The Production of Caustic Soda Solution from Trona Citation Kuffel, George Cole (1923) The Production of Caustic Soda Solution from Trona. Bachelor's thesis, California Institute of Technology. doi:10.7907/R9VQ-3S68. https://resolver.caltech.edu/CaltechETD:etd-02242005-113025 Abstract This problem concerns itself with the possibility of producing a reasonably pure concentrated caustic soda solution using trona as a starting material. It involves a carbonation process, purification of raw material and causticization with lime. Since soda ash costs a little less than three times what trona costs, the process has good commercial possibilities. Item Type: Thesis (Bachelor's thesis) Subject Keywords: Chemical Engineering Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemical Engineering Thesis Availability: Public (worldwide access) Research Advisor(s): Lacey, William Noble Thesis Committee: None, None Defense Date: 1 January 1923 Record Number: CaltechETD:etd-02242005-113025 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-02242005-113025 DOI: 10.7907/R9VQ-3S68 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 731 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 28 Feb 2005 Last Modified: 03 Oct 2019 23:01 Thesis Files Preview PDF (Kuffel_gc_1923.pdf) - Final Version See Usage Policy. 823kB Repository Staff Only: item control page

THE PRODUCTION OF CAUSTIC SODA SOLUTION FROM TRONA Thesis by George 6. Kuffel In partial fulfillment of the requirements for the degree of Bachelor of Science in Chemical Engineering. CALIFORNIA INSTITUTE OF TECHNOLOGY Pasadena, California. 1923. THE PRODUCTION OF CAUSTIC SODA FROM TRONA This problem concerns itself with the psassibility of producing a reasonably pure concentrated caustic soda sol- ution using fona as a starting material. It involves @ car~ bonation process, purification of raw material and caust- icization with lime. Since soda ash costs a little less than three times what trona costs, the process has good com mercial possibilities. frona is a double carbonate of sodium whose formula has been given by Lunge and others a NafCOZNaHCOZHe0. It is a white crystalline solid which loses CO, easily and is decomposed in solution at about 21°C. into the two carbonates and water. It has been made in the crystel form in the lab- oratory. The largest supply of the salt is et Owen's Lake in California. This lake has no outlet and is fed by the Qwen’'s River which brings down large quantities of the carb onates each year. In 1898, Lunge calovleted the amount of trona in and around the lake to be 50 million tons. The lake water is not yet saturated so that a system of solar evaporation is used. The material from the evaporators is sacked and shipped under the name of Owen's Leake trona. w Methods of Analysis in order to have some basis on which to work it was necessary to obtain an analysis of the raw trona. In 1922, Wede Beman” worked on the same material toward the same end py the LUweg process. His analysis showed the absence of appreciable amounts of potassium so that all salts are con~ sidered sodium compounds. Since most of the acid constituents are carbonates of both normel and acid carbonate form, a method was found that would show the amounts of these quickly and accurstely. A modification of the method given in"Quantitative Analysis" by Treadwell and Hall, was used. Five grams of the material was dissolved in weter and diluted to exactly 500 cc. in a standard volumetric flask. Twenty ce. portions were then pippetted to ah Erlenmeyer flask and five drops of phenol phthalein indicator added. Since sodium bicarbonate is neutral to phenolphthalein all the red color is due to the normal carbonate. Very dilute standard hydrochloric acid is then measured into the solution mtil the red color just disappears. At this point the normal carbonate has all been changed to the bicarbonate. Methyl orange indicator is add- ed and the solution reititrated until neutral. In the first titration it is essential that no COp be lost. As e sefe~ guard against this, the acid is added very slowly and the point of the burette is held under the surface of the liquid. Unless the acid is added very slowly and the normality is less than 0.3 N, it is very difficult to obtain re~ producable results. The method also serves as a means of determining the relative amounts of the two carbon» ates, in which case the strength of the acid is not necessarily known. The chloride content was determined by grav- imetric analysis using silver hitrate as a precipitant according to the method outlined by Blaisdale in his "Quantitative Analysis". Several volumetric schemes were tried in an effort to obtain an analysis quickly but no satisfactory results were obtained. The grav imetric method consists essentially in preparing a sol- ution of known amounts of materiel, precipitating