Oxalic Acid from Sawdust OPTIMUM CONDITIONS FOR MANUFACTURE Donald F. Othmer, and Joseph J.
Carl H. Gamer, Jacobs, Jr.
Polytechnic Institute, Brooklyn,
Th
N. Y.
Oxalic acid is now made commercially by the action of nitric acid on cellulose, or other carbohydrate material, by passing hot carbon monoxide over a mixture of sodium hydroxide and hot coke to form sodium formate, which is later converted into the oxalate, and as a by-product in the fermentation of citric acid (both acids occurring as the calcium salts). The object of this investigation was t o determine whether the commercial production of oxalic acid and acetic acid (a valuable by-product) from waste sawdust by fusion with alkali would again be profitable by modern chemical methods, and t o determine the optimum procedure if such methods are used. Although the literature gives yields of oxalic acid per pound of sawdust when treated with alkali, this is not the vital factor. The value.of dry sawdust, on the basis of its heat content, ranges between 50 and 60 cents per ton. Caustic soda for the fusion, lime for recausticization, and sulfuric acid for conversion are the raw materials needed; all three are cheap chemicals but are expensive compared to sawdust. Lime and sulfuric acid are required in almost stoichiometric amounts; any excesses are almost wholly due to caustic losses and may be proportional to such losses. The feasibility of the process depends more, therefore, upon the efficiency of caustic recovery than upon any other single factor. Since none is represented in the products formed, the make-up requirements are due only to losses, which should be minimized. A further requirement for commercial success is an economical process for separating the materials resulting from the fusion. Methanol and acetic acid as by-products of this reaction have been mentioned in the literature without extensive data as to amounts, effect on production, or cost of production. KO mention is made of the evolution of combustible gases during the reaction.
e fusion of sawdust with caustic soda to produce salts of oxalic and acetic acids was the commercidl source of oxalic acid for many years. Because of current prices and demand for these acids, the Fusion operation has been studied from the standpoint of more efficient recovery of by-products, as well ds an increase of yield by accurate control of the process variables. The optimum values of the variables investigated were: ratio of NaOH to sawdust, 3 : l ; concentration of NaOH used, 50 per cent; time and temperature of fusion, 3 hours and 200-220° C.; depth of fusion mass, as shallow as possible) type of wood, dependent on cellulose content. Carrying out the fusion in thin layers or blowing a i r over the mass increased the yields of oxalic acid. Runs made on seven types of w o o d showed that oxalic acid yield w a s dependent upon the cellulose content, although the yields were always higher by a constant percentage than the theoretical yield based on the results of fusion with pure cellulose. Since the sodium hydroxide i s the most expensive r a w material used, its efficient recovery i s a requisite of the successful operation of the process. Runs showed the necessity for a considerable make-up of sodium hydroxide. However, a careful material balance indicated that it should be possible to reduce this to 3 per cent make-up. Methanol and formic acid are produced in quantities large enough to warrant their separation and recovery. Enough data were collected to provide a basis For the study OF continuous manufacture, which, w i t h its attendant advantages, seems the desirable way to operate the process.
