THE CONSTRUCTION AND OPERATION OP A PILOT PLANT POR INULIN EXTRACTION A Thesis Presented to the Faculty of the Department of Chemistry University of Southern California In Partial Palfillment of the Requirements for the Degree Doctor of Philosophy by Kenton James Leeg June 1940 UMI Number: DP21728 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. Dissertation Publishing UMI DP21728 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 T h is d is s e rta tio n , w r it t e n by ......... I ^ U T O ^ ............ u n d e r the guidance o f h h L - F a c u lty C o m m itte e on S tudies, a n d ap p ro ve d by a ll its m em bers, has been presented to a n d accepted by the C o u n c il on G ra d u a te S tu d y a n d Research, in p a r tia l f u l fillm e n t o f re q u ire m e n ts fo r the degree o f D O C T O R O F P H IL O S O P H Y S'F" /y V -V •/ f.. ; - - - - - ............ D ean S ecretary 40...... C o m m itte e on Studies ACKHOWLEDGMEim The author wishes to express his appreciation to Dr. LeBoy 3, Weatherby, who first suggested this problem, and to Dr. B* E. Vivian, fheir kindly interest and timely criti cism in directing this problem were a source of encouragement. fhanks are due Mrs. Dorris Bucher, Mr. Bill Coleman, and Mr. John P. Hanna for their indispensable assistance in construction and operation; to Mr. Paul Carnes, Mr. Michael Lowry, and Mr. Bay Van Best for technical advice and assist ance; to the Messrs. Pfluger of the Bainbow Dahlia Gardens, Mr. Armstrong of Gardena, and Mr. Andre Menou for supplying tuberous roots of dahlias for the experiments. TABLE OF COBTENTS CHAP TEH I* II. PAGE INTRODUCTION............................. 1 Statement of the pr o b l e m.................. 2 Discussion of the source • . . . ........... 2 REVIEW OP LITERATURE...................... 6 Physical constants and chemical nature of i n u l i n ............................. 6 Purification of inulin . . . .............. 7 Value of I n u l i n .......................... 8 Previous attempts on semieommercial prepar ation of i n u l i n ................. . • • • III. PRELIMINARY WORK 10 ........................ 15 * .................... 13 Inulin by modified sugar beet method . . . 13 Test of method of analysis.............. 14 Quantitative Benedict's 16 Methods of analysis ................ Brix spindle........................... Methods of extraction and c o n t r o l ........ 16 16 Tests to determine pH of the solution con taining roots with varying amounts of added calcium carbonate or sodium bicar bonate ............................... 16 Wash boiler m e t h o d ...................... 20 Small scale diffusion battery and results. 21 V CHAPTER PAGE Experiment # 2 ....... .. ............. 22 Experiment # 3 ................... . . 24 Experiment # 4 ....................... * 25 Experiment # 5 ........................ 26 Filtration methods andresults • • • . • • 29 Controlled pressure inulin filtration Experiment.......................... 29 Inulin filtration in small area filter press. Experiment #2 . ............... 30 Filtration of hot, crude inulin solution to remove filter cell and decolorizing charcoal........... .. ............. Experiment # 3 .......................... 30 Experiment # 4 .......................... 31 Experiment #5 .......... 31 .................... 32 CONSTRUCTION OF PILOT P L A N T .............. 33 Materials of construction• • • • .......... 33 The c e l l ................................. 36 Assembly................................. 47 Auxiliary a p p a r a t u s ........... 48 Experiment # 6 . . . IV. 30 Cutter Filter press ............................ 46 ............. 52 Air compressor......................... 52 vi CHAPTER PAGE Steam generator Drier ......... 53 . . . . . . . . . . . . . 54 Miscellaneous equipment.......... 54 P a i n t ............... V. 55 DIFFUSION STUDIES OH A PILOTPLANTSCALE Experiment #1 . . . . ......... . 57 . . . . 58 Experiment # 2 ...................... Experiment #3 60 ............. 62 Experiment #4 and 5 • ♦ ............ 65 Experiment # 6 ...................... 65 Extraction curve of cells insystem . . . . 68 Determination of hydrolysis ofcrude inulin Experiment # 7 ...................... Explanation of l o s s ............... VI. 70 71 . . . SUMMARY AND CONCLUSIONS............ BIBLIOGRAPHY................................. 74 77 82 LIST OF TABLES STABLE I* II* PAGE Aoid Hydrolysis of Boots • • * • • • • • • Determination of pH of Boot Suspension with Varying Amounts of Salts III* IV* V* VI* VII* VIII* IX* 15 * * * * * 17 Determination of Inulin Recovery with Varying Amounts of Salts • • • • • • • • 18 Degree of Extraction of Inulin • • • • • • 80 Data 84 Experiment #8 • • • • • • • • • * Data -- Experiment #3 • • • • • • • • • • 85 Data --- Experiment #4 * . • • • • * • • • 86 Data 87 Experiment #5 • * • • • • * * • • Composite Results Obtained with Small Scale Diffusion Battery • • • • • • • • 88 X* Filtrate Time vs* Pressure * • • • • • • • 31 XI* Filtrate Time vs* Pressure * • * • • • • * 38 XII* Composite Results of Filtration of Hot Inulin Solutions * • • • • • • • • • • * 33 XIII* Experiment #8 • • • • • • • • • • • • • * 63 XIV* Experiment #3 • • • • • ................ 64 XV* Experiment #4 • • • • • • • • • • • • * • 66 XVI* Experiment #5 • • • • • • • • • • • • • • 67 XVII* XVIII. Determination for Extraction Curve * * * * 69 Determination of Hydrolysis of Crude Inulin • • • • • • * * • • • • • • • * • 71 viii TABLE XIX. XX. PAGE Experiment #7 . . .. ........ . . . . 72 Composite Basalts of Experiments #3, 4, 5, 6, and 7 ......... 74 LIST OF FIGUBES FIGURE PAGE I* Taboroas Boots of the D a h l i a ............. 3 S. Small Scale Diffusion B a t t e r y ........... 23 3. Inner Cell (Upside d o w n ) ................. 37 4. Diagram of Cell (Side v i e w ) ............. 39 5* Diagram of Cell (End view) • •» ........... 40 6* Diagram of Cell (Top v i e w ) ............... 41 7* Diffusion Cell (Open)..................... 43 8. Diffusion Cell • • • ..................... 44 9♦ Complete Diffusion Cell ShowingInsulation Areas • • .............................. 45 10* Filter S a c k .......................... .. 46 11. Pilot Plant Diffusion BatteryOpened * .• . 49 13. Diagram Top View of DiffusionBattery . . . 50 13. Cutting Machine and Pilot PlantDiffusion Battery in Operation .................... 14. Mahogany Disc . * ................... . . 51 61 15. Extraction Curve ....................... 68a 16. Flow S h e e t ............................... 81 CHAPTER I INTRODUCTION Inulin is a polysaccharide composed of many levulose 18 units; the exact number of which is as yet undetermined* Until recently it has been a comparatively rare substance used only in the laboratory, and even there to a rather limited extent* Research as to its preparation has been in progress in several laboratories during the last few years* At the University of Southern California investigations in this field 23 have been in progress by Weatherby and co-workers since 1928* The chief value of inulin is its use in the preparation of the sugar, levulose* The production of levulose on a com mercial scale has been the dream of many chemists. Various methods have been attempted to produce levulose from its five important sources* These are the Jerusalem artichoke(Helian- thus tuberosus L * ). Chicory root(Ciohorium intybus L.). the Dahlia(Dahlia variabilis(Willd.) Desf., Sugar Beet(Beta vul garis D*), and Sugar Cane(Saccharum officinarum L.).^ How ever, in most of these attempts, the purification has been tried on levulose itself* Difficulties with crystallization have occurred in every oase* Where the raw material was su crose these difficulties have been surmounted, but the cost remains high* It has been the contention of Weatherby, in his researches, that a syrup of a high degree of purity would crystallize readily, and that the logical method of obtaining this syrup was to purify the inulin before hydrolysis* Statement of the problem* It has been the object of this study to develop a method of obtaining relatively pure Inulin, in large quantities, from dahlia tuberous roots and to build a pilot plant to determine a commercially feasible method of extraction* Di so us si on of the souroe * The dahlia is a genus of herbaceous plants of the family Oompositae* The genus con tains ten species indigenous to the high, sandy plains of Mexico* The species, Dahlia variabilis is the one from which the majority of the forms, now common, have originated* 2 The ordinary, natural height of the dahlia is about seven or eight feet, but one of the dwarf species grows to only eighteen in ches* As is the case with many other plants the dahlia will do relatively well with a minimum supply of moisture, but the plant does better as the water supply approaches the optimum* numerous varieties of dahlias are now raised for flow ers, examples of which are: cactus dahlias, show dahlias, and pompon dahlias* They are propagated from seed and by divid ing the large, tuberous roots; in so doing, care must be taken to leave an eye on each tuber* The best and most general mode Figure 1. tuberous roots of the Dahlia. JX&0B* 4 of propagation is by cuttings. Some of the varieties, the "Queen," produce large, succulent tuberous roots, These are most satisfactory for inulin extraction. Dahlias succeed best in an open situation, and in rich, deep loam, but there is scarcely any garden soil in which they will not thrive if it is properly fertilized. Dahlias, on a large scale, succeed well in most climates, especially if the 17 rainfall is moderately heavy. Investigations in this laboratory have shown the tuber ous roots of the dahlia to contain from 6*5% to 14% inulin by weight. At the time that sugar beets were first considered as a source of sugar, they did not yield over 8% sugar, The sucrose content has now been increased by selective breeding to from 16% to 22%. It may be assumed that similar increases may be obtained with dahlias. It has been estimated that a yield of from 400 to 460 bushels of roots per acre, or from 9 to 10 tons per acre, might be obtained, and that large varieties might yield 700 bushels per acre. A clump of dahlia roots weighing 70 pounds was raised by Father Sohroeder of Mission Santa Barbara. Close planting, in commercial dahlia raising for inulin ex traction, might bring the yield considerably above that of sugar beets which average about 11 short tons per acre. This 5 holds without the added advantage of selective breeding which would further increase the yield. CHAP (EBB II BE VIEW OP LlfEBAfUBE Physical constants and chemical nature of inulin. Inulin is a white powder resembling starch when pure. odorless and tasteless. It is It has two crystalline forms; i.e.t white, double refracting, sphero-^crystals and acicular crystale. 18 Pormerly it was thought to be made entirely of fruc tose units; the number of which was in dispute. Becently, it has been reported that a very pure inulin, on acid hydrolysis, yields 3% glucose and this has been suggested as a means of investigating the molecular weight.13 However, if the hydrol- ysis is achieved by enzymes, the per cent of glucose formed is less. 16 fhe molecular weight has been estimated by Haworth and by Pringsheim.