The Lowig Process for Caustic Soda Using Trona as a Raw Material Citation Beman, Willard Jarvis (1922) The Lowig Process for Caustic Soda Using Trona as a Raw Material. Bachelor's thesis, California Institute of Technology. doi:10.7907/6J5Q-Z168. https://resolver.caltech.edu/CaltechETD:etd-04192004-143549 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 1922 Record Number: CaltechETD:etd-04192004-143549 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-04192004-143549 DOI: 10.7907/6J5Q-Z168 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 1414 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 22 Apr 2004 Last Modified: 03 Oct 2019 23:08 Thesis Files Preview PDF (Beman_wj_1922.pdf) - Final Version See Usage Policy. 664kB Repository Staff Only: item control page

Alles, Gordon A Ames,Panl R Beman;illerd <3: Bridgeford, Prank R. Burks, Gillies, Robert W Jesse C Knight,Alfred ¥ Reynolds, Maynard 5S Ritchie,Chéerles Fisher Vesper, Warner, Howard G Lester 0° THESSS 1922 A new acridine dye. The specific heat capacity of heavy pétroléum products. ™he Tiowig procese for caustic soda using trana as a rew materio?, A process for the utilization of pyrite ash. Seni-works development of a germanium extraction process. An attempt to dehydrogenate cottonseed oil. Perchloric acid from ammonium perchlorate and oxides of nitrogen. The utilization of a denosit of magnesite located nesr Porterville,California as a source of raw materiel for the manufacture of C.P. Epsom salts. An investigation on the production of sodium thiosulfate from trona. The preparation of carbon monosulphide and hydrocyanic acid by the use of activated carbon. The reaction between nitrogen and methane in the silent electric discharge. California Institute of Technology Pasadena, California CLASS OF 1922 Department of Chemistry Wndergraduate Theses of Gordon A.Alles Paul R.Ames Willard J.Beman rank R.Bridgeford Jesse C.Burks — Robert W.Gillies Alfred ™.KXnight Maynard 5.Reynolds Charles *isher Ritchie Howard G.Vesper Lester 0. Warner THE LBWIG PROCESS FOR CAUSTIC SODA, using TRONA AS A RAW MATERIAL. Thesis by ? Willard J. Beman In partis! fulfillment of the requiremente for the degree of Bachelor of Science in Chemical Engineering. CALIFORNIA INSTITUTE OF TECHNOLOGY Pasadena, Californis. I Introductior Caustic soda is one cf the more important heavy chemicals and has wide industrial use. It would probably have greater use, if it were not fcr ite high present ecst. Trona is a combination of Na,CO,, NaHCC,, and H,0. The formula to which it mest nearly corresponds is Na,CO,-NaHCO,.2H,0. There are large watureal deposite of trona in various parts of the world. Large quantities are crystallized from Cwen's Lake, Inyd County, California, during the summer months. There are severe) plants around the lake for this purpose end the preduct is relatively cheap. This raw meterial offers the possibility of producing caustic soda at a lower cost provided a suitable process can be developed. The quantities of the various salts in the brine have been estimated, 29 in alllions cf metric tens, to be as follows: WNa,CO, — 46.6; NaHCO, — 11.08; Na,SO, — 19.43 NaCl — 51.53 KCl ~ 5.53; Na,P,0, - 3.58.+ II Previous Work There have been two methods used for making NaOH from Na,C0,. The most used method is by causticizing with Ce(OH),. The other jethod has had very limited use and is known as Lawig's process. The following description of Léwig's process is from Lunge” (Brit~ ish patents 7364 of 1882, 591 cf 1897, 4227 of 1897, and 1974 of 1887). chemical and Vetallurgicel Engineering, 24 (1921) 683-8. - See also: Lunge, Sulfuric Acid and Alkali, vol. 2, part 1, 71-7. Jour. Am. Chem. Soc., 28 (1906) 1164-70. “sulphuric Acid and Alkali, vol. 2, part 2, 821-35. See alec: Jour. Soc. Chem. Ind., 2 (1883) 280. . 