STIMATES indicate that more than 8,000,000 tons of E sawdust are either burned as fuel or wasted annually in this country. The utilization of sawdust and wood waste in
METHODS
OF OPERATION AND ANALYSIS
Thorne (6) found that 2 parts of caustic alkali (80 per cent potash, 20 per cent soda) heated with 1 part of sawdust t o 200" C. yielded a quantity of oxalic acid equal t o 36 per cent of the dry wood substance (d. w. s.) in the sawdust. By increasing the caustic:sawdust ratio to 4:1,he obtained a yield of 42 per cent oxalic acid; and when he reacted the latter quantities in a thin layer, the yield was 52.15 per cent. He leached the fusion cake with water t o produce a solution of sodium oxalate which was evaporated until crystals formed. They were separated from the mother liquor and redissolved, and the solution was boiled with milk of lime. The calcium oxalate formed was separated and treated with sulfuric acid to produce oxalic acid, which was subsequently crystallized and purified by recrystallization. The fusion step was operated according to the methods later described. Early work showed that the caustic potash gave higher yields; but the greater cost did not warrant its
the production of oxalic acid, of which there is a shortage, by the fusion of sawdust with alkali is therefore a study of potential commercial importance. The production of oxalic acid from sawdust is not a new process. It was manufactured in 1829 by Gay-Lussac and applied by Dale in 1856 (4). The technique remained unchanged until comparatively recently, when all plants utilizing this method had stopped production. Oxalic acid is an important organic acid; in the United States some 10 million pounds are used annually (8). Laundries consume a large amount as an acid rinse. It is used in the production of celluloid and rayon, manufacture of explosives, purification of glycerol and stearin, leather manufacture and dressing, tanning, calico printing, bleaching straw and wax, preparation of certain dyes, extraction of rare earths from monazite, manufacture of indigo blue for laundry work, etc. 262
INDUSTRIAL A N D ENGINEERING CHEMISTRY
March, 1942
use. Consequently, all reported work was with caustic soda alone. The fusion mass was put through the steps shown in the flow sheet (Figure 1) to determine yields. This scheme of operations was used as an analytical method in the work; but it is a process by which large-scale steps could be operated and is similar to that of Thorne. The main constituents in the solution which was made by taking up the fusion mass in boiling water were sodium hydroxide, sodium carbonate, sodium acetate, and sodium oxalate. Insoluble "humus" impurities were filtered off as a slimy mass and discarded. On evaporation of the filtrate, the scum of crystals forming on the surface was continually stirred into the solution. Evaporation was stopped at 38' €36. and the liquid was cooled to atmospheric temperature, whereupon a mixture of sodium acetate and sodium oxalate crystals was thrown down. Considerable difficulty was experienced in filtering these crystals from the strongly alkaline mother liquor; finally a small, high-speed centrifuge was used. Analysis of the filtrate indicated the amount of alkali which would be re-used in subsequent batches in a plant operation. This alkali value, added to that obtained after the subsequent liming step, is reported as alkali recovered; and the difference between this sum and the amount originally used represents the amount needed to obtain the given products. The dark brown, sticky, crystalline mass resulting from this fltration was dissolved in water: and the solution was
I
LIMING TANK
SULFURIC
RECOVERED CAUSTIC
CALCIUM SULFATE WASTE
t CRUDE ACETIC
AND FORMIC ACID8
t CRUDE OXALIC ACID
Figure
I . Flow Sheet of
Operations
263
strongly agitated as it was treated with lime. In addition to caustic soda, there resulted calcium oxalate which is extremely insoluble and calcium acetate which is moderately soluble, as well as some calcium carbonate from the sodium carbonate present. When these reactions are carried out at the boiling temperature with a minimum of water, the mixed calcium salts precipitate. To prevent an injurious excess of lime, which is also insolubIe and will contaminate the precipitate, the following qualitative test for the sodium salts was employed: The lime was added in small quantities to the boiling solution of the salts. At intervals, samples were withdrawn and tered. To the clear samples was added acetic acid in excess over that required for neutralization. Calcium chloride solution was then added. As long as turbidity resulted, sodium oxalate was still present, and more milk of lime was added to the batch. After the lime treatment the material was filtered on a, Biichner funnel. The alkali content of the filtrate was determined by titration of a sample in order to calculate the recovery of alkali as mentioned above, and the bulk of the liquid was used in a subsequent batch. The sludge from the filter and some water were added to a distilling flask supplied with suitable agitation and condenser. Sulfuric acid was added from a dropping funnel t o