the chloride with silver nitrate from the solution made acid with nitric acid, boiling the mixture in order to coagulate the precipitate, filtering onto asbestos in Gooch crucibles of known weight and weighing after drying to 200°C. The other more important constituent is sul- phate. Gravimetric methods were also followed in this case. Gooch filters consisting of asbestos in the ord» inary Gooch cruchble were prepared and ignited in a blast lamp. After cooling in a dessécater, they were carefully weighed. Te a boiling, acid solution of the material, a hot saturated solution of barium chloride was added in sufficient quantity to be sure that the barium was in slight excess. The precipitate was allowed to stand over night. The supernatent liquid was decanted and the precipitate washed into the funnel. After drying at 110° until all the water had been driven off, the crucibles were again ignited, cooled in a desg- ¢cator and weighed. Using the above outlined methods, the fol- lowing analysis was obtained: Total carbonates 74% Naptos 41% MalCO- B3% Nacl 4.2% NapS04 9% No attempt was made to determine the amowut of weter contained or the quantity of insoluble matter present, both of which were considerable. Need of Purification It depends entirely on what uses of the caustic produced whet impurities are allowable in it. In the case of soap manufacture, there would be no ob- jection to the presence of chlorides. However, in order to have a generally marketable product, sten dard requir- ments must be met with regard to the purity of the caustic. The only practicable method devidsed was simply washing the very soluble impurities from the relatively insol~ . uble sodium bicarbonate. Since the total carbonate content was over half normal carbonate, this portion had to be converted into the acid form. Two general methods of earbonation were tried, the wet and the dry. Wet Carbonation The trona was dissolved in as little water as possible, below 70°c. The temperature was raised no further because it has been shown that sodium bicarbon- ate loses carbon dioxide at that temperature in apprec- jable amounts. The solution was then filtered warm to remove the insoluble matter, largely sand and other Silicate materials. The filtrate was then cooled and earbonated at the same time. Carbonation was effected by bubbling carbon dioxide through a jet in the solution. The carbonated solution showed marked supersaturstion. Even upon cooling to as low as 0°s., violent agitation was required to precipitate the bicarbonate. It then came down in varying forms, but on standing for a short time was easily filtered and was mostly erystalline. A preliminary run showed a yield of 45 grams of bicarb- onate from 100 grams of raw trona. This is only about 56% efficient. The next modification tried was the use of a, saturated salt solution as a solvent for trona and then following the same procedure as before. Saturated salt solution dissolved only 6 grams of trona per 100 cc. of solution when warmed to 50°C. and continually stirred. A precipitate formed in the vessel as the trona went into solution. The salt solution was carbonated but with practically the same difficulties as before. The sol- vbility of the bicarbonate was cut down and was easier to precipitate but there was a correspondingly smaller amount of it in solution, so that supersaturation was still a hindrance. Because of the low efficiency of the process and the necessity of the use of heat and subsequent cooling to O° or lower, this process was abandoned for the present in favor of the dry carbonation. Dry Carbonation ' A reference was found in Lunge stating that unhydrated normal sodium carbonate will combine with equimolal quantities of carbon dioxide and water to form two mols of the bicarbonate. From the formula assigned to trona it would appear as if there were enough water present for the reaction. This was tried by running carbon dioxide into a mass of the trona ery-~ steals. There was no appreciable absorption of the gag. Obviotisly it was necessary to add water. In order tc carry out this process an apparatus was set up as shown in Fig. 1. 