16 (1887) 1015. | An intinate mixture of sodium carbonate and ferric oxide, when ex— posed te a bright red heat, expells CO, and sodium ferrite (NaFeC,) ‘is formed, which is quickly decomposed by hot water into NaOH and Fe,05+ “The Fe,0, must de pure, especially free from alumina end silica, and is used as a granular powder. The heating is done in two stages, the first being at a cherry—red heat for $ of an hour, which prepares a mass held together by fused sodium carbonate — it is then powdered, well mixed, and heated in a muffle for another § of an hour, by which time all of the CO, should be criven off, The melt is dark green, heavy, harc and of a sandy ard grainy tex— ture. The mass is then washed with cold water tc remove soluble " Gmpurities, after which the NaQH is leached out with hot water at a tempereture of 70-8C° C. The leaching takes about a half hour. On cocling, a few flakes of Fe,0, settle out. The Fe,0, can be used over and over again. The process immediately yields a caustic solution of specific gravity 1.32 —~ 1.38. Lowig® later found that ' the process is practicable only with a very pure soda, the impuri— ties preventing an intimate contact of the Na,CO, and Fe,0,, and their fusion prevents the escape of CO,. It is necessary to re~ move the CO, or the reectior would soon te stopped by the atmos— phere of the pure CO,. Precipitated Fe,0, is not suitable for the process, since it yields e ferrite difficult to extract, on account of the oxide being separated as a fine mud. (1888) 122. *, See élsc: dour. Soc. Chem. Ind., (1883) 280. jw [ni Jour. Soc. Chem. Ind., ‘The article on British Patent No. 364° stetes that soluble im purities removed by the cold wash are "Na,S0,, NaCl, Na,cO,, etc." | It was decided to try the Lowig process for several reasons. .The Lowig process yields a solution of specific gravity 1.52 -— 1.38; the lime process a specific gravity of 1.075 — 1.100. A much purer - product is secured by the Lowig process for the comucn impurities in NeOH made by the lime process (NaCl, Na,S0,,5 etc.) are removed by the cold washing. The main disadvantage with the Lowig process is, appar—- ently, that of the heat required. iil Experimental A qualitative analysis” of an average sample of trona showed: Nat large arount Kx? none Catt small anount co, large amount S0Q,~ about 1% (estimated as SC,>) oa about 2% (estimated as C17) B,07 possible trace There was a relatively smell amount of dirt, insoluble in acid. In all succeeding work the trone was dried over conc. H,S0,. Sulfate was cetermined gravimetrically. A known weight cf? trona was dissolved in water, acidified with conc. HCl, heated to destroy HCO,” and COs, filtered to remove insolubles, and finally the sulfate was calculated as so,~ and results were 1.98% and 1.92%. 5 Jour. Soc. Chem. Ind. 6 (1887) 438. S Wethod cf A. A. Noyes, Gualitative Analysis, 8°" edition. ‘The insoluble material in the sulfate determination was filter— ed thra previcusly dried and weighed filters. The filter papers and residues were dried and weighed. The insoluble matter thus found was 1.36% and 0.53%. Na,CO, and NeiCO, were determined by the following procedures. . Tote] alkali wae determined by adding an excess of standard HCl to a solution of the sample, boiling to drive off CO, and then titrat- ing the cold solution with standard bese for excess acid, using methyl crange as an indicator’. To determine NaiCC, an excess of standard base was added to a soluticy of the sample. The carbonate was then precipitated with BaCl, and filtered off. The excess base was then determined using standard HC] with ehenolphthalein as indi— cator. This method was net entirely satisfactory as « few very fine particles of EaCO, would come thru the filter and were net removed by repeated filtration. Boiling before filtretior did not help. However, the same amount of these particles seemed ta be present after the end—-paint was reached as when the titration with the ecid was started. Na,C0, is calculated by difference betweer total al- xali and NaHCO,. 7 consists of adding The determinatian of CI~ volumetrically ferric ennonius sulfate to the solution and a few drops of standard KCNS te impart a red coloration. Standard AgNO, is added until the solutior is decclorized and then a few additional drops are added. The AgCl is filtered off and the filtrate and washings are titrated with KCNS for excess AgNO,. Severel analyses gave widely divergent ? Methecd from Solvay Bulletin No. 3. results. It was then found that the colcrsatior dismppesred before all of the chloride had been precipitated. This method was there fore rejected because of the difficulty of determining whether or not all the Cl~ had been precipitated without adding a large ex— cess of AgNO,. It was thought that K,Cr0, would be a better indi- cator. The first run was made using 19 grame trona and 27 grams com mercial Fe,0,. The mixture was placed im an iron crucible and heat— ed with Tirril burners fer 25 minutes. Judging by the appearence there seemed