^’*^ !2he evidence tends to indicate that there are probably between thirty and sixty units of fructose 8 16 per molecule. * As in the case of starch and proteins, this large molecular size leads to the formation of colloidal suspensions. Consequently, though the solubility of inulin in water varies from 0.014$ at 0° C. and 0.23% at 30° C. to 38.3% at 98 it is necessary to have a solution of about 8% inulin to insure precipitation at 20° C. before fermenta tion becomes appreciable. Inulin forms a colloidal suspension in water and precipitates very slowly if not present in a high concentration* This is especially trae with impure solutions in which there may be protective colloids. As a standard of purity, it has been suggested that: 1) the inulin be white, 2) should give less than 0.1$ ash, 3) should not reduce Pehling's solution, and 4) should display n 3 a specific rotation of approximately -38 . Air-dried inulin contains about 10$ water, but this is 18 bound water and cannot be removed without decomposition* Por a complete discussion of inulin and its physical and chem ical nature, see Pringsheim, "The Chemistry of the Saccharides” page 296, and Thorpe, Chemical Dictionary, Vol. II, page 53, 18.22 1921. * Purification of inulin. The first attempt to purify Inulin by any standard method was given by Kiliani in I860. His method involved the precipitation of the inulin from sol ution by freezing after the removal of albuminous substances with lead acetate. Essentially, this same method has been fol lowed in most of the later investigations. H. A. Spoehr has suggested a new method of purification which would not, however, be satisfactory for commercial practice. 21 In his meth od the juice extracted from dahlias is stored for five days at -20° €•, then thawed and filtered. The inulin is redis solved in hot water and stored again at -8° C. for several 8 days. The resultant product has an ash of 0.16J6 to 0 .2$> and a specific rotation of -35.5° to -38.1°# Other methods of purification involve precipitation 23 with alcohol or allowing the inulin solution to stand for a long period of time with a preservative added. In neither case could the method be applied commercially because of high operating costs or large losses of inulin. Purification of inulin has been developed to a high degree at the University of Southern California. Bartel pre- pared inulin by successive recrystallizations from hot water. 3 He was interested in obtaining inulin of the highest degree of purity# I'or this purpose he found it necessary to wash the very pure product with alcohol and finally with ether. To obtain inulin with a very satisfactory degree of purity without going through several recrystallizations, Brobst rec ommends purification by the addition of 1$ Norite and heating with five parts boiling water for one half hour, adding 1% diatomaceous filter aid and filtering hot, cooling, and allowg ing the inulin to precipitate. Pure inulin can be obtained from a commercial stock by this method without a repeated recrystallizati on•* Value of inulin. At the present time purified inulin has very few uses. The inulin of Jerusalem artichokes can be 9 assimilated fey the human body# However, the mechanism fey which it is hydrolyzed in the feody has not feeen definitely determined# It is apparently hydrolyzed fey feacteria in the intestines, fey a combination of enzymes and hydrochloric acid in the stomach, or fey both. Some animals may possess inulase enzymes# It has feeen suggested fey A . Goudberg that inulin is 7 an excellent food for diabetic patients* He has shown that inulin is utilized over a long period of time in the feody# Shis allows the feody to oxidize a relatively large portion of the carbohydrate with a minimum of glycogen formation and a small increase in blood sugar concentration# Ho verification of or addition to this work has feeen done; further investi gations are advisable# A test has feeen made on the relative value of levulose and glucose in the diets of normal, growing rats* 'Pwo groups of rats from the same colony were fed the same basic diets with the same calorific values. One group received most of its carbohydrate in the form of levulose while the other group received an equal quantity of glucose# ence in weight gained was observed# Ho appreciable differ It was found in every case but one that the livers of the levulose-fed rats were considerably heavier than those of the glucose-fed rats.1 Work has feeen done on the utilization of levulose as compared to that of glucose and other sugars fey depancreatized 10 dogs. 5 It has been found that a diet in which the carbohydrate used is levulose is utilized to a marked degree as compared to a similar diet but with glucose as the carbohydrate. This difference is pronounced at first but becomes successively less each day until at the end of thirty days there is no noticeable difference. After a period of time, during which levulose is not fed, the animals' ability to assimilate this 5 sugar again becomes appreciably greater. The utilization of levulose, prepared from inulin, by depancreatized animals is being investigated in this laboratory 20 by other workers. Previous attempts on semicommercial preparation of in ulin. In 1934 inulin was prepared on a large scale in this 19 laboratory by Bieger. fhe method developed by him was model ed after that of the cane sugar industry. Since inulin is not soluble in cold water, the roots were placed in a pressure cooker and heated to a pressure of about 15 pounds steam at which point the pressure was suddenly released and the soft ened pulp was pressed in a hydraulic press. The expressed juice was collected and the pulp was saturated with water aid put back through the process again. It was held that the superheating made the inulin soluble in the cell sap and the sudden release of pressure ruptured the cell walls releasing the inulin in solution. The inulin solution thus obtained 11 was purified, cooled, and set aside to crystallize, fhis method was tested using steam pressures of 5, 10, 15, and 20 pounds. Fifteen to 20 pounds gave the most complete removal of inulin. In these cases the extraction was in the neigh borhood of 97%. Duplications of these experiments were obtained by 9 Bolzman in 1938. He then modified the method by shredding the roots, boiling the pulp in boilers, and removing the juice by pressing. Shis was followed by a second extraction with a smaller amount of water. Ho pressure cooking was used, but it was necessary to shred the roots, and in order to maintain a concentration of inulin favorable to precipitation it was necessary to return much of the water extracted back into the press again. By this method he was able to obtain extractions as high as 93%. Inulin has been produced on a large scale from Jerusalem artichokes at Iowa State College by McGlumphy, JSiohinger, Hixon, 15 and Buchanan. However, in their method the inulin was not purified, but the extracted juice was immediately hydrolyzed to levulose which was precipitated and purified as calcium levulate. fhe calcium was removed by the addition of COg, the syrup was concentrated, and the levulose crystallized. Levulose has been prepared by similar methods at the United States Bureau of Standards by Jackson, who obtained U.S. Pat- IE ent #2,007,971 in 1935 on a method for the purification and crystallization of levulose*^ The method of obtaining the levulose is not specified, but Jackson has prepared levulose from many sources* Most of his work has been on the prepar ation of levulose from Jerusalem artichokes, but experiments have been conducted on the preparation of levulose from dahlia roots and it has been suggested that a satisfactory product might be obtained from beet sugar molasses and from the hydrol yzed juice of the Chicory root* In obtaining levulose from dahlia roots he suggested that inulin might first be precipi tated and filtered and the rest of the levulose, as levulose and levulins, should be hydrolyzed and recovered as calcium levulate* He showed that the inulin could be obtained with a satisfactory degree of purity by filtering the hot inulin solution through kieselguhr.10 Ho indication was shown of this having been done beyond laboratory scale. CHAPTER III PRELIMINARY WORK I. METHODS OP ANALYSIS Inulin by modified sugar beet method* This procedure is designed to determine the total inulin plus levulose minus glucose in the ratio of the specific rotations of levulose to glucose; however, this is justified, due to the small per cent of levulose and of glucose* A sample of 18*8 grams of ground roots was placed in a 250 ml* Erlenmeyer flask; a homogeneous sample, including juice, was assumed to contain 13*8 ml. of water since the roots averaged 76% water. Eighty-five and two tenths ml* of water and 1 ml* concentrated hydrochloric acid were added to bring the total liquid volume to 100 ml* The hydrogen ion concentration was lower than would be expected due to buffers in the roots. About one fourth gram filter cell and about one half gram decolorizing charcoal were added; this was heated in a water bath with frequent shakings for fortyfive minutes, using an air reflux. The first few milliliters and the residue were filtered and discarded* The filtrate was polarised with a sacoharimeter using sodium light. If a two decimeter tube is used, the reading is directly in per cent* Direct percentage readings are possible because the saocharimeter is built to read directly in per cent when the 14 correct amount of sample is taken* It is designed primarily for calculations of sucrose solutions* If £6*048 grams of pure sucrose is dissolved in enough water to make 100 ml* of solution, the sacoharimeter is designed to read 100% sugar if the reading is made through a two decimeter tube* If a sample of solution weighing £6*048 grams is diluted to 100 ml* and the reading is found to be 45%, then that sample contains 45% 14 suorose. !Zhis is an arbitrary scale, but it is convenient for plant operation. Since levulose has a specific rotation of -9£° and su crose a specific rotation of +66*4°, it would take less lev ulose to read 100% than sucrose* fhe levulose would, of course, read to the left instead of to the right* £6.048 x 466*4 -9£ equals 18*8 grams of levulose in enough water to make 100 ml* would read 100% if read through a two decimeter tube* fhen a sample of levulose solution weighing 18*8 grams in enough water to make 100 ml* will read directly in per cent levulose* fest of method of analysis* In order to determine whether the acid concentration was high enough to insure com plete hydrolysis of the samples, six duplicate samples were hydrolysed with varying amounts of acid. Each sample was hy drolyzed for one hour in a boiling water bath, fhe pH of the solution was determined using a Beckman pH meter immediately after mixing and again after hydrolysis. Each sample contained 15 18.8 grams of roots with water and acid to give 100 ml, of solution* fhe acid concentration was calculated on the basis of acid added* Ihe pH was then determined because it was to be expected that the plant salts would buffer the solutions* fABLE I ACID HYDBOLYSIS OF W O T S Sample # 1' ' Acid ' * Cone. Mix. ' pH after