A critical study of the Gooch and Havens hydrochloric acid-ether method for the quantitative separation of beryllium and aluminumкод для вставкиСкачать
A CRITICAL STUDY OF THE GOOCH AND HAVENS HYDROCHLORIC ACID-ETHER METHOD FOR THE QUANTITATIVE SEPARATION OF BERYLLIUM AND ALUMINUM A. Thesis Presented to the Faculty of the Department of Chemistry The University of Southern California In Partial Fulfillment of the Requirements for the Degree Master of Science Edwin Ralph Calderon August 1941 UMI Number: EP41532 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. UMI Dissertation ftAfishfeg UMI EP41532 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 LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 C 4 %C J1his thesis, written by .........E^F^.|i...Ralp.ii...Gal.djer.aa......... under the direction of h ..l3 F a c u lty Committee, and a p p ro ve d by a l l its members, has been presented to and accepted by the Council on Graduate Study and Research in partial f u l f i l l ment of the requirem ents f o r the degree of .MASTEE...QE...S.GISECE. Dean Secretary AUGUST,1941 Date .........:.... Faculty Committee Chairman * j TABLE OF CONTENTS PAGE A CRITICAL STUDY OF THE GOOCH AND HAVENS HYDROCHLORIC ACID-ETHER METHOD FOR THE QUANTITATIVE SEPARATION OF BERYLLIUM AND ALUMINUM . . . . . . . . . . . . 1 Historical Introduction • 1 A rapid Survey of the Literature . . . . . . Scope and Nature of this Investigation Experimental Work . . . . . . ..... . . . .. . . . . . . 4 11 . . 11 Preparation of standard solutions of beryllium and aluminum chlorides .... 11 Test of Gooch and Havens Method . . . . . . . . . 13 Time and Convenience Studies of Gooch and Havens Method Compared vith other Methods . . . . . 18 The Generation of Hydrogen Chloride Gas . . . . .. 23 The Keeping Qualities of the Wash Liquid of Ether and Hydrochloric Acid . . . . . . . . ..... 24 The Investigation of a Possible Substitute for Diethyl Ether • 24 . . . ♦ .. . . . . . . . . . .. . 27 The Investigation of a Less Corrosive and More Conveniently handled Reagent than Hydrogen Chloride Gas The Detailed Application of the Gooch and Havens Method to Mineral and Ore Analysis * . . . . . Summary .......... BIBLIOGRAPHY . . . . . . . . . . . . . . . .................... .. . . 29 38 40 A CRITICAL STUDY OF THE GOOCH AND HAVENS HYDROCHLORIC AC ID-ETHER METHOD FOR THE QUANTITATIVE SEPARATION OF BERYLLIUM AND ALUMINUM Historical Introduction The discovery of beryllium resulted from the Abbe Hatty*s (1) observation of the close similarity and probable identity of beryl and the emerald. At this suggestion Vauquelin (1) made some very careful chemical analyses of these two minerals, and found in 1798 that they are indeed identical, and that they contain a new earth, which he named glucina, but which is now known, as beryllia. In speaking of the discovery of beryllium, Fourcroy once said, It is to geometry that we owe in some sort of this discovery; it is that science that furnished the first idea of it, and we may say that without it the knowledge of this new earth would not have been acquired for a long time, since according to the analysis of the emerald by M. Klaproth and those of beryl by M. Bindheim one would not have thought it possible to recommence this work without the strong analogies or even almost perfect identity that Citizen Hatty found for the geometrical properties between these,two s tony foss ils. (1) As a result of his analysis of a Peruvian emerald, Klaproth had stated that this gem had the following composition: Silica,,Jsilexw 66.25$ Alumina, ^alumine or argilw 31.25$ Iron oxide 0 .50 $ To explain his extravagance he said, ,3For the specimen of emerald sacrificed to this analytical process, I am indebted to the liberal kindness of Prince Dimitri Gallitzin, whose zeal for the study of mineralogy is most honourably known."(2 ) Beryl had also been analyzed by Bergman, Achard, Bindheim, and Vauquelin, and was supposed to be a calcium and aluminum silicate.(3 ) The identity of beryl and the emerald was not suspected, until the famous French mineralogist, L ’Abb£ Hauy, made a careful study of their crystal forms and physical properties and was so struck by the similarity of the two minerals that he asked Vauquelin to analyze them chemically. Although the latter had previously overlooked the new earth because of its similarity to alumina, he found in 1798 that the hydroxide that precipitates when caustic potash is added to an acid solution of the beryl does not dissolve in an excess of the alkali. It also differs from alumina in other respects, for it forms no alum, it dissolves in ammonium carbonate, and its salts have a sweet taste. Vauquelin*s paper read before the French Academy in 1798,(3),(4) proved that beryl and the emerald have the same composition, and that they contain silica, alumina, and a new earth, a sample of which he presented to the academy. At the suggestion of the editors, 1 of the Annales de Chimie et de Physique, he called the new earth Glucina, meaning , 1 Guyton de Morveau, Monge, Berthollet, Fourcroy, Seguin, Chaptal and Vauquelin. sweet. The specimen of beryl that Vauquelin analyzed was presented to him by "Citizen Patrin, whose zeal for the advancement of the sciences is well known to every one of their cultivators."(3 ) Vauquelin believed that Bergman’s incorrect conclusions as to the chemical nature of the beryl had been caused by the unwillingness of his "active mind to submit to the details of experiment." Thus Bergman, and Bindheim as well, had entrusted their analyses to young pupils who were incapable of distin guishing a new substance when they saw it. According to Bindheim’s analysis the beryl consisted of 64 per cent silica, 27 per cent alumina, 8 per cent of lime, and 2 per cent of iron (total 101 per cent) (3 ). When Vauquelin analyzed a Peruvian emerald (5 ) after his discovery of chromium and glucina, the results differed greatly from his previous ones and from those of Klaproth. Silica Alumina Glucina Lime Chromium oxide Moisture or dther volatile matter Total He found: 64.60 1 4 .00 13 .00 2 .5 6 3.50 2 .0 0 99.66% J. F. Gmelin’s analysis of a Siberian beryl soon confirmed Vauquelin’s conclusions as to the essential constituents of that gem, for he found no lime, but only silica, alumina, glucina, and small amount of iron fexide (6 ). Since yttria, as well as glucina form sweet salts, Klaproth preferred to call the latter earth beryllia, and it is still known by that name. Beryl and the emerald are now known to be a beryllium aluminum silicate, Be-^AlgC SiOj )g. A RAPID SURVEY OF THE LITERATURE Today, as at the time of.Vauquelin et al, the satisfac tory separation of aluminum and beryllium is fraught with difficulties. Of the numerous methods proposed to solve this problem, there are three fairly satisfactory methods in use today. A brief resume'’of the literature from the time of the discovery of