to have been a small conversion to sodium ferrite, leaving much uncenverted carbonate. The mass was heated at a high— er tenperature for en additional hour, at the end of which the coaversior appeared much larger. The mass was then removed fron: the crucible (it was found necessary to use a hammer end steel spatula to lcosen the materia]), pulverized and placed on an as— bestoe suction filter and washed with 30 cc. cold water. Hot water (40 cc.) was then passed thru the filter with apparently no effect. The mase was transferred from the filter to a beaker, the hot filtrete added and heated for 25 minutes at a temperature of 70-95° C. and then refiltered thru asbestos. The filtrate cole tained a smal] amount of Fe,0,. Repeated filtering thru quantita tive filter papers did not entirely remove it. The filtrete was diluted to 100 cc. and portions were analyzed for total alkali and for NaCH, Na,C0O, being obtained by the difference. Total alkali was determined by titrating with standard HCl] using methy] crenge as indi- cator. NaOH was determined by precipitating the carbonate, filtering off, and titrating for NaOH using standard HCl with phenol phthalein as indicator.® The analysis showed 5.6% Na,CO, and 0.98% NaOH. The re- sult for Na,CO, was calculated as a percentage of the Na,CO, which would be present if all the sodium were in that form. The result for “NadH was calculated in a similar manner. In ths anelysis given above this means that 5.6% of the total sodium was present in the hot leach as Na,C0,, ond 0.98% was present as NaOH. Thie left 93.38% (100 — 5.6 — 0.99) of the total sodium unaccounted for. It may have been in the cold wash or it may have remained with the Fe,0, residue. All the results in the following enalyses srs given on this basis. It might be said here that, in this and all succeeding runs (unless noted), a clear filtrate was obteined when the precipitated Bacto, was filterei off for the NaOH analysis. Qualitative tests showed the presence of apprecisble emounte of C1™ and SO, The second run was made using larger quantities of trona and commercial Fe,0, ~ 30 grame trona, 33 grams Fe,0, (10% excess over theoretical). The mixture was placed in a crucible and hested as in the previous run for 45 minutes. There seemed to he but little conversion snd practically no tendency to fuse. The second heat— ing of the mess was for one hour. This time » Meeker burner was used and the crucible and burner were surrounded by esbestos form ing an air beth which geva a much higher temperature. Judging by appearance, a fair conversion was obtained -— also there seemed to Method from Solvay Bulletin No. 2. ‘have beon some fusion curing the second hesting. The mass was trans— ferved Lo a beaker, 55 oa. of cold water was added and ellowe4 to stand over night. After filtering thru asbestos the filtrete had a litile Fe,0, which was filtered out thru a filter paper, the re- sulting filtrate being clear and having a yellow color. On stand— ing for some time & precipitate aettled out which on filtering was Pound to be light yellow in color. The filtrete was still yellow. Analysie showed the solution to heave 7.7% NeOH and 23.0% Na,CO,. Hot water wae passed thrn the filter. No eppreciable change vas seen in the emount of the precipitete but the water, on analysis, showed 0.07% NaOH and no Ne,CO,. After analyzing the cold wash wa— ter some of it was heated to boiling for several minutes. A light yellow crystalline precipitate settled out snd the yellow color remained in ths solution. The mixture, after the cold wash was filtered off, was traneferred to a beaker, 45 cc. of hot water was added sand kept at a Lempersture of 80-95° GC. for 40 minutes and wag then filtered. The filtrste was freed from Fe,0, by a second filtration (thru paper) and analyzed. The snalysis showed 3.2% NaOH and 3.3% Neco... At this point re?srences were found indiesating that the natur— al iron oxide should be used. However, the third and fourth runs were made with the commercial Fe,0,., to see. what effect, if any, strong heating of the Fe,0, before use would have on the results. The Fe,0, used in run No. 3 was strongly hestel in e muffle fur- nace for 35 minutes before use, wherens that used in run No. 4 was not heated. In these runs, and the succeeding ones, shallow iron eT pans; rather then iron crucibles, and 6 muffle furnace, hmvin euction draft to remove the gases formed, were used. In other re— epects, these runs were similer to No. 2. Analysis for total al- kali in the hot Leachgs showed 3.6% in No. 3 and 7.7% in Yo. 4. While waiting for some red haematite to arrive & couple of rans were made using heated Limonite (2Fe,0,-3H,0).