2 gram Carbon dioxide was generated in a Kipp generator and mixed with steam in an ordinary T tube. The reaction tube was set in motion and the mixed gases sent in at a moderate rate. The solid did not completely absorb the gases as there was a considerable volume lost thru the exit tube. After a short time the reaction began and could be followed from one end of the tube to the other by the "hot spot". In one place the tube was sev~ eral degrees higher than it was an inch on either side of the spot. Difficulty became apparent during thes first run when the material was removed from the tube. It had formed an extremely hard cake on the walls of the tube and apparently was impervious to gas flow. In run #1, 50 grams of trona was placed in the revolving tube. Carbon dioxide was generated from an excess of marble chips and 100 cc. of 6N HCl. Twenty-five cc. of water was sent into the apparatus as steam. An analysis of the product showed only 90% of the total carbonates in the raw material. This was due to the addition of waterwhich was not entirely eliminated before sampling. The two cartonates were divided: 33% Na,CO,and 40% NaHS0,. Calculation showed that only 26% of the CO, produced was retained in the solid and that a large excess of water was used. The results of this run were better than those of the wet method but unsatisfactory for use in a commercial scheme. In the next run a different method was used. In order to maintain equimolal quantities of carbon dioxide and water the gas was bubbled through water held at a constant temperature at which the vapor press- of water equalled a half atmosphere. This temverature was calculatted to be 80°. lo attempts were made to economize on gas as the rate of Flow was fairly rapid and the carbonation extended over about three hours. In order to break uv the enke which formeé on the suf» face of the tube a number of steel rollers were into- duced. They did not accomplish their purpose after the cake had once been started. The results of this run showed the normal carbonate to be present as 17% of the total 74% of carbonates. This therefore, showed a dec- ided improvement on any previous run. The next run was the final small scale experiment and was carried out in practically the same way as those preceeding but with better control on the amount of water that was mixed with the gas. The temp- erature of the bath surrounding the flask through which the gas was bubbled seldom rose above 70°, The cake in the tube formed as before but was not nearly so hard and tenacious. Results of this run showd 90% conversion of the normal carbonate to bicarbonate. The success of this run may be attributed to the fact that the tube was heated by an improvised steam jacket consisting of a piece of stove pipe surrounding the reaction tube. Stemm was blown into the space between and the reaction was accelerated to such ean extent that the time taten for this run was less than either of the others and yet the conversion was practically complete. The only piece of large apvaratus approximat- ing the conditions expected in specially designed carb- onators that was available, consisted of a steam jacketed vacuum crystallizer,fitted with scrapers. A diagram of the complete set up is shown in Fig. 2. Steam was piped to the jacket and the tempsrature read by thermo« couples. Carbon dioxide was introduced through a water trap which was kept at 15. The gas was provided from a cylinder. During this run there was a leak in the Fig 2. ‘Boi /er Therrro couple CO, ~~ SS yz Conreetron ( | Heat, vg —> Stearn * v | g \\ Crain a SCVYAP CFS J —— fn — | | [= Oriving Gear sacketed Cysallizer Used as a Carborator ae 5 4, which was net located sc that no control of the pres- ae eI sure within the machine was possible. After heating the inside up to the desired temperature of 60°C., the heat of the reaction and the entering steam kept the temperature at that point end even reised it two or three degrees. Under these cond- itions the gas was passed in for a period of 6 hours. The sample used was 25 pounds of trona put through a Braun Uhipmunk jaw crusher. Seven pounds of carbon dioxide was used. It wae aprarent that water was being carried over by entreinment so that it did not enter the reaction space as vapor but as liquid. Since the materiel caked very easily if wet, the whole interior of the machine was found tobe coated with a hard layer of the carbonated trona. Analysis showed that 96% of the normal carbonate had been converted to bicarbonate. This part of the process would have been satisfactory hac it not been for the inconvenience of removing the finished product. In the first run the materiel accumulated on one of the agitators and left the other idle. To obvi- ate this difficulty twice as much material was used for a charge. Since the excess water was known to be caus= ing the caking, two inlets were used, one for steam and one for carben dioxide. The steam was generated in a graduated flask so that control could te kert on the il rate of steam consumption. The carbon dioxide entered directly through another opening. This time the pres- sure was under control and served as a means of showing the speed of the reaction by the rate in which the gas was absorbed. During most of the run a pressure of one half inch of mercury was kept. This varied because it was impossible to keep step with the changing cond- itions with the erude controls on the gas flow. Fifty pounds of trona were carbonated with 12 pounds of carbon dioxide and 4 pounds of water. No test could be obtained for