release the organic acids. A large excess of sulfuric acid was used (to ensure complete liberation of the oxalic acid) because of its comparative strength. (The dissociation constant of the first hydrogen of oxalic acid is equal to 3.8 X while that of the second hydrogen of sulfuric acid is 2 x 10-2). The material in the flask was boiled, and small additional amounts of water were added to drive off all of the acetic acid. The acetic acid and water vapors were condensed to give a crude dilute solution of acetic acid, which was measured and titrated for acid content. Oxalic acid, being soluble, remained in the solution, and calcium sulfate precipitated as a finely divided powder. During this reaction any sodium hydroxide retained from the former steps was converted into the corresponding sodium sulfate which, being moderately soluble, remained in solution with the oxalic acid; calcium carbonate, resulting from sodium carbonate passing to the liming step, gave more calcium sulfate. The calcium sulfate was filtered from the solution and washed free of oxalic acid. The filtrate was titrated with potassium permanganate to determine the resulting oxalic acid product. I n many cases pure oxalic acid was prepared by evaporating the filtrate to 15" BB., precipitating and filtering off the calcium sulfate, evaporating this filtrate to 30" BB., cooling to separate oxalic acid crystals, and recrystallizing to remove all trace of sulfuric acid. Figure 1 illustrates this method, which was a t once an analytical and a process method. Its chief disadvantages as a production method are (a) sodium carbonate formed in the fusion (as found while this work was in progress) is recausticized in the liming step and the calcium carbonate passes to the acid step and thus requires sulfuric acid; and (6) the salts of the volatile acids may not be recovered as such. It has a major disadvantage as an analytical procedure also, in that a certain amount of the salts may recycle and thus not show up in the product. This would be of little disadvantage from the production standpoint or for identical runs after several cycles, but would not be desirable for absolute quantitative results. Nevertheless, the relative but not always absolute effects of different variables may be studied from the results obtained by this straightforward method.
a-
264
I N D U S T R I A L A N D ENGINEERING CHEMISTRY
F U S I O N M E T H O D S AND EQUIPMENT The caustic fusion process was carried out to determine the optimum conditions in three different types of apparatus:
DEEPR m METHOD, A gas-fired, closed fusion kettle with a capacity of 1.2 cubic feet was used. It was provided with a stirrer, tipping device for emptying, thermometer well, and outlet for a condenser. Besides the one pound of chestnut sawdust (on the dry basis) used in most runs, the same amount of purified cotton linters was employed in other runs to determine the effect of substantially pure cellulose. The relatively large size of the kettle in comparison with the size of the batch allowed the subsequent operations of leaching, liming, etc., to be conducted in the same unit.
Figure
2. Effect of Ratio
of Caustic to Sawdust on
Yields of O x a l i c and Acetic Acids and on Caustic Recovery
Figure 3. Effect of Caustic Concentration on Yields of O x a l i c and Acetic A c i d s a n d on Caustic Recovery
MULTIPLE RUNMETHOD. Four batches of 0.25 pound each of chestnut sawdust (on dry basis) were fused simultaneously in a battery of four cast-iron glue pots heated in a sand bath. Agitation was supplied by individual stirrers driven by a.single motor. SHALLOW RUNMETHOD. Several runs were carried out t o determine whether the depth of the fusion mass.affected yields. An iron baking pan 36 inches long, 24 inches wide, and 1.25 inches deep was used. Heating was somewhat uneven throughout the mass although three large flat gas burners were supplied, and agitation was supplied manually with an iron rake.
In each case dry sawdust was first put into the reaction vessel and the caustic solution added. I n most cases the sawdust completely absorbed the caustic solution so that the mass was merely wetted, with no excess of water. As the heat was steadily applied, the following reactions were observed: At 100-120" C. boiling set in, and the mass became more homogeneous. The mixture turned reddish brown; and while the temperature remained in this range, considerable water was evaporated. At 170-175" C. the sawdust began to decompose and became a stiff yellowish mass, almost dry in appearance. Gas was given off which did not condense but burned with a blue flame (later shown to be composed in part of carbon monoxide). At 175-180" C. the exothermic reaction began and heating was stopped. The temperature mounted rapidly nevertheless, and a pungent odor was evolved. Large volumes of combustible gases were observed, the color of the mass changing from light yellow to a sickly greenish hue, and the mass became pasty and fluid again. Above 200' C. no more gas or condensate was observed. The liquid mass bubbled quietly and the temperature could be adjusted readily, since the exothermic reaction was complete.