mixing jfinuiin pH after * hydrolysis * read 1 0*01 U 3.22 3.80 5.9# 2 0.01 H 3.39 3.74 5.8# 3 0*03 N 2.25 2.37 12.8# 4 0.03 H 2.25 2.37 13.0# 5 0.10 U 1.45 1.36 12.9# 6 0.10 JJ 1.40 1.35 13.4# The rotations of the sugar solutions were read at be tween 18° and 19° C* !Ehls chart indicates that hydrolysis is oomplete in one hour with a pH of less than 2*25 or with a calculated acid concentration of 0*05 N, but is not complete with a pH of 3.74. Since the rate of hydrolysis varies direct ly with the acid concentration, the samples are probably com pletely hydrolysed in less than thirty minutes* determinations have been made to verify this* Satisfactory therefore, forty minutes hydrolysis, under the conditions where the acid 16 concentration Is mixed to 0,10 U, is entirely adequate. Quantitative Benedict1s. glucose determinations. This is the standard used in Since levolose reduces only 94JJ as much copper Benedict's per gram as glucose, this correction must he used. Brix spindle. This instrument is essentially a hydro meter which is graduated in brix or total per cent solids. It is one of the instruments of control used in a beet sugar plant for rough, rapid calculation of the concentration of solids in sugar solutions. It was necessary, in the work on inulin, to have some means of determining the concentration of the liquors obtained during the preparation of inulin in order that any variation in control might be made immediately. It was not expected that the Brix reading would give an accu rate estimate of the inulin concentration, but was used for rough control. II. METHODS OF EXTRACTION AND CONTROD Tests to determine pH of the solution containing roots with varying amounts of added calcium carbonate or sodium bi carbonate. Since it has been shown that inulin decreases in stability with the increase in hydrogen ion concentration, it was desirable to know the pH resulting from adding differ ent amounts of neutralizing salts in order that a control of 17 pH could be maintained* A pH between seven and eight was assumed to be best; however, it might be possible that a slightly higher pH would be more satisfactory* In order to verify the former the following determinations were made* Nine 50-gram samples of ground roots were each diluted with 100 ml* water* To the separate samples were added the follow ing amounts of calcium carbonate and sodium bicarbonate and the pH checked in each case: TABLE II DET1BMINATI0N OF pH OF BOOT SUSPENSION WITH VABYINGr AMOUNTS OF SAITS amount of salt 1. 2• 3* A. 5. 6. 7. 8. 9* 0*5 gm. CaCOglppted) 1*0 gm* i» 2.0 gm* «t 4*0 gm. 0*5 gm* HaHCOg n 1.0 gm* i t 2*0 gm. n 4*0 gm* no salt added pH oi solution after boiling 2 hours pH pH pH PH pH pH PH pH pH 7.03 7.72 7.38 7.42 8.61 8.56 8*57 8.78 6.88 In each case the inulin from the samples in which sodium bi carbonate was used contained much more coloring matter* An attempt was made to determine the relative amounts of levulose to inulin in these determinations, but no satisfactory method was found* Accordingly, another set of determinations was made in which the inulin was separated and weighed in each case* Again 50 grams of roots were dilated with 100 ml* of water per sample; the previous pH investigations indicated that lower amounts of salt were satisfactory* TABLE III DETERMINATION OF INULIH RECOVERY WITH VARYIHGr AMOOTTS OE SALTS grams of inulin obtained from 10 ml* of extract amount of salt 1. s* 3* 4* 5. 6. 7. 8* 0*£55 gm* 0*340 gm* 0*360 gm* 0*355 gm* 0*350 gm. 0*£85 gm* sample discarded* 0*330 gm* 0*25 gm* CaC0*(ppted) 0*50 gm* w 1*00 gm* w £•00 gm* " 0.25 gm* HaHCOg 0*50 gm* " 1*00 gm* * no salt added *Sample spilled* In every case the filtered mother liquor had a light tan color in each of the samples in which calcium carbonate was used and in the blank; those in which sodium bicarbonate was used were much darker* The higher pH of the sodium bi carbonate solution apparently favors the extraction of the coloring matter from the roots* The weights obtained in the chart were those of inulin precipitated from 10 ml* of the liquid by the addition of 90 ml* of absolute alcohol, filter ing the inulin, drying, and weighing it* In every case where 19 calcium carbonate was used the alcohol brought down about 10 ml* of flocculent, inulin precipitate; about 15 ml. were ob tained in sample #6, 7 ml. in sample #5, and about 2.5 ml. of a granular precipitate in sample #6. The granular precipitate from the sodium bicarbonate solution was more easily filtered than the flocculent precipitate from the calcium carbonate solutions* The inulin obtained from the sodium bicarbonate solutions had a distinctly disagreeable flavor while the cal cium carbonate product was tasteless. able sweetness. Neither had any notice The sodium bicarbonate inulin would be harder to decolorize than that obtained using calcium carbonate. However, due to the ease of filtration of the sodium bicarbon ate inulin, it might be advisable to make added determinations of the relative merits of these two neutralizers. It has been assumed in some cases that the volume of precipitate obtained under the conditions of alcohol precipi tation gives an approximate idea of the inulin content of the solution. It seemed desirable that this assumption should be checked, and that results obtained in Table III should be put on a quantitative basis. J'or these reasons precipitates were filtered, dried, and weighed. A comparison of the results obtained by the two methods showed that the volume of the pre cipitate does not necessarily vary as the concentration of the inulin. Apparently the precipitate is more voluminous when the solution has a low pH than when it is higher* There fore, it is advisable to weigh all precipitates in calculations of this kind. Wash boiler method* This method was attempted to de termine whether it would be feasible to use a small number of cells in a series with final extraction of inulin being ob tained by pressing the pulp. were ground in a meat grinder. About 30 to 40 pounds of roots To half of the roots was add ed 8 pounds of water; this was boiled for forty-five minutes. The liquid was then poured off and the pulp pressed. The ex tract, as well as the extract from the pressed pulp, was pour ed upon the remaining roots. This was cooked again for forty- five minutes; the extract was again poured off and the pulp pressed. The resultant liquor was clarified with filter-cell and charcoal. The filter-cell and charcoal were removed by filtration through a centrifuge which was lined with paper, Filtration was poor, due to gumming of the filter paper. Two gallons of juice were recovered; the inulin and levulose con tent of which was 12,6$, as determined by the saccharimeter. TABLE IV DEGREE OF EXTRACTION OF INULIN stage 1, first pressed pulp 8, second pressed pulp 3, original roots % inulin content tM £1 fhe results of this experiment indicate that the method might he possible if four or five cells were used, but the labor costs of several pressings of the roots would be prohibitive. However, for laboratory scale production of inulin where high recovery is not essential, the method is simple and adequate. Small scale diffusion battery and results, fhe diffus ion battery is one of the older methods of extracting soluble materials from solids. When the problem is to remove a high per cent of the soluble materials without obtaining a dilute solution, it is one of the most obvious methods to be used. It is used with notable success in the extraction of sucrose from sugar beets, fhese extractions have been carried out on the counter-current principle, that is, the solvent is fed in at one end and the raw material is fed in at the other. Know ledge of these principles led to preliminary extractions of this type by Black, Hieger, and Holzman at the University of A TQ Q Southern California. * * 2Che largest battery built by these men was that of Holzman which consisted of six cells of about 500 ml. each. It was his suggestion that this work be carried further. A battery of eight cells was built using glass bottles of £50 ml. each and connecting them with glass tubing. (Figure £) In each case the entering liquid came through a short tube while the outgoing liquid entered a long tube at the bot- zz tom of the cell as in a trap* tate rapid changing. Extra cells were made to facili The whole battery was then immersed in a boiling water bath which kept the contents of the cells be tween 90° and 100° C. A trial extraction was attempted using the eight cells with a preheater and pressare supplied by the water main* Difficulties were encountered with the solids plugging up the tubes* This would build up a back pressure until the stoppers blew out* In order to obtain results the stoppers were all removed and the liquor poured from bottle to bottle at regular intervals. The data would be affected little by this as the increment of difference of concentration was still fairly large. The liquor obtained was difficult to filter and large quantities of filter-cell and high suction had to be used to obtain a product, The solution was decolor ized with charcoal and filtered while hot. precipitated by adding alcohol to about 35$. grams of good inulin were obtained. The inulin was Seventy-nine She experiment was of no significance quantitatively and was of value only in pointing out the difficulties. Experiment #8. In this experiment eight bottles again were used in series, with the inlet at the bottom and the out let at the top. A pad of glass wool was tied over the opening to keep the roots out of the tubes. Ten changes were made to insure equilibrium in the system before starting the experiment. FIGURE SMALL SCALE DIFFUSION BATTERY This was necessary because fresh roots were used throughout the system while consecutively weaker roots were in the cells when the system was in equilibrium* Nine more changes were made with BOO grams of dahlias and about 2 grams of calcium carbonate in each bottle; the liquor obtained was decolorized by animal charcoal and filter-cell* and filtered hot. The filtrate was allowed to cool and the inulin was allowed to precipitate during thirty-six hours, after which the inulin was filtered and dried. inulin were obtained. Seventy-nine and one-half grams of If it is assumed that the dahlia con tained 11$ inulin. which is a high value, a little over 40$ recovery of inulin was obtained. TABLE V DATA — 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. EXPERIMENT #2 Cells in system 8 Charge of nahlias per c e l l ......... . BOO gm. Total Dahlias used ................ 1,800 gm. Time of Draw • • • • . .................. 15 min. Number of Draws per experiment. . . . . . 9 Average volume of Draw ............ 150 ml. Total inulin in Dahlias ofexperiment . . 180 - 200 gm Temperature in c e l l s .................. 90° - 91° C. Dry inulin recovered.................... Recovery................................ 39.55® - Experiment #3. This experiment was made using the same principles as used in experiment #2. Eight bottles were 25 run through the battery to establish equilibrium; then* an experiment of eight changes was made, and the liquor was clar ified and filtered hot as before* The inulin solution was evaporated to one third the volume and the Inulin was allowed to precipitate, filtered, and dried in the drier to be describ ed in Ohapter IT* Hepresentative samples of roots from the experiment were polarized and read 10*4$ and 10*9$ inulin. TABLE VI DATA — 1* 8* 3* 4* 5* 6* 7* 8* 9* 10. 