sodium carbonate, showing complete conversion to the bicarbonate. In this run the material came from the machine in small round pell-~ ets averaging about 1/8" in diameter. It showed no further tendency to form a cake except on the flat sur- faces of the iron where a comparitively loose coating was found. The results of this run show beyond 4 doubt that with proper control of water added, trona can be carbonated on a large scale. There appears to be no reason why a long rotary carbonator would not be suc~ cessful. Provisions for heating do not have to be elaborate because after the reaction has once been well started its om heat keeps the material up to the proper temperature. Moisture control would be a neessity, 12 otherwise, the rolling action of the kiln-like carbon- ator would hot be enough to keep the trona from coating the walls and forming a solid cake. The last run in the vacuum erystallizer showed that it was possible to con= trol the amount of water added until there was little or no tendency for caking. The source of carbon dioxide offers some prob- lems. The raw trona is 33% NaHCO, and 40% Naelog. After earbonation it is practically all changed to the bicarbonate. For the purpose of causticization a given amount of lime will produce twice as much NaHO from NaecOZ as it will from NaEcoOgz. Therefore, between the carbonation and the causticization a heater ean be placed which will drive off all the CO, held in the form of bicarbonate and leave the normal carbonate to react with the lime. Using this source of COe the larger part of the gas required is provided for. The remainder can be obtained from either a burner or from limestone kilns. If the size of the plant warranted, a lime kiln would probably be usé@d to produce the CaO used in caust- icization. The gas derived from either burning or from the lime kiln has the disadvantage that it is contamin- ated with nitrogen and carbon monoxide to say nothing ofp the possibility of some volatile material from the fuel. For this reason, it would probably be advantag~ eous to carbonate in two steps. First with the impure gas and then finish with the pure CO, from the bicarb- onate reduction. Washing Since the sodium bicarbonate is relatively insoluble as compared with the compounds which are im- purities in this case, it should be possible to wash from the bicartonate the chlorides and sulphates which render it of less valve. Twenty-five grams of the com- pletely carbonated trona were put into a Buchner funnel and washed with 25 ec. of pure water. Vacuum was applied and 22 ca. were drawn out of the mass. The solid showed a mSium test for chlorides by use of silver nitrete. The same procedure was followed in a second trial.. This time the liquid was put through three times, being drawn off each time by vacuum. After thie 10 cc. of pure water was put through. The solid showed no test for either chloride or sulphate. After the first large scale run, a washer was constructed as shown in Fig. 3. Fer F Botton To Vac. Furnp. 14 It was constructed by soldering a false bottom onto a can after perforating the upper bottom. Material was put into the can until the cake was about 4 inchee thick. Nater was added a little at a time until a liter had gone into the solid. About half of this volume wad drawn off by vacuum and put through twice more. Two hundred fifty ec. of pure water was then put in and the vacuum applied. Tests of the solid showed marked chlor- ide content. As a small centrifugal was available, a run was made to see whether it would prove practical aga means of washing the material. Seven pounds of carbon- ated trona was added to enough water to form a thick sludge. This was put through the centrifugal and washed with about a quart of water. The solid, which was re- markably dry, gave a strong test for chloride. Since both of these methods failed and the laboratory experi- ment was successful, some difference in the conditions must have been respnsible. The greatest difference was in the thickness of the cake being washed. In the lab- oratory experiment this cake was approximately $" in thickness and in the other two, much more than that. If a piece of apparatus could be devised to wash a large quentity of solid without forming a thick cake it ought to do the work. Such a machine is the rotary filter 15 which runs continuously and forms only a thir caxe. By specially designing the outlet ports, and mechanically controlling the wash water feed, it would be poscible to use one of these filters with two different kinds of wash liquor. The sludge can be made up from a saturated solution of bicarbonate without seriously affecting the solubilities of the impurities in it. The first wash can be another saturated solution to te used only for that purpose, While the last wash should be pure water, Time and apparatus were not available for