Vol. 34, No. 3
O P T I M U M CONDITIONS
The conditions of this process which were varied to determine their effect upon yield were: ratio of sodium hydroxide to sawdust, concentration of sodium hydroxide, time of fusion, maximum temperature of fusion, depth of fusion mass, relation of air conditions, and species of wood. The tables and graphs from these data define the optimum conditions for the process, as operated in each manner and type of equipment indicated. Other methods and types, or sizes of equipment, would probably give other values; but it is expected that to some degree a t least they would be relative. The yields reported in terms of the caustic recovered are true only for a first run. Commercially the second run in a batch process would use the caustic recovered from the first run, together with enough new caustic to make up for any losses. From the use of the method of analysis described and from a study of this and subsequent work, it is apparent that the considerable caustic loss reported in most of the experiments described later in this paper is not due to actual losses but to the sodium content of salts dissolved in the filtrate after the first evaporation and filtration. Thus, both the relatively low yields and the relatively high caustic losses are due to the same factor. This point is developed in a subsequent paper (page 268). In every case the weight of the sawdust represents dry wood substance and is corrected for the 16.7 per cent moisture present in the chestnut sawdust used, except where otherwise noted. For convenience results are indicated in terms of 100 pounds of sawdust or 100 pounds of caustic soda. Acetic and formic acids are usually represented together as the acetic acid equivalent, since they were not separately analyzed in much of the work reported in this paper. To determine the exact amount of formic acid present in the solution distilled off after acid treatment, a sample was first titrated with standard alkali and the amount of formic acid determined on another sample by Fincke's method @)-L e. precipitation of mercurous chloride. Analysis of several representative runs showed that the average yields were: acetic acid, 8.6 to 9.3 per cent, and formic acid, 1.8to 2.0 per cent of dry weight of wood; thus, in general, about 20 per cent of the volatile acids obtained might be expected to be formic acid. Qualitative tests for propionic acid gave negative results. EFFECT O F R A T I O O F CAUSTIC T O S A W D U S T
In this series of runs the constant factors were: 50 per cent caustic soda concentration, 3-hour run a t 200" C., and 1 pound of dry wood substance. The ratio of caustic t o sawdust was varied with the following results: Ratio, NaOH: Sawdust
1:l
2:l 3:l 4:l
Lb. Acid/100 Lb. Sawdust Oxalic Acetic 3.5 9.2 16.8 8.8 11.4 33.0 36.2 12.0
Lb. NaOH Recovered 0.3 1.37 2.58 3.2
Lb. Acid/100 Lb. NaOH Oxalic Acetic 13.1 5 14 26.6 78.5 24.9 45.2 15
The percentage yields of both oxalic acid and acetic acid increase with the increase in caustic used (Figure 2). The yields per pound of caustic also increase until more than 3 pounds of caustic are used per pound of sawdust. Although the yields of acid are higher, the greater loss of caustic (i. e., pounds caustic per pound acid) did not appear to warrant a greater caustic:sawdust ratio than 3:l. EFFECT O F CAUSTIC C O N C E N T R A T I O N
The ratio of caustic to sawdust was kept constant a t 3 to 1, the runs lasted 3 hours a t 200 " C., and the weight of dry wood substance was 0.25 pound, Figure 3 and the following table show the effect of varying the caustic concentration:
March, 1942 N~OH Concn.,
%
10 30 50 70 90
31.3 31.6 33.1 32.0 9.2
Lb. Acid/100 Lb.
7
Lb.Acid/100 Lb. Sawdust Oxalic Acetic
Na8H Recovered
Oxalic
85 85 86.1 81 54.6
69.5 70.2 78.5 39.6 16.9
9.2 10.0 11.3 8.0
....