beryllium to the present day will be presented. Vauquelin (3 ) first used ammonium carbonate, but his first separation depended upon the solubility of beryllium hydroxide ini potassium hydroxide and its precipitation on boiling. Gmelin (7 ) and Schaffgotsch (8) both used this same method, but it is far from accurate. Scheerer (9 ) first pro posed the separation of the last traces of iron from the ammonium carbonate solution by means of ammonium sulfide. Berthier (10) proposed the use of ammonium sulfide as a reagent but the method was shown to be valueless by Bottinger (11). In 1850 Rivot (12) proposed the Ignition of the oxides in a current of hydrogen, whereby the iron oxide was reduced to metal and could be dissolved out by dilute nitric acid or its mass determined by the loss in weight. Debray (13) developed a separation depending on the action of zinc on the mixed sulfates, precipitating the aluminum as a basic sulfate, but the method was never made quantitative. Joy (14) made a comparative study of all methods proposed to his time. Gibbs (15), in 1864, for the first time suggests the use of sodium flouride to quantitatively separate aluminum from beryllium, and Pollock (16) showed that'the flouride separa tion is exceedingly sharp. Cook (17), after reducing the iron in hydrogen, volatizes it in a current of hydrogen chloride. Havens and Way (18) accom plish the same result without previous reduction of the oxide. Rossler (1 9 ) succeeded in separating beryllium from small amounts of aluminum by precipitating with ammonium phosphate in the presence of citric acid. Vincent (20) uses dimethylamine to precipitate beryllium salts and finds that the aluminum compound is soluble in excess of the reagent. Iron acts like beryllium. Renz (21) confirmed this, and stated the same to be true of methyl, ethyl and diethylamine, and claimed the results to be quantitatively accurate. Zimmermann (22) in I8 87 returned to the old KOH and sulfite and thiosulfate method without any special addition. Schier (2 3 ) in 1892, Atkinson and Smith (24) in I8 9 5 , and Burgass (2 5 ) in I8 96 separate iron quantitatively from beryllium by means of nitroso - B - napthol. Lebeau (26) pre cipitates the iron in nitric aifcid solution by potassium ferro- 6 cyanide, removes the iferrocyanide by cupric nitrate, and excess copper by HgS. Hart (27) removes the major part of both iron and aluminum by careful precipitation of the sulfates with sodium carbonate, the beryllium being the last to precipitate, owing to the great solubility of its own hydroxide in its own sulfate. Gooch and Havens (28) separate beryllium and aluminum quantitatively by the insolubility of hydrous aluminum chloride in a miacture of hydrochloric acid and diethyl ether which has been saturated with hydrogen chloride gas. This is the method here investigated, and is the subject of this paper. Haber and Van Oordt (29) dissolve basic beryllium acetate in chloroform, leaving behind aluminum and iron as acetates, layers(3 0 ) removes iron electrolytically from a slightly acid solution of the sulfates using a mercury electrode. Parsons and Robinson (31) separate basic beryllium acetate in a pure state from other acetates by means of its ready solubility in hot glacial acetic acid and comparative in solubility in the same reagent when cold. Parsons and Barnes (3 2 ) separate beryllium and aluminum quantitatively by taking advantage of the insolubility of aluminum hydroxide and ferric hydroxide in a boiling ten per cent sodium bicarbonate solution, and of solubility of beryllium hydroxide in the same medium. Glassman (33) separates aluminum and beryllium quanti tatively by treating a slightly acid mixture of their salts 7 with an excess of sodium thiosulfate solution and heating the mixture to tolling. as hydroxide. The aluminum is completely precipitated Wunder and Chiladze (34) have developed a separa tion "by treating the tared dry chlorides with a saturated solution of either potassium of sodium hydroxide, until dis solved. 400 ml of water are added and the whole toiled for one hour, the volume teing kept constant. Beryllium hydroxide is filtered off and determined in the usual way. The acidi fied filtrate yields aluminum hydroxide with ammonia. To detect and separate aluminum from teryllium, Browning and Kuzirian (3 5 ) use amyl alcohol on a mixture of the nitrates, aluminum nitrate teing completely insolutle in amyl alcohol. It is not satisfactory for a quantitative method. To separate teryllla from the argillaceous earths, Wunder and Wenger (3 6 ) fuse the mixture with sodium cartonate, leach out with water, aluminum and chromium oxides dissolving. Iron and beryllium oxides remain insolutle. To separate alumina and beryllia, the authors fuse with dodium cartonate, dissolve in water, filter and weigh the ignited residue as BeO, deter mining alumina in the filtrate. Bleyer and Moorman (37) have developed a volumetric determination of teryllium. Normal beryllium salts hydrolyze yielding acid which, acting o m a mixture of potassium iodide and iodate, produces free iodine. The iodine is distilled in a current of hydrogen into potassium iodide solution and titrated with standard thiosulfate. The results are claimed to he accurate. Wunder and Wenger (3 8 ) state that beryllium is precipi tated quantitatively by sodium carbonate, while aluminum re mains in solution. Kling and Gelin (39) devised a scheme where by basic beryllium acetate is sublimed in vacuo at I65 degrees, while the acetates or basic acetates of iron and aluminum are non-volatile under these conditions. The results are fairly satisfactory. Copaux (40) sinters one part of beryl mixed with four times its weight of sodium fluosilicate in a clay or graphite crucible. The fritted mass is pulverized and extracted with hot water, the sodium fluoberyllate dissolving, leaving insoluble sodium fluoaluminate. The reaction is Be^Al2 (SiO-j)g + 6 Na2SiFg= 3 Na2BeF^ + iWa^AlFg + 3 SIF^ + 9 SiOg. Britton (41) investigated the isoelectric points of aluminum and beryllium hydroxides and has developed a rather clean separation. For the determination and detection of minute amounts of beryllium Fisher (42) adds an alkaline solution of quinalizarin (1, 2, 5, 8 - tetrahydroxyanthraquinone) to an alkaline beryllium .solution giving a distinct blue precipitate which is satisfactory for the detection of beryllium in the presence of iron, aluminum, phosphate, tartrate and magnesium. The precipitate is filtered off and determined colorimetrically. Moser and Neissner (43) base their separation upon the 9 different behavior of beryllium from aluminum when tannin is added to a dilute solution containing ammonium acetate. Aluminum is precipitated quantitatively, but beryllium remains in solution. For the colorimetric determination and detection of beryllium, Kolthoff (44) uses curcumin. reagent will show 0,05 Mg Be/liter, It is stated that