Dehydration of Alcohol by Potassium Carbonate Citation Prentice, Leland Busby (1925) Dehydration of Alcohol by Potassium Carbonate. Bachelor's thesis, California Institute of Technology. doi:10.7907/ADR8-HW51. https://resolver.caltech.edu/CaltechTHESIS:10172011-143206819 Abstract No Abstract. 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): Unknown, Unknown Thesis Committee: None, None Defense Date: 1 January 1925 Record Number: CaltechTHESIS:10172011-143206819 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:10172011-143206819 DOI: 10.7907/ADR8-HW51 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 6717 Collection: CaltechTHESIS Deposited By: INVALID USER Deposited On: 17 Oct 2011 21:57 Last Modified: 03 Oct 2019 23:53 Thesis Files PDF - Final Version See Usage Policy. 2MB Repository Staff Only: item control page

DEHYDRATION OF ALCOHOL by POTASSIUM CARBONATE THESIS by Leland B. Prentice In partial fulfillment of the requirements for the degree of Bachelor of Science in Chemical Engineering CALIFORNIA IUSTITurE OF TECHNOLOGY Pasadena, California 1925 INTRODUCTION There has been considerable investieation into the feasibility of concentrating alcohol in fairly dilute solutions up to 95% alcohol, the product ordinarily sold on the market. by means of solid dehydrating agents such as anhydrous salts or metallic sodium. This investigation has been inspired by the hope of eliminating large and expensive fractionating columns and other distilling equipment. The use of calcium oxide as a dehydrating agent is a possibility, but in many cases where a pure product is required, the Bomparatively high solubility of the calcium compounds prevents the use of this substance. Copper sulphate has also been investigat- ed. The purpose of this research has been to inquire into the fea~ibility of, and to study a method for, the use of potassium carbonate in dehydrating alcohol, with a view to putting this process on a commercial basis. No record of previous work on this subject could be found in the literature. The pDinciplesunderlylng this dehydration process are the~e: When anhydrous K2C03 is added to a dilute alcohol solution a certain amount dissolves. As more is continual- ly added, two liquid phases are formed, the lower containing 2. mo st of the K2C03 and but little alcohol; and the upper containing but little K2 C03' and decreasing amounts of water. When the liquid system will not dissolve any more KaC03,the a lcohol layer has been concentrated to 91% by welght of alcohol ~ (hereafter all alcohol percentages will be given as per cent by weight) and shows but a trace of K2C03 . The aque- ous layer has dissolved about its own weight of K2C03 . If further anhydrous salt is now added, a new solid phase appears which is a hydrated K2C03 and probably has the composition, K2C03 . 2 H2~; and the aqueous phase disappears . If, however, the aqueous layer is first removed and anhydrous salt then added,this new phase forms with the removal of water from the alcohol, and much less anhydrous salt Is needed than is used when the a queous layer is not first removed . 3. EXPERIMENTAL METHODS AND RESULTS The first problem that arose in the execution of this investigation was to decide on a method of analysis . Since there is no simple chemical method of analysis of alcohol water - K2C03 systems, physical methods were relied upon. The density has been taken as a measure of the ratio or alcohol to water . The solutions of interest in this work contained 80% alcohol or above , a very little K2C0 3 , and the rest, water; or else they are about 50% water and 50% KZC0 3 with but little alcohol . For the latter case no estimation of the amount of alcohol will be made due to the lack of a convenient method . A mixture of 92% alcohol saturated with K2C03 showed a density of . 8170 . It was then distilled, leaving a very slight residue of KzC0 3 • . 8175 . The distillate showed a density of In the following work the density has therefore been assumed to measure the per cent of alcohol in concentrated al cohol solutions saturated with K2C03 , and the effect of the K2C03 on it ha s been neglected . A sample of the commercial KZC03 used in this work was weighed, heated for two hours in an iron dish and reweighed . A loss of weight of about 3i% occurred due to the prese nc e or moisture . All references to "anhydrous KZC0 3 "Wil1 be under- 4. stood to mean this salt which actually contains 3i% water. A preliminary run was first made to determine qualitatively and semi-quantitatively the effects when anhydrous K2C03 is added to a 50% alcohol solution . 