NaOH Acetic 20.4 22.2 24.9 10
....
Comparatively dilute caustic solutions merely evaporate to
a higher concentration, without great differencein yields. The lower yields obtained when a concentration of more than 50 per cent caustic is used are probably due to insufficient water for the hydrolysis which is probably an initial reaction, or to charring action from localized overheating. When 70 per cent caustic is used, the effects of this charring are very evident in the decreased amount of caustic recovered. When 90 per cent caustic is used, the charring has such an effect that no acetic acid is formed and only 54.6 per cent of the original caustic is recovered. It is interesting to note that the boiling point of a 70 per cent caustic solution is over 180" C., of an 80 per cent solution about 210" C., and of a 90 per cent solution about 245" C. Thus, localized overheating could easily take place, since before the temperature of the boiling point of the solution was reached, serious overheating might result. Furthermore, the small amount of aqueous solution did not serve t o wet completely or uniformly the sawdust when high concentrations were used. EFFECT O F TIME AND TEMPERATURE
At constant conditions the ratio of caustic t o sawdust waa maintained a t 3:1, the concentration of caustic at 50 per cent, and the weight of dry sawdust a t 0.25 pound. Figure 4 and the table show the effect of varying the length and temperature of the runs: Lb.Acid/100 Lb. Sawdust Oxalic Acetic
N38H Recovered
fall off even at 230" C. due to decomposition of products once formed (or recycled with caustic). The caustic recovery by the method used, however, does fall off by continuing the heating too long. EFFECT OF DEPTH O F FUSION MASS I n these runs the ratio of caustic to sawdust was 3 to 1, the concentration of caustic was 50 per cent, and the time was 3 hours. Results are shown in Figure 5 and the following table: Depth of Mass, In. 0.25 0.5 0.75 3 6 9
Lb. Acid/100 Lb. Sawdust Oxalic Acetic 52.6 51.0 34.2 32.8 33.1 33.0
12.9 14.2 11.2 11.4 11.1 11.4
N%H Recovered 88.6 87.1 85.9 84.1 85.7 86.0
Lb. Acid/100 Lb. NaOH Oxalic Acetia 131.0 132.0 81.5 69.0 78.5 78.4
31.0 86.7 27.6 24.0 24.9 24.2
The first three of these runs were made in the shallow pan, the last three in the fusion pot. The fact that the shallowdepth runs result in such large yields of both acids and also in over 80 per cent caustic recovery may be assumed t o be due t o several factors: (a) heating and agitation give more constant conditions throughout a mass of this nature; (a) probably more important, oxygen from the air is more available for oxidation of the thinner layers; (c) carbon dioxide formed during the reaction can diffuse out and therefore re-
Lb. Acid/100 Lb. NaOH Oxalio Acetia
Time, HI.
Temp., C.
1 3 6 12
170 170 170 170
2.1 6.8 34 54
2.2 3.4 8 11.2
90 87.2 70.6 46
7 17.7 38.6 33.3
7.3 8.85 9.1 6.9
1 3 6 12
200 200 200 200
12 33 68 71
4 11.4 12 12
89.2 86 68.3 41.7
37 78.6 69.4 40.5
12.8 24.9 12.3 6.86
1
230 230 230 230
47 67 64 66.8
8 12 13.8 13.8
72.1 68.7 59.2 31.6
41.6 71.4 52.5 32.7
7.08 12.8 11.3 6.78
3 6 12
265
INDUSTRIAL AND ENGINEERING CHEMISTRY
While the curves for 170' and 200" C. show an increase in yield with an increase in time, that for the highest temperature (230" C.) does not. This and other considerations indicate that a rapid increase to a maximum temperature, followed by a short space of careful control to prevent overheating, may serve to give optimum yields as well as a longer period of time at a carefully controlled lower temperature. The lower caustic recovery may indicate a larger amount of carbonate formation and larger attendant losses. Probably a better control of temperature than was possible in this equipment and better recovery or analytical methods would be necessary t o demonstrate these relations conclusively. It is worth noting that the highest temperature showed lowest caustic recovery, both in per cent and in pounds of acid per pound of caustic. The yield per pound of caustic is of much importance, as already noted. Furthermore, the length of time required for the highest yields (as indicated in these experiments) would require comparatively large equipment for a given throughput on an industrial operation. One valuable point indicated is that even if a long time is unnecessary for the reaction, at least the yield does not
FUSION TIME IN HOURS
Figure 4.