the Kolthoff and Sandell (45) treat a solution of beryllium and aluminum salts with 8-hydroxyquinoline in dilute acetic acid solution at about 60 degrees. Dilute ammonium acetate is added until a permanent precipitate is obtained. The precipitate of aluminum oxyquinolate Al(C^Hg 0N)^ is filtered, and beryllipa is determined in the filtrate. This is said to be one of the best methods yet devised, J^lek and Kota (46) use guanidine carbonate in presence of tartrate at proper PH to precipitate beryllium. cipitate is filtered, ignited and weighed as BeO. not precipitated, The pre Aluminum is Gaspar y Arnal (47) ■uses calcium ferrocyanide to precipitate aluminum in presence of beryllium, Pache (48) developed a method to separate small amounts of beryllium from large amounts of aluminum. The alloy is heated with dilute sodium hydroxide to dissolve most of the aluminum and zinc. The beryllium and other hydroxides are dissolved in nitric acid, 40 per cent sodium hydroxide is added. The mixture is neutralized with ammonium hydroxide to precipitate aluminum ShdsberylHum hydroxides. in 10 This precipitate is dissolved molar potassium hydroxide, diluted and heated to 10 40 degrees. Beryllium hydroxide precipitates and is finally weighed as BeO. Nichols and Schempf (49) use tannin to separate aluminum and beryllium. They claim that aluminum is completely preci pitated by tannin at Pjj 4.6 in presence of beryllium! beryllium is determined in the filtrate. The The precipitates yield the oxides, AlgO^ and BeO on ignition. It is said to be accurate within 0.4 mg. This covers the previous work up to the present time. From survey of the literature it would seem that many satisfac tory methods for the determination of beryllium in the presence of aluminum exist. The practical fact, however, is that there is scarcely an element in the whole mineral kingdom which is more frequently wrongly reported in analyses than beryllium. Of all the methods proposed there are actually very few that will lead to accurate results, and, unfortunately, these seem to require much study and experience before they can be used effectively by an unpractised analyst. Among the very few methods that could be designated as "hopeful11, that of Gooch and Havens occupies a prominent place. seems to be seldom used. And yet this method This may be largely due to the dis inclination of analytical chemists to work with ether and hydrogen chloride gas. of close study. None the less, the method seems worthy 11 SCOPE AND NATURE OF THIS INVESTIGATION The following Is a brief outline of the research. 1~, The accuracy of separation by the method as origi nally given by Gooch and Havens. 2. Time and convenience studies versus the 8 -hydroxy- quinoline and sodium bicarbonate methods. 3. Comparison of the Gooch and Havens method "with the Parsons and Barnes method for accuracy. if. The best technique for carrying out the generation of, and saturation by, hydrogen chloride gas to avoid dis comfort in the laboratory. 5. The keeping qualities of the wash liquid of ether and hydrochloric acid. 6. The investigation of a possible substitute for diethyl ether, which would be less volatile, less expensive and less lethal. 7. The investigation of a less corrosive and more con veniently handled reagent than hydrogen chloride gasV 8. The detailed application of the Gooch Hand Havens method to mineral and ore analysis, EXPERIMENTAL WORK Preparation of standard solutions of beryllium and aluminum chlorides. Forty-one grams of aluminum chloride were 12 dissolved in water acidified with enough hydrochloric acid to prevent hydrolysis and made up to a volume of two liters. This solution was standardized by precipitation of the aluminum hydroxide from 2 5 .0 ml portions by the addition of ammonium hydroxide in slight excess’, which is conveniently indicated by a slight ammoniacal odor. The precipitate was filtered, washed and ignited to constant weight-of AlgO^. The following results were found: grams No. 1 0.2163 ml No. 2 0.2158 Average 0.2160 Eleven grams of beryllium hydroxide were dissolved in cold dilute (1:1) hydrochloric acid and diluted to one liter. 2 5 .O ml aliquots were withdrawn, about 100 ml of water added and brought almost to boiling. Six normal ammonium hydroxide is then added dropwise with constant stirring using bromethyl blue as indicator. The precipitated beryllium hydroxide was filtered, washed with a hot solution of two per cent ammonium nitrate, ignited and weighed as BeO. It is a good plan, after filtering the beryllium hydroxide, to redissolve the minute quantity of the precipitated hydroxide sticking to the beaker in dilute (1 :1 ) nitric acid, and to reprecipitate with a few drops of ammonia, adding this to the main precipitate in the funnel. grams The following results were found: BeO/ 2 5 ml No. 1 0.1099 No, 2 0.1102 Average 0.1101 13 TEST OF GOOCH AND HAVENS METHOD A typical experiment will now be discussed in detail. Into a 250 ml Erlenmeyer flask were pipetted two 10 ml aliquots each of the standard aluminum and beryllium chloride solutions. Then in a small beaker were mixed 10 ml of twelve molar hydro chloric acid and 10 ml of ether. The ether is readily soluble in the concentrated hydrochloric acid, forming a single phase. The first product of the reaction between an ether and a strong acid is an oxonium compound. The freezing point curve of ether and anhydrous hydrogen chloride indicates the existence of two addition compounds, (CgH^JgO'HCl and (CgH,_)20*2 HC1. In these compounds the hydrogen atoms of the hydrochloric acid are attached to the unshared electron pairs of the oxygen atom: CoHl- : 0 : H+, Cl" 2 5 « c2h5 and C0Hc : 8 : H++, 2 Cl“ 2 3 u e2H5 The ready solubility of ethers in concentrated acids is due to the formation of these oxonium compounds. They are readily decomposed by water, the ether being thrown out of solution on dilution. This mixture of ether and hydrochloric acid is then added to the standard solutions in the flask. Two phases will be found here but on saturating with hydrogen chloride gas, a homogeneous phase will be found. The flask is provided with a good 2-hole rubber stopper (preferrably a new one) carrying 14 an Inlet tube reaching almost to the bottom of the flask. It is advantageous but not essential to have the end of the tube drawn out to a fine point. The stopper is also provided with a short exit tube. The flask is then placed under the water tap and a rapid stream of pure dry hydrogen chloride gas passed in until the crystalline precipitate of hydrous aluminum chloride is ob served. The gas stream is stopped and 20 ml of ether are added and again saturated with the gas. As much heat is evolved on dissolving gaseous hydrogen chloride, it is essential that the entire flask be kept cold in order to avoid loss by volatilization of the ether. This is best accomplished by rotating and shaking the flask under the rapid stream of