500 c.c. of 51% al co- hol were used, and to this was added K2C03 : 10 grams K2C03 gave 2 6 . c . water phase 20" " n 25 " " " 30" " " 60 " " " " 70 n " n " " " 100 " " n 60 " " " " " 120 " " " 70 " " " 135 " " 80 " " " 1§0 " 100 " " " 80 " " " " " " 125 " " " " 205 " " " 40 50 250 required to saturate . 500 c . c . of 50% alcohol were shaken in a liter separatory funnel with different amounts of anhydrous K2C03 unt1l equilibrium was attained , and the resulting phases were then analyzed . 5. The following results were obtained: 1K2C03 Water 500 c . c . of 50% Alcohol used Layer Alcohol Layer b rams c. c. s ~~CO~ ~H20 60 349 75 25 56 339 80.5 19 . 5 c . c.s ~lcohol %H20 \ 100 I 178 40 \ 125 \ 205 44 1150 224 45 55 329 84 . 5 . 15 . 5 175 242 51 49 318 87 13 1200 260 51 49 307 90 10 I I I In the above table the K2C03 in the alcohol is very small and is neglected . An attempt was made to distil the liquid of the aqueous layer off from the K2C03 and then determine the ratio of alcohol to water from the density of the distillate . However this was found to be impracti cable because of the large In an extensive investigation of this problem some such method would doubtless have to be resorted to , as the losses of alcohol in the aqueous layer are apparently quite appreciable , and might even prove to be an important deciding fac tor in the htteI'uln'eS1! of the whole process . An investigation was next made of the dehydrat ~on of alcohoI by contact with the anhydrous K2C03 ' but one liquid phase being present . Eighty- six percent alcohol was used, as it 6. was thought that that concentration would probably be attained by the salting out process, in which two liquid phases are present. 200 c . c . of 86% alcohol was shaken with an excess of anhydrous K2C03 , (75 grams were used) in a 500 C. C. bottle . The shaking was carried on vigorously for 3 minutes and a concentration of 94% was attained . The mixture was then shaken 4 more minutes and the alcohol concentration rose to about 94 . 3% . It was thought that some time might be required for this dehydration in a commercial scale process , as much more inti mate ' contact would be attained by hand shaking than is possi ble in any ki nd of machine agitator . Accordingly , a 500 c . c. bottle was wi red to a wheel geared to an electric motor so that the bottle was made to revolve and cause a tumbling over of the contents . A sample of 200 c . c . of 86% alcohol and 75 grams of anhydrous K2C03 was mixed in this device for 40 minutes and an alcohol of 93% was obtained . After 25 ~ore minutes of shaking, the alcohol attained a concentration of 94% . Since K2C0 3 is comparatively expensive, it must be recovered after use . The problem therefore arises of dehydrating the dehydrating agent . To observe the qualitative effects, a concentrated solution of K2C03 was heated in a small iron dish over a bUnsen burner . Except for a spattering just 7. before the free water was all driven off, no d iffi culty was encountered . The anhydrous material forms a hard cake which adheres to the container, so that it may be expected that precautions will have to be taken to prevent a serious reduction of the heat transfer in this process . Having learned these things about the process# it was next necessary to devise a procedure for the whole working process . It would obviously be desirable to carryon both the prelimi na ry dehydration ~ by salting out , and the final dehydration by contact with the anhydrous salt, in the same container , and thus avoid the problem of transportation of the final pasty mass . 8. The following process was decided upon : I water I r-- concentrated K2C03 solution to be evaporated 50% al cohol (raw material) I _ ~ r-- K2 C03 2 H20 and adhering a l cohol l 85% alcohol J I anhydrous K2 C0 3 ! 95% alcoh ol (final product) 9. For preliminary work this process was carried out in a glass jar about 5 inches in diameter . Stirring was accom- plished by a brass propeller 4 inches long of thiP gauge . The propeller was placed almost touching the bottom of the jar in order to prevent the formation of a cake on the bottom and thus reduce the exposure of new surface . In api te of all precautions, however , tendency for cake formation coul~ not be entirely prevented . One dehydration test was made in which the anhydrous salt was wetted with 95% alcohol before it was mixed with the more dilute alcohol, and the cake formation was reduced materially in this way . The first test made was started with 800 c . c . of 90% alcoAfter stirring, 550 C. c . of 94% alcohol were decanted . 