Effect oF Fusion Time on Yields
of O x a l i c and A c e t i c A c i d s and on Caustic Recovery
duce the amount of sodium carbonate formed. The direct recovery of caustic would then be higher. The variation i n the caustic recovery is probably too small, however, to warrant deductions, although i t does show up sharply in pounds
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
266
of acids per pound of caustic. Above 0.75 inch the depth of the fusion mass had no appreciable effect on yield. RELATION OF AIR CONDITIONS Other tests were carried out with variations in the contact of air with the fusion mass. Three-hour runs were made a t 200" C., using three times as much caustic soda as d. 'w. s. and a caustic solution of 50 per cent strength. Several runs were carried as far as the 38" BB. solution and the crystallization of salts. These salts were analyzed only for oxalic acid in most cases, and this yield was used as a criterion for the efficacy of the given set of conditions: DEEPPOTFUSION.Oxalic acid yield, 0.33 pound; acetic and formic acids yield 0.114 pound. SHALLOW P.+N USI ION. Oxalic acid yield, 0.51 pound; acetic and formic acids yield, 0.142 pound. SHALLOW PANFUSIONWITH ADDEDAIR. Oxalic acid yield, 0.515 pound. An electric fan was placed 6 feet from the center of the pan and slightly above it, and kept running to sweep the air from the surface of the fusing mass. SMALL POTMETHOD WITH ADDEDAIR. Oxalic acid yield, 0.37 . The electric fan circulated air over the surface of the our pots throughout the entire run, from a distance of less than 6 feet. DEEPPOTFUSION WITH ADDEDAIR. Oxalic acid yield, 0.456 pound; acetic acid yield, 0.117; formic acid yield, 0.025. A compressor was used t o pump air to the interior of the fusion kettle and thus sweep out the gases formed during the fusion. No attempt was made t o introduce the air beneath the surface of the mass. Agitation was supplied by the mechanical stirrer.
Vol. 34, No. 3
I n another run samples of condensate were collected a t different temperatures during the fusion to determine the methanol concentration in the condensate a t each temperature and stage of the fusion process. These successive analyses probably have little practical significance, since the only important item during a commercial run would be the total amount of alcohol formed. They are shown in Figure 6 and the following table: (Ratio, NaOH:sawdust, 3 : 1: concentration of caustio 50%; time, 3 hours, and temperature 200° C.) Temp. of Concn. Temp. of Concn. Evolution, of CIlsOH Evolution, of CRaOH c. in Condonsate c. in Condensate 160 100 0 3.9 120 1.1 180 3.2 140 3 200 Trace
The yield of methanol is greater than the corresponding yields from destructive distillation of the same wood; but it is less than is theoretically possible from the methoxyl content of the wood, as determined by the Zeisel method (6). Fig-
pound
It is apparent that the yield is increased because of some action such as oxidation of the original lignocellulose by added oxygen. The shallow pan run with air blown in does not show an appreciable gain over a shallow pan run without added air, owing probably t o the fact that ample contact with air FUSION TEMPERATURE- 'C.
100
Figure
6. Concentration of Methanol Recovered during Fusion
80
8 60 I u)
n
wr 40 -I
20
h
' 'A '
O ' o b I I I 3 DEPTH OF MASS- IN.