water, exposing all parts of the flask. Using a rapid current of hydrogen chloride gas, the mixture may be completely saturated in thirty minutes. Satura tion may be considered complete when no more heat is evolved when the stream of gas is passed through the flask and the flask held in the hand. During the saturation, a Gooch crucible is prepared in the ordinary manner. It is then washed with strong hydrochloric acid followed by washing with the wash liquid which consists of concentrated hydrochloric acid and ether in equal parts saturated with hydrogen chloride g&a'. is now ready for use. The filtering crucible filter. The beaker and filter are then washed with a hot 2 per cent ammonium nitrate solution. Usually there Is a minute quantity of beryllium hydroxide sticking to the sides of even the cleanest beaker. This may be swabbed out with a small fragment of filter paper and added to the precipitate in the funnel. The paper and contents are transferred to a tared crucible, preferrably of platinum, ignited to constant weight and weighed as beryllium oxide. Some authors recommend that the crucible be covered during the final stage of the ignition. If one desires to do this, the cover should be inverted, as it has been found that conduction of combustion gases into the cruci ble under the wide flanges of a regularly placed porcelain cover leads to an increase of weight, probably due to formation of beryllium sulfate. The writer prefers platinum crucibles, as the paper burns away much more rapidly. It has been claimed by some that beryllium oxide attacks platinum very slightly, but in this investigation no evidence of this has been found. Platinum crucibles used almost daily for the. past nine months have suffered only a few milligrams loss in weight. Now to return to the aluminum. The crystalline preci pitate of hydrous aluminum chloride AlClj * 6 H20 contained in the Gooch crucible is dissolved by applying gentle suction and pouring 50 - 60 ml of water containing a few drops of hydro chloric acid through the crucible. The filtrate is rinsed into a 400 ml beaker, about 150 ml water added and brought to 17 a gentle boil. Six normal ammonium hydroxide is added drop-wise with constant stirring in slight excess which may be conveniently determined by carefully smelling the vapors arising from the heated solution. After allowing to stand about five minutes, the mixture is filtered through a nine centimeter number 40 Whatman filter. It is then washed with a hot 2 per cent ammonium nitrate solu tion, ignited to constant weight and weighed as aluminum oxide, AlgO-j. Gooch and Havens (28) determine the aluminum by a novel and unique method. After filtration and washing with the wash liquid, it is dried for one half hour at 150 degrees. . It is then covered with a layer (about one gram) of mercuric oxide and first gently heated and then strongly ignited (Hood ). The reaction is 2 AlCl^ + 3 HgO = AlgO^ + 3 HgClg, but this method was not used in the present investigation. The following results were obtained in trying the method of Gooch and Havens. AlgO-^ taken 0 .1 2 9 0 0 .1 2 9 0 0.0864 0.0864 AlgO-^ found error 0.1286 -0.0004 BeO taken 0.0660 BeO found Error 0.0664 40.0004 0 .1 9 9 5 40.0003 0 .0 6 6 0 0 .0 6 5 7 - 0.0865 4-0.0001 +0.0005 0.0440 0.0440 0.0438 0.0445 -0.0002 +0 .0 0 0 5 0 .0 8 6 9 0.0003 This table shows that the method of Gooch and Havens is reasonably accurate for the separation of aluminum and beryllium; A distinct tendency toward the contamination of the aluminum by beryllium or toward incomplete precipitation of aluminum could 18 not foe detected. TIME AND CONVENIENCE STUDIES OF GOOCH AND HAVENS METHOD COMPARED’WITH OTHER METHODS The total time required for a separation such as just described may be apportioned approximately as follows. The time necessary for the complete saturation of the mixture by hydrogen chloride gas is about thirty minutes. For filtration and washing ten minutes should be ample, as this operation should be conducted quickly in order that too much hydrogen chloride gas shall not escape. Evaporation of the filtrate to determine the beryllium takes the most time, as it must be done slowly so that no loss takes place due to spattering. This takes from thirty to forty-five minutes. Precipitation, filtration and washing require about a iialf hour, while ignition and weighing will take from one and one-half to two hours. The aluminum may be determined during the evaporation of the beryllium filtrate. The total time, approximately, for the Gooch and Havens separation is three and one-half to four hours. The 8-hydroxyquinoline method (45),(50) is very fine for the separation of these two elements. The aluminum is precipi tated quantitatively as the oxyquinolate Al(C^HgON)^ and may be weighed as such or may be ignited to the oxide Al20j. The determination of aluminum does not offer any difficulties and 19 Is rapid, accurate and easily performed. The accurate recovery of beryllium in the filtrate, however, presents some difficul ties. Apparently the beryllium In the filtrate forms a complex of some sort with the excess oxine and incomplete precipitation amounting at times to 20 per cent of the total beryllium oxide will result. The filtrate is treated with a slight excess of dilute ammonium hydroxide to precipitate the beryllium hydroxide in the usual way. use. This is filtered off and reserved for future The filtrate from this precipitate is evaporated nearly to dryness or to a very small volume and treated with very powerful oxidizing agents such as filming nitric and perchloric acids. This is to destroy the organic matter from the excess of oxine. This is tedious and time-consuming. Some workers have reported privately to Doctor Brinton that sometimes the oxine will work perfectly and it is not always necessary to make two berylliuirt precipitations and sub sequent destruction of the organic matter. This erratic b e havior has not yet been satisfactorily explained. The complete time of separation, and determination, is quite long - much more time-consuming than the Gooch and Havens method. In addition to its being so laborious and time- consuming, this method is very expensive, as the oxine is diffi cult to prepare, and extremely difficult to prepare in an analytically pure state. This compound has not yet been 20 produced in quantity and its expense is prohibitive for many laboratories. However, with all its attendant difficulties, this method in experienced hands is comparable in accuracy to the Gooch and Havens, ether method. There is one other method that is in general use for the separation of these two vital elements. This is the sodium bicarbonate method of Parsons and Barnes (32). This method is based on the insolubility of aluminum and ferric hydroxides in a boiling 10 per cent sodium bicarbonate solu tion and on the solubility of beryllium hydroxide in the same medium. In