1000 c . c . of 50% alcohol and 200 grams of K2C03 were then added to the pasty mass of hydrated K2C03 and alcohol, and the whole stirred until the solid phase had disappeared and equilibrium was reached . hol solution was 89% . The resulting alco- The lower aqueous layer was then siph- oned off and 200 grams of K2C03 added . After vigorous stir- ring the alcohol was concentrated again to 94% and 500 c.c . removed . To be sure that the whole process had come to e- quilibrium as far as the amounts of substances involved were agais c.c. of a concerned , the cycle was again repeated and~47 94% alcohol were obtained . With the object of reducing the amount of K2C03 needed per 10. given amount of 95% alcohol obtained, additional runs were made in which all K2C03 used in one cycle wae s.dded in one batch, instead of two; and the alcohol was not brought up to ae . n~gh a concentration as 89% in the first step. The next run was started with 1000 cc of 50% alcohol and 300 grams of K2C03 added to the 89% alcohol remaining. Aft- er again stirring, 380 c . c. of 94f. alcohol were removed. Another 1000 c . c . of 50% alcohol were then added to the pasty mass and the mixture stirred without the addition here of more anhydrous salt . After the solid phase had disappeared, the aqueous layer was siphoned off and 300 grams of a~~ydrous K2C03 were added . were obtained . After stirring, 600 c.c. of 92% alcohol On repeating the whole cycle again 500 c . c. of 92% alcohol were obtained. From this was assumed that after the process had settled down to a uniform running condition, the average output would be 500 c . c . of 92% alcohol per 1000 c . c . of 50% alcohol per 300 grams of K2C03 introduced . The small amount of final alcohol (380c.c . ) obtain- ed in the first cycle may be accounted for by the large amount which must have adhered to the hydrated ~2C03 present . A run was next made using 350 grams of K2C03 per 1000 c . c. of 50% alcohol. Running conditions settled down after three cycles and it was found that on an average about 470 c . c . 11. of 94% alcohol were removed per 1000 c . c . of 50% alcohol per 350 grams of K2C030 Some difficulty was encountered in the above run before it was successfully finished. The anhydrous I 2C03 used for the fl8$1 dehydration had been pulverized to a fine dust so that after the alcohol had been concentrated to 94%, the K2C0 3 remained as a finely divided suspension and the alcohol could not be decanted . It is therefore probable that some sur- face will have to be sacrificed in this step to avoid the necessity of filtering instead of decanting the alcohol. Appa- rently about 60 mesh particles will give enough surface for the dehydration and still be large enough to prevent the formation of an undesirable suspension. SEMI -- ~ORKS SCALE OF EXPERIMENTATION On the basis of the above laboratory experimental work, the transition to a semi-works scale was next attempted. Both time and equipment proved inadequate for a thorough investigation of this phase of the problem . It is believed, however, that some valuable information was obtained, much of which is of a qualitative nature . The mixing operations were carried on in an oak barrel provided with two stationary hori~ontal paddles and 3 moving paddles revolving at a speed of 60 R.P . M. shown in Fig . 1 . This apparatuB is "- /1/1/11 \ \ \ \\ 12. L I I I l I J l \. I I \\ I '/ F/g../ The stirrer shaft was made of ! inch iron pipe . The stirring bladee were of wood 3/4 inch ·thick , and presented a stirring surface 1- 3/4 inches wide . These blades were fast- ened to the shaft .by inserting a t inch cross in the shaft and wiring to this . It was found in the first test, however, that such a large torque was developed in the final dehydrating step by the large quantity of solid material which must be moved in addition to agitating the liquid , that this type of construction was not sufficiently rugged . of ! inch pipe Ac cordingly, a piece - about a foot long - was screwed into each side of the cross and wired to the wooden paddles in the manner of Fig . 2 . 13. Fig . 