11* 12* 13* EXPERIMENT #3 8 Cells in system* ........................ 200 gm* Charge of Dahlias per c e l l ........ Total Dahlias u s e d ............... * • • 1,600 gm* Time of D r a w ............ * 15 min* Number of Draws per experiment.......... 8 150 ml* Average volume of D r a w ............ Temperatnre in cells * * * * ............ 90 - 95 C# Average inulin content of Dahlias * . ♦ • 10.7$ Inulin in extracted roots (pulp) ........ 1.' Degree of extraction ................... Total inulin in Dahliasof experiment • • 171 gm Inulin recovered................... 87 gm Recovery • • ........................... Experiment #4* Due to pressure difficulties that had been encountered in the first three experiments, the pressure system was removed and pouring from flask to flask was sub&ti tuted* Eight Erlenmeyer flasks were used in the experiment* Again eight flasks were run through to bring the system to equilibrium* The liquor was decolorized and filtered; the 26 inulin was allowed to precipitate; it was then dried and weighed* TABLE VII DATA — 1. 2* 3. 4* 6. 6* 7. 8. 9* 10. 11* 12* 13. EXPERIMENT #4 Cells in system * « • ................ Charge of Dahlias per c e l l .......... Total Dahlias u s e d ........ • • • • • Time of D r a w ........................ Humber of Draws per experiment • • • • Average volume of D r a w ........ * • • Temperature in cells ................ Average inulin content of Dahlias * • • Inulin content of pulp................ Degree of extraction • • • .......... Total inulin in Dahlias .............. Dry inulin recovered ................ Becovery • * ........................ Experiment #5* m m t m m v m In order to check the diffusion in a longer system, a battery of sixteen bottles was built* These were again immersed in a boiling water bath and pressure from the water main was used* Sixteen cells were run through to bring the system to equilibrium and twenty cells were run through in an experiment* This experiment verifys the conten tion that a longer diffusion battery leads to higher extrac tion and higher recovery* If the pH of the solutions were kept lower by the addition of calcium carbonate, higher recov eries might be obtained* On this experiment very few diffi culties were encountered and praotically no solution was lost 27 due to leaks in the system* TABLE VIII BATA — - EXPERIMENT #5 1. Cells in s y s t e m ............ ........... . • • Total Dahlias u s e d .......... • • • • • Time of Draw .......................... Number of Draws per experiment ........ Average volume of D r a w ................ Average inulin content of Dahlias • • . • Inulin content of pulp ................ Degree of extraction .................. pH of liquor • • • • .......... • • • • Brix of liquor at 20 C.. . . . . . . . • Temperature in cells • • • • • • • • • • Inulin in liquor from spent roots • • • • Total Inulin in Dahlias of experiment • • Dry inulin recovered • • • • • • • • • • Recovery • • • • • • • • • • • • • . • • 2 . Average charge of Dahlias per cell 3. 4. 5. 6. 7. 8. 9. 10. ill. 12. 13. 14. 15. 16. 16 202 gm. 4,040 gn. 30 min. 20 125 ml. 10# 0.8# 92# 5.45 14.2 90° - 95° C. 0.0# 404 gm. 255 gm. 63# The results obtained in these diffusion attempts show that it was possible to obtain a liquor of high inulin concen tration by this method. It is also indicated that rather com plete extraction of the roots can be obtained at the same time, but that more than eight cells were necessary to achieve this result. The extraction attained in these experiments would indicate that a battery of twelve cells should be adequate; this would be a very convenient number, because this is the number of cells used by sugar beet plants in the extraction of sucrose, and it has been the purpose of Weatherby and co-work ers to adapt a beet sugar plant to an inulin extraction plant TABLE IX COMPOSITE RESULTS OBTAINED WITH SMALL SCALE DIFB’USION BATTERY _______ gxperiment____ _______ 1 2 3 4 5 SfillS_lD_SY£lSB__ ______________ __8_______ 3______ ___ 8________ 8_______ __1S__ Average charge of Dahlias 1^0-150 _ 202 _2§r_c§ll_in_grflms_ _ _ ___ 200 200 _ Total Dahlias used in grgms 1^800 ____________ _________ 15____ Draws per experiment Average volume of Draw in. ml. % Average inulin content of Dahlias Inulin content of gulp in Degree of extraction in # pH of liguor 1^600 1*160 4a040 _15___ __ 15_____- 3 0 ____ 9 8 8 20 150 150 100 125 11± 10.7 10.5 10 _1*8___ _3±5___ 83 66.7 0*8 __ 92_ 5*45 Brix of liguor .14*3___ o 90-100 90-91 Temperature in cells in 0. 90-95 90-95 90-95 Inulin in liquor from spent roots in § ^pO.O Total inulin in Dahlias sf_expsriiDSDl-in_grflms.-________ _______ 1S0=2QQ._ — 121__ __ 122___ — 404____ Pri-iBuliD-recQyersd-ln-graas____ .22_____ 22*5___ _82___ _26*5___ _255___ £9 in its off season. It was therefore decided to build a pilot plant extraction battery of twelve cells with a capacity in the neighborhood of 50 to 100 pounds of roots per hour. III. FILTRATION METHODS M B RESULTS Preliminary to building a pilot plant diffusion battery, it was considered necessary to obtain more data on the equip-* ment necessary for handling the large quantities of liquid which would be obtained. Since many difficulties had been en countered in the filtration of inulin solutions, particularly with reference to the removal of the decolorizing charcoal, arrangements were made with the Dicalite Company to test fil tration pressures on inulin solutions using different grades of Dicalite Filter Aid* A solution was made of two and one half pounds crude inulin in boiling water to mahe two gallons; this was allowed to stand for eighteen hours. This inulin suspension and hot inulin solutions were filtered to determine the most satisfactory filtration conditions. The results are discussed in the following paragraphs. Controlled pressure inulin filtration. Experiment #1. In eleven minutes £6,090 ml. of filtrate were obtained through a press of 60 square inches of filtering area at 5 pounds per square inch pressure. The filtrate contained 0.07$ solids, and the cake, which was about two thirds water, weighed 560 30 grams; the cake thickness in the filter press was one half inoh plus or minus one thirty-second. Filtration was at room temperature. Inulin filtration in small area filter press. Experi ment #g. She filtration area was 1.77 square inches; 134 ml. of filtrate were obtained in nine minutes. one half inch thick* She cake was about She pressure varied from 0 to 30 pounds; higher pressures did not appreciably increase the rate of fil tration. She cake was gummy which indicated that the use of filter aid would be necessary in the filtration of crude inulin. Filtration of hot, crude inulin solution to remove fil ter cell and decolorizing charcoal* A solution was made using 1.350 grams of crude inulin in two gallons of hot water; 135 grams of Dicalite Special Speed Flow and 13.5 grams of Norite were added. Shis solution was used in the following experi ments. Experiment #3* precoat of Speed Flow. trate were obtained. She solution was filtered through a In twenty-one minutes 975 ml. of fil She filtering area was 1.77 square inches and the pressure varied from 5 to 40 pounds* 31 TABLE X FILTRATION TIME VS. PRESSURE minutes pressure 3 6 9 IS 16 £1 filtrate in ml. 5 lbs. 10 " £0 * 30 11 40 * 40 " £70 ml. 430 " 600 * 737 " 836 " 975 " A five ounce sample of filtrate contained S3.5 grams of inulin. Experiment #4. The filter area was 60 square inches. The solution was filtered through a preooat of Speed Flow. In one and one half minutes 1,665 ml were obtained at a tern* perature of 91° C. under 10 pounds pressure. A six ounce sample of filtrate contained £5 grams of inulin; this inulin filtered slowly. Experiment #5. The solution was prefiltered through Special Speed Flow; then, 1$ Norite and 1$ Speed Flow were added. In fifteen minutes 870 ml. of filtrate were obtained at a temperature of 91° 0. inches. The filtering area was 1.77 square An eight ounce sample of filtrate contained £8 grams of inulin; the inulin filtered slowly. 32 TABLE XI PILTRATIOR T i m VS. PRESSURE minutes pressure filtrate in ml. 10 lbs. 20 »# 30 it 40 tt 40 n 3 6 9 12 15 Experiment # 6 . 110 ml. 280 n 495 n 725 tt 870 IT The solution was prefiltered through Special Speed Plow; 1$ Rorite and 1?S Special Speed Plow were added, and this solution was filtered using a press of 60 square inches filtering area. The filtering time was two min utes at a temperature of 90° C.; 1,300 ml. of filtrate were obtained. The maximum pressure was 10 pounds per square inch. Six and one half ounces of filtrate contained 37 grams of in ulin; this inulin filtered well. These filtration results all indicate that it is advisable, when filtering crude inulin, to add about Speed Plow. of filter aid of the type of Dicalite Special This filter aid is a calcined diatomaceous earth and is designed to give fast filtration. It is not necessary to prefilter the solutions before adding Rorite, but higher Rorite economy is obtained by this method. It is advisable, in using Rorite, to add it several minutes previous to the addition of the filter aid because the filter aid tends to 33 TABLE XII COMPOSITE RESULTS OF FILTRATION OF HOT INULIN SOLUTIONS # of experiment Filtering time in minutes Maximum pressure in pounds 3 4 5 6 21 1.5 15 2 40 10 40 10 975 1.165 870 1.300 Filtrate in ml. grams inulin/ 100 ml. solution Area of filter press in sauare inches 15.9 14.1 11.9 19.3 1.77 60 1.77 60 Type of preooat S.F.* S.F.* S.F.* S.F.* 1S.S.F.** S.S.F.** S.F.* S.S.F.** 1.0 0.0 1.0 1.0 92 91 91 90 adsorb the Norite and decrease its efficiency. About 1$ Type of filter aid % Norite Temperature °C. S.F.* S.S.F. Speed Flow ---Special Speed Flow Norite, based on the dry weight of inulin, is usually adequate to remove all coloring matter from the crude inulin from the diffusion battery. 