Figure 5. Effect of Depth of Fusion M a s s o n Yields of Oxalic and Acetic Acids and on Caustic Recovery
is provided in either case. The deep pot fusion with air blown in approached the flat pan yield closely; but the blowing of air over the small pots did not have great effect, probably because of the baffling action of the walls of the pot.
ures presented by Zeisel show that 30 per cent of the methanol is formed by simple hydrolysis and the other 70 per cent by more drastic action involved in the fusion process. This might indicate that there are methyl esters and ethers in the original lignocellulose. Using the Vorlander and Hobohm test for traces of acetone on the condensate, a slight precipitate of dibenzalacetone resulted, showing that traces of acetone are formed in the fusion. Qualitative tests for ethyl and propyl alcohols gave negative results. NONCONDENSABLE GASES FORMED DURING COMBUSTION
Some of the gases given off during the fusion were not liquefied in the condenser used and will be referred to as noncondensable gases. For several runs, four samples each mere taken and analyzed by the Elliott apparatus for carbon dioxide, illuminants, oxygen, carbon monoxide, and nitrogen (by difference). The data from one typical run are as follows:
FORMATION OF METHANOL AND OTHER VOLATILE MATERIALS
During a run made a t the optimum conditions, all the condensable vapors were collected. An unidentified waxy substance and about 15 cc. of unidentified wood oils per pound of d. w. s. were separated from the aqueous layer. These oils boiled a t about 185" C. and amounted to about 0.03 pound per pound d. w. s. By fractional distillation and the specific gravity method, the total methanol content of the aqueous layer was determined to be 5.5 pounds per 100 pounds d. w.s.
Per Cent Carbqn dioxide Illurninants Oxygen Carbon monoxide Nitrogen (by difference)
24.5
Negligible Negligible 72 3.4
cu. ft. Gas/Lb. D. W. 9. 0.615
... ...
1.80
0.osj ~
Total volume of gas
2.500
The high concentrations of carbon dioxide in the gases might explain the low direct caustic recoveries in the deep pot fusion. There was evidently a high conversion of sodium
March, 1942
INDUSTRIAL AND ENGINEERING CHEMISTRY
hydroxide to carbonate due to absorption of carbon dioxide; whereas in the shallow pot runs, absorption of carbon dioxide was much less. The small heating value of the gas (about 500 B. t. u. per pound of sawdust) indicates that little attention should be given these gases as a source of heat. CELLULOSE AS THE CONTROLLING FACTOR
A run to determine whether only the cellulose in the wood reacts with the caustic was carried out with 0.52 pound of purified cotton linters, the cellulose equivalent of 1 pound of chestnut sawdust according to Bunbury ( I ) , under the optimum conditions described above. The yields were as follows :
Acid Oxalic Acetic
From Celluloee (Equivalent of 1 Lb. Chestnut), Lb. 0.304 0.102
From 1 Lb. Chestnut, Lb. 0.33 0.114
It thus appears that a larger quantity of each acid is formed from the wood substance than would be expected from the cellulose equivalent of the wood. Some of the reaction products must therefore come from the lignin and the other constituents. This was checked by various other runs. This other work also appeared to show that not only was the volatile acid fraction low, but the amount of formic acid was larger than the amount of acetic in the cellulose runs. In some cases it appeared that the formic acid from an equivalent amount of cellulose was the total amount to be expected from a given amount of wood, although the acetic acid yield was very low; this indicated that the formic acid may come from the cellulose part of the wood and the acetic acid largely from other parts. Runs were made on the sawdust from several species of wood in order to check these results and to compare the action of the different species. The woods varied from balsa and yellow pine through the range of hardness to sugar maple and hickory, and included white pine, white birch, and chestnut. No important difference in the fusion operation was noted for any of these woods, and all gave satisfactory results. It was not considered necessary to carry the runs through to the final acids; but comparative amounts of total acids were obtained stoichiometrically from the salts formed. As had been expected, the amount of acids calculated from the cellulose content as shown by Bunbury (1) compared well with those actually obtained and were slightly lower. CAUSTIC ECONOMY
Since the efficient recovery of caustic soda used in the process is important to the economy, several runs were made to determine the losses in the laboratory equipment used. I n one run a sodium balance over the fusion part of the system showed that 3 per cent of the caustic soda used was unaccounted for. This figure would vary, depending upon the type and size of equipment used. Obviously, efficient operation of this process would require that caustic liquor containing small amounts of dissolved organic acid salts be recycled to the system. There should, therefore, be a negligible amount of loss of organic acids. Since the yields reported in this paper generally disregard the residual salts in the mother liquor, in actual commercial operation the recovery would be higher. OPTIMUM CONDITIONS
The above experiments indicate that under the methods pertaining, the optimum conditions for the simultaneous production of oxalic acid and acetic acid @re: 3 to 1 ratio of caustic to sawdust; 50 per cent concentration of caustic; 200’ C.