the present investigation a series of experiments comparing this method with the Gooch and Havens method and the oxine method for time and accuracy has been carried out. Following is a detailed description of this method. Equal portions of the standard aluminum and beryllium solutions were pipetted into a 250 ml beaker. The solution was made up to about 100 ml,neutralized as nearly as possible with 6 normal ammonium hydroxide to a slight permanent pre cipitate, and 10 grams of solid sodium bicarbonate were added. The solution must be cold or spattering will occur. .The beaker was kept covered with a watchglass and the solution was brought to boiling as quickly as possible and allowed to boil not more than one minute. COg is not evolved very rapidly until the boiling temperature is reached, if the solution is 21 neutral, but evolution of gas must not be mistaken for boiling. The beaker is now placed in cold water, and when cool, the precipitate is filtered and washed two or three times with hot water. The precipitate is dissolved on the filter with the least possible amount of 6 -normal hydrochloric acid, and the solution is allowed to run into the original beaker. This solution is made up to 100 ml, made neutral by careful addition of dilute ammonium hydroxide, precipitated again with sodium bicarbonate as before, cooled, filtered and washed with hot water, running both filtrates together. A cloudiness will be observed in the combined filtrates after washing, which looks as if aluminum hydroxide is coming through but this need cause no concern as the cloudiness is due to the addition of water to the strong sodium bicarbonate solution. The aluminum hydroxide is again dissolved in 6 -normal hydrochloric acid, precipitated with ammonia and determined in the ordinary way, as it is impossible to wash out all of the sodium bicarbonate from the gelatinous aluminum hydroxide. The filtrate from the double precipitation containing the beryllium in solution, is now carefully acidified with concentrated hydrochloric acid in a covered beaker. Just before neutralization a portion of the beryllium will be observed being thrown down as the hydroxide, but this will immediately dissolve on adding more acid. The solution is 22 then boiled to remove CO2 in order that no ammonium carbonate be formed on addition of ammonium hydroxide in slight excess to precipitate beryllium hydroxide. The precipitate is filtered on a 9 cm, number 4-0 Whatman, or other filter paper of like quality. The filter and contents are then washed with about 25 ml hot 2 per cent ammonium nitrate solution. The precipitate is transferred to a tared crucible and ignited to constant weight and finally weighed as the oxide BeO, AlgO^ BeO Taken Found Error .1 2 9 0 .1351 .1311 .0895 +.0061 +.0021 .1290 .0864 + .0 0 3 1 Taken Found Error .0 6 6 0 .0648 .0653 .0401 -.0012 -.0007 -.0039 .0660 .0440 From this table it is seen that the results are too high for aluminum and too low for beryllium. The same results are brought out in the following table, which are the results of Parsons and Barnes.(32) AlgO^ Taken BeO .1323 Found .1266 .1336 Error +.0014 +.0013 Taken .0818 .0629 Found .0810 .0594 Error -.0008 -.0035 .0 9 0 7 .0 9 4 5 + .0 0 3 8 .0 5 7 3 .0 5 6 3 - .0 0 1 0 .1 2 5 2 The advantage of this method is facility of operation, and it does not require expensive or corrosive chemicals. It is economical in time consumption - the separation just de scribed may be done in about two hours. 23 THE GENERATION OF HYDROGEN CHLORIDE GAS As previously stated, the chief reason for the relative unpopularity of the Gooch and Havens method is the use of hydrogen chloride gas. The method used in this investigation for the genera tion of hydrogen chloride consists of dropping concentrated, technical hydrochloric acid on concentrated, technical sulfuric acid contained in a large flask. The flask is provided with a -well fitting 2 -hole rubber stopper carrying a dropping funnel and a delivery tube bent at right angles. The gas is passed through a wash bottle containing concentrated sulfuric acid. As hydrogen chloride attacks rubber, it is well to have the connections as short as possible. The gas is then conducted by means of glass or rubber tubing to the saturation flask. This consists of an Erlenmeyer flask of the appropriate size fitted with a 2 -hole rubber stopper carrying an inlet tube which extends to the bottom and an outlet tube. To the latter is attached a long rubber tube, which can be extended into the fume hood, or be inserted into the drain pipe of the laboratory sink. By this simple, yet effective, technique, this gas may be comfortably handled. 24 THE KEEPING QUALITIES OF THE WASH LIQUID OF ETHER AND HYDROCHLORIC ACID This is prepared by mixing equal volumes of chemically pure diethyl ether and chemically pure 1 2 -normal hydrochloric acid. The mixture is placed in the absorption flask, put under the tap, and a rapid current of hydrogen chloride passed in for at least 30 minutes. It is then placed in a tight-fitting glass stoppered bottle and kept in a cool place away from direct sunlight. On exposure to direct solar radiation, the liquid slowly turns to a light yellow color. It has been found that it will keep indefinitely if stored in a dark bottle kept in a cool place. THE INVESTIGATION OF A. POSSIBLE SUBSTITUTE FOR DIETHYL ETHER The purpose of this investigation was to find a sub stitute which would be less volatile, less expensive and less lethal. The first solvent tried was methyl alcohol. 10 ml each of the standard solutions of beryllium and aluminum were pipetted into a small Erlenmeyer flask. To this was added 10 ml of 1 2 -normal hydrochloric acid and an equal volume of absolute methyl alcohol. The procedure for this experiment and the following ones were carried out in exactly the same way as were the experiments verifying the method of Gooch and Havens, 25 The following results were obtained: T&keh AI 2 O5 Found .0864 ,0751 +O.OII3 Taken .0440 BeO Found .0748 Error +O .0 3 0 8 It is obvious that methyl ale ohol could not be used. Carbon tetrachloride was next tried, and the following results were obtained: BeO AI2 O3 Taken Error Taken Found Error -.0034 .0440 .0475 +.0035 .0440 .0458 +.0018 Found Commercial CC1|, .0864 .0830^ C. P. CClk .0864 .0845 — -.0019 An attempt to substitute benzene for ether did not lead to encouraging results, as the following data, show. - — „— ... Taken .0864 .0864 .0864 .0864 .0864 .0864 .0864 .0864 A.1p0-2; Found .0796 .... A.— -y*---- »— .. .0 8 3 9 .0851 .0870 .0784 .0 8 3 0 .0855 .0799 Error -.0 0 6 8 -;0Q25 -.0013 +. 0006 -.0080 * .0 0 3 4 - .0 0 0 9 - .0 0 6 5 Taken .0440 .0440 .0440 .0440 .0440 .0440 .0440 .0440 BeO Found .0 5 1 7 .0451 .0386 .0447 .0498 .0451 .0435 .0 5 0 1 Error +.0077 +• 0011 +.0054 +.0007 +.0058 +.0011 - .0 0 0 5 + .0 0 6 1 These erratic results are probably due to the two immiscible phases. 