4 CHAPTER IV CONSTRUCTION OF PILOT PLANT I. MATERIALS OF CONSTRUCTION It was desired to construct a battery having a capacity of 50 to 100 pounds of dahlia roots per hour* The presence of tannins in the dahlia roots makes the use of iron containers impractical due to the formation of inks. Tests were made on the solubility of copper and tin in solutions of the juice from the dahlias. After ten hours of boiling copper turnings with dahlia roots no test for copper in the solution, was ob tainable. Similar results were obtained when tin was subjected to this treatment. However, when dahlia roots were boiled with copper plus tin, a trace of tin was detected. Therefore, it was decided that containers constructed wholly of copper or tin would be satisfactory, but it would be advisable to avoid the use of both in contact with the roots. Lata were collected on the costs of construction of suitable containers of copper and/or copper alloys. To force the diffusion liquid through the battery, in a small scale operation it has been shown that pressures as high as 60 pounds might be necessary. Therefore, it was considered advisable to build a battery capable of withstanding pressures as high as 125 pounds. The cost of building such cells wholly of copper would have been prohib 35 itive* Building the cells of iron and electroplating them was suggested, hut this plan was discarded because of the pos sibility of small exposures which would interfeafc with the pur ity of the product* fhe use of ice-cream cans, which are of sufficiently heavy construction to withstand the pressure and are heavily tinned, was suggested* fhe use of these contain ers would require that the connecting tubes be made of block tin or of iron tinned on the inside* However, tinners will not tin the inside of small tubing because of the dangers of explosion* It was desirable to avoid the use of block tin tubing, if possible, because of the softness of the material* However, after extensive searching, no other satisfactory tubing was found, and it was decided to use a heavy, block tin tubing* Half-inch block tin tubing of a weight of 6 ounces per foot is capable of withstanding a pressure of 600 pounds; it is fairly soft, readily worked, and, with care, can be handled successfully* fhe next problem was the heating of the solutions, and steam coils were the obvious means of accomplishing this* Since the steam coils could be mounted on the outside of the containers and thus avoid contact with the roots, malleable copper tubing seemed the best material of construction* It was not possible to calculate the amount of contact surface necessary, but the estimates obtained from the Chemical Engi- 36 nee ring Department indicated that six or seven tarns of halfinoh copper tubing around each cell would probably be satis factory* I'rom calculations based upon the sige of the con- tainers, it was estimated that about sixteen feet of tubing would be necessary for each cell* The copper tubing was avail able in twenty-foot lengths which allowed extra tubing for con nections* It was estimated that a flat top, for the container, capable of withstanding pressures of 60 pounds, would have to be of one fourth to three eighths inch iron; this lid would also have to be tinned* To check this estimate a trial con tainer was built to meet these specifications. It was tested and found to have a factor of safety greater than two. In order to hold this lid in place, it was planned that eight strap iron lugs would be riveted to the can, and bolts insert ed in the lugs through corresponding holes in the lid* The lid then could be secured by wing-nuts. Estimates were obtained for the construction and tin ning of the lids, the building of the lugs, and the insertion of the tin tubing into the bottom of the container and into the lid* When this work was completed the steam colls were bent into shape around a circular block of wood one inch small er in diameter than the containers; these coils were then sprung upon the cans and soldered in place; the ends of the 37 Figure 3* Inner Cell (Upside down) c • Container LOT - Liguor Outlet Tube L - Luge SIS - Steam Inlet Tube SOS - Steam Outlet Tube ST - Steam Tubing 37 FIGUBB 3 38 coiled tubes were bent out perpendicular to the sides of the can by means of a bending tool(Figure 3)* II. THE CELL It was next necessary to determine a means of insulat ing the units. Many insulating materials were available, but the simplest one that seemed satisfactory was rock wool. The insulating material had to be held away from the steam coils lest it reduce conductivity of the heat. Cans having a suit able diameter to fit over the steam coils, were purchased. These were put over the coils and the whole supported in a box (Figure s 4, 5, & 6). The top of the outer can was flared and nailed to the top of the box(Figure 7). The space between the outer can and the walls of the box was then filled with rock wool. The thickness of the insulating material was one inch in the thinnest places and was packed to a density of IB pounds per cubic foot (Figures 4 & 5). The insulation box was built with a false bottom to allow for a fitting on the end of the tin outlet tube(Figure 4). This was to facilitate cleaning of the connection lines in ease any plugging occurred. Tin tubing was cut to length, filled with sand, and bent to shape to connect the outlet tube of one cell to the intake tube of the next. The connections in the tin tubing were made by means of flared fittings. heavily tinned. The couplings were made of brass, ft to ,0000 o •rt ;zz:z 's~2fs-sizjzzs:sy*\ LOl 40 S c a / e - . / ’= «3 ” End View FIGURE 5. *9 U A D U P**s ;d/oof m*u *1* L It i 4Z Figure 7* Diffusion Cell — Open FS - Filter Sack IS - Insulation Space Id - Lid LI3? - Liquor Inlet Tube L - Lug SCS - Steam Chamber Seal SGCC - Steam Coil Coyer Can SIT - Steam Inlet Tube SOT - Steam Outlet Tube 42 FIGURE 7 43 To facilitate rapid operation it was necessary to leave about two inches of the top of the can above the box(3?igure 8)* To Insulate this large area and the surface of the can tops a cover box was constructed; this had a doable top to allow for insolation(Figure 9)* So insulate the sides of the cover box, a coarse, matting material was obtained and tacked in place* A small space had been left between the insulation box top and the walls of the diffusion cell* So prevent the entrance of water into this space and the loss of heat, a mixture of cement and asbestos was pressed in place* well insulated, waterproof unit* Shis gave a compact, The boxes were varnished and painted with aluminum paint to protect the wood* It was necessary to have some type of removable con tainer within the cell to hold the roots, keep them from en tering and plugging the lines, and to make it possible to re move the spent roots without too much wasted labor. Many types of containers were considered and it was decided that a sack made of ooarse filter cloth would be satisfactory* fhese sacks were made with a flared top to fit over the top of the cell (figure 10}* Hight holes were made in the flared top to fit over the bolts that held the lids in place thus holding the sack firmly in place* It was found by a prelim inary test that this canvas material did not serve as an ade quate gasket to hold the pressure* Gaskets were made by cut- 44 figure 8* Diffusion Cell IS - Insulation Spaoe Li - Lid LIT - Liquor Inlet Tube L - Lug SCS - Steam Chamber Seal SCCC - Steam Coil Coyer Can SIT - Steam Inlet Tube SOT - Steam Outlet Tube m - Wing Wuts 44 FIGURE 8. 45 Figaro 9* Complete Diffusion Cell showing Insul ation areas DCC - Diffusion Cell Cover IS •» Insolation Space Li - Lid LIT w Liquor Inlet Tube L - Lug SOS mm Steam Chamber Seal SCCC - Steam Coil Cover Can SIT - Steam Inlet Tabe Steam Outlet Tube SOT WN • Wing UutB 45 FIGURE 9. 46 Figure 10* Filter Sack 47 ting rings of graphite-asbestos gasket material, These were about one thirty-second of an inch in thickness and were en tirely satisfactory. It was found, after the preliminary ex periment with this set up, that the filter cloth sacks were too finely woven; when the material clogged with fine solids, the back pressure developed to such an extent that operation difficulties occurred. A new set of sacks were made using the same pattern but of unbleached muslin to allow faster flow. These were quite satisfactory and were used for the remainder of the experiments. III. ASSEMBLY The steam from the steam boiler was connected to the steam main and the steam inlet lines were run off from this line in parallel. The drips lines were connected to the drips main in a similar fashion* compression fittings. All connections were made using A steam trap was connected to the drips line and a pipe was run from this to a container for measuring the condensate. It was later decided to use the condensate container as a preheater for the water inlet line to the dif fusion battery. The water inlet line was run directly from the water main to the preheater, at whioh point a needle valve was inserted to allow fine control of the flow. From the pre- heater the water line was run to the center of the diffusion battery, which was arranged in a circle, where a rubber hose 46 was connected so that the inlet oould be readily attached to any cell in the system(Figures 11 & IB)* twelve of the four teen cells would be connected together with a rubber hose con nected to the outlet of the twelfth cell. Shis rubber, out let hose was run to a water cooled trough operating on the counter-current principle for the cooling of the diffusion liquid. From the outlet of this trough the liquor dropped in to a graduated container to measure the quantity of each draw* The two cells not in the system were to be used for emptyii^g, cleaning, and refilling. IV. AUXILIARY APPARATUS Cutter(Figure 13). In building a diffusion battery on the desired scale, it was necessary to obtain a cutting ma chine of adequate size with relatively easy operation. were made on the rate of cutting of several machines. Tests A large power meat grinder was tried, but the fibers in the roots covered the cutting edges after a few pounds had been forced through and stopped the operation. Several vegetable cutting machines and shredders were next tried, but in all cases cut ting edges were soon obstructed by fibers. Attempts were made to obtain a beet sug§r, cossette cutter, but small ones were not available. At this point a vegetable cutter, manufactured by J. 1, Smith and Sons of Buffalo, Rew York, was suggested and on trial proved quite adequate. It was found that about 49 Pigure 11* Pilot Plant Diffusion Battery opened CP - Conneoting Pubes DL - Drips Line HWL - Hot Water Line IHV - Inlet Heedle Valve ISY - Inlet Steam Valve OSV - Outlet Steam Valve P - Preheater SL - Steam Line PW thermometer Welle 49 DLSL - tiWL FIGURE 11. T op V i® W of Diffusion Battery |«T< r t=l FIGURE 13. 51 figure 13* Cutting Machine and Pilot Plant Bif fusion Battery In operation CM - Cutting Machine 13? - Insulating fop 101 - Liquor Outlet Line P - Preheater BOB - Botary Cutter Bowl 51 FIGURE 13. 52 150 pounds of dahlia roots per hour could be cut in a machine with a fourteen inch bowl. One of these machines, with a six teen inch bowl, was obtained* £his machine is designed to be emptied by hand, but this was considered too dangerous so a curved piece of cardboard was cut to be used for the removal of the roots, ^his machine was used throughout the experiments with complete satisfaction. Filter press. A plate and frame filter press with twelve by twelve plates was available. 3!his type of filter press can be used on small scale work where washing is unneces sary and where it is not required that the cake be easily and quickly removed. Since no other filter press was obtainable, this had to be used. It is suggested that an Oliver Continuous Filter Press would prove much more satisfactory. A three quar ter inch bronze gear pump was also available; this was capable of developing heads up to 60 pounds with a satisfactory volume of flow. A pump of sturdier construction would be more satis factory because the filtrations of the suspensions of diatomaceous earth wear a pump of this design and require frequent replacements. Hevertheless, for small plants, this pump is inexpensive and adequate. Air compressor. In filtering the inulin suspension it was necessary to blow the cake dry with compressed air before its removal. Otherwise the cake was so wet and gummy that 53 removal was too slow* An air compressor had to he obtained to perform this operation* ]?