- The limonite was crushed snd heated in a muffle furnace to a bright red for 1 hours. The heating change’ the limonite from a yellowish brown color to » dark red. Thirteen grams of trona were used with 15 grams of the heated linonite, corresponding to 10% excess F2,0,, on the as— sumption that 211 of the water had been driven off from the limonite. a] @ he mixture was hented at a low red for 45 minutes, was pulverized, and heated for 40 minutes ot a high red heat. Cold water {45 cc.) was added to the mixture ond it was allowed to stand over night. The cold wash water was filtered off thru asbestos and then fil— tered thru paper to remove a little Pe,0,. The filtrate came thru clear and was yellow in color, but on standing became cloudy. A second filtration cleared the solution but it again became cloudy moved the cloudinesn and left a2 clear yellow solution. A little KCNS was added giving :. deep red coloration. The yellow residues on the ?ilter paper were apparently insoluble in hot water but dissolved renudily in dilute HC] end gave a deep red coloration with CNS. The yellow coloration was eppsrently due to colloidal iron. .The analysis of the cold wash water was net made. The hot leach showec, on analysic, 10.8% NaOH and 26.5% Ne,CO,. The hot leach solution wes vary slightly yellow and gave a test for iron. The sixth run wes similar to the fifth, reutilizing the Fe,0, from the fifth ran. After adding cold H,0, the mixture stood for several days. Analysis showed (in percent): Wadd Na,o0,, Total alkali cold hot, total cold hot total wash leach wash Teach 19.6 19.0 38.6 28.1 15.1 43.2 81.8 The cold wash was yellow in color and it was found thet by boil~ ing for & to 10 minutes, cooling and ?iltering, a clear, color— less solution could be obtained, which cid not cloud up on stand. The seventh and eighth runs were made at the same time and similar to the previous run. Red haemetite (Fe,0,) was used in both runes. Run No. 7 used 10 grams 0.P. Na,CO, with 17 grams of hasmatite vhile run No. 8 used 14.5 grams trona (Na,0 content was the ssme as in the Na.,00, ) and 17 grams haematite. A second hot leach of 2 hours duration wes made to gee if it would account Por any mors alcalis This second hot leach (and ell subsequent leachinge and washings) were made in e beaker, useing en invarted 't' stirrer driven by a small motor. This geve continuel stir- ting instaed of oceasional stirring by hand as in previous rune. ig iy a 10 The results of the analysis were: NaOH cold - Ist hot 2nd hot total wash leach leach No. 7 C.P. Na,G0, 2.4 37.5 0.0% 39.9 ‘No. 8 trons 0.3 33.0 0.7 34.5 Na,CO, Total Alkali cold Ist hot 2nd hot totel waeh leach leach No. 7 Na,c0, 7.0 17.5 15.7" 40.2 80.1 No. 8 trona 4.7 15.2 10.3 30.2 64.7 * Probable loss dus to spattering. The ninth «nd tenth runs wers made together — No. 9 using the Fe,0, from No. 7 with C.P. Na,cO, and No. 10 using the Fe,0, from No. 8 with trona. The time for the cold wash was reduced to § hour and the hot leach was lengthened from 4 hour to 2 hours. Except for the conditions noted, the runs wer? made the sams @s the previous rang. The analyses on the rune were spoiled so that, no results were obtained. Runs Yo. ll snd No. 12 were made together and were made the seme as No. 9 snd No. 10. Run No. 11 used OC. P. Na,CO, with the Fe,0, from No. 10. Poth the cold wash and hot leach waters had to be boiled to ?rse them from iron. Anelyses for the rune showed: ll NaOH cold hot total wash leach No. 11 G.P. Na,t0,* 0.5 20.4 20.7 No. 12 trona 15.4 41.7 55.1 Na,Co, Total Alkali cold hot, total waah leach No. 11 G. P. Na,c0,* 4.8 8.1 12.9 33.6 No. 12 trona 18.0 10.6 28.6 85.7 x very small loss due to breaking bottom of iron pan while transfercing to bexker after heating. The solutions ?rom runs No. 7 to No. 12 inclusive were tested quelitetively for chloride and sulfate. The Na,CO, runs showed no sulfate and only a trace of chloride, due to the reagents or water. With the trona the hot leaches appeared to have about 0.1% each of chloride and sulfate (estimated as C1” and 50, ) while the. sold wash weters in general showed from two to four times as much chloride and sulfste