26 Iso-propyl ether Fas then tried. grade and had many impurities. This Fas the technical It Fas purified by distillation through a Vigreux column; only the fraction boiling at 6 7 -6 9 degrees was collected. The results with this solvent are given here: Taken BeO Found Sfrror -.0029 -.0023 .0440 .0440 .0456 .0478 + .0 0 1 6 +.OO35 -.0004 -.0013 -.0005 .0440 .0440 .0440 .0445 .0442 .0449 +.0005 .0 2 3 0 + .0 0 1 0 .0229 +.0009 Taken. a1 2°3 Found Frror Technical .0864 grade .0864 .0835 .0841 .0864 .0864 .0864 .0452 .0432 .0 8 6 0 Purified ether .0851 .0859 .0421 .0422 - .0 0 1 1 - .0 0 1 0 .0 2 2 0 .0 2 2 0 + .0 0 0 2 +.0009 This table shoFs iso-propyl ether to be a reasonably satisfactory substitute for diethyl ether. Di-iso-propyl ether, or 2 -isopropoxypropane-1- has a boiling point of 6 7 -'6 9 degrees, and because of its lower vapor pressure is less subject to fire and explosion hazards than diethyl ether, and it is therefore much safer to use. It does not possess the toxic qualities of diethyl ether, and its cost is much lower. Iso-propyl ether does not form a thick, almost sirupy liquid when mixed with an equal volume of concentrated hydrochloric acid saturated with hydrogen chloride, as is the case with diethyl ether. This 1 Name approved by the International Uninn of Chemistry. 27 together with its lower vapor pressure, greatly facilitates the filtration of the aluminum chloride. THE INVESTIGATION OF A LESS CORROSIVE AND MORE CONVENIENTLY HANDLED REAGENT THAN HYDROGEN CHLORIDE GAS In an attempt to solve this problem, several experiments were tried and are presented here. Into a 250 ml glass stoppered Erlenmeyer flask were pipetted 5 ml each of the standard solutions of beryllium and aluminum chlorides. To this were added 5 ml of 1 2 -normal hydrochloric acid and an equal volume of diethyl ether. The mixture was shaken, being cooled under the tap. Pure anhydrous lithium chloride was added in small portions and shaken under the. tap until the salt was dissolved. The salt was added until no more was dissolved. The mixture was then filtered through a Gooch crucible and washed with a wash liquid consisting of diethyl ether and concentrated hydrochloric acid in equal parts, saturated with anhydrous lithium chloride. The precipitate, which consisted of hydrous aluminum chloride and excess lithium chloride, was dissolved by applying gentle suction and passing-water through the crucible. After diluting to 500-600 ml the aluminum was determined in the usual way. 28 The ethereal filtrate containing the beryllium was carefully evaporated to remove the ether and most of the acid. It was then diluted to about 600 ml and the beryllium was determined in the usual manner Al^Oj (C2h5 )2° (C3H7 )20 Taken .0432 .0432 BeO Found .0337 .0331 Error -.0095 -.0101 Taken .0220 .0220 Found .0311 .0319 Error +.0091 +.0099 The incomplete precipitation of aluminum is due to the inability of the lithium chloride to yield sufficient chloride ions. An experiment using barium chloride was next performed. Anhydrous barium chloride was used in precisely the same way as the lithium chloride. The crystalline precipitate of excess barium chloride and aluminum chloride was filtered through a Gooch crucible and washed with a mixture consisting of ether and concentrated hydrochloric acid in equal parts saturated with anhydrous barium chloride. Gooch crucible was rejected. The residue in the The filtrate was carefully evaporated to a small volume and then transferred to a 2 liter beaker. It was diluted to 1200 ml brought nearly to boiling, and hot 6 -normal sulfuric acid was added dropwise, with constant stirring,in slight excess. The precipitated barium 29 sulfate was allowed to settle for an hour and then filtered through a number 1, 24-cm Whatman filter. The barium sulfate was washed with hot-water and then discarded. The filtrate, about 1.5 liters in volume, was treated for beryllium in the regular way. In this experiment .0452 gram Al20-^ and .0220 grams BeO were taken. The weight of the Be© found was .0652 grams, which is the sum of the two taken. Therefore, no separation took place. Calcium chloride, which is soluble in ether, was con sidered, but the problem of separating 20 or 25 grams of cal cium chloride from a few milligrams of beryllium chloride would be more troublesome than generating hydrogen chloride. THE DETAILED APPLICATION OF THE G00CH AND HAVENS METHOD TO MINERAL AND ORE ANALYSIS Beryllium is found widely diffused in nature in many minerals. The only commercial ore is beryl, the others being very rare. The composition of beryl according to the formula Be^Al2(SiO-j)g is A^O-j 19 per cent, BeO 14 per cent and Si02 67 per cent. Alkalies (NagO, CsgO, Li20) are sometimes present replacing the BeO, from .25 to 5 per cent; also chemically combined water, including which the formula becomes (R2R)0. 6 BeO. 2 A120^. 12 SiOg. 30 Beryl Is most commonly found in granitic rocks, either In druses In the granite or in pegmatite veins. It has also "been noted In tin ores together with topaz and In mica schists. The emeralds found at Muzo, Colombia occur in a bituminous limestone - a most unique type of occurrence. It Is common practice in silicate ore analysis to fuse 1 part of ore with 4 to 10 parts of sodium carbonate. This large amount of sodium salt thus introduced into the analysis together with the long time required for fusion makes this method undesirable. The desirability of using smaller amounts of flux was recognized by Finn and Klekokta (51), who showed that certain ceramic materials, for example, clay, feldspar, and high slumina refractories, can be successfully broken up by sintering at 825-950 degrees C. with a portion of sodium carbonate approximately equal to the weight of the sample. Hoffman (52), of the Bureau of Standards, has fully investigated this fine method of decomposing ores. The fusion (more correctly, sintering) is made In a 75 or pre ferably a 100 ml platinum dish instead of a crucible. The same dish is used for not only the fusion, but for the disintegration of the melt, the evaporation and the ignition of the silica. This not only more than halves the time and labor, but minimizes the danger of loss due to transferring from one vessel to another. 31 Using this method, of decomposition in conjunction with the Gooch and Havens separation, the writer presents a detailed system of analysis of beryl ore. The ore should be ground to a fineness of the order of 100 mesh. ,5 0 0 0 gram of the ore is weighed out and placed in the platinum dish. .5 0 gram of anhydrous sodium carbonate is added and the two mixed very intimately by means of an agate pestle. The mixture is then brushed into the center of the dish and the charge flattened out by means of the agate pestle so that it covers a space about 5 cms in diameter. The mixture is then covered as evenly as possible witfjr: an additional half gram of sodium carbonate. It is now ready to be heated. Instead of the costly electric furnaces of the Bureau of Standards, this writer has found the following method to be perfectly satisfactory. The dish is covered with a large number 14 poreelain crucible cover and placed on a large silica triangle supported, on an iron ring. The height is adjusted so that the bottom of the dish is about 6 or 7 cms above the top of a Fisher blast burner. It is advisable to also use an ordinary Fisher burner at the same time. The dish-is heated for 15 to 20 minutes at a bright red heat and allowed to cool with the cover left on to prevent loss by flying fragments that might be ejected on cooling. 20 ml of 6 -normal hydrochloric acid are added to the cooled melt, the dish is covered, and allowed to digest on the steam hath until disintegration is complete. The action of the acid may he hastened hy occasionally crushing the layer of, insoluble, silica that tends to cover the unattacked ■ portion of the fused mass. It is evaporated to dryness on the steam hath, and kept there*for at least a half-hour to render the silica insoluble. The dish is then removed from the steam hath and allowed to cool. Then 5 ml of concentrated hydrochloric acid are added and the dish is again placed upon the steam hath for about 5 or 10 minutes, hut not longer, as silica is appreciably soluble in hot concentrated hydrochloric acid. The object of warming is to dissolve all basic salts. Then 15 ml of hot water are added and the dish is gently heated over a small flame while stirring. The solution is filtered through a 9 c m , number ifO "Whatman filter, collecting the filtrate in a 600 ml beaker. All the silica is carefully washed onto the filter by hot dilute hydrochloric acid (2:98). About 200 ml is used. The next step is the ignition of the silica. Even though we are not interested in determining the silica, it is necessary to ignite it, treat with hydrofluoric acid, and fuse the residue with pyrosulfate to recover the very appreciable amount of beryllium present in the silica. This cannot be washed out and mu3t be recovered by the process to be described The filter and contents are transferred to a platinum crucible, and the paper is burned off, leaving the white silica The silica, when cool, is moistened with 2 - 3 ml water and 1-2 drops of 18-normal sulfuric acid are added. Then about 10 ml of 4-8 per cent hydrofluoric acid are added and the crucible placed on the steam bath until most of the liquid has evaporated. The remaining liquid is carefully volatilized by heating over a small free flame until 50-^ fumes cease to be evolved. It is then heated to a red heat for a few minutes. There is quite an appreciable residue. This is dissolved by adding about 5 grams of potassium pyrosulfate and gently fusing for about 10 - 15 minutes, or until a clear melt is obtained. The melt is allowed to cool, the crucible is half filled with water containing a few drops of hydrochloric acid and gently boiled. dissolved. This is repeated 2 or 3 times until all is The solution is passed through a filter (a number 1 Whatman is satisfactory) into the main solution. This now contains all the beryllium, aluminum and iron. This solution is now heated almost to boiling and 6 -normal ammonium hydroxide is added dropwise, with constant stirring, in slight excess. The precipitate, which consists of the hydroxides of aluminum, beryllium and iron, is filtered through an 11 cm, number 1 Whatman fitter, and the filtrate is rejected. The filter and contents are then transferred to the 600 ml heaker in which the precipitation took place, and the precipitate is dissolved in the least possible volume of 12-normal hydrochloric acid. About 20 ml is required.. This solution is then transferred to the Erlenmeyer (absorption) flask, and an equal volume of pure isopropyl ether is added. The mixture is saturated with hydrogen chloride as previously described; 20 ml more of the ether being added after the appearance of the fine crystalline precipitate of hydrous aluminum chloride, and again completely saturated. The mixture is filtered through a Gooch crucible, and washed with the wash liquid as previously described. The aluminum may be determined, if desired, exactly as described previously. It is in a pure state; the iron being in the filtrate with the beryllium. The filtrate is evaporated on the steam bath to remove the ether, and then on the hot plate to remove most of the hydrochloric acid. The concentrated solution is then diluted to about 250 ml, heated almost to boiling, and 6 -normal sodium hydroxide is added dropwise with constant stirring until a good excess has been added. This is to separate the iron from the beryl lium; beryllium being amphoteric, its hydroxide is soluble in alkalies, while iron does not show this property. The beaker is allowed to cool somewhat, and the iron precipitate is then filtered through an 11 cm, number 1 "Whatman filter. The filter 35 and contents are then returned to the beaker in which the precipitation was made, the precipitate is dissolved in dilute hydrochloric acid, and a second precipitation is made, the second filtrate being run into the first* This is done because the gelatinous ferric hydroxide occludes some of the soluble sodium beryllate. The filtrate, now containing only the beryllium as Na2 Be02, is carefully acidified by adding concentrated hydro chloric acid dropwise with constant stirring. When neutrality is reached, a portion of the beryllium will be observed to be thrown down, but this immediately dissolves on adding a little more acid. The solution is now on the acid side and is now ready for the final precipitation. The solution is heated almost to boiling and 6 -normal ammonium hydroxide is added drop by drop, with constant stirring, until its odor is very distinctly perceptible. The mixture is boiled for a moment and allowed to cool so that the beaker may be grasped with the hand. It is then filtered through a 9 cm, number 40 Whatman filter and washed with a hot 2 per cent ammonium nitrate solution. The filter and contents are transferred to a tared platinum crucible and ignited to constant weight and weighed as BeO. This system of analysis is not only easily performed, 36 but it is capable of obtaining good results. A sample of ore -was furnished this worker by Doctor Brinton and was subjected to the analysis. The results are given below: Weight sample bottle before Weight sample bottle after Weight of sample Final weight crucible + BeO Final weight crucible Weight BeO Run I Run II 16.6044 16.1043 13*5535 13*0576 .5001 .4959 25.5649 25.5200 25.5845 25*5212 .0449 .0453 L^O^^ilOP) = 8.72$ BeO This compares not too unfavorably with the result of 8.52 per cent BeO, obtained in Doctor Brinton’s laboratory, by the most refined oxyquinolate method. It was thought worth while to try decomposing beryl by passing silicon tetrafuoride over the finely powdered ore at a temperature estimated to be between 650 and 850 degrees C. This has been attempted by Claffin (53) on a commercial scale. This would be of great advantage to the analyst, as only the beryllium would come into solution. The reaction is 2 BejjAig (SiO^ )|? + 5 S i % = 6 BeFg + 2 AlgO^ + 15 SiOg. The silica and aluminum oxide are insoluble in water, while beryllium flucride is extremely soluble. are no alkalies introduced from any fluxes. Also, there 31 The experiment was performed by weighing a small quantity of the same ore as used in the former experiment into the platinum dish and passing in silicon fluoride. The dish was provided with a copper cover through which passed a tube of copper which was connected to the silicon tetrafluoride generator. The method used in this investigation to prepare silicon tetrafluoride consisted of a mixture of fluorspar and white sand in equal proportions heated in a thick flask with thrice its weight of concentrated sulfuric acid. action is: The re 2 CaP2 + 2 H2S0^ + SiOg = 2 CaSO^ + SiF^ + 2 HgO After passing the gas over the heated ore for one hour the dish was removed from the flame. "When cold 25 ml of water were added and the solution was vigorously stirred while it was gently warmed. The liquid was decanted through a filter; the residue was washed and then rejected. The colorless filtrate was evaporated to dryness in the platinum dish and then evaporated to strong fumes with 5 ml of concentrated sulfuric acid to expel the fluoride. It is absolutely necessary to eliminate all fluoride before precipitating beryllium, as it forms a complex, BeF^J and thus escapes complete precipitation as hydroxide. After about half of the sulfuric acid was volatilized, the solution was diluted to about 250 ml. This dilution is necessary, as there must be no large amounts of sulfate present 38 when beryllium is being precipitated, since the hydroxide is rather soluble in the sulfate. The beryllium was then pre cipitated with ammonia in the usual manner, filtered, ignited and weighed. The weight of the BeO, from .5001 grams of ore, t , was only .0 2 3 1 grams, or less than 50 per cent extraction. However, the writer believes that under a slight pressure, say 2 or 3 atmospheres, this process may be more successful. Summary. In this paper an historical account of the discovery of beryllium and a brief survey of the literature concerning the analytical separation of beryllium and aluminum have been presented. The accuracy of separation as originally claimed by the discoverers of the hydrochloric acid-ether method was verified. The technique for the safe and comfortable manipulation of hydrogen chloride gas as used in this process was described. A satisfactory substitute for diethyl ether has been found; namely, isopropyl ether. Attempts were made to sub stitute metallic chlorides for hydrogen chloride gas, but were unsuccessful, as a sufficiently high chloride ion concentration could not be obtained, A new method by Hofftnan (5 2 ) was cited for decomposition of silicates using only 2 parts of sodium carbonate to one part of sample, and has been found perfectly reliable. 39 This is applied to beryllium ore analysis, in conjunc tion with the modified Gooch and Havens method. The Parsons and Barnes sodium bicarbonate method was compared for accuracy and convenience of time and labor. A method of decomposing beryl by passing silicon tetrafluoride over the heated ore is described, but was not found satisfactory under analytical conditions. BIBLIOGRAPHY ' Mellor, "Comprehensive Treatise on Inorganic and Theoretical Chemistry,".Vol. 4, Longmans, Green and Co. London, 192 3 , p. 204. . Klaproth, "Analytical Essays Towards Promoting the Chemical Knowledge of Mineral Substances," Cadell and Davies London, 1801, p. 325* Vauquelin, "Analyse de 1 ‘Aigue marine, ou Beril; et decouverte d ’une terre nouvelle dans cette pierre/' Ann. Chim. Phys. 1 , 26, 155 (May(30 Floreal), 1 7 9 8 ). Jagnaux, "Histoire de la Chimie," ref. (2) Vol. 2, p. I6 9 . Vauquelin, "Analyse de 1 1Emeraude du Perou " Ann. Chim. Phys. 1 , .26, 259 (June(j50Prairial), 1798). Gmelin, Johann Friedrich, "Analyse du beril de Nertschinsk en Sib^rie, et examen de quelques caracteresqui distinguent la glucine qu 1il contient," ibid, 1 , 4_,27 (0 ct . ( 3 0 Vendema ir e),18 03 Gmelin, J. F., Ann. Physik, 5 0 , Schaffgo.tsch, Ibid., 175. 50, I8 3 . Scheerer, Ibid., 5 6 , 479. Berthier, Ann. Chim. 3 , Phys. 7 ,7 4 . Bottinger, Ann, fxl, 397. 3 Rivot, Ann. Chim. Phys., Debray, Ibid., 2 , Joy, Am. J. Sci., Gibbs, Ibid., , 3 0 , 183 (1 8 5 0 ). 44, 1. 2 , 36, 83. 2 ,3 7 .,• 346 (1864) Pollock, Trans. Roy. Dublin Soc., Cooke, Am. J. Sci. 2 , 42_, Havens and Way, Ibid., 4 2 , 1904, 139. 78. , 8 , 217. Rossler, Z. anal. Chem. 17, 148. 41 (20) Vincent, Bull. Soc. Chim. 33» 157• (21) Renz, Ber., 36, 2751. (22) Zimmermann, Z. Anorg. Chem., 1 5 , 286 (1 8 8 7 ) (2 3 ) Schier, Chem. Ztg., 16, 420 (1892) (24) Atkinson and. Smith, J. Am. Chem. Soc., IJ_, 688 (I8 9 5 ) (2 5 ) Burgass, Z. Angew. Chemie, 596 (1 8 9 6 ) (26) Leheau, P., Compt. Rend. 121, 641. (27) Hart, J. Am. Chem. Soc. 1 7 , 604. (28) Gooch and Havens, Am. J. Sci., 4 , 4_, 111 (1897) (2 9 ) Haber and Von Oordt, Z. anorg, Chem., 40, 465 (1904) (30) Myers, J. Am. Chem. Soc., 26_, 1124. (3 1 ) Parsons, Charles L., and Robinson, Ibid., 28, (3 2 ) 555 (1906) Parsons, Charles L. and S. E. Barnes, Ibid., 28, 1589 (1906 (3 3 ) Glassman, B. Ber., 39 ^, 3366 (I9 0 6 ) (34) Wunder, M. and Chil&dze, N. Ann. Chim. Anal. 16_, ( 35 ) 205 (1911) Browning, Philip and Kuzirlan, S.B., Orig. Com. 8 th Intern. Congr. Appl. Chem., 1_, 87 (1 9 1 2 ) (3 6 ) Wunder, M. and ¥enger, P. A. Zanal. Chem., 51, 470 (1912). (3 7 ) Bleyer, B. and Moormann, A., ^The Volumetric determination of beryllium. 11 Ibid., 5 1 , 3 6 0 ,(1 9 1 2 ). (3 8 ) ¥under and ¥enger, Ann. Chim. anal, 1 7 , 3 63 (1 9 1 3 ). (39) Kling, A., and Gelin, E. Bull. soc. chim., 15_, 205, ( 1914) (40) Copaux, H. ^Determination du glucine en beril,*' Comptes rendus, 168, 612 (1 9 1 9 ). (41) Britton, Hubert Thomas Stanley. ^Electrometric studies of the precipitation of hydroxides.M J. Chem; Soc., 127, 2120 (1925). Fischer, Hellmut, Wiss. 99 (1 9 2 6 ). Veroff. Siemens Konzern, 5, Moser, L. and Niessner, M. Monatah, 48, 113 (1927) Kolthoff, I. M. J. Am. Chem. Soc., 50, 393 (1928) Kolthoff, I. M., and Sandell, Ernest B. Ibid., 5 0 , 1900 (1928 ) jflek, A. and Kota, J. Collection Czechoslov. Chem. Comm. 3., 336 (1931) / /v Gaspar y Apnal, Teofilo. Anales sociedad espanola f-Lsica quimica. .3 2 , 868 (1 9 3 4 ) Pache, Erich, Chem. Ztg., 61, 880 (1937). Nichols, M. L. and Schempf, John M. Anal. Ed. ILL, 278 (1939). Ind. Eng. Chem., Knowles, H.B., Journal of Research, U.S. National Bureau of Standards 15, 87 (1936). Finn, A. N. and Klekokta, J. F., Ibid., 4, 809 (1930) E P 180. ~ Hofftnan, James I., Ibid., 25_, 379, Sept. 1940. Claffin, Harry C., U.S.Pat. 2, 081, 984.