• E. Holmes of 665 Del Monte Street, Pasadena, has constructed and patented an air compres sor of radical design which has several obvious advantages over the accepted types of air compressors* It is capable of developing heads of 100 pounds per square inch without the disadvantage of pulsating flow of a reciprocal pump or the fine clearance required for a centrifugal pump* The pump has an infinitely variable displacement at constant speed, and it is built with only three moving parts all of which are rotary* Since the compression chambers rotate, it is not necessary to provide any cooling system because a stream of air is carrie d over the compressor by its motion dissipating its heat by con vection* It is capable of producing a rather large volume of air per pound weight of pump and has been found to be quite satisfactory in our experiments* Steam generator* A small vertical tube steam generator was available* It was designed to develop steam as high as 100 pounds pressure and had an output of 5 horsepower. An iron pipe was run from this generator to the diffusion battery and this was purposely not insulated in order that wet steam only would be used* SEhus, the danger of superheated steam with temperatures higher than those calculated from the steam pressure was avoided. A little difficulty was encountered in attempting to fceep the steam pressure constant, but this 54 ooold be overcome with practice or the insertion of a needle valve in the gas line. In all other respects, this generator was quite satisfactory. Drier. After the inulin had been filtered and blown as dry as possible in the filter press, the cake still con tained approximately two-thirds water. For this reason a shelf drier with a fan and heating coils had been constructed several years ago for the purpose of drying similar inulin cakes. The fan was mounted at one end of a tunnel which con tained the heating elements; this tunnel led into a shelf com partment. The construction was simple, but it was quite satis factory for 10 to 15 pounds of inulin at a time. When fully loaded, the drier would dry the cake in about twenty-four hours. Miscellaneous equipment. Other equipment that was necessary in a plant of this size included various containers and a set of hot plates. It was found that several buckets should be included in the equipment to facilitate the handling of solutions. For large quantities of liquor and pulp six 50- gallon drums were obtained. blue scale linings. These drums were of iron, with These would not be satisfactory because the iron would rust even with the covering of blue scale. It was therefore necessary to obtain some covering that would be non-poisonous and unaffected by the solution. Most paints or enamels would tend to peel when in contact with hot solutions. 55 The Amerioan Concrete and Steel Pipe Company kindly famished two gallons of their special acid and alkali resisting enamel which served to coat the drums on the inside; this enamel holds very well under the conditions used and was more satis factory than any other container with the possible exception of stainless steel or aluminum. Paint. "Amercoat11 is a chemically resistant coating manufactured and marketed by the American Concrete and Steel Pipe Company of South Cate, California. It resists the cor rosive action of practically all inorganic solutions except concentrated nitric and sulfuric acids. It also resists the action of most organic solvents except the lower fatty acids, aldehydes, ketones, esters, chlorinated hydrocarbons, and the ring compounds. It is sufficiently hard to withstand a considerable amount of abrasion and yet is plastic enough not to check, crack, or fracture when subjected to vibration or shock. In order to heat the liquor for the hot filtration necessary for purification, several tin-lined copper boilers were obtained. Gras hot plates were used as a source of heat. To purify the inulin, it was necessary to remove plant dyes which were extracted in the process. To accomplish this 1fo of Uorite and 1$ Dioalite were added to the solution. 56 These were filtered from the hot solution after one half hour of heating* For this operation 15 pounds of Norite and 100 pounds of Dicalite were obtained* Dicalite is also added to slimy, Inulin precipitates to facilitate their filtration* Shis Dioalite was removed in the succeeding hot filtration* CHABTBB V DIFFUSIOK STUDIES OK A PILOT BLAST SCALE Before starting a preliminary diffusion study the appar atus was cheeked for possible difficulties. The apparatus was set up as for an experiment, the steam generator started, and water was run through the system to test for possible leaks and to determine the rate of flow, A few leaks were found in the steam lines, but these were stopped by tightening the fit tings* ho leaks developed in the battery itself and water flowed through very readily. The steam colls proved to be en tirely adequate, as the temperature of the system could be raised to 90° C, or above in less than half an hour. It was found that about 10 pounds of cut dahliaroots were a suitable charge for each cell* This was adopted for the standard charge throughout this experiment* It was desir able to obtain a liquor containing as high a concentration of Inulin as possible. cipitate readily. If it is above 15%, the inulin will pre Since preliminary investigations had indi cated that most of our dahlia roots would average 10% inulin, a draw of about three and one third quarts for each cell should give approximately this concentration. Twelve cells were fill ed with cut roots and water, and the trial experiment was started. 58 Experiment #1. Uine changes were made* The changes were to he at half hoar intervals, hat after the first two draws were made, a back pressure developed to such an extent that the fall pressure of the water main, 60 to 70 pounds per square inch, was not enough to force sufficient liquid through in a half hour* Three and one half quarts of liquor obtained were considered the draw* A new cell was then added and five more quarts of liquor were then forced through to fill this cell before removing the spent cell. Five quarts of dilute solution were discarded each time with the spent fibers, but tests on the concentration of this solution showed that the inulin losses were negligible* By operating the system at maximum pressure, it was possible to complete an experiment of nine changes, but it was not possible to make the changes every half hour* The time of changes varied from one half hour to one hour and fifteen minutes* At this point it was decided to discontinue operation and attempt to eliminate the cause of the back pressure. How ever, the system was approaching equilibrium and sampling of the spent roots and the liquor being obtained would indicate the efficiency of the system under equilibrium conditions* Two samples eaoh were taken from; 1) the original roots, 2) the last draw, and 3) the most completely extracted roots* Since the system was in a state comparable to equilibrium the 59 last draw was a liquor containing inulin approaching the con centration of the equilibrium liquor* similarly, the spent roots were extracted nearly as completely as those from the system in equilibrium. fhe samples, 16.8 grams each, were to be analyzed by the modified sugar beet method. Since this method assumes, in the case of roots and extracted roots, which will for con venience be refe^ed to as pulp in subsequent references, 10.8 grams of water per sample; it was considered advisable to take samples to be dried in order to check the slight variation in this assumption, therefore, two 60 gram samples of roots and two 60 gram samples of pulp were taken for analysis of water eontent. It was later decided to ash one of each of these samples as an added check on the variation in roots and as a check on the extraction of salts. fhe specific gravity of the last draw was tested with the Brix spindle and found to be 10 Brix. Since a solution of 10 Brix is rather dilute for inulin recovery, 50 pounds of fresh roots were added to the liquor and boiled in a boiler for one half hour, then pressed in a cider press to remove the solution and allow recovery of the inulin. All the sol utions were very dark, indicating that long boiling probably removes much of the tannin, fhe solution was set aside to allow the inulin to precipitate. 60 All of the palp was now removed from the battery. It was found that the pressure had forced the bottom of the sack against the bottom of the cell, pushing a small section of the sack into the outlet and reducing filtration area to about one square inch# It was decided that a barrier should be construct ed which would separate the bottom of the sack from the bottom of the cell and allow free flow of the solution. Various ma terials of construction were considered and ridged mahogany discs were finally selected(Figure 14). fhese were heavily varnished to protect the wood from the boiling water. Experiment #8. The results of experiment #1 showed that more roots per cell would give a more satisfactory sol ution, so 15 pounds of roots were used in each cell in this experiment. It was decided to attempt to make a change every twenty minutes in this experiment. With the new wooden discs in the system, operation was more satisfactory and experiment #8 was started. In order to obtain concentrated liquor from the start, water was run through it and the solution was ran into cell #8 until it was filled; this process was continued until the whole battery of twelve cells was filled# Since the battery contained fresh roots throughout, the resultant solution would be more concentrated than an equilibrium sol ution. Consequently twelve changes were made to bring the system to equilibrium, after which the true "experiment” was begun and data were collected. Fifteen grams of CaCOg were Figure 14• Mahogany Disc 61 FIGURE 14 62 added to each 15 pounds of roots to reduce acidity. As the experiment progressed, hack pressure again de veloped and twenty minute changes were possible only through the first five draws. four minutes. The average time per draw was twenty- Samples of solution were taken from alternate draws for Brix: readings and analysis. Samples of roots and pulp were also taken for analysis and for determinations of water content and ash. A record of the cell temperatures, as well as steam boiler pressures, was kept throughout the ex periment. It was desired in this experiment to collect two and one half quarts of solution per draw, but it was found that a large lag occurred between adjustment of water pressure and rate of flow of the outcoming liquor. It was impossible, consequently, to control the volumes of the draws, ied from two and one half quarts to six quarts. They var The average volume per draw was four and one fourth quarts. Experiment #g. To avoid the difficulties of experiment #2, new filter sacks were made of unbleached muslin which is a much more porous cloth, and as was expected, maintenance