as the hot lesaches. Theres still seemed to be & considerable loss of alxali and an appreciable amount of the NeOH is obtained in the cold wash. It was thought thet when working with large quantities £t would be possible to cut down on the amount of cold wash water usec with the result thet the amount of NaOH removed by the cold wash might be decressed. Runs No. 13 and No. 14 were made together, run No. 13 using three times as muck C.P. Na,CO, as in No. 11 and run No. 14 using three times as much trona as in No. 12. The excess of Fe,0, over theorstics] reguir- 12 ed was inereaged from 10% to 20%. Tron crucibles were used instead of shallow iron pans. Fifty cc. of water (same as for previoue runs) wes used for the cold weshe For the hot. leach the water used was increased from 30 ec. to 70 cc. It was necessary to heat both cold washes and hot leaches to free them from iron. Analyses showed: Nadu cold hot total wash leach No. 18 0.P. Na,co, 1.0 7.5 8.5 No. 14 trons 5.7 15.4 19.1 Na,00, Totel Alkali cold hot wash leach total No. 15 C.P. Na,CO, 22.4 11.2 35.6 42.1 No. 14 trona 32.4 16.7 49.1 68.2 Qualitative tests on No. 14 showed considerable chloride end sul-- fate in the hot leach, but ebout one-quarter to one-half ee mach se there was in the cold wash. Runs No. 15 and No. 16 were made the seme as No. 14. Both runs used about 20 grams trona (No. 14 used 43 grams trona). Run No. 15 was heated for 2 hours, No. 16 wes heated for 3 hours, whereas No. 14 was heated for 1 hour 26 minutes. Shallow iron pans were used in these runs. Considerable trouble was caused by the stirrers sticking while making the hot lesaches so thet the water was in contact with the mixture much longer than 2 hours. However, the water was kept at 80-95° C. for only two hours. The cold wash waters were boiled several times, sbout 10 minutes each time, but thet. Prom No. 16 did not lose the yellow color cue to the colloidel iron. At the end of 39 minutes (the total time of the boiling) the cold washes were Piltered and the filtrate from No. 15 was then analyzed. ~The filtrate from No. 16 (still yellow in color) was boiled for an additional 9 minutes, cooled and ?iltered. The filtrate was still yellow. Previously 10 minutes was the mexinum time required to Pree from the colloidel iron. Analysis of this solution was then made, after precipiteting the esrbonate (as BaCO,) no trace of the yellow coler vould by seen in either the filtrate or the precipitate. Ana- lyses of these two runs showed: Nadu cold Ist hot 2nd hot wash leech leach total No. 18 0.09 2262 0.00 220d No. 15 0.00 14.2 0.00 14.2 No. 14% 3.7 16.4 _- 19.1 Neco. cold Ist hot 2nd hot , waeh each leach total No. 15 8.5 9.6 19.5 37.6 No. 18 7. =14.9 9.5 31.4 No. 14% 32.4 16.7 -_ 49.1 *Run No. 14 is given hers for convenience. Total Alkali 14 In these tivo rane the Bac0, precipitate came repeatedly thra the fil~ ter 2s mentioned under the analysis of the trone for Na,CO, and Nalco, A qualitative analysis showed for both No. 15 and Ne. 16, thet practical— ly all of the chloride was removed by the cold wash. In run No. 15, the cold wash remove! from 2 to 4 times as much aulfste as was remov— ed.by the first hot leach, while in run No. 16, the cold wash remov- ed from four to five times as much sulfate xs the firss hot leach. The second hot leach of both No. 15 and No. 16 showed a trace of eul~ Pale. IV Summary * The Lowig process has been studied with reference to its adapt— ation for tise of trona as a raw material. The following conclusions haves beon reacheds 1. Commercial Fe,0,, which is obtained by calcining precipi-~ tated Fe(OH), is unsatisfactory for this process. <. Limonits, heated before use, and rad haematite give about the same results. 5. The cold washing, for the purpose of removing impurities, 4. The hot lesching for the recovery of the NaOH would be satisfactory if the cold washing ramoved the chloride, sulfate, and carbonate. 5- Colloids] iron mey be separeted by boiling Prom 5 to 10 minutes, cooling, and filtering. 15 6. Such impurities as are Pound in the trons. do not hinder the process as shown by comperative sune with C.P. Na,C0,. 7. The conversion his been unsetisfectory, renging from 19.1% to 88.1% in the experiments using trona and red haematite. 8. The process is impracticsble without further work. If the gonversion enc cold washing can be improved the process would prob~ ably prove to be practicrble.