the larre scale testing of this part of the process. Causticizing After the carbonated trona is washed free from the chlorides ard sulphates, the wet cake which will be about 1o% moisture, is run into a heater where the water and UO from the NadCo, is driven off. The gas is led pack to the carbonators and NaetOg is left. At this point the product is identical with material being hendled commercially at the present time. Further experimental work was therefore considered unnecessary and the following is an outline of possible operations to complete the production of caustic soautions. The first problem consists in getting the concentrations and proportions such that maximum yield is obtained in the shortest time using the least amount of Ca0. It hes teen found that concentrated eartonate pe solutions react faster with the Lime but do not sive as good yields. lLunge states that the optimum concentrat- don for liquor to be cauatizgized is one of 1.1 specific gravity. Such a solttion contains 9.25% Ha v0, and is approximately 2 normal. In sn erticle* on the same subject it was stated that &£.5N solution reached equil- ibrium with the lime in } hour if well agitated. From these two stdements it seems safe to assume the action to take abovt 40 minutes or an hour, if a slight excess of lime is used and the material well heated and agit- ated, using 2 normal solution of Napv0,. Apparatus hae been designed to causticize continuously. It consists essentially of @ series of agitated tanks through which the mud and solution are run in succession. Dorr thickeners are used as a first separation of solid and liquid. The nearly clarified liquor is filtered and run to the evaporators. The advantages of such a system sre that a uniform solution is obtained and little attention need be naid to the *Le Blane and K. Novotny "Z. ancorg. Chem." 51 181-201. 4 wd puts 16] outfit once it started and regulated. Batch caust- icizers are commonly used. They are simply tanks fitted with agitators run mechanically or with steam jets. The lime is suspended in the tank in a eprid toe retain the large pieces. With these machines the process follows this general path: The carbonate solution of proper strength is put into the causticizer and brought up to boiling by steam heaters. The lime is then added always in the unslaxed condition in order to utilize the heat of the reaction. After agitation from + to 11 hours the suspension is allowed to settle for + hour. The clear liquor is drawn off by & swivel pipe and led to a settling tank where it is left perfectly quiet. The mud 2 ‘ is agitated with enough water te form a sludge ané a; is, &llowed to settle. The thin liquor is drawn off ané used fof the next run. After all these settling operations, at a ¥ the amount of solid in the liqvid is so small t filter press is practicable for senarating the last of the suspension. Concentration The eaustie solutions so proéuced are very dilute. They might be used as such if the demand and the Supply were close enough together but for the purpose 7h of shipping it is necessary te concentrate the solution. Many systems have been propesed and tried ovt. Ine few cases the dilute liquor wae used as Dotler feed. This % ee per*eat cgontrei wn ~ 7] ct at) 3 es ia} cs “ ;# 2) {Q ce Py 2D ta th t S | we {9 $4 s ms WIS S wept on the tcoiler but careless oneretion caused rreat Multinie effect evaporators are generally uséd. vevaration ig cerried en in these systems up to the point of eoncentration causing the precipitation of Summary This process has been developed with the idea of using Owen's Lake trons as a starting material for the preduetion of caustic soda solutions. The following results have been ohtained: 1. The trona contains no impurity rendering it impractical for use in this process. 2. ‘Trona can be converted to sodium bicarben- ate by [Cy and steam, quantitateively on a large scale, but requires careful remvletion te orevent the preduct from absorbing too mech water anc becomine tarved in tie asposratue. oe The chloride ané sulphete sortent can be removed py washine after earbonation. used in incustries already operating, so that the commercial epplicstion of the process is assured if the operations involveé do not prove tco exrensive. a 2 rey, S 4 ~, pa a) ay ae a ne ae ' Ca Pa, professor of Applied Jhemistry st ths California Inst- be be ns, if * wag * 4a es » oA fay ee ed oe Gitute, for his invaluable help and advice in gatheri: and putting togetrer the data here set forth. Bibliography. Little materiel could be fomda on this auhe jset, even on the widely tised commercial process pr The following Iterature was consulted: Thesis--"The L'weg Proeass for Caustic Soda using Trons as a Raw Material.” Weds Beaman 1922. "Sulphuric Acid end Alkelid™ by Lunze. 1909 weveral articles and analytical books noted aa the t: he i) xt. ov