxalic and acetic acid were obtained. The paddles used tended to aerate the mass more effectively than was previously possible and may have accounted for the higher yields. The presence of wood fatty acids in the fusion mass was shown. Definite color changes were noted in the mass as fusion progressed. These colors were a direct indication of the yield of oxalic acid. An excellent arbitrary control method for unskilled labor is here indicated. A complete material balance was made around the fusion trough to show the disposition of materials. A method of separating the salts was to leach the fusion mass with water, evaporate to 38' BB., and precipitate out most of the sodium oxalate, sodium acetate, and sodium formate. After centrifuging, t,he liquor (containing all of the free caustic and some sodium acetate) could be evaporated and recycled. The flter cake was treated with milk of lime, and the calcium salts of the acids were precipitated for a further separation. The regenerated caustic was then evaporated and recycled to the fusion operation. The continuous fusion of caustic soda and sawdust has been shown to be feasible, and it appears to be capable of industrial exploitation.