*
267
as final fusion temperature; 0.25 to 0.5 inch as depth of fusion mass; and 3-hour fusion time. It would be expected that these conditions could all be maintained during commercial manufacture. But because of the difficulty of stirring a mass only 0.5 inch thick, the apparent optimum conditions obtained with a layer of this thickness would probably be discarded as impracticable; or equipment for continuous production would have to be devised to give these advantages. Further work indicated that the time of fusion may be dependent upon the rate of heating and therefore is a function of the design of apparatus. The maximum yield seems to be obtained shortly after the exothermic reaction takes place, regardless of the time required to attain that point. The higher yields of the thin layer fusion were obtained in the ordinary deep pot fusion, however, simply by blowing air over the mass while fusion is taking place. It thus appears that from sawdust, caustic soda (to the extent not recovered), lime, and sulfuric acid in approximately stoichiometric amounts, the following materials may be expected to give the products listed: Lb. Material dry sawdust sodium h droxide 34.7 lime (Ca& 61.1 sulfuric acid (100% HsSOa)
Lb. Product Formed 4 5 . 5 oxalic acid 11.7 acetic acid 2 . 4 8 formic acid 6 . 5 methanol 85 calcium sulfate (waste) 3 wood oils
100 9
CONCLUSIONS
Under optimum conditions as outlined in this paper, it is possible to fuse sawdust with alkali and produce high yields of oxalic acid, acetic acid, formic acid, and methanol. These yields may be high enough to permit profitable operation of such a process. Higher yields can be obtained by fusing the mass in thin layers. However, carrying out the fusion in the normal manner and blowing air over the mass gave yields which closely approached those obtained by thin-layer fusions. By carrying out fusions on several different types of wood, it has been shown that a consistent ratio exists between the yields obtained and the amount of cellulose in the wood. The fact that in each case the yield was slightly higher than that obtained from pure cellulose indicated that the lignin in the wood undoubtedly entered into the reaction. The evolution of noncondensable gases containing mainly carbon monoxide and carbon dioxide was noted, as was the presence of a small amount of wood oils and an unidentified waxy substance in the vapors. Material balances showed that there should be no appreciable caustic loss other than by mechanical handling; and since caustic soda is the most expensive chemical used in the process, efficient operation from the standpoint of caustic recovery should permit the successful manufacture of the various products by this method. LITERATURE CITED (1) Bunbury, H. H., “Destructive Distillation of Wood”, N e w York,
D. Van Nostrand Co.,1923. (2) Chem. & Met. Eng., 46,107 (1939). (3) Griffin, R. C., “Technical Methods of Analysis as Employed in Laboratories of Arthur D. Little, Inc.”, 2nd e d . , p. 457, N e w York, McGraw-Hill Book
Co.,1927.
(4) Hilbert, Chem. & Met. Eng., 2 2 , 8 3 8 (1920). (6) Hubbard, E . , “Utilization of Wood Waste”, 2nd ed., London, Soott, Greenwood & Co., 1913.