of flow was easier. in the system. The experiment was made with twelve cells Fifteen pounds of roots were used in each cell and about two and one half to three quarts of liquor were col lected per draw. Sixty-seven and one half grams, 1$, of cal cium carbonate were added with each 15 pounds of roots. TABLE XIII 63 EXPERIMENT #2 12 CELLS — 26 MIN* PER DRAW *Craw Draw Draw Draw Craw Craw Craw Craw Craw * #2 #3 #4 #5 #6 #7 #8 #9 * #1 fime of braw ’ in minutes Volume of Draw in auarts Average dell0 Temperature C. Brix of Draw Inulin floho. in per cent Quantitative Benedict’s Pulp Sample in grams Inulin Co'nc # in per oent Quantitative Benedict’s Dry Weight of Pulp in grams Ash of Pulp in per cent Root Sample in grams Inulin done* in per oent Pulp Sample in grams Hoot Sample in grams Bry Weight of Roots in grams Ash of Roots in per oent Extraction % 20 20 20 25 25 35 35 30 25 4*0 6.0 6.0 3 f0 3.0 4.0 4.0 2.5 3.5 91 90 47 78 66 84 78 12.0 11.5 11.4 12.7 10.3 5*0 6.1 5.7 6*9 5.9 37.6 37.6 37.6 37.6 37.6 0.6 0*6 0.8 0.1 0.5 1.49 1.35 1.85 1.39 1.39 18.8 18.8 20.0 20*0 20.0 20.0 20.0 20.0 18.8 18.8 18.8 6.9 7.4 7.4 20*0 20.0 20.0 20.0 3.81 91*5 3.72 92.0 4*04 89.0 3.93 98.8 3.96 94.0 TABLE XIV 64 EXPERIMENT #3 12 CELLS — 22 MIR♦ PER DRAW *Draw Draw Draw Draw Draw Draw Draw Draw Draw * #1 #2 #3 #4 #8 #6 #7 #8 #9 *1* Time of Draw in minutes Volume of Draw in auart8 Average bell Temperature C. Brix of Draw Draw Sample in grams Inulin Cone, in per oent Quantitative Benedict1s Pulp Sample in grams Inulin Cone, in per oent 5ry Weight of Pulp in grams Ash of £ulp in per oent Rooi Sample in grams Inulin Coho, in per oent Quantitative Benedict’s Pulp Sample in grams Soot Sample in grams Dry Weight of Roots in grams Ash of hoots in per oent Extraction % 25 20 20 20 20 20 30 20 25 3.0 2.6 4.0 4.0 3.5 4.0 2.5 4.0 8.5 90 66 91 ?1 91 94 89 92 92 , 11.2 9.6 10.0 10.0 10.1 18.8 18.8 18.8 18.8 18.8 7.3 6.9 7.1 7.3 6.3 37.6 37.6 37.6 37.6 37.6 0.1 0.0 0.7 0.0 0.4 1.37 1.32 1.17 1.07 1.07 0.430 18.8 18.8 18.8 18.6 18.8 7.6 7.6 9.3 8.1 7.9 7.0 20.0 20.0 20.0 20.0 20.0 20.0 20.0 20.0 20.0 20.0 3.07 2.98 3.13 3.74 1.040 98.8 100.0 92.5 3.05 1.181 100.0 95.0 65 Twelve changes of cells were made to bring the system to equi librium after which the experiment was begun* Samples of roots* solution* and pulp were taken from every alternate draw throughout the experiment in order to determine the efficiency* The experiment involved nine changes* The conditions of experiments #4 and #5 were the same as those of experiment #3 except that experiment #4 employed a system of ten cells and experiment #5 used a system of eight cells* while #3 was operated with a system of twelve cells* Experiment #6* It was desirable to compare sodium bi carbonate with calcium carbonate in controlling acidity in an experiment of this kind* It was expected from previous exper imentation that darker inulin* of an inferior grade* would be produced through the use of sodium bicarbonate* but that this inulin would be filtered more easily* This experiment was for the purpose of checking these assumptions* The conditions of this experiment were similar to those of experiments #3* #4* and #5 except that one half of 1# sodium bicarbonate was used in place of Vjt> calcium carbonate* The system was not brought to equilibrium before starting the experiment because of a shortage of roots. i*or this reason quantitative data on ex traction would have no meaning. Eowever* for comparison of the types of inulin obtained, the experiment was entirely satisfactory* Twelve cells were filled with roots, and the TABLE XV 66 EXPERIMENT #4 XO CELLS --- 20 Mil. PER DRAW * * * Draw #3 * * * Draw #4 * * * Draw #6 16 20 20 26 15 6.0 6.0 2.6 2.5 4.0 92 90 85 90 91 18.8 18.8 18.8 . 18.8 9.9 10.0 10.6 CD Extraction % Draw #2 GO H Brix of Draw braw Sample in grams inulin Cono. in per oent Quantitative Benedict*s Pulp Sample in grams Inulin dono. in per oent Quantitative Benedict's bry Weight of Pulp in grams Ash o£ Pulp in per oent Soot Sample in grams Inulin done, in per oent Quantitative Benedict's Pulp Sample in grams Hoot Sample in grams bry Weight of Roots in grams Ash of Hoots in per cent * * * 7.5 7.7 37.6 37.6 37.6 0.0 0.1 0.2 7.6 9.4 Not significant 1.07 0.93 1*066 . a? Tim© of Draw in minutes Volume of Draw in auarts Average Ceil Temperature C. Draw #1 18.8 8.4 8.1 20.0 20.0 20.0 20.0 20.0 20.0 18.8 8.0 H 03 * * * 9.0 3.26 3.17 1.090 100.0 2.71 — 1*106 98.8 97.5 67 TABLE XVI EXPERIMENT #5 8 CELLS — * * * Tim© of Draw in minutes Volume of Draw in Quarts Irerag© d©ll Temperature °C. Brix of Draw Draw Sample in grams Inulin Cfono* in per oent Quantitative Benedict *s tul'p Sample in grams inulin 6 one. in per cent Quantitative Benedict1s Ery Weight of Pulp in grams isk of Pulp in per cent boot Sample in grams inulin done* in per oent Quantitative Benedict1s Eulp Sample in grams Hoof Sample in grams l>ry Weight of Boots in grams i!sk of Hoots in per cent Extraction % Draw * * 25 MIN. PEB DRAW Draw #2 * * * Draw #3 * * * Draw * * Draw #6 : SO 30 25 20 20 4*0 2.5 2.6 4.0 6.0 86 90 90 83 88 9*6 9.5 8.9 18.8 18.8 18.8 7.2 6.9 7.1 37.6 37.6 37.6 0.0 0.3 0.4 7.5 Not significant 1.09 1.12 1.24 1.590 1.800 __ 18.8 18.8 18.8 8.1 7.1 6.8 20.0 20.0 20.0 20.0 20.0 20.0 8.8 3.36 3.10 1.235 1.055 100.0 3.25 95.7 94.0 68 experiment was immediately started* Five changes were made and water was continued running through the system until the liqupr obtained dropped to a concentration of less than 8 Brix* The liquor obtained was treated in the same manner as those in experiments #3, #4, and #5; it was allowed to stand for an in terval of thirty-six to forty-eight hours and the inulin was removed by filtration and dried* It was somewhat darker than that from calcium carbonate experiments, but it was not appre ciably easier to filter* From this it would appear that so dium bicarbonate offered no appreciable advantage and had some di sadvantages• Extraction curve of cells in system* Upon the comple tion of experiment #5, the system was disconnected and pulp samples were removed from each cell in the system in order to determine the degree of extraction in the consecutive cells of the system* Samples were analyzed for inulin plus sugar, while other samples were analyzed for water content and ash* Cell #1 was the cell containing the freshest roots, and cell #8 contained the roots which were most completely spent* fhe results of this experiment are given in fable XVII, and Figure 16* fI All 'E-i ±1 : ie: 69 TABLE XVII DETERMIH ATIONS FOR EXTRACT!OH CURVE # of Cell * $ Inulin * * * Pulp * Sample * in &rams * Dry * Weight * in «:rams 1 5.5 30.0 3.220 2 4.2 30*0 2*455 3 3*4 30.0 2.470 4 1*4 30.0 2.550 5 0.85 30.0 1.880 6 0*8 20.0 1.890 7 0*55 20.0 1.650 a 0*5 20*0 1.650 cell * * Ash per cent 1.07 1.21 1.13 #1 -- Quantitative Benedict's — - 6.75fo The average original inulin content of these roots was in the neighborhood of 7*3$ which indicates an inulin extrac tion of aboat 93$ under the conditions of this experiment. Therefore* with roots cut to a rather small size, a battery of eight cells gave sufficiently complete extraction for all practical purposes. However, the number of cells required in the system will vary with the time of each draw and with the subdivision of the roots. If the roots were sliced or shred ded only, more cells were required to achieve this degree of extraction with the same conditions of time, temperature, and 70 flow* For general purposes of extraction using a cutting ma chine* eight cells should usually he sufficient. Determination of hydrolysis of crude inulin* Determin ations have been made on the hydrolysis of purified inulin in boiling water under various pH conditions*0 However, it was considered advisable to make a check of the hydrolysis of crude inulin as obtained from the diffusion battery, because this would be more comparable to the conditions existing in the battery. Six 40 gram samples of dry inulin were taken from experiment #3, way, fhis inulin had not been purified in any Each sample was dissolved in 200 ml, of boiling water and placed in a boiling water bath under air reflux. Samples #1 and #2 were removed after ten minutes; #3 and #4 after thirty minutes, and #5 and #6 after three hours, fhe inulin was allowed to precipitate during forty-eight hours, after which the samples were filtered and dried* She results of these determinations showed that rather large leased occurred in all cases* fhe increased loss with Increased time was small as compared to the initial loss. From this it would appear that slightly lower losses of in ulin can be expected if it is in the battery for a shorter period of time, but the increased loss with the slight increase in time is not large. It is possible then that the increased extraction with the increase in time might more than balance 71 2ABLE XVIII PEfERMINATI OH OF HYPROLYSIS OF CRUPE INULIN Weight * Average Lost ^Weight Lost in grams _ Number of * Weight of * fime in * Sample Sample Bath * in grams * in min* * in grams * 1 40*0 10 5*85 z 40.0 10 5. 26 3 40.0 30 7*06 4 40*0 30 6*43 & 40*0 180 6 40*0 180 the losses through hydrolysis* ____ 7*60 5*65 6.75 7*62 7*63 In order to determine optimum conditions for inulin extraction it will be necessary to carry out a great many diffusion studies* Shese studies will prob ably not be completed until the process is in commercial use* Experiment # 7 * It was considered necessary to determine the obtainable yield if a battery of eight cells were run much more rapidly than in the previous experiments. In order to do this changes in construction to facilitate rapid operation were imperative* Hose couplings replaced the flared fittings* and sacks with screen bottoms replaced the filter cloth sacks of former determinations* A clamp was placed on the rubber hose to serve as a valve to stop the flow of liquor during changes. 72 TABLE XIX EXPERIMENT #7 8 CELLS 10 MIN. PER DRAW *Praw *Braw *Praw *Praw *Braw *Braw *Praw : #x * * #4 * * #5 * * #6 * * #2 * * #3 * Time of Draw in minutes Volume of Brew in quarts Average deli Temperature C. Brix of Draw Draw Sample in grams Inulin tiono. in per oent Quantitative Benedictfs Pulp Sample in grams I&tilin done, in per oent Quantitative Benedict1s Pry Weight of Pulp in grams Ash of Pulp in per cent Root Sample in grams inulin Cone, in per oent Quantitative Benedict vs Pulp Sample in grams Root Sample in grams Pry Weight of Roots in grams Ash of Roots in per cent Extraction % 10 10 10 10 10 10 5 2.7 2.7 3.2 2.7 2.7 2.7 4.3 18.8 18.8 18.8 18.8 90 - 95 12.2 18.8 18.8 18.8 6.6 8.6 7.5 6.9 37.6 37.6 37.6 37.6 1.2 1.2 0.5 0.5 1.766 1.640 1.530 1.280 18.8 18.8 18.8 18.8 9.1 9.1 10.6 10.6 20.0 20.0 20.0 20.0 20.0 20.0 20.0 20.0 3.68 87.0 3.16 87.0 3.72 95.0 3.92 95.0 73 With this equipment ten minute changes were made and satis factory control of liquor flow was possible* An experiment of seven changes was made after bringing the system to equilib rium in the usual manner* Charges of 15 pounds of roots per cell were used throughout the experiment* and analyzed as in previous experiments* Samples were taken fhis experiment in dicated that operation at such a rapid rate is not so satis factory as when changes are made at twenty minute intervals* One per oent calcium carbonate was used to control the acidity of the solutions* She quality of the inulin obtained was not noticeably different from that obtained in other experiments in Which calcium carbonate was used* 74 TABLE XX COMPOSITE RESULTS OF EXPERIMENTS #3, 4* 5 t 6t & 7. * * * #5 9 # of Draws Total Roots used in lbs* 133 Average Inulin 8.1 Content in % Soial Inulin in 10.9 Boots in lbs* try Inulin recovered in lbs . 4*31 Recovery in % Average Inulin Content in % 39.5 0*24 Extraction % 97.0 bugar in liquor 2.9 Filtrate in % Volume of Filtrate in qts* 32.0 loss thru Hydrolysis 17.0 pH of Liquor Average CellQ Temperature C. 5.96 * * * #4 * * * #5 * * * #5 * * * #7 5 5 75 75 210 105 7*6 7.3 7.7 9.85 5.7 5.5 20.8 10.34 2.56 2.69 9.0 3.33 44.9 48.8 0.1 7 43.3 31.8 0.85 0.23 98.6 97.0 91.0 2.9 2.6 2.5 17.0 18.0 17.2 17.0 5.90 Explanation of loss* 5.68 53.0 20.0 10.7 7.70 90 - 95 In th© r©salts of all the experi ments on th© diffusion battery, it will be noticed that appre ciable losses of inulin are not explained* A large part of this loss is due to mechanical imperfections of the battery* In every ease leaks developed at some point in the system dur- 75 ing the experiment; these and mpchanipal losses due to rapid handling account for part of the unexplained loss* Another important source of error is the transformation of levulose to glucose during the experiment* If 1$ of the levulose is converted to glucose in a solution of levulose, the saccharimeter reading will he reduced by 1.6$. Xhis is due to the fact that the reading is not only reduced by l$f from the loss of levulose, but it is also reduced because the glucose is dextro rotatory and its rotation is therefore subtracted from the levulose reading which is levo-rotatory* fhe reduction in reading is 1*6$ rather than the expected 2$ because glucose has a specific rotation of 52*5° while levulose has a spe cific rotation of -90°. She rate of transformation of levulose to gluoose, with varying temperatures and under varying pE conditions, has been investigated at the United States Bureau. *^a It was shown that 1$ of the levulose, in a water solution of levulose, is transformed to glucose in four and one tenth minutes at a temperature of 100° C* and a pH of 7; at a temperature of 90° C, and a pH of 7, twelve minutes are required for 1$ trans formation; at a pH of 6, twenty-five minutes are required at 100° C*f and seventy-one minutes at a temperature of 90° C* Sherefore, a large transformation of levulose would occur in a solution at a temperature of 90° to 100° C. and a pH between 76 6 and 7* These were the conditions maintained in the oper ation of this plant, and since an appreciable quantity of levulose is formed during extraction, a large part of the levulose would be in the form of glucose in the liquor fil trate of the experiments. Therefore, the readings obtained by the Modified Sugar Beet Method, on these filtrates, were low and the hydrolysis of inulin was larger than that indi cated in these data. The results obtained in determinations of total sugar in the roots by Quantitative Benedicts reagent were consider ably higher than those readings obtained by the Modified Sugar Beet Method which would indicate that considerable hydrolysis have and transformation of the inulin may*already occurred in the fresh roots. The levulose present would undergo transformation throughout the extraction adding to the apparent loss. Since it appears that considerable inulin was hydrolyzed before the diffusion process was started, there was less inulin in the roots than these results indicated* A larger per cent of the available inulin was recovered than is evidenced in these data. Ho method has been found for determining inulin in the presence of levulose or levulose in the presence of inulin as they occur in the roots. CHAPTEB VI SUMMARY AMD CONCLUSIONS fhe purpose of this investigation was to construct and operate a pilot plant for the extraction of inulin from dahlia roots* A review of the literature showed that very little work had been done in this field* attempted in this laboratory* Small scale operations had been At first, preliminary experi ments were performed using 6 ounce bottles coupled as a dif fusion battery* fhese verified previous work and indicated that extraction by means of a diffusion battery might be a promising method for obtaining inulin on a large scale* In order to develop a technique for handling solutions of the type which would be obtained from a diffusion battery without large losses, inulin solutions were prepared from dahlia roots by boiling them in wash boilers and pressing them to obtain the liquor* With this liquor the available purifi cation equipment was tested* filtration methods were examined to determine quantities of decolorizing agents and filter aids necessary* Further data on time and pressure required for filtration of these solutions were gathered* A pilot plant was built according to a design based 78 upon that of the usual type of diffusion battery* It was de signed to produce about 2 pounds of inulin per hour, using 30 to 60 pounds of roots. Seven diffusion studies were carried out with this bat tery in order to establish, as nearly as possible, the optimum conditions. The temperature was kept between 90° and 100° C, in all oases and variations were made in the time of extrac tion and the number of cells. It was found, by making four determinations with a sys tem of twelve cells and about twenty-five minute draws, that yields of inulin of approximately 40$ could be expected, and about 97$ would be removed from the roots. At about twenty minute intervals and with ten cells, the extraction was equally thorough, being 98.6$ and with eight cells 97$. fhe recovery was 44,9$ with ten cells and 48,8$ with eight cells. This in dicates that eight or ten cells allow for adequate extraction and that higher recovery, with decreasing number of units in the diffusion battery and shorter time of draws, can be ex plained by the fact that the inulin is exposed to hydrolysis for a shorter period of time* However, a system of eight cells with ten minute draws showed an extraction of 91.3$ and a reoovery of 31.8$. In all of these determinations the roots were cut quite 79 fin©; the particles averaged about one eighth inch in diameter* It is to be expected that larger particle size would result in slower diffusion rates and would consequently require more cells or slower rate of operation* This would again lead to Increased losses through hydrolysis hence smaller particles Should give more satisfactory yields* From this research it would appear that the following conditions lead to the most satisfactory extraction and re covery of inulin from dahlia roots by the diffusion method* 1* The roots should be cut into pieces of one eighth inch diameter or less* £• A battery of eight to twelve cells is adequate* 3* The time of draw should be about twenty minutes* 4* The pH of the solution should be maintained as near 7 as possible* 5* The temperature of the solution should be between 90° and 100° C. 6* The liquor obtained should be cooled rapidly in order to reduce hydrolysis to a minimum* 7* The inulin should be separated from the mother liquor by filtration after an interval of thirtysix to forty-eight hours* The inulin obtained under these conditions can be satisfact orily purified by redissolving it in hot water, decolorizing 80 it with 1% decolorizing charcoal, allowing it to stand before reprecipitating, filtering, and drying# It has been shown that inulin can be obtained from dahlia roots in reasonable quantities by extraction in the diffusion battery# Yields obtained in the operation of a pilot plant Indicate the feasibility of converting a beet sugar plant, in its off season, to extract inulin from dahlia roots# With commercial methods of operation and control, higher yields could be expected# She conversion of inulin to levulose syrup or crystalline levulose might be accomplished by utilizing the sugar end of the factory# fhese possibilities are dependent upon the development and production of satis* factory dahlia roots in large quantities# Waste In, to J e vet\ ----- . I ----- 1 /?o. s : !/?r- ;/ €nrs,'_ v Wauhcj o/ c ok - ^ on ConI VQyor ^ r n r ^C j i'f vor > C & C03 |ad ded \ . \Pressed J !P u /b i / D r 4; CO fSOo/ro L D'S/fo/*jt. Baticr^ Hot | L, / C^uah OeCoion* -}/*$ C*4»* C£>n!*dfal J t# < J 't c i - i t I u . n end of factory In u h n P a rt a CarKt hydnl/..s CQfke to dr/er lt)Ut/n *— filte re d filtrate btorcd $ 36 tirj. fi/trail| to t*r-j /nrnI* /o^ j bif ctis til/artt'oh Plow tnation Sheet I n u h n f xhr&c f/on FIGURE 16 f % ///- t e n & rd (y d d c d wofjrr. CGffi'C /IICohrl /nv/sfl i r* BIBLIOGRAPHY 1. Bachman, George, John Haldi, Winfrey Wynn, and Charles Ensor, Journal of Nutrition, Vol. 16, p. 229, 1938. 2. Bailey, Liberty H., Manual of Cultivated Plants. 3. Bartel, A. W . t "Purification and Study of Inulin," Unpublished Thesis, University of Southern California, 1937. 4. Black, Le Roy G., "The Commercial Production of Levulose," Unpublished Thesis, University of Southern California, 1927. 5. Bollman, J* L., and P. C. Mann, "Utilization of Various Carbohydrates by Lepancreatized Animals," American Journal of Physiology, 107:183-9, Jan. 1934. 6. Brobst, W. H., "A Study of the Behavior of Inulin in Solution," Unpublished Thesis, University of Southern California, 1939. 7. Goudberg, A., Zeitung. Bxp. Path. Ther., 13:310, 1908. 8. Haworth, W. N . , E. L. Hirst, and E. G. V. Percival, Chemical Society Journal, Transactions, 141, II, 2684-88, 1932. 9. Holzman, Josef J., "Studies in the Production of Inulin," Unpublished Thesis, University of Southern California, 1938. 10.. Jackson, Richard P., United States Patent #2,007,971, Process of Making Sugar Products, July 16, 1935. 11. Jackson, Richard P., C. G. Silsbee, and M, J. Proffitt, Industrial and Engineering Chemistry, 16:1250-51, 1924. 83 12. Jackson, Richard P., C. G. Silsbee, and M. J, Proffitt, United State© Bureau of Standards, Scientific Paper, Uo. 519, p p . 587-617. 13* Jackson, Richard P., and R. McDonald, United States Bureau of Standards, Journal of Re search, Rft 2^S . 14* Kahn, Allen Ray, and Leroy S. Weatherby, Sugar. Sugar Publications Co., Los Angeles, p. 37. 15. McGlumphy, J. E., J. W, Kichinger, with Hixon and Buchanan, Industrial and Engineering Chemistry. S3, 1202, i'9#L. 15a. Mathews, Joseph A., and Richard P. Jackson, United States Bureau of Standards. Bureau of Research. Research Saner. SFTSTT.------------ -------------------16. Ohlmeyer, Paul, and Hans Pringsheim, Bar., 66B, 1292*5, 1933. 17. Peabody, Lawrence, The Dahlia. New York, Orange Judd Pub* lishing Co., Inc•," 1931. 18* Pringsheim, Hans, She Chemistry of the Monosaccharides and of the Polysaccharides. Sew York, W » r a w*Hili, 1932. 19. Rieger, Wray M., "Commercial Preparation of Levulose from Dahlia Tubers," Unpublished Thesis, University of Southern California, 1934. 20. Shelly, Plorence M., Research in progress, 1940. 21. Spoehr, H. A., Plant Physiology. 13:207*8, 1938. 22. Thorpe, Chemical Dictionary. Vol. II, p. 53. 23. Vivian, R. E., "The Preparation of Levulose," Unpublished Thesis, University of Southern California, 1922.