2 . This bracing proved to be satisfactory . The two stationary paddles were 2 x 4's drilled 7/8 inch for the stirrer shaft , and nailed to the side of the barrel . Solutions were removed from the barrel by s1phoning with a rubber hose , as in th1s way either the upper or lower phase could be removed with ease . The barrel was provided with a wooden cover to reduce evaporation losses of alcohol . The most satisfactory apparatus available for drying the concentrated solution of K2C03 to the anhydrous salt was an iron crystallizer which was s t eam jacketed and closed , except for a vapor line leading from the top, and a door about one foot square 1n the top . This crystallizer was provided with a heavy two - bladed stirrer revolving slowly and just clearing the bottom and sides . The stirrer was effectual in maintain- ing the bottom and wall free from cake formation, and therefore 14. suggests ideas for an agitator for the final dehydration step. Lack of equipment prevented the removal of the water vapor by the jet condenser connected to the vapor line . To ac- complish this function, a blower w~s therefore trained on the surface of the evaporating liquid through the door in the top , of the crystallizer and the air and vapor expelled through the vapor line and two ports 1n the top of the kettle . In this apparatus the K2C03 was dried to a moist crys talline mass , which was then removed to a sheet iron plate . It was !pread upon this and heated by gas burners until all water had been driven off . The resulting lump! were pulver- ized to about 60 mesh and the mater i al was then ready for another run . For the first test on the semi - works scale , 73 pounds of X2C03 and 25 gallons of 50% alcohol were used . These figures correspond to the 350 grams of K2 C03 and 1000 c . c . of 50% al cohol found advisable in the work on laboratory scale . The test was started with 22 gallons of 80% alcohol and After stirring, an alcohol of 92% was ob talned . This was drawn off and 25 gallons of 50% alcohol were added . 15. The test was made through two cycles. The first cycle produced 15 gallons of 91% alcohol, and the second cycle produced 15 gallons of 90% alcohol. (In the second test the re was a shortage of 12 pounds of K2C03 . This might bring t he al- cohol up to 91% . ) Lack of time prevented the further investigation of this problem. CONCLUSIONS On the basis of the semi - works scale experimentation, re- commendations will here be made for the design of a plant to operate this process: The stirring tanks wlll have to be operated in small units because of the dlfficulty encountered in securing efficient agitation . The agitator blades should be horizontal rather than vertical . The bottom blade should be of sturdy construc- tion, as it will be heavily loaded . It should barely clear the bottom and sides of the tank. The revolving blades should all widen at their ends to cover the space between the stationary blades al most completely . It is very important that these precautions be taken to prevent cake formation in the second step of the dehydration, as this materially lowers the efficiency of the p rocess. 16 . It is recommended that the recovery of the K2C03 be carried out in two stages . In the first stage, the concentrated solution is to be evaporated in salting-out evaporators designed for continuous operation. A double seal or "salt catch" is to be used for the removal of crystals . Multiple effect evaporation may be used i f the size of the plant seems to warrant it . It i s recommended that the moist crystalline mass be dried in thin layers on a hot plate , The plate may be heated by gas or oil burners , but the mass must be maintained above 135 deg . C. For large quantities the process should be made automatic by moving the plate and providing means for scraping off the dried produc t. For pulverizing the dried product a ball -mll1 1:s recommended . Many thanks are due to Dr . W. N. Lacey for his v8.1uable aid and suggestions , and to Dr . A. A. Noyes, at whose suggestion the use of K2C03 ' was tried for this work . 17. SUMMARY 1. Investigation was made of the system Alcohol - K2C03 - Water. and a process devised for the dehydration of alcohol on the laboratory scale. 2. This process was carried out on a semi - works scale with the view of designing a commercial plant for the carrying out ot